From 37f812763c15650c2780bb3b4a908a8801520194 Mon Sep 17 00:00:00 2001
From: Morten Hjorth-Jensen <morten.hjorth-jensen@fys.uio.no>
Date: Sat, 9 Oct 2021 19:16:35 +0200
Subject: [PATCH] setting up week41

---
 doc/pub/week41/html/week41-bs.html           | 320 ++++++++-------
 doc/pub/week41/html/week41-reveal.html       | 205 +++++-----
 doc/pub/week41/html/week41-solarized.html    | 387 +++++++++++--------
 doc/pub/week41/html/week41.html              | 383 ++++++++++--------
 doc/pub/week41/ipynb/ipynb-week41-src.tar.gz | Bin 87344 -> 87692 bytes
 doc/pub/week41/ipynb/week41.ipynb            | 230 ++++++-----
 doc/src/week41/week41.do.txt                 |  10 +-
 7 files changed, 876 insertions(+), 659 deletions(-)
diff --git a/doc/pub/week41/html/week41-bs.html b/doc/pub/week41/html/week41-bs.html
index 54e48bba2..abd36d462 100644
--- a/doc/pub/week41/html/week41-bs.html
+++ b/doc/pub/week41/html/week41-bs.html
@@ -1,15 +1,15 @@
 <!--
 Automatically generated HTML file from DocOnce source
-(https://github.com/hplgit/doconce/)
+(https://github.com/doconce/doconce/)
 -->
 <html>
 <head>
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
-<meta name="generator" content="DocOnce: https://github.com/hplgit/doconce/" />
+<meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 41 Tensor flow and Deep Learning, Convolutional Neural Networks">
+<meta name="description" content="Week 41 Constructing a Neural Network code, Tensor flow and start Convolutional Neural Networks">
 
-<title>Week 41 Tensor flow and Deep Learning, Convolutional Neural Networks</title>
+<title>Week 41 Constructing a Neural Network code, Tensor flow and start Convolutional Neural Networks</title>
 
 <!-- Bootstrap style: bootstrap -->
 <link href="https://netdna.bootstrapcdn.com/bootstrap/3.1.1/css/bootstrap.min.css" rel="stylesheet">
@@ -41,106 +41,160 @@ Automatically generated HTML file from DocOnce source
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Plan for week 41', 2, None, '___sec0'),
-              ('Setting up the Back propagation algorithm', 2, None, '___sec1'),
+ 'sections': [('Plan for week 41', 2, None, 'plan-for-week-41'),
+              ('Setting up the Back propagation algorithm',
+               2,
+               None,
+               'setting-up-the-back-propagation-algorithm'),
               ('Setting up a Multi-layer perceptron model for classification',
                2,
                None,
-               '___sec2'),
-              ('Defining the cost function', 2, None, '___sec3'),
-              ('Example: binary classification problem', 2, None, '___sec4'),
-              ('The Softmax function', 2, None, '___sec5'),
+               'setting-up-a-multi-layer-perceptron-model-for-classification'),
+              ('Defining the cost function',
+               2,
+               None,
+               'defining-the-cost-function'),
+              ('Example: binary classification problem',
+               2,
+               None,
+               'example-binary-classification-problem'),
+              ('The Softmax function', 2, None, 'the-softmax-function'),
               ('Developing a code for doing neural networks with back '
                'propagation',
                2,
                None,
-               '___sec6'),
-              ('Collect and pre-process data', 2, None, '___sec7'),
-              ('Train and test datasets', 2, None, '___sec8'),
-              ('Define model and architecture', 2, None, '___sec9'),
-              ('Layers', 2, None, '___sec10'),
-              ('Weights and biases', 2, None, '___sec11'),
-              ('Feed-forward pass', 2, None, '___sec12'),
-              ('Matrix multiplications', 2, None, '___sec13'),
-              ('Choose cost function and optimizer', 2, None, '___sec14'),
-              ('Optimizing the cost function', 2, None, '___sec15'),
-              ('Regularization', 2, None, '___sec16'),
-              ('Matrix  multiplication', 2, None, '___sec17'),
-              ('Improving performance', 2, None, '___sec18'),
-              ('Full object-oriented implementation', 2, None, '___sec19'),
-              ('Evaluate model performance on test data', 2, None, '___sec20'),
-              ('Adjust hyperparameters', 2, None, '___sec21'),
-              ('Visualization', 2, None, '___sec22'),
-              ('scikit-learn implementation', 2, None, '___sec23'),
-              ('Visualization', 2, None, '___sec24'),
+               'developing-a-code-for-doing-neural-networks-with-back-propagation'),
+              ('Collect and pre-process data',
+               2,
+               None,
+               'collect-and-pre-process-data'),
+              ('Train and test datasets', 2, None, 'train-and-test-datasets'),
+              ('Define model and architecture',
+               2,
+               None,
+               'define-model-and-architecture'),
+              ('Layers', 2, None, 'layers'),
+              ('Weights and biases', 2, None, 'weights-and-biases'),
+              ('Feed-forward pass', 2, None, 'feed-forward-pass'),
+              ('Matrix multiplications', 2, None, 'matrix-multiplications'),
+              ('Choose cost function and optimizer',
+               2,
+               None,
+               'choose-cost-function-and-optimizer'),
+              ('Optimizing the cost function',
+               2,
+               None,
+               'optimizing-the-cost-function'),
+              ('Regularization', 2, None, 'regularization'),
+              ('Matrix  multiplication', 2, None, 'matrix-multiplication'),
+              ('Improving performance', 2, None, 'improving-performance'),
+              ('Full object-oriented implementation',
+               2,
+               None,
+               'full-object-oriented-implementation'),
+              ('Evaluate model performance on test data',
+               2,
+               None,
+               'evaluate-model-performance-on-test-data'),
+              ('Adjust hyperparameters', 2, None, 'adjust-hyperparameters'),
+              ('Visualization', 2, None, 'visualization'),
+              ('scikit-learn implementation',
+               2,
+               None,
+               'scikit-learn-implementation'),
+              ('Visualization', 2, None, 'visualization'),
               ('Building neural networks in Tensorflow and Keras',
                2,
                None,
-               '___sec25'),
-              ('Tensorflow', 2, None, '___sec26'),
-              ('Using Keras', 2, None, '___sec27'),
-              ('Collect and pre-process data', 2, None, '___sec28'),
-              ('The Breast Cancer Data, now with Keras', 2, None, '___sec29'),
+               'building-neural-networks-in-tensorflow-and-keras'),
+              ('Tensorflow', 2, None, 'tensorflow'),
+              ('Using Keras', 2, None, 'using-keras'),
+              ('Collect and pre-process data',
+               2,
+               None,
+               'collect-and-pre-process-data'),
+              ('The Breast Cancer Data, now with Keras',
+               2,
+               None,
+               'the-breast-cancer-data-now-with-keras'),
               ('Fine-tuning neural network hyperparameters',
                2,
                None,
-               '___sec30'),
-              ('Hidden layers', 2, None, '___sec31'),
-              ('Which activation function should I use?', 2, None, '___sec32'),
+               'fine-tuning-neural-network-hyperparameters'),
+              ('Hidden layers', 2, None, 'hidden-layers'),
+              ('Which activation function should I use?',
+               2,
+               None,
+               'which-activation-function-should-i-use'),
               ('Is the Logistic activation function (Sigmoid)  our choice?',
                2,
                None,
-               '___sec33'),
-              ('The derivative of the Logistic funtion', 2, None, '___sec34'),
-              ('The RELU function family', 2, None, '___sec35'),
-              ('Which activation function should we use?', 2, None, '___sec36'),
+               'is-the-logistic-activation-function-sigmoid-our-choice'),
+              ('The derivative of the Logistic funtion',
+               2,
+               None,
+               'the-derivative-of-the-logistic-funtion'),
+              ('The RELU function family', 2, None, 'the-relu-function-family'),
+              ('Which activation function should we use?',
+               2,
+               None,
+               'which-activation-function-should-we-use'),
               ('More on activation functions, output layers',
                2,
                None,
-               '___sec37'),
-              ('Batch Normalization', 2, None, '___sec38'),
-              ('Dropout', 2, None, '___sec39'),
-              ('Gradient Clipping', 2, None, '___sec40'),
-              ('A very nice website on Neural Networks', 2, None, '___sec41'),
+               'more-on-activation-functions-output-layers'),
+              ('Batch Normalization', 2, None, 'batch-normalization'),
+              ('Dropout', 2, None, 'dropout'),
+              ('Gradient Clipping', 2, None, 'gradient-clipping'),
+              ('A very nice website on Neural Networks',
+               2,
+               None,
+               'a-very-nice-website-on-neural-networks'),
               ('A top-down perspective on Neural networks',
                2,
                None,
-               '___sec42'),
+               'a-top-down-perspective-on-neural-networks'),
               ('Limitations of supervised learning with deep networks',
                2,
                None,
-               '___sec43'),
+               'limitations-of-supervised-learning-with-deep-networks'),
               ('Convolutional Neural Networks (recognizing images)',
                2,
                None,
-               '___sec44'),
+               'convolutional-neural-networks-recognizing-images'),
               ('Regular NNs don’t scale well to full images',
                2,
                None,
-               '___sec45'),
-              ('3D volumes of neurons', 2, None, '___sec46'),
-              ('Layers used to build CNNs', 2, None, '___sec47'),
-              ('Transforming images', 2, None, '___sec48'),
-              ('CNNs in brief', 2, None, '___sec49'),
+               'regular-nns-don-t-scale-well-to-full-images'),
+              ('3D volumes of neurons', 2, None, '3d-volumes-of-neurons'),
+              ('Layers used to build CNNs',
+               2,
+               None,
+               'layers-used-to-build-cnns'),
+              ('Transforming images', 2, None, 'transforming-images'),
+              ('CNNs in brief', 2, None, 'cnns-in-brief'),
               ('CNNs in more detail, building convolutional neural networks in '
                'Tensorflow and Keras',
                2,
                None,
-               '___sec50'),
-              ('Setting it up', 2, None, '___sec51'),
-              ('The MNIST dataset again', 2, None, '___sec52'),
-              ('Strong correlations', 2, None, '___sec53'),
-              ('Layers of a CNN', 2, None, '___sec54'),
-              ('Systematic reduction', 2, None, '___sec55'),
+               'cnns-in-more-detail-building-convolutional-neural-networks-in-tensorflow-and-keras'),
+              ('Setting it up', 2, None, 'setting-it-up'),
+              ('The MNIST dataset again', 2, None, 'the-mnist-dataset-again'),
+              ('Strong correlations', 2, None, 'strong-correlations'),
+              ('Layers of a CNN', 2, None, 'layers-of-a-cnn'),
+              ('Systematic reduction', 2, None, 'systematic-reduction'),
               ('Prerequisites: Collect and pre-process data',
                2,
                None,
-               '___sec56'),
-              ('Importing Keras and Tensorflow', 2, None, '___sec57'),
-              ('Running with Keras', 2, None, '___sec58'),
-              ('Final part', 2, None, '___sec59'),
-              ('Final visualization', 2, None, '___sec60'),
-              ('Fun links', 2, None, '___sec61')]}
+               'prerequisites-collect-and-pre-process-data'),
+              ('Importing Keras and Tensorflow',
+               2,
+               None,
+               'importing-keras-and-tensorflow'),
+              ('Running with Keras', 2, None, 'running-with-keras'),
+              ('Final part', 2, None, 'final-part'),
+              ('Final visualization', 2, None, 'final-visualization'),
+              ('Fun links', 2, None, 'fun-links')]}
 end of tocinfo -->
 
 <body>
@@ -170,7 +224,7 @@ MathJax.Hub.Config({
       <span class="icon-bar"></span>
       <span class="icon-bar"></span>
     </button>
-    <a class="navbar-brand" href="week41-bs.html">Week 41 Tensor flow and Deep Learning, Convolutional Neural Networks</a>
+    <a class="navbar-brand" href="week41-bs.html">Week 41 Constructing a Neural Network code, Tensor flow and start Convolutional Neural Networks</a>
   </div>
 
   <div class="navbar-collapse collapse navbar-responsive-collapse">
@@ -178,68 +232,68 @@ MathJax.Hub.Config({
       <li class="dropdown">
         <a href="#" class="dropdown-toggle" data-toggle="dropdown">Contents <b class="caret"></b></a>
         <ul class="dropdown-menu">
-     <!-- navigation toc: --> <li><a href="._week41-bs001.html#___sec0" style="font-size: 80%;">Plan for week 41</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs002.html#___sec1" style="font-size: 80%;">Setting up the Back propagation algorithm</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs003.html#___sec2" style="font-size: 80%;">Setting up a Multi-layer perceptron model for classification</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs004.html#___sec3" style="font-size: 80%;">Defining the cost function</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs005.html#___sec4" style="font-size: 80%;">Example: binary classification problem</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs006.html#___sec5" style="font-size: 80%;">The Softmax function</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs007.html#___sec6" style="font-size: 80%;">Developing a code for doing neural networks with back propagation</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs008.html#___sec7" style="font-size: 80%;">Collect and pre-process data</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs009.html#___sec8" style="font-size: 80%;">Train and test datasets</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs010.html#___sec9" style="font-size: 80%;">Define model and architecture</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs011.html#___sec10" style="font-size: 80%;">Layers</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs012.html#___sec11" style="font-size: 80%;">Weights and biases</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs013.html#___sec12" style="font-size: 80%;">Feed-forward pass</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs014.html#___sec13" style="font-size: 80%;">Matrix multiplications</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs015.html#___sec14" style="font-size: 80%;">Choose cost function and optimizer</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs016.html#___sec15" style="font-size: 80%;">Optimizing the cost function</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs017.html#___sec16" style="font-size: 80%;">Regularization</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs018.html#___sec17" style="font-size: 80%;">Matrix  multiplication</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs019.html#___sec18" style="font-size: 80%;">Improving performance</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs020.html#___sec19" style="font-size: 80%;">Full object-oriented implementation</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs021.html#___sec20" style="font-size: 80%;">Evaluate model performance on test data</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs022.html#___sec21" style="font-size: 80%;">Adjust hyperparameters</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs023.html#___sec22" style="font-size: 80%;">Visualization</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs024.html#___sec23" style="font-size: 80%;">scikit-learn implementation</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs025.html#___sec24" style="font-size: 80%;">Visualization</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs026.html#___sec25" style="font-size: 80%;">Building neural networks in Tensorflow and Keras</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs027.html#___sec26" style="font-size: 80%;">Tensorflow</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs028.html#___sec27" style="font-size: 80%;">Using Keras</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs029.html#___sec28" style="font-size: 80%;">Collect and pre-process data</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs030.html#___sec29" style="font-size: 80%;">The Breast Cancer Data, now with Keras</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs031.html#___sec30" style="font-size: 80%;">Fine-tuning neural network hyperparameters</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs032.html#___sec31" style="font-size: 80%;">Hidden layers</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs033.html#___sec32" style="font-size: 80%;">Which activation function should I use?</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs034.html#___sec33" style="font-size: 80%;">Is the Logistic activation function (Sigmoid)  our choice?</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs035.html#___sec34" style="font-size: 80%;">The derivative of the Logistic funtion</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs036.html#___sec35" style="font-size: 80%;">The RELU function family</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs037.html#___sec36" style="font-size: 80%;">Which activation function should we use?</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs038.html#___sec37" style="font-size: 80%;">More on activation functions, output layers</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs039.html#___sec38" style="font-size: 80%;">Batch Normalization</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs040.html#___sec39" style="font-size: 80%;">Dropout</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs041.html#___sec40" style="font-size: 80%;">Gradient Clipping</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs042.html#___sec41" style="font-size: 80%;">A very nice website on Neural Networks</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs043.html#___sec42" style="font-size: 80%;">A top-down perspective on Neural networks</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs044.html#___sec43" style="font-size: 80%;">Limitations of supervised learning with deep networks</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs045.html#___sec44" style="font-size: 80%;">Convolutional Neural Networks (recognizing images)</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs046.html#___sec45" style="font-size: 80%;">Regular NNs don’t scale well to full images</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs047.html#___sec46" style="font-size: 80%;">3D volumes of neurons</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs048.html#___sec47" style="font-size: 80%;">Layers used to build CNNs</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs049.html#___sec48" style="font-size: 80%;">Transforming images</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs050.html#___sec49" style="font-size: 80%;">CNNs in brief</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs051.html#___sec50" style="font-size: 80%;">CNNs in more detail, building convolutional neural networks in Tensorflow and Keras</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs052.html#___sec51" style="font-size: 80%;">Setting it up</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs053.html#___sec52" style="font-size: 80%;">The MNIST dataset again</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs054.html#___sec53" style="font-size: 80%;">Strong correlations</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs055.html#___sec54" style="font-size: 80%;">Layers of a CNN</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs056.html#___sec55" style="font-size: 80%;">Systematic reduction</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs057.html#___sec56" style="font-size: 80%;">Prerequisites: Collect and pre-process data</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs058.html#___sec57" style="font-size: 80%;">Importing Keras and Tensorflow</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs059.html#___sec58" style="font-size: 80%;">Running with Keras</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs060.html#___sec59" style="font-size: 80%;">Final part</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs061.html#___sec60" style="font-size: 80%;">Final visualization</a></li>
-     <!-- navigation toc: --> <li><a href="._week41-bs062.html#___sec61" style="font-size: 80%;">Fun links</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs001.html#plan-for-week-41" style="font-size: 80%;">Plan for week 41</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs002.html#setting-up-the-back-propagation-algorithm" style="font-size: 80%;">Setting up the Back propagation algorithm</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs003.html#setting-up-a-multi-layer-perceptron-model-for-classification" style="font-size: 80%;">Setting up a Multi-layer perceptron model for classification</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs004.html#defining-the-cost-function" style="font-size: 80%;">Defining the cost function</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs005.html#example-binary-classification-problem" style="font-size: 80%;">Example: binary classification problem</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs006.html#the-softmax-function" style="font-size: 80%;">The Softmax function</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs007.html#developing-a-code-for-doing-neural-networks-with-back-propagation" style="font-size: 80%;">Developing a code for doing neural networks with back propagation</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs029.html#collect-and-pre-process-data" style="font-size: 80%;">Collect and pre-process data</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs009.html#train-and-test-datasets" style="font-size: 80%;">Train and test datasets</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs010.html#define-model-and-architecture" style="font-size: 80%;">Define model and architecture</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs011.html#layers" style="font-size: 80%;">Layers</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs012.html#weights-and-biases" style="font-size: 80%;">Weights and biases</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs013.html#feed-forward-pass" style="font-size: 80%;">Feed-forward pass</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs014.html#matrix-multiplications" style="font-size: 80%;">Matrix multiplications</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs015.html#choose-cost-function-and-optimizer" style="font-size: 80%;">Choose cost function and optimizer</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs016.html#optimizing-the-cost-function" style="font-size: 80%;">Optimizing the cost function</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs017.html#regularization" style="font-size: 80%;">Regularization</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs018.html#matrix-multiplication" style="font-size: 80%;">Matrix  multiplication</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs019.html#improving-performance" style="font-size: 80%;">Improving performance</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs020.html#full-object-oriented-implementation" style="font-size: 80%;">Full object-oriented implementation</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs021.html#evaluate-model-performance-on-test-data" style="font-size: 80%;">Evaluate model performance on test data</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs022.html#adjust-hyperparameters" style="font-size: 80%;">Adjust hyperparameters</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs025.html#visualization" style="font-size: 80%;">Visualization</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs024.html#scikit-learn-implementation" style="font-size: 80%;">scikit-learn implementation</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs025.html#visualization" style="font-size: 80%;">Visualization</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs026.html#building-neural-networks-in-tensorflow-and-keras" style="font-size: 80%;">Building neural networks in Tensorflow and Keras</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs027.html#tensorflow" style="font-size: 80%;">Tensorflow</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs028.html#using-keras" style="font-size: 80%;">Using Keras</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs029.html#collect-and-pre-process-data" style="font-size: 80%;">Collect and pre-process data</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs030.html#the-breast-cancer-data-now-with-keras" style="font-size: 80%;">The Breast Cancer Data, now with Keras</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs031.html#fine-tuning-neural-network-hyperparameters" style="font-size: 80%;">Fine-tuning neural network hyperparameters</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs032.html#hidden-layers" style="font-size: 80%;">Hidden layers</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs033.html#which-activation-function-should-i-use" style="font-size: 80%;">Which activation function should I use?</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs034.html#is-the-logistic-activation-function-sigmoid-our-choice" style="font-size: 80%;">Is the Logistic activation function (Sigmoid)  our choice?</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs035.html#the-derivative-of-the-logistic-funtion" style="font-size: 80%;">The derivative of the Logistic funtion</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs036.html#the-relu-function-family" style="font-size: 80%;">The RELU function family</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs037.html#which-activation-function-should-we-use" style="font-size: 80%;">Which activation function should we use?</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs038.html#more-on-activation-functions-output-layers" style="font-size: 80%;">More on activation functions, output layers</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs039.html#batch-normalization" style="font-size: 80%;">Batch Normalization</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs040.html#dropout" style="font-size: 80%;">Dropout</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs041.html#gradient-clipping" style="font-size: 80%;">Gradient Clipping</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs042.html#a-very-nice-website-on-neural-networks" style="font-size: 80%;">A very nice website on Neural Networks</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs043.html#a-top-down-perspective-on-neural-networks" style="font-size: 80%;">A top-down perspective on Neural networks</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs044.html#limitations-of-supervised-learning-with-deep-networks" style="font-size: 80%;">Limitations of supervised learning with deep networks</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs045.html#convolutional-neural-networks-recognizing-images" style="font-size: 80%;">Convolutional Neural Networks (recognizing images)</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs046.html#regular-nns-don-t-scale-well-to-full-images" style="font-size: 80%;">Regular NNs don’t scale well to full images</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs047.html#3d-volumes-of-neurons" style="font-size: 80%;">3D volumes of neurons</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs048.html#layers-used-to-build-cnns" style="font-size: 80%;">Layers used to build CNNs</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs049.html#transforming-images" style="font-size: 80%;">Transforming images</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs050.html#cnns-in-brief" style="font-size: 80%;">CNNs in brief</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs051.html#cnns-in-more-detail-building-convolutional-neural-networks-in-tensorflow-and-keras" style="font-size: 80%;">CNNs in more detail, building convolutional neural networks in Tensorflow and Keras</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs052.html#setting-it-up" style="font-size: 80%;">Setting it up</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs053.html#the-mnist-dataset-again" style="font-size: 80%;">The MNIST dataset again</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs054.html#strong-correlations" style="font-size: 80%;">Strong correlations</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs055.html#layers-of-a-cnn" style="font-size: 80%;">Layers of a CNN</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs056.html#systematic-reduction" style="font-size: 80%;">Systematic reduction</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs057.html#prerequisites-collect-and-pre-process-data" style="font-size: 80%;">Prerequisites: Collect and pre-process data</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs058.html#importing-keras-and-tensorflow" style="font-size: 80%;">Importing Keras and Tensorflow</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs059.html#running-with-keras" style="font-size: 80%;">Running with Keras</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs060.html#final-part" style="font-size: 80%;">Final part</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs061.html#final-visualization" style="font-size: 80%;">Final visualization</a></li>
+     <!-- navigation toc: --> <li><a href="._week41-bs062.html#fun-links" style="font-size: 80%;">Fun links</a></li>
 
         </ul>
       </li>
@@ -258,7 +312,7 @@ MathJax.Hub.Config({
 
 
 <div class="jumbotron">
-<center><h1>Week 41 Tensor flow and Deep Learning, Convolutional Neural Networks</h1></center>  <!-- document title -->
+<center><h1>Week 41 Constructing a Neural Network code, Tensor flow and start Convolutional Neural Networks</h1></center>  <!-- document title -->
 
 <p>
 <!-- author(s): Morten Hjorth-Jensen -->
@@ -274,7 +328,7 @@ MathJax.Hub.Config({
 <center>[2] <b>Department of Physics and Astronomy and National Superconducting Cyclotron Laboratory, Michigan State University</b></center>
 <br>
 <p>
-<center><h4>Oct 10, 2020</h4></center> <!-- date -->
+<center><h4>Oct 9, 2021</h4></center> <!-- date -->
 <br>
 <p>
 
@@ -310,13 +364,13 @@ MathJax.Hub.Config({
 
 <!-- Bootstrap footer
 <footer>
-<a href="http://..."><img width="250" align=right src="http://..."></a>
+<a href="https://..."><img width="250" align=right src="https://..."></a>
 </footer>
 -->
 
 
 <center style="font-size:80%">
-<!-- copyright --> &copy; 1999-2020, Morten Hjorth-Jensen. Released under CC Attribution-NonCommercial 4.0 license
+<!-- copyright --> &copy; 1999-2021, Morten Hjorth-Jensen. Released under CC Attribution-NonCommercial 4.0 license
 </center>
 
 
diff --git a/doc/pub/week41/html/week41-reveal.html b/doc/pub/week41/html/week41-reveal.html
index d02dd41e1..30cbb8176 100644
--- a/doc/pub/week41/html/week41-reveal.html
+++ b/doc/pub/week41/html/week41-reveal.html
@@ -1,11 +1,11 @@
 <!DOCTYPE html>
 
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
-<meta name="generator" content="DocOnce: https://github.com/hplgit/doconce/" />
+<meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 41 Tensor flow and Deep Learning, Convolutional Neural Networks">
+<meta name="description" content="Week 41 Constructing a Neural Network code, Tensor flow and start Convolutional Neural Networks">
 
-<title>Week 41 Tensor flow and Deep Learning, Convolutional Neural Networks</title>
+<title>Week 41 Constructing a Neural Network code, Tensor flow and start Convolutional Neural Networks</title>
 
 
 
@@ -13,7 +13,7 @@
 
 
 
-<!-- reveal.js: http://lab.hakim.se/reveal-js/ -->
+<!-- reveal.js: https://lab.hakim.se/reveal-js/ -->
 
 <meta name="viewport" content="width=device-width, initial-scale=1.0, maximum-scale=1.0, user-scalable=no">
 
@@ -76,10 +76,10 @@ document.getElementsByTagName( 'head' )[0].appendChild( link );
      .reveal .alert-block > p, .alert-block > ul {margin-bottom:1em}
      /*.reveal .alert li {margin-top: 1em}*/
      .reveal .alert-block p+p {margin-top:5px}
-     /*.reveal .alert-notice { background-image: url(http://hplgit.github.io/doconce/bundled/html_images/small_gray_notice.png); }
-     .reveal .alert-summary  { background-image:url(http://hplgit.github.io/doconce/bundled/html_images/small_gray_summary.png); }
-     .reveal .alert-warning { background-image: url(http://hplgit.github.io/doconce/bundled/html_images/small_gray_warning.png); }
-     .reveal .alert-question {background-image:url(http://hplgit.github.io/doconce/bundled/html_images/small_gray_question.png); } */
+     /*.reveal .alert-notice { background-image: url(https://hplgit.github.io/doconce/bundled/html_images/small_gray_notice.png); }
+     .reveal .alert-summary  { background-image:url(https://hplgit.github.io/doconce/bundled/html_images/small_gray_summary.png); }
+     .reveal .alert-warning { background-image: url(https://hplgit.github.io/doconce/bundled/html_images/small_gray_warning.png); }
+     .reveal .alert-question {background-image:url(https://hplgit.github.io/doconce/bundled/html_images/small_gray_question.png); } */
 
 </style>
 
@@ -132,7 +132,7 @@ MathJax.Hub.Config({
 
 
 
-<center><h1 style="text-align: center;">Week 41 Tensor flow and Deep Learning, Convolutional Neural Networks</h1></center>  <!-- document title -->
+<center><h1 style="text-align: center;">Week 41 Constructing a Neural Network code, Tensor flow and start Convolutional Neural Networks</h1></center>  <!-- document title -->
 
 <p>
 <!-- author(s): Morten Hjorth-Jensen -->
@@ -148,32 +148,33 @@ MathJax.Hub.Config({
 <center>[2] <b>Department of Physics and Astronomy and National Superconducting Cyclotron Laboratory, Michigan State University</b></center>
 <br>
 <p>&nbsp;<br>
-<center><h4>Oct 10, 2020</h4></center> <!-- date -->
+<center><h4>Oct 9, 2021</h4></center> <!-- date -->
 <br>
 <p>
 
 <center style="font-size:80%">
-<!-- copyright --> &copy; 1999-2020, Morten Hjorth-Jensen. Released under CC Attribution-NonCommercial 4.0 license
+<!-- copyright --> &copy; 1999-2021, Morten Hjorth-Jensen. Released under CC Attribution-NonCommercial 4.0 license
 </center>
 </section>
 
 
 <section>
-<h2 id="___sec0">Plan for week 41 </h2>
+<h2 id="plan-for-week-41">Plan for week 41 </h2>
 
 <ul>
-<p><li> Thursday: Building our own Feed-forward Neural Network. <a href="https://www.uio.no/studier/emner/matnat/fys/FYS-STK4155/h20/forelesningsvideoer/LectureOctober8.mp4?vrtx=view-as-webpage" target="_blank">Video of Lecture</a></li>
-<p><li> Friday: Playing around with our own Feed-forward Neural Network and introduction to TensorFlow. Start convolutional Neural Networks (CNN). <a href="https://www.uio.no/studier/emner/matnat/fys/FYS-STK4155/h20/forelesningsvideoer/LectureOcotber9.mp4?vrtx=view-as-webpage" target="_blank">Video of Lecture</a></li>
+<p><li> Thursday: Building our own Feed-forward Neural Network.</li> 
+<p><li> Friday: Playing around with our own Feed-forward Neural Network and introduction to TensorFlow. Start convolutional Neural Networks (CNN).</li> 
 </ul>
 <p>
 
 Reading suggestions for both days: <a href="https://github.com/CompPhysics/MachineLearning/blob/master/doc/T\
-extbooks/TensorflowML.pdf" target="_blank">Aurelien Geron's chapters 10-11</a> and Hastie et al chapter 11.
+extbooks/TensorflowML.pdf" target="_blank">Aurelien Geron's chapters 10-11</a>.
+For amore in depth discussion on  neural networks we recommend Goodfellow et al chapters 6 and 7. For CNNs, see Goodfellow et al chapter 9. chapter 11 and 12 on practicalities and applications
 </section>
 
 
 <section>
-<h2 id="___sec1">Setting up the Back propagation algorithm </h2>
+<h2 id="setting-up-the-back-propagation-algorithm">Setting up the Back propagation algorithm </h2>
 
 <p>
 The four equations  derived last week provide us with a way of computing the gradient of the cost function. Let us write this out in the form of an algorithm.
@@ -246,7 +247,7 @@ Here it is convenient to use stochastic gradient descent (see the examples below
 
 
 <section>
-<h2 id="___sec2">Setting up a Multi-layer perceptron model for classification  </h2>
+<h2 id="setting-up-a-multi-layer-perceptron-model-for-classification">Setting up a Multi-layer perceptron model for classification  </h2>
 
 <p>
 We are now gong to develop an example based on the MNIST data
@@ -288,7 +289,7 @@ of our network.
 
 
 <section>
-<h2 id="___sec3">Defining the cost function </h2>
+<h2 id="defining-the-cost-function">Defining the cost function </h2>
 
 <p>
 Our cost function is given as (see the Logistic regression lectures)
@@ -356,7 +357,7 @@ The back propagation equations need now only a small change, namely the definiti
 
 
 <section>
-<h2 id="___sec4">Example: binary classification problem </h2>
+<h2 id="example-binary-classification-problem">Example: binary classification problem </h2>
 
 <p>
 As an example of the above, relevant for project 2 as well, let us consider a binary class. As discussed in our logistic regression lectures, we defined a cost function in terms of the parameters \( \beta \) as
@@ -418,7 +419,7 @@ In case we use another activation function than the logistic one, we need to eva
 
 
 <section>
-<h2 id="___sec5">The Softmax function </h2>
+<h2 id="the-softmax-function">The Softmax function </h2>
 In case we employ the more general case given by the Softmax equation, we need to evaluate the derivative of the activation function with respect to the activation \( z_i^l \), that is we need
 <p>&nbsp;<br>
 $$
@@ -446,7 +447,7 @@ which in case of the simply binary model reduces to  having \( i=j \).
 
 
 <section>
-<h2 id="___sec6">Developing a code for doing neural networks with back propagation </h2>
+<h2 id="developing-a-code-for-doing-neural-networks-with-back-propagation">Developing a code for doing neural networks with back propagation </h2>
 
 <p>
 One can identify a set of key steps when using neural networks to solve supervised learning problems:  
@@ -468,7 +469,7 @@ One can identify a set of key steps when using neural networks to solve supervis
 
 
 <section>
-<h2 id="___sec7">Collect and pre-process data </h2>
+<h2 id="collect-and-pre-process-data">Collect and pre-process data </h2>
 
 <p>
 Here we will be using the MNIST dataset, which is readily available through the <b>scikit-learn</b>
@@ -521,7 +522,7 @@ images.
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%"><span></span><span style="color: #228B22"># import necessary packages</span>
+<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%;"><span></span><span style="color: #228B22"># import necessary packages</span>
 <span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
 <span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
 <span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn</span> <span style="color: #8B008B; font-weight: bold">import</span> datasets
@@ -568,7 +569,7 @@ plt.show()
 
 
 <section>
-<h2 id="___sec8">Train and test datasets </h2>
+<h2 id="train-and-test-datasets">Train and test datasets </h2>
 
 <p>
 Performing analysis before partitioning the dataset is a major error, that can lead to incorrect conclusions.
@@ -586,7 +587,7 @@ collected from 12.00 to 24.00.
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%"><span></span><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.model_selection</span> <span style="color: #8B008B; font-weight: bold">import</span> train_test_split
+<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%;"><span></span><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.model_selection</span> <span style="color: #8B008B; font-weight: bold">import</span> train_test_split
 
 <span style="color: #228B22"># one-liner from scikit-learn library</span>
 train_size = <span style="color: #B452CD">0.8</span>
@@ -618,7 +619,7 @@ X_train, X_test, Y_train, Y_test = train_test_split(inputs, labels, train_size=t
 
 
 <section>
-<h2 id="___sec9">Define model and architecture </h2>
+<h2 id="define-model-and-architecture">Define model and architecture </h2>
 
 <p>
 Our simple feed-forward neural network will consist of an <em>input</em> layer, a single <em>hidden</em> layer and an <em>output</em> layer. The activation \( y \) of each neuron is a weighted sum of inputs, passed through an activation function. In case of the simple perceptron model we have
@@ -672,7 +673,7 @@ which is inspired by probability theory (see logistic regression) and was most c
 
 
 <section>
-<h2 id="___sec10">Layers </h2>
+<h2 id="layers">Layers </h2>
 
 <ul>
 <p><li> Input</li> 
@@ -724,7 +725,7 @@ weights to the output layer.
 
 
 <section>
-<h2 id="___sec11">Weights and biases </h2>
+<h2 id="weights-and-biases">Weights and biases </h2>
 
 <p>
 Typically weights are initialized with small values distributed around zero, drawn from a uniform
@@ -743,7 +744,7 @@ The bias weights \( \hat{b} \) are often initialized to zero, but a small value
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%"><span></span><span style="color: #228B22"># building our neural network</span>
+<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%;"><span></span><span style="color: #228B22"># building our neural network</span>
 
 n_inputs, n_features = X_train.shape
 n_hidden_neurons = <span style="color: #B452CD">50</span>
@@ -763,7 +764,7 @@ output_bias = np.zeros(n_categories) + <span style="color: #B452CD">0.01</span>
 
 
 <section>
-<h2 id="___sec12">Feed-forward pass </h2>
+<h2 id="feed-forward-pass">Feed-forward pass </h2>
 
 <p>
 Denote \( F \) the number of features, \( H \) the number of hidden neurons and \( C \) the number of categories.  
@@ -798,7 +799,7 @@ $$ a_{j}^{L} = \frac{\exp{(z_j^{L})}}
 
 
 <section>
-<h2 id="___sec13">Matrix multiplications </h2>
+<h2 id="matrix-multiplications">Matrix multiplications </h2>
 
 <p>
 Since our data has the dimensions \( X = (n_{inputs}, n_{features}) \) and our weights to the hidden
@@ -844,7 +845,7 @@ $$ output = softmax (\hat{z}^{L}) = (n_{inputs}, n_{categories}) .$$
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%"><span></span><span style="color: #228B22"># setup the feed-forward pass, subscript h = hidden layer</span>
+<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%;"><span></span><span style="color: #228B22"># setup the feed-forward pass, subscript h = hidden layer</span>
 
 <span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">sigmoid</span>(x):
     <span style="color: #8B008B; font-weight: bold">return</span> <span style="color: #B452CD">1</span>/(<span style="color: #B452CD">1</span> + np.exp(-x))
@@ -884,7 +885,7 @@ predictions = predict(X_train)
 
 
 <section>
-<h2 id="___sec14">Choose cost function and optimizer </h2>
+<h2 id="choose-cost-function-and-optimizer">Choose cost function and optimizer </h2>
 
 <p>
 To measure how well our neural network is doing we need to introduce a cost function.  
@@ -919,7 +920,7 @@ you got the correct label. The probability of category \( c \) is given by the s
 
 
 <section>
-<h2 id="___sec15">Optimizing the cost function </h2>
+<h2 id="optimizing-the-cost-function">Optimizing the cost function </h2>
 
 <p>
 The network is trained by finding the weights and biases that minimize the cost function. One of the most widely used classes of methods is <em>gradient descent</em> and its generalizations. The idea behind gradient descent
@@ -965,7 +966,7 @@ The various optmization  methods, with codes and algorithms,  are discussed in o
 
 
 <section>
-<h2 id="___sec16">Regularization </h2>
+<h2 id="regularization">Regularization </h2>
 
 <p>
 It is common to add an extra term to the cost function, proportional
@@ -1000,7 +1001,7 @@ calculate the gradient efficently.
 
 
 <section>
-<h2 id="___sec17">Matrix  multiplication </h2>
+<h2 id="matrix-multiplication">Matrix  multiplication </h2>
 
 <p>
 To more efficently train our network these equations are implemented using matrix operations.  
@@ -1054,7 +1055,7 @@ $$ \nabla b_{h} = \sum_{i=1}^{n_{inputs}} \delta_h = (n_{hidden}) .$$
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%"><span></span><span style="color: #228B22"># to categorical turns our integer vector into a onehot representation</span>
+<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%;"><span></span><span style="color: #228B22"># to categorical turns our integer vector into a onehot representation</span>
 <span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.metrics</span> <span style="color: #8B008B; font-weight: bold">import</span> accuracy_score
 
 <span style="color: #228B22"># one-hot in numpy</span>
@@ -1127,7 +1128,7 @@ lmbd = <span style="color: #B452CD">0.01</span>
 
 
 <section>
-<h2 id="___sec18">Improving performance </h2>
+<h2 id="improving-performance">Improving performance </h2>
 
 <p>
 As we can see the network does not seem to be learning at all. It seems to be just guessing the label for each image.  
@@ -1147,7 +1148,7 @@ Andrew Ng goes through some of these considerations in this <a href="https://you
 
 
 <section>
-<h2 id="___sec19">Full object-oriented implementation </h2>
+<h2 id="full-object-oriented-implementation">Full object-oriented implementation </h2>
 
 <p>
 It is very natural to think of the network as an object, with specific instances of the network
@@ -1156,7 +1157,7 @@ being realizations of this object with different hyperparameters. An implementat
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%"><span></span><span style="color: #8B008B; font-weight: bold">class</span> <span style="color: #008b45; font-weight: bold">NeuralNetwork</span>:
+<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%;"><span></span><span style="color: #8B008B; font-weight: bold">class</span> <span style="color: #008b45; font-weight: bold">NeuralNetwork</span>:
     <span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">__init__</span>(
             <span style="color: #658b00">self</span>,
             X_data,
@@ -1260,7 +1261,7 @@ being realizations of this object with different hyperparameters. An implementat
 
 
 <section>
-<h2 id="___sec20">Evaluate model performance on test data </h2>
+<h2 id="evaluate-model-performance-on-test-data">Evaluate model performance on test data </h2>
 
 <p>
 To measure the performance of our network we evaluate how well it does it data it has never seen before, i.e. the test data.  
@@ -1277,7 +1278,7 @@ where \( I \) is the indicator function, \( 1 \) if \( \hat{y}_i = y_i \) and \(
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%"><span></span>epochs = <span style="color: #B452CD">100</span>
+<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%;"><span></span>epochs = <span style="color: #B452CD">100</span>
 batch_size = <span style="color: #B452CD">100</span>
 
 dnn = NeuralNetwork(X_train, Y_train_onehot, eta=eta, lmbd=lmbd, epochs=epochs, batch_size=batch_size,
@@ -1298,7 +1299,7 @@ test_predict = dnn.predict(X_test)
 
 
 <section>
-<h2 id="___sec21">Adjust hyperparameters </h2>
+<h2 id="adjust-hyperparameters">Adjust hyperparameters </h2>
 
 <p>
 We now perform a grid search to find the optimal hyperparameters for the network.  
@@ -1307,7 +1308,7 @@ Note that we are only using 1 layer with 50 neurons, and human performance is es
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%"><span></span>eta_vals = np.logspace(-<span style="color: #B452CD">5</span>, <span style="color: #B452CD">1</span>, <span style="color: #B452CD">7</span>)
+<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%;"><span></span>eta_vals = np.logspace(-<span style="color: #B452CD">5</span>, <span style="color: #B452CD">1</span>, <span style="color: #B452CD">7</span>)
 lmbd_vals = np.logspace(-<span style="color: #B452CD">5</span>, <span style="color: #B452CD">1</span>, <span style="color: #B452CD">7</span>)
 <span style="color: #228B22"># store the models for later use</span>
 DNN_numpy = np.zeros((<span style="color: #658b00">len</span>(eta_vals), <span style="color: #658b00">len</span>(lmbd_vals)), dtype=<span style="color: #658b00">object</span>)
@@ -1332,12 +1333,12 @@ DNN_numpy = np.zeros((<span style="color: #658b00">len</span>(eta_vals), <span s
 
 
 <section>
-<h2 id="___sec22">Visualization </h2>
+<h2 id="visualization">Visualization </h2>
 
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%"><span></span><span style="color: #228B22"># visual representation of grid search</span>
+<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%;"><span></span><span style="color: #228B22"># visual representation of grid search</span>
 <span style="color: #228B22"># uses seaborn heatmap, you can also do this with matplotlib imshow</span>
 <span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">seaborn</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">sns</span>
 
@@ -1375,7 +1376,7 @@ plt.show()
 
 
 <section>
-<h2 id="___sec23">scikit-learn implementation </h2>
+<h2 id="scikit-learn-implementation">scikit-learn implementation </h2>
 
 <p>
 <b>scikit-learn</b> focuses more
@@ -1394,7 +1395,7 @@ performance overall.
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%"><span></span><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.neural_network</span> <span style="color: #8B008B; font-weight: bold">import</span> MLPClassifier
+<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%;"><span></span><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.neural_network</span> <span style="color: #8B008B; font-weight: bold">import</span> MLPClassifier
 <span style="color: #228B22"># store models for later use</span>
 DNN_scikit = np.zeros((<span style="color: #658b00">len</span>(eta_vals), <span style="color: #658b00">len</span>(lmbd_vals)), dtype=<span style="color: #658b00">object</span>)
 
@@ -1415,11 +1416,11 @@ DNN_scikit = np.zeros((<span style="color: #658b00">len</span>(eta_vals), <span
 
 
 <section>
-<h2 id="___sec24">Visualization </h2>
+<h2 id="visualization">Visualization </h2>
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%"><span></span><span style="color: #228B22"># optional</span>
+<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%;"><span></span><span style="color: #228B22"># optional</span>
 <span style="color: #228B22"># visual representation of grid search</span>
 <span style="color: #228B22"># uses seaborn heatmap, could probably do this in matplotlib</span>
 <span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">seaborn</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">sns</span>
@@ -1458,7 +1459,7 @@ plt.show()
 
 
 <section>
-<h2 id="___sec25">Building neural networks in Tensorflow and Keras </h2>
+<h2 id="building-neural-networks-in-tensorflow-and-keras">Building neural networks in Tensorflow and Keras </h2>
 
 <p>
 Now we want  to build on the experience gained from our neural network implementation in NumPy and scikit-learn
@@ -1473,7 +1474,7 @@ NumPy arrays.
 
 
 <section>
-<h2 id="___sec26">Tensorflow </h2>
+<h2 id="tensorflow">Tensorflow </h2>
 
 <p>
 Tensorflow is an open source library machine learning library
@@ -1506,7 +1507,7 @@ To install tensorflow on Unix/Linux systems, use pip as
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%"><span></span>pip3 install tensorflow
+<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%;"><span></span>pip3 install tensorflow
 </pre></div>
 <p>
 and/or if you use <b>anaconda</b>, just write (or install from the graphical user interface)
@@ -1514,7 +1515,7 @@ and/or if you use <b>anaconda</b>, just write (or install from the graphical use
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%"><span></span>conda create -n tf tensorflow
+<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%;"><span></span>conda create -n tf tensorflow
 conda activate tf
 </pre></div>
 <p>
@@ -1522,14 +1523,14 @@ To install the current release of GPU TensorFlow
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%"><span></span>conda create -n tf-gpu tensorflow-gpu
+<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%;"><span></span>conda create -n tf-gpu tensorflow-gpu
 conda activate tf-gpu
 </pre></div>
 </section>
 
 
 <section>
-<h2 id="___sec27">Using Keras </h2>
+<h2 id="using-keras">Using Keras </h2>
 
 <p>
 Keras is a high level <a href="https://en.wikipedia.org/wiki/Application_programming_interface" target="_blank">neural network</a>
@@ -1538,7 +1539,7 @@ If you have Anaconda installed you may run the following command
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%"><span></span>conda install keras
+<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%;"><span></span>conda install keras
 </pre></div>
 <p>
 You can look up the <a href="https://keras.io/" target="_blank">instructions here</a> for more information.
@@ -1549,7 +1550,7 @@ We will to a large extent use <b>keras</b> in this course.
 
 
 <section>
-<h2 id="___sec28">Collect and pre-process data </h2>
+<h2 id="collect-and-pre-process-data">Collect and pre-process data </h2>
 
 <p>
 Let us look again at the MINST data set.
@@ -1557,7 +1558,7 @@ Let us look again at the MINST data set.
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%"><span></span><span style="color: #228B22"># import necessary packages</span>
+<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%;"><span></span><span style="color: #228B22"># import necessary packages</span>
 <span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
 <span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
 <span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">tensorflow</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">tf</span>
@@ -1604,7 +1605,7 @@ plt.show()
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%"><span></span><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">tensorflow.keras.layers</span> <span style="color: #8B008B; font-weight: bold">import</span> Input
+<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%;"><span></span><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">tensorflow.keras.layers</span> <span style="color: #8B008B; font-weight: bold">import</span> Input
 <span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">tensorflow.keras.models</span> <span style="color: #8B008B; font-weight: bold">import</span> Sequential      <span style="color: #228B22">#This allows appending layers to existing models</span>
 <span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">tensorflow.keras.layers</span> <span style="color: #8B008B; font-weight: bold">import</span> Dense           <span style="color: #228B22">#This allows defining the characteristics of a particular layer</span>
 <span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">tensorflow.keras</span> <span style="color: #8B008B; font-weight: bold">import</span> optimizers             <span style="color: #228B22">#This allows using whichever optimiser we want (sgd,adam,RMSprop)</span>
@@ -1625,7 +1626,7 @@ X_train, X_test, Y_train, Y_test = train_test_split(inputs, labels, train_size=t
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%"><span></span>epochs = <span style="color: #B452CD">100</span>
+<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%;"><span></span>epochs = <span style="color: #B452CD">100</span>
 batch_size = <span style="color: #B452CD">100</span>
 n_neurons_layer1 = <span style="color: #B452CD">100</span>
 n_neurons_layer2 = <span style="color: #B452CD">50</span>
@@ -1646,7 +1647,7 @@ lmbd_vals = np.logspace(-<span style="color: #B452CD">5</span>, <span style="col
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%"><span></span>DNN_keras = np.zeros((<span style="color: #658b00">len</span>(eta_vals), <span style="color: #658b00">len</span>(lmbd_vals)), dtype=<span style="color: #658b00">object</span>)
+<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%;"><span></span>DNN_keras = np.zeros((<span style="color: #658b00">len</span>(eta_vals), <span style="color: #658b00">len</span>(lmbd_vals)), dtype=<span style="color: #658b00">object</span>)
         
 <span style="color: #8B008B; font-weight: bold">for</span> i, eta <span style="color: #8B008B">in</span> <span style="color: #658b00">enumerate</span>(eta_vals):
     <span style="color: #8B008B; font-weight: bold">for</span> j, lmbd <span style="color: #8B008B">in</span> <span style="color: #658b00">enumerate</span>(lmbd_vals):
@@ -1665,7 +1666,7 @@ lmbd_vals = np.logspace(-<span style="color: #B452CD">5</span>, <span style="col
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%"><span></span><span style="color: #228B22"># optional</span>
+<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%;"><span></span><span style="color: #228B22"># optional</span>
 <span style="color: #228B22"># visual representation of grid search</span>
 <span style="color: #228B22"># uses seaborn heatmap, could probably do this in matplotlib</span>
 <span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">seaborn</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">sns</span>
@@ -1701,12 +1702,12 @@ plt.show()
 
 
 <section>
-<h2 id="___sec29">The Breast Cancer Data, now with Keras </h2>
+<h2 id="the-breast-cancer-data-now-with-keras">The Breast Cancer Data, now with Keras </h2>
 
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%"><span></span><span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">tensorflow</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">tf</span>
+<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%;"><span></span><span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">tensorflow</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">tf</span>
 <span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">tensorflow.keras.layers</span> <span style="color: #8B008B; font-weight: bold">import</span> Input
 <span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">tensorflow.keras.models</span> <span style="color: #8B008B; font-weight: bold">import</span> Sequential      <span style="color: #228B22">#This allows appending layers to existing models</span>
 <span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">tensorflow.keras.layers</span> <span style="color: #8B008B; font-weight: bold">import</span> Dense           <span style="color: #228B22">#This allows defining the characteristics of a particular layer</span>
@@ -1876,7 +1877,7 @@ plot_data(eta,n_neuron,Test_accuracy, <span style="color: #CD5555">&#39;testing&
 
 
 <section>
-<h2 id="___sec30">Fine-tuning neural network hyperparameters </h2>
+<h2 id="fine-tuning-neural-network-hyperparameters">Fine-tuning neural network hyperparameters </h2>
 
 <p>
 The flexibility of neural networks is also one of their main
@@ -1904,7 +1905,7 @@ will only be able to explore a tiny part of the hyperparameter space.
 
 
 <section>
-<h2 id="___sec31">Hidden layers </h2>
+<h2 id="hidden-layers">Hidden layers </h2>
 
 <p>
 For many problems you can start with just one or two hidden layers and it will work just fine.
@@ -1924,7 +1925,7 @@ common to reuse parts of a pretrained state-of-the-art network that performs a s
 
 
 <section>
-<h2 id="___sec32">Which activation function should I use? </h2>
+<h2 id="which-activation-function-should-i-use">Which activation function should I use? </h2>
 
 <p>
 The Back propagation algorithm we derived above works by going from
@@ -1953,7 +1954,7 @@ learn at widely different speeds
 
 
 <section>
-<h2 id="___sec33">Is the Logistic activation function (Sigmoid)  our choice? </h2>
+<h2 id="is-the-logistic-activation-function-sigmoid-our-choice">Is the Logistic activation function (Sigmoid)  our choice? </h2>
 
 <p>
 Although this unfortunate behavior has been empirically observed for
@@ -1983,7 +1984,7 @@ better than the logistic function in deep networks).
 
 
 <section>
-<h2 id="___sec34">The derivative of the Logistic funtion </h2>
+<h2 id="the-derivative-of-the-logistic-funtion">The derivative of the Logistic funtion </h2>
 
 <p>
 Looking at the logistic activation function, when inputs become large
@@ -2019,7 +2020,7 @@ fast to compute).
 
 
 <section>
-<h2 id="___sec35">The RELU function family </h2>
+<h2 id="the-relu-function-family">The RELU function family </h2>
 
 <p>
 The ReLU activation function suffers from a problem known as the dying
@@ -2048,7 +2049,7 @@ $$
 
 
 <section>
-<h2 id="___sec36">Which activation function should we use? </h2>
+<h2 id="which-activation-function-should-we-use">Which activation function should we use? </h2>
 
 <p>
 In general it seems that the ELU activation function is better than
@@ -2068,7 +2069,7 @@ bootstrap to evaluate other activation functions.
 
 
 <section>
-<h2 id="___sec37">More on activation functions, output layers </h2>
+<h2 id="more-on-activation-functions-output-layers">More on activation functions, output layers </h2>
 
 <p>
 In most cases you can use the ReLU activation function in the hidden layers (or one of its variants).
@@ -2087,7 +2088,7 @@ It is a bit faster to compute than other activation functions, and the gradient
 
 
 <section>
-<h2 id="___sec38">Batch Normalization </h2>
+<h2 id="batch-normalization">Batch Normalization </h2>
 
 <p>
 Batch Normalization
@@ -2106,7 +2107,7 @@ mini-batch, from this the name batch normalization.
 
 
 <section>
-<h2 id="___sec39">Dropout </h2>
+<h2 id="dropout">Dropout </h2>
 
 <p>
 It is a fairly simple algorithm: at every training step, every neuron (including the input neurons but
@@ -2121,7 +2122,7 @@ hyperparameter \( p \) is called the dropout rate, and it is typically set to 50
 
 
 <section>
-<h2 id="___sec40">Gradient Clipping </h2>
+<h2 id="gradient-clipping">Gradient Clipping </h2>
 
 <p>
 A popular technique to lessen the exploding gradients problem is to simply clip the gradients during
@@ -2138,7 +2139,7 @@ Normalization is preferred.
 
 
 <section>
-<h2 id="___sec41">A very nice website on Neural Networks </h2>
+<h2 id="a-very-nice-website-on-neural-networks">A very nice website on Neural Networks </h2>
 
 <p>
 You may find this <a href="https://playground.tensorflow.org/#activation=tanh&batchSize=10&dataset=circle&regDataset=reg-plane&learningRate=0.03&regularizationRate=0&noise=0&networkShape=4,2&seed=0.29243&showTestData=false&discretize=false&percTrainData=50&x=true&y=true&xTimesY=false&xSquared=false&ySquared=false&cosX=false&sinX=false&cosY=false&sinY=false&collectStats=false&problem=classification&initZero=false&hideText=false" target="_blank">website</a> very useful. Thx a million to Ghadi for sharing.
@@ -2146,7 +2147,7 @@ You may find this <a href="https://playground.tensorflow.org/#activation=tanh&ba
 
 
 <section>
-<h2 id="___sec42">A top-down perspective on Neural networks </h2>
+<h2 id="a-top-down-perspective-on-neural-networks">A top-down perspective on Neural networks </h2>
 
 <p>
 The first thing we would like to do is divide the data into two or three
@@ -2189,7 +2190,7 @@ supervised learning.
 
 
 <section>
-<h2 id="___sec43">Limitations of supervised learning with deep networks </h2>
+<h2 id="limitations-of-supervised-learning-with-deep-networks">Limitations of supervised learning with deep networks </h2>
 
 <p>
 Like all statistical methods, supervised learning using neural
@@ -2216,7 +2217,7 @@ Some of these remarks are particular to DNNs, others are shared by all supervise
 
 
 <section>
-<h2 id="___sec44">Convolutional Neural Networks (recognizing images) </h2>
+<h2 id="convolutional-neural-networks-recognizing-images">Convolutional Neural Networks (recognizing images) </h2>
 
 <p>
 Convolutional neural networks (CNNs) were developed during the last
@@ -2255,7 +2256,7 @@ Another good read is the article here <a href="https://arxiv.org/pdf/1603.07285.
 
 
 <section>
-<h2 id="___sec45">Regular NNs don&#8217;t scale well to full images </h2>
+<h2 id="regular-nns-don-t-scale-well-to-full-images">Regular NNs don&#8217;t scale well to full images </h2>
 
 <p>
 As an example, consider
@@ -2283,7 +2284,7 @@ would quickly lead to possible overfitting.
 
 
 <section>
-<h2 id="___sec46">3D volumes of neurons </h2>
+<h2 id="3d-volumes-of-neurons">3D volumes of neurons </h2>
 
 <p>
 Convolutional Neural Networks take advantage of the fact that the
@@ -2323,7 +2324,7 @@ dimension.
 
 
 <section>
-<h2 id="___sec47">Layers used to build CNNs </h2>
+<h2 id="layers-used-to-build-cnns">Layers used to build CNNs </h2>
 
 <p>
 A simple CNN is a sequence of layers, and every layer of a CNN
@@ -2347,7 +2348,7 @@ A simple CNN for image classification could have the architecture:
 
 
 <section>
-<h2 id="___sec48">Transforming images </h2>
+<h2 id="transforming-images">Transforming images </h2>
 
 <p>
 CNNs transform the original image layer by layer from the original
@@ -2366,7 +2367,7 @@ are consistent with the labels in the training set for each image.
 
 
 <section>
-<h2 id="___sec49">CNNs in brief </h2>
+<h2 id="cnns-in-brief">CNNs in brief </h2>
 
 <p>
 In summary:
@@ -2388,7 +2389,7 @@ and the slides of <a href="http://cs231n.github.io/convolutional-networks/" targ
 
 
 <section>
-<h2 id="___sec50">CNNs in more detail, building convolutional neural networks in Tensorflow and Keras </h2>
+<h2 id="cnns-in-more-detail-building-convolutional-neural-networks-in-tensorflow-and-keras">CNNs in more detail, building convolutional neural networks in Tensorflow and Keras </h2>
 
 <p>
 As discussed above, CNNs are neural networks built from the assumption that the inputs
@@ -2404,7 +2405,7 @@ matrices, typically 1 for each color dimension (Red, Green, Blue).
 
 
 <section>
-<h2 id="___sec51">Setting it up </h2>
+<h2 id="setting-it-up">Setting it up </h2>
 
 <p>
 It means that to represent the entire
@@ -2419,7 +2420,7 @@ $$
 
 
 <section>
-<h2 id="___sec52">The MNIST dataset again </h2>
+<h2 id="the-mnist-dataset-again">The MNIST dataset again </h2>
 
 <p>
 The MNIST dataset consists of grayscale images with a pixel size of
@@ -2437,7 +2438,7 @@ single neuron in the first hidden layer.
 
 
 <section>
-<h2 id="___sec53">Strong correlations </h2>
+<h2 id="strong-correlations">Strong correlations </h2>
 
 <p>
 Images typically have strong local correlations, meaning that a small
@@ -2455,7 +2456,7 @@ fixed, and known as a <a href="https://en.wikipedia.org/wiki/Receptive_field" ta
 
 
 <section>
-<h2 id="___sec54">Layers of a CNN </h2>
+<h2 id="layers-of-a-cnn">Layers of a CNN </h2>
 The layers of a convolutional neural network arrange neurons in 3D: width, height and depth.  
 The input image is typically a square matrix of depth 3.
 
@@ -2477,7 +2478,7 @@ layer.
 
 
 <section>
-<h2 id="___sec55">Systematic reduction </h2>
+<h2 id="systematic-reduction">Systematic reduction </h2>
 
 <p>
 By systematically reducing the size of the input volume, through
@@ -2492,11 +2493,11 @@ classification.
 
 
 <section>
-<h2 id="___sec56">Prerequisites: Collect and pre-process data </h2>
+<h2 id="prerequisites-collect-and-pre-process-data">Prerequisites: Collect and pre-process data </h2>
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%"><span></span><span style="color: #228B22"># import necessary packages</span>
+<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%;"><span></span><span style="color: #228B22"># import necessary packages</span>
 <span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
 <span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
 <span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn</span> <span style="color: #8B008B; font-weight: bold">import</span> datasets
@@ -2541,11 +2542,11 @@ plt.show()
 
 
 <section>
-<h2 id="___sec57">Importing Keras and Tensorflow </h2>
+<h2 id="importing-keras-and-tensorflow">Importing Keras and Tensorflow </h2>
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%"><span></span><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">tensorflow.keras</span> <span style="color: #8B008B; font-weight: bold">import</span> datasets, layers, models
+<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%;"><span></span><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">tensorflow.keras</span> <span style="color: #8B008B; font-weight: bold">import</span> datasets, layers, models
 <span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">tensorflow.keras.layers</span> <span style="color: #8B008B; font-weight: bold">import</span> Input
 <span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">tensorflow.keras.models</span> <span style="color: #8B008B; font-weight: bold">import</span> Sequential      <span style="color: #228B22">#This allows appending layers to existing models</span>
 <span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">tensorflow.keras.layers</span> <span style="color: #8B008B; font-weight: bold">import</span> Dense           <span style="color: #228B22">#This allows defining the characteristics of a particular layer</span>
@@ -2572,12 +2573,12 @@ X_train, X_test, Y_train, Y_test = train_test_split(inputs, labels, train_size=t
 
 
 <section>
-<h2 id="___sec58">Running with Keras </h2>
+<h2 id="running-with-keras">Running with Keras </h2>
 
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%"><span></span><span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">create_convolutional_neural_network_keras</span>(input_shape, receptive_field,
+<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%;"><span></span><span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">create_convolutional_neural_network_keras</span>(input_shape, receptive_field,
                                               n_filters, n_neurons_connected, n_categories,
                                               eta, lmbd):
     model = Sequential()
@@ -2608,12 +2609,12 @@ lmbd_vals = np.logspace(-<span style="color: #B452CD">5</span>, <span style="col
 
 
 <section>
-<h2 id="___sec59">Final part </h2>
+<h2 id="final-part">Final part </h2>
 
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%"><span></span>CNN_keras = np.zeros((<span style="color: #658b00">len</span>(eta_vals), <span style="color: #658b00">len</span>(lmbd_vals)), dtype=<span style="color: #658b00">object</span>)
+<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%;"><span></span>CNN_keras = np.zeros((<span style="color: #658b00">len</span>(eta_vals), <span style="color: #658b00">len</span>(lmbd_vals)), dtype=<span style="color: #658b00">object</span>)
         
 <span style="color: #8B008B; font-weight: bold">for</span> i, eta <span style="color: #8B008B">in</span> <span style="color: #658b00">enumerate</span>(eta_vals):
     <span style="color: #8B008B; font-weight: bold">for</span> j, lmbd <span style="color: #8B008B">in</span> <span style="color: #658b00">enumerate</span>(lmbd_vals):
@@ -2634,12 +2635,12 @@ lmbd_vals = np.logspace(-<span style="color: #B452CD">5</span>, <span style="col
 
 
 <section>
-<h2 id="___sec60">Final visualization </h2>
+<h2 id="final-visualization">Final visualization </h2>
 
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%"><span></span><span style="color: #228B22"># visual representation of grid search</span>
+<div class="highlight" style="background: #eeeedd"><pre style="font-size: 80%; line-height: 125%;"><span></span><span style="color: #228B22"># visual representation of grid search</span>
 <span style="color: #228B22"># uses seaborn heatmap, could probably do this in matplotlib</span>
 <span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">seaborn</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">sns</span>
 
@@ -2674,7 +2675,7 @@ plt.show()
 
 
 <section>
-<h2 id="___sec61">Fun links </h2>
+<h2 id="fun-links">Fun links </h2>
 
 <ol>
 <p><li> <a href="https://arxiv.org/abs/1604.07316" target="_blank">Self-Driving cars using a convolutional neural network</a></li>
diff --git a/doc/pub/week41/html/week41-solarized.html b/doc/pub/week41/html/week41-solarized.html
index d50fe5c91..d49686ea2 100644
--- a/doc/pub/week41/html/week41-solarized.html
+++ b/doc/pub/week41/html/week41-solarized.html
@@ -1,19 +1,19 @@
 <!--
 Automatically generated HTML file from DocOnce source
-(https://github.com/hplgit/doconce/)
+(https://github.com/doconce/doconce/)
 -->
 <html>
 <head>
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
-<meta name="generator" content="DocOnce: https://github.com/hplgit/doconce/" />
+<meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 41 Tensor flow and Deep Learning, Convolutional Neural Networks">
+<meta name="description" content="Week 41 Constructing a Neural Network code, Tensor flow and start Convolutional Neural Networks">
 
-<title>Week 41 Tensor flow and Deep Learning, Convolutional Neural Networks</title>
+<title>Week 41 Constructing a Neural Network code, Tensor flow and start Convolutional Neural Networks</title>
 
 
-<link href="https://cdn.rawgit.com/hplgit/doconce/master/bundled/html_styles/style_solarized_box/css/solarized_light_code.css" rel="stylesheet" type="text/css" title="light"/>
-<script src="https://cdn.rawgit.com/hplgit/doconce/master/bundled/html_styles/style_solarized_box/js/highlight.pack.js"></script>
+<link href="https://cdn.rawgit.com/doconce/doconce/master/bundled/html_styles/style_solarized_box/css/solarized_light_code.css" rel="stylesheet" type="text/css" title="light"/>
+<script src="https://cdn.rawgit.com/doconce/doconce/master/bundled/html_styles/style_solarized_box/js/highlight.pack.js"></script>
 <script>hljs.initHighlightingOnLoad();</script>
 
 <link href="https://thomasf.github.io/solarized-css/solarized-light.min.css" rel="stylesheet">
@@ -48,10 +48,10 @@ pre {
 .alert-block > p, .alert-block > ul {margin-bottom:1em}
 .alert li {margin-top: 1em}
 .alert-block p+p {margin-top:5px}
-.alert-notice { background-image: url(https://cdn.rawgit.com/hplgit/doconce/master/bundled/html_images/small_yellow_notice.png); }
-.alert-summary  { background-image:url(https://cdn.rawgit.com/hplgit/doconce/master/bundled/html_images/small_yellow_summary.png); }
-.alert-warning { background-image: url(https://cdn.rawgit.com/hplgit/doconce/master/bundled/html_images/small_yellow_warning.png); }
-.alert-question {background-image:url(https://cdn.rawgit.com/hplgit/doconce/master/bundled/html_images/small_yellow_question.png); }
+.alert-notice { background-image: url(https://cdn.rawgit.com/doconce/doconce/master/bundled/html_images/small_yellow_notice.png); }
+.alert-summary  { background-image:url(https://cdn.rawgit.com/doconce/doconce/master/bundled/html_images/small_yellow_summary.png); }
+.alert-warning { background-image: url(https://cdn.rawgit.com/doconce/doconce/master/bundled/html_images/small_yellow_warning.png); }
+.alert-question {background-image:url(https://cdn.rawgit.com/doconce/doconce/master/bundled/html_images/small_yellow_question.png); }
 
 div { text-align: justify; text-justify: inter-word; }
 </style>
@@ -61,106 +61,160 @@ div { text-align: justify; text-justify: inter-word; }
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Plan for week 41', 2, None, '___sec0'),
-              ('Setting up the Back propagation algorithm', 2, None, '___sec1'),
+ 'sections': [('Plan for week 41', 2, None, 'plan-for-week-41'),
+              ('Setting up the Back propagation algorithm',
+               2,
+               None,
+               'setting-up-the-back-propagation-algorithm'),
               ('Setting up a Multi-layer perceptron model for classification',
                2,
                None,
-               '___sec2'),
-              ('Defining the cost function', 2, None, '___sec3'),
-              ('Example: binary classification problem', 2, None, '___sec4'),
-              ('The Softmax function', 2, None, '___sec5'),
+               'setting-up-a-multi-layer-perceptron-model-for-classification'),
+              ('Defining the cost function',
+               2,
+               None,
+               'defining-the-cost-function'),
+              ('Example: binary classification problem',
+               2,
+               None,
+               'example-binary-classification-problem'),
+              ('The Softmax function', 2, None, 'the-softmax-function'),
               ('Developing a code for doing neural networks with back '
                'propagation',
                2,
                None,
-               '___sec6'),
-              ('Collect and pre-process data', 2, None, '___sec7'),
-              ('Train and test datasets', 2, None, '___sec8'),
-              ('Define model and architecture', 2, None, '___sec9'),
-              ('Layers', 2, None, '___sec10'),
-              ('Weights and biases', 2, None, '___sec11'),
-              ('Feed-forward pass', 2, None, '___sec12'),
-              ('Matrix multiplications', 2, None, '___sec13'),
-              ('Choose cost function and optimizer', 2, None, '___sec14'),
-              ('Optimizing the cost function', 2, None, '___sec15'),
-              ('Regularization', 2, None, '___sec16'),
-              ('Matrix  multiplication', 2, None, '___sec17'),
-              ('Improving performance', 2, None, '___sec18'),
-              ('Full object-oriented implementation', 2, None, '___sec19'),
-              ('Evaluate model performance on test data', 2, None, '___sec20'),
-              ('Adjust hyperparameters', 2, None, '___sec21'),
-              ('Visualization', 2, None, '___sec22'),
-              ('scikit-learn implementation', 2, None, '___sec23'),
-              ('Visualization', 2, None, '___sec24'),
+               'developing-a-code-for-doing-neural-networks-with-back-propagation'),
+              ('Collect and pre-process data',
+               2,
+               None,
+               'collect-and-pre-process-data'),
+              ('Train and test datasets', 2, None, 'train-and-test-datasets'),
+              ('Define model and architecture',
+               2,
+               None,
+               'define-model-and-architecture'),
+              ('Layers', 2, None, 'layers'),
+              ('Weights and biases', 2, None, 'weights-and-biases'),
+              ('Feed-forward pass', 2, None, 'feed-forward-pass'),
+              ('Matrix multiplications', 2, None, 'matrix-multiplications'),
+              ('Choose cost function and optimizer',
+               2,
+               None,
+               'choose-cost-function-and-optimizer'),
+              ('Optimizing the cost function',
+               2,
+               None,
+               'optimizing-the-cost-function'),
+              ('Regularization', 2, None, 'regularization'),
+              ('Matrix  multiplication', 2, None, 'matrix-multiplication'),
+              ('Improving performance', 2, None, 'improving-performance'),
+              ('Full object-oriented implementation',
+               2,
+               None,
+               'full-object-oriented-implementation'),
+              ('Evaluate model performance on test data',
+               2,
+               None,
+               'evaluate-model-performance-on-test-data'),
+              ('Adjust hyperparameters', 2, None, 'adjust-hyperparameters'),
+              ('Visualization', 2, None, 'visualization'),
+              ('scikit-learn implementation',
+               2,
+               None,
+               'scikit-learn-implementation'),
+              ('Visualization', 2, None, 'visualization'),
               ('Building neural networks in Tensorflow and Keras',
                2,
                None,
-               '___sec25'),
-              ('Tensorflow', 2, None, '___sec26'),
-              ('Using Keras', 2, None, '___sec27'),
-              ('Collect and pre-process data', 2, None, '___sec28'),
-              ('The Breast Cancer Data, now with Keras', 2, None, '___sec29'),
+               'building-neural-networks-in-tensorflow-and-keras'),
+              ('Tensorflow', 2, None, 'tensorflow'),
+              ('Using Keras', 2, None, 'using-keras'),
+              ('Collect and pre-process data',
+               2,
+               None,
+               'collect-and-pre-process-data'),
+              ('The Breast Cancer Data, now with Keras',
+               2,
+               None,
+               'the-breast-cancer-data-now-with-keras'),
               ('Fine-tuning neural network hyperparameters',
                2,
                None,
-               '___sec30'),
-              ('Hidden layers', 2, None, '___sec31'),
-              ('Which activation function should I use?', 2, None, '___sec32'),
+               'fine-tuning-neural-network-hyperparameters'),
+              ('Hidden layers', 2, None, 'hidden-layers'),
+              ('Which activation function should I use?',
+               2,
+               None,
+               'which-activation-function-should-i-use'),
               ('Is the Logistic activation function (Sigmoid)  our choice?',
                2,
                None,
-               '___sec33'),
-              ('The derivative of the Logistic funtion', 2, None, '___sec34'),
-              ('The RELU function family', 2, None, '___sec35'),
-              ('Which activation function should we use?', 2, None, '___sec36'),
+               'is-the-logistic-activation-function-sigmoid-our-choice'),
+              ('The derivative of the Logistic funtion',
+               2,
+               None,
+               'the-derivative-of-the-logistic-funtion'),
+              ('The RELU function family', 2, None, 'the-relu-function-family'),
+              ('Which activation function should we use?',
+               2,
+               None,
+               'which-activation-function-should-we-use'),
               ('More on activation functions, output layers',
                2,
                None,
-               '___sec37'),
-              ('Batch Normalization', 2, None, '___sec38'),
-              ('Dropout', 2, None, '___sec39'),
-              ('Gradient Clipping', 2, None, '___sec40'),
-              ('A very nice website on Neural Networks', 2, None, '___sec41'),
+               'more-on-activation-functions-output-layers'),
+              ('Batch Normalization', 2, None, 'batch-normalization'),
+              ('Dropout', 2, None, 'dropout'),
+              ('Gradient Clipping', 2, None, 'gradient-clipping'),
+              ('A very nice website on Neural Networks',
+               2,
+               None,
+               'a-very-nice-website-on-neural-networks'),
               ('A top-down perspective on Neural networks',
                2,
                None,
-               '___sec42'),
+               'a-top-down-perspective-on-neural-networks'),
               ('Limitations of supervised learning with deep networks',
                2,
                None,
-               '___sec43'),
+               'limitations-of-supervised-learning-with-deep-networks'),
               ('Convolutional Neural Networks (recognizing images)',
                2,
                None,
-               '___sec44'),
+               'convolutional-neural-networks-recognizing-images'),
               ('Regular NNs don’t scale well to full images',
                2,
                None,
-               '___sec45'),
-              ('3D volumes of neurons', 2, None, '___sec46'),
-              ('Layers used to build CNNs', 2, None, '___sec47'),
-              ('Transforming images', 2, None, '___sec48'),
-              ('CNNs in brief', 2, None, '___sec49'),
+               'regular-nns-don-t-scale-well-to-full-images'),
+              ('3D volumes of neurons', 2, None, '3d-volumes-of-neurons'),
+              ('Layers used to build CNNs',
+               2,
+               None,
+               'layers-used-to-build-cnns'),
+              ('Transforming images', 2, None, 'transforming-images'),
+              ('CNNs in brief', 2, None, 'cnns-in-brief'),
               ('CNNs in more detail, building convolutional neural networks in '
                'Tensorflow and Keras',
                2,
                None,
-               '___sec50'),
-              ('Setting it up', 2, None, '___sec51'),
-              ('The MNIST dataset again', 2, None, '___sec52'),
-              ('Strong correlations', 2, None, '___sec53'),
-              ('Layers of a CNN', 2, None, '___sec54'),
-              ('Systematic reduction', 2, None, '___sec55'),
+               'cnns-in-more-detail-building-convolutional-neural-networks-in-tensorflow-and-keras'),
+              ('Setting it up', 2, None, 'setting-it-up'),
+              ('The MNIST dataset again', 2, None, 'the-mnist-dataset-again'),
+              ('Strong correlations', 2, None, 'strong-correlations'),
+              ('Layers of a CNN', 2, None, 'layers-of-a-cnn'),
+              ('Systematic reduction', 2, None, 'systematic-reduction'),
               ('Prerequisites: Collect and pre-process data',
                2,
                None,
-               '___sec56'),
-              ('Importing Keras and Tensorflow', 2, None, '___sec57'),
-              ('Running with Keras', 2, None, '___sec58'),
-              ('Final part', 2, None, '___sec59'),
-              ('Final visualization', 2, None, '___sec60'),
-              ('Fun links', 2, None, '___sec61')]}
+               'prerequisites-collect-and-pre-process-data'),
+              ('Importing Keras and Tensorflow',
+               2,
+               None,
+               'importing-keras-and-tensorflow'),
+              ('Running with Keras', 2, None, 'running-with-keras'),
+              ('Final part', 2, None, 'final-part'),
+              ('Final visualization', 2, None, 'final-visualization'),
+              ('Fun links', 2, None, 'fun-links')]}
 end of tocinfo -->
 
 <body>
@@ -186,7 +240,7 @@ MathJax.Hub.Config({
 
 
 
-<center><h1>Week 41 Tensor flow and Deep Learning, Convolutional Neural Networks</h1></center>  <!-- document title -->
+<center><h1>Week 41 Constructing a Neural Network code, Tensor flow and start Convolutional Neural Networks</h1></center>  <!-- document title -->
 
 <p>
 <!-- author(s): Morten Hjorth-Jensen -->
@@ -202,25 +256,26 @@ MathJax.Hub.Config({
 <center>[2] <b>Department of Physics and Astronomy and National Superconducting Cyclotron Laboratory, Michigan State University</b></center>
 <br>
 <p>
-<center><h4>Oct 10, 2020</h4></center> <!-- date -->
+<center><h4>Oct 9, 2021</h4></center> <!-- date -->
 <br>
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec0">Plan for week 41 </h2>
+<h2 id="plan-for-week-41">Plan for week 41 </h2>
 
 <ul>
-<li> Thursday: Building our own Feed-forward Neural Network. <a href="https://www.uio.no/studier/emner/matnat/fys/FYS-STK4155/h20/forelesningsvideoer/LectureOctober8.mp4?vrtx=view-as-webpage" target="_blank">Video of Lecture</a></li>
-<li> Friday: Playing around with our own Feed-forward Neural Network and introduction to TensorFlow. Start convolutional Neural Networks (CNN). <a href="https://www.uio.no/studier/emner/matnat/fys/FYS-STK4155/h20/forelesningsvideoer/LectureOcotber9.mp4?vrtx=view-as-webpage" target="_blank">Video of Lecture</a></li>
+<li> Thursday: Building our own Feed-forward Neural Network.</li> 
+<li> Friday: Playing around with our own Feed-forward Neural Network and introduction to TensorFlow. Start convolutional Neural Networks (CNN).</li> 
 </ul>
 
 Reading suggestions for both days: <a href="https://github.com/CompPhysics/MachineLearning/blob/master/doc/T\
-extbooks/TensorflowML.pdf" target="_blank">Aurelien Geron's chapters 10-11</a> and Hastie et al chapter 11.
+extbooks/TensorflowML.pdf" target="_blank">Aurelien Geron's chapters 10-11</a>.
+For amore in depth discussion on  neural networks we recommend Goodfellow et al chapters 6 and 7. For CNNs, see Goodfellow et al chapter 9. chapter 11 and 12 on practicalities and applications
 
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec1">Setting up the Back propagation algorithm </h2>
+<h2 id="setting-up-the-back-propagation-algorithm">Setting up the Back propagation algorithm </h2>
 
 <p>
 The four equations  derived last week provide us with a way of computing the gradient of the cost function. Let us write this out in the form of an algorithm.
@@ -291,7 +346,7 @@ Here it is convenient to use stochastic gradient descent (see the examples below
 <p>
 <!-- !split  -->
 
-<h2 id="___sec2">Setting up a Multi-layer perceptron model for classification  </h2>
+<h2 id="setting-up-a-multi-layer-perceptron-model-for-classification">Setting up a Multi-layer perceptron model for classification  </h2>
 
 <p>
 We are now gong to develop an example based on the MNIST data
@@ -329,7 +384,7 @@ of our network.
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec3">Defining the cost function </h2>
+<h2 id="defining-the-cost-function">Defining the cost function </h2>
 
 <p>
 Our cost function is given as (see the Logistic regression lectures)
@@ -389,7 +444,7 @@ The back propagation equations need now only a small change, namely the definiti
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec4">Example: binary classification problem </h2>
+<h2 id="example-binary-classification-problem">Example: binary classification problem </h2>
 
 <p>
 As an example of the above, relevant for project 2 as well, let us consider a binary class. As discussed in our logistic regression lectures, we defined a cost function in terms of the parameters \( \beta \) as
@@ -437,7 +492,7 @@ In case we use another activation function than the logistic one, we need to eva
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec5">The Softmax function </h2>
+<h2 id="the-softmax-function">The Softmax function </h2>
 In case we employ the more general case given by the Softmax equation, we need to evaluate the derivative of the activation function with respect to the activation \( z_i^l \), that is we need
 $$
 \frac{\partial f(z_i^l)}{\partial w_{jk}^l} =
@@ -459,7 +514,7 @@ which in case of the simply binary model reduces to  having \( i=j \).
 <p>
 <!-- !split  -->
 
-<h2 id="___sec6">Developing a code for doing neural networks with back propagation </h2>
+<h2 id="developing-a-code-for-doing-neural-networks-with-back-propagation">Developing a code for doing neural networks with back propagation </h2>
 
 <p>
 One can identify a set of key steps when using neural networks to solve supervised learning problems:  
@@ -475,7 +530,7 @@ One can identify a set of key steps when using neural networks to solve supervis
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec7">Collect and pre-process data </h2>
+<h2 id="collect-and-pre-process-data">Collect and pre-process data </h2>
 
 <p>
 Here we will be using the MNIST dataset, which is readily available through the <b>scikit-learn</b>
@@ -524,7 +579,7 @@ images.
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%"><span></span><span style="color: #228B22"># import necessary packages</span>
+<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #228B22"># import necessary packages</span>
 <span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
 <span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
 <span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn</span> <span style="color: #8B008B; font-weight: bold">import</span> datasets
@@ -570,7 +625,7 @@ plt.show()
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec8">Train and test datasets </h2>
+<h2 id="train-and-test-datasets">Train and test datasets </h2>
 
 <p>
 Performing analysis before partitioning the dataset is a major error, that can lead to incorrect conclusions.
@@ -588,7 +643,7 @@ collected from 12.00 to 24.00.
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%"><span></span><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.model_selection</span> <span style="color: #8B008B; font-weight: bold">import</span> train_test_split
+<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.model_selection</span> <span style="color: #8B008B; font-weight: bold">import</span> train_test_split
 
 <span style="color: #228B22"># one-liner from scikit-learn library</span>
 train_size = <span style="color: #B452CD">0.8</span>
@@ -619,7 +674,7 @@ X_train, X_test, Y_train, Y_test = train_test_split(inputs, labels, train_size=t
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec9">Define model and architecture </h2>
+<h2 id="define-model-and-architecture">Define model and architecture </h2>
 
 <p>
 Our simple feed-forward neural network will consist of an <em>input</em> layer, a single <em>hidden</em> layer and an <em>output</em> layer. The activation \( y \) of each neuron is a weighted sum of inputs, passed through an activation function. In case of the simple perceptron model we have 
@@ -665,7 +720,7 @@ which is inspired by probability theory (see logistic regression) and was most c
 <p>
 <!-- !split   -->
 
-<h2 id="___sec10">Layers </h2>
+<h2 id="layers">Layers </h2>
 
 <ul>
 <li> Input</li> 
@@ -710,7 +765,7 @@ weights to the output layer.
 <p>
 <!-- !split   -->
 
-<h2 id="___sec11">Weights and biases </h2>
+<h2 id="weights-and-biases">Weights and biases </h2>
 
 <p>
 Typically weights are initialized with small values distributed around zero, drawn from a uniform
@@ -727,7 +782,7 @@ The bias weights \( \hat{b} \) are often initialized to zero, but a small value
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%"><span></span><span style="color: #228B22"># building our neural network</span>
+<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #228B22"># building our neural network</span>
 
 n_inputs, n_features = X_train.shape
 n_hidden_neurons = <span style="color: #B452CD">50</span>
@@ -746,7 +801,7 @@ output_bias = np.zeros(n_categories) + <span style="color: #B452CD">0.01</span>
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec12">Feed-forward pass </h2>
+<h2 id="feed-forward-pass">Feed-forward pass </h2>
 
 <p>
 Denote \( F \) the number of features, \( H \) the number of hidden neurons and \( C \) the number of categories.  
@@ -773,7 +828,7 @@ $$ a_{j}^{L} = \frac{\exp{(z_j^{L})}}
 <p>
 <!-- !split    -->
 
-<h2 id="___sec13">Matrix multiplications </h2>
+<h2 id="matrix-multiplications">Matrix multiplications </h2>
 
 <p>
 Since our data has the dimensions \( X = (n_{inputs}, n_{features}) \) and our weights to the hidden
@@ -809,7 +864,7 @@ $$ output = softmax (\hat{z}^{L}) = (n_{inputs}, n_{categories}) .$$
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%"><span></span><span style="color: #228B22"># setup the feed-forward pass, subscript h = hidden layer</span>
+<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #228B22"># setup the feed-forward pass, subscript h = hidden layer</span>
 
 <span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">sigmoid</span>(x):
     <span style="color: #8B008B; font-weight: bold">return</span> <span style="color: #B452CD">1</span>/(<span style="color: #B452CD">1</span> + np.exp(-x))
@@ -848,7 +903,7 @@ predictions = predict(X_train)
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec14">Choose cost function and optimizer </h2>
+<h2 id="choose-cost-function-and-optimizer">Choose cost function and optimizer </h2>
 
 <p>
 To measure how well our neural network is doing we need to introduce a cost function.  
@@ -879,7 +934,7 @@ you got the correct label. The probability of category \( c \) is given by the s
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec15">Optimizing the cost function </h2>
+<h2 id="optimizing-the-cost-function">Optimizing the cost function </h2>
 
 <p>
 The network is trained by finding the weights and biases that minimize the cost function. One of the most widely used classes of methods is <em>gradient descent</em> and its generalizations. The idea behind gradient descent
@@ -919,7 +974,7 @@ The various optmization  methods, with codes and algorithms,  are discussed in o
 <p>
 <!-- !split   -->
 
-<h2 id="___sec16">Regularization </h2>
+<h2 id="regularization">Regularization </h2>
 
 <p>
 It is common to add an extra term to the cost function, proportional
@@ -952,7 +1007,7 @@ calculate the gradient efficently.
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec17">Matrix  multiplication </h2>
+<h2 id="matrix-multiplication">Matrix  multiplication </h2>
 
 <p>
 To more efficently train our network these equations are implemented using matrix operations.  
@@ -994,7 +1049,7 @@ $$ \nabla b_{h} = \sum_{i=1}^{n_{inputs}} \delta_h = (n_{hidden}) .$$
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%"><span></span><span style="color: #228B22"># to categorical turns our integer vector into a onehot representation</span>
+<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #228B22"># to categorical turns our integer vector into a onehot representation</span>
 <span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.metrics</span> <span style="color: #8B008B; font-weight: bold">import</span> accuracy_score
 
 <span style="color: #228B22"># one-hot in numpy</span>
@@ -1066,7 +1121,7 @@ lmbd = <span style="color: #B452CD">0.01</span>
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec18">Improving performance </h2>
+<h2 id="improving-performance">Improving performance </h2>
 
 <p>
 As we can see the network does not seem to be learning at all. It seems to be just guessing the label for each image.  
@@ -1086,7 +1141,7 @@ Andrew Ng goes through some of these considerations in this <a href="https://you
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec19">Full object-oriented implementation </h2>
+<h2 id="full-object-oriented-implementation">Full object-oriented implementation </h2>
 
 <p>
 It is very natural to think of the network as an object, with specific instances of the network
@@ -1095,7 +1150,7 @@ being realizations of this object with different hyperparameters. An implementat
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%"><span></span><span style="color: #8B008B; font-weight: bold">class</span> <span style="color: #008b45; font-weight: bold">NeuralNetwork</span>:
+<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #8B008B; font-weight: bold">class</span> <span style="color: #008b45; font-weight: bold">NeuralNetwork</span>:
     <span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">__init__</span>(
             <span style="color: #658b00">self</span>,
             X_data,
@@ -1198,7 +1253,7 @@ being realizations of this object with different hyperparameters. An implementat
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec20">Evaluate model performance on test data </h2>
+<h2 id="evaluate-model-performance-on-test-data">Evaluate model performance on test data </h2>
 
 <p>
 To measure the performance of our network we evaluate how well it does it data it has never seen before, i.e. the test data.  
@@ -1213,7 +1268,7 @@ where \( I \) is the indicator function, \( 1 \) if \( \hat{y}_i = y_i \) and \(
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%"><span></span>epochs = <span style="color: #B452CD">100</span>
+<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span>epochs = <span style="color: #B452CD">100</span>
 batch_size = <span style="color: #B452CD">100</span>
 
 dnn = NeuralNetwork(X_train, Y_train_onehot, eta=eta, lmbd=lmbd, epochs=epochs, batch_size=batch_size,
@@ -1233,7 +1288,7 @@ test_predict = dnn.predict(X_test)
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec21">Adjust hyperparameters </h2>
+<h2 id="adjust-hyperparameters">Adjust hyperparameters </h2>
 
 <p>
 We now perform a grid search to find the optimal hyperparameters for the network.  
@@ -1242,7 +1297,7 @@ Note that we are only using 1 layer with 50 neurons, and human performance is es
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%"><span></span>eta_vals = np.logspace(-<span style="color: #B452CD">5</span>, <span style="color: #B452CD">1</span>, <span style="color: #B452CD">7</span>)
+<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span>eta_vals = np.logspace(-<span style="color: #B452CD">5</span>, <span style="color: #B452CD">1</span>, <span style="color: #B452CD">7</span>)
 lmbd_vals = np.logspace(-<span style="color: #B452CD">5</span>, <span style="color: #B452CD">1</span>, <span style="color: #B452CD">7</span>)
 <span style="color: #228B22"># store the models for later use</span>
 DNN_numpy = np.zeros((<span style="color: #658b00">len</span>(eta_vals), <span style="color: #658b00">len</span>(lmbd_vals)), dtype=<span style="color: #658b00">object</span>)
@@ -1266,12 +1321,12 @@ DNN_numpy = np.zeros((<span style="color: #658b00">len</span>(eta_vals), <span s
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec22">Visualization </h2>
+<h2 id="visualization">Visualization </h2>
 
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%"><span></span><span style="color: #228B22"># visual representation of grid search</span>
+<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #228B22"># visual representation of grid search</span>
 <span style="color: #228B22"># uses seaborn heatmap, you can also do this with matplotlib imshow</span>
 <span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">seaborn</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">sns</span>
 
@@ -1308,7 +1363,7 @@ plt.show()
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec23">scikit-learn implementation </h2>
+<h2 id="scikit-learn-implementation">scikit-learn implementation </h2>
 
 <p>
 <b>scikit-learn</b> focuses more
@@ -1327,7 +1382,7 @@ performance overall.
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%"><span></span><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.neural_network</span> <span style="color: #8B008B; font-weight: bold">import</span> MLPClassifier
+<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn.neural_network</span> <span style="color: #8B008B; font-weight: bold">import</span> MLPClassifier
 <span style="color: #228B22"># store models for later use</span>
 DNN_scikit = np.zeros((<span style="color: #658b00">len</span>(eta_vals), <span style="color: #658b00">len</span>(lmbd_vals)), dtype=<span style="color: #658b00">object</span>)
 
@@ -1347,11 +1402,11 @@ DNN_scikit = np.zeros((<span style="color: #658b00">len</span>(eta_vals), <span
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec24">Visualization </h2>
+<h2 id="visualization">Visualization </h2>
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%"><span></span><span style="color: #228B22"># optional</span>
+<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #228B22"># optional</span>
 <span style="color: #228B22"># visual representation of grid search</span>
 <span style="color: #228B22"># uses seaborn heatmap, could probably do this in matplotlib</span>
 <span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">seaborn</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">sns</span>
@@ -1389,7 +1444,7 @@ plt.show()
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec25">Building neural networks in Tensorflow and Keras </h2>
+<h2 id="building-neural-networks-in-tensorflow-and-keras">Building neural networks in Tensorflow and Keras </h2>
 
 <p>
 Now we want  to build on the experience gained from our neural network implementation in NumPy and scikit-learn
@@ -1404,7 +1459,7 @@ NumPy arrays.
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec26">Tensorflow </h2>
+<h2 id="tensorflow">Tensorflow </h2>
 
 <p>
 Tensorflow is an open source library machine learning library
@@ -1437,7 +1492,7 @@ To install tensorflow on Unix/Linux systems, use pip as
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%"><span></span>pip3 install tensorflow
+<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span>pip3 install tensorflow
 </pre></div>
 <p>
 and/or if you use <b>anaconda</b>, just write (or install from the graphical user interface)
@@ -1445,7 +1500,7 @@ and/or if you use <b>anaconda</b>, just write (or install from the graphical use
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%"><span></span>conda create -n tf tensorflow
+<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span>conda create -n tf tensorflow
 conda activate tf
 </pre></div>
 <p>
@@ -1453,13 +1508,13 @@ To install the current release of GPU TensorFlow
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%"><span></span>conda create -n tf-gpu tensorflow-gpu
+<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span>conda create -n tf-gpu tensorflow-gpu
 conda activate tf-gpu
 </pre></div>
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec27">Using Keras </h2>
+<h2 id="using-keras">Using Keras </h2>
 
 <p>
 Keras is a high level <a href="https://en.wikipedia.org/wiki/Application_programming_interface" target="_blank">neural network</a>
@@ -1468,7 +1523,7 @@ If you have Anaconda installed you may run the following command
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%"><span></span>conda install keras
+<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span>conda install keras
 </pre></div>
 <p>
 You can look up the <a href="https://keras.io/" target="_blank">instructions here</a> for more information.
@@ -1479,7 +1534,7 @@ We will to a large extent use <b>keras</b> in this course.
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec28">Collect and pre-process data </h2>
+<h2 id="collect-and-pre-process-data">Collect and pre-process data </h2>
 
 <p>
 Let us look again at the MINST data set.
@@ -1487,7 +1542,7 @@ Let us look again at the MINST data set.
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%"><span></span><span style="color: #228B22"># import necessary packages</span>
+<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #228B22"># import necessary packages</span>
 <span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
 <span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
 <span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">tensorflow</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">tf</span>
@@ -1534,7 +1589,7 @@ plt.show()
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%"><span></span><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">tensorflow.keras.layers</span> <span style="color: #8B008B; font-weight: bold">import</span> Input
+<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">tensorflow.keras.layers</span> <span style="color: #8B008B; font-weight: bold">import</span> Input
 <span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">tensorflow.keras.models</span> <span style="color: #8B008B; font-weight: bold">import</span> Sequential      <span style="color: #228B22">#This allows appending layers to existing models</span>
 <span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">tensorflow.keras.layers</span> <span style="color: #8B008B; font-weight: bold">import</span> Dense           <span style="color: #228B22">#This allows defining the characteristics of a particular layer</span>
 <span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">tensorflow.keras</span> <span style="color: #8B008B; font-weight: bold">import</span> optimizers             <span style="color: #228B22">#This allows using whichever optimiser we want (sgd,adam,RMSprop)</span>
@@ -1555,7 +1610,7 @@ X_train, X_test, Y_train, Y_test = train_test_split(inputs, labels, train_size=t
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%"><span></span>epochs = <span style="color: #B452CD">100</span>
+<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span>epochs = <span style="color: #B452CD">100</span>
 batch_size = <span style="color: #B452CD">100</span>
 n_neurons_layer1 = <span style="color: #B452CD">100</span>
 n_neurons_layer2 = <span style="color: #B452CD">50</span>
@@ -1576,7 +1631,7 @@ lmbd_vals = np.logspace(-<span style="color: #B452CD">5</span>, <span style="col
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%"><span></span>DNN_keras = np.zeros((<span style="color: #658b00">len</span>(eta_vals), <span style="color: #658b00">len</span>(lmbd_vals)), dtype=<span style="color: #658b00">object</span>)
+<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span>DNN_keras = np.zeros((<span style="color: #658b00">len</span>(eta_vals), <span style="color: #658b00">len</span>(lmbd_vals)), dtype=<span style="color: #658b00">object</span>)
         
 <span style="color: #8B008B; font-weight: bold">for</span> i, eta <span style="color: #8B008B">in</span> <span style="color: #658b00">enumerate</span>(eta_vals):
     <span style="color: #8B008B; font-weight: bold">for</span> j, lmbd <span style="color: #8B008B">in</span> <span style="color: #658b00">enumerate</span>(lmbd_vals):
@@ -1595,7 +1650,7 @@ lmbd_vals = np.logspace(-<span style="color: #B452CD">5</span>, <span style="col
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%"><span></span><span style="color: #228B22"># optional</span>
+<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #228B22"># optional</span>
 <span style="color: #228B22"># visual representation of grid search</span>
 <span style="color: #228B22"># uses seaborn heatmap, could probably do this in matplotlib</span>
 <span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">seaborn</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">sns</span>
@@ -1630,12 +1685,12 @@ plt.show()
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec29">The Breast Cancer Data, now with Keras </h2>
+<h2 id="the-breast-cancer-data-now-with-keras">The Breast Cancer Data, now with Keras </h2>
 
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%"><span></span><span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">tensorflow</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">tf</span>
+<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">tensorflow</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">tf</span>
 <span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">tensorflow.keras.layers</span> <span style="color: #8B008B; font-weight: bold">import</span> Input
 <span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">tensorflow.keras.models</span> <span style="color: #8B008B; font-weight: bold">import</span> Sequential      <span style="color: #228B22">#This allows appending layers to existing models</span>
 <span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">tensorflow.keras.layers</span> <span style="color: #8B008B; font-weight: bold">import</span> Dense           <span style="color: #228B22">#This allows defining the characteristics of a particular layer</span>
@@ -1804,7 +1859,7 @@ plot_data(eta,n_neuron,Test_accuracy, <span style="color: #CD5555">&#39;testing&
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec30">Fine-tuning neural network hyperparameters </h2>
+<h2 id="fine-tuning-neural-network-hyperparameters">Fine-tuning neural network hyperparameters </h2>
 
 <p>
 The flexibility of neural networks is also one of their main
@@ -1830,7 +1885,7 @@ will only be able to explore a tiny part of the hyperparameter space.
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec31">Hidden layers </h2>
+<h2 id="hidden-layers">Hidden layers </h2>
 
 <p>
 For many problems you can start with just one or two hidden layers and it will work just fine.
@@ -1850,7 +1905,7 @@ common to reuse parts of a pretrained state-of-the-art network that performs a s
 <p>
 <!-- !split  -->
 
-<h2 id="___sec32">Which activation function should I use? </h2>
+<h2 id="which-activation-function-should-i-use">Which activation function should I use? </h2>
 
 <p>
 The Back propagation algorithm we derived above works by going from
@@ -1879,7 +1934,7 @@ learn at widely different speeds
 <p>
 <!-- !split  -->
 
-<h2 id="___sec33">Is the Logistic activation function (Sigmoid)  our choice? </h2>
+<h2 id="is-the-logistic-activation-function-sigmoid-our-choice">Is the Logistic activation function (Sigmoid)  our choice? </h2>
 
 <p>
 Although this unfortunate behavior has been empirically observed for
@@ -1909,7 +1964,7 @@ better than the logistic function in deep networks).
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec34">The derivative of the Logistic funtion </h2>
+<h2 id="the-derivative-of-the-logistic-funtion">The derivative of the Logistic funtion </h2>
 
 <p>
 Looking at the logistic activation function, when inputs become large
@@ -1945,7 +2000,7 @@ fast to compute).
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec35">The RELU function family </h2>
+<h2 id="the-relu-function-family">The RELU function family </h2>
 
 <p>
 The ReLU activation function suffers from a problem known as the dying
@@ -1972,7 +2027,7 @@ $$
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec36">Which activation function should we use? </h2>
+<h2 id="which-activation-function-should-we-use">Which activation function should we use? </h2>
 
 <p>
 In general it seems that the ELU activation function is better than
@@ -1992,7 +2047,7 @@ bootstrap to evaluate other activation functions.
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec37">More on activation functions, output layers </h2>
+<h2 id="more-on-activation-functions-output-layers">More on activation functions, output layers </h2>
 
 <p>
 In most cases you can use the ReLU activation function in the hidden layers (or one of its variants).
@@ -2010,7 +2065,7 @@ It is a bit faster to compute than other activation functions, and the gradient
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec38">Batch Normalization </h2>
+<h2 id="batch-normalization">Batch Normalization </h2>
 
 <p>
 Batch Normalization
@@ -2029,7 +2084,7 @@ mini-batch, from this the name batch normalization.
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec39">Dropout </h2>
+<h2 id="dropout">Dropout </h2>
 
 <p>
 It is a fairly simple algorithm: at every training step, every neuron (including the input neurons but
@@ -2044,7 +2099,7 @@ hyperparameter \( p \) is called the dropout rate, and it is typically set to 50
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec40">Gradient Clipping </h2>
+<h2 id="gradient-clipping">Gradient Clipping </h2>
 
 <p>
 A popular technique to lessen the exploding gradients problem is to simply clip the gradients during
@@ -2061,7 +2116,7 @@ Normalization is preferred.
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec41">A very nice website on Neural Networks </h2>
+<h2 id="a-very-nice-website-on-neural-networks">A very nice website on Neural Networks </h2>
 
 <p>
 You may find this <a href="https://playground.tensorflow.org/#activation=tanh&batchSize=10&dataset=circle&regDataset=reg-plane&learningRate=0.03&regularizationRate=0&noise=0&networkShape=4,2&seed=0.29243&showTestData=false&discretize=false&percTrainData=50&x=true&y=true&xTimesY=false&xSquared=false&ySquared=false&cosX=false&sinX=false&cosY=false&sinY=false&collectStats=false&problem=classification&initZero=false&hideText=false" target="_blank">website</a> very useful. Thx a million to Ghadi for sharing.
@@ -2069,7 +2124,7 @@ You may find this <a href="https://playground.tensorflow.org/#activation=tanh&ba
 <p>
 <!-- !split  -->
 
-<h2 id="___sec42">A top-down perspective on Neural networks </h2>
+<h2 id="a-top-down-perspective-on-neural-networks">A top-down perspective on Neural networks </h2>
 
 <p>
 The first thing we would like to do is divide the data into two or three
@@ -2111,7 +2166,7 @@ supervised learning.
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec43">Limitations of supervised learning with deep networks </h2>
+<h2 id="limitations-of-supervised-learning-with-deep-networks">Limitations of supervised learning with deep networks </h2>
 
 <p>
 Like all statistical methods, supervised learning using neural
@@ -2137,7 +2192,7 @@ Some of these remarks are particular to DNNs, others are shared by all supervise
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec44">Convolutional Neural Networks (recognizing images) </h2>
+<h2 id="convolutional-neural-networks-recognizing-images">Convolutional Neural Networks (recognizing images) </h2>
 
 <p>
 Convolutional neural networks (CNNs) were developed during the last
@@ -2176,7 +2231,7 @@ Another good read is the article here <a href="https://arxiv.org/pdf/1603.07285.
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec45">Regular NNs don&#8217;t scale well to full images </h2>
+<h2 id="regular-nns-don-t-scale-well-to-full-images">Regular NNs don&#8217;t scale well to full images </h2>
 
 <p>
 As an example, consider
@@ -2204,7 +2259,7 @@ would quickly lead to possible overfitting.
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec46">3D volumes of neurons </h2>
+<h2 id="3d-volumes-of-neurons">3D volumes of neurons </h2>
 
 <p>
 Convolutional Neural Networks take advantage of the fact that the
@@ -2244,7 +2299,7 @@ dimension.
 <p>
 <!-- !split  -->
 
-<h2 id="___sec47">Layers used to build CNNs </h2>
+<h2 id="layers-used-to-build-cnns">Layers used to build CNNs </h2>
 
 <p>
 A simple CNN is a sequence of layers, and every layer of a CNN
@@ -2267,7 +2322,7 @@ A simple CNN for image classification could have the architecture:
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec48">Transforming images </h2>
+<h2 id="transforming-images">Transforming images </h2>
 
 <p>
 CNNs transform the original image layer by layer from the original
@@ -2286,7 +2341,7 @@ are consistent with the labels in the training set for each image.
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec49">CNNs in brief </h2>
+<h2 id="cnns-in-brief">CNNs in brief </h2>
 
 <p>
 In summary:
@@ -2307,7 +2362,7 @@ and the slides of <a href="http://cs231n.github.io/convolutional-networks/" targ
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec50">CNNs in more detail, building convolutional neural networks in Tensorflow and Keras </h2>
+<h2 id="cnns-in-more-detail-building-convolutional-neural-networks-in-tensorflow-and-keras">CNNs in more detail, building convolutional neural networks in Tensorflow and Keras </h2>
 
 <p>
 As discussed above, CNNs are neural networks built from the assumption that the inputs
@@ -2323,7 +2378,7 @@ matrices, typically 1 for each color dimension (Red, Green, Blue).
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec51">Setting it up </h2>
+<h2 id="setting-it-up">Setting it up </h2>
 
 <p>
 It means that to represent the entire
@@ -2335,7 +2390,7 @@ $$
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec52">The MNIST dataset again </h2>
+<h2 id="the-mnist-dataset-again">The MNIST dataset again </h2>
 
 <p>
 The MNIST dataset consists of grayscale images with a pixel size of
@@ -2353,7 +2408,7 @@ single neuron in the first hidden layer.
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec53">Strong correlations </h2>
+<h2 id="strong-correlations">Strong correlations </h2>
 
 <p>
 Images typically have strong local correlations, meaning that a small
@@ -2371,7 +2426,7 @@ fixed, and known as a <a href="https://en.wikipedia.org/wiki/Receptive_field" ta
 <p>
 <!-- !split  -->
 
-<h2 id="___sec54">Layers of a CNN </h2>
+<h2 id="layers-of-a-cnn">Layers of a CNN </h2>
 The layers of a convolutional neural network arrange neurons in 3D: width, height and depth.  
 The input image is typically a square matrix of depth 3.
 
@@ -2393,7 +2448,7 @@ layer.
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec55">Systematic reduction </h2>
+<h2 id="systematic-reduction">Systematic reduction </h2>
 
 <p>
 By systematically reducing the size of the input volume, through
@@ -2408,11 +2463,11 @@ classification.
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec56">Prerequisites: Collect and pre-process data </h2>
+<h2 id="prerequisites-collect-and-pre-process-data">Prerequisites: Collect and pre-process data </h2>
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%"><span></span><span style="color: #228B22"># import necessary packages</span>
+<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #228B22"># import necessary packages</span>
 <span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">numpy</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">np</span>
 <span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">matplotlib.pyplot</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">plt</span>
 <span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">sklearn</span> <span style="color: #8B008B; font-weight: bold">import</span> datasets
@@ -2456,11 +2511,11 @@ plt.show()
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec57">Importing Keras and Tensorflow </h2>
+<h2 id="importing-keras-and-tensorflow">Importing Keras and Tensorflow </h2>
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%"><span></span><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">tensorflow.keras</span> <span style="color: #8B008B; font-weight: bold">import</span> datasets, layers, models
+<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">tensorflow.keras</span> <span style="color: #8B008B; font-weight: bold">import</span> datasets, layers, models
 <span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">tensorflow.keras.layers</span> <span style="color: #8B008B; font-weight: bold">import</span> Input
 <span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">tensorflow.keras.models</span> <span style="color: #8B008B; font-weight: bold">import</span> Sequential      <span style="color: #228B22">#This allows appending layers to existing models</span>
 <span style="color: #8B008B; font-weight: bold">from</span> <span style="color: #008b45; text-decoration: underline">tensorflow.keras.layers</span> <span style="color: #8B008B; font-weight: bold">import</span> Dense           <span style="color: #228B22">#This allows defining the characteristics of a particular layer</span>
@@ -2486,12 +2541,12 @@ X_train, X_test, Y_train, Y_test = train_test_split(inputs, labels, train_size=t
 <p>
 <!-- !split  -->
 
-<h2 id="___sec58">Running with Keras </h2>
+<h2 id="running-with-keras">Running with Keras </h2>
 
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%"><span></span><span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">create_convolutional_neural_network_keras</span>(input_shape, receptive_field,
+<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #8B008B; font-weight: bold">def</span> <span style="color: #008b45">create_convolutional_neural_network_keras</span>(input_shape, receptive_field,
                                               n_filters, n_neurons_connected, n_categories,
                                               eta, lmbd):
     model = Sequential()
@@ -2521,12 +2576,12 @@ lmbd_vals = np.logspace(-<span style="color: #B452CD">5</span>, <span style="col
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec59">Final part </h2>
+<h2 id="final-part">Final part </h2>
 
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%"><span></span>CNN_keras = np.zeros((<span style="color: #658b00">len</span>(eta_vals), <span style="color: #658b00">len</span>(lmbd_vals)), dtype=<span style="color: #658b00">object</span>)
+<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span>CNN_keras = np.zeros((<span style="color: #658b00">len</span>(eta_vals), <span style="color: #658b00">len</span>(lmbd_vals)), dtype=<span style="color: #658b00">object</span>)
         
 <span style="color: #8B008B; font-weight: bold">for</span> i, eta <span style="color: #8B008B">in</span> <span style="color: #658b00">enumerate</span>(eta_vals):
     <span style="color: #8B008B; font-weight: bold">for</span> j, lmbd <span style="color: #8B008B">in</span> <span style="color: #658b00">enumerate</span>(lmbd_vals):
@@ -2546,12 +2601,12 @@ lmbd_vals = np.logspace(-<span style="color: #B452CD">5</span>, <span style="col
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec60">Final visualization </h2>
+<h2 id="final-visualization">Final visualization </h2>
 
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "perldoc" -->
-<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%"><span></span><span style="color: #228B22"># visual representation of grid search</span>
+<div class="highlight" style="background: #eeeedd"><pre style="line-height: 125%;"><span></span><span style="color: #228B22"># visual representation of grid search</span>
 <span style="color: #228B22"># uses seaborn heatmap, could probably do this in matplotlib</span>
 <span style="color: #8B008B; font-weight: bold">import</span> <span style="color: #008b45; text-decoration: underline">seaborn</span> <span style="color: #8B008B; font-weight: bold">as</span> <span style="color: #008b45; text-decoration: underline">sns</span>
 
@@ -2585,7 +2640,7 @@ plt.show()
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec61">Fun links </h2>
+<h2 id="fun-links">Fun links </h2>
 
 <ol>
 <li> <a href="https://arxiv.org/abs/1604.07316" target="_blank">Self-Driving cars using a convolutional neural network</a></li>
@@ -2597,7 +2652,7 @@ plt.show()
 
 
 <center style="font-size:80%">
-<!-- copyright --> &copy; 1999-2020, Morten Hjorth-Jensen. Released under CC Attribution-NonCommercial 4.0 license
+<!-- copyright --> &copy; 1999-2021, Morten Hjorth-Jensen. Released under CC Attribution-NonCommercial 4.0 license
 </center>
 
 
diff --git a/doc/pub/week41/html/week41.html b/doc/pub/week41/html/week41.html
index bfaad0323..49e0c5776 100644
--- a/doc/pub/week41/html/week41.html
+++ b/doc/pub/week41/html/week41.html
@@ -1,15 +1,15 @@
 <!--
 Automatically generated HTML file from DocOnce source
-(https://github.com/hplgit/doconce/)
+(https://github.com/doconce/doconce/)
 -->
 <html>
 <head>
 <meta http-equiv="Content-Type" content="text/html; charset=utf-8" />
-<meta name="generator" content="DocOnce: https://github.com/hplgit/doconce/" />
+<meta name="generator" content="DocOnce: https://github.com/doconce/doconce/" />
 <meta name="viewport" content="width=device-width, initial-scale=1.0" />
-<meta name="description" content="Week 41 Tensor flow and Deep Learning, Convolutional Neural Networks">
+<meta name="description" content="Week 41 Constructing a Neural Network code, Tensor flow and start Convolutional Neural Networks">
 
-<title>Week 41 Tensor flow and Deep Learning, Convolutional Neural Networks</title>
+<title>Week 41 Constructing a Neural Network code, Tensor flow and start Convolutional Neural Networks</title>
 
 
 <style type="text/css">
@@ -53,10 +53,10 @@ hr.figure { border: 0; width: 80%; border-bottom: 1px solid #aaa}
 .alert-block > p, .alert-block > ul {margin-bottom:1em}
 .alert li {margin-top: 1em}
 .alert-block p+p {margin-top:5px}
-.alert-notice { background-image: url(https://cdn.rawgit.com/hplgit/doconce/master/bundled/html_images/small_gray_notice.png); }
-.alert-summary  { background-image:url(https://cdn.rawgit.com/hplgit/doconce/master/bundled/html_images/small_gray_summary.png); }
-.alert-warning { background-image: url(https://cdn.rawgit.com/hplgit/doconce/master/bundled/html_images/small_gray_warning.png); }
-.alert-question {background-image:url(https://cdn.rawgit.com/hplgit/doconce/master/bundled/html_images/small_gray_question.png); }
+.alert-notice { background-image: url(https://cdn.rawgit.com/doconce/doconce/master/bundled/html_images/small_gray_notice.png); }
+.alert-summary  { background-image:url(https://cdn.rawgit.com/doconce/doconce/master/bundled/html_images/small_gray_summary.png); }
+.alert-warning { background-image: url(https://cdn.rawgit.com/doconce/doconce/master/bundled/html_images/small_gray_warning.png); }
+.alert-question {background-image:url(https://cdn.rawgit.com/doconce/doconce/master/bundled/html_images/small_gray_question.png); }
 
 div { text-align: justify; text-justify: inter-word; }
 </style>
@@ -66,106 +66,160 @@ div { text-align: justify; text-justify: inter-word; }
 
 <!-- tocinfo
 {'highest level': 2,
- 'sections': [('Plan for week 41', 2, None, '___sec0'),
-              ('Setting up the Back propagation algorithm', 2, None, '___sec1'),
+ 'sections': [('Plan for week 41', 2, None, 'plan-for-week-41'),
+              ('Setting up the Back propagation algorithm',
+               2,
+               None,
+               'setting-up-the-back-propagation-algorithm'),
               ('Setting up a Multi-layer perceptron model for classification',
                2,
                None,
-               '___sec2'),
-              ('Defining the cost function', 2, None, '___sec3'),
-              ('Example: binary classification problem', 2, None, '___sec4'),
-              ('The Softmax function', 2, None, '___sec5'),
+               'setting-up-a-multi-layer-perceptron-model-for-classification'),
+              ('Defining the cost function',
+               2,
+               None,
+               'defining-the-cost-function'),
+              ('Example: binary classification problem',
+               2,
+               None,
+               'example-binary-classification-problem'),
+              ('The Softmax function', 2, None, 'the-softmax-function'),
               ('Developing a code for doing neural networks with back '
                'propagation',
                2,
                None,
-               '___sec6'),
-              ('Collect and pre-process data', 2, None, '___sec7'),
-              ('Train and test datasets', 2, None, '___sec8'),
-              ('Define model and architecture', 2, None, '___sec9'),
-              ('Layers', 2, None, '___sec10'),
-              ('Weights and biases', 2, None, '___sec11'),
-              ('Feed-forward pass', 2, None, '___sec12'),
-              ('Matrix multiplications', 2, None, '___sec13'),
-              ('Choose cost function and optimizer', 2, None, '___sec14'),
-              ('Optimizing the cost function', 2, None, '___sec15'),
-              ('Regularization', 2, None, '___sec16'),
-              ('Matrix  multiplication', 2, None, '___sec17'),
-              ('Improving performance', 2, None, '___sec18'),
-              ('Full object-oriented implementation', 2, None, '___sec19'),
-              ('Evaluate model performance on test data', 2, None, '___sec20'),
-              ('Adjust hyperparameters', 2, None, '___sec21'),
-              ('Visualization', 2, None, '___sec22'),
-              ('scikit-learn implementation', 2, None, '___sec23'),
-              ('Visualization', 2, None, '___sec24'),
+               'developing-a-code-for-doing-neural-networks-with-back-propagation'),
+              ('Collect and pre-process data',
+               2,
+               None,
+               'collect-and-pre-process-data'),
+              ('Train and test datasets', 2, None, 'train-and-test-datasets'),
+              ('Define model and architecture',
+               2,
+               None,
+               'define-model-and-architecture'),
+              ('Layers', 2, None, 'layers'),
+              ('Weights and biases', 2, None, 'weights-and-biases'),
+              ('Feed-forward pass', 2, None, 'feed-forward-pass'),
+              ('Matrix multiplications', 2, None, 'matrix-multiplications'),
+              ('Choose cost function and optimizer',
+               2,
+               None,
+               'choose-cost-function-and-optimizer'),
+              ('Optimizing the cost function',
+               2,
+               None,
+               'optimizing-the-cost-function'),
+              ('Regularization', 2, None, 'regularization'),
+              ('Matrix  multiplication', 2, None, 'matrix-multiplication'),
+              ('Improving performance', 2, None, 'improving-performance'),
+              ('Full object-oriented implementation',
+               2,
+               None,
+               'full-object-oriented-implementation'),
+              ('Evaluate model performance on test data',
+               2,
+               None,
+               'evaluate-model-performance-on-test-data'),
+              ('Adjust hyperparameters', 2, None, 'adjust-hyperparameters'),
+              ('Visualization', 2, None, 'visualization'),
+              ('scikit-learn implementation',
+               2,
+               None,
+               'scikit-learn-implementation'),
+              ('Visualization', 2, None, 'visualization'),
               ('Building neural networks in Tensorflow and Keras',
                2,
                None,
-               '___sec25'),
-              ('Tensorflow', 2, None, '___sec26'),
-              ('Using Keras', 2, None, '___sec27'),
-              ('Collect and pre-process data', 2, None, '___sec28'),
-              ('The Breast Cancer Data, now with Keras', 2, None, '___sec29'),
+               'building-neural-networks-in-tensorflow-and-keras'),
+              ('Tensorflow', 2, None, 'tensorflow'),
+              ('Using Keras', 2, None, 'using-keras'),
+              ('Collect and pre-process data',
+               2,
+               None,
+               'collect-and-pre-process-data'),
+              ('The Breast Cancer Data, now with Keras',
+               2,
+               None,
+               'the-breast-cancer-data-now-with-keras'),
               ('Fine-tuning neural network hyperparameters',
                2,
                None,
-               '___sec30'),
-              ('Hidden layers', 2, None, '___sec31'),
-              ('Which activation function should I use?', 2, None, '___sec32'),
+               'fine-tuning-neural-network-hyperparameters'),
+              ('Hidden layers', 2, None, 'hidden-layers'),
+              ('Which activation function should I use?',
+               2,
+               None,
+               'which-activation-function-should-i-use'),
               ('Is the Logistic activation function (Sigmoid)  our choice?',
                2,
                None,
-               '___sec33'),
-              ('The derivative of the Logistic funtion', 2, None, '___sec34'),
-              ('The RELU function family', 2, None, '___sec35'),
-              ('Which activation function should we use?', 2, None, '___sec36'),
+               'is-the-logistic-activation-function-sigmoid-our-choice'),
+              ('The derivative of the Logistic funtion',
+               2,
+               None,
+               'the-derivative-of-the-logistic-funtion'),
+              ('The RELU function family', 2, None, 'the-relu-function-family'),
+              ('Which activation function should we use?',
+               2,
+               None,
+               'which-activation-function-should-we-use'),
               ('More on activation functions, output layers',
                2,
                None,
-               '___sec37'),
-              ('Batch Normalization', 2, None, '___sec38'),
-              ('Dropout', 2, None, '___sec39'),
-              ('Gradient Clipping', 2, None, '___sec40'),
-              ('A very nice website on Neural Networks', 2, None, '___sec41'),
+               'more-on-activation-functions-output-layers'),
+              ('Batch Normalization', 2, None, 'batch-normalization'),
+              ('Dropout', 2, None, 'dropout'),
+              ('Gradient Clipping', 2, None, 'gradient-clipping'),
+              ('A very nice website on Neural Networks',
+               2,
+               None,
+               'a-very-nice-website-on-neural-networks'),
               ('A top-down perspective on Neural networks',
                2,
                None,
-               '___sec42'),
+               'a-top-down-perspective-on-neural-networks'),
               ('Limitations of supervised learning with deep networks',
                2,
                None,
-               '___sec43'),
+               'limitations-of-supervised-learning-with-deep-networks'),
               ('Convolutional Neural Networks (recognizing images)',
                2,
                None,
-               '___sec44'),
+               'convolutional-neural-networks-recognizing-images'),
               ('Regular NNs don’t scale well to full images',
                2,
                None,
-               '___sec45'),
-              ('3D volumes of neurons', 2, None, '___sec46'),
-              ('Layers used to build CNNs', 2, None, '___sec47'),
-              ('Transforming images', 2, None, '___sec48'),
-              ('CNNs in brief', 2, None, '___sec49'),
+               'regular-nns-don-t-scale-well-to-full-images'),
+              ('3D volumes of neurons', 2, None, '3d-volumes-of-neurons'),
+              ('Layers used to build CNNs',
+               2,
+               None,
+               'layers-used-to-build-cnns'),
+              ('Transforming images', 2, None, 'transforming-images'),
+              ('CNNs in brief', 2, None, 'cnns-in-brief'),
               ('CNNs in more detail, building convolutional neural networks in '
                'Tensorflow and Keras',
                2,
                None,
-               '___sec50'),
-              ('Setting it up', 2, None, '___sec51'),
-              ('The MNIST dataset again', 2, None, '___sec52'),
-              ('Strong correlations', 2, None, '___sec53'),
-              ('Layers of a CNN', 2, None, '___sec54'),
-              ('Systematic reduction', 2, None, '___sec55'),
+               'cnns-in-more-detail-building-convolutional-neural-networks-in-tensorflow-and-keras'),
+              ('Setting it up', 2, None, 'setting-it-up'),
+              ('The MNIST dataset again', 2, None, 'the-mnist-dataset-again'),
+              ('Strong correlations', 2, None, 'strong-correlations'),
+              ('Layers of a CNN', 2, None, 'layers-of-a-cnn'),
+              ('Systematic reduction', 2, None, 'systematic-reduction'),
               ('Prerequisites: Collect and pre-process data',
                2,
                None,
-               '___sec56'),
-              ('Importing Keras and Tensorflow', 2, None, '___sec57'),
-              ('Running with Keras', 2, None, '___sec58'),
-              ('Final part', 2, None, '___sec59'),
-              ('Final visualization', 2, None, '___sec60'),
-              ('Fun links', 2, None, '___sec61')]}
+               'prerequisites-collect-and-pre-process-data'),
+              ('Importing Keras and Tensorflow',
+               2,
+               None,
+               'importing-keras-and-tensorflow'),
+              ('Running with Keras', 2, None, 'running-with-keras'),
+              ('Final part', 2, None, 'final-part'),
+              ('Final visualization', 2, None, 'final-visualization'),
+              ('Fun links', 2, None, 'fun-links')]}
 end of tocinfo -->
 
 <body>
@@ -191,7 +245,7 @@ MathJax.Hub.Config({
 
 
 
-<center><h1>Week 41 Tensor flow and Deep Learning, Convolutional Neural Networks</h1></center>  <!-- document title -->
+<center><h1>Week 41 Constructing a Neural Network code, Tensor flow and start Convolutional Neural Networks</h1></center>  <!-- document title -->
 
 <p>
 <!-- author(s): Morten Hjorth-Jensen -->
@@ -207,25 +261,26 @@ MathJax.Hub.Config({
 <center>[2] <b>Department of Physics and Astronomy and National Superconducting Cyclotron Laboratory, Michigan State University</b></center>
 <br>
 <p>
-<center><h4>Oct 10, 2020</h4></center> <!-- date -->
+<center><h4>Oct 9, 2021</h4></center> <!-- date -->
 <br>
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec0">Plan for week 41 </h2>
+<h2 id="plan-for-week-41">Plan for week 41 </h2>
 
 <ul>
-<li> Thursday: Building our own Feed-forward Neural Network. <a href="https://www.uio.no/studier/emner/matnat/fys/FYS-STK4155/h20/forelesningsvideoer/LectureOctober8.mp4?vrtx=view-as-webpage" target="_blank">Video of Lecture</a></li>
-<li> Friday: Playing around with our own Feed-forward Neural Network and introduction to TensorFlow. Start convolutional Neural Networks (CNN). <a href="https://www.uio.no/studier/emner/matnat/fys/FYS-STK4155/h20/forelesningsvideoer/LectureOcotber9.mp4?vrtx=view-as-webpage" target="_blank">Video of Lecture</a></li>
+<li> Thursday: Building our own Feed-forward Neural Network.</li> 
+<li> Friday: Playing around with our own Feed-forward Neural Network and introduction to TensorFlow. Start convolutional Neural Networks (CNN).</li> 
 </ul>
 
 Reading suggestions for both days: <a href="https://github.com/CompPhysics/MachineLearning/blob/master/doc/T\
-extbooks/TensorflowML.pdf" target="_blank">Aurelien Geron's chapters 10-11</a> and Hastie et al chapter 11.
+extbooks/TensorflowML.pdf" target="_blank">Aurelien Geron's chapters 10-11</a>.
+For amore in depth discussion on  neural networks we recommend Goodfellow et al chapters 6 and 7. For CNNs, see Goodfellow et al chapter 9. chapter 11 and 12 on practicalities and applications
 
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec1">Setting up the Back propagation algorithm </h2>
+<h2 id="setting-up-the-back-propagation-algorithm">Setting up the Back propagation algorithm </h2>
 
 <p>
 The four equations  derived last week provide us with a way of computing the gradient of the cost function. Let us write this out in the form of an algorithm.
@@ -296,7 +351,7 @@ Here it is convenient to use stochastic gradient descent (see the examples below
 <p>
 <!-- !split  -->
 
-<h2 id="___sec2">Setting up a Multi-layer perceptron model for classification  </h2>
+<h2 id="setting-up-a-multi-layer-perceptron-model-for-classification">Setting up a Multi-layer perceptron model for classification  </h2>
 
 <p>
 We are now gong to develop an example based on the MNIST data
@@ -334,7 +389,7 @@ of our network.
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec3">Defining the cost function </h2>
+<h2 id="defining-the-cost-function">Defining the cost function </h2>
 
 <p>
 Our cost function is given as (see the Logistic regression lectures)
@@ -394,7 +449,7 @@ The back propagation equations need now only a small change, namely the definiti
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec4">Example: binary classification problem </h2>
+<h2 id="example-binary-classification-problem">Example: binary classification problem </h2>
 
 <p>
 As an example of the above, relevant for project 2 as well, let us consider a binary class. As discussed in our logistic regression lectures, we defined a cost function in terms of the parameters \( \beta \) as
@@ -442,7 +497,7 @@ In case we use another activation function than the logistic one, we need to eva
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec5">The Softmax function </h2>
+<h2 id="the-softmax-function">The Softmax function </h2>
 In case we employ the more general case given by the Softmax equation, we need to evaluate the derivative of the activation function with respect to the activation \( z_i^l \), that is we need
 $$
 \frac{\partial f(z_i^l)}{\partial w_{jk}^l} =
@@ -464,7 +519,7 @@ which in case of the simply binary model reduces to  having \( i=j \).
 <p>
 <!-- !split  -->
 
-<h2 id="___sec6">Developing a code for doing neural networks with back propagation </h2>
+<h2 id="developing-a-code-for-doing-neural-networks-with-back-propagation">Developing a code for doing neural networks with back propagation </h2>
 
 <p>
 One can identify a set of key steps when using neural networks to solve supervised learning problems:  
@@ -480,7 +535,7 @@ One can identify a set of key steps when using neural networks to solve supervis
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec7">Collect and pre-process data </h2>
+<h2 id="collect-and-pre-process-data">Collect and pre-process data </h2>
 
 <p>
 Here we will be using the MNIST dataset, which is readily available through the <b>scikit-learn</b>
@@ -529,7 +584,7 @@ images.
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%"><span></span><span style="color: #408080; font-style: italic"># import necessary packages</span>
+<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%;"><span></span><span style="color: #408080; font-style: italic"># import necessary packages</span>
 <span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">numpy</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">np</span>
 <span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">matplotlib.pyplot</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">plt</span>
 <span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn</span> <span style="color: #008000; font-weight: bold">import</span> datasets
@@ -575,7 +630,7 @@ plt<span style="color: #666666">.</span>show()
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec8">Train and test datasets </h2>
+<h2 id="train-and-test-datasets">Train and test datasets </h2>
 
 <p>
 Performing analysis before partitioning the dataset is a major error, that can lead to incorrect conclusions.
@@ -593,7 +648,7 @@ collected from 12.00 to 24.00.
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%"><span></span><span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.model_selection</span> <span style="color: #008000; font-weight: bold">import</span> train_test_split
+<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%;"><span></span><span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.model_selection</span> <span style="color: #008000; font-weight: bold">import</span> train_test_split
 
 <span style="color: #408080; font-style: italic"># one-liner from scikit-learn library</span>
 train_size <span style="color: #666666">=</span> <span style="color: #666666">0.8</span>
@@ -624,7 +679,7 @@ X_train, X_test, Y_train, Y_test <span style="color: #666666">=</span> train_tes
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec9">Define model and architecture </h2>
+<h2 id="define-model-and-architecture">Define model and architecture </h2>
 
 <p>
 Our simple feed-forward neural network will consist of an <em>input</em> layer, a single <em>hidden</em> layer and an <em>output</em> layer. The activation \( y \) of each neuron is a weighted sum of inputs, passed through an activation function. In case of the simple perceptron model we have 
@@ -670,7 +725,7 @@ which is inspired by probability theory (see logistic regression) and was most c
 <p>
 <!-- !split   -->
 
-<h2 id="___sec10">Layers </h2>
+<h2 id="layers">Layers </h2>
 
 <ul>
 <li> Input</li> 
@@ -715,7 +770,7 @@ weights to the output layer.
 <p>
 <!-- !split   -->
 
-<h2 id="___sec11">Weights and biases </h2>
+<h2 id="weights-and-biases">Weights and biases </h2>
 
 <p>
 Typically weights are initialized with small values distributed around zero, drawn from a uniform
@@ -732,7 +787,7 @@ The bias weights \( \hat{b} \) are often initialized to zero, but a small value
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%"><span></span><span style="color: #408080; font-style: italic"># building our neural network</span>
+<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%;"><span></span><span style="color: #408080; font-style: italic"># building our neural network</span>
 
 n_inputs, n_features <span style="color: #666666">=</span> X_train<span style="color: #666666">.</span>shape
 n_hidden_neurons <span style="color: #666666">=</span> <span style="color: #666666">50</span>
@@ -751,7 +806,7 @@ output_bias <span style="color: #666666">=</span> np<span style="color: #666666"
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec12">Feed-forward pass </h2>
+<h2 id="feed-forward-pass">Feed-forward pass </h2>
 
 <p>
 Denote \( F \) the number of features, \( H \) the number of hidden neurons and \( C \) the number of categories.  
@@ -778,7 +833,7 @@ $$ a_{j}^{L} = \frac{\exp{(z_j^{L})}}
 <p>
 <!-- !split    -->
 
-<h2 id="___sec13">Matrix multiplications </h2>
+<h2 id="matrix-multiplications">Matrix multiplications </h2>
 
 <p>
 Since our data has the dimensions \( X = (n_{inputs}, n_{features}) \) and our weights to the hidden
@@ -814,7 +869,7 @@ $$ output = softmax (\hat{z}^{L}) = (n_{inputs}, n_{categories}) .$$
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%"><span></span><span style="color: #408080; font-style: italic"># setup the feed-forward pass, subscript h = hidden layer</span>
+<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%;"><span></span><span style="color: #408080; font-style: italic"># setup the feed-forward pass, subscript h = hidden layer</span>
 
 <span style="color: #008000; font-weight: bold">def</span> <span style="color: #0000FF">sigmoid</span>(x):
     <span style="color: #008000; font-weight: bold">return</span> <span style="color: #666666">1/</span>(<span style="color: #666666">1</span> <span style="color: #666666">+</span> np<span style="color: #666666">.</span>exp(<span style="color: #666666">-</span>x))
@@ -853,7 +908,7 @@ predictions <span style="color: #666666">=</span> predict(X_train)
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec14">Choose cost function and optimizer </h2>
+<h2 id="choose-cost-function-and-optimizer">Choose cost function and optimizer </h2>
 
 <p>
 To measure how well our neural network is doing we need to introduce a cost function.  
@@ -884,7 +939,7 @@ you got the correct label. The probability of category \( c \) is given by the s
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec15">Optimizing the cost function </h2>
+<h2 id="optimizing-the-cost-function">Optimizing the cost function </h2>
 
 <p>
 The network is trained by finding the weights and biases that minimize the cost function. One of the most widely used classes of methods is <em>gradient descent</em> and its generalizations. The idea behind gradient descent
@@ -924,7 +979,7 @@ The various optmization  methods, with codes and algorithms,  are discussed in o
 <p>
 <!-- !split   -->
 
-<h2 id="___sec16">Regularization </h2>
+<h2 id="regularization">Regularization </h2>
 
 <p>
 It is common to add an extra term to the cost function, proportional
@@ -957,7 +1012,7 @@ calculate the gradient efficently.
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec17">Matrix  multiplication </h2>
+<h2 id="matrix-multiplication">Matrix  multiplication </h2>
 
 <p>
 To more efficently train our network these equations are implemented using matrix operations.  
@@ -999,7 +1054,7 @@ $$ \nabla b_{h} = \sum_{i=1}^{n_{inputs}} \delta_h = (n_{hidden}) .$$
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%"><span></span><span style="color: #408080; font-style: italic"># to categorical turns our integer vector into a onehot representation</span>
+<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%;"><span></span><span style="color: #408080; font-style: italic"># to categorical turns our integer vector into a onehot representation</span>
 <span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.metrics</span> <span style="color: #008000; font-weight: bold">import</span> accuracy_score
 
 <span style="color: #408080; font-style: italic"># one-hot in numpy</span>
@@ -1071,7 +1126,7 @@ lmbd <span style="color: #666666">=</span> <span style="color: #666666">0.01</sp
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec18">Improving performance </h2>
+<h2 id="improving-performance">Improving performance </h2>
 
 <p>
 As we can see the network does not seem to be learning at all. It seems to be just guessing the label for each image.  
@@ -1091,7 +1146,7 @@ Andrew Ng goes through some of these considerations in this <a href="https://you
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec19">Full object-oriented implementation </h2>
+<h2 id="full-object-oriented-implementation">Full object-oriented implementation </h2>
 
 <p>
 It is very natural to think of the network as an object, with specific instances of the network
@@ -1100,7 +1155,7 @@ being realizations of this object with different hyperparameters. An implementat
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%"><span></span><span style="color: #008000; font-weight: bold">class</span> <span style="color: #0000FF; font-weight: bold">NeuralNetwork</span>:
+<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%;"><span></span><span style="color: #008000; font-weight: bold">class</span> <span style="color: #0000FF; font-weight: bold">NeuralNetwork</span>:
     <span style="color: #008000; font-weight: bold">def</span> <span style="color: #0000FF">__init__</span>(
             <span style="color: #008000">self</span>,
             X_data,
@@ -1203,7 +1258,7 @@ being realizations of this object with different hyperparameters. An implementat
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec20">Evaluate model performance on test data </h2>
+<h2 id="evaluate-model-performance-on-test-data">Evaluate model performance on test data </h2>
 
 <p>
 To measure the performance of our network we evaluate how well it does it data it has never seen before, i.e. the test data.  
@@ -1218,7 +1273,7 @@ where \( I \) is the indicator function, \( 1 \) if \( \hat{y}_i = y_i \) and \(
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%"><span></span>epochs <span style="color: #666666">=</span> <span style="color: #666666">100</span>
+<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%;"><span></span>epochs <span style="color: #666666">=</span> <span style="color: #666666">100</span>
 batch_size <span style="color: #666666">=</span> <span style="color: #666666">100</span>
 
 dnn <span style="color: #666666">=</span> NeuralNetwork(X_train, Y_train_onehot, eta<span style="color: #666666">=</span>eta, lmbd<span style="color: #666666">=</span>lmbd, epochs<span style="color: #666666">=</span>epochs, batch_size<span style="color: #666666">=</span>batch_size,
@@ -1238,7 +1293,7 @@ test_predict <span style="color: #666666">=</span> dnn<span style="color: #66666
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec21">Adjust hyperparameters </h2>
+<h2 id="adjust-hyperparameters">Adjust hyperparameters </h2>
 
 <p>
 We now perform a grid search to find the optimal hyperparameters for the network.  
@@ -1247,7 +1302,7 @@ Note that we are only using 1 layer with 50 neurons, and human performance is es
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%"><span></span>eta_vals <span style="color: #666666">=</span> np<span style="color: #666666">.</span>logspace(<span style="color: #666666">-5</span>, <span style="color: #666666">1</span>, <span style="color: #666666">7</span>)
+<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%;"><span></span>eta_vals <span style="color: #666666">=</span> np<span style="color: #666666">.</span>logspace(<span style="color: #666666">-5</span>, <span style="color: #666666">1</span>, <span style="color: #666666">7</span>)
 lmbd_vals <span style="color: #666666">=</span> np<span style="color: #666666">.</span>logspace(<span style="color: #666666">-5</span>, <span style="color: #666666">1</span>, <span style="color: #666666">7</span>)
 <span style="color: #408080; font-style: italic"># store the models for later use</span>
 DNN_numpy <span style="color: #666666">=</span> np<span style="color: #666666">.</span>zeros((<span style="color: #008000">len</span>(eta_vals), <span style="color: #008000">len</span>(lmbd_vals)), dtype<span style="color: #666666">=</span><span style="color: #008000">object</span>)
@@ -1271,12 +1326,12 @@ DNN_numpy <span style="color: #666666">=</span> np<span style="color: #666666">.
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec22">Visualization </h2>
+<h2 id="visualization">Visualization </h2>
 
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%"><span></span><span style="color: #408080; font-style: italic"># visual representation of grid search</span>
+<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%;"><span></span><span style="color: #408080; font-style: italic"># visual representation of grid search</span>
 <span style="color: #408080; font-style: italic"># uses seaborn heatmap, you can also do this with matplotlib imshow</span>
 <span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">seaborn</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">sns</span>
 
@@ -1313,7 +1368,7 @@ plt<span style="color: #666666">.</span>show()
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec23">scikit-learn implementation </h2>
+<h2 id="scikit-learn-implementation">scikit-learn implementation </h2>
 
 <p>
 <b>scikit-learn</b> focuses more
@@ -1332,7 +1387,7 @@ performance overall.
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%"><span></span><span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.neural_network</span> <span style="color: #008000; font-weight: bold">import</span> MLPClassifier
+<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%;"><span></span><span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn.neural_network</span> <span style="color: #008000; font-weight: bold">import</span> MLPClassifier
 <span style="color: #408080; font-style: italic"># store models for later use</span>
 DNN_scikit <span style="color: #666666">=</span> np<span style="color: #666666">.</span>zeros((<span style="color: #008000">len</span>(eta_vals), <span style="color: #008000">len</span>(lmbd_vals)), dtype<span style="color: #666666">=</span><span style="color: #008000">object</span>)
 
@@ -1352,11 +1407,11 @@ DNN_scikit <span style="color: #666666">=</span> np<span style="color: #666666">
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec24">Visualization </h2>
+<h2 id="visualization">Visualization </h2>
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%"><span></span><span style="color: #408080; font-style: italic"># optional</span>
+<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%;"><span></span><span style="color: #408080; font-style: italic"># optional</span>
 <span style="color: #408080; font-style: italic"># visual representation of grid search</span>
 <span style="color: #408080; font-style: italic"># uses seaborn heatmap, could probably do this in matplotlib</span>
 <span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">seaborn</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">sns</span>
@@ -1394,7 +1449,7 @@ plt<span style="color: #666666">.</span>show()
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec25">Building neural networks in Tensorflow and Keras </h2>
+<h2 id="building-neural-networks-in-tensorflow-and-keras">Building neural networks in Tensorflow and Keras </h2>
 
 <p>
 Now we want  to build on the experience gained from our neural network implementation in NumPy and scikit-learn
@@ -1409,7 +1464,7 @@ NumPy arrays.
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec26">Tensorflow </h2>
+<h2 id="tensorflow">Tensorflow </h2>
 
 <p>
 Tensorflow is an open source library machine learning library
@@ -1442,7 +1497,7 @@ To install tensorflow on Unix/Linux systems, use pip as
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%"><span></span>pip3 install tensorflow
+<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%;"><span></span>pip3 install tensorflow
 </pre></div>
 <p>
 and/or if you use <b>anaconda</b>, just write (or install from the graphical user interface)
@@ -1450,7 +1505,7 @@ and/or if you use <b>anaconda</b>, just write (or install from the graphical use
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%"><span></span>conda create <span style="color: #666666">-</span>n tf tensorflow
+<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%;"><span></span>conda create <span style="color: #666666">-</span>n tf tensorflow
 conda activate tf
 </pre></div>
 <p>
@@ -1458,13 +1513,13 @@ To install the current release of GPU TensorFlow
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%"><span></span>conda create <span style="color: #666666">-</span>n tf<span style="color: #666666">-</span>gpu tensorflow<span style="color: #666666">-</span>gpu
+<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%;"><span></span>conda create <span style="color: #666666">-</span>n tf<span style="color: #666666">-</span>gpu tensorflow<span style="color: #666666">-</span>gpu
 conda activate tf<span style="color: #666666">-</span>gpu
 </pre></div>
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec27">Using Keras </h2>
+<h2 id="using-keras">Using Keras </h2>
 
 <p>
 Keras is a high level <a href="https://en.wikipedia.org/wiki/Application_programming_interface" target="_blank">neural network</a>
@@ -1473,7 +1528,7 @@ If you have Anaconda installed you may run the following command
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%"><span></span>conda install keras
+<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%;"><span></span>conda install keras
 </pre></div>
 <p>
 You can look up the <a href="https://keras.io/" target="_blank">instructions here</a> for more information.
@@ -1484,7 +1539,7 @@ We will to a large extent use <b>keras</b> in this course.
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec28">Collect and pre-process data </h2>
+<h2 id="collect-and-pre-process-data">Collect and pre-process data </h2>
 
 <p>
 Let us look again at the MINST data set.
@@ -1492,7 +1547,7 @@ Let us look again at the MINST data set.
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%"><span></span><span style="color: #408080; font-style: italic"># import necessary packages</span>
+<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%;"><span></span><span style="color: #408080; font-style: italic"># import necessary packages</span>
 <span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">numpy</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">np</span>
 <span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">matplotlib.pyplot</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">plt</span>
 <span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">tensorflow</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">tf</span>
@@ -1539,7 +1594,7 @@ plt<span style="color: #666666">.</span>show()
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%"><span></span><span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">tensorflow.keras.layers</span> <span style="color: #008000; font-weight: bold">import</span> Input
+<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%;"><span></span><span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">tensorflow.keras.layers</span> <span style="color: #008000; font-weight: bold">import</span> Input
 <span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">tensorflow.keras.models</span> <span style="color: #008000; font-weight: bold">import</span> Sequential      <span style="color: #408080; font-style: italic">#This allows appending layers to existing models</span>
 <span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">tensorflow.keras.layers</span> <span style="color: #008000; font-weight: bold">import</span> Dense           <span style="color: #408080; font-style: italic">#This allows defining the characteristics of a particular layer</span>
 <span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">tensorflow.keras</span> <span style="color: #008000; font-weight: bold">import</span> optimizers             <span style="color: #408080; font-style: italic">#This allows using whichever optimiser we want (sgd,adam,RMSprop)</span>
@@ -1560,7 +1615,7 @@ X_train, X_test, Y_train, Y_test <span style="color: #666666">=</span> train_tes
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%"><span></span>epochs <span style="color: #666666">=</span> <span style="color: #666666">100</span>
+<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%;"><span></span>epochs <span style="color: #666666">=</span> <span style="color: #666666">100</span>
 batch_size <span style="color: #666666">=</span> <span style="color: #666666">100</span>
 n_neurons_layer1 <span style="color: #666666">=</span> <span style="color: #666666">100</span>
 n_neurons_layer2 <span style="color: #666666">=</span> <span style="color: #666666">50</span>
@@ -1581,7 +1636,7 @@ lmbd_vals <span style="color: #666666">=</span> np<span style="color: #666666">.
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%"><span></span>DNN_keras <span style="color: #666666">=</span> np<span style="color: #666666">.</span>zeros((<span style="color: #008000">len</span>(eta_vals), <span style="color: #008000">len</span>(lmbd_vals)), dtype<span style="color: #666666">=</span><span style="color: #008000">object</span>)
+<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%;"><span></span>DNN_keras <span style="color: #666666">=</span> np<span style="color: #666666">.</span>zeros((<span style="color: #008000">len</span>(eta_vals), <span style="color: #008000">len</span>(lmbd_vals)), dtype<span style="color: #666666">=</span><span style="color: #008000">object</span>)
         
 <span style="color: #008000; font-weight: bold">for</span> i, eta <span style="color: #AA22FF; font-weight: bold">in</span> <span style="color: #008000">enumerate</span>(eta_vals):
     <span style="color: #008000; font-weight: bold">for</span> j, lmbd <span style="color: #AA22FF; font-weight: bold">in</span> <span style="color: #008000">enumerate</span>(lmbd_vals):
@@ -1600,7 +1655,7 @@ lmbd_vals <span style="color: #666666">=</span> np<span style="color: #666666">.
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%"><span></span><span style="color: #408080; font-style: italic"># optional</span>
+<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%;"><span></span><span style="color: #408080; font-style: italic"># optional</span>
 <span style="color: #408080; font-style: italic"># visual representation of grid search</span>
 <span style="color: #408080; font-style: italic"># uses seaborn heatmap, could probably do this in matplotlib</span>
 <span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">seaborn</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">sns</span>
@@ -1635,12 +1690,12 @@ plt<span style="color: #666666">.</span>show()
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec29">The Breast Cancer Data, now with Keras </h2>
+<h2 id="the-breast-cancer-data-now-with-keras">The Breast Cancer Data, now with Keras </h2>
 
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%"><span></span><span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">tensorflow</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">tf</span>
+<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%;"><span></span><span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">tensorflow</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">tf</span>
 <span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">tensorflow.keras.layers</span> <span style="color: #008000; font-weight: bold">import</span> Input
 <span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">tensorflow.keras.models</span> <span style="color: #008000; font-weight: bold">import</span> Sequential      <span style="color: #408080; font-style: italic">#This allows appending layers to existing models</span>
 <span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">tensorflow.keras.layers</span> <span style="color: #008000; font-weight: bold">import</span> Dense           <span style="color: #408080; font-style: italic">#This allows defining the characteristics of a particular layer</span>
@@ -1809,7 +1864,7 @@ plot_data(eta,n_neuron,Test_accuracy, <span style="color: #BA2121">&#39;testing&
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec30">Fine-tuning neural network hyperparameters </h2>
+<h2 id="fine-tuning-neural-network-hyperparameters">Fine-tuning neural network hyperparameters </h2>
 
 <p>
 The flexibility of neural networks is also one of their main
@@ -1835,7 +1890,7 @@ will only be able to explore a tiny part of the hyperparameter space.
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec31">Hidden layers </h2>
+<h2 id="hidden-layers">Hidden layers </h2>
 
 <p>
 For many problems you can start with just one or two hidden layers and it will work just fine.
@@ -1855,7 +1910,7 @@ common to reuse parts of a pretrained state-of-the-art network that performs a s
 <p>
 <!-- !split  -->
 
-<h2 id="___sec32">Which activation function should I use? </h2>
+<h2 id="which-activation-function-should-i-use">Which activation function should I use? </h2>
 
 <p>
 The Back propagation algorithm we derived above works by going from
@@ -1884,7 +1939,7 @@ learn at widely different speeds
 <p>
 <!-- !split  -->
 
-<h2 id="___sec33">Is the Logistic activation function (Sigmoid)  our choice? </h2>
+<h2 id="is-the-logistic-activation-function-sigmoid-our-choice">Is the Logistic activation function (Sigmoid)  our choice? </h2>
 
 <p>
 Although this unfortunate behavior has been empirically observed for
@@ -1914,7 +1969,7 @@ better than the logistic function in deep networks).
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec34">The derivative of the Logistic funtion </h2>
+<h2 id="the-derivative-of-the-logistic-funtion">The derivative of the Logistic funtion </h2>
 
 <p>
 Looking at the logistic activation function, when inputs become large
@@ -1950,7 +2005,7 @@ fast to compute).
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec35">The RELU function family </h2>
+<h2 id="the-relu-function-family">The RELU function family </h2>
 
 <p>
 The ReLU activation function suffers from a problem known as the dying
@@ -1977,7 +2032,7 @@ $$
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec36">Which activation function should we use? </h2>
+<h2 id="which-activation-function-should-we-use">Which activation function should we use? </h2>
 
 <p>
 In general it seems that the ELU activation function is better than
@@ -1997,7 +2052,7 @@ bootstrap to evaluate other activation functions.
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec37">More on activation functions, output layers </h2>
+<h2 id="more-on-activation-functions-output-layers">More on activation functions, output layers </h2>
 
 <p>
 In most cases you can use the ReLU activation function in the hidden layers (or one of its variants).
@@ -2015,7 +2070,7 @@ It is a bit faster to compute than other activation functions, and the gradient
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec38">Batch Normalization </h2>
+<h2 id="batch-normalization">Batch Normalization </h2>
 
 <p>
 Batch Normalization
@@ -2034,7 +2089,7 @@ mini-batch, from this the name batch normalization.
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec39">Dropout </h2>
+<h2 id="dropout">Dropout </h2>
 
 <p>
 It is a fairly simple algorithm: at every training step, every neuron (including the input neurons but
@@ -2049,7 +2104,7 @@ hyperparameter \( p \) is called the dropout rate, and it is typically set to 50
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec40">Gradient Clipping </h2>
+<h2 id="gradient-clipping">Gradient Clipping </h2>
 
 <p>
 A popular technique to lessen the exploding gradients problem is to simply clip the gradients during
@@ -2066,7 +2121,7 @@ Normalization is preferred.
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec41">A very nice website on Neural Networks </h2>
+<h2 id="a-very-nice-website-on-neural-networks">A very nice website on Neural Networks </h2>
 
 <p>
 You may find this <a href="https://playground.tensorflow.org/#activation=tanh&batchSize=10&dataset=circle&regDataset=reg-plane&learningRate=0.03&regularizationRate=0&noise=0&networkShape=4,2&seed=0.29243&showTestData=false&discretize=false&percTrainData=50&x=true&y=true&xTimesY=false&xSquared=false&ySquared=false&cosX=false&sinX=false&cosY=false&sinY=false&collectStats=false&problem=classification&initZero=false&hideText=false" target="_blank">website</a> very useful. Thx a million to Ghadi for sharing.
@@ -2074,7 +2129,7 @@ You may find this <a href="https://playground.tensorflow.org/#activation=tanh&ba
 <p>
 <!-- !split  -->
 
-<h2 id="___sec42">A top-down perspective on Neural networks </h2>
+<h2 id="a-top-down-perspective-on-neural-networks">A top-down perspective on Neural networks </h2>
 
 <p>
 The first thing we would like to do is divide the data into two or three
@@ -2116,7 +2171,7 @@ supervised learning.
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec43">Limitations of supervised learning with deep networks </h2>
+<h2 id="limitations-of-supervised-learning-with-deep-networks">Limitations of supervised learning with deep networks </h2>
 
 <p>
 Like all statistical methods, supervised learning using neural
@@ -2142,7 +2197,7 @@ Some of these remarks are particular to DNNs, others are shared by all supervise
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec44">Convolutional Neural Networks (recognizing images) </h2>
+<h2 id="convolutional-neural-networks-recognizing-images">Convolutional Neural Networks (recognizing images) </h2>
 
 <p>
 Convolutional neural networks (CNNs) were developed during the last
@@ -2181,7 +2236,7 @@ Another good read is the article here <a href="https://arxiv.org/pdf/1603.07285.
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec45">Regular NNs don&#8217;t scale well to full images </h2>
+<h2 id="regular-nns-don-t-scale-well-to-full-images">Regular NNs don&#8217;t scale well to full images </h2>
 
 <p>
 As an example, consider
@@ -2209,7 +2264,7 @@ would quickly lead to possible overfitting.
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec46">3D volumes of neurons </h2>
+<h2 id="3d-volumes-of-neurons">3D volumes of neurons </h2>
 
 <p>
 Convolutional Neural Networks take advantage of the fact that the
@@ -2249,7 +2304,7 @@ dimension.
 <p>
 <!-- !split  -->
 
-<h2 id="___sec47">Layers used to build CNNs </h2>
+<h2 id="layers-used-to-build-cnns">Layers used to build CNNs </h2>
 
 <p>
 A simple CNN is a sequence of layers, and every layer of a CNN
@@ -2272,7 +2327,7 @@ A simple CNN for image classification could have the architecture:
 
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec48">Transforming images </h2>
+<h2 id="transforming-images">Transforming images </h2>
 
 <p>
 CNNs transform the original image layer by layer from the original
@@ -2291,7 +2346,7 @@ are consistent with the labels in the training set for each image.
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec49">CNNs in brief </h2>
+<h2 id="cnns-in-brief">CNNs in brief </h2>
 
 <p>
 In summary:
@@ -2312,7 +2367,7 @@ and the slides of <a href="http://cs231n.github.io/convolutional-networks/" targ
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec50">CNNs in more detail, building convolutional neural networks in Tensorflow and Keras </h2>
+<h2 id="cnns-in-more-detail-building-convolutional-neural-networks-in-tensorflow-and-keras">CNNs in more detail, building convolutional neural networks in Tensorflow and Keras </h2>
 
 <p>
 As discussed above, CNNs are neural networks built from the assumption that the inputs
@@ -2328,7 +2383,7 @@ matrices, typically 1 for each color dimension (Red, Green, Blue).
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec51">Setting it up </h2>
+<h2 id="setting-it-up">Setting it up </h2>
 
 <p>
 It means that to represent the entire
@@ -2340,7 +2395,7 @@ $$
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec52">The MNIST dataset again </h2>
+<h2 id="the-mnist-dataset-again">The MNIST dataset again </h2>
 
 <p>
 The MNIST dataset consists of grayscale images with a pixel size of
@@ -2358,7 +2413,7 @@ single neuron in the first hidden layer.
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec53">Strong correlations </h2>
+<h2 id="strong-correlations">Strong correlations </h2>
 
 <p>
 Images typically have strong local correlations, meaning that a small
@@ -2376,7 +2431,7 @@ fixed, and known as a <a href="https://en.wikipedia.org/wiki/Receptive_field" ta
 <p>
 <!-- !split  -->
 
-<h2 id="___sec54">Layers of a CNN </h2>
+<h2 id="layers-of-a-cnn">Layers of a CNN </h2>
 The layers of a convolutional neural network arrange neurons in 3D: width, height and depth.  
 The input image is typically a square matrix of depth 3.
 
@@ -2398,7 +2453,7 @@ layer.
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec55">Systematic reduction </h2>
+<h2 id="systematic-reduction">Systematic reduction </h2>
 
 <p>
 By systematically reducing the size of the input volume, through
@@ -2413,11 +2468,11 @@ classification.
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec56">Prerequisites: Collect and pre-process data </h2>
+<h2 id="prerequisites-collect-and-pre-process-data">Prerequisites: Collect and pre-process data </h2>
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%"><span></span><span style="color: #408080; font-style: italic"># import necessary packages</span>
+<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%;"><span></span><span style="color: #408080; font-style: italic"># import necessary packages</span>
 <span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">numpy</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">np</span>
 <span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">matplotlib.pyplot</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">plt</span>
 <span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">sklearn</span> <span style="color: #008000; font-weight: bold">import</span> datasets
@@ -2461,11 +2516,11 @@ plt<span style="color: #666666">.</span>show()
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec57">Importing Keras and Tensorflow </h2>
+<h2 id="importing-keras-and-tensorflow">Importing Keras and Tensorflow </h2>
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%"><span></span><span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">tensorflow.keras</span> <span style="color: #008000; font-weight: bold">import</span> datasets, layers, models
+<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%;"><span></span><span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">tensorflow.keras</span> <span style="color: #008000; font-weight: bold">import</span> datasets, layers, models
 <span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">tensorflow.keras.layers</span> <span style="color: #008000; font-weight: bold">import</span> Input
 <span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">tensorflow.keras.models</span> <span style="color: #008000; font-weight: bold">import</span> Sequential      <span style="color: #408080; font-style: italic">#This allows appending layers to existing models</span>
 <span style="color: #008000; font-weight: bold">from</span> <span style="color: #0000FF; font-weight: bold">tensorflow.keras.layers</span> <span style="color: #008000; font-weight: bold">import</span> Dense           <span style="color: #408080; font-style: italic">#This allows defining the characteristics of a particular layer</span>
@@ -2491,12 +2546,12 @@ X_train, X_test, Y_train, Y_test <span style="color: #666666">=</span> train_tes
 <p>
 <!-- !split  -->
 
-<h2 id="___sec58">Running with Keras </h2>
+<h2 id="running-with-keras">Running with Keras </h2>
 
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%"><span></span><span style="color: #008000; font-weight: bold">def</span> <span style="color: #0000FF">create_convolutional_neural_network_keras</span>(input_shape, receptive_field,
+<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%;"><span></span><span style="color: #008000; font-weight: bold">def</span> <span style="color: #0000FF">create_convolutional_neural_network_keras</span>(input_shape, receptive_field,
                                               n_filters, n_neurons_connected, n_categories,
                                               eta, lmbd):
     model <span style="color: #666666">=</span> Sequential()
@@ -2526,12 +2581,12 @@ lmbd_vals <span style="color: #666666">=</span> np<span style="color: #666666">.
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec59">Final part </h2>
+<h2 id="final-part">Final part </h2>
 
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%"><span></span>CNN_keras <span style="color: #666666">=</span> np<span style="color: #666666">.</span>zeros((<span style="color: #008000">len</span>(eta_vals), <span style="color: #008000">len</span>(lmbd_vals)), dtype<span style="color: #666666">=</span><span style="color: #008000">object</span>)
+<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%;"><span></span>CNN_keras <span style="color: #666666">=</span> np<span style="color: #666666">.</span>zeros((<span style="color: #008000">len</span>(eta_vals), <span style="color: #008000">len</span>(lmbd_vals)), dtype<span style="color: #666666">=</span><span style="color: #008000">object</span>)
         
 <span style="color: #008000; font-weight: bold">for</span> i, eta <span style="color: #AA22FF; font-weight: bold">in</span> <span style="color: #008000">enumerate</span>(eta_vals):
     <span style="color: #008000; font-weight: bold">for</span> j, lmbd <span style="color: #AA22FF; font-weight: bold">in</span> <span style="color: #008000">enumerate</span>(lmbd_vals):
@@ -2551,12 +2606,12 @@ lmbd_vals <span style="color: #666666">=</span> np<span style="color: #666666">.
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec60">Final visualization </h2>
+<h2 id="final-visualization">Final visualization </h2>
 
 <p>
 
 <!-- code=python (!bc pycod) typeset with pygments style "default" -->
-<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%"><span></span><span style="color: #408080; font-style: italic"># visual representation of grid search</span>
+<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%;"><span></span><span style="color: #408080; font-style: italic"># visual representation of grid search</span>
 <span style="color: #408080; font-style: italic"># uses seaborn heatmap, could probably do this in matplotlib</span>
 <span style="color: #008000; font-weight: bold">import</span> <span style="color: #0000FF; font-weight: bold">seaborn</span> <span style="color: #008000; font-weight: bold">as</span> <span style="color: #0000FF; font-weight: bold">sns</span>
 
@@ -2590,7 +2645,7 @@ plt<span style="color: #666666">.</span>show()
 <p>
 <!-- !split --><br><br><br><br><br><br><br><br><br><br>
 
-<h2 id="___sec61">Fun links </h2>
+<h2 id="fun-links">Fun links </h2>
 
 <ol>
 <li> <a href="https://arxiv.org/abs/1604.07316" target="_blank">Self-Driving cars using a convolutional neural network</a></li>
@@ -2602,7 +2657,7 @@ plt<span style="color: #666666">.</span>show()
 
 
 <center style="font-size:80%">
-<!-- copyright --> &copy; 1999-2020, Morten Hjorth-Jensen. Released under CC Attribution-NonCommercial 4.0 license
+<!-- copyright --> &copy; 1999-2021, Morten Hjorth-Jensen. Released under CC Attribution-NonCommercial 4.0 license
 </center>
 
 
diff --git a/doc/pub/week41/ipynb/ipynb-week41-src.tar.gz b/doc/pub/week41/ipynb/ipynb-week41-src.tar.gz
index b28180744245f882772936a1d7ee10283f760990..60560e5358ae55b2b865a069578abc60af3de38a 100644
GIT binary patch
literal 87692
zcmV(rK<>XEiwFQ;&S7By1MIs8SW{n@FMN;^I?_8qK|zsTr6hs^A_Ag95Rj;#Gy&-W
z0zr@>RYl}4Dkz{J2oa<RgdjD5NEZ;KB!F~Cr~y)LeBYV({l9Z(=FXivcb@NizHdEQ
zIXgM~th3iXYwh)0d!K#B(>?I6r`zp7RTa3euZq{b+wQ=BHL0no>8Pp6{yysJX#eG?
zrlIxM<8RW`(2-Tw)YVke*4EY5QkPXz*Vb0k1Z36z%UklVP;g+-%>YJId_25v`v%_j
z{kQ5ds@=Ks@5uS9Dzbk&{=bD-ML@tI0I<60Dr<8^)<O24==XmB2ZJ>jT)^PpzlZ;0
zB8{(Jz5Gwb-&g#>V939y!u0q5g<bu8RBqn8clWl6fAGzKo4!GwzPAB?RW&s|bpwXo
z+`6Nusc8`CaWlZv*WJzFw2q0IhN<y6WxaDITFP2x=X8}#jP=x&)l4)s&Y7H3*EBWL
z`!DF0|F6)Uzxn_FE${zd_`imlj+VOmzx00%UG@Lp|NlFPK1!bk4qq^{FawyFe)}ck
zi;f3wnIb$H+FDrwCmEI40ahjffEoCw^Gp%|EAUqzzQ82;xAJ8s#lP1%B>4C8KM~WP
znLZAj1(**V`u$-L3**De&dSQd!paE(v9WV=a&vQWa&htS3LNI)<>%$%IxKRSUr<O`
zSeTnnR7^xjOh8Cj=(i9iX2vxvtQ@SY96~%?JVO85ME?Ttvok$l7G-8S3LN5RV&-R}
zw*p{B8(EqDLVs)jAx0b7K<pfxT#O1;hk-*(%*=;an18pLaW;~19$?{T6*!`9%qD1i
z6Li#HNaOzFH|%mI<xRqu`-$?Jw*q1~I7LLo#3ha$S5Q20QcGJ$SMRj`*>k35=I1Rg
zT)Apz4|BNY=<0U+j=P7aS71<ZNGKvKJoZ6cd_v;Gq$e3qGoNMs@jN>(zu@h=!lL)Z
z6`v}rs%t*i)-|`ZwzcCrI=co2hrSPwjQ$wI6J}=Te$6i|E|E4ix3+h7$$R_1^<n~;
znf{snRlWEbdL3e6VP*mS){E&-=<f>rEUZV=*#wMjK{x#ck80d!7czPLro4$mPV+KR
z_*Otar--~3{ut@EYJbt}Kc-mBe@e5zDfV~0#sF?+CdR{K<_Dkv^-5lnBJjV@@uZ|{
zJzeK-N**@%J@Gd5#_>E%E=_>PhK4GGuH*CnUPRj*Kv3Oj&{FgAWg%Dw9cVUwPX~Ip
zaa4*H>d$5O-n~bWUPkE-|CXfu--2S7ZCAqEglFjh@sSs5lj=_gP(0v2mw#;1+`x|M
zuEUCG1v>B`_6i-yT~0qBd;G%>T5s;A?EX-MOu1<olp-&YOQ8eKsHqn7KbLoqbl^E?
zIDd3INvW4|tYmUx`B9tpP({04McHiZM)npbq13cLOX<$muVSUwc}Yslm;TKjtL2kk
z(zVyTmK^c$2zP0{D}PSr*FWj6y>Pjy>x0nX&tI>CIRYyGT@^j4xzJzd{vX^|opb;%
zJ46TO_h)D{i7gxM-eWxpo~OzN&UXcL2iSCb=Tu8vJzDVNt$tU2h1pvnTW9w(wx&YB
ze-~dBr30iHMi)4b?_J$yr2`biPt=GG&CYg$V+FA-WD9pj+%y%s#AJ$GC7JL1B8N^(
zBgyV(CDL1M#6JCzahCYDwjQOcH`)Vno{uT1O$h#Q9&(H}5_LLd0^zh1Hb^nJ)pgN_
zP<l&UCBiqQlPmq*Brba{=lE0U7dolpU$3UWS43x-C(;3*)I-opuUr<49QsUZR=Vl1
zT)fpEEAn*#HZIZIQvR0~=s-<%2L4+(5AzJ@zM&;C^x7W;-QiLfrxOOZ@30r7&;gNe
z*wPDnbs8a`CpeSeY?}&OnhtN#0sck%x}1zipZAM3=jCJ8dhDK^xJm~E>YlLg3WwBn
zA4*bu%`?S92RhpPXv=h9;dp}zX0Qpmq3%Qn>gd2e2ig+N45rvEggf}f{t~h^4Z}J7
z8~w@DLNTQSaM#~*)SV+k!2GC@(Z5LPhJoRh4`|$|vce2FIYluqX<3a^=_xs#bfp>2
zUdddFz`;NDX-$laI2G;{+;zX}q0uus?-dhSGM@||ztRpCxsAZGBl$_@={~xDWL+N;
zU6Ba8alRvf^YKbg!VqWsTdyFjj<aVMiW5?ihjsBeyE9vNI~)8Uim4(2s4DF#s<0e+
z+@Pv64^EKYxAUe0W1>6pWP2IEhp10!Zy}lD-DA12>z65Ju_b@>b+(`0d!;`4(V<Iu
z>=&BPHV9Ijhsl`EU9}}n&1^-%;;Oz&)V=o%vNDe9jTF(Rt>pf}yAt;_e57>5*{uLA
z8pxI%!Y~^<cBLWYF*<N`HS3^|4wz`qeP4nI(1FJxU#Ww1;ElwlIvr?nDRhyXiHdI!
zvUTBKWEHYK^WVnCC^{gc3*P?Lm}zqfP6u$QIBKBZoSk%E4ldo+RCpT~Z|?6e{1)@6
z_vBfY3oaIJH;R+kEaZuR^zzK8HdF@HbQbATOm_O>_-ZO9ip^87s;r!_Ae5Zy*{;c2
z?Z5GO{Pkt=&01H3+-hi|cH_gGlSG@9)#<N;x6z2TE634@=#1M1?`D+L#PH1)?P2^H
zxsC8(q&nFfPX}UYO#OzEcyHmsO68+JoyJX0%Lls$QmnhLJ+J>HIEg{tIa0WiziB|+
z%B-CYgfpY=tS7)f<tAiV5W(@KL8%=g1Pc0jSm>0d^_0P8&(BqGGeBt2Nm9ZY#C)W9
zPW$m8&4}AQafDX_Ev*vT9j(>+sde83WpLiiF$Vk{OYEgxip-k{gsBz$jLoS2&L%;(
z34|emoBI{U49^nxH(<{{5NHO;?x*i6E`{CxqWnwNAS#i3g=E=62gETPZ^<<+x+t%K
z^_ic{H!Qf)8-$WL#+DFrBykqOt@lj#Sq|~BNKgn^Vl*{@4&05f611{)C!z@5DZ@JX
z5bLaxGSJi5ypm^jm((Ah<b4c#`=C|N!o*4QNyD*Y-8GPEMv$2_91Q&zj;ckpp`_T;
zfwe|jjQ7Cu7aScB4wrc*IlULhurL2%hWi5l8sB4#6Myp6>j|y!?H%Y^<EAifWB?D1
z1uws!9iho_lKjTpcOF-gIbJUg*G0Xg9Vh*2Q2W$pc-pV_ZT%k~FSRNv+&G2db$FEj
zUT(N*HQ;NDgh>7euPdcxuL{R4LIVwg(iDg4%l8$JW230iQIamm01~$Gz=r&*uVJl3
zt6GNlW39M^Z?3rd^GiY8xBxb>MJ3eN3oq?Zu9D^<`dZZd5jP(sGd21ciLD0IIyDAa
zf+m6w`vT8+*zgast<!3|T91O|KQy#+-^hrbpNIN5jkme`2^~K7$XMV3<K8Fp?neVs
z0Lbe}(%$fRsC!XnrLn0H!%EXvPaJSkjHquOZzsNnaFSy=Ha}o)zEQ{smHIO0a_KD+
z{|G|j22BK`naD>-<L$Y8ScRG|y=f$z_e{}9NCH8Ov~xXt>eUCa8(|MlG<`W`Y|lBv
zvo8)eq>WE~k^)zvn+}pAPe*;7`36)M;$K-G-<XSE1^g8CW>38^$wAym{m$xVR$0`w
z(Dhc^_S%afSV~kUSW2F}OLVdv(PrI1n!ByO^R}PU5nDg{B>+YaJz<XXA+Ngh!fnJ)
zv{?ij#=pG&2ds!}Ni!W+-L;`T@cvrzkHgQrD;<b$u>PRiPUQ(#%)`hHa4P0S|7Y27
z!$!vI!pfnh=<^2&6wgLt&Jt=OmPUroIB!iY!>Kb?WxQdbkkXq)<9Ks(I>513NF9Z)
z$I=*RFM@$~Lii4*=|F1W)64T&{#3{Gjr@ED*x0>7TLJGj(E$l4-iU!5>OoW@s=c;L
zaX#zbeieA`-<4$BtAgM5f_9~)(*!4M4nCr$KopOE5IV1zpL4xsX$X}m4W=m6<?h~~
z16V$C|DS|v_t=^n)+c-~q5~%ll)2hP_tQ%Kd=wv;ACX<X%x4nmU5!>I=2B90;Ok&2
z%gB3}-SM#@lKRk$b>1cWK1O64-Hc(x)I)#f*Nt=_T{3<D*Hl+wad_L1j-%b|uAn)~
z2|>{(_MFA)Lm6L+(-v%8I~LIC+f!rk_AyXmm84j(wC<Ja>e?rIg%12ru5z1vR6HY_
zBBNMMMI&8BSn8ru^?K!-PL@*&aEmT<V8qj<Xa4zi*sH7+yTbH`sOsER9+JeCIA4NW
zY_r6Tdly=Zx)sKA5`rsoPE(gd45_Pd4m!}JwEkiJ6ir{u_Q@W_=hW_z5$+G97}$cm
zAi(E!P=OKunr(f(_rl3e*}SBaoPT-vd!jVX%;~cP^baM+5g%RCQiRF=pv&NzYsV7>
zi*vqbdClZ9#|KXuF@dkc9@54IM$aA`N80JX&~T`Zz4t`RLppx$%Sk@T3Ny|7F9Y=Z
z*!Ws<MHJGbWXP_q{Y6)YgX(=uuZ9S;Rxr#}rYs42h#X0(Y0D8MCM0Y{f}A4wc*h;S
z)FI(FlT4nqtKB*sG-=k%Yat)&GBr(gBG{zMZfe;i=?jnFe^#U8U(mUI7Pz_HZnV7M
zNtiJb2;D<HpX-WED_w|$P3P{3mC}I;6yq_z9qE8b5n(vbDvxJh3B2j7O*~bT?HV8u
zbYsoCoDRfA9-OBG{g?V0iPssRQ4aPzB}?0~%7?d(vu|>~zY^91C!!lY!TTpAL}^@k
zbYRvCiibxHK)0@HsZvB4u&CpEb5O2XKm0VR8A~ldEfjgc$tuYVG*I+EOTPsA4@n$i
z+VT<J%%Ui*kD1ew;^XK*%tGUyL^>HZRhdo|=Ha3Pto^7BZ6H$86gfu+I1hFn(}BD<
zZ&1^9sx-5Fn(*$b933z{Yr?RDAUGxU4EVpgfj^O<r0R@mC1p+9;X-fk6UWrTO2i{S
z{ERV~otZOY4U_;SX{8H|MHZePDh(EksKUJ2IZ%^|UGbi0K(u^DC>06f<6KO@MJazI
zi9FXH9-XPEN|Tei?+|mLHQ}}IFRb_Awo=3KMU}h0g%!0n_1c(}v_=#5qYWn=YRpdR
z^v%9v)tT6aFw}@iAy;g+Onp*$`*yLL5h7d1Qr{);4!3~SyVLTivk!+3)X_CqF$6CT
zAKH@3jeuUfRIM-b!mqIP%t5nW^BQC}#$wF^-GNSR6s|&vdZPF~AYnzlsGB!ucr<>o
z?YSPen&p?9DSUQY_R)DG(qs6i%=qIKrI6d@rHcCgpBSD~%k$3z179y+oV^n91&x!L
zH2+0s^^*ljxS=t51>Ym=78xIz|N6LPmY8%t+!|7dehM)nzrrw(BC7=Edu2V>D{s3&
z3ZKtTvLf46NAcNB3psf;WiDQQcS9LSDiNAq2~Tv?_ma@=d#)$Y{L_*P6+BW-<-6rQ
z)pXT5JND<MI=`1Q&eSCzuev9Cw{Zfx&yQ9?+Y+-V$<^?+Tq^s}%<m{+9#5WngEkd{
zJqS1cO9J$tV^GRBS`{6b)EW!=YfM0E9hiQ!HAN__avWU!K~Wq&g%rjcNuXqqSHBap
zL*s*{uV>urcoKlgn-;-cW{a}>%9i2V@w`WC=9_|BMAi47B02-k&flaPFkaW@GGmIF
zhMph}VM5!$)$UEIpk8_X^wwAWbLmH(HQP5dNt>@RZ>}!m9On#A&;gcRQX$R)PZC{K
zof)6F*!l1lQ8m8s0P@1U$HV|VC3{0U_B41IUK*)1JvBNw?H>JMV9K0)1#ib*e{Q7i
zZs@0($!jgO6sPQks}Y}YIk(o=%-122jL_ld7ahLxRk)_rTk-h|58Y>3_9eL=$`zTX
z-bSd#Aw3yjR@yWF67g4Uq4$nLdgWr4g}+Rs?h#qA*J_o`kBn(rmB;YN6emSIkZIm~
zK?Eb^D!Ug3%IKKS-diSsDFXBO*^yK<^7wESIu@&%J|DAbih1T_?`Lr&bfR*rSwH<!
z_>~}-*~LI{_|(HyEm$W_hME^8OY6bvRFESNEIfyIED)3Sk4aqC&Ym^1sNP!ufBD+I
z0yke%uy*8ispOMN`5-f(<V1(u5KJqNu+r<2>0zI)CAOrPS1>yK55evk*@k8&sk)oa
zz?VZM73s}Lnk&V0fq@}?!~SY0!t*Yg)HhgZGFP}4_C)!?g;_}_9bHZ41$(15n}gBU
ze5QHWAcl#(A*rUNXAwv6&&o@23LM2=pR1}sBd5pAzYh(dnybk3Br*3kh|wqF?FX^P
zTUzdAoZOE+ZVcWf#Q_lsNrr*=DVFiDeOk@U0y8t%Y8IEc->&g#zh+D*j9Ou+Y_z;M
z$<treDDCCC*7o*CU3(brR7;3aGryyZ7&RR$<B<iu`(4@35cXdE=dR<#SIgS^C&^Vd
z6n2RnX}NQQ80ekq?CWu5FFYDUxX#-v;`*;Losl^-hGv9RS=8&rY}J<Qr}r%pgkS66
zi)y{!4-YUhxuZ_r_t$B%h!Hrc>gUaZ!(UIwj7kYcKfeV!_|SM^cNOkX-lz5TmHFlR
z`1a->+!@Z+cmZwcPVG%d{Y<ZD)Q~7FTlIblO_ChJ<Thkzp&^vL#arI2CkwXze9#tH
zS@jvLK=CF{B~+p%U0SABetL%V9jQO;`$XwlsE`t7KE1Y6;nywAv05~MxQ%h0v63_c
zR_kl8W=Nk4`<mNyR$}*uq3TUSgJ+aDSwqtspQ>;AqE(_;s^M#F>uoxaID6Q@O^RCq
zlFk5A(Jo-3YMObX*0X1N!{1+9p0@CpN&q^_tivv~fDRaN<4JQ}#xEy0d>AMB;f<tG
zuQoMv-Ic6x5Jhn!NBUQfSdnPCpof)_Gak=8YQp*pv2L|`(r;GbsF5$Z)u`si!(RL5
z36K*ACNZQLIU+>clI5YU_T1+q@1l$5$j))JBMoQ1j(i7yg0J%6(PM#l>#P|#Psd?v
ztliGb9tkI|*P!QIRF00jJY9APj`MAogO(uO!zp{y=~ozu(W9RX7zQZKfXOFwQ{Prn
zbM8W<J<nMY{1UUjze~R9!9Kb4Nxp1qZO!4*<J|51VeFR??QaGx652cHk}2bNiy>-N
zZQtjHV)WjmnqB*m4gegT=$^*7aK=cs`sTpeg%a3Ul|6;8c`$4x*Cdi<yQ-@Z^lXKJ
zMDNR(vpx>|Bh}}^ixO?2bF*(me*B=gQFEr#zl3YiB;&FfR#J!5adG4}`c|p_08W;O
zn;CKoYzxbcIYyjHY)h?4GW;g6f2Tc{&~TcP*8`35-wlmTDQ^kj&^wYo7QBgSLotm~
zj7H|azx}A2!#r|~qvTpEX0#_jgrK_i@`<_cwQ!B3ZIaMccxBpzHK(1VbIK~_bP?mB
zm=?F`I~KN1NXSMxRAe=&cdl8i(*b3^&u9;FC?f$ACGdRtco0EGIq{TlF>It)F)GgP
z;TX>7{=^b1Lbb&YJVooR+sX7J7!-~~Nl!#;O++y@ls+C>f#K|MdRudWsrid=3FL7i
zxIOixQt1`TitkSd+7XfV{V`pa(lej9&6cU(e_~Y9-AaiJSqiyKwx%WR)zF0KK$~Ce
z=?pwB8I%;Hw+H7%B*t_#PNZGDzbA$(FGX?FhQ|ygvKMj>2Qd=egt3L`vNw|2CAoGP
zKy>(QDuz1a{0`U9Z&$f$-Rn@1w9&lAwQOOwu%=ufU^X+PCeOcBqV<mxRSr?gr1tyM
zM%rk~*^)6&k$k;~=O$T@qy~or#p4*0&7D^xv?CLig-AK#?Wsi0a<dWbb7~H_^-l5a
z7L6Rp;%mM)eyToO%h!B;=aYuVj!no9>YX!QvY6&1u8Tws(_~_==qto%jjty{WaZ;(
z23%ba&D3Ur<O)x+<NU9Mpg!?7ka(Ar%?ax%YZMbZOk3X;PR*e;QjghC>b9K?7%}p0
zDKv+Lnq$asH2>oR=5&6`{W-^0mMj6+xl5pLa=Djn0?q^3z+T}DG!}atd6xLIuBNrr
z^#bwPV?UuRojUE0_8hV-3fX&07R>9q6*1m3ANbRhmQ|BvWWUpnFOm>+fFO*i14mDH
z_8V$ReHe>Hah;x~OaxMYW`5brM#nX}SCJFmID2DN!?vh!+m<n>v6}+zhd-z4@Ta(f
zRmA{aUhfL6>ke(k-qgr;-n>{Ou9pTSYX%bNz~~@uf3?6j*xMj6*z$M^{+YqqI(rcM
zXINd}Q{nxa2T<hYYVrutrpZOG`Z4^#+cN&wjaLFQzCReaBf6_XmBNOZYRciQCT7J{
zuJhtmR#!_Y&n?Y#x~(rhc;CxyDHS-34WK?VjG8Kk;~K6^LzBy7rtl+4>ynp=I^_ZW
zQkv(_MCb;7={(u+aat;h!e6GqyV$G&*VMPWTF}3*%)m4+gX8`(2L|he)=aIchZ5(T
zL#DVWEV#?$+L=uKs28`rX1<ljT>Ow_L+#OJagAobvnYAx!oI9w>`w3yu04J1U}m>%
zf&GPP+L9{A!tRMKOCctf4+#gQvn>jZFI@6x(qlvIv+EHhUv%Zpk|gHdRylJ;^beJ&
zp0R0U*>2h^f$fBoH47<+$QG~38comKKP5-b4WvZAo1=NUZ8zOI1A6qv+-oa*|3Z*J
z!5d{daHZpw*{|7jU9unP$w=EsfJgSX_e467tlZOCIi7<4<IG&XLJ>}HIo<YKq~@QH
zT@FN^qBsDN8_VylYk}4L8o64%cPBB;%!Z^A5p_<+gCTFwiEJ5MNIpxzK;1vRP<l%)
zQ?qL5T3bA^Vr*E$;{Vxo-cYY6m{r`nY~G3&cAOAtWU3*a|2gj1!(oS4XXV~>eKd^W
z^S43EuKgd)CfxDcM$<^|2H(*EM>?=B;S*LkrkpNAEru8tuFwHu@%**QXfK)LL&vV`
znhaKF>FPdlQ|LN<^0kN@_Ry?|@L*JHZiO0I5Py&^j`&6gY9Q)}H&(;-AAa)7jjtQf
zzVa^=V|<lb)M!huWz_Mdax_NP!fEQ~9d<wmxG1s%0@NFR*}C5qI6EfvZcnBdNeo4d
zEZBuyNzLqi(LNk}Qg4faZ|re7zo*OmQDP82GHj+MS{uqn{(8~M#uV#fOgJwO4sggV
zK3_1awhNMd6eUH?Mw%1g(YYd3=|9i}o@Mw?&wG+npykUk(8|i53%K4OySv=4>W?`S
zv=z2ou=0#7B_E+OZ$S>l^QDG!Y9L<Vu~mK{ZfDbVx;zC=yZT*n+X$a!)$otk7EISD
z?>8yQ*FLss)RWXp+^O{@_%L4_J_YZt7*fuwzfT7mq!OiVW=%a{Ig!4<Y925Ww3`)l
z0P&YO{tK8Rv||XTFX0Uh?G0r=7E8oWb;>`_khD2Jatg(C)t3(3<x-BCTcLpwnl7Y<
z4g=(Z&luOe2!U2jCfWvixL%Q>(0hkRzjz}3jn}i%QqO(Xj7hK1oiU%h$i=l^4*BpB
zbpRGk2X52)ruaKWv(u7^<5n3TLXuQ`cc=j^Q1i1et=!S67ExVK)Qpv%6Y1AG$X4B}
znF9BP;!vg)=cJz--@rn&Q8;g<WUTwU$?N2j%L-nO3(wlGe_NUNdQ4k{o^>!MxZ_k4
zF$mZfogCi5pVM&iv&*-c%lV1sRAO^OeKU=fxE|xvt<RQlGyjd4U-|flyED{0HYd{Y
z8V}}q;{#Q4E>5AFoI%os4d+F{FAn(o_-wW7Fp^|I3YJG!-9eB;r&X1=&|D5Lp3Yu(
zQ5d@e8q#`x5;f5lt}O&a0{}A+^7c2@ctXx1S+-1ZVD$HKiGHRS=MOrIn^yOxBRfWF
zsh|MX0Ju2HLs}PK@NvP>s)U@?zWdoySzu75Dcvkfs(2SSY4lN}<t5qk0J5atm^g^b
z74{uzl~F<(m1W->4)<6pF4?`7Py@_TzHMFt85JP7+UP(PTcHn{a>jz9J1zFps*vm=
zdO~@WCVt$P9W)YS77_>YpCoUR*x($ap1E8Q=`mB$hua>=#MhtBJbX8HV=ZG2&)e53
z^F?H<7e%TkY^OK%vUL_;j9I_>8R|Y5MEdI7W6sR{DnI7@D(<q<V=@mh=AhZp9z6pV
zJ&hslCfXJ@WLsN(xT_)S(jq2~V`ehv<nXpX9@xl31CJtgi4E8cH<l-Slg>XM$LCyz
z+<|uRe&Jm5$ir5v+uA$H$obTNL_fu#h`P8&AnMo@1y0S$QLUcJ)ohJo`7C>FY5C=|
zhX}6YmO4xz`BlTGs1lD4O9QD68y!VfO%qbkFL1IVym>t(fJe{~3+17NGKY!gh?6Y2
zt_k?Y-OAkg@N$-Eqp_HCKTO)q^!AOemLvB&j$-Y|=mR690eR|fIeD<TTEDxa-W4O;
z7A&@P?NQ17)t3>l#Bn;%1ofdDZHtnjxRM`FD_yqPKmMvzEzaoa@n+Ca&h|_%vn@3A
z3RdrrhApUB367>m8=m4sxkp06FrI`1NlP+gfF3SW(dJ0M@$twh&A~t)+le^jLps2%
zNr()X@Ig8qm?JNfgdcW189%sn)gEjib4hQTt$`U2Td^g^;(30;_HE*h4psfpgD>{B
z_EX@Y&P{E|MdZDIpT<@awG)d-RIW1o)EbInpr`4)y3CIxwp<jkewyWqa(B38^*0+A
z&VbY-D+%#Zccp|k*?A$A#Q6lRUUm=tsE7I2Obb2j50^HS7H~aC(iY={_E>q^y0l5~
z=aowLNMuTz7ui|CE>w<ajlY8Iv(|p9O=&x2t#@1QvHe0>YInLQ*|~a(hY}toO752%
z!bq2Lh@P*-h@ZuPf2yy9^2eyh%cl5>Ev=ox*c`}u^#<R_a#0{#2I?%9YPigtKV^Mk
zo*dZ=dNYReU!XV<D_RIU87grhbin=d$$=k}MSG=%cE_9}++z%4!ln4%FyMj>hNN9x
zNL<F*o<xk8ob?Ii>EKpfB?4JF;FGY>-B>o73M5dJ_rl8t<u^-xSz5l~^KnR60eRfG
zayZCfp-Nh3G1S!1Ja&oQFh>9IJ>%(qf**$mR&x++JrY!Z-LbP)ZK73*c}cVHAkBLo
zEsWGBDXfg6T@YXKm3;5X7eX|>Aws9lr-mt?f<*IO%-p(-k8H#jwo;i1xcE226ljRT
zeeAUxNRHAese<H;*nOGTkH7g#gC2W9E86p?lYeZ(ALOVt5K{+Vt-f_R{ki<<4vWtD
z{f~iLMX#v^m}|uLpbB!$bV;f{_XGH0kCJ)E^WleD6nw*pKj^>>SpRqFHVo{Z1vjan
zuY+;Dkj6(NV^Gm+ZdEplF>$I@8M=y<8urmf0Ib;F<1U6<ya~U~u6&VL@NU7l9~?u#
zU_-p+c*_^P03t5H71Q?<Clg6v0gN7%27a6AO$D=jPe`s*#2!Ti9O)^~OU<%<O*?Uz
z_n!XTndK}krmalj&AIi}s%eX$uUB4=Ezx90RIulRoFa6t#qQWEDA56=D5SkenrkMZ
zz5A8ZrC2;Xz5q&^zlEGIIk$C(g&zm{WwXlzglDy;r_ngbc7mx9PI0Pdx)z70u(f_R
zkE7NiotP&vl@v>I3n^cnEZJf)?Q`ht{W5ZN6m{gF9~V0muEb3T4x`VI{SUMm0luIl
z)cq|T@bWsSc?GjSN|>&>-XiBEgDC<HWpAGZh`AiJP8q4{q4ug7AF~4jFBYflUs9{a
zOXhc*4eU?mJEQdT)Jo~rvO{U9>~Hu!E~WPRobduM&6$lUnStX<_~IUqFFHn+AQIZc
z;{|Q{mG8UXLh#8tKf=bAxH8uQl3-D+rU#e}1Eszx>2k-sP`dr@i*e2(RVc*v@iRm2
zU3unNtRUH`m}2~$uXR0DRk)$_rH_9h**8i#h<xU!%WWfT;Fygea%w_av>6;d_P|A!
z4lv>~!D?U`d-`)r3)kV{{abW^(+ngTAD$+()%jJ%n~+nBIxt0=LAj8+K|}=X;^R9c
zof44Q$sZz1;qG$ab?zDV0H!Yak=)AkRz=F~_fDOl=%6qc9=5LAV=5EpCp4Ld+_-n9
zRSldV_18s{f(*bso^>jj@0^b1<hnD(v&)IG#{zt`#_)OX1)~3jB=7fgFrRO)(T3zj
zqG}7clJC9>Qj_3)s{<A08OeT#(1k=+v`!?h>GMBApwH&jOI)_U!%?}Go*5+-HH_k?
z1FqLrk-`Tiqf>~&FAD@7C4x(q<WCJVYhgdpXY$;rL2F(XjCLfBIKs#pVzBB9o+V_3
zHVOBKAJ|z72G>N)5_3X2a>I2P7~+fE0EV{TK7Qp>_Q1xZQvJ0)+$q7=L62=%ui3DG
z&%$-cBQ%bAVsh(PzE|C`ocGhx&QF@xo3q}HfW~yVlZ+nQW5Pd!A_m=q)RjDQUu&Or
z9;xi*UJP7vY|r;BSnN|-M5`7$+)N4oR1WaHNB724lQAeidE3#bA*gWasY@RZP7nKX
zFi;}D#{D!ppbCje29p(RVzhm%rOk0P6;f{N*Uvp6<XBip;udX9ex6Fp?D=T-V)Q%V
z#H;ysRkGqlTFD2SY3Pr2b5s)6=&2YTsPuV62i!a-Fh7pz3EW>QFdv+!3BEeei|Q5y
z^HhCvB8a9}Wz5(lhwrHtEe8fX>l*D=JH#dmsGN~fwByVNjUsu7>n&jJt*ZluCti;t
zP}Wy6SRZhSh0r$nzv>=bwAd|6YB(u&egHY}{bi<7?ZWScPFPsC@J}#kVbXxe60Xxw
z*z(?Ug#B!b+Cr$5Mow|dCvJXcqww?xjkDJoNHK_k6wNT*DB{Tbk>At7n@^wRXj31r
z$v8o)$Rl|B)n4EE;9pN_XJZHbqPij;C@3tzSrE%-$;o)KHkxmV-%Ce;RHqrATA%-Z
z&cW?o*lftrn(HE_C%<TXY+`Zahcm+Pv`Q^;zEzV^Sm`XBs`-Y(ClGq}%;k^HvKmFX
zfD}Ak5}J+V7K)X=(s(G5{SakS<%daDd~G_>$rHT0bJ@qG)Sd6w8t<@#GAC_w0>9^d
z{xeK^-Z<QN(Eh^u1=X7lU_7gS<$6>@Pz<~DccRuX>~qz&w0nI0IO&t;>`a0EhI1kg
zZcc@&)+tT*j$U3tfOoLecNFHJN$35z?V6;!Ge51#b1Iv9o!fQM{(ZAD(R`TYzO)YV
z{`>cxkQK%q^VNR842G7ZR#dW*R9A9I0c{!GaxJ2}?9l?2ZOj*<nWh5xXblm^ey`iG
z@{uYbEY#E}h3j7x(=1nY4~`4>aHj?=GubY0)C;cg0LW5e?oQ$`Mv$1_l*1Kk^D*+u
z`9ayiw?k-ONk;Zme<nI^dRo!NVzn;_eW{<KP4G){g2Tv#?mMA#mFwKy7}Z_NLxr?`
z9v9^z3deV>G(~}^ka!8R!6Vb^&2bMK`Z(yO)k@R-pK%+!#`DYfN68Qzek%D`<scH4
z$H(QZ%Puo+`|&ekRaI!qt@C8T%bj*(anPL9s*;NWTKwv-*^<2<QQ-GmqAfN*M!tlL
zG`hL<UvYR^KA}<^)n6YIaEX#jUX4`tbZ5);k#6|FKkn47_Sob;zxiwC`*1lAIACaA
zur%C>2S0QieuPiYlwRLI=Ndb78%77Dp%>QpC|X&uqcb#qKj|;f)v;U94T0Fbq`L6Z
zwZIwCQv-8zefu&=ds8$CR9lW1fat~ZB?gUVJXF={-70G2d#bqa2X&kx?yklWYv4Df
zvRAT=nwEJBz2!}9?l~NarR|qmGmm+Efb)^ZTP9DP*;lE5x71~Z>MNUhK?jT#gsaZ5
z3fV_Bpcf}a2UM!89CxIl)>T%U&er3$pplV}(iXCu0QNla<Q^z|Zy84iWVUl95oxGH
z$Rqc)ue%|$CXX2YSX{E^;E54SV>k&@`bC3v0Wxl<yfMp?Sno`>!9VbI#S9l*uhD-5
zdYTdfdZ*9ys+E|z-pEIj_5gDm8t^9Lmsu(*t*_U0YVL+fmZ)zV>2m|DKmdvj%j7Y|
zk(W)gC>TM)!k%GARHHl#T)(pPTzC(e&6y>oZMH!a2a%%qWvZV;os*G%m)xm=ni~d8
zKQp3%*(a=Cnyi<*&j4;2W7zB8*W9V4*yFDxy3FBmIpUxBB7()6m$J_0osfo=srLs=
zTvI&6K+nBgOK@pJM3f2{`~CI6^l?>gkIjIiDKZK-f6l*I`(em>VV^Oo(+qu~%P&wT
zgG)c?=zJV^YJKF?9HU1_H7QmU;{+A}v9>=|Dq4H6V;G*Kh%Pf!%X|FC(Py`W&veR9
zGr^B~Lk~wCi5i!I4$-8Mf<w72aK51TQN0A4#3GJOKE$cf!xnOe?^!<^WU9eAtMbhZ
z2ejgJZGXtTA!Y2ek*^a?widI8>w;AeuZ?Oc3qY$TeR~d7#E9se7US)*Mtr-c1FZtL
zGt7)hmyuH6;S-wh{Fack3$K?5%}&re;+Ddkdue_fUDW7~9v#8^t&9|!7^>pQzE#su
z6v4_)Wajd#^ht;(SQa{wa`Hg%TSOdUu>E_+ZjixDLd?{a&1Rc)R3$D!xvi(+K-nua
zq$PpYJ}eMsZgkG$iTX^<l9J_@shW|N)Cjl@FQRVBfea_yGO2zw!=>p_pH9_+o=gwZ
zflXDlBW=mL)9I-$GFj(HjqyS20-kFHCnuXF&Of9n7KkCnsRgnJLpVuCRhU1<P-aX@
zyP*{pLb73buO(1~m`jaSdN0nx-^3F7UTKC7B!W4jj-s0M*Y1#PVpOWYi6|yy^L}nT
zT7FZt@|eQ&n@Cg%%eI)<!3@ojugH*f<{KZu?}1jYAla8@WN9{ZOVHR~e6_EH4}V+s
z{gn>c*nx&5vBVKa-+My^iw)K>Ns49pg8%UDQd!_m64~axH~a=^H<qJ{G@pQqU`lkG
zY}RlT0k__BV*Y#wP7qt52+{#L;l=|r9eAC+RYz2(18pt|Z{x`jcP^qgq5TCZxk;yP
zE={%$T!*btepdA|auzbtw8p;v$gr0HZ&LRAGI=bQpwN8B$;dlvhynraLkeDgx(UUi
z>eMs39!*845^^4*4)iJJP$A@LuT3)SN^nB!m^1T(Q}<Maf_kh`lk)Do{>!u@<e6zW
z7s9Le`Vi9MHc#A&lD~J=?pFpFx~%;*_bd)nrqrX~)yPKES!fii43n?49}KRCX4v+Z
zSu0fCyN8e!aVAPK_eBRNRgEBx-yvMBvMzn!=JL^KS8enUJ=q>RF2B8lPDR?PCerjM
z#+9VpW-yyqlr*-g(@*ec*Rygsz{jQ|3D;{6_`$*^wg_#4vUxzoD7T9oDc8PlzjDIu
zBYPxU#wy=Q2HW=BvV*EETzGFg<#{C=*rNLFe3X$Jy3>ypA-m84oPP1EcPnOnKhA~C
z{W1L*7a6_d^)45Pl0`aYtU`s6p2w=QS3e{CT5w;Z`PS-ZKr;AVLpAj4%IkdA#9yz3
z<1=siyc=!XRW3d1kPpu*I)uIiO7%uns5VpK#>p2GT4pRgj5p99?gZCfS-SpkiViSy
zgYLqPBjrdn(`E3M>~6TAp#j)R7iOU9YI80i;GP~>eX$ufpmhn9NLJdZB*r%sq1+Hp
zl+cF|_t~$V(jHQm^)o&J@R7wX?!&5Dv_kpi%k{rXrgkj^h=s>Gucu-q$?gQ5+#RTZ
zM|JjxGlx?(qBfwrQa3YbhfI-vIvHq>DFIcaBoWjW{(U>Cl@5&Yf7P(K>1K^cbA7~T
z`x*Gb%#G6ncGpu-VBg9}A44u8ts9Yy+eIlni`LRdv|DyrFw06!zvL}#XZEBBkwHGh
zNz3%Z-dXOgQTjs~yLI2%o!=N;KD8k#B{($mv6+!JA0siFQeRxQ;=)bxd^c1obvYk3
zNVxWNqAmP~!gC=v#T7xslFe!DHAb#OgOZ(^P&|dC+s_K!OK}QocaMB}1z)|A^ALA2
z*ms(F3WyGzs)V(rnx7MqIYhjn>taX}UAy6su5<5Y{PT3;OSYyJW?kau7*c|4^Nu1z
zw5xvTc5c_|V0oDUiFn_!_wpcw`6Dydf`M$@{o*8^rT9@IXAFOmH(TKix!qvW97mJG
zRTRe;=d>m})Pwo)BIX2gG7XEPj?iv>-d}+nr32`4IuJ|;{wRZBvp6UWKzrThcq2*7
z)-+6YRr<hm0s09Yn?K}7$gmr!64e<~iS_>0{>J|c=j%fzF<)yx`ac3rb@5#+cwPIp
zJjvR+V@p*8(Q=k}Ji-41IT=>buT{k6hG~3tD>ahMAths!*T!qFc$db>nn%vR@_7&<
zh%_Xnx7Ck2C|TuuR+xK1%Isc6m5Rx1*AgRk1WBSeEBir$&8l&P3c9HCU{x2tU0Evf
z>g|!abH%Bf5cNzDbtmvqV<0)ACEVsgP=t@d*}+p=J}N&<jBnpO)WmJMQ1lvr#9$mh
zFY0=Eb*#sWt0t1gqcmyXMN6f28Y?It$I4yjoDnhnDh9kJkId^fSAWl$`k=%px@<Pd
z47}n!U3ywC3s#=Ky7O?U`#|mzBn7t&&)0gu#&Ve5@UZtQS0x3Kq`uJ9**VmMTriS}
z^wrwCk(K;x(7|KKa5(ypwHs2AWM`0VM#>D!*6{{5%=iu?MT`kUC=q-}q|lbd7kKm5
zWqZvr?@2KHsQBBh&v9z(zQ#c<$R9p4bCBByP+C7$9s<FVwMki4=`D@IL2r*s#x;*E
zKT08Z@B*i<`Zg2K+yYK^VPA3ID&BL?*IVklH*5g5v^5o8c=_hvenVvBc{)vETc0>3
zxuw03mqbiIzu}jX`{`FEN#}f3sm?DO9`ZmhpL^Lgf%MOj-!pu<_)r)FDK{A>{O&4i
zG1;kA9U)0rE%t~yidWWtedvdmQlAcf<2?kUf;htW6>bdEB!Td_1eziu{JLbT4o8F+
zC4cLQ=nARE&TqbN*6kAL0=w+t0KRa?y^xVs7cFXrFJX{k=8L>ZY?$RA_vSBKE>)8e
zmIqHrt05LEV>e*^G)a<57J(`<0_#V=844QDoDy#3$-KF^#J*rWE5ieL_Wx~x(B9|5
zH?flIXT+oqyNQEKj}QEWmETD!p}xOO2cY**Q{R8Tn^<+Kaej^OV83z3@_8s%-gjZx
z!R4>CsA)RTSVGf6d0j*G{sixCquSGM|5Qo7Y6oAnRNXs>rP0vj{(q2kI_qAYvaKEJ
z?*h$t-eys#O>_VSqj>zf-n>v{YDXJlq+_|Cq3cU@AQe3J48FMwqEg|c{D0-<?~BIH
z<1a6X+`1=po<nP`KCb^xlHxnHKkY$bZol)-i!AzoU3D4Bj?5d~)7*JoTSt9KgCO;(
z`C7|i@A_995D{%3oy=>qo9&Mr!v(zU(J;GqQ@kxI!K3b5dg4}uG|aL-i4Is{@2mKf
zy}r@^O|Zx<SguGmhOYsEO2R0=MU1zEKf7Oc&9rDyB8A5%*jufuBSp=;2=X9+m%T}N
zqp+DTW%3Bau0s*U&;+nR&h9m`I4Yo}#YNSjd%%;m)79^ukh|Zhdu-tqhd8b-y)`Ja
zM>wn)q%?d8iO78`WRFUqoPA41n|>PBx^6aNKY*4K7<c(`v7$voCZ(((29o*aI_S~w
zH5=i*)q`7zM^QtlDsT*h*<($GqJ)=>BdMmir~gXtHcs@h*dCiZznwd*=@NHfNBfzI
z|8|=uND)keN+Ml{xSZ~LRy$b<1%<f>ZiFRMo9`9exd;g3sacF6Gy?wPph>eO7%?`L
zQdNtcB1nhaQ9U*{==O1PO5OC-Xs`Sy^`=`&Co|?;+(QI;m#x>`!l`ow5^)8?BkJ-y
zn@7^l7l^|ys~Jl-32(Y5VvZCdHHI7Uk|gDNFEDTL!(@la)6`#?_a@Fn6&pqH_v+n<
zdj0_|gmAScrNxC7cs=4+S%Z1=E=6eMbo0rM$lrFIDzzL_DjnH&pA9ol*zOgf^+=R{
z*iH7IfhLf2<~}F<s;fvzqZKo)g%H)Qyp8mp&Xw_i@&}0pr}KGTTC&%;=lDBsFJa{E
z;fHyUwM2u~nnoVTk!wgj5;D)*LaZ$L{e{M3t!{^p+})iAWD@2UYm9bN8{MsJnkBZd
z>SQPqD?P&ZE%;4&TPhi35R!8^2LV+z6fzuf(}6??`?_WM<l4wBTxcSRMq{)IL{yC-
z^qtpiif=<@nmyR$%VRR^6;hs>U%4*jWrfcbxYqap6X=Vqt$gt_uY2-i`EpLZjh>;e
z)JXYof;d4&mYRwc_4fJYw6H62tUZ*Y@@wbe*wMIKT~H>6nH>*OVVs2@D+NAsHy>W>
zUV<JYP_6sv020=>Ju@W^^>DIbf}h3b^8C<ELWYJT>wS)opE@FcZ;-!_w?2H@Z56By
zgoUO@DN{VCxhth*NLU1(PhhN9AxoyY|GL$Zosjup%A*4Aq|II-#9(XVVa#>cQi=-^
zl{8H>`Y1T&F=iE8S}uQYpgyy+X?EW5kxp#vIvwDGy4SxU@R6;t$0k)DVtmy3glsk)
z7h8%8y!#&KeEwp@7UjA&#6)o6p_X9z;ny5Te|X<pP;ZUV`>?hu^W&)PmrEvlH~8+L
zVSSkEZxHWPHO+>DsMpkA4eMH{fp<H#En87MloJHv)OeKSK$V$q#r13LYkWnwH26Mi
zZ|trr69e(-EvqWu42|AQIFV;ye;5~xE{0Yd8V~^eVViu@SoVgi{C%7}5ZVYf3@awD
zw-S+Hq9OTv6-!TTQK8aJkEs18`T_;+KQhf#aTOFO@pvY2q}dUbQ0ajzQPLW(#Zt<`
zJf(K;oL1P70+i;VEK!|UD3SHa5>}d=L48cH938MS6F>E!uTw#R%}n|8A)8~?0XEV~
z?`T8R0z>J?{sjnEoXB$LlL7drZu9CHvEJ&z)idC8f8MtR>&5MS8bWH`pYe;ER3L7k
zc>?`J$8PQ?+=vC6pd=P5bH{q|eh*jX(o_X;Hx10pGmBc}#X4x4D!rY>`E`PlrU>hX
z16$8YDROe?TT3CwtOUDT97T&89OBm=Y0A`JkGi4PBxV#2D<?;^*1?mtI<|R^?n|5W
z?^|%WgZk2nBhz>yFP?19_foxz7<iCZ*1ws)`fA1p)f$>}n9M(I(G5{2C!enVTzfIY
z^<Db*#z(2u+XC+3CaDGz+?^so*2XtR!+EseEC~BPC(rP^#*QYS7l9|}z%|2y$oard
z=ovMDnUR$JV73Me(*fVcv?+d7AxqGMBQ?ZlPo)z0bsVg(^u^!bWI(Xd@R2FI&C%?%
zx3C{FZ^Z(PS%qw!-!+*EU1EXRf1Dh*^7H2o75P2hyuH-1WXAc?@US_H1O?xRF`W8H
z2Ud|&yhk77$rdElNIDQdU<Pq{Rg}V^{51DU6|g2V^)g;zD-CPJ7_JKuBG*wn&b1`p
zm)}5%hCIdBmiHsO&+a~ce6p1cx$ia^Nq}D%&NRQ^b^c+SwaTgOVx?C^RLTnF$}|sG
zo0Yu%GyQij{N95ee>rDy>l&K_DMi;2Pcc3?=gAR129G<XLUAqG;vYS-0sD#NHLr^?
zYKt^1HsM^&(SPtZ&%B<jKr>T9?ZWARvGx@LbX9ZOCVr#gReK?|{6lPcP^m1}!x~g@
zqQRvtd?F-z<RLi}pWX^j-mknzZftdL4ma!?uPkM!ZD)Z_u0J`wt6~%v$wi!>_KS%U
zZZB-vmpVMj6yNwVtVVhHpncZ8|DI7vQ@dsM+f^UF3o(^GfgRf%Sgoi&9hi>#Vn{BJ
zp)?{@sPw)0Jly%*m-s7^+$m|T3NB(Jxd<p88qbIi@YJ1b^VF^S>k-DonJ`ll&lbGC
z=4Ei&`A$U+d7f4&Wa=X>*VW=#h!Rp`0IN(sdC*c<j6W_p{h2GP`j5y?(VwBxO>zzQ
zUMV;i3!4-HOwk~(jo#E;rpooC9io>`Ms>mr50tOa+kOvb=A3CR_vHEf_wcmoP538V
z>o@x;+XgaxCVqo*^L1+VE-z|U&I{DFWHWi=^Za{f!49w8Y=G$ZAXg?>%+Le^Mq1qu
zvXL-;P=i4p!xq&x*Xc3`-k9bU+iKb~$Bgihyr*{)XzB>W;#li(gQAL^@Dw@0vCbb)
zcV&tZ&`aza>4%VT5~vByKaKzqy5CMe+6#v!TVK?*czLoLmml<1#>{MucY|pw&NMKP
zClga6Do}9v+}qh__MAoePb>@pBd=$T1S8GnedQi>3H<EMcTNh=mAk(^upM81`FlnR
zKxa?-2}X@NzDZ+dJk|#x1Nhm4EdBZu;><;y?7Pg^Ut*WIAjin}2{vPX&UM7BmfE=i
zLqX;}$Ht^2jpsO)HomjHg{=4Iy>eenK%|h`eGp-?m$YEZ&aJT@5YJ<N+1IA(JEmH3
z8`1#4<cJj~wr5UH@D7^y-@Q3lIeztrFh(ib$YNW5!(#S{#jh$wX32Jqg%Fz9P$B80
zqs5yRF3S+rd;!ys+%Mm=fZyqFp_8LzXgwK7_wU>3V6wA9%QtC1mK>SrFYcQXI={^9
z{hs=|`~YT|;^qepgNr5@2u0B5zLxhxIda(R?uzBH%=#NH-F#mU>E2_P4e7Wu?=`>Z
zqia>wU%BkfC5c$QyPEHCGv-)xdy-l)==Gs!NGKhM`%Z#R<JFNDi4RA%JbL&Z8tO3v
zjgwCx00r6j2JJ$QBqcB)rjXBr9H|jy^jl_C52vr6)2R#N+#sy_`t`>^410%BNk`O?
zsIM+;q=~OyIy(UfcNd;t$KU7aKEl%hOc&&G;Lx->jrf{3m!P_T0N#b~CCy~Ua>uau
z=%FRarKJ=*YVIkY7tX^UF#h=6J2juMb!-NoM`O%r^Y6k`2aC`mWaWc1E@Z(=5z1Z`
z!91KxSLB|ATKKL6^k82<v4FP*zC6%US1V+1HfMFzc>413j!NNs$AvMU$v2Vujz1jF
zZT1=#a!7e`X$p=492BKr=+R{S&#B7Ywr_zu&q<~4b1iCWGc1WB&jU22KkyqUUH5_F
z>hFO*J-22XIvF+9C;%---B!L{Y4G%OR@)V)h2ad7>4zdAKc{65@+?GYtgCP@q$D}e
zotTU_e}ECal~}zm^PDwpR!l%&mhZ@}`_I<hQ(o%ZzC$A8={!7A&wns7r`cD(M_VSi
zX4FLECqYUi`9@uD{8<O&CE`Q<!uV|izh@zjAtBWrOY&nn`+8>w??Es2p8<oTzUE3D
zrM|{;FJP6_OR**KpT7G%x^Xs4h9#P}siF*YI~bI`oxKN}ewUNW4)sN)H*&9LboFoF
z1SM~gqnpitt!O@wR>#h&fQ%30F6<X$^ZBlf4PK7OF>=_jDiHq1eWx2s+h7h<NBwQu
zA|2>U4fxBxFsqM$&*y$GiNvoBa2;GLEqjypYs^i(le<V<`h1f;D29E-`xhM$K+2F{
zUwrN?x1l(=TowjuUc5)>-*LA9mij__^qDq!#}mH7Ep5-ppoqyc?g<d}MLz$6?%}c2
z&uzal1LHnd`z9z&1okI@&pt9bgmEIQudItERRxn!v8uffD3N!O9oy^>hq5%tN|*lp
zyA`Q4F|{wkQNmIE8k5T<)47S&%*{D^$?SQy6=QUuKkx3e$clzBO&}cL3;fH%ir()X
zSIg=93sXG}G8h6Kph3%^%lRn%+19&hUgJkO74tr){ZE_pOR}MqI3MV8F&)q^glVDn
zILA;UcC=ty@?+ZxnK)A+*oW{V|I6(S_V9-o^#Vk%38p`ILT4N{;CC43?;dve$!XEP
z*uB`%Xj3oS^AT&v=irqL5@=fIeP~Hk4@%G%qu_1u{4Eic$WcEq>pJ3jV-E49(;l_c
z7F3;EwZ3{~+HVUbLhIFsr-(<{M&NxsoC+{9Az-5f>*+`~(@1G&(If~2sYSH=>g0`u
z5GON(O3O>jf*tJbj5c1>@O9OsTKLrd`rdRY07pQ$zkDSunu=*;0RHhkN70qSBG28s
z7e3Tsu_Ef*uo#)s%b)FetESh-&^xKArPw2sGXq#5ggx@ZJ~)Oyd+q94I)rdWF81Ly
zo6{zGX!`*=;FTL&*S*YxZk^)NDm#BQ%P#(?wg3Iv1pM+?gGS4xs_%CR^%b@88w0j=
zr&nir(~QpY^61*TJh;+ysbM{^7t2oTg|Ja`D28LO*vhy{PZMg1?aRHaP`4nB&a{X}
zuUW<QY%^RkOod}Ze{RbO3wx9vN>VJ)%pswhzw#w#dwGo5>^NbbpIU!}+!=_|ySb#n
zX70bnBnbu}s$Su#)b*A|X}$`?<~d2O4r^_(tNt2-yD!8P4!*>L+fwNO2Oa2WxlJQ6
zoWbfc^)-ashQ!w1+X|8RV=>3DuQ&IdOZzzY=hJGni?KbLUlel0|FZF6=hl<r1|B+)
z3}-RaBqd*T2u^lo%V(zp=f3;g%hH)IYC#Filda9)@v?9D7!a};H*z;J;SnVS1tKz<
zRhpHcqK-$x<W@xdhjP*1M%b(;(9P>(XH7`v(W={ig3c}YZ>}cP_L~1e*_($$9sd9R
zBO>dN?8_*Ggd*8z+K`G#iL6s0TlQ@*cCv)AgtCV$6_f1yGImK>lCjJt`!Zu0gE5Xi
z=bY<&uj^dr{C?-}`DfnqzUTRTJn#GUdfg6QeuDN)iG!u7f)S^-{OtoIWkxDDX|O|^
z+19*lO8yyyY?5KM&E00V+E?PYl7{-AO`5sgr$zG2D~Ld(@Mce<GnAPoZ0xHDXeXO|
zGQ|u4_gFzesNo2>*3A_i5!_-&1v~C2>YHWtO56Dsq4r57=H(5+oH|w+=d>R)+nv&}
z!vTKfB(0uy0IHY_>=J48Rgl8(0$E}GcIF)b{@7iG@U^+<m?4{dQVel1CPWb}U0W3_
z*=;%{kR`S^<jKw@0w4xP_TK2<5Ku1w-Y<-QP1)kq{1Mb#o;vsY495Irdc+sUqRZWR
zVE`jX#(XN0%wj-=;n|M?+^fp_esx1zz3m}Rz}Z9=Kij@~DH3OZz`vT%LsPPQY-O7w
zbp=E%Q`PNFdA2WKYQcSLo#lhQwtuJ8)_MYv%fc$;H-uL|PF2tjiH~``C|6?soh{L5
zUfkYZ$Hc=ZkVDBIRdxdaU<Rz;VAp*-;roV@ZD3PXYt>cr7kFpS`?4`=;$66n4godI
zh#!gB2<a=uTEk0?ZEF>L?t5QfSYymNe+&pq+bICLR_ga0130wL6PoXrtA=tcYew~6
z2m4Sw-{0(eNS%Q&+UTC7zT3o@Zdh7ZJFub!^kR|mA^CFYcgE^7{idKjH1bDL)6x4D
zbE0K{8(5NNIN@t2YrfMe!R}&}<T>U9qI?Ghx_GMKLU}E+K|IcW3+L*$HtdYg+5WvD
zh~r4>BRFFtX-<`J8I$|NYzHp)UrQ*<d{0GgeDis}OOyXdoy86FcIIg`jM22bDb2Dt
zgNq3o$<qe8UlOp^rPCquUym~BC)uj&F+qi(MfG<Iv_wij^;GC%3ORG7ZwA%sqn9)5
zc6ZZTO})%pCuxt)>Lm5(Ci(A21IdZ$J~Ph1)N)fZSv92)raTYPMKPC+PgDVl?sytr
zX&p-*!vDo_-z2Il(9Dag!wO%XPQV8nIQ=~av~vV|?@idMHw^b=3cyPu;$_QZ|I@(H
z-R_$568jC-;~COlS9+sy=7%!F^j40)RuvLqRyWxPBRv6%-Y=woKMPkgWOnDKZkksA
z{Gw2_V4!kokSZRf$9)W#G+vQqB(}2KG&Ybnqj#p1q_fd}b^Jub+BT;sc4u4Wi<pL2
z=Pr4VFSNa*20^-6eNNs{ls1)jO;9AGTU~-O2c~T9HC9fWALZ{D!tm`e2xY-(8%N<+
z!ojag&*F(ZdSwOS3FXscTb(G8QEefY68Oh!oPGf`7UDHRpVgRE#YgITI_?x4B=<0o
zTO@Mz(&G;bfxFZ5p5DF1J~}8gjMN6HB^(OSE<ha7uG@&&RxX$7V?e^&V?epczgE+H
zxPRX<fENs0`87<iW~B25^Ygi4k`k_Z#6@P<G3Ia?nE)xDv2+nyFbUG5d*|nF!|MWc
z?)B|GwR=H{iF;Km(J`%FFWSJ(zSoOLJNu$QnqBG9UF^S3(E;c|O8b^-Fr%)QNInQ*
zKs6*|do@em6M|w9N?h$9Ub@_`5H^==^Z>idFYH_K=E!>vh_Mc1UxWDDG!Yh~ZYUoO
zgfPlMqrR!SCCRZ;VaA8+z4{L>pwXF?Gaw7I;e27l)p7(68AF4S604!l<#^!vci%+k
zUAUnVK2;q7eTf#0*awXM*XPZn`DGKn!+=UL-8>}Pk^{q&?ge+X!hQAqmLEoJlppZn
zAY@{w<0i{VyrSn5(p!ad0*ftohD2UyAJ*ac7^!|~uVps!G;!j*C!pF{5lU!}Q=m&A
zEEBZLx0k_ul1JhLTxXoP(=yaqoE8tUvk!r3s&KHhB~`u?Nod#@36JV<`uxTHP>`iG
zOY&&pJ4a<`jF$Rvu<Zwf!<+XD|Fk8Ah}XS+2K^?YqLlXcnd?-2VngrMo?wQGC5zCw
zYe5zACe`tWMokGlsx?DzeRZm6$QKNJ+lIq2amN5VPK*1)8j3AYQpSSo*wlP$Z}QWv
z!O+GHxDMCv^RX-j#{i*|@c&4Su7|ot4JS(yKaRu|QXSp2)ek7u3}(W@0}Gft`ifgu
zy92EIw39Uir456S-JuvIblW*9%$<<SzoOt5Y!aZU^F`=oW6p`a8E*rM6FYE80_CSp
zNaGlAMv=x_Y6CCoQFLz!ATUxA^%fpygg>3X(fV5bhk8H2qcBW+6W@mpO*#g!wJ0%1
z;t$=KZrY1nOW3+LEL9oUT<`re2mrvbo!}!;8nUAIue*7N^BlwW-v7R`0N?rUUj;Cv
zLY>L&;`E6;R#Gr-?Ap`<IizLL&cZFKEkJ4NJ~ERe65DqVuwP0!;i>4Vlkd0?tBtq@
zS;N+<rTSPO16KV+-DlUEL5T?T@3vlQUlHvQb?Sq{<eSSoC;-)KN_+z|xV-oMUOD>S
ztXVlz?mL!^Eb2N>vG1KnKl&I<?9!Gk?kPwere-zksdX#7+#=p4U3_0=1%t$_(>#lZ
zMIYjd{W(TCyvMHY&ohjNZ^G*M_Tvs8A{ZM++h16lD_nR<7Dwu|$Gx-zDYxjK^<EZL
zI)6~Ev$`XR8|i+dlmKF-NfRxZP3tGG{ZaBm(W?D!F33#4yz+fjeai>4?^(x2jc?jk
z|JT`2jNA1@-;L4}`aXTJCBKXE&^b>Uddl*2_4_V8(orE)&D+fia**4o!%9Kj0fBa5
zz}z+k5?6I`A!+jZo5h~p0XE1*)x}UoRVAtwq;9WoGoF8Abg9YZCZ1E-^Sf89A1g(i
z=-UK2s$*#leqyt5Hs>hS*;IRF6seQdxR?bzQ56Piq~jY{#E-|nwf;I@JqDb5Ca|Tt
z;}EXOZVm1Ftui%(_GjQ=`g&Ik`aLK~zI%5wMqGfbZu`lRU%*SvJnQt?P^>KnL^B3@
zl++^a`JA~FzYen@y)It4oB5zkOJ7EaGlTZd6pnT3Z(MJD8e-!x7%ZuaSP2nAz)1!p
zF>O*$;h%nK9IC}C`1RALNo%2-RrgOka)_`P*+BO1(x^R?sfE&3(|CVlcdPH6=k9YG
zjk#c6UASrZ;Y*s@gv@nJ?(Yxy><+&g;o7uv&ZP%I9S3#t6ErpA2q(3N(J_#c{dwMj
z-`_wUSjHkrF5q1G*Kk4zxrC4)VS725mYa@BmU%B?4n!4tszU=r_9K&nlG3M*Z-A~Y
zwV$OuKfLao(I!gGu%D!f$)EAr6u6Nv%)OSD1cg3&`V7~wypxgn$$T|1xRsvo%eS81
z#d^LW99<kA$gR)xSha#5-EHyYNA@RkbZF_!ha?x5D6XWoDB=C)$0SffyZw<K(Ta3-
zgn$IxU2r*h>cPJyHDdEoTX)3^coW<>m~jj{y8$OeK?}7Coz$cIqDVu8`WUbRHCWXh
zXcPAnAXA&hu2Y?lPf_Vy9g!eL=-T}o<Fng$<m0|*^)4}G*wXG_(I3ws-!`@vGW`ZW
z*_G%`=gAgmBV!Y@pwikJct`N0hZLRob>Cz$q97AAi@Hv(?+z8GV_@fep;7s#chyLn
zVLO>y2y-$U-(?-n*3YL@a(K3qVJzth{O+g|rbSKJ;wRUJ@I7M)fe(G0^RfKf7EPxt
z4}Wjiy-mR7`6+Vy8C&}-Va>&^d<?cjR0F%+r$K}#?HHOI338JJ&Q|1bCmh`m?iZAK
z*PWu7mmdhG?)Eh9Y)*f^{nqeA3HG`+dGrJnTwOh{**7i;%_BLN4Eid9H+6l2Wc{vF
zJl8@3h7R4R)A-Zle(mCItniqQjOK24f`SbFqOju&_)w(@&^~3P0QHeBMl&LzdQhBb
z6cu7!J-PE9&Wi!xQ`@sP5)wJe03h#C4u|A}^+sJR8!j}}zUS~lE3+7!OW?ED=GM1`
zank&bUi%`W+b%Rvs}7a#<?)fNB!|aof-%B2wPyw#N0pkA0-5jugYW*>{nccpN!?2g
z0vQ?G=pU-ke_;gH<#3?&R!xD?-E5)LCDk8hKv<_e39nL#voS0#&m5Szh0OIU&3Fy9
zR8l@`UHLy`Q6D1te4kGs-#KHOs-@`vGR{d{MeShn7yuu08Vjum-AlaQGPcg&X(NAl
zn?9Z=PG#jxdRONUmLvCmxf2MDt|_otSsv)h0Gy3C=unqVy7juVw|!lnj6qWZsE^Cx
zpBAKB#_zKmDnVtKx7N4+_VuX?UjDP_T&K|kssgSd{Ye`l$Ty&KW{2)XXgmuBrRX+(
zb@YO-h-;x3PZv(Z%2q_+uT@>m^0qTYn`R^br}F8>F@SNnMdwhI=6Y260%=CwBu=+8
zz}<+K+W@4XFYbTy-gtt|4R%e>?ykhdQ%+Rz#Jtu26(}%S>Wa0atsy4q%)&&A49m%6
zqR<8XOW^o|AhXdLvlj-h5Id!E%>y$eS6d43C|`B@w!1ee6K$*)G=(`zKL%_z#MiId
zoF}+bFJ2aBY|OSg%5Rf&!LsjuMRb0govGbO_d=_i^L+cGZP%U!0Qsx#>7Lt@Ruk|E
z2|(x^eL_HSO0lL`_E1L{S064^$cnq^Wh=umAkMT>CCqs1*46&`vyIYs6Tg<OSYW9q
z!Gy63RFjF$K3Ti)woW&Nl1IPmV~f`LRfvDH`04yC^?E{&&PbW{g?<}B(#%M!eVAjw
z8ku}=X%^m$z8Y7fV`IczRT9c52MI{`2uw@t*cRa(dX!tM-x$@?*zWK+B;g<<UbfQq
zk7|p|2{3VH9`^m)2(KW!6Cp7nT#gZY<~c(FS1ybgMluJL>=^%X5VGyqQ)lUJu_Twk
zIN+(1Xgc?a$3|uLj#b@#^<uv$JQi}TSo~@ZQ8GN$#M^qZ3o^B0C}0V`PCh!3A88w%
zhV!)jHpk04{CvwPM{590S<$p916AHdZ&tY{ra_V&2yc>EA02Gjo*+CKbB$V=0e2gT
z;^#jLbpkzBJR?0Ao*eQ*WzU1Zt%Z7@%y^TUj&s#|N?N*mneB{W7NF$UI7gR>$8~ao
zhu#@n{F=)GC>9JFhrCo0A0qR1QBa<^VACg19>=d-3V-rs5_}k?hy_gPOUD4k#dpCv
zHc)aX#s4Q=#wcs}?I_%(m;X<YwLf%UibFFXjHe}N{>qwX=Sk?QMRpwnxc=Psn7jDK
zIH1Vfwr;G*>F@d1HzYVDy$2ha#DS^<KnyPdg6ze{w6nYO&+Z#O?nchv(Xaeo1|bPR
z?g5}|v=A~&*Sa&d3nuicC+dfvIsDVCw-cNN(0`*V|8aOi;$SNjba;p6@{4|fS^|nO
zwYgaHizdg;+H$*+^)@TtU5G=B+k===HZDsP+e8m7a1<;DKLM@Vq|i8ROP4RyHx}=0
zuw9tTBF`eO7)fV5kn8B|z3uVEJ9^P*kMeRN)M40YtcTlCGqgvHO;+IWKD4p!`_C(Y
zmX!1T&(27RiZTw8o5FNN?jDuWAf?NuZNgMzoV9mP@*h2kD2wT9*ZWz#x>X!<QU{fd
zyoaq0S=Kqpk`Sl1Um?QFsT_jqZ>?t6!N-6rBMo+tia=x+Bp%E|&Bm8iYRhO}?K~Bo
zW_YIRzEeKlv2S)WLK&!jvH25##3@fvBTdW_R>V~&eMiJrZf}A=`V9!+KKTPnwlvMY
zQY|}dL`b^rX$6iJA<0qC6F)Zfj?lLmGQjg$bCsce;@UsyqVT|Amu>;_GKI&pao<Vg
z+l3_aBp~C@ys#_)rnxzy>$gkx`=YwKXpU4v`RF?jmCfUD0IjV3&nFHkfOt$^ebi1u
zb>&htPq*1lAYhl$Z#3+ua_vi&fNgUUixxipJ}^swxMbI~OhjWN+xR43M|2uZx0<n>
zKS5;wwvZ*F8*v+0=!4{Nvf(0qTRKpP(savqVU#i@De~!Y8EhD6r+OJkr5dU>@q&X$
z`JJ1baA?<_g_Pg@0KT0drAkG;WjF65;vav2&x*!E6Off^f$j_eRsX8#|L&JjUhD7m
zJut&|Ce;99F<IF5?Mi~5e*FC0+CGvpOj3)zW=6{I12R~~Av{SdxY-)E;_O-V2*vZS
zyly;N1kD6^4)9UbI@%LMM2-sRf{tzKq}u*$!*Sc8j_!cEG%l%#hnkw=z_LyuCf-4A
zL2kX9E1ZZp6itB|eImGe6JQw{>%V_|WS8dW{W&SBhXJdMhKdD4&Xu8OveGJHVq862
z<aZ7->*u#Mm!908NjWgkg@e|%58Y^&j*1lu>wo&rk0Ok~mPihlJC?BS{mv)1JY`v0
z?5k38#O1N_y*ejEi)&Lv_>)*Va?h;HX^~Cy@zak~ME#W)0!E4-R@&TN`{c-2*Va}y
zcDYhXN%2BWO?_fX7QvrY^ls7R?|oqarc^k<Y*8k<MIW->d}uj9)_HWBl!@<5Ga}XJ
zL|rUGp!YIwLZ5lyNH?CU_dk=PZW8pnL#eX6qf6o<RpTKdAtOV|trQfz&E^B)wCQ1N
zqr^phfc;IO7u`%C6Dp`vSOy)+x|}CWT8|2JODyR+Uss2ex|CF$Hgr`gwJ+5`Ce0>~
zm*cPWBTLDoR+G%j{%S7SWNg^A(AXx)<)+s5cP;NqAFw`i-<0$_@X=W<E^O_6oBn?B
z;#iaR8+>VZq9qQNu%@j==EBW5p5=dIL_W_{dBx+~u4I|$R_j&^(uW#%s0C*ye9K-P
zzN(-qcLTl~EElDcHV(#bwNHe9nGQX(EUxMBOLa(gm-G6{b@TISM`N#9%AN%<d2t5(
z%_}FkPcfRjflkb=c4Jv9rKeoVPGdcbGi1)-7*f*4KCCooJl<|-*fOcz{7)yGyX`}&
z1JS=zyDs_@ntU5=%yHtTD|R1gj#~0*866!Ul^}`A{ElNlZA@f~LFiEmiannR9(~5(
z+PW`?gjzvz7-X}x_`nS1*fbFnTl3lV{{5EM#{g*KO)0hO8)9O?H~HQW^IzKtP}82d
zY8w<<ti%d98(ezvwD0gNhccbXjOV9MReze^H@!h}1IL$AiO8RkHuuSs#EeYR#xDfK
z>4i3wux^fcV|Yq>n>~i8TsD7xfcV*qMM!dS^W4+beC7}-OSl@YK0(#h!MH(zg5(?#
zwzmAk_`^bfpSO(8eFLQ8rQ(k3OQ>-V&%4x?N$88It&s0MNcaNl5l+o6vwJ)xq(7*2
zHhVkiDF6WIp}OK&LtnPX=Lt=KrJSxma>8`K)Y7@7rhL0-w%ShK;V+XzTECFA0j9TY
zm5KX|21~0aSo)W?9C;DHK9@@wa<$cKhtAGU-TNk3Ec*E~2Y#Fa#wDDU|9<kjVnHi&
zXUUwCT_DlnV*u}woOdy-oxN--d~02%{E4W4hs0K5GO%D(q!%nPs;z(zhMG)%j29Ef
zFuSB(;rQ~d)4Ft%|Hr948L`!8$qZ6BuHCJTdT+IL2lDMhu0eeedy9&PIDJJWdVE9w
z&Ld6K!!z`=mekZR?a0nHM)GxJnLfE5O=qvm?4lgJoy^D+O3D2gKb{~GJ2`GEk$@|E
z)_kK6)6UT*>j>g+QQ@jBux2e(H($rbxfb_6aBGd-T?&lQl}QMZf<y4&G@eP{AO(Jm
zdRM1Q{v4lim}Hz+Hx~=zskU29tz@<Mw)1oFRh)j*NyM2dOvYM2sOobn;~AJ@r+tj0
z^SzY9ML$lx0iONs+p2|bLew8r?AB%5a}Ui}4_5WIo4z;=w`n{ypI*&+o27&XaXO3Z
zOvEq~WHBP@-W@N!R1)^I^oz{$5FFGXJQj1-+&4c~{L(~=IU$I}zLB(kC}Q$W0!NV6
zkk4EF>JZV0s_;<)F6Rl3y%^g`1W!MxLT*@M?o<`*b%+HoF)se4@{mOc+X{U>5Fv2>
z$94!Q5$?0^X?yq$a+|ujie}px*6PA?oxDOMoArt*<+`}o*gEm(XghIiXsJYRs8ObH
zehbcrB5M&c!bxdp<yDKLKd;ZHN6czuEIhE2a+x{Y5i>Z|^0+IMgJy83>q2R}M$Khh
zo)ZjjnFV!e#Xf2w81ylblZq6UK5r#+-R@R_BPk^l6}}33bBpM9jt_p<O)Ivb<?E1O
zWT}xpqo7sJ*A@2=PO9mu6^ULhp3V0KBiGH-@xgf9;hATWU;THW)uP65kP;G~G_(Bc
z&YN=vB9|_3TnY~Zh(&v;?g-xV@JVwC-PRr)OU*V&uT{{THXEz_qV(*3wpLjlD=yjI
zE-~Enk<JN%e#cgsAxn?aLzG=0)v$!h_t!Y}{_a8<R_Ya^vNWD%)C>?kUnQ%5k1%Q;
zgBY8=rT>Cex5pd<9NSM<B0t?dcv;~fgYl{Gn$X9*k`X-rmUJ(Nb0{T;G5J<X$;Ftu
zr0hc3af-V&5><v8IdLCsz04bkL$J{hesW#sxlYN9<uZ)t3&Clv_ME6vo{1e;^Q%ac
zfbmd%Q;QN|7T71O&{l<?ebHw<TFJbG@>zYGCm=Y&6%{#d_G3^{aajD5)8`_Kc)4$P
zlY2++4#pN#tB3+v9vR;4GaJIz=dWwXoM>ANE8d-re;j;?I;07UsTtB6?$(2*R95}D
z=uy0j@Z-=^>f5Bs=0~+~mgO_k6lx&?;ES^h@`g&c!NIxB$ACbB9g%(DQdIui*H@A6
zJYdz6YghuOTCYD=0DNU7j(~cTRdC*pLws5<EP{z2hbUN=HZTrjebRq<Wb9Vd=X*o8
z!s&df3;v<ipiU9k)E;gw1}`{uah%8Scd1V#9jqkPnNmH&#SPjhoc#1lH~8|4!nbC(
z2VCp@{_m%@0e_AGc+}(1oaFY~Hu18kGdl?tL0wrg@daC|72;^q7{RHaS7`v9dPbYa
z1lXT#d`zs>5OX3p+BYX|C%ZH@a;>6~Stp~lG9Ivxr3(=WN|AK;(+HahCWCOXq*eh;
z5#yNWWx#>71i5sb1?Qb;+nFVk_=N~j?S11tKX?p?6r~%R{kry$WyB9#XA}9+C}U2A
zgo@J^@bdspe|e(^timIY$Tn$s_oQ|p&EcyPSaEbQ{Rcm*1NQ^$-FXc5qu=Denxqy1
z?mkbZNZ`3oLQey^;i?OBvPW59A$W6;4fR$Rt7ZQ6PQxOH$vxeh5UZaoC7+oe4&G$4
z<C)87^LL**3*|gZXQU<)U~D}7^#L@h%JSE9#MLHbRmiV7R!Czj3ER#7M4eTW-7=zR
z`EMEcZ}wc663y@^cVnG)YfhaC3v3f4$J>|GK2RtrP{}Z}39Apeo;CoWO$<`CD3N2e
zX5(M9pL3Hk<97b~w|;Bc=c$V@FO)%vuFn%o@WZc;0cev|IxiD^8xM>1Q6GzG9h+Q#
z2zs2#cjMqH>g_SW>12WE_k{ABk9<JW!(CZHPFaAoyKCLaQ0q@u!*lFv?H`F!^t_R(
zDZBMuoAH6rp5U{@hhvOgu8WTL_3w~ZHM9hY)r!DL$sDr;`Yh<&E{YQ_a*OJZ>vN|<
z+azI?ktG3kuU}uhR6U$Bh|{dx#BZ~Qa>Cs^f$_?;%jDYD?dt34-myvq8*d>99Jnbr
zA0pWrI2k4hO@NS0v_20=Syo8mM<PcpFO>R(IyK)p-}kD<%yr~Q1!i=`q~dXZ&*K^z
z^2@(6rZ76_D7)>fV;(bU?f(3fp7_SL+hl9gGwmttY%DI6E!wy%ey2&c5?%H`8!Ktl
zBb4^lV*p3op&r$i5=B~5a>WDBSu4g7n`O`Kn}^u8{DHjrdl@PO?+dVmSm|FFj9=>f
z6?SRR>&z}+=HYXiMkJwqt&^;T_P%p5)s>&HFa@R2I@%B{bh=*pe0!Bl(nso>LwSV8
z#D<je?8WS(e3feZ_Z7JnTXXYwMDiL(0*T%FPNEY<<CKHOqGLe6=@`8Uwf)K0TmKl4
zTv2-D3tP0RqNa?9{aP|H)(z|{{I7$N;otxJ!G5XX|2XL1hSpeu<T*J*ogh@%k~n<x
zgbVTvLS{_BvZ?+B$D<Xmw~jY=d#RQzba(^(0&po2p>PafTCCH|?9Im%b*qWkY{7&=
z?IV*+<sBMLgbuD81KL8;_Y4DhOyl(j%(8RM4ffb%3(GqGa{`9Xr^=n^=0#FS3{DAo
zz@dw{O*(uUt|Sh5JL5Fg&qMHDUQjL&Pa9Z;d7Hfn!PtmH&xH(x3S72hyG3=TEV(Z?
z31-=^uL>48)ojEEx>Y-TcKqE`TeEZ!JltfhtD*jwSN7S|PUkTE7!cbg;}XhS>qJu_
z5#>&aWzI=GyYmp`$+HnGQceeLnhj^Mbi4C9=W`<Xa3baS_&)tHg_`|Q`H$gl7}*Kt
zeHcJ#@Cuiv-S!w3A%=s*3NJ*vJ2(-<K%Gr&z>H~hhoQ{Pf#j>w$i0>>LoY8c=aJA=
z#r9^V_21avF`p2ZNvOy+rh(Hv3m;t^uDy3@wS4^EHq|=<Dn(NHrN0tTu^vR}BU!1%
zKuN866~y3&q`Cv0>~xu<1whKir->dakYY>nAWIixmRiY%V|8|!CJiq+JVK*AQ`WjC
z#{vaj*C5%a-WZD9uO6+?gLvDY>dH?Iw|qRI1Jx4Du}e#=(H4Y%rbJMaLJ3fOt0ew+
zuEgN1^H#v#CCF^&;#48X0s;GWXYO*k8(5Nh46r66!d~-x4p4%gfU~`@)D6tRh4^tC
z_|BNNT4_lCNhVF%6;u4Z*&#0RP9dXSn{7|{!EBM!-PvmLUys=*;d`425CNa|s1W9C
zn?`wm?E>9WjNX)*_mx4}KGL3)k9a9yiDo*6Q2mV1nS4eLNYvJ$ybLP0Q@<MC@2is|
z%GamS%diX+2>G#hy$RI=Vd*2T<{AYH&tdM694KFIy3SvCA)uYoU<<ILbuxQ!a|4Uo
z9o(ZZ8S#^JF!{qF45@wla4O-=@LL&?=f;0<iu%v^^k=0GgU>x=iJ?6_%6>6XJ3@V+
z&3Iyq@VN`Hb$+Zu>y&nSI-pOdpGn+b-mikRN=eD!^wkz2-!>8llD)W7X6|oIVf{Fe
zzYGr2Wj>4y>MXQ9KPpCe7SbLy9s{Bs@S#9UDz+!dXf=giY9N6xR*txW4%73ayvlzY
z7MR})O~Qb>L}Wjt!_3HH3;gKYHT$IvFFcCFxF$w~I%Nu&$Tp$4(D1CiZY=jSl}Kd&
z_;-|c;2)hy*STzxlen77x}2`Dq{hFbJ&B$>MFy><6B=k{1Vb*MnB%wUK8}OGdHaer
z-_FTlL`emR#*V>gk+hBc3Yq{hmz^LU$yJ44*Au(fWuw*|V&}H95@0GG#dt#`_Kk4#
zLq3ydfGdD9)G>hAOQbi}Z<VjG9|Nf5hGRgbEA3w;@gMSvY++jisq5c_S7GYV;RS>u
zby;POibSWNE5x&YXB-3UTUe?bY@{UXNPkeOF_fFs&$#|PYzM3co46CKVfN~}aIY_u
z?he!7r}if|4<^6TtdBmn3DY<X+pdzW^Kt{t0-dT}2Gz5C*s?txrt>aLXSH|j7~swc
z?foE4^(BD_ki=JZ3tH5j-uT(c!_HrJf;&I70w+1`U$<{6gsxWoGMuPydYVabIVzy(
zww|njC-77)2Z(my^?Zu{^90We51K;(!GXUg(VKrn@a~bjRVbHaXBan4@2QrvjbhW`
zbA{Evxbu`bL8L`@<I%Ua^Hjt>UFzr^U1~Yd@zIFCi@kLJ4V_S*PY`~-Nrb{tV+c3R
zWSsYY9hny%WaY4R4EWZfFYu=XoAZjJyLUM6`scKDV}s{P8y3faNw1Ps%Q>AD+0-+m
z>c4Cg0Z()V*;sCz2FJ_)(zx}%UsS>VXv|eL7fk3qz{n42MO`&y>KBrFbmb;X8A|8B
z4#NLcS^1%=E`zY(xV}z!v~^$p^qq&w1~OrQO&ewi_4Ci2uUMAv`!P>ReR1t}!!|@4
zQjQt;Jk{3oeCETS_cOB%cT+ETZy`4WIUqL(i}9h33YJ8j9Wao3T{@MpAf0AR7A0RZ
zUjB7mSwQmqK?1Q(Kqfu1yG@5Iv#wJy3%`%w8C@1<4yRSXdM-yB_7i=s>}V%V{ig!;
zeb!>fCU?k>Nn>g?US(#*&-hl{7r67rZJB9fW{&bC)WlZ}rrbsr`+z?4H4sY(n<I>Q
z=VI;H+d1b<^`IjQ9HY+Iu7*ygM%QA}zEm2<MOZPPzzKHYeAJ6I{`Sr_b#6sERZV>5
zCUYY!>O_-hI^;L#H=i8+neM)`{45m9i)kDg#QCYKOGfU8|Cyb;Svqat?W^tgo18du
z`3DboU@H7nn6?1yD0QtS*T~QvUmEKR1gy%W@wFr>6t_~o@g}pbrxqR75@F(QAs5XZ
zDZKDSG>okI^P_+;ED52mT-Fg8_ZFCLAT+>Q6Lu@_<{prr{(DZOcKjGng%Bv6qunGk
zib~fdM|A+pGg&0_J_!b<*eH-Et*G_>_@MxVbwwY(+&4hCE)%y2%$-zw-~kC-F~BIR
zY+pBEbY7G=h3s~pR$S~0W%EG+ms6mnRODmkY4bdcvY&oe3ZQ==Q%9v^a4}imm9&V)
z^4D&jahek)g`@$y@>y;*xZXAve_85d#6D;S*svX$HB#pclmCGTI=rw7Vy6o^&RzN1
z+MI230^p;*WtWRFKlScv6+G|1nk2YA+W?p>g@Xvdo=rBYrP1WhHcKz)G*o;(i~GgI
zdB%vOmMUL~2ngH@Oez5Ckw9xjxDmFwL`!o^q%b`Copx466WtFsuIJU-pwNJ{pxEig
z@$uZckZ`H2gYl=?TZG0@DT1AR*0;(~w?@ZZ3v}<dK>u6FtxwRwKArH9uPE~ZATM=@
z#O%)F!ULb|m@Go)LB8$(CO|Ueage_`MI}(fZf<>!Go3b%#wbHIzYt0R_JrV5lo5jJ
z+cz)g#N+vo0rzqe<cpp~-wN^y5b>YqokloQ4?$s-1*E>$uaI$w;4(x;){}<I=lj)D
zt$Y|OXCiJ4Tmn1eB#x3*fjmp~Hs{G|@hD)A$a_<Yg2mk9zvg#azBn^07bWRk$GLWZ
zt}t(+mBKM23EGAN-Ox82S+CndT(xdVh`3%=z8BFFfi4rpZ5g<kEoqWqQSIWY@=zXN
zOgnFZwLow3a|M(oyZa|4T|2kz>AW1?<SQY@n<oEnSIj847~)+;db1RIQ&7vMfj=@m
zt`9i`pBm;pdJ+yoI`4lrIR-=^;AVGz%lrllP{(?(tf~fIE$KY#kD%3gttMZ+$H@x}
zKd!<aI%id6#FUPNmd#l%Q?4lPySaUQN=OVge;3Bmh{~dw9|KNvQ7u2N8WzG$b-W3B
z^|_y&@YXt&j39>>A7XSKdsGGvCbkFJTDX<|xdfWpq`a->bagm<;=E=4^~tuPkG)eE
zXkF;(;`oNp=99|y=Y}u>QxToNuTMz~al4wcVY%Jux?hraj^LX?z%jtzX}Q10Jcd$F
ztqu~vTt2ZC_t;d6df73wlo5LJdE|wHH*_{3{3ge2rxRHTKbb@q7?xvSbP3eXyOl@7
z-3^8(OrA64ItVGf4Syp}Ce`CI!`s-FWsFF$*I#PXpT017{%CCYgmI)_SRVs$4?Ho=
z@e5!O&4^UmW6kYM=OLetiM>W!cY0OAW$V7tbTse&5-<t1$q8`XSo3VgJ}ivC6!rfR
zU*G*FzM}qDeEnayCS1xuRV>ees$E(i(I*XSV68ecL+5N)zTYTduQ7{(7g5k8=0tCb
zg0SVLz{)GXkCks;jWWBTgP!rwH&%L1czKaH*_~>Jo&%227)mt=d$6?c4UF{0l;j{>
zZ%7Tp&YdxRcitiSFUAEkKNcDE%T@QEAKsaHGm=1;Yq-B^c#G9-=(g(HSn@tx#7wUD
z20(nDMMHaX>I!|JzU@kiTZlB`=5LzmBF8I*`9syhI`6O+GJWaZel;bZ>V$`!p=RG7
zr3o<@K9G9j`_b}RT2tXxs_2FU(a$Ro;5jqo=a{kf3-@FbMluTWETsR22=077yjc4?
zCT1t@eR105eq$@K);0szUaeQ_S&N}BPrwOFs-pRe)+cZN@@cIX2moD9%%44Z&{Gm=
z^uhWH{{#9-uwj#-vpvDV&duaYqK^T**JMaNJ!nw6yRF?-#*H2Y?$EZ`@bX0~TH%7#
zOz9HeIb#d{|Cw@LzzHkzz|*UHYFrm0#J~gt(VZteRWe(AWEmRoMLEeMr1r%hA5HIn
zrra_UiNX985^srs0G%l?s;qxc8JH(zz(%uaJHpvww#ZQ!af7xs?8D(&Kl?(top!3=
zD636mfP}=!gd?OqL1{syPM+)yo&A=X;&wzjU2F3C`NMKS)4-=YpZ<^W%$S-o^#2gg
zwCJ4zFj7p&1gw@NymIE_@Nbkm!Sd+{xlbc~-W@#>76P8wiw^Om|Fo#4e}~k{pxcqZ
zbd_k50{sXD!ltyrl>7PbkFSX4aEq+EMSeD0s!g94PRtSTtJe2)(RR7~_)cf)o#*NH
z{~BF6{X6I6^5tS9OHGaK#*KB&zyGJ%bbFC@j<NS4RfM8WEf2m{BI#QU&J-Ve{j$lv
z9uRm8sL-IW)CRM_03oyD=b;!i;xPb?otnJs?Oo!O=|pfBG7ucReM<3p{R>TgGi{&m
zKl5fTu2q`6Sd`{#zCsf_R2cFmK|2By-%cZ*W=R*<Q!sH#eB;Ax)t}FLePNZ)qpk&%
zw}{|IPP;VqQ4~M$SCBS)J#zCMBFGsU*}t!+c`lY&ZKp&`?7sL^c*3?fIsnfYtWZe0
z&=TzIHSD()%G7&?Eu{-x!>lLb(yEhmbEi_(cx|L8<E;Es>18Y}>G3V98Cn14L`7e8
zh}2PuRlcHvANIR(C$8}<sGGsMo+bmU0YzOss=(<8^wgE_mHw)vy%!<9l-WvZSww@i
z$@**isHxjT-TdwlpcyMy>#L6o^(<l1Vn#aqI|`$1%_MAG4NB1QH>U1QDD%fsE5)S^
z1*})T@tXcEpajXKFn0}3ZL{%7buoO8Poe+8_`U9e2vbFR#JRK8v7Sx~w;(wE=ion&
zvo^Ff4j&+{ko7v9h+<$?5vtNu#TjsZE1Sv=O@2&DH(f~VcF`%Jr*=F`E`Na-{oAC?
zM8xwFnkvaXy2eS{0Woy_!>68?@fg>3rc7VJhGbsgsVmB>7C9$I#Jj>v{|!2kIXvH}
zn^=etJn;F&in1Gjto1xXo~Fdi*<3k86v@s{Z4Dxfbm{;J$S^ZBk|WT2szB3vTrSdX
z<x2C7zM(J%8{*6<?G4bMv@T)$8erk-+=ZXMcL^pgtL#5t>oI27PtR;eOP&ku%B4=?
zFNb!ik)2e_GU-p@@ne4aAClvq9_h(4!f-4-vybTXZ9F!qvxdA7&*q+{KK{(OyB|8I
z-U*pwx<bDkWb~oTV)?qt+}w<=lY`r91Bcu$DQD*UD&hd~k~YDk!b<QZCvHc&6xlDP
z&?&1e#iIyHz8)7?b&s|<bhSDo*;k$I>d1?ZZA|7mTe}EC-`0VwN#DLra?R=;b83>}
zFkSIXc%SWuV`U0sz^DsqI0N<!{Rs~7#{efoi$yeuQ`i;QN!EyYO!n#TA5B^Y;I_0p
z`Xr@UV2)IC5)z$%MjL#s_0*UxSa^YAb#2cG5;xtBU<gz8U^euI0qGiu3$7Q&?vobd
zd|VZKPxGK1t{!s3o_&iRkZIRA(dcr_Y-jK(sOvR<?uNku-q)t$-4B^kFAOFmefSm5
z@kRN}j_NixG1S4;tkYtC?kZI)u!;F=NKggCkIU|Fy-eirtc#4%ixIglOU|;4eQjq$
zel@|6ssg0Bc1JodxwSgLdub0VlZ7siM$YYG0{R{#x%~cLk+71NfbH1ipuHGNS9S5L
zn~(r4W!09PU6@&v<fPrt$7}oc0#)bzu;*Ib4|N9HK-7GWw)3QxFtja^{f>iue^*7v
zYEB?+*28W^c$PJ{mhbJH+^7J0r1BWxI9->NB9ur_ccgHYN}@_M7n5p3MmIzbpSrnz
zG+1A_DdpAcRatDu9TXwITKOLTC#&rNd9oK33#rNP4%$3n|LUP0CVrVAvP`!egNxy{
zvp17i01>n+9LfePFHgN^-doQmt8`zdfQ6{~I2L#6(@-A8_Ua1z#MtYm?@t8`h@zMh
z|KRT=XxuwMQ@p7Rc+^xW-X;x~&(YEL<XfDQQvB%*`h55T^rB{l1V97x!NjAm<;zRy
zo<nuS56ro=z4-uC2aqbW*@X&^R<+Dkq4Tzj-0gKiorAyEo#;*zHz(>3_6%qzv~dYd
zR|G5R@)7dF%+e8VN10#&`p*Nqj5!Tb;_GrRvsZz|p8IM)J$r|u&|z1=kM8AgMB04T
z1%x=kvu{$1{%yC7D=f6MkB7+7N8B%TyDeBB<j0sKe-hofx&`RvP6@jp#e9J7qRWmV
zT=DA=>z}n}R1GiC&+F9wZFV!@-UIbV_De)Y$j|(zf-g{qlJ;q;v&<>;r%vs}$3d2O
z$J~2Ont5!=cO?!c7xmGJ7dI(#N9AsGwmGZ&N@yivR#h0Yl~tqrH)yJT%}j8om6yR5
z(wsB!-GlpvkF$RgQ3k5RysLw<Um$|iF$0o#oZ{y%*6cw(coVO!dOKbEMaYqN7%R%P
zmd`I>0CvY$*etdN%Q^$D_+ryjx?ESKHSDSN{NXdYT@HXnD*1H4X9I%!G|Lk?5Qe5i
zipRWCIU34+{x(UIRDkqlIsAU51$zPnj4Rk0hc~ZOkwR$2Y+S>;;pYXP`CM{1V&0Pc
z@@<M81X*O3Ld6g-Ndld@V0lBcC$&|Ml02QoQ?dc+H$+VC<onnP$}rBg`3rypGwW|p
zd=P*7XO*Uk&sFNgh?OA!jg?)lD?2ND-U^@t^@2WvN^LoV_^6XNN!PR7X*WMxIPvF%
zOOJn0%hwfqdI&Itt5b`}K}0>?ju1f@)6!XS)nfp+tcs@R>m=nf^Sr;TLTlvRyF$M~
zWANjd7UbU2U@{9#{bjakN@7#K?i(2iuPfaNetg|N;xYtx9xM1Zp(aFu%8H+LbpEAs
zA~!3yq+_718@DBKZm+f*%C*Is+ngt;_wi3pK1;ONp&61(DuTW&r`6};y7Hgjz2cc#
zRJP_9{me6M1Y$e&blRyd#R6m5<Kl8QDoK$0!iA(TSsd_VfI#nN*iSeHPBX_hlkuTq
z%byo4SI8g*dp7WVx33`4UtLr*NtW?*6S5Nq&}IoPyZPLyz_HGFGsV@GVZklFab=?q
zV3*dv&{B_6i`a$5<z3Du3M9G?T3bi7iH}JQSjyY@^)qbTR96=of{-}^^drC$R`lPn
zvosC-F+kL2`M@u&u-6TCuP%A*x8mEBlyO^ksZ${j4@1cEd#5gR$H#u)(c?3D#;_y{
zNZ~iJ^W{o7WB)u^{{t%1>i(f5!kDA^#I8)>C5sE9>7#BVTMvNc#xs2*<LS}F&@;<m
zEmGyj4=x^GFs}2u(gU4F_|CP#8xoWYVOK&VXo4hY&*z5+ltkAcm4dBZL5ROcktoll
zE&T!w@rkV;AIspY_<ZC7HHd`l4MZm1jnK_BY<{g_cE6+dmsV~Vs&0ZQ?54IUEHXrv
zI!ecfu)tk<guT&}<yY$n^F=YN?L%lL&xXgDN5y%!${9IfHZ*A}<lk+Od7{=zcf#J=
zp{=ps8S<9<$QnCB>c$-hsdrg^K&nTOGyxLwaq~AaCHE86g}`@MZL<nKAD3fgU%h3J
zP?%So5;@*T1lcTO(o0bGe18+i)#}m|mlo@j4MdO+JCgNB%c%KXnlg#k`o?x=6(>!<
zx$oQia(e*FMlJ>Sa>Cs{94a{HF}G%#Pz{LeBWyEQ=P)|kH5BDMb&UZr&2No6J%B)D
zPkxjzORo}Fu8Mh7G1-1?_Y6F56uNnS2WJmu5ij7GpJSpn6F?Dk$vP6SCypcI8%Rl?
z><%^@<T!2cbm#eOTz~&;^6mEu?M~2hv}+AXXu@K|hEKyc9>mQeRQ}5&z-0#YUG*0M
zhY_use`$<|ci@VI0mU_hLGc3ycj>-tH|oA6C{RXdgJBIW^^gh6J!_ZR`S!J`o<wA_
zK3~6C@(KnS*Zwlt0*#)NYs`jGLnY6R))unge_0A@O6D45ff|{8FvehMhNOl3FH4%X
zE5`SIE1qEj^u4|%`&CI6yf}q?6r2Q(VSNPdxkxz%s0^V5e#|D7w2YkqY~KiH$d5UO
z2C`agv6D42&9?MbXlSqo(-)`Zm8YblJR^o3m*%opt}h!K(<SPog$<-rB0k9qT{c((
z2n{2u52fL(?XU8a4%*il2#2=bICqXYa3;?B!e4LR7Kg<aJDpApT7tHS+Ayu@o}^qE
z^D`6<8MRouPa31>kQdj3eC>&w7-jmlF6m2xqEF~+3%e1wl<p9s^;a~DWGc(1#^0jD
z$PFX_p-jr`s^Gj9$^v(PK>G2nVYuzRbFzFgcza#QZ};VZ&alt8*^hfhZd-wQpYC1i
zzf@WN(cC2vK@m>OZTn!Wxt4)hkxrEp_>~^c6zn{~Ckehxoj7qF{I8@6rKAu>XA}uZ
z5C!z7IBg7lgo!Q=Wr*K#_Fz8uiui8CqUR_V;*@*4Ph4u8E=0Tc3mO@s<ap!Ftl|xR
zk$KP-W@dgv@O&&&M6q&))%syw+ry(a$V<!icsgedH3s+WAhc(M5xLGOO!>*6Fmu<x
zwI~$9(@>zX|E*Q()$T*oCFQNF<pw4;4n@CgTn{yAFRQ4Bq);?nrP?IIrN5<l`pK)K
zyk)K_+udG#`iQhSG;k;D;BKu{mUc##-}C>=%%|M{%zXMEJ;DtzVVV~Sgq~flgy(f(
z@XDNDpo>Pp4~c45Xg5xxe9}t%{;y+zbA?Oq_vch664sn7Xj|i~wn}k%0&Z?(3O08<
z3ALW&5dw6WfE1mgN<!1pN~vBK#*_{(-na>hO57;`x{zVrKu);u^az=``FZh0<<sLU
zN!X1_FKn195tF|j_hH2hRZ8VWTPz|~8FGSOPZY(FUwjPPn?eRokV@AgNt@_Ni0k;P
z%9bbPXN5Lu2fj~dH_Bd#p)&!0x4wBW`lZJ5TFuzn`kI2|VY20&s@`_phkycLehU&K
zK!6<#)VRM6w0n2j_SC|*(<bdUBwo0&IHV6O+}L7xySHhfo;ZA0^u4Nl{oVG{j~jy7
zQ`T*c0speJZ6_E+dV-qT5##UEj|vkspZ;RsJ||(1Y#eqS_ZQ$XpB~cQ`tLIj^w`P@
z>{6ishPgH;GNc>Jd%~N4u~A^b;ftAgN6_E|&8njkpJyBSK|?%>e}sgRV)Q12_9dFx
zkmYG$>SH!%`lT0-_Dt2UXT4bDxbMZzb#z1fdG_CPZ9)j)F}hIjCxX+oIay)sYTIu`
zuhDzme<RXT&*dZXO6zpo@u6KPAVQKf69^7>ykI08RDR`UW9Vl8?1&6w!4lxhn_~c#
z{_l7rk+z!A>?Gy=+f^z7$Vt6@mlWk+wakC@&&R(*v)`+HoJ54c=%mPSOiMeZp2~);
z0d=mf^FrU5!;7Hls)lSk*U_A7=CW$O4?Di`PJ6id#Z%(t6Aene-(Sxi(L9PhSt(`h
zj-dc>3W4V+oeb?d29TA=if=r$8>4>ZzQ#XBru5G<Y=QqS*VK>tk|x)A4lg^?&yzo9
zo+z&7C{>2A@vfb`)vi#75(EUONiv<iG2IYxrt7POoFOzSl=Mi4P0SQ>8WHfzfvQ7h
z1DrJ-mvmqITs`*vb;vC_iTrnO`wWCz1x+67zj|4Wta1FZ&eClGjj!piRS<Yw{rlQ-
zK+;ARUM7y76csu7yT;}5jj|2J3)S@*0e|mG#N8jboWN&%V?Zhoz}Om~-;2bw<Mlh(
zMdA=nL_;-bYn(slMb@ajE%8p*F{J^b)M>GPR1)Gb5&8y8GL|zl7_C{?$AdbAtVEhu
zzhC@JIO7G@HO?(VfBlX~wRrHV+{@VPg{!W0sv#i`;=YNgLqGj$mWgxknlAjER(Ho@
zuA=iS0B&a`C-j3y&vOwMoNu*>9*H>#7f*1rDqV8|V@0DohMG#3*x_42^VL`XzK@ft
znol*TjBmlDE&8iFC^~rjpZeafQd6$}|L%LQ8lfHoI<l0yvCPhNellkv))7<8JV)T|
z?3`E5`(Dg!y22KrX*Qf+J-N<V8PY9&xrQo7{<J<%^_NtJL;AeruJmwiW?=ADTb~w1
zj{%>Sym?HaUsiL;d+1Fr#6wamHU{ye*wV#9=Vy1uq<k{)<s%JKvG;c+t18Ze^}Jqq
z3~@JK3ubllKYjE?&9qwI+g15;iXr=J1MY;^kJKe1&Ri%#{m_xYLHZz5f%sx7E0O9u
za^`bQJ1K%XrrarRbPp}t`!kr~&Fs6JNi!hroD5+zQqsj86!IM`BVY`ivw^>u$68+9
z-IugCJ_fi7H;of04kIS$iIStGeE2UI+t4xKv@-ePmrdYn$V+6LaBT;pV{5(3-A;-8
z-0nwALV^RwfK@ZC3Tif`^{B{fO-<PmTYaL07uC~u_CuyV4+yP^Jza;~W<A&``ODXQ
z@tm_E=BmTjkjeJC+5FDePzD6J?nHVWCSvQyc0_cyqtmVsTyi`<=zr%<zB|8eRqJ^S
z_*S|`hIB$WSF+mU+<&x^C3|>tZ+BaKPi!rIALlM-*;M{ggMLwJ2g8i&fKIt|oCL$!
zFI!7~)z0AaEdy<2t}?ps`Veo%5b^J$08GQk<Fw(xBd_uD;le+A9t}|2&5_W4*~0ja
zokSq;Bo#T6EUj;f4cAPk-hFvvX~iT$Q2Y$|b%Xm^1S?ssbNwZ;2gagExv{>0h3l;k
zSz3ip4i4ZBI3gg?Ra<$-fUeN3D^sB@p*y_RnMiH#VwyUs<O4y>IrKvB51npcyQGiF
z&DnQ2hRq+$hJMCWX#V%m*jC_~6T5x2uIzqlrZTs97K6B>qL>eQz45_T%3mQ_L|x=9
zHymFbcH@Uj+fhE4bAU8-vvGX4vAqW2Qqhj7b=emQa+9im3oKkr)vSZwD1h%p(HWeN
z0X5eGQQav16@d_;NmWyWYfM#I7G39#0cMh1>~;)ECPamtuYHg4*rXGBRrx#(hY|V8
zucxY|KPMdCQE>NQxv9v)x#A({agVrAF(P}Q*HTYZ>YD}O?|3@>L|<$rOwULsMLT3<
z=((M1W4KhjLzMZdO+->-4)0*$Nb6%jlqCM#Rd`x=n$qTOY=4k~4$urcSS7SN7#Y3|
z%4o`ugjG}ClLifHTaEtG42Ui>ts9L2nZd(*8JFkF)}W@Fy#T;G9fUab8(~Yhi|{A=
z`AqC~06IX$zc4;sl3b`B%Pcz`z=2NE5Sq`>j;0=ws!e|S_S5C4XJ;+D`FV&>M!Obt
z9*qM8l|y+?i)kh^fRrmkq6)*bp<UvfRNcc{)a%pTSK#*zd>i)N9toPV$6gei6=|`w
z^ypvIjsV^^SIe&ojlNuIvD(#i42Y$Hd?$=@L+$~;yPtsJDlb7MrWQDr-S>^AN2(~A
zp(>fAJnjxuXd*Mn#1X|<qz(W^nvkI#hnSX|rBu|NZxcJ&#jRtg3%obAgT<gLz!yM?
z2u=VTLc8RiWW&P*_a}U;Y(KRnK|Y=T9sRRS>J?oQJibY_fQTVL#{fwdv<T>|C5kx|
zIOqCuZ|c1Wgv66|ARZ6c-@sCSl78eWPs35@5NT5MoX1e;i+hZ_{q8isX4y7%ngTVk
zaPFx-AyN2%{hhh+$iwHXZ+XUs-pRR_$d9?XerO5DH^j+A_X_{JiDQqYB4F~wCmq^j
z#ARw4Qx7{a7DH9YY$xDUPL6|gA;#cf%rY)NFO+{{4WUKIbw}s%jVtQEwlHyW<;S%!
zKVvu_d@E05M1CA+_9<it#!vZ1Re6#hnK5RcyrW|QS}-GvN3;c!6;j3!x63V#QobOv
z-*a1n*SZ-+m4-(cC$qHGN>;Ke8|_^S(b@+eQY=o_jl2^Sea%vLHHXscQo{F<tz~HA
zh~wjFS@V2d`rv^V?Zr`hn=+iTsWL@n9>jJX-0Sx`pJ7PwgL(X#0T$05Ok&eXDB9WK
zM|Z(GR)$)`pYIc!^Xr7p%)p*q_F%>JkK*=DwXJ+{HW+k!Fe;xRclw;M`P=_&dXrKA
zjj}fnhx+~7het%Rlk7$)%2u*wnY0jvR0yGxkR*h`n2~)Kl0wEViOHUA24g9*FOzkQ
zv4vsAGTxXOkI(nM@8kY{pX2vCp5ysr{+i>Mx!%`tU9a;z-`DHBPMV-mtEwZT&kE10
z*83euB~Yte%Kkh`YWn*8)@P$y2TjP%PwR2ZImg-&D<AkTN5L@~^9x;w_ixQLZH~0o
z*4CfGvs;NgCK}BBzMhG;+zI5>Ym_?$zbB^SH8bEyzTZ`Q6{TM!!#Gu^^$S8^CJ<4#
z34F{G3QvNl<0<z>%$yCu^gL&U?b2}=cGutKE3G$qiyMo5qo5rbMR0pAr0VKfY<IKJ
z96BWHZL9b~`sAAa`V{w{B`s}K5n3`y+%@o;DM5u9xWASXA6!1R=j^7JVXl_CEf>2p
z>SmpmF&OZOKz)&+qr$c9cp9#}@*V5-1%<Ugeb=l<+F>tMq;uBbetcrddQo#70kbwZ
zYaUryxKRFyUiK}lxyJrM%#sSud*0)Ews($XV8-SVyDcTg2R50y%r+l$kJlIpYGImP
zV%Xwb4ngcllHeie#vup`0VE;7j@R5j{(a4S@*yaBn@ezOmv#<OZA}+A1aTgMW}$~5
z;s~P6<=`e_5JK)}uq83{9z`w<sBfirc?z!}PthEOTJgvG?zy$3w%ZoK%0AbL<^Nna
z5Rup{ov!Gn-Js{7)dA&T$PEk`74@yA`DJFXzL<7zN2}9j)C2zHR+;0G>^rSr^!Wh@
zF{w=na@Q2n(jb}p<%Kot;=%n9^|L{7QL@SP;rVY<yjI$Gb9n41$hAfSm><e!GN87c
z7=y@bXH74i<a~5pN~i9Uu(C*HdtizITgDPvnC3OsbY|`$S`udSV3>AQ-yUKwce)|t
zS;Sd@De3BKRp=SJ&?72(=MdzI+LvR99I*__QBf)FW9YsBpK7{JrXDZ^Hkg2W!-jQ|
zPw^5r6l*u%aPje{V!q$yUJsXi&$P2xSYqos5!|t}^V`bLZ`nB`AnStRKg`>}QFT>P
zj-p9&DV=iF7x=J*#dCVhUwR=*xt97uzoMO6|Nh@)Z~8Uxry7hH?G}zEK#<VsxaVL#
zTYqlS`CLvfi%_Or#0RgQ3q(-GtoMmdqtQTC425`e@a4B7{lMU>86re;xR$BBzaj8t
zGNvLdYYFXhRYIj5G9HJWsKe^$akSMAE?+^%c5WD9qLZbkIK_X-&FFNO+kH86>mbJ}
zX>Es+=on+mHI`Q1MdcB5^`zR{c3u_y^q#bJZJ%Ge%dr1x{UU;JnVyGPMgm7Ghxwm`
z0h9t|CW=ybn^Ns~YPdDfA?d4|-x8_oPNU6f>T&Uyr(5hW`L3Qm(0Wb?PmL=ooyJ3{
zBO=-n+^yx%pjNaHX1GW(&mizYDRbzzqQ?)CuIYpg#Nv-2s-GHr2+BhckTf1JEigoH
zb8$j!GC=C6B!=F+KrzWpd%JQ%m(SrW5j?KYwx6Bni#&0}o0!`be)|V1ypPkT90Q5J
zxt5Ppzs?YVOeip74ndI!vMz$M0{MSGLk^EJfSv1)!E?MhJq(w2n=ne)K2HpC55A|t
zm;O}W;=*Hrc;;P$75~&;nE$Cw14Zx7wG_kaGUr?GaT|aWf;A6pp<#0$N(sHE!4gn^
z(}8Wu8{5;v#apmn$HZCpCdE&8-PytkZgIJt*T`$+O=)zggAe1ui)ebxPT93?Lv!9$
zC7#InTR-&Q6tG`qu4<E&p|AO&8qv%!<tjjd?9z&6t4xqdx-opF_Fd5>*-0O@D2bVC
z>%o6;^N@BW{e^1B<F{eyUeG{d=l&h|OE1=#^#{LN`&4h3Yq#j#>V;2?rG4!#HM5T?
zKF(2aR6%_*{>RV@{*0(v&(P<Dk3+z4emlm-jrwe`kekK|{8q~A$jA8uB6CTM@ss;s
zjqeFh7)w~3a{iPEk}SY=AA-accibSe1?e$_+aKQ?f&?Bx`KQ>|PYLTLRU0o>B^0C|
zf5L%_r?kin%FwgRb&aI^i3p*UZ0VQl2a4wNlO@|6{ot!L>rY;okIQ#z(vq%J%k=xL
zy@AUCd$UO#LW%azbq>;qYx}Ra{vhw5x_L$;TwNie=nGJsacvD*2<uW^g*`q7Gm=Xa
z5rL={S!+>Umg-o`0Ve_~f+l)udH-x-!Pkp&<7meNVBB-88Q<o(sXUHoKy1s{5G~Da
zERNAnVmG(c1&&l8Asa1CGuUHD)6APrquMjfqlA8jqjF!$Lt?>Y{toQ)PjE4SrzZ%^
zIcYyF2YE3PHBr$21^=#?KA<q-iW1)tjjs4&$z*8=Vg{Y`{crtZ;~+;dOq=cH-gr2M
z${Vhpw8Sg5cYy+#rc$6FZ{wD4=CNQwm>xL^>+LQ|{}ZGAP|2w^W8n}~FS@H6pB+{r
z-DHi0a5kA97_NTERslY9W>`n$AT)20V^bO<p(*gF<P|1Qs&yo&_z)y$O$z&h=)s3G
z_^-<y)Y-OL(LyQfUyUeyn{*jM631T$&H<8I@aaw^9n%{7WV^PDY$bZX<WkIKvoZ#1
zYa_d76oQ~BC#3Htcko9%uxJ5sgCx_GK!QvE+<;0yqWkt@d$@dD*8k1--wSLI>LTEB
z%e{rrI_!jhboEx&^)Ghca@kjY-s;voFfrIIPl;uPrGG`VUS^!``0@)ukZ|fa1o6Cj
z7->NuIj5aU+FvO@D!=3}a)+Q$--qs;$*V|F`1w_h>pEq{I0H@|i!GVgK@>;(LE9kX
zU6pM6eip#`fP;1#xb7pkm^wS@ar;5^y6JU0r=Sbj9XCGRkg2d_n`vH;y&&f4E#~<;
zFZTS23yVpIpfh&n`zx5flF?o;<613pXB6@I0bKThzfrEmO~F;k=zk0GWg=ewFO&P{
z|2Dn5_v#P6JZ0fvafA%i_vENGQckNcs#j^f0%qctAsD)so<8&p#22J|>v$5gmWb@J
zv%$QDXOH*FJqv7p>dw$4FNIU+b*R6A8i5E`2P4)%L~zs>+UgR73&uhJCv9xo-nvU)
zK6O^*?JGk-S3CFEMEjs*sGDk;$c5MbeE5s8+Z!hjL9OQzF3l&{eG@pWDcL!8L3sCx
z-;2g#i4U(1o}pRS_qpd<1luf_UyAh$f92cY*@0qq*ObPaQ;<;`<vh2_YHrDT{(6+L
zs?fiK2hOprjhEXAV+i&MSMKmu{drMW(QrWQyMUOziu>s9NHzyYgIS+kiCWvE#6A4Z
z%hf}G2;Wj(lVb)1+vfDOz>RXrTjiIp=Z~j~cR1o>wu1$?tlPj7i;GKLx5-!{II;k6
zp%`cs_~F>-Z1-f}6;0ep-g0_dA0{dIhpeCWP1MWJrPQq8MOOyrk={$aejDr0XX|dY
z)h~UW)9&;9Kkshu8Ov^i{85*N^N`dZ0Fqw#Iz^=o$vGuh6VYl>YR@w#(?AbA&kSmG
z`b6n0`~rJIAz_ZTHP@q>nv%!9Bs2XRmk_6}?rO~@&PObZqN8v$i&@k-mavdzt$HrK
zz{Y6aX1rs|^ffkxFR+T=`{Ax^ib8ElwRu)ouUYwLZv6dGMP0o@^I(JzE3$Z^)4^u(
zadLZDy<CXGqKW{3Y)90Ljxa>X3H5^vu`YT0>+PDgdqHjAPCT@1@Mdah{rdwot?>rQ
z1T<3fN6KoBAV-lQ1|Y;@sy8^>7oInQv9evEOcH?BL9W3RJZ<PrrV@u~?eO;M!<R2T
z-WhQ<o#2zP7yw<d>dJr}0oF-mC!(^KG*(Am{Y^oH7f<YGhxxHhbPoPbuKY4}weO<p
z<?dQly()rvXA&pu=@8dUL_pk+_hMxlL=K;qhzSScV5GrL-%WqtIlLg|`2j?5Wpxg7
z!cjA<$%jeR=*xuj(;uG3?|u8PwLA;;XBEH8Rz2tl0XV2h%>%y6G!e>+7N}!;*nHRn
znbVrYoY#jSU-q4Y`u=H%Qjm4Du1y})tcMghhH@&Hbdt84439A%s1TvaAEWQ2p$lE~
z3a9v{Ci)Uieecdq9*7r|+F`N&eZ`{}zZOL~;pU%m8`6g0of2^Kk-3a`HYMqIiLtR|
z(Wq_JZ--+1VX)UVYDkLKjyAb}mE^}8RXr^@r~-Mde;w@xs$nnaIcglud^Ev4Koe`-
z&5Uk5|6vI<&&zh9CB)E)h4-Q>^Iv?;b?OtJB_pbLBW)V{e9A@pZ@fQBg@Pm*kcV3i
zhBn&AaJ#pYsU9ly^P4*SpDZC87g3+y@Ps=V1BK3H;@ncSZ%=~qt-H|R*%6U9t6AbF
zzR$kbZi`^yy(%$+EY}qw-*3V2v`KJaMT4_`?Y^>wXdIPK4@VzHHPZruJb>rZh55mB
z0sA1TOMm)TAOYRaF4^dv!{7%O-t-~6%<U$XIj46Xg7^<X6>Hs>B^0n00gub|S8p8y
z_d`(y7yhjqyrgX4)86#l#<=I*%&xJ0zmTm%(9_#5Q_V~8-w{f?7hm6E_7P_efy8}Z
zM27%A<f?0cb5Stzb{*bV`mNq}j!2RIQsurpaV4O<Eq4C}H1N&*f?&zK$+x0!{aFYn
ziL0=S^jd_gN|be?lR1A4Ca$`9>Db+lgwFo^4YSPnT_v30ZVepD7}uAeuwn7MfC4p~
zY+BMt5ejjZc~E&LEVxix|HUN{ox5tEf7sY#Lae7CC*kry5YY%bEzk)+-E2n^u&NVi
zql@quCUlV-E?&S^ur?{0YMPbW_nXAm&nDK!%9r1N{I9z<-yI2%61a}OFg>-#lgB1Z
zxt=I-ZP?&@ond>*WfR%gY_IQ~bG-AT{}H&nAo)x}L2??O!Sm-@Ant+;+H7hEggk|)
zRl>xq-F^f+1KgLnUO0v5QMzv_E1<gnHO-L4FZOcR$D3|u+yx`n_vFr=sK307XQe@g
zX?o-yY&K6*P2GrMb(u2G?2@@Kbm}I5(QCwXua_-qZTEC);6C!LdQU(53d?*;;<MV+
z@%LLvPCtUIzYjW^U$~o6W%RCCel0*V60Hu!5R@yJY1hf^bq&vKtZ-`TTfE0E)enbg
z>roHRqY7ciDRylLfp6$xLh~!!y_PAmf=4!FH~*z7{Z?0@4};_GjMUqQ$4>r`+TAmC
zf^Jdolda=Wh9!kqNZ8J{r>fHJYY3@Q>eH}={sxln4vIrWNc6+2eaQqi=uB57?h-|2
z?sOWtv2~XrJtjNp{sJRw0uPvn_^BVM9%N4nl(p#^(H@v2OU@YAL@UANp=P5Tzk--H
zJ#!KaZ)>f*RyF5;vk>7WDGtQt!bibkiiMQZ`JVZ}?duQH?~tjvR|+yBdi<Y!E(_&;
zAkcD`RiyXzXSpJHKa$7SQ^0)}A*AaiM0;S{a0qHI|JABcY|FP9!IpdBwGK%(+rD&J
z^ePC%6ry9DNpk^+`7nv0yn}1{#i*Ht26%$UZydK7la$MgrFdd?h(HI(17;2s6OP0}
zwn(}SZav&$^ShnVZ=9-ActQ{o9#$e{9Shthq{l#96X3W`qyPzTMjlaGbY<|$Mojie
zdWD!lb>r*u%1^DrE&&o{(7N^t+I8UjuP@LDv<Kw+s(aBnGuEkh4_{whVoR~92OA8q
z%Eju6<N-Zce^Ks=VuOH!UhoHxPghTZ+S0nFWh#VaYG8leN`Or2)lsMKquKx1=U+LG
zd%#%_5@aj144J>qZ!ZUyb3dS)TGmKh!9D!`qXJzQyWTga5$k&H?k{P+&;rSvSWw)}
zu{B+zLfe%<`zg)+;34)i^>48<s4p_K*oLc8p{hJD_JVB4bupM{(@&S$CHW$W?GVm&
zS=E&*Oi#Dm#`~P3vuHW~+rsd-T?247%2*6V4EteX<}ECJOdzSLNzG$gBCwQu_7Eg3
z&%D&nJJEk22fKSTkBM#P#R!NP7xt2g8-?k2>h)wTC4?{%h6`Qnj%K#@tq6Cn-s)do
z|Ge+MEx>AhPgR(2gx4qIHzhG9A4qDODEfpdP<6#!RJCXSbAhEx=?8u2O61yAy;34R
z3S9jX$>QmYh_ZUZkhPthKGCG<Q}mlXHDK=aOY!*Rgxv4`OldM8um|oV&6u9w#6fN?
zs2%ZjMvtbyEp*H8|DiN~i|qtrM!Jbi(*t@ar7A=f-Y)VX2wR!>#`MH`eX^eHJAwGb
z`lpaTf`9Q`3??X;U|WzO8OJwr%BREuzQ2&<(?p-zm$D#T1#8kCBVr)ck(nMGfFYO-
z`s$pfMwFG`fj#z%l=x#o7ENDfWRp9V!Aav7(U)H73^qLfO-;<^WWU0B>$JD6377FQ
zEYy99V=m&~8pB1zjZb<c2nxxPEZ(MgBW$UzPoCW4%St?%R>-cEzkr%tB~)V8kU@u_
z6ANFeQ`<EC8b%B5N}anT@ZMjf|3b>o`E0Y{tVQ$^U_)G0iA8WVsSvre2iIzPHSl-i
z&-`>$z1U%%d(Y%u01L1C@117R{`Nq1M(UTAI@@4+(}%i?km(e;d*#J{LXC)yI*b-P
zpE4RmyFuYuYbClD8!N7#9ia&zyk9p&ghv#(KG>c*;orhg(RG;@>6b9+HT9@?=r*lY
zwJ{-|*OXuVU(QOjFTL&%<OX1#Z-fsWf}ZgOQv}*g>7NLn&z;>mm$C5EQS<Tqx1hIw
zLpub3FU(8G@QF~E9UTMLr5RMWH)w=;coM|4yDqF}>e7pI+dusF*13glqFXrV94$E>
zFbIBfnITaxZ67#WoBwT!b;`N<J+oXd&F<>M)^%RVoi)pCAQ0Z25jRCy>B;uU`H!RG
zbn7p`ZN|+>m0fd2xO+O7roQ6wPkt!sNRosH8vLWd92_b6#J*JWhY2zH++fcRf9xw{
z?u2t;T1;Ej+muq|%i0s5St1g^99(_mW;+Ln5l9~?a!Ic$BV@4Gi}#y~*ViRpo3UA~
zaVNDRLQWsamGUBrN;HZm1{Q!wZya4`_u_k1!X3yx`JG+-wZ1K=1$aURTjH8bki2k(
zI~kbRk5QFFhFnD->$^jmLa}vsd9<JJMPdHCtyu7<J=B)!{^{BM(_`LPr4@J@ZqTOh
zWKgnv$ty1;ImR*YNFfD-$@vUEg^h=pjeLgS_@K?f3BI+w|LfOA-$R1rvs`G>)-x3d
zyb(?QXkVKgjHw(fAA%zM-dL<jNp{PL94*1L-EWS&gGY1{Lf}X5#4+|E1pCjYf=;ki
z#&jau&IYdv(YFz8Nx%(M^V35RItj3LxhdJk^q+rSvV92Rbg*+YQ20odiPSn8b)oP3
z`wt#}ZoL;fa<qh|{``N&pXx4ALxAyoMn9gz7a_a~Za&>DU0uZ$*;xN_&r4+_)OhV{
zSB(#(!f_4xRl*dm19LGaA!8@z%KDIQZ;E@!^wRU2bBUMRe$~_y&ws`Q1$gRHBtwza
zc-krEHV3Arj{Y#Xs2j1pp(fu)e7F2ARxB}Lm(<5&xn<<zm{&7yqG7G?FJ?1R^0W46
z%^Y#;e>HsXl5`GKfa}x(00D`j8NICC&pNs3Cc1W>ky<pxxA`-i*T$Q5`4^VyPsEeW
zL(p0w#l7Q0_F0NtD>$|fL3jp0==R%DkO``Ifd5H+=hbIDzK>VTd(oJy?gR*M9uXl6
zcm<G`sIdcOKgV>1zG9t9_aNsRhUK0X3Ljh+?xMpqbG~E+q$PgSb1`(=UI}J*aG_+*
z>ER+h%5))w9Zd(NF`hIxPT0o>WM}=4OCrS3!1HAr;;`F8FrnG}IqkD{r2+X?RS9CO
z3t7&`4*2?lI5YP8TQfGIY1ikJuC3mP@O1>g$fwIac!$V$9gsYBUR#W;tvd_K5ooAC
zBev2X#AjnTGPi@k;s5NY@IaI9&@<Ows+pb}q*)S42H57$XouS|f5b&j{5o;DZATCK
z>&c2E7s7>PWJ@^(*9B1C8Hb?QNgnqie$jqYiE9xD->{55HsmMti!MNk+<zUoU()e)
zMe~>Mw>vf)HO<$q7NoGhybMnBS@8}+nqfQeHT_t}lig$FRh~J8HIkh@zs&1L)RQzr
z1Z#PQhuRwS$B%B+*cTHXf2c1Mmdt@i2w}#V5M4p?=o=29*%uCS?{iz1+b(}+7<7Qi
zeGYn*ze*ILxx`#Ace(vu1?(yndSZJz<ivg#GH`^TXVi%lhp~}?26wdl_nIGf)@u^Y
zue=T>Ci(ERTYoe@c=$s9qF)#RgKhTdP&Bs0G?-V}p5K}N*;RE)eG<4#eFk{UK2CBb
z#M$+vEs<<L9z6K8BewY!F@3DrfbZ$e8+W7gPky}!+CIg!xf_G#f;|CtT?Ffr)E#-C
zj)wdJam9aZ%0?*P*M*ndx<qUOmZY2T?Z~PoBooY3C|cJbN3(;6xkL0@W$7=8-Vkv6
zVAMo6Yv6);aN4f@nSX^Sd|{3OjLM<lcKuVaTP<$I*9$e6Tb~?jdx}{98FH*_XEiH6
zEQ7SVbfi8ReVPs2jdJm5emp9Ic1z!FyCR|jBVlrCIeb5Oinxx<s9~I944sB6j83*?
zt5kThK^Iy1eihp<90|jR4|n?S7R-lgd=MXCv!|pJB-1IWt;4RnhoI$MFQi5%T}ajT
zHyfv(!Lqe((lsV=$QqsEY1xcv0(^-TRC9^tu*d#?zqGLN=Q7T9vhT2H?pCb(&O@r%
zS_9D1Zq&d-P}DibZ^}$0jIO$`@#k4<f6CGTwf|bMNF~$i=<!1k4LXIuX)r7o@l1@%
zSc(gF_-62lVO^U-&GDLGrB>_MyME~>15V!3MX<9-e$r1x96hooc&8VW_Q3yIkfh!b
zJm4gSf8{353B^9JdFp#%nRuavY#`AA3uK5>j(%U@pi8tkUSH_%eV}U=U8*@zv%%6n
z{$m)N|Hw{UpvI^OZV@GHnGEm9`(9?Kp6@1l|I6RTeDhfq5`>JWrVa?)Cu>HNHWnY?
zeG2ciDXg0P`EbmiDO;1XWUaH=3M~j!9Pq#{1G(SoqLsd_(}d=F7cFtHOOWw%TM%?N
zYg@1#Q=)nAqbV>O;suIH+78g6rg}W@c*qBid-HMKc`|vkdPs2{|0rx5bsBo1B2Xba
z+>1A886IV=M?^(ChyYH6$%LSehMlOWH&A{0V*cyzUD@62kjs+Ys^*W~37JVOjB&}c
zh^iMD@U{eZYq~o+CZ}y^cV6J+MTn80P}Cvlu?okEEzRK!P%?+-r#;2xcm9GI2cM}k
z^NEP_@~CI9dj9}<lq7OjXMe@|{w*&u?*LB-bV2GCbNy#|!g(E`_h*N-YjHIHAf{M*
zGv%b?l^0pgiyVhjSlBj4(mV_Fg<-1X-PS`8i(9q;x%pkikfOkg^E%mva!N9F6ditJ
zecI#4)&@mxezxb3o==Kx4W2l?3f1!Pc=~WI#^|QyT@dqt<yWfw6QY5?%W;$P{LPc{
zc@4i!|I@gu+%){};jiD!CuHV!Hp0HG!Q==xG|{>g_Dw~pxZ`C5WEamnKB3T`vgD#B
zP6Wj|_wN%l&!?X>XNqzmrN0YH;0a&1#ow-zc0BchyeS>e$4D$L@)_}e$$hLjP$e4j
z92KQ6OP>A^=SygGB~*skP1@VRR~FTo!xmK(efoRf8OtV`J_5kx$e!7aFzI2ZC%&bR
zlD$e+*gn)}?P%+gM5K61R!m<Xf&>{lFEY}jCkY#E%R&z+;h%0dU*(*1bEn<u4dvkZ
zWyYqoiOcfMwwNn(>^uZj6(_vGpMdgoRzlb9roO?2M{+bTkH<TEcf&2>YncoZO^6wu
zElDC!^EOF+^r*g)Mrkq3tl?cUVot$PdFHJ4-7_-b-IV%}yP%U-`{3b#0vUW8Fm}<K
zA4b>DHPX<zWz+w#rC%H=CgtY33rTX{CF)|EL~@}-Qw>o@s$1!V_?*s{Fp^wD_SfJ}
zE|A+Fyxj}GWP6*boBr+>))S%PSG0CGwWNztpH&dfBzFRRDkHrOBH{pjWvGL{Ig&N?
zusc|rB~hoce(6g=HNA*-lf1a71ex0iUEfyfdb-+h@6%d=e!(%X)y*nN_;L<UzC219
z2XX6<v_1V+vYeq*(G+R<+lp$w^hkUA63-{|1A<YB;8NFIR|hJd*aHL+EC^hzj*W9*
z-Vj;mwZX1X$XTy0i>s1bbC}c14qQXe&=l5Ux57UlgOtsyxjznaBp)7draB7L#lNaf
ziU}-^#<=u?t(%8L<$`T{*M$FiYq2b<|C9-Phf#1jW5{m1i{N1Vx+1d@R}MX}#sCqh
zEP-={G8}I<B9i2!-eCDLYq7fuT>3bZs>q@42{D~p)-PY%b@Ho8#^#6Uzh~MsYC{}@
zkj#sbywHug<1_W*{7*eO+`85(Y&FaVrQ>~>Bk!_z;bld4Hlo2*n&B`b@}eX6@)b*D
z6iTUoLhWGkA@!GFTT<F*#uoEHigB0T)o-40j>pt>RQNR8RU!q75{O#fyH+yCCLcI_
zD{t1U!?4Vf@#?<d<8WO!<Y|xM@AYr1Awq6@rbgEr_M8*mVpW-Yl%cxL=4=&GBFHLC
z$0S~O+!1Qt;rg;EfIysgob16qlA1L`*Ar#EQg{pn*J=phYI7{t0|y(JDOu|xFnra1
z6DvCZ;xtw#-x(Ybn)-I6+`4=!P_qpYvv1vjKW+n!GI^Y5QMou$>oD<{b$Rk{R@fA8
z&;ZMS*lg{i+8uYOkaf(KU>9zP00{%HmdxWAGAAj>2~Vo$cX|#@#f2L5S67dy$M0F|
zs(+@oyVpgl6;w%^WIt)fZD9GNd-gzGa9F;iSmfo?%(w<!wpk<6PS9-dt4RWaq~NN{
zM|=p19PYRmqE_b2cI>{qn3GehQk#qRX8$ioPq+Pq@-nC6q=es^wf-k$Bj;ueWWCRb
zaDLO0N}NLnQ8utioG^=lvxB-exi;?mkFxfJD%k$~>73WIXcoC8GlurArAPT<A~h@R
zVS+&KTz}-i<pAWL&0C(|G+iJopWcH@qx6E^+e_bO+5cThx+eS)6L)?4c`OZx!eVNh
z<N57wQhM5~ONMkZn@<Y9=VTi6@al5#$=>^_iLVmsQi*k&r`@L)s2`MF>XaLzoqw6}
z(uaBEvQf>$AhjEC%f9bvnSiPN3-^)@WHpksItvb|ghW+;@QeoYyCAB@0i&y)R}i|}
zHsj0zkBAGcKFQ%S$#ag(a29A!rz1a@7}a$<x`roOlfBA#;OXKpmh4j_W8d!Cs{4%8
z!~c8j);2N}Q^kW#3OU^ly^j4p`GPz1+-P|cGC+Qvv#dMpZJKFmm&uOpHW^Si-7>fG
zwk%&PlL(EKN8Oq~`JWzL-@v4B%)EW&A?VFf-8O9cjUeMnaQ~&{1kJ6-6sx(ES1m@$
z^D4jB#P_JTuL;UnjHEj7{XPA8OkJ`4!k2$Vv}*$!M3rcLWoYbDi{ptu63E6;n9U(b
znHn<?o!Ka9NchD7Z^W0Gx6fRlm#OO{2(?u{0$d2;{A$+w#Ph~Gim#pG5pVVb#W8o&
zw!v?k*m7xFWD+`taByNx@yvXLNPt*lqwjOAm0-{^D2J0L9CCb{C#*rn_HS}#nF(<T
zJzCECN*P%qOMSF$(XuO>iXZ%62kHNx&;QwNs&@!#)$!o2qO4NCP;PdY+GFAbj1oRB
zr0BT(c*L#6a%zLdJS?}~4L}&|djUrRuZwi);2v;+^xOZQI~fR_kvr&@z}{KMzHQ+*
zzyn?N5U6#oL^5SZrC(w9$<0%iBL6xSh9;5f2oGs42eL4m;b#vXqa_=fEo36@QvXuJ
zWwTjDrM!c$B-W5dXC3QMak*scHGj%eh8xRM;*E4u2d@te(to(4N-_^Yqhlcuh}vf#
z2g;Ft;v>-S^aRnv3MSmV?8Z}W3+LT%VM%K0fK$XdUwb(k1DD&t0{5@SnGURNCxP5J
znMK%zguIu4Pw1BwhoF$zHyt{&jxAu$6(o}`(;<jm`Q59UUvG4b*VhZuhGu(YP1Ypd
zhqj;Y(*8SBm$WwSUf|qxX>K|QTW)r(%L{$!5X2;AOf&WF_6n4lDvl}@>f6D%xa`4;
ziX$M;mE_tbrK%93s|}dAh37gcbA-I;c!3ZmMYh3+;PV7<WVRq#{0R&~phJvbdZ<pe
z<XM49kQvYRjJ?cj<o?cnt`+LiO^RW{+_mt%*DJM*`<Gf*9=H^j;$C9J;VRHOOmnEP
zEqu-aPG6#aD{90L{iJI&!rk1XPu$kz#bE7c^E`PJ^Uxv4!#rweik@zpWJ9hQI2c^1
z<vQX=wa)#E#{;dS^i=_&jz-qhxoPQjql52p)yC%Zs;6SxWo81E{}U@!E-+Wuir|Dj
z89H9n=o0ti?SXQkXGDE1WUbC?rC4g;g|)Fko_SFR9G!iI{%M^sgg#TiP$XY^B3K;j
z!P_*3)w0a;)oVyWH1x+?4?0LvB-7`5CGnPnsGO1N)re`qj>br<oqfK6u%CJ_rb!94
zg>LF)(@oY^Rp*T#e`d}vSrIX30bh5#0z}Qa#Oiao3+c-PI&Vx15*8hy`lKGaDegsS
zbGCtJe0^eIC6FsE5ODqm4Ql75JtVUP>6#azXm{2GbbNL0vd%sTUOail_lyBpwLDmb
zfNP%y&*4L4D6C0x>kBa{LC+sQt9`H1D|V3NFdM3+9z{k_-wx<2vMXvB=R%AA@cHN9
zX{nvq4UeB<EnT|Dbly5Vp+K-#kZS~@9V$9opeBXbI#IGoGs<gW{d1dNe^7W?Tbn8y
zPeP=X4Sbricu;Nsyr=z&WqDz7U!8`vylm`)<(5N`%%8;0avZKW*n$vo%DTt#)FH?f
zFHAEA(jcw<<C~pao}J0Z)RFvCkN@tAy14UCMf%ERg);43Yh19tkmh8fJEY9(<7d=B
zGXyZRsBqgCcxYGqgn!iR>pSiN7x`}|q8~XaFO3_ZkSNG(0fuWsSuU(CTbxpyXnQ$H
z^TuFI+?UnGE8kb?c$1w&Q0O6uY>JA4b2ORr(~MFO@%s9iOAVtmvpZA*U8c^v`+=XT
zr%dfk*ecg$lgdHGr<KFsNNb97Vor56F-{3=IhG)vwx$5$BL2rWe<LnI%H^#FPe2|O
zSuNxVn4m?DxJCg-0`5v=^l3E~fe+`d8kx1=%3yNzuM9z&`v)D*lR#VN>OHT{PreNk
z%h`md?-ca)tYv1<ttb$HKR8Jf1t=^-tNOe(L+WE^iMZ|Kr!@zEjrZgny9fE3_oGnU
z_Xd5zt4U3=vssWrEuC|W#eWA!E2^PluO#%hBSUwqINZTvMgr>85AC#yX=P(6pOFeF
zXW@5FVB6*Wj-+U{Qd9Fgjgf~Sv!)oAx`x`<lpZn8QU&f4QadQy1gMpD4EBn5(OS=Y
z3(JiS(stwd4dun#K@WL<DOQSKkQb4~8GO3m4iQl64{xySimtid{5atQhx1Fl_KsV(
za4Uh?EdmqB16|nVk)g$dIYc5RfT-lF+n#7q?KW1xRq++sVDdJbaFy-z_-0aMMJpw9
z?h16P<G3P>Jx2s=TG#wI`PNTWmsH)yf-DCS;w0@=bx(dElUg&nuBLMGmk$9SLEQ`p
zjQsKQp2uwor)a(w>b5pKm#6RPQvdx=9!~$VK4OTTaqu$$VNM6hM3c+>`8|4VO!pCF
z^lH@Cww6t^6D;0Sq!zJN9QqVYntX&1cy8si+(`z7jI$mJacMwo=mc_RsTgEmN%7Ma
zm<<*p;G>k@=pE}&R6Fo$Q&C(CY^+%LxO=YJH^I}KOhqa&N`gI$ZEjpX+c8oq{z7j&
zTCn|%kef(o$#W@?sxiqCa-^(kJ{MJ#+|GktCNJg^r=L9&$w|noiu*FvXL#%9wxwQ1
z6>iq0Jt>}34#+2-9;*JNtuIRcd*UNwEybH#kJg8Q-et2V9)c!{TlF~`uQ*4v&nV*l
zyid9_Fx>x?M*Buw5Kp{%Hfq;BzF`oq0Sl^2k+2}ag#eQUljb~?&j>XaUCnEH+;@V}
z_Z|9=)Mhpx@WfS$JQL+!TyKx^7J_Z^PzzpG?7K*YHtd7;@>Z#W<gyQ+o#mo@=}v|e
zu1|Q$FMX9!@Osx4htIXiU5`fc=OCTtp|f%cG)-`lE(AP-Q{#&ZceIXY%Hs631d}rM
zO=3Y?_@jj;N#mnfY2mJo5j1~m50_(ny!GJMchRNE8x9c;juaW{E3#bsxa7)#iKt(>
z8R6^SXLeJs7+bs=^}CNv12$+fcyWF9+2x5W$f+iEA;^;vp7bw<*|GBrjZ9k*Gv2#w
zLzYJ~BZ_9aPF(#6GbNN(5inpkAjs6}*xfr1#Eq1%gbn+#GlwO2Sdmpoj<fhvj)-Tg
z&xx%LXAC#zhkZvbk`M-GT4-W!`e|#+s}~gV*$hRj?!PnolEWhes8gtpd@GxPLf2`h
ze{S*V)~gPCY2AY8xn+Lm-Q{Y5<nDPFPaz^*b%Xi@_=x9yqZ4qBEZKaaV~$IespDG#
zttw|t)mU;NJz+J;8t5fozX@~{7}lMwo(}9^8s5o#5<|-JZi?sswDc;tl`Uh-k(FlM
zG`z#k5TLnNM|hUcX^dqwVJLq~uX*SDctz=D#%FVn9kfX15+&{p+f@(laBxJgua)Os
zzOHMeAgwWULO<O4l7n%fe*S?zun&m}B>hg86u|f+u^I78>#v?Pu<BjT;`Eahe*)|@
z?CXlB=A|qLSPol?os=(^A%Z*2IbrN%b$l$%CeTA|`9z`?(SG<*w@<BnI`DD!0K_f}
zy0zwU(?093##-W#&Ak&d#u82mOwawaPO{xIh0!Sj)a@fhytwo29OdcienxG71r&a|
z*<-v@w}U?8)2X>M@#7H0%;<**VqO6xngzw>0S0g*e;^9DWwQ@~dzwaaWfac-4u4(X
zA>*B5=vNTH4Ic(f542#afXEXr)LWWh!LpKQgl4i`RAXv9Y?tQ~n8DK76#IkjVJb0z
z_07~*bOhv|X9fvOyJ6V$<sJY2{7Uj`5G8@Ob#GzWD*Dw62^)a?QY*>)o;~DIrUO}z
zA{!uib|{-)TPBt7o&4<PEx~=khx)2yo=CX#U!Fo}DR$o_dK=g0p>C(O?zA8A`l>Fs
zQK8Q_-v>JtrP!8D6)pg2L`h6$EzMb}wH^DU_1pS`5Qnb$wzI70Fbau2cT{9_cfgT>
zY5fV9=l?jO#o902!AbHp{cw8|d>_2o&u#vpE9;eFPI8jU1_!bVLDFICv{8p}0O_~V
zTM+`=;OSLKk!Gs|oG`2Fb??EfL~Vs1&)!*P4&}xrc=|5)nY@3bZks>tnispS-PUaM
zzkPcS_O_C004|3nMwW=^E9edQ`lwVXG4+*?-tf8>$fw39#7^9$zD55z1aKc&#lk!X
zNJr>K=OLh?l-2ur9nYr4>vC>xqPh8mN-wpxfbWN}qmHBz2eg=4IjXdSKV8@r$q+G+
z%Pd}&UpO~nMD95+Ji9af_g(Yt&z?*)K^)@(kisKLv$v2f*>Z!Q`mca5)$=RU?DzYw
zP9qyIh9D4#lWC$!X6V9STg&W-qDi-X-)~k{8b4(sXIM@oe^~}eT$uh}gA>zD7`>b^
zI)sN-S^RU9i8;1{;8RxXx`el%_^m{G`uJAMyL^{-MpDdS@rmw5fZoAPH^ix$BYMs4
z=q5BHK#9z5Tz_MF=Lwba!LVhD)#xsd)NG=AO7?{W8~HYC^yKO?6(RF^?tF>2qupdT
z9aTe1vde4gA{>G`QGglxJV0#4{J}pt1cf^j&1>nUzYomT4neG~Rl<y|d$2nI?htfx
zbxtLOra(`i=N>U{)1FB>isWED?pMjV_GglD2)ctZ`GG6eYV!&-EohqbqaCDO7>IW{
zw;flt=@Re-nrS>#<DqlAX~d&!N9pD#4<*T~XS2S3@F*P#VUpbzEOtCabEXu>ksvR5
z*rx5E+_!s38XVDTJ!eWfi!GyGV*9wQ`CdOHy+o$ZG{?=QlWy{~lI&Wka9Q%m(TsWg
z(VHF!UB0c9^)+Yj<l9RMLT5BzS@Z>3pH&RjzIMN4z1&4GfA0<62tFaVl!+1^DG_na
z_3&Nw9Q`}(Ua|h(y6--GtN|18O6>q8K~o-U0uwU-vUs`kKjKJItef6)I#HyrEtN7Q
z&K$dr)}Z;s5m=)LW%szw`vU6C*A#@i?e{KxU3qv*%xPU%LaQ$jXVO08<&ExD`aJ9*
zn=$P`-0xK|u^v=tmYk;Ha#wu078cN--PA`%D3f;ABIYw(PDJOYJj@137FE*R=2U<i
zy5MqmjLU%|)%iUR>4LBNy*W@j{YR_sQkHn2b>AWAG+fmk%!hfFnZ*A$o5!UVr}XL!
z`sT<H5q;kNt!FXiu1J8zW(be|Xtsi&xZZ_S+wM!vHCuJvtJZ@rwl#d>ZR1uEZ8+NC
zc6nUcLm<hdc4WKf%f?IP`@bXS1l}hH@dhOY`jvjQbIq_{%nwpYb#ddEGCemm$@id2
z9Aw<W7IN7^;Xn$S-hQR4EkU%)l3co+VIPS+=C+^dP=K|#->4`95~jRumD>oiE`nJe
zf>NH$9fA&6?2m+fdYOXNJhr>~%DK_u<%Tj5VH2cH__^(7)E0^1+|L5wZvb(GQhNjz
zhnWz)Iq1dJ^`7!-qQh>%*Gkw~CV8UHeF-ssy+hP96!WZ1%zjMyW>Mzj^wj=KB5iS>
z0=eh6LB6slOL$23M6MuH4|go&xG&uW`%Ea3VOPPKD#bV(ZMYD1P%v^^$Rs04XS873
zbb^syxpT2F=3XKD$c(+<AxOz)`?DZ%MwFhf2=gVo*irD2Z`?N4tE1;NubllobK@%|
zq%X4#L;l?Iw%?4~>kmvCNOhz55Ehf>O11QVrVpl`ElK@-@zyuXVL#T_220E`Y=VQL
z-?Qqa*VmB0)_RGyS$hEvyLCN?#)?0K4mPvk#CO0}mxIpGnW3R!19JY}W%tTyT9wJ=
zrhp*O-A1rd6Pc+?_cfMUJxaUp8>ETFM@--o-iR5C%8j-kf}9hr9MJbZ{=t-My)wMo
z`<U`Rz)9IN#@==2Gs(Shv3+$ov24eEv92yC0BmUC5JdJNK98axZ#DiHep)|Vqo?mE
zb9cyd*tw%1i;VptD-M2of`!jP4e^S7qj049sC1%B<|N>qjGzEDl7@e#N7D_SY8^$E
z@<rL>6JFH|4bMvAa&M<+IPy2tOb^w0TGG$SMHl6Kbu=pOuh`KOOe5GoMeqVHCZozP
zWrxaDJe+<6iDarUIdfGcvdL)aTG*G_jAt5}c-^zRY!xeCVDW$Kw}pM8q&j+zds2S_
z2uk^*)>({z-65z=ZtfFqz4gbU$kaYpMcvSN6dYmblzcAXSwh-ih_kaci)6UyUfmoD
ztHeVT3Q<R%fJWl`hjz5?ys{RX^Xk48zU};27;NDnaiG8$7Oh+RHKzj!0ZVLH4QWAq
z2|S`^e&hIwXW@}H4+$brOS$3`f7+^chSElpkX?6QBdhHozv;*G-to5Z8nC^)sq`w~
z$-e@*nGvl9G!yt0m^I*apy}$ZyXqedj@GrMl8)G8E#I+wnlibMZ9=s6nvynRsmO^f
z>hA%>^Jz&X0h4HG@!W^+ELXr0Yh9leOlU@mD+Q_@jl#@}K5d5}0hJU+HyGrLj2xTT
zP1Hp|)ste;7c)s2qkkVCf?V!-*ZwpTcDP)(@Dt>0Vd|poonOUaFosN-+6D4%$=o8|
zOjKjCJoQ*!0(5F$zMOf*)bro-hy&9CJi`w`^G_;|ru|=sCnk;ZG`apoNV`LX(*2vm
zJ3R;)i;CJ2sY6h<_1$l!dl#jfCfgB}{cU&w*2{olU*Gkg!CKAb!)q1qFdsVtYXs{T
zz`EdZHU{6K(kwEXah{ASfVghuN=}~YN?tQAIKmt|YAhMbW=?U5YI61fPlr9I-E@Ak
zZfOmP|Afku&*91D>FSSl;6I~lNA_d9vDlgh;6%~|J*qs6r!?M@1f~uHUO`%Qqt8ZR
z>b_ZOwuNEGit#=8x9eVjb<-HUf<RYk*I*cy7LBjGM)WKE<hP`K<hv>i7X>P2`-MG3
zHDlF2Uo-s6Y3;VikT)ajS=%@m&6$rI5j<n+s?maV-_sh5i+)lk;j!T=44p^mG*vx(
zWXsXO<>nBvx*A`-znf}KoY{YQU$_gW#<XoJxYaU%aqVaQ;BL+fb?-Xe)7F7t8()-R
z9OH$*a)S`^^kskt$0>-`-Urf$pl^)>iwF03g1-tg?F52I@Qd*N44#hXarcd72Xw97
zBdc*Q^=*>V*Ed?lT3WiK65D-{8kNBC95_-%E)=cq_Del&%G2F#F8b+~k_vB$@JFYT
zWizIw#YfIb=6Uc5FqiuIUUaVi7r<{>qKv!r>B@tB0ZB51U!ksvVvgO2gNro{;j9N}
z#x6d5T+fxAl4H8S=lm6ze`>La^O?Rn2P-Lh!A6BqbS>yPGc<MFvn{1p$?<Vc3Y^5(
zeGig;b0!+}7eTG(mZFZJi*)<ykTM*n^j`YDcX+@(+T_DC45<Q?9yj#oBH9$Ff6ZXW
zsiRn213N|-BF6d2YYXqL6#U?P0KO@gyrTU}<eO)>gNi;I5S9NT{y-_bO$=#%o?#`o
z%)Rk7cE>69&C=Q!?x0yWQS+vm?zDLTy;5IgmBC5wjGA?1%I~Mhb)_T5MOE8c^cu3*
zUHZu+PLL--sSz*@7CTU6{AP&LKQK;@LmhA4Y#cp%EZE8RP5VUJAXqIvFoNY>Txw0Q
z@pH%}MxUrTO&J5od$w<gIiQe*G_~Q!WNON{i!u#akPPBTynP8Z*=N8SKiU)fb>%^D
za*eRJK|sVQi$~81N1LA&5=WEChaO->kx{VgMXg8B_@2wd-`Tq{|AWQ~Pw^L-cUPrC
zkHdpt0(GN^XFL-M_vqO*tG_my1j}SIFmKfr@;~sU^eW|RnqqnWZk>4yf#ty8;X06H
zOEBvcs)fWc_aPqb7=spf|IkoV))xOg()iwaf$7^Hg%1pUI^AEWOPnu}x$^FQoKX#~
zA0njd;Yne3_(2}AO`;;gw*B~5_w%wP-tIj%-?$_d!s^|U^+daol0*%y8KxaCKyvpX
zs++Dq5u8_L-m;xEc^if_Jo8;oTQ3rp+B$)Vf<atde2JM=XPY>mSCeZp_EP)L&;EMN
z#*xLT@a|U2K@-@C&&K#bi*^V?tyW-7KBw#KNp_GATK)0WhGj*z)03tIWJK*FTeo+g
zQ}AE8HF*P{`O+%Au`7<eP~6e3I<BkZ?)ha!6cMF139X&9Imaw0ksssT3o!TF?qRW4
zqB@YV8ShWeNY+FV(22`G>*`|At39n?X(xp)Ga(T|=~xrjfgGTT)x=LUY)rsB=h$@S
zN7R?L-g;Vy{5%Aq>-bh39At=l2uGR9^fn)r0|Q`VuBWE1E$Mi7Mi75-d$yG11C9}A
zAuX-0Qt*S)K*(CojNHhRG5rg`*F#X3Xv$`-f{cMPpWd~vvqbuI*of#p*MBs3tSo3;
zuIAx(a}fL|BqsO@?si4e`sDwF#2kzz{5$nOb(c+rh>Dav4Wv6!9r>3(G~~UlIh*X)
zBYX&YdK0TMdlTA(tlCI|3tWCglcy9^x^>?^f;~+XQiFLa{N-bp<>EwJ^t-$e{lI@N
zLD?cCTvGzi6-Ri|xJop;>+&?RyYF2|bN<!knGW5NZU%Qs@{wIg$7g-qbg2zvpOkGU
z7@f|3b*aKUV)2Bi*YI$^@~(%2YKI`+zngzve9+iWc%RR2H+`eh5Ig;l&p`+U|Mj6k
zBmZ0#xu-oNDx}8eSEz`nopzEIXZSp4_QHEVTM#8FR@ASCpdha=PP_ZVoeoO<Cs#B6
z*oTXJZHmPx(Bn{VCfziwBATC`t3(6mOuI6)1B{~8^3GVj<;1))tv42I@V>Og&ootB
z-*E_vKN?F@00GI>n@f90j!3L<S5&?3{_St?$8+z)zBf>+=Urr+K&r=KL5iEERlk$x
zPvi>u2-Fx|)hn$P{4887gxaCYWIq_0tlyU-Km;eCRc*F#47XHGL!H2PKjFB8mTcsC
z76Fo@D!d=g3a;}fv6s+nD7s*G;O1;l=P*p6;8BjNuUJiM%EAb_Ti<Bj=OYfO)VhTb
zhj}7f!4>{V+l_ABFE(#Wra9GZylm2W#(_U`dN%Y5zrwaZ^%Zd2ovbqJ7}~_+!QVMy
zmU+tm>{UtU9<D76cqT22p<XV>yK?Q4z3(%N!k%;e#(A*{y;uIZ)%_S+i;lrOUp)lz
zxmh8_(BdWEfvVYujB)XIXCM<FEarPWrIXOtgf3k9u9&qh2Jb46x+oQ5FdJs7qmwhW
z=RfP;(R`Vh-qy?9nQ?cMFp~4TcA;`?t3*N3h>qjDJoR|Dd>w5@|ESgh$Q)R}0s)1S
zdsmd2Y4CG{O7FPdO-@}9GFU4;Tgk*|MOje#J3bpd8pF39f_U=nuiJTkSiQgM`Lwxj
zJt7JvC3(@D`R{q0SxQzN-CX*?L}sjp{yCasT$}YJ*DDfr)Qz}u#zQp#{)A8FOhK1H
zqE~B|9CHRq{}_zfUK&dCB`TnRTQCnlP1)2ZdEdoDD!4PFOsCmyGN%M2MLM!qZ$w!Q
znLYC`cWBr{sozd~mYgK39H@*Cn99Di?eg!3hCjSwU!A<aQRfNBnX*Z`2xn|t+zeoQ
z{Xb?dE4f6u`dYFTM*RbPToU=3Oj>O+?C=8SYX6E{FR=oZ2utmLbHDHFW|HC8;J%^E
zzmSsdwB-3mW2_rI&ggT0T(mgHSXloPOq6=bc?fDbW~I_8PpqU@!}%Y%0rn<yV-MG<
zzOGl_apB*lZc<Ar@P05a;}kT`vH`e0*L<PNVDy}(fUAR4%!IJ49`Tu#J~m<&<o{Uo
zuRcG#zoGUmu?H2cuTSBJmUdQQ=dTUiT$r>mG(6dP@p7l(>mSE|$fi2Hak4^<&dw6o
zrtBtWR%+6N7<Sng3h!z8M=XpTY}cu8tS(H=A5p=rqzNQ3Fq{1Yb5i>-#4X?yV-SB#
zW=b>Rnw0{c_10^|bvsFYI+Er9L_oX0Sq!s90}*6Q`-ERo$x5=KZ6flUVdu;vkm||O
zorz*G0~UU<EGrTEN3@hXM4(B4h!KFIvG!xtlM9)5lek!T?Q)Mx*WOH8)K)h3`1EsE
z?j&!i@RKiEkm+8H?V9(Q4SPCtf`g#=rBfqzOCIH89uUlJM2jNr{vjxJ5->dkJ?=r$
zkIiLL7Q1`y(n}d9vG6bCr-U}lb!pmBNh1>6ivT)Vr(;Hw(iH)7pVjlj8dA<hCLKEj
zxs_Z})d*mb+1k1&g*wv^Z=K2*oZ-X4>?nxY?1&O~g<@!i`}x!Mc4qgi6sjj?UOpyi
z&D3{^2U^`*Dj*24Av}C*>zd0BL2OSc=@#uet|RJyi!ZqD_D{slbT3g6ls5bYHHBiG
zBF0cC=d+*<j@Z4-x%gclcn|oPELyHtv@oX!(SNa$<p>v!;l8XQnb0aU>)ZAmb1I3|
zK7<p_Hp*uP%EpyFsaZ*v<#U6fvdn$MZ<B5Tqhv8tpn{CXw`o)*@EmC8rQgX`o2n&=
zzQ1cNl{R6rv#nPeSp9XBn)cp<rlX8XF*DtBpgO#@&7wRSv)Vl?5!8)OtJ~5<Yjh{W
z#3|Uu;}&@i-iXyWTtA%q5cD28a|rr#ZubymB3o9v@fITn3nu)csS`b;uNVyIhDNA;
zd}t2qhWJ(8dTL+#Lzz_nlPzql5y=eWKo1>)o__#b${g*Kwrj05QtZc1r^u@6RDPd2
zmGw%5g*$51r1{E^k8|3Oaq9Y(!IQgaQg7bUy&CNw)hQq7c$Gad$0~E$JIFD_A7zyV
zdvJ&2-Z@E?L(un|ct%X%EW(8+3c<$^cH>{xzc{uX?0Bsz;>^%}Gp@xpW{vBqqR$>D
zlH#q40NhPvdo~yDO{PnNvgevyO-!MzluN?<(^3NQ*{`_*g%e-=vDxgHu?X}X&zKb`
zHpx2ojrfdVMdSH|<~geH3=VqKn`!e-(Koi8u|)(ivkzzAJ^J}-esh48sPo-tQiNnT
zBSi7+LY;QC4Ip>DrOC}E@%(y)d}9A}(bz#=BjMx2ez3CEGqB0vAjbJ1ZetvdBOs!E
z#aw;mC5=-KLFUtn7ClQ%(JdFRVYx#8ANJk@9<Huy939c3gy=0yqC^lxuOTFe2%>j_
zh~7J6CVD3bq9;mp6ETS134-X+jR_KUMjcGK@_z68ywCgI@4H|5fA`+scmH?&W;w?>
z`|Q2f-g~XR);W8xbx$_ePr~dG+8%WV6M#@US>8u(<Q4W^Z;)-saprDLiMaOe+Ltmm
zcec<k+o^#yOstE0f^uH{iw){?&R<W*ch@&OCL@iN-<Ue`R96SDxCWV_7CW&K@3Sog
z)?P3jJoFcH9<6H?h$*%y7SX{@c!&Nn-|s14PoB#IO9Tk*b`QLdo)Lfa`qG>*@hf&N
zsO&;#KrOM^3WQD_1e>VD#J=4UTx^+Tb$<@$lM%1$Zp?4#4Xz~?_@v|>K?ShoU=GIu
zIz>nAXj9@hk@a#vug=&eARBZBH~3uqZ)zl*3)g#(CYtKsLzNaUhUENeUcoInDn9Bh
zl4W*xm~A7I%s6e$Zp@p!*io45%ESYxIxcO|>x*EdbI|(xsr_P{D6bFceLP?)rJ%eV
zch#YT!v?JhhoB!FTLB)zH$RyO(*6<xtyZUkMru<!xc9oYv(HP7zKLV~m0HH#T~l2}
zx0b5y8o9{FKkW+a_JI&^JRlUv0>0kkT?mpz^~ZX{jZ%IzG@|H*H%;ASZjr05o*$EI
zX^p2@NI1wEEWRrl9A>UDY82clNYBTfE{?<3HT9hK)5Uk&mhy0}PNrxUWJ`ZO8~S(+
zN{_CzLgr2Hl|JWep9Cp--%8;j;Y=DA6&kPC=H3+E*N0edU!{e_`M)Qfw?omtu0y0&
z7>Ay<fPH^L2yj=_ku}?<D3Z3S##oSJP5T8_GtDV9_slc->9M>$#Vf#-v=bln{VEJ+
zbU$B}`FHa?d*ifcjTe+j7q)Jpr=0NiERJKJJ5TbSNF~0pmD>)My3=B{k5Lvc3D8K#
zxNcWA-c4Uw4ZcRNW7)LzZwKNwx9Up{U#3Mb!DU+I5Z5>bjkF=zlI|I1fg-&r((BAY
zNo`b0H&5Q^d=xLP8cPGo!QJO=Paa!+0!hxg;6y<3;aIgOd2QazGLv5F#VDI0X`RqR
z*Q2-=Hg`0?U1Q`)7fN8&zgL^YGK^uQI@QuF{Iw{zH`V%PMfvVbduU;ULH6PyzunP~
zeJ0zk8a!aN8c4Ml8{X4Fh2suLekSlHY(XVaHH7Rip1MoZJ9~DQ_`(qP&0!uA)=FrN
z6nl{Cnr4BzyW%jF9(?9W&ql_cphjAhRDZNizhpm7ccYo?VxkESA%MI{rUWrg)yQjv
z1P5NqHvc*2eMaXbyQY*cJ%PdTfzOY4d16*;Zx}V}s61bA+c|<?oiwjCR8pbWy*q;}
zz$8do?E|qyv*kt^Xl<M-oA7BeF3fqV@>6m*2xp5?`AM&%;uff+@{jsD$;(D~BCz-4
znQVKSZAD)~pnt=lC}NWGTA)=t<OMXc`MAD!1(|p|$AjUmMyWP8GvX6r9Uj0AwJ#I?
z!+G<EMrHeTl2YV?#gx2ydCz+X**07O`&8}1=*;Lv5v6iQ$h#M>1l%>SIpCX4O2$76
z$Ei_E=KTfey|tB+kO7t8cmj&0hQJk*%txSOJb(x~>ALHRlUocNQz*tsX0tV>xvl+T
zz8kQ~VLK8Q6#DJP>{30_Xh#OrK?QNqo|l8S&3s=lY6|ab_Fy4EyiQXPn<d$}HK1IL
z4%xQ6ix}qR7<kX*^HxWNftdKO3xeuxZOg!e5C)s6h^Rpyw%8X3wXH5}UFN9Vvi{)N
zMx#LpQs0V$H@w3Z51@0bwL8|Pw=n%_c>Uet#2|N}+B3NVem=;JXQYD)0JPFt9;6$`
zKJDYmajJ*B_`VQN5wn-VvYR1c+)$UQN~2Gi6kZaX9Okn^3sN+2<z3ye*`Z`xNL1&}
zy;Qqm&cN}Zmea_NZmv$?H4q=XAX8Q$-MKV1k1UQy8A4k;t8JyEef52X8y+BhAm3!<
zHC|0EhMesK#x=~h>Y9Q0FP?Tva^a>Dy}S0~cJKhWC&mE}D7@^R^q5QIs>S`E77#Dm
zc$UB1e?J@T*rQ^*xsNu7uqytH#{<@F6p=+ZCSZT^IVT=)W}J)>_{k~|@k;@W2Sg5_
zif}O<I76<#?5_F85-xe<)&JZ?WY{W@D~hf!pJAd9r5y@^(@t79FY{bDnFoGyeO@`B
zMPc;yy4kR=s0yXKc#U6$Duo#E0jVig4*O1?xCHZP5;h`cbU0G%g)MTS{#a`X;3Z9d
zd(?GWqtT5cM4pDh4~{H2XSETIsp4M~?o;H}Vl;!qKc-eQ2>kRKZT>LdP`XLouNckR
z5q;D_V2;kwZk~XRsP)a;wF-FT%I912eodBZi1`|wGPvqLuqy}-#sj)kF%rR5;1Pe*
znOx=-w3%Vuto2-OINRHJPwaHo+Y3HuvjPWtba+kb6-H*QS|Z%oImoxYh=j@`wZDdR
z%B{~jrm$=*omIC0mN@tAk}ZZ@{<csL?2QRV_MYPb7<$`hlRo#g5UMl2ii8V1)4ldr
zT`;GI>U(A)-R~%U&r^>4LqWtv7`bt}r{lYrGPLMd%%B_|5V+8UC4DZiIToE6zE+B%
z*G42K3NDOI$P1P(1ZXX5U8TDg^}KZYMdMqbChquHF!(#aquyq4{!32H;Wq+PARat`
zb2sy%6c13ASfBpKNGQg?BcJHN+ogUwkE<}iIbQ$CvBe)~M6os&J};F3zJ)gHH)@m`
zcgkm9IeGjpg-Eyq`j(tgF`Rs_tn3Y!6cbNf^MP}@Xk4y2cg=6<U`PB{Co;d0zP^of
z#cCXSG+e#%vMqFX{kzA4#gkl)0g~2*)s>yeRXiY!sd4t4cL92=QGal$qv~hbW!Zj_
zV{$43ye32N66Om`msPdBxnEf69w=ZQt1RX?Xt{8iwQ0t`wlos@;^30)@*VL2P@KXT
z6XsC+;)juN?cx3UEa4v*$SQ}FN-*!-+CL}J*#Qk;i#yU#c)+WGC5D{>`*>VnDfZyb
zg}l|_h~j>VTJ0hAE)S+48AI6r6Z+zh%YV=6e)3z0<G)iz&xJaK7l8)=P|X5!e_THa
z=-AALtZ*jVZUE;s&J=-Y(zkz}e9Ge~oxEt2j;|D$Da*$V<i$ddWBE+GROk((Z~k+)
z>Hqf2g7Cxx%oon|kN>nOMrVm^C*YV~$I3jweh659Qcw+IMaQoy&MS7$q*_@7C@Wd7
zu`z3p_?JBHm1-Q`cr5B2&e`|~+AdFwsHR(U4;P0p1i0buYd@T>PTiW=66-0Vitu~x
zMV_xj#Gw-Ci;ZlkwDq<Nc4t`1`|3(Lb9mDyo%#CEP{7SFCDIya#4ei-suV+K-)9=R
z7yH2~VBy4f+D&t5FNbtr7$ha|T&!RHtmxOU#{+`6#!;N47}fr+?V%j<*r18!+*ZY7
z<tk1pfDO?Y#WjneSkn-`{%eQyt>BN|nH@-&LlTE19$?ogu{itMpP1ed+_rX2tb6|c
za&^^~WV2SLBTL5G$+=w=^E2O<emC!M6n`U$1bYRH_37XifDK7G9BEPuV0zH4v{C1>
zu{{aF=ec3gMHQk#)Lb+&q_hbO=;;d+H1won`WP4qtm`OnhDVJw#^AWe=mv=Hr5yIT
z=F%F!u#a-ddvA1_zKaKtC{UrjSJirU<+>G^8}5BK4Lca&6<vDj+3rZ~=R*A)KnMWb
zr@YNRm73e2CKyy9knZL$uA>q{u))?&cFCu4E7;k;e0s46EDeMFuxZ2teDQ!II?K#I
zg^RmW7<ZWT@oy_U{|O1B(eb$U0qiW!=JWZE!X-yQYyXnT+Zu)TOfn2J9$=w;y933N
zwdPJ`9MIE~{UNfVQwTPqHDe}MW>(b4nQnbXXUTS{MIg~x{>T;!6mVw{VV1_x0T0}0
zL5_^{*k0&8^>b8E4!E1sxJ!D=Budfe`#EDP@5y^?4jaNCDt!$|xs?ou#J00SU_WJ#
zD0R}i7H_>y@}rfO-{sD|Zel9pb06l;{_W?~55His-7Nux_FN;5oWQQe=@)VVkyr>4
zL)Y3hFtLVS|L&BR@z{>nrp|V6)0tMj!DZL#Qz#_?(e*$gqcQQBgkyikxI4uJmNXYH
z!IO-Nb);Q+4=t1%75Jybtl_)&uI>8<AjtiE?_^OecM2N<3Sb?(naINvn~ASM>v+IP
zW8CJmyYd1b?q%S9E$m$>QZS{%xQ1fC4FE8G$apwYQoAh3C>l{Z@YrxgCjy980WoSY
z`$)&zz3^2$m<4bI4laIwUXe3i#6Phfo4Zk7S=Wk9!jUb5XwV{dcGYRZ@xECFCbPVL
zi@Uy(WU6lK^!)jc?=mOV?_IK^_SG^K6(9LPDc=8dyyd?KtoqL((aWnP$pFgib=(b%
z7ILgFk3`eCJ3I5FChNwo)8rX#Pb~eF;RY4GPU}&T5ub^V#<?k<iFrBd4+lD-3Fize
zJKJj=Z0O+53^2pTam#bR@w;wDc0D~;zL*8X`eiVi?@AAV(+72%OZ74|xeI?B`Hxoz
zn^`o@4}4qo8npE=sMIUvdm>g!KN%P2y<wONK6G2&HWg5C9X~Q)IG9PfaN5F2esgti
zr9tMHp?n6(+lvK-$QR~@R6*&PLkiBJVGnese4&PDYSf`B`VhXoy`(U@tv&^tYac0d
zG<p}<FyWn~W+45XieTvVgFel>uH=(HB5F?OwpyVbU$(wV8%YoS)TW7lKg=v%vaj;F
zv2Y{d#X9}DWNMH}OVQ*{XuABZ3lAqCjq7?HvY;EQ-OIrnwWd;_S>JJmB7lj&E_J<-
zKqIg7_6mIa3&XDh;tq!e>bUdeYIvT;!*(Pj(I<T+@VyhSH_?|f1nAoKp|CVCfabNg
zo*FQ^SGvNhL$Xp8I;%PZZC>&fAMnWV&6q4VVo)B+|78BnE3U06*Or%uX~J+lWX!6t
z6$(!t0HTdo=-O|hA70R1DB@`3VpcEg=P@hwZbD*pHaq>JS#B`U6ZGXsCO=#&?YWdc
zcz_MW0n%+Go)Xfy3$~i`V_=#4xIenTFZ^Uix<V|1U)5A|&b`uYC2JF6{Zi5FJh=|L
z&wC#Vj82=kJZ<c%hJrKNqal#BJT8qH*a*kIZbRd=ujtz7;QAHE6Fz_bw#VefY2UyS
zl~&{+E7Vw`prY%_)fF?){nSIf$-6bxur_h0+FR4EC@+O?+yXJ5U%Docor&Z-?_ycE
zw(tP8+NSn8S>K&+YTPF;m2QP8zk>xwUxTHJ%C6ORtz`%lXm|})XJgJ!cc+fLmw_3Y
zQC({{rrO#hq#Tvzn$%71H0-`;+br3G`b>18qWgf90i-6#y)>v3#Cl3z$@}EKD6npr
zcb;c4zVS}6uqf}{2b)lm=87}3lTJJ!b@)Oa@_k70D5N^by}w{yKohNA<*2|88~3GV
zL=PD1o|rObG^M<@-}Kvb!c>1D-K*^<kf7=B7lJnU>16UR?=Qt4w#>e7{p|IDi!5`D
zfJLi@Q-X)b*KSZo^>ke2PGuvnTjS&7A%P9s3!5$6_1U9J#aIZ11ccyM_M2gMs;b#|
z|0_+nGn#h*Uuo`FWL?k)sLg>i<OM=4T%M-v=W|E!9aRn2;N7=m9%LNq(e^U8Z{;Dd
z4v4sU>3KT!a<?fzXQNM!mC&p~qsY1-2`(-DG_QnG#!@mqRr162H!l#p46S-z2lrpb
zt>UCG#&u|3<g5Y8bsf5Z8?3$ciS|r&&B&u7^rPogs)q-`+DDA{h4L>e3LMK%%iAc+
zJWt!o?dyFf?o@M)^Fo&X&-xj-h5Ysi-o7H$QU@%e^c|b{h87z-(ofD>TsxMN<tS1&
znCE>e;J;+}Q#}%^{khW|1(pZJ7wR(&gZdZ_AFb5mI^bN+WdWD+i#kT09UdIZsJL4o
zZ*f)X24A)a>)Z1q?#Ab2Yb|&{QUP$M{TfDYLf+cl8W;oHRXtgrmrU9p)aKW>rz5?3
z5yd1uAhT^!<;LD^AX8Q3yd*w-((Lr}h~J~ZU7JSj+KxMa&A0V8JS-B;8wt5?%^&uf
z^J`TCq`D#V-g@YI#L+|*_=a;fBw~*<BUq8aGCGMO=9@>?QzfwND}Yzv4_|0t_0O3o
zPL7G3<|F4r`Kn=ozIBESARCnlEDy?5v_G_zpI!R0x%eMn13dff{w{F<>+Qa-H}xkC
z%s|bDw-zyVIk$VlBPAc1=%h@^-n*3^rNj}Hrvd<tAy&Stb;}mSzChXcw0WDLl@KJ0
z5Ew{6?jc333=f<*0Fj>fnqzg!(6FEZ`YI;;DLQL48EC9_(7((%nb_8U-)tzs9`(AQ
zIz&#zOPk_}+b)|b3{f(cRU@x-;T4cVT#I7rQ6wuT_1(>@s(UQ!W5Si9zW;@A`BM=#
z`dB-*xxS$qwX)V2)6O?Y7;8Ho5jCKh*!1OKJQ=EgFjgI6q4Lzj$7bI-aklI{OMwxY
z5x*KjwV;Z`r8^e>G8}GTlVdi#Gl?WD-wR3`taiFnv(#{p{7&I3DG(hVz{SvU0mTE}
zWgj&kYX6h<{KsC}-v+}byY)Z4Px<3JC^d=+8<Knh#q@a|e+T_J@@9w-9a;p}oUaL4
zyH^+|P}K=n5X4X+)dmEwI$DZ<6ReNXQ+Txd=5r%&t5*#R_|+=nwR1OYY%Um~*2R4t
zLytbLFm@P6kw$|ouB^;=cPy#$!$~hv2=}wkyW_e`$bG@wj;yIsU(YIpS#gUE{0+OF
z$V#0%kEPt%zX*@%0<5q^hQ6mgkonoQPi*dNAn!?ersAQFRv1zx)Fg(B7_?g}o9j>2
zN!KZHEWt4PQD5)-^cp`07t-kDKH3+lrXxEY8QI^lHb^19%ssntVX$3yqmJ(;KiwGr
z`xt5N*+v}|zkZQ0m_d?1FR$O_aC;W^k4s+vAy$UiZzm_d|BgxbA3y^30Zi~P?ibDg
zMZI&Wh1|Y{(!h?`;{lH++E>ykQ)H7$9&fb!UQR{Q=*&)Folydg!-;8Le|`qmxV9YZ
zSmONRaCb(0d4$<IOr|?u_GPP!1#4P4FGKKT_E{sFa3_{J+7oRqWEN9mS(s2_M$KJd
zK&qF8sex{}f%_QmeaC%mr#N_aD_btM_@>uaH=aSocNl?aYzC*XqC>|#3m1Xjg(>5_
z?2ADU?Jdn<#ZKzz=HzZ@IF1eZp<1H4zvx2`x3|WNopBLIi~fnBVq{Ml!`DkgN4Odz
zhpt2E^|7YU6etv|ResrXDXLGIH=Z`tee^!%fnqDrr!lBpx1ga128&AtTkfVT#f;pm
z=ifXyDZlXFg9Kv>MJI7fn&%wBmrC(&?WcISeViLy|H5aPGf(ISL5tG>M<4GuHyg!e
z@s-$6S4H=+$O&UmV<xNkb`Tx9kjxv__#+4p@Ch<X1%I);g3j`W3)2P^z#HR^mL%7Y
z%d3RN6a5&2Wx}}KQlnH}8gV#Nk#e!f8%=Z&E84f2b%Q0R6>bGw^zyBBaBR_p<Z2QH
z3%a~VoO*<jniJ$z_N&U+fEH=XV|w5>->tnLzI{~xCde5oWjViij=PmX%{S)#w7Z61
z%=eb8Q})K@MlQW8H)5!WegErTs)hv<(}iea2W)@$N@@7Xg?IR`W+(eyUVh+w-72MZ
zsBiOUo3>6t4(gu}{#n%i&MsmXh@dJ*F~Jq%U`xPfGd9M@_l3=3v*&f+UF)on*>uWv
zC1U^XBEpbJy;F-K)kQb0YR7<iQR&GszSWXD!{zS{!<b!XIrJrCYFAgY$OP-;qFDwd
zRPzG#-beeaG>I=v%xt=Co9$fq%kbG8RAj}Ct|x}+_I&mVUfgn~-pRFUc~JvP-`PuD
zQ3Gk>0dHPmc=3Rb>|?n2R%k`#=GZ*-DsKYD83pVvxURr~4%Zw9X}|B{sgcmPrBGdH
zBH}uX<tAeI*+WEQiGaVYH<p(v8K}`D?uTX63Lee8iz&{%dB{RQNN&dMAq~|dW|{Q0
zmagBIEKn_QpF+#y8rWDe*Xzp+;@|KE%cxm83~G6`7}jT})@c@qgTDh0Od$5j7c6SH
z<oaqzydtCZgo3PDKyW9JY6oM7(z>tcbN*&0W82n6Rk^h^t%op7AFxHhQ){6%kQ~;$
z*#`t-1ZKr6s}+_fyPG>sBDX9bZ*!-dy>8$nPoS{5ebJ}JRqzxTiwD$gCd?vuF@n@l
z9GsKl1=p8MtMiL&tNV9dy<XiJ{b9&;i))o&=nC_TxWfwFUR7V7t1YszOQx9CSPQE$
zt&^r>lb2F_l@XoK`lge`@5~scWk<u(4H2wKjF|%Q?MY*gj+$2SfQF}d0E;`0-q-@8
zoZM|{7~|_;$S$WMrtbCfV4dG$F}?HIiVu!q?=od{s2fNdDItLs_HtgP%NBH+aKN2c
z*p74^Q5<4z83r)s*bD`JkT<eucy0g2FFNOwn#2zd=OA6PU(aYeA0H_6r>gS3ZnC5X
z(OErfr4f!`1zwLgh%nxnA^<H!9-xH3cG`)wq>{2fC@+mH*JXY?B7Nxfx#_-JTAkng
z$ZemOEGIxh4gSjTJQ@9->9Tcl^|E=(F_o`S9`0Q8=9bF(mpCnSC_-A%NihaU1mbzo
z{xaFqCI9ZsnpGB!{&h~<o#DIo%PJKs_5A1}T`UyL=!j#>(TLWG*9tu9pC!!U`Me}z
zv!Enb%W6xwPgDs(YHLRsHoGRrgWRd}1k-mHeg%m)a#6Vyd3{|$-sZo`j*!@ez<77?
z0C!Z2=7f-|gb7G)t&Iv&?`Yaf@wxkqs1l&W|D^nj4AB*_FQR1zy>PDBx{WvKsu@X3
zv>{V2vsu4v#LOF`(Mc@k>>sCY=@y73p4l6y59ucDrw=M?1gqrRA}$3IczpRTQmd&5
z9c_WOWsD7W<d*VkYDrLIv9$75i4>R;`KjtWs_3wPbxVN+8!b=$&Z#*ss?OiGvX7>1
zl2^=Z5O}Vhz?H81K)G_Vmf;u2O*A{|bS*z2pAj|C3#6Jyo9b4ENJhKh0WPg7H+P@g
zs<3p9PLP&yg-_wW0qYFLk`+iCfYczG+XW>(#TE;L)F*Fh%GcFy#Cd&5&`4cPOVZME
za)R3%=9#F{POAiu?#>$p`Dh&5cG??1?O=LUWJU8ZLt<K;|DF@g6Cr-x91SB#ByXW6
zY$w@o@~-uy_=1L<!wg$+fG#sa=~YDqfbctkVL83^gCqL-W7nlc43(~zoA&Q%l2iF_
zmhSu3hdjkwMR&h+QtQQK$2ra|QuE^h5%&6cKmiMcUlGnSt+?55HPmJpcZBQE`y`F4
z=GaV!bc<C$#u7Txfyjidf<uJxx@Dyb(5ifnbh>Eq0E$-i)jmUI9u(nd^7L>ZJ51%Y
zL(%rJ+|6@XaJ~T^aCNjC4^R-r>Hm>WP&IuBxby|kFxh*k<8Ngq|1BS=L18j<YA1Ya
zd@s-l!zEjaK@Xql{5>CoEtt;ndEV;~)BB^O-9bFqj3&yu`9b>Uk|xoF;TFgAMhs3h
zu%@hWE%NM7m#WuC@!$H_zlJiv8^8R65*niBN#Ald>=u<i+4yr4?3X6w-MPaJlPOge
z%M`Q!*W1DQvMMS6>y(WBTm7JL*$vc{YEA1_e=Ah&Z#fxNjH}IV`6(ScoXS`J(fHq5
zrQCTw#5&V|u}x7LYSbC5mKDfvxNWdj#O<sV_hIk1dfD&xU)gK_2S#U`-XR>rQGyu0
z;Yb4f@Bo!>ZuJx`W(#K$WI&x1_THqj;wv%XZr44-f(|W1RT${utk0f>eKO^OGQ~ra
zp?9@zyXj?Aw|&2?iv1t&YmL$|p4i-05fmP9!5k2`ZvbO8NgHOh>wd&zbjReyVq(fg
zeg8c*`Pm8`<LM%;)<FLLpZlE%l*vm@=~i|%@|3phB?}WA<sh;&(x-zcz+3y9kB0ci
ziK<)I%@zhy8Pv+<Q?=B2Q`LFXOs90rZlv>T!Pa$)+%%UOD_!>%_I|G>`}$`WZ2w<C
z0mMEbNOh(V4+x6~nT$(V(XEt5h>uZp`E%q2mPvnvxeI-P6-v4bQjh-lNVqR0pM62n
zAzi-`#ws;XfaV+P6{P5`_Lhp;6*tqkSW=*fj&u(*o2yW{rA(%|2Ysa`5{IP@FSGUp
z&WrRb7~hr8agrY~^*N4nbsl$WoG$iSGt-~6K%5orTU`a|BJF!wYHlGL<6*+IBQ9d%
zL8_(K79Pohx+SGC&Z7#q9O#z8lR&C^adaavYPG&T&#F@pZ0qXHC%jLj;1w%rFLSWK
zVWpF#@ElvNKo+2OVuv0_^7gM>6YoQ&uW!cMOzc<hyzpe`y*nO8o^jKTGS2ff+<T&B
zEV<(#m3nR;*Oj?)&cD&-a{FO)1U!DYa(K@0eN5_RL||Ge3!e%!NOnP8(BP=<kydfK
zY;9RaoUm5;&~l}INbs5!!_R!miQ*}YQkAiTsmlw})~TA*Wd`1NLO%%3W(Tbr#`^6u
zAD|U0K=P<bg@UqxnVF%j(;n7#o2G>DX5IiRsy5N$lwI>&Fy*BU;Uc_HIWX^@M3fB;
zZD(6hakhR3t9h;Ng{Iu2Pgu{;4SI3+`Hw5qzzY}La61iJ3SkwoEP#}V#&9*e6|J=h
z2&$>or{IRUyG|d9_N$Y3J*w~=-~zDE9ewua8alu%kVW?!ru7bSOlGFYI_KxSdA^W-
z)@16x{>jhsCG$|4k8_!Dq`JasL&Fiq1)Ss*l$RfGegEXWKXGDuGt2F!rWD&G33ijL
z7(j6Lnfu8kdc9-|_{l#>0vgqE-DS^LfXwC1mID1u%0|PIFKx=kgR`gQTxDNwe}>Md
zX+!3LC&QQK+(U!rrsU3yY-e=1(WizteJ2APZc{c?9z+iBk{+adxw)ETk0U3%V8<vR
zg)g;%3iMcKGdGxa%E#NCVbRnlDaC5O74<$zfAJxMH(S*9n@xG!<fwzEb!2bXwMD9h
zT`iyG+Pur9H{WHD{E@fm7F+fK8lsgom75FKv`Tm0n%-%nv$`)o*3eFEWxfTuilGY(
z_7q%N>W{V)pQURG<n{}t5EJ>4s?MF?^uZY9fDWk$h;U%EltYdzH(L#}nzQz`EV<$K
z-vIqWYOQF%b9exg8;IpX1VfJ0bvji5PPAIt+HusEn}pdVYsW1Uh?d>RJ?ng}fK+Z{
zj$6g(tgU>v(nGs?!`t|0_SL6SzgOosQRVobeBY!vcrrnAi22T=Jv<BWjPLf)7|=D|
z`c{2sCU+_%UC5V#Mfy<NVCE!lX0T-^$;tQ3aWc#aQ@+&Vj&%a(qT_p;)xwXQ^Q<hV
zU0|BGzw_ne0Xm;G0yu|>*^0dwym|@Jrb3Nb_xs(WaZJJVGazqxHP8N<`Yz*zQDelS
z>SE<UQ~&Y(m@g7YDU2Z27Quv}jqjJrzuA(hGkp{s;}Z+Dcbzc5>z@H-6OpfTNuf_0
z6i6F{nqvzTts#;5B4}cGo>hPIjeI_oVyt+4@b=Qn^F3QnmJFU3k7J(G*y>h%W~aBm
zFECQw{h6|e+OWaf(s2?hD_P1QaL_NhVLmEp2yPK@oB7x~1m&%>qII{-a0J?!Evf{%
z-#|V|>*9}Aw|zderK&{ap~U{{CHJEr9)8*A1sqvmxf2bD3soAqFKpKaLypA>!M3^h
z29wy!{F0SgTWqHjL(-(qvlN#kP)!>~*!ZlO;j^<d_O$w%yllT;MW*cfx3^Se1ms{?
zy*f;CDr7awQVIo324C3@ijF<T>h=1g(5YKQ&C^is=5*WZoysIBvE&rAB%XuE(DxV}
zQ~~@n^cZ+KSZUDPJ(j?ZPDQ(EiQr&Ik04OZDlYv%3PUHQIaQxL$2GvnDp@A3X}WDO
zbqh8%50u(Tt%nit%%!wA-b8SfcquR;$98;?irt4l@c>$RgLj}usN&Qo=M6BMrYWBM
zmww#BX&YC+j8ZY*2s7U3cslBJmqRhen|@kddg@cDYJq&!%%@;g5yM8qg*ibjSzjJj
z^|N8aBh$JgT17^ORJ14NLH20=Ir)QG!HVh$HiK9t!CRSf%CE=-b17N<qcA!}0mG`{
zMo1jgw#HG4tu+TLxh`}w&6&aV(=C1vB~sF`tI#mWYL3)ARW%|x8BB}ZT)2&PI=!y_
z!Lj+R%K&A+&vS!8MNuNX{Kp1k67%3r|HI3k@$VZ8m2K;yh70P{-xu6;d+{*`c<%@h
zs1>p)aX8mV_E0EbOMmOq5~<LuGjz*)R5Rb=^vo+RwZ*jV@fw>>TRSh4D|Ec(7q0Al
zx78Isg)$AWuE&aio*+ZYeQX1P3zYIW<ii&5DD_B!r+%w7>S0$0wIvbCDBef-nm9s4
zsJ8r}!wcZfS%MhasK21jJx)7T0uQ*9NQLc<Qiio^zSi&VzN*<4kZ#NkBBkl1QjE21
z2V)Dsg1AxEr1Xk_4#d*PN<#bX?M83Sfcf;uzORzCZUQ$vBiR&wl44&gClEtTqpx_6
z)?-Ull=Ko1S~m`T?KvD(1d5BUmGI6!svp;&8>*?U%GNF^YJuEuDX2PEt+igeCVB0l
zMtj3ll@Jq1$}qL6_Hlc~)?==p&N;2<zEyVFS%an))IhYLD=Y{4aZ60I(yY*A@kR1&
z8!MW2gLcENRa75{`T_{DLTX1Vy5x)zgE*%pC<-FMcPD=|Yu-lj=uaMBML4nz(_+5Q
z(Xi~XEAc+>(|!f4vybd>D45SYiQ<rr+crRMM6U&xMyn+qm<8B1WpAto*4Wy)RP=?0
zDWJwmG~LL%*QY(jzE*0{mrs@&1jp)o<~GTgv#D!7L7}Td+-(^7h7`6m554w3l(#^;
z@c^--9JLwP@0Ir%=Nah`*YaPp6zyy+G8x7tRtCP@G<(-<Ku^(G@UaLEiYa-cwUI}T
zLF}*eB)SrdWnPk3RJ=>oxDFs}GeI(i!rY_4R1+XU__p@PpUwJLmS?W@4ZS<I9pp|`
zmrqr%1al)B5)tfT7jl*pRmXYqU((n#(2Nf8MI@?sI>R0URMm(Gt=<l7rS#Q+?Nu}e
z)%)=P@#5JE?WQFTH>==a5EN&mK!6UsFhwH}(A@##40ruZD-gMMGJ5VR{P=SUoV5!0
zvVckNo!pDzzJuf}Bl{usVc__s)w?zkaj8AM$1oTW=m*&R1dsqYEU+O^E7yK)3dNnR
z&$9e)c$9_Ho!yu9P1}APO}EXBsi+&2eb_s~119t??Tijt+3r?~6BJ4ax{3$PJsikx
z53gAMnOzU_`#^z}h3{_0OL>$pa9$ry=f1%a5qCO&)`KJG#RG1;qv&u=CV0Rv7lxxn
zk*iZ<AFpRO`<8zwuma!o1yX>;(Rw`)5-@9HVryWLW6*L5d`SVGV>QDWwef3tiQ+M6
zmU2BuM*hb}n~`3IpjuC(mzfjILeKq5m_O=^q-qHj4<)ry-W(|k`dZOm@w*yaEf!+G
zLb0WA?;Be$juS<d=<4rZX=U-%K1QAVEg~Z7ibNuL;i`{L?orhknm50sOzA=nqmL|E
zI%=HBXP&8Nm+ku?!$+M;HU3k|-6g0ATK4teX+xD~Kx=brZkfRor)N(DwKRDt)LU*>
zJM#G2e;8n-D$!o#6XdbmtUo@kQ;5L6Ia0?cB7^!z5o<mWrWa=A-_eBvlYX~{@qo%h
z%1`$8OZ|J71f;F4guX{TMt;o~M;-*E6b!S<Ienz)URgTu-ydrbIP8FD_9Ysv?qg_c
z_Km*PSPq-CPP>T@n|+R@Cd)=pZK<i!M0c@MQ9GTp!J!zs9{6>%K*=xJ(`Y6g6vJ_Z
z_ewy-*v(cZt<dbp1&TKf75NcF1fM0fg2i&9i6i=KP+9Gq>axdt1F1AEo^@~LTWt!P
zGF#XeK7CSU4jR>mUlow@uk`zrX6bIIcyigkF(6xiRGIVnNombB#eRRo23y_MRK(7=
zFwYyo-*=SeREa}q*4*PQNs)%tB$;bK7Boa@Y7@|BSHAVyO!Fo$p_5#g_l^ME>U*X6
z6qpc1f-R{Rh#f7d3NSM6xE(m?4m0l}1dQZsZ9dZqajwQDDZT&#Fi+99>=1ZBtSLVl
z(a#|&=ET$-RV|+MGm(p*-#Tit{vjh!(l}uiAQzC#GhFl*;Z13$#(F(Z)>D5i=S`l3
zK=Gokx<Gq1thGDR#On^v{u8}39$v@@w{HhYtM3}S6^{HGs-Wl8^48QoNXE8kUem5b
zGnhNWZ)ea+LxsjBPHRKvnb_y^*-6}8)L0aT_<82?Z4^m5e1X?B-n-Irz?nrMwD`H0
z6$P%pfT9pBa^VNEMTXR`_w5ywI=rUf6Z#?gIAe5XpVBVbB4#Ljj`RttfE*Ku=0!;%
zqLW*_FO)1Topt-bCkm|JO6v31Mn?x(UfI-Qic><v`~7Q)K614HyWbpwx`ru6WW_zu
z{uiGuy*DECUD@VjocSb|QiE&zVdcSU_ZN5DbF9GgV2iP7MGAC}5~}@eb7r(}o1E2T
zR$P3W(4O2|^`qs81e^p>z5=pW_V^^0H|FTgOqF2!f`6&Rk@DK|O7e<2$L^PcduK6(
zDPt06De$qx6*76^3p0$=G{pX*W#JlA?`(M?x}ax+>|LGb-PjwVh3bneEFW75nfs+A
zLkQP^UAt0<0u23_C9=Qm{w#*;VNb?vb6#Y9Pq*+|;;lK#gb4P0XVqSNG|4>=wG_S3
za_d^!^Ck6(q8&REZ{=ww?l*R1$|vp3awP(NOSgB6J*5&(+~iyC1I0AfH$vv%!g1~7
zDD2e9GBYYFbWJ1PGwc0?gtSn;$mCrv8#RQ5CfFbRU6BC9ejyPsygJ*f7!{<8Rx{w>
zYnXJi{g~csMDuy+kk5v6C@GqGk^5F%A2?>Vf78T}jx@T*_9Fuizrj@1htxOb?Ad;<
z5*3c6v-+-D#d+#_+0)!dhaG1?Zt$X|B#J4bod-O|*v!#U;V?5_8o$=i5v|>tO4JfF
z;zsFJagSDrC2Ho&kMr4S)BeWN0DrBVXVdPxKEqd4f4M<#8y6-7)rkPWx7+0|#d0Vf
zFsP5sQrz*2#sk#Mj2>QCV0e*6aX2otRG$?Sj`P>KT8|(H%N4ZpiRvq>CCtEAs!t?o
zVOa*|uOjTJrjJ(UzPtEYz0s}aCB8G`+E3-3X29dnFA#V?eO{xIOufH2+X!Y`I;z%z
zHbMx$lJ_mkQ|z&nL9WEOciwgNskU6+FsCYYl2I&ivJBYoF@?lcVMH!jgMXp7v6f}m
z#}R6yF^FLM^oz`Or-SQ&04NZ|iur(rwmM#jmqx)7^;Vx*$PJt;ln|j-eo9IiejH$4
z)H+Eqpm4it@P;lW0U>d8x{EEg3UF3No1wNJN3_?j<~^C$fCGm&=DrR$8RG%|U^<R8
z2|t6?Y#L}-c)F<u!!8%|bXdgn<C0<D5L>A`#YYZ3+qCAtOzJu8>=%446b9z9`|C>4
z{!f~xj?K1R#WZ0FBc6eQqJ5NiWbW?u5)(G@GBBpC-RHF&zd7SGqC2KDz8OCn{|xq7
z-StQE#WrVZNq~HK+!@yi8vP0<k2XW%0hg1!Q12-oz#t)D?7E)J9TG0`Wdb$s^(_rI
ziU&M7bYIx+&!+CdYOJ1ivF)0k_9#-!%*Y&XOr~c$R=wxEW8M8D<)Rcp<i_92UjM^R
z9w<M`Y%KG6s%lHlYkBIIaFF5B;NTOHWW#E2lQk~3L1zXI`$NCDGdv*WlsXzLf+p>o
zA9iGAN|KkVb{fu`VfCwa_mmzm?@uyIE^!Gc?UyADJ?fhRsns~P11INIA}qEtJr^Fm
zaM{(|d$8!?A(j@0X+aMn3cgPoqpd?+SJ{F)DA1FdjZ&<P<)y}!4`r8Br5<tUord27
z)TOF=UFDLP#f@LmTd$n}L_T#VL^m3}8LZZ<SWxoyTsC>0x<<!$>Mrbgdb@42^TzbY
z)F;*azD^aI1|Qyw&$~X6t>>NyshX~D`wEZsauXUk7_5}rf!bsJ6_s7vdC<zIeO_qG
zl|EB$98bK$%S6T^3d`FhTGrQEJ;$`tzYTNtuzlGie^;x}-5!W7=&;$3hFo=%zmH&s
zZRb5)qxW+;ZN{!2O{#l)ih^;<3F`zrxmW0KaHz{j9BG2>vnPgs+T72#&^5T3QKO^N
zn&$sW%W`|Xa(n60ZsSkc%>EP{6nIuZgt=FTuD$SaTE2ry9P1sU0SO?9qe_KxiX4_%
zU&5N-+$8;)@jyvlbW`qA-z#!@`x3P9<!lqPaz)W89ulgcfO1*64!iIFG(`Ksta~O_
zQm{~!Pk>efQvQa$uj#51ARVX$vPAd5twP($T2b2Z$KsQ)1<!iAaEY&Pd2C{3TXH=}
z-*dkr0}6wO+DTCE0|iM&cmS**S$y_eRS8UzgL<52EQ{eiQQ&UAhd}`m_E6v>fuH@+
zyiD>*l2u~w8WLF&ns9?pEknw{rHP(*sFaU4P(K~VcIcL5`f=UaY@`LMv0eHCc6_KX
zQy3?JN*n*TUKRf@Bf1jqk0}vb!vnB!4ISq~-R;wbZ97nL+(;Dvjojt=oYJ=HKvko;
z6?i>tC4{o#;k=5#vG$(7g5af;ONM;0T4ll}ovfVG>Yz?>x=I)K@qq857kGe*`Ru&D
zQA^lba)q^nbz34fPxfek#QEaI^zLw};K>YjoPp2#a2s_8+N@<&ogC2`5?q1n>!@#x
z1K(M3U`?t+Ru1@$*OY{mC~Yca`sIKfNk6QP46@MXzW6|$1)c%Ijuc^*O~`j>sfW|i
zJ-bqUC%4^coFqRi3$$#wFZEq@Gv+7L0P2d85_f2oK{W{bnVlHQY-^>39b-#~i52|m
z*NO%i5kuHxcA2K?l12COM38tzf3d9q-}1SYVb%QcHrg0HaB^wR$F^Swk8_Y<-?z%W
zvkWLwo7gE(Lvdt%_;3-?pz!jx0s8~m#*ZQ&9uuYfC}@D}mKI_l*!+4tfEpz6?RZ4*
z=}7%Rj6_Zg7jZ)#`54RNVY;u0!f)<L_pga99DRrz5L{~&Vl<XTF9PeUBR@qnE$SuM
zeSK3jw70gf?pW#gwKzThQ}JTSohOS6*qP>?9gxh0DhP<OPrw80lP^azl0AM4cNYzD
zE#3O~E2Z4NDf&xjE@zTO_(PU?87>bNX%bu|9zZOMjNFglg&|KPLE643N#m!IIJyv#
z`NuV1I;nd0Z@MdePKFbqPD&U?v}5o2{?zN)2vBq>f1(@t0d6PbEv>PXDIQoqnZ9*A
z74s}z!AQ%{k*9vLVRWFk1Ezsd`FuGV+yW}L*NQVX`~9EIkNFbA_^vk^wucWVhM`)8
zq->kG8k~P=&-LIm_u-k?cv*K`nwn*zN}%-O;dWMV=LG}BOrDkh;~Iy4`8dgOAq#d)
z_4MMMgCS9`wlve@wt5V8e8*}Q<N$ykzyoaCR&cM<PkdIkUf=;flRs<B55;kFU!68?
z<mR4CoYtqB&H3#(YjKR?Tsztt&kM3wJ}E|nG}l(pyheRXz^)CwHGwumoSeg|bgIse
z1r{#tll*%>?|(PAw1b>a($MV4Xn0g4ya63t(Sx$mj{ed<;VLgaxz*VC3iI`o>)6*N
zhPMpAUIC!v6#}~==8kIj&KE_DO}q@tcU@Uqs}(FQO%&9|*nD87{&hw&?KQK4`IR}o
zn9CU_+Iv;;yRgx0F}6xUp)>)O28S{@Q#e<)C*F-aFzekr;Q!$}pVm!wEdMH?3X10A
zNP%Lw(TVRcI>;c=rGQekR;1;J*T@O++Ix{vGQv^Z!JX;nH|OR(6HZ1zX1^n872LoB
z92Ao*X;Gy8OB4^0&~Oa3*XcPpt9bG5%Dhn^hkh~Hi<^7~?r+Nht`vtN95p(<^tFNO
zPjStqPts+dh<#WG`;;Ht=@wr0_e{z@a9GU$F-ljeRg&|jcn>!BbMj0gIDZA%assYJ
z`tLMZBl8j02PDIV^2#)rWb)S}FCHda<*RWaq3M?c6f$pyW6tUJ263X7-fgq8jJuV}
zDzrGhn4DWNYfa=G>Pb5kyRI7GA~{hE{5B+jxYw#$La9kPk=$5n=~M)3;;s<57*8)w
zpUQUZ`_RIQwYZ~v&(C8HQ?<XmISi?T4k*T2!PNS%x%V12OAJJ{U!6BSWA)lP;S^|l
zvYEVG<hw9T7p8Bmru!WMv8FzgPM(M2`n&Le5~li-dirh8YfKNBqI#?iqzxmbxeL1A
zO9%`uth}kNoDANnQnclDbF?V8LOKmM-~rR!ad)6`vf6QAA;^$UI|bx>oi83xxWB#Y
zbO#zNJ_3Q2!yOO2F36qoPuoiPtnA$5W@cTF#h$^|6*n4JQ?XYrpC|ezut|-+Q-I0p
z=+!=7x4*2&N2Bp0qw;jyjyf?)1-Z<eIW4jJU9btS^W6MN{SFpzSn!WLV<iagv^vnd
z>|^Z<RtaG&ot5vnW!v4;Ii3FT6QlFq_W~>M^_=vEnbSrMH*2lTI{kVyE4V=u)Tk(h
z1AypS(RctJK-AZa%+IX5aM_YF%JxVPtJG%bM@Hn6AX!8M*ggdry#rke^04S%mW)Wo
zJh(6uTCC3#WIicqJ6X_k{a2D*N0Z7`D9cXG170pJUV@Kg7h!Ryp5`V|UrS?5lIZet
zw}z@{%huq03t8S??GCW?>B6ICht1MRpCS&E;PK|RW~pFPEM{!1<GWthd}d#A1O(<D
z1=25=Yo3jf*xtUcuFAGEE85-t^!aN)IgaRhxmD6#XXgu^0CHdigavaS>X6TgL=3x^
zn$OKi^s-iws5)RWh+be8lrD;if7}Q?CsvU+VtV$~*~U<gY(yFMUgGLldbdY14?@VO
z$VgMVO|W`;fKn{Npk$wyHy9$;q8YN{d3JWe8Zcmi4!`ilJVyazU~EuID<@NiR+OVu
zkITtP??OoKV>y`^)n325PK`a#8%w|adE@v3@PHmmxo;AVU`BMZzN?uEN;1dEK8gfP
z>4j+Tj8j`Cxu=}&q4v{`ej62e!yu@(GQ!HJTq{pTPtR$2kcUk5(kgKkDk*YY@ycMI
zYZML~&c^vQEH&a%P<X(lCE;YnhMVYM2gx}j3!b!2w$nL=mY_&u>Z_$Wt-P+rU0Dq8
z)Cq6dd}&9Z>MbvT2k;rAGO@8Dc!1oK3$3`rj+wg>Os358oTXYmcIraT!^^2@x$DFG
z!y2v(&OFY};2}+gK#PVphU_3=IUoz9OH=WO4qEL}%1&qE;xpPu3geKdz|xsdJ3@r1
z%Y9>J`=gWd^QKEnn0WnNLj(6_jXhUfY2OO+nc^9LK^-qP=^GyKbK7ZE(ebbZBZ+Mi
z?*M{qkR!9P?ev31JRdT>ToTvwB0F!KrR0|1!t~E8DYhB{b`eL)ap$1A_>Q`taO7UJ
zxZ`A{X1#)mgrWIJzd1Bj!}(&?Xwr<%Fu8uCWr^oh5)UW_KgTVsu#4{}cGN94L%NfR
zw$zcQ>TQi3)1as4_9r2)&RWafhZ^&x#1EBxRCq>8bR)T4BAM)$A{!3S9xfg$nYN}$
zz^Oh<3COsvt8%Jq5dE3q=k!_uY%E4W>Sxf)6kr#G;^#6~Q_CTIihJ7rarSV;K}j-b
zywW#~YvYG7^Qe<at60G5!B--#XrEG<e(%beCqjpBQq2c@RIBo1RnP00pD-{uJ~=V0
zZ{06FbB1)H_i<z>IFKHsf?ncyk9Zuff((9~);+`D`?^t6etW=1my7d;A5Hej*n&TN
zGeH~}g9ls#%c9vH74*(Whph+rYI06%XiKD}rN$BP5^U(bdVEOuJksg7V~%%~yQ6Nl
zTfmh<w?$K<cRuoK_J_t0mb-P98j0U|QyG+Nw8UV!lxuB;Er;!b+OhU;;an`^aXwSN
zI&WB=z>b<jUKR$kw&e6oAK&o!l$;h<6}GDjXQ6<Xn#OAJIWQgrluSk@_q21*f_a^2
zu=&GJGM{jNf4az&lR;|FOX$H(`C3#}3PVp0WKy~y4F~}Rdpr9g2+O~>e>m}sh<(Ya
ztD}rM2xH<3(@&XcKQ3%-6{fX2a}%bB8Ai&-rK{*1YG->jjBoxhueE5hY;G8hnapoJ
zXqz-M^L1|D-@`nacgKWRL$*^Pw{)DdM@pl8EFKzmqz%@7Y)st>6v?YC%&<iRoQ@fg
zd^OI44z61K<RU@eF~42wIBQ2MGU4P_FptpN9QBosrsgg3!f-aWNBR9T@;~Ie@*Xyl
zKklMj0vi|ry)l<FXjl~JE_wk){TQhoEm)7P57GF_xN*`i4d*b|<g$y1NKU7cK7y8(
zK_}{Jcn1&ToPcwO(cChQC02`3*%H(*pf#Un4ub2F#QFQRK<8~o7^yj*-+{I_)I9!+
z3qJoRBLRL6@V_U#bhP(&aeQItEi7bZ>*glp<Y8z3-(AMNd-o*o-MjhwC?zTJm*YKA
z@qan~1~E~|n<8RTV)rB@q$I>eZr&4-kQA2$+`RYSU6Fqce7$|Fy)LWb>fmJO=56Qp
zZ>76T_VVTbMLz#R%FREH|5JdNkrZ$R0MN9yy{WBt)A;6J;rBHF*(HizV(}&Z{(t5_
zmZg%uzV2U7e?Rf|5*7X_3Blj5GuXPj3R!!2xY!AK`dWKgyZJb}*#SI-@7<FYk-Ho>
zHZP^c#N@mkti2rF>|e;qNGjhGRZ)5*DE&xTTu@x~k(8jalC+56J!LV`N6L>x#8gzJ
z{~Ma+e;4>W|Nq~`{r?{SU-JCr#ed>|aS74?pa1`#0Q`6SD&YDPRdrPW07pQ$zX8E-
zzP$W{hXZU>0vs=?)zk#=UnaT=ASR&wz1R3(?h~*Ahynk?!%qm<{y5hqxbyckSLptJ
z{#PLQGvXHj4*`T%uKfO80@3A<_$o0m5fL#tDJjWSN^(j{3UUgHYgDw?uTjxZQBYiG
zyiP+$&%nSyNzKH}NY6}5&p`h>69U4^XNZW&h>6MQuTfm1|4#$H6F_qnFmy$Rkbo0#
zg@%BTh5+9S0AALSnBXt)NByr{){%ttDj7M&WrF(afGY%qgja|Nf7kl*Zph_*01*u_
z?JW@{5;`4gQch2L(a@ClSGkmHyBKuGk=$Z7USVY9j7-cdtUR}Q@9^=9OGrvd%g8=_
zq@t?!SpA8fzJZ~Uv5Bee3%i&04vtRVKE8hb0f9l`5s^{RF|l!}X>ZaqGT&xp7Zw$l
zd?+m|ulQ64t8Zxh+|=CN)7#fS@MUmlVsdJFX7>BsJbVrDV|`=u=dUf)!Qs*I$tn8m
z{5M?$078Pl!M{ou%_Ut|h=>S@NPp8saK-<30vaOXTOuU1N;;(0o^+g|p;zgZQ{LBh
zk#ULXA{lJF#>pAE#o;`t-=zJ8vi}%iVgDt{{vhn{bj<@O2?;JckB|nS2*B$7pZBpV
zKOjnbCHD>R7?A2g8id?6I?2x1H=nwTVObqsVRf7l4Wx5wnDy);)Wm48JogVB_$rOL
zr$wENfs}!~r}r>iaJHtqtt>7Z?9RRSZm{)W3};3b{p^3LHLh%nyRvB_Kg5$1ly)&h
z<v4OoB?jp0400FrrUo5o+VL$s@_*JNTbXUj#i2bf-2be9LLC|Z?9nrmtooQs@x#4a
zt^KM4Rt+5xv^AbA{8lo)F<GeYFEKIt=}Mp0Rh@64+qtu$poi#aUsM<_ghDxA!;195
zg|IL4=*$$eZ;}{xOmF4at6;@Bg`sOw2-5i90W1aA#&^@XlwOe>YKgkerI{vJb2zw!
z*!}kN!2JKNs(tI`74=T`p8#%vx-?-M2#&x5fLp*D$H5&}(W-kI!X#!g(WWR$QWE6v
zZ$x$e>{;hDyQgD7G#>B_y;-@0R_m7XG6@i!`zkvd^ITdg8eLtT^Lc0NTe5q{jKjQS
z4qTJ*nxb}?z%|K<`{k_2vZ7RV#aQm+AK63R4Ke#Y)@`L7_dD*5!(64hv+uHk3y2I^
zU5kTtCEz-0+^G;+_3c54!_!{H7=J5f;mvO_z902SAL>u1Itjb3O3}s$#l_?MxVJ>;
zXcW~pf*5no7%898P8E7$xh0ZUU3PE#hW$l;X~!g>eQ4;i1_EZr859ATPO9Mm@&R|f
zDV$I|9CK-{5~shpEY+}~_AVnzXeyBez_0)w?l>^T*x>=sbKJ3%v$rfDJ1uws547Lv
zgscG%_%>po7}LX2BAj(>N3UbS8}(mXg|1K1QPR*be{ZUqs^%lM_}7~6#e;b1L;2Rd
zg?uSnmJ%Nro!?@I3ccRtFuC>oD$y%#rR@h4D)!$4!ZoJd_jju51vG{iNc3A88uKhQ
zURtNq-jPvOyLsa!jPG`KGsE6xDI$@7a@z7ALtR<`cP_#5t1Exu(%DfCMHKEmYRsqw
znH=Hv;}g=Z?yWEMdZ&jwHOkCP%5$#ZJ;w^@|6uPuqni4<chLwas1y;DCPW29KtZKR
zje>Lm1?eRqok)`wArPc@1Qe92R1qPx&>{39(wl^ufb^0egg`=a@OSP#<Gk;C&;6e<
z?zm&z5BEbpWMpOUHTRlpJ<qfDoNMB#fW!lq1pnrRFc`m|U|Q(i%#KoM1Z^l39)q54
zO!#@iKp>qs#OFD;u#Z@e?Gta7kGb-rlMokTWCc=)$$D0r-;3m(63Eg#=zu>nf0%gG
z#$dv^ZAI%iRhvSzI+EY&3j1qKA-q6#QwkXLU(P_iDEOtFmV4l#FHQ!1ki|8i{Z1Yi
z8Xd}9(G9QP2spWSaqIi%G1r?&WVCf{v~RxCJ-~P014TXTS_^Q90x&gwT!>bMKrXkr
zU*6UjwmxGpqN>@__8)HW`pOjLUtDGK{ozATRQpkm(R0-+qtJg%fPG-9&yfA!OV0TH
zEjgoSW5wY`(zD!X2wx@Nn5P?8FwZmpVQpz8TDOhOn%aG1({1{A$uGz`|97sk>)j#0
z+$IhRH?SMEuOTH)=4qn>I!CAgCYB}eu>f}(g=vurz`C4dQUM=g(-nvO|2JAY5xgiy
z*G6^%Cax@^aKjWSDj@f9F??~Jg>oyA=mh7c0!FTnqj2u@Q~;ZLRnA`dN}|{NR|E}?
zW&6dv7_roTQBkFxZOuQcH$@B`=0|fgNwUtmq+6Nqi0hhh%ea&)*WV-9{^J~zekFdG
zYdsDX_lnGmfnT&Bj&)Vd1lDMCx+PEYKwwL-s)sjT`!jlNC*|BDIxcCWI4M)<&4VSM
z=ZLljRk)AHrZI$ora~>pg=tfc-`*k{DvS=ubEqQNos&|SP5`!~*E>tTi(9pp;85U?
zeR#1WJeiFj_6<6Cp0_c{<k&IBJ0}K?RF6&R;25z|Q@A3Q*$`fJE=IEn^M^lylWEcI
znjKIJCurvOV#8`rzj%l6JN`)%0lB-N8=0kVad6FERPc6*GZMSIFf_s&2d{RIf<2o$
zj(n@cwZSKbYkrToxCuf(LkE8T(vbnpy!xKmLro&i2U`t$PI|i5px34ikCroxYbeVS
z>@*rR?H5XtWE_3v1OHIo>(vv~L&tqYTJ-&Zkq1Ul0Wa~uZt<%`dh-#xY$laQI{u;p
zwLZR}hVZ33kLiGEvM*J&vHd`nHF(9uO3~*IESF@!E^FIYTC7oYnHU`QHXnTFOa;Nt
z&Mq@MOyA_&x2&Vd5x)`iUd;MP;lBTR%gRMo!?vN)!tKKNTl<^dTUq;x-`vp#rl(6j
z!L1ZK6v>?y(Zuof=ev$FZ8-a=vH++N66?eBg`Qv5HX#dIX3*Utf)4}M+<JQ@F?{TK
z1n*^&?U!;5YAG{Kp+D<?GpHHtZ7tV~4Cs;6VPIThCp_c<6yA;=NQ?4WU{gfa<P@4}
zRa&?|$g^JpN(Lxr?$FZ5q;#5PMFvv=F7U{0#D*^#b`isMM|EyF_x8Lsai%_s>*3Od
zBE(TfWF%2>+#%T?0`05bDlJHzBsP|WJJ%aH>u%2ajp+r6?rb0sjU~>%T5<4BCcpL^
zV9<kg3K!0DH4YijtUT#{OSJC$VuWN=h+VoadveFc5M7pnfW$Ub+<lmF)#S~(kusPp
z!l@Sia(q9dm@b~~W3R{};D?8YhaBH8p(r9_9)+W}N3C^;@ZsIiNrgZ7tqPE@Mb4HI
zaus-d<6F6|M?B{bWBu@BXE$^6vLW+@<yV<sWCZqoZk7JYIB1@lJIp>%Wj8MO|M!Cj
z|BFvp@OdGC-)F$PGXjgu2Y({}cbJg^eCb4GDI4L30W!?>*)R4x-rKag5xh7L#Macf
zEmeV(6<eqP(>y1%d$^>D<cSi)cgYzILO!HDH$hW?B9Jnb3=tkg^P`v|2-#!;xkGKF
zdf|dox$`lM)x^0bX<K6|T=iPACRtMmAuu(xaEWqb4rYzYQ64Tzo@yIb>Yr1SKqAla
z+Qv}<UgL+y)@)c{r#vE#<IkEI`<@w5j($4eYqxLB=Z^25IXnK@RKQajlIZQzK9v5$
zp)8BJ83Cfa&3XpJ2Jy&0R6y`mqi3i0HA8onn|C(qOL78U>_69cFF&O=jG~B>H25--
z*iUJbUO46|Q*NLFru`_|;5gJ4n=2LYC4Dq|^_&b5EV1$9c5Z_n<NujzB%J)?-IG0y
z&1+5<CF)zB$t4F|5d3gGdO2zG;qX;Qui(!<0)#7njN=q+0!gYaDQo`M{+N)ZtKav0
zb@9BM3}G(%_)_r;{jLn&u$|dGRiGxftfQOnUdu@?QgdIC`OWtr9y2Z5ajA#Vno^Ml
zKkxGmhaP4SOPa>KXDoqwwG|b$36q0!be<}gw_iLTKBJaVh+aDmY>tm_@~!b*<B#D?
zbIDBzzEEW~C!&>HIlGuS`#1fp{^LKXXNhDpBGL;IC=}Ktc(V_2YI4Cs;M)G$kk{so
zt>%VwrMob*dSpvVyVd0+yQw+&=x)otzqe!k!8?N0oQ_8}5O-1=c?T9Hqsd>942jlV
zzxD3GSF!JVh8L{jS9QIQ9Tjkn406Q`p=mLhm)w%oUG+7~1m*M7Rh*_z+ac;Vv8c$L
zJx>%P6|f^j1?0PeU5pc>G!32MX1vu|iy0Sg=hhGWk37y8;ws_$2~sSGXX9!r-}%OB
z`;gr&ne_0^mnJI>uUu>1dVRX~+T7`3Zy7|BI7TwqG%E3QUEF*hAI!fk$6EUeNU%f0
zwf=0{+N!}4E(gV|Efq({%f7CwT}y0OLkuSJQGS7VDOWQLmM$_hNw|C64&fOvP`hVV
z9hrNy9d+tJG=;=kLJx6i(b8OF6j}G`JJQuA#u@D~W@MgGDR<;~Lu}!hC~%t-33mjM
z>H0UzDVl*jYf30gDSwyIt$y>-@5tDjZ7oA^x0Ifw;cX8!7belRY~XMiwqeZsu7qBF
z%)(aE0L1Z{?mM@+-{+n&=*(_kbIdgE_Jwntu5Bfoo(e%MpPmsi)L3Os_Bw&1btqQK
zuLWKOUaY0~BmCL-a?GXQw62$jz8D$!+W0P$E_%7WUI=hjNZ{@G`1kg#onkF20K?YR
zwt*3^BQ01Ey)Ym)7h(8!C|GR*@eWXHjD{I#efI8nkSrR0QUs&@PJTqxJr;`C#|OZ~
zO9jn+QXdA#3!E*t`~amb4EnIgTtc!XzaD9FcE((cHFs7p^yA23j)fEi=5FT=E$rpZ
zwm}u%(bJDMLMeAC6R&5^_k=DpsCl53OW+FvtGF&l1$$vJV=c%F3SV*IQ~Zk!wdc!e
zP3ivLE<L}{;I|4gc`kokCi(bAzxz1<Sp}P$s$Sc3EL*>#a*dRp*2%{_tYA)X-yUGm
znGE_lvq%}PPJjzH8$;#wXp_J>rKf%KitM=um0lTS0(BX)R@WS>eaYuBj~=GcR;zTK
zzjyw~t5v9?0N1k}ji2hKT%Yg;Gn;f_nbfsD&Mfp4IIaA?D|n4PRbwxzHRPU#ES;}Z
z(bPJ6{)Z72z(A?h9aKo%0V%&y02QnoPfgXBxn^bkCW+3?nzu#^cHD0N#K~As8?5W8
zZG$-x*G9FpRF1jc6?dPu+F4<^B2aJ7X~2Ib?YhVLK~Ze%LL3Nbl=Et=fAQ!Opzo4X
z3k1G!Tk}M8=oQ=z9D5v&h!(rNq!q3p(_NY7tLc6#Yl_cb!!WOpj;YlygkEyeMo(Km
zjY0z|dmI=Y*t4#B<>+t`j)o8th$;Ez8!bGz{f5FckH*Coo%6}((<TI9qAVUd+Dm$9
zbwV=nO}>Ro%XQprL(nI;2+dtkdAf$p^NFmn8|w9~c_c01pB*^yJlHaEPkTq{l<uJv
z$sm#e42}nfFXC?{Iv2Uf(c!t4#!I(vI$dDgkw09trMtF&Jv7c(og@)7OSyqx?a`6J
z)W$5<pqeK2-ERimP6bv-8tr=cC1ROl-zJ2y_#b#&Xt6hXQ9gca)>ath+oyK^hrWBR
zCTuQwBLAVS@DD9%my?mhyQd1s@4wb3|GysjU+R%lHgNBsLeMCJB^8i7wIoiWn}M%6
z)L+=70_+!#?gl-e0?Ku2DG?B@fK1K>3~Z?66y??x{`ps-w#Z|)KK}2XpYIH=o_VyP
zxph1BMTk(cy2;Wb)2u<7x8o_lj0ddl;lmHNuYA+{vgS2li$jkZ@O>%n)YAMemu(Y8
z2m>}PbQv?qwwNTNV;WgFq3t9>%!JVJzU)c`(bZWv@!2?5P+*c*?_OtsGRwR>w(O72
zn5%c7(tLOKyu7)*H)|K8HuAapqqE#@8IZhn0zq0DkDi|SoN)W4<+m6)fXG0Acd>|8
zzgLKt*QuH^TKPv(S%L5n?8HFne)eVD``BA8o0PP`!@RLP#i8$!m4po^A`<ViGI^|b
z9;&690`{Q-v|C6c>m%8#lPc2zK2q&acn4wrTE*V=Ie~=k{{zG(ejVPbbRF31c{%!f
zA~NzX()9yxGL33%*`;F*8Z`42EfZfubYn!HBpW@E2{Z+D3>`gOh=3omQ30dU2&Cx$
z1^PutQ$#5ZeBt76Pkq^<`^!HS!{o^Z#G*hvj}Zz74mn`iI^i1=#<H;I0p5w@&0FUf
z!|Wuy%~G)7&H+7aME#pZFwMMFVTRkhQ8U(rg_Q+cQ|q?J7alo>NJlw>r>TIp93)3v
z%Z(YbT62BVD47K-8qcOhgfMNQXUxnyjqSI8s2Q%lNihE@P~5?o6mDtDXE-GIg=ucD
zw3CP^hB{($I)}ooK*T_qNCiKZw`aFY7eep)#M;8vJ>HsaOl$KVUaVfWu7Z8`PxKn4
z0u~G500*#;C0@E<vDuWp8m%6+FW(>V(Ze%hhNsZ$*L?+`)|MC*U?V<#e5fn^049L@
z(QfRv_MKes83+Ha4^%oDfUSBiYO_o`67q2B9bkaVVYUFyJ?51=LR^8Vc<uz*BsYoQ
z4OuxaAebAYcI#f*n;Quxi+5|e$1KHk%M)MkZqt)&i2;N(qA3;>2|llI|0{G7$Z95?
z<AZH>>DG~dJenKC)R|%-l5#U}|JmoAXKET3pn;&SRA5zFH{u+Tq*7l_94jz|ijB3Q
zv&urhn$JOe&z%GBH$r7rR8#0*i!o~o2A35PUJ%bOF3iZZBk<rQ<^oV6O!L+kFBbC6
z%U0UVS(ub-aj~uhdwV5^3C3={CNACcM;UAJUId1dZ1UrRmK-l!PDgorf9=nTVDwL=
z<W9pjvdf!1cI<~ARjv|gRa8Jo24(Oyy7zEiKI2JaxQbrUhX37t<4Y-I!6yO|o+B51
z_>fp|_0%>?LC`eiTqI^Ucwe3bVwhV%&uxG%leOJXKG_s@G3M$_x=K$b`!c++7P~U!
zAn4rjbpuod>Y@UobC_XV-z><_aH|nyE>Mj343j<>IY;zA(j6cBIj1ET?g?DDSW8<X
z<43wr=|wRH$&+Pp3tjLibPCxK?-+GwbPv00(BxM!w(P-}`p%r6?lBuI+=Q+Ee7mEQ
ztSx&%bC>UoIuJtz#QQsLZIYgnRqzANnP4_D*HYbgt0hC6&ri98qAz`}tcm@}bLIh$
zI+*Hp*>5i<#Sn<6&Qt`fvNb_==CHO5uK8+0q~pJyA0hsDzxO~xh%PQJbm&Sm7C{tR
z-6COUSgLezJ?f)#x@$Yth5_fXZa?|18oYb6%8zalz^iwG*vYryQ9vedao(0%OfyTD
zKQLm8BDC<$Vs`DS%#W3I`y3uS$m<Zr=^vjMj&z3+bH%sjbD|*ZN$WY>O0w&4WMU^U
zBD!!r1xnwV&5EZ};!o=d5{K@8+I_0kYWAx~n0chv)UUqv2V>&<bBP4gldmut(skmQ
zATH>xRyvr)S;p_ODNF^I4??Fq<gr&NgqWwCTaB^ian|R27h2ABOxp^?=_LlIGw1Mb
z4~5^r6AGZQcPQ8Ks{Jlr^T|yb+R2#-uVd)7zTi~|XNY{qs;g0w5LR5I#n{V$k1$hA
zM)%l2#Dq{l8HV+7T+O40^o;cb>XNPy=D|bWKq}y@%t=-*GJb!TXBHYlu)oD5uQlRc
zQ2!CX7k=-n!H}9L_!`UunMSC__rxjP{NXJgK3bOOoWEwpiFYV)cMU}F1dKO`e^s{N
zoTz!D6`#`nC%kWr$cxRgC3NE>LzMW)s(8hra0OXE%x1h1X5WDuHZPMdI_)}s^pQ@@
zHu0xZD`U`W0FPE6Dv#0zubPi6x*K2H?#`^qz{&K+&8=WP@gO-B-XmqT1%$UsC2zf%
z3WYBgf-bu5_lvqm5#*-Y10gTFC)3M5Np)qEFP-qM(0M0=)Need{ZoO36NPiy)hio9
zF=VsdE=yO|iXta&olZ|-TKU3krxE@*I9vR$Yl~AKf+j(<O&p}FP%B_G>b%r>7*A<m
zvAMbSBD_`s?H;Xfnj{!(t}%dYzwg=?{H-Wh<0ci*1&SJ4I^lT{Kqm+9*ycX(cH#D3
zLg9$8n>x?Sv;KaYQ#%uKp2byD6%SIP_ZB+9aVf-r4zt~GA0`&d<`MphohEH8uY8-C
z?cB5z7OS&PS9ya*6xzN@tFW`##w)XRh~)(uo=*2vD9CJ7>}A7<6ho}3&MPG&^^t6;
z)<uNErAPeN5=6I$h(atKBrpS|&x&oFkIvP+>!l_9Mcr*$i++!-7q#V4YM=dC{E4=U
zR6s|gco-Fcdl3u2;OtbsZp6pm3=#Dvl|Q~)J(jP%mHRAj9sCtXH(QWC7N8&q{~>fs
zMA{?mFp0zheh9T)YTn5rJtI5dsDRie1KR>3Plplr(@B1PNPvl}71U4TX{WG!MpZ@6
zE8~#^JSZZFg{X~9BeMgSq^4rMwLzEK@}2#)XG}Jp-yzJ~rS>XxN53_)W`A3E%}wzL
z@e%WbQ*Ix>%ia8Iq<K_VlVqlaZ~i_v>pmjWG6Ylc4Dp=DlSc@^2$<;?;*@6M3ml=U
zXcKK0_RVTNP1EZl>o~pg<DSV%DR^wq-**pDdBakAT*LsKDSMIPf_ay6Rqy4K;hT<(
zUNbRA7u)XwTlL~l$by`ZJD8FEW}yzF^zDQX5KeO^vGDf^ru;yWkl95>Z_ZYlro%<Y
zV5yu-WZJQ=sjXsX{8c?xbo2H?oICk>;l*ub&Di^ynqm%N@w1E}`CfnQa&jXXY&eu|
zo-zn8QknUW_ktv}OJU2XXJTDedaRosMD$a`wvrmMac45sZwy?G;`K$%qPX0V0XG&G
zwsQEKmCpWwsXFtHOc=49l%APu=>|%|OH<+-mE7NB8{9d_$~Z*FrjwY00oD@OVf3d$
zqhMkeTW!iyXTT58D@TM$o$&Mo@tkDC9pZVg0!;K#-n-Fg%Wu2QG27|Wb{V}tE96CO
z7_92s<E9E=w{W1y7J14<4zCXi=tQs*%IV<nO=nIj+`M%6=T+$z^D4O~i)D5;jlL57
z)_V{+vTVec2p`Clet|mL>czNu*0=7=H}(@U0=u`4jH0v}ePK0jKZY7tPi`L6GytDk
z!IX!QiZbT0%J{2kUtK26oLlahT6QT^sNqv%j2>w+{Q8l6Z?4KpmLx_QZE!%56p8fa
z#Ehlw3u9gBvj&aQ#cv=scSp>NZS|9wf^pxJ47WF}Y;gvd!B8B6kbjad$->y4>i+;5
z^kapGGeLj2pRE`FnQ~?Jr*Z<on*PcM5?|2I97eJ!=`z~ocr{G2+17k=OZ<1;gt%*}
z`PZ0!N$2I>)@pM@4jzWp?7~J7yLzRIXH1PUXRqM~A3RG`{5D!Fjr6~pq1jl6crm`A
zTP<A5yWCV0idw48Dc)T``fcB!-#Ws8Smx04PAD|G8v(Ju#$C{>Q}caM&!f0r`;Lp6
z`A)Hq&@f3l9I>8#d-Sf4CC~HfiT%cc1@{0Oxgom#OnY8UBQE-PkUs&&5iA_=HYK>@
ziMlDWc6=9uSB=O{lX><%`EFfPHt6<!x-DUDVXLs)Ls}<-qx%`WEP5dwI#sn}Th^BE
z(CE$zRqkVvuK<<;sD;XV%!AA;BA0X~<IQmI8?>kURGl044(vusMJCv5bXk4>xPgBk
zy-8imjUm!D{ma=ozY#HDZdawdki($!!!UK{T&V7^Bmc>|DQEVy>>9$fLh-Uh0?XL$
ziFg^jm*W`~fOrGiX9M%WSc%RU9WRW-+rY2T0KKNBbfE`&k{1S-H5Tno0N+=|Uie_X
zlALEMHn=&W6<9H)x2B$sNi|;{?4G7kW%L`dQ9n88Q`A&QehulS+#sto5ZmBN;a~Iw
z5uu+0(;DI&Aufz^JzSL4;@$mL#iBS08&SHy?RdMBk8A8kw&?lggeNu6ezZ$p4^8+y
zl>t4Y3>8(q007#n1j{E4Sf?XLwzM9sJWHKd>-vw^cZ%#UGPc;CtJ40D#HIg`L+}1w
zGF2DVx5ZO9ejZoG+%8M$xL{YA&#I6t*TYTobH{_7IX}BS$4ZZ6_yb27YI)!^mlR*^
zqLq1#D~~RaK_E!-%qr!-mPwW?|BLNB{G0^A5BGem)WpoHaBU(=`zW}NOW|Az7&kb7
zkqXFxPyuAlfm4(+Zq^E0zk3;lAkz+JnGEa%FkM+%`1eN7sQ||9Mbr@ibt(<w7RZL~
z{RWk!Q>w+~rGh#*%<fQy9+D+$L#TjN^xx?Ia}sd;3wVb-RGbosI7Jj&87kn4`d!`r
zT`C~>+VJj4;A;8Qf3>%$M|5;QO9c!esQ?P20~LTR>1$y}ce!I6>>ly%6waTR;+ARj
zg3j6t%71MZv#@aWb|?NXG+6aoZndVMownaZIQ#Rb=P7*fzHYJv6;PK$NkRq|gRhVk
zsDPAsq9YaH2|D3-A?Y_M=;6T!!~4Q8hEtbck-C6$N29v3Z<miY$zwnlEYxiSv2TiM
zlJS{&`JSI046T_*Tay~OIG4)Qo2G^qs=pW0cY?zzH8!s~zX|%e<-yLtA&@?ruAd&e
z(U($fu0ZOziegjX_+suXrx!`S;b`-!&X9v``M8R2`9NSQ6{bz(K#-mAcY=f>7F2M%
z@veC$V_kgAQ*L*B*dGr|M!%Z_oE?7$00fIdFhcFx3p2mK(8HdFvSJ&Y%}xOlw;Ec#
zS$y|gDZB1S^W0M5l*rHTSI%&@E!&E3@Clr<250h>^tJ+=)rYTSDZIjUWLi{{%r}dI
zuEql{<D?H$H(ux41wXPMgwzzc_$_N9Pn{jG8ZS#WmuS63q$R1(qccPi3E83$8tD(A
z7JH68dI$Z?TZ`F2*&n80H;Cd`z2FKLnRNO{g?m2c_(uJygFvm7mE+xOE1H?BT!44?
z>bbfh<{0Ddct|_9UuWV4m_>aUv=)>8=|E9`@@|ETzt<D{kUgnZo-Fo|(7ZFlHFWTE
zAbV0GdZB{=W~0bkfallJ(WQrx6h8EPziW}nw9EiqEuXoljzrm1-*vz*{;ZPx=J4Ca
z?T2>a&eBU_kxWa#>7h>4B?Yk85|~M+h{)D8RwIVorP(~Zsy|yU+xdf*Kl9+n@5U5E
zcerCBn!}=obhQd!+hyF~CXH;aX>m1ErEK<HyYcZFAn?rr!WW-IW`kqqJ|s;jUFwJ9
zB72NV8v0E}5JkycoUi%oL)7h4Z~Dj41T4dfhE7?%u}*H#kd*}9F`>Gv5})=~tY}1P
z?$$YU<;MNW9?rNwVIQ&;?)CCya9#X4`a)FZ=l32N{dA-8W*h?36HNg?9Rg+ATdv?7
z;vMgVK>GW3-S`5ES$I}H^*PE4aXsXENGr7}gQ{GJ*=bVo!zM<i=ffGLc#arrUlZ9n
z*=Ee<WH;=>BI8cjZx2ar0U1FyJm8pP)uK2mz-fIK=8iqlJUa??E?cP?t-j)O<cf_D
zvK;*8G1$4Bu*?Agv%;-;g<9oYhhpH4u_dTTrH7SzN{V8l(}knRtj{@BvmK)4Jd`yL
zwNbCQJqe#;nIB`nDNV7@(kG9zMN5JPK)f&;hH5;#YvD36QcFR*P$-ITa314zD9HNc
zBBDm!4T?}5Y2k%f(GVRdXOTFWjv1D7PfOVwTz`&9%ABpgrE=qzy8PqW&e|Wci!W0P
zNM=f3NK$J`GB`paA(01~uqV#E9>kZX0C}x3!pZXmosclQ^!h8;*|%kfZ9tqZm~PYP
z{wA2m4bgI!NZ6)Jt8ED1$D5kfFLyc9J&j31JWfi93;u+b{<=0?D>&oeLHM8~7WsDV
z^nrGP7=+y;P?wPKGftzIN1V{=(YNZxUiQiEVLt(R{dW}U3zT*VPl+i@yvp-SSoK&b
zV9eEWx!wl#s3uj<ldQUH5z=@O+Fo8=X2A8B17IAuQUFyY9(lkL4XXOxMTn7n;i_VS
zyVA8aCFg)}39fFJde^A>@L89s=03nNec|%y2r24yW-}0-i^w7-4)~%j4vlJ|W-e_8
zScPEMcu8my4e+D&pOnzJE1EW|95OyO0~4{&&1S#g9WiPpc+vI-a3t9fBTh48>in7M
zmUh7O*2);cwXG)U!bd#v;WdwXdYZa95CL}br8+>WYNX{Mv~WRtCM;qJEb}yaTHiRV
z9N?k&a3Aqm;XV~WH%U~ub9J?cV3ybgysAqJhDLUpedtSdmaFL<Rn-_1*nW1N7Eom?
zwSwZLoSP5oxavvPB$~p%S-`|9v!(E$;N6|&#)!#@*&B@&&KKn#Ts`|PrS7#Z2QAQ9
zEx&nxd;SzJE-q&pHw>jJNs{kQQ22-mZBiMHQ~)nbe?ka3R&A*Qo)8=_OyAf5i{0)}
zx^l&1iw0pWF%SUXW+U-KLkTuSK3HHL`7vH5F26kdF4?xgW$ezRUrowOVf*Sx+TQyd
zXN?73UCR_^0(Nifcf)542=H;V)9R?0RI~h#kC$xy-fP|MJ?LkRA*mC38>*^LQd;gQ
z;-8<#6&7&stglE_eFrD(eLVc`LfOfCw&u(dBwW>xTFh9uDp?vpcsA-CaH2dvW{kU@
zqcNG6ZJj%P)6eO*U`14%AcI%<k%{APDLh9TFxnS&_3zeKgL}N|c}pQv{Mn`TKN^h+
zn|2{iH(<csI}l=e5zdkD7+<vg33t5E-EhrcC|q)zU%O^FFHn73p8VVsa^X(Pn+R1J
zmix-|9Wp=Qjc-5TXuD{;z%_^M(?At4qt!2~j)M8Ms4qs{*g3lhjbL7Y-9z5DTq~$&
zxFhlJEBEzackPM%DeQA!C;~y##;HV^ML78;n6KqJS&M=ufysiYHPQ>yIrh^8aFu$M
zUq;jJ+!T*7-@(QU;}!4wKIOwD7xXKO`AnUViIVoNzLUaprt9cc@7%)gR6q%z4rU?P
z_gJR;RB*L`8YbBEk@(ac$D;l`O-H1ShV9<W19v5B&ur;3#m%30;I{5uL~)H4+~3j^
z31(AaL41}9;oj#Zqr~0rEWJw+@I6du!y42r0g^J8D+ErSs|K(`V;|(s#6Ow%TzXqs
zYkcYs@J{Xowpzbd`go()J9qAQp5j+&rj#yN@rQyW0i8g&GRjq=M2A^sI&;b*nAMc{
zmKy@E?uQLmHI`)cA@iNNVuK%T)wS{&<K>pxJt$)>Vi_GkjzZ#j{AGDjje}_9rl_X(
zKoBk8%T)GkLt|)9ule=3OZ4+X+gdA4wz#1-D&SIA9o{%x62x{9cEoQo_N8&}&y@Se
zB4D_`tlwqhA~XG-L4#NW^T(UA!kRxM;_{wLBlR~O(w$jF?l1q2sp((Nq5>=rl@w26
z$=vC?-QYy@$|r2eF87*ToRPmiu^Vrgrr)&pake_(k$%)qns4MUS^$kUL_rHZjFk!q
zm%srviS8%JlM2E`^H}$7P~{O1d0HTF-iXkx@YB|FP~M@LO>M-R3SeGOJ@)-z<67uU
z*sSEyerUv*2Ub-DoY}wPTRupPWlf*Th-}m$DJss)zn8r^qX}`C(Y!skoqIvD`F;ID
zmoJ6nw_sng5LEPYFQ_X?ha^G_Sj6Ms<L1|7AYV<0$@8tEg6*@K4`hRTU(<gy<~izq
zEoa<F28~GVGDGE`j=}V>Afh&M##_{<G%);5%lJ$*)AARQk93T*Gm4vEa+{jmyyAZf
z|Hu|`(Ht~X;#|mvv+0QDAA0Jj5quj|mE&&o|EMvt%0Ub@{*6q77PkFQK@Dely(o_w
zB2NPP{#F(^iX{Y|l5=?s;`p}^CEowo&w>hgC-#L3u-b#K-D-Jxk^^5(AVMhhnp8lZ
zRe32cyUdboyiQTqSsFTGZBnHItT%H`*e5VSG*m#6`000q2^H`mpePZvtcK?}2`z)~
z8yJTZFH)W~b5NRXDTZM|Y!k&~)iw9$s_fnWbC}-0M|e$L*FEbCu6*%eg9^9|b4HkT
z?yW7@g><?5Vr6d`^^4?STQ?J*Z-{CIWCuwh9Ni^0(r<sIjBw;EzlmOTUHHAjQ?ts|
zQy4I%iM?!`Oqrkp_S}O`*XlA>_!@C@vZN30LIoVjpg5_3DF)O&%==7Y0lZgih*DQX
z2@#nv97Qd2oKjy411jM06_=LgSJ20*><0ftRw2?T98^H>SJc5D5GBs#IHUzk1XBSw
zb5AezlL|P#Uy_RG0jeMKs9@myrMWS}?0*f^!~LP?%3yL%I;nvD_>yfcA_B8Y&UEhp
zk)o-9bsl5%2MiV-bK1kmFWQTbRj+Vn^UFUHP<imxiw{#LEI^z1G<2w<%vN9PD2LEJ
zQiSJ2Eb`t$>)l#8@qw7+4@%3UUFpB9X>38(V{kyU3PsDf9dvxsLIsRmG#j1F4}8u>
zH2rW2R$F+>#3(AjE|K<EYwnCbZeQQYtTdFS5^eREU$p0DwqcYAkqUU_y$;r#YDZmG
zV&4E=B-?)@8xfuH%pDkC1dM&N2EC{@jC?6kqaEe|7Sq++WnV=o|J38k3mOJ-or1IG
zP*`G*UT8LF%x_qI0nz8qp<L72?1N6R90t#BV)yU6*>$C>;l*5gc@HP!^1S#7iKA{x
z56QMCMLSIbg@hS3oMCN*9mK2NU5azMEI<C<y}`FnpS)JN^l%_>gd_>wRU@Y0=v%od
zp`N_mIi9)bLVnpjJB<$l(al`ErWyxke^|oc=M`jedaZ<@tGKQb^Iwk!?xZ}OS#9`K
z|E2(r^?lg4^*0Wj?>=^j2C`5|K$H-BFyJm07P7*K)SH1*e+eVIJsx&!P^f-avxt-M
zOafZb18(z5c_hDBa-CecSC*M|oby>AoGuw~Eg7<>j&0Oj+GPnESzzD4_(n?ne9zL{
zgJZg28u8NoF>Xs8po2;>so_#&9l~G?2FGg?8e-fpLKdzC@i<%Yy<;Eu&I$XYDR@GD
zD40}GXDguPe>phh485$>eoh$q4CZv$ptpM3<&@IgA($C?_lhoSW0=DO=Ec@k->r(5
z(|1f#GSf4quj)4y<d$ztYW4B$Vi2tBE%$JQj54RGqz!J}0(aiIf^i|?JGx&!M2@o>
zxEwd9zAdnF@&7PxB<GVU?&;;@f#HvkI{Rrl5db(ZghnYMb!57M75-W|OfBM2PzU8g
z<A*{d4%~TUGL4qMhMIO%XAA2S>c2*ttSON0magDMa+-++0?BqbtGvZJyUQ+omxzWE
zQHYQvq+3za&vu1iEnE6iV9576F)9KH#s^l*zDX3o*^d&vP7#8M^1$W3CS);M>Aoc~
z*<DfVqNUuPQHW^qnYFL2*Rnm9cp(SvvX`AEHKr{&&lwrKd%X4-k$p-*p#he-r<6<0
zqjw!}<CAb}R~c)QFdpBqXG%K&PJTLT))3K^W|c1RWw>oD1kOk_UIKSCCSJI?J-W{l
z^G74i3?_@YnLZa*nY+&u!EVhQ(%{Kysw%8*$yjWpyPo3%dx@V<!u4Q`V-*Y$I*f9<
z9}0&&LD?RNup73EA=(u$0I#(HvQkGVlB{zYkQQc00&XgaF*z4nIj5IJjm?#f*!l%F
z1iWxN*X@B-f;V^#r|;)55kZ|`Wb^zdEGQ<2T|r_ghtb*ZqL2IH0-%D@NqRImRqtyn
z@N4z_t^1+B%9YmfAPhV*=pK~+8LlXLOLZp+`uU1MjfTbRWVMc+hU~{==Ev6?9%^c3
zW>&Yf7h4L8dMmw0(3~_WxE2uk3EITe<J4)piR$QmZJw7h3Wt6g!<}E7O+%H)%pPF`
z#78Oj_#94@gKuK{J(xF92Mdw6zWfWOn+lbJK!bVoRbCyXgdQ$0F9L1G$snTbH<$-r
zHxjRcg?G9NVZ*!MZ{G?pAUa*gQ9g-tCEq(Qq4lr>CEBC^P5sV2q@a?4vr$aq{bdwe
z3-jzIAKx%p2>y+Ur&9@tb4$6~Qofs50X93^+0_W>ZLQ5?vjBCm%!GF3)VK$`1+i({
zeu27gNpA;Yd6kC@|3sh&&+INu7mgqWpc8|p<@6T9OlwpCa_5uJ5<J0*ndk%Y6#O3a
zuxZTOd_DJNS&f=w@{;qlRTo>{w(C<P3$=-@uC&iil994apnUW?cUniu{ZgW@xrTDa
z737)acn_teCkYq!1%zcqgr__W@4AL$&ARze0p(Y-><cK}sPmZenPus?l$r@q(lqaK
zTH^8A&`?nG=3?}nyOIk;(O2cuYDvoTcQrrRuGu9|Td5p~{nV?MT7y@rZ-;>#^g>eS
z8Ss9gWWCZ~`RWEZ*PU=XceZC$4JTyZvQ~0(UT=R&H>dA2^HoHIw3Tr6NrMs}BuL!W
z9gDh5OjxiOcg=;BR$0te?48*+A(Y9#mc4Qej!a6Ms;H<Ty7c4JmBh4DmA;-d<i?2Z
zg*M&(`aAnJw^Z_zxZasd5@+hmb?J*+h4{MjKhu1=Q%t8PkWkSbTqU1*mbO@^ylCnN
zs2Y(-*2XUI>ftukN2Np&EG7aHn~_@f92}6FJ0_uQ;mY?|yoPK-GHP<&dh%x}Qy|9A
z^{E1rVr7#u!mDQMV|_fVZRoap>UG#WDj=eT2sy=1CT%L<tTB+R>X8oWVgh#@Th?S0
zk_D#jY`MNuXJl1){9JwM_^>1g*8n?h`J}NEc(^By{~lPfP5cy~g80wt0C7_bOz^cE
z#qh;%e@~GrQ=xF^MZ?$9prI6sDn%0BzK$0^xx9M`zR-*6PD2KwE)gP6A(|-=H$@5I
zPNQ5X0}*3D7(@#<6)>!#PX)9Y_D}(=7LjDJoKtZ*B8I0wA;<o$(;Gg91s@B}U#9|4
zN$_~S2?ALK%?l?>9d>|^8C<|bkvb{>d;34T_!{`H*5v(b4+D&6@GMTJe)Vsb<Waau
z7{m#klsNf{1ss2+_g_t^O9aBK@u0X}ew=+_e86JyPU*nJ_QY(ABS~!gW30RKKvMQE
z$qDrtJWzI@3g~c;wc;cqi{|Eu4oAB-)$dh6pJ+RKJM9?naV?EDUO37l+!?#OZ_;wz
z1bri6e&I!or2cD<n+J%3O}pLVvfuwt*kZ;1YiyD2?c;}6fNx<Ji^%~~f@9U|_T@O&
z%nT{TmrC4=hqb|n0zPiRz|RgTDV6t2j3o00uH(et`hXJsmm7Su?$q!xRi!u-eOK?W
z;3MGpmJ>>cY4@23juB*;La2&L)yRIp!!5}fjZTl5&U@@|{r_GDAYDNQ(=5@>&6}N=
zH{D}Xh*sl|n7<L6cbxqC@0qg-lUY6jx!)6WiuDcc2l%_;OE=&fpWwgMT5L{A1c@0e
z31kt<&A=N{m{(N5^frdp8_07sNCjm0`nzs}KYuu1SqrJ~wvL(&+w1H)hpQi)oAb<`
zP*d;E0f?cTJ_Y@h>U9MxFLL13^f%o=k&e3;1A18|o8<|mLu84mxCa1`W9;>7R6sv;
zb8_?l1Jt$z^D|axTM#!dWj(6akkSyE{LS?mP2id9iR~>D2o^T7@rk$M1u%_5&>@9m
zcSeW=B#Xi;-^4E;Q2}T^ho9ueif+$1Y>VzmVG%LI9iR5Z)pW}i??`x|JWRYdn>hDo
z%oQjzgz_aX#kn3LM!3UF&11V8FFg@Uzt5UuJ^$>o^3G;lvgZ!u<M>g>BGc41?e*GI
zO80HPty=d1`fz90mkOx*9Y1|axAUa(NV@Zni=ZTo;oQ%Xv(+4CVU<5tFGSUBXezbR
zf|ZR2;7CH4!gItaHH^!CHV#L=BDZSHteAxjsw+4mA=<>GgC}|z$M-=4KpM#f^l#&(
zw5~>RqK0HFDY_?=%nSdse(Vz>uU0j{GMDOnp+dfmv#sq&IglQIkswLDix=vs{0l@`
zG_6;ZR0;7mSK{YN2++c4!}lR)dPF=PFktK4i9*JS+X|0KGHWf&^)t5ET^bG?p*~+}
zB7M>%-$zXZ6h2JD!zu8vPedp1k;XI4>u36I<?6pqc|Gea@I7&e3IM>KPof-2;8?3O
z#FTG3Soxs{-_RH3lLb7ElainK9fK<bR_sN8-#-wMy<V|J1vr4F`A;p&T}N>SJ#?<2
z$l!OQ7Il=zo=zAHj#Z1wo+IA<Tmvgy{9X`CqXM-2IWLdbev(G0Abuig5F^`uyX^<S
z8C4gv_$ar2nq@5u%>HDp3c0G2w6mr{^FRXZ6x0X0D(N1VQ?(j$FmTiEK}V5F(Z<y`
zovFs}x+?}~5r!hIOV{nSYVNf+l<4LcKi0P8R1<k!5sr5!0EwEX_quORG}gm8M!9yM
z$aH3NAa#_EyKR2RRX_F%hKKLTK5T0PN;re7r=lp_<TJ#GmIrU8+@o_?d~V@>_Qv>b
z)IR-@YJ5gr-u|-arF*(P*BOIBDu~_f>JpxenVG24JKK5Ag>|3W1vEaVI@q=XXf%~R
z9H}hb71t;B;Z~QXy5UT}`aNyF61)2?NAB`6xGMK(-Hx!93*Hd4mRnJH87H}V^+^8y
zWSp7B+)z1%CaA-T4cC&!65S%~OvqtV5^+h2J_vN|yf3)yyLYdzwkEnQ;o(x>GsD+!
z*78NBiJPPgM2;SR+s*7Vqgxa%^3$<rLwV4|4v*3VA;Tkm?>M^DTdhX3`P;qpT=a8F
zJTTSapu14ECwO>-yAc(jGO?Yyg0_iUJTL-el@*Ef{|b>eMDP%&o)En-yyl%0{_&fm
z)rul>J7qho4^2kYm6o#s&-(-b=dT+J&h5C`Lx5-IC}R|1y!zGxroLMH_NyQ<Xqtr{
zZI81EWuxEtLA8NaN7mPv`0InpI~L2dIR#40ln(gCAWo+*NH~iLlnUsYvi&*HWPpu*
z(bCwGxwlK2!%7XE;V9?2^STFvVEMI?%|O({9WJ#L6bvzi0ojD`gLy^UKU33n_9LWI
zoJd{IIv1Jm9X$k0ZLGwWnm`jZDkr=fdwrOs3G8(>jWr(l0Um(iL?qp%$LAXG5}BEA
zNc`N-V^HE=%e!a;g6v2&(<!`TBik&cr&>X53Lx4A^{cTAwUv`EAGermXL!>{&wnXz
zrL8N|^IU0pKwKy#d*C8ZmB<CMRFfP_REuay<AxKgXXLw0oZ@~A^aKm*EsA^Rp<<m_
z4st@o$qBXQga+NlCp3DoUz#@Utr^xZ_L=Lyaj7+CRzlv|mOd_}RqBX@t9_J1c-D0C
z0cL~UmkWxvcMQaw1O+5}7HEInNk8{4*;l0ZXTt2+i*ZXdF%p^11M219)Wn|G6IMI)
zel;!*wePa17n%_1)h2~NQiCN3)1I$jD~&hKg6W1YH3c#mZw~>OMcY9x;?vj;g2UJj
zMGnR}mWLbYj?#B$gBs|9#dRjVPZqV>0m`a(T(lz|0WQq4P3=5f*32{IxA(T@GSz)d
z|K(hZ&q)jAmI;)F;DmRJ)0%@Ob{oAWIzAaSSji@=tizJ|tw=`7o%@`;z4WOMJe_;I
zd{c}Zd%@Ly@mrE0G|S@YfD60_MO4Rb^XoS?*o4QlW9Fqlp3r@RbCWm-2gF#LMc_0&
zR%dxHucabi`^&W2_lp)f(899}5`QiUI4%*NC%PPgu*~6@p$LBvfg?VfCz07Ho2GJ1
z>gv18f*P$+;X9^LXFLb7_XkDWq*yL&qOZnc8WGX{ES&)?=N8O$Tye)>j!o8uU#6Ct
zw1Jt54Z*I<-E=RXE{cDiZT|3OKz|#(!v5O{-P1tn0aXN|P63Kd7hjtQ)5gl*N}nOg
zv|l~eA(`F!;QUHyh<8{OBw2=+;GJbne4d!o!V8AJr_7;P_CcIlwN|%qL(!$xJe%W|
zCBBn~PfH&1C4}=V$nJuRz~xu-?7sz_DutEuo_7ro9?h84?%Mh#Siv`QQM?Vlh(U4g
z&2o#sl+$Wc&l@_0<x6wuJQM$8viyhAm(ax0pcKr{%qoSa;`nFZQF{iAW?HO4PAA!j
zn-d@mj0$j``+J06djn<YPtMkY3kja_+f80(d&&2Hre`L$**Av#c~?pJq-+7(1ZVb6
z#ec%ZtoMKKv{t$%^K#kZjfCq^M)zFfu~$~h0hj=9I?1ym$^5cSb*0<xfYz~Io$t~P
zBF1@7CxKsCn6~*2SL4(rlqs*&o-o?NSnIup7n71UTwDMpg&7?-XU4CdZT_Wzd1cbm
z*fQFBEfBIg13r??i3S}k6~fo@z-q>WzNkGf&}H~$&(l*{t#(QZCFF8w<u`8UHjM@U
zO8`Zyk{}h(*M2(vURF{8>*iCasR1|-lGr4VN27fs%<EThv%ZAMnl&@+==5zu_h-L{
zrOqWaN%>NaB|*I)JP9vEX~$CmYu}C^G({5Jh~N^k0&b-^5KA8~)$G3pwlG!sVFs%+
zIdojHmdpOM=*XqiY?VC1+yCVG6lnS4SNK#@&hdG~N0j5j5&vIdP^J%Gn?e-%3?0#I
z01tMBC?Sv)@##Z~#%=^^^L@^-)F$`B`Az$UuArP{?|Fq&PRr8?CNS5Z8meQ(u_%6Y
zc^`ayka%R=0R2OG;?4^HCAUfis7gV>!Kh{J5%}gL=(mYFM$s&%(<<VC{rS^>Vy+)Q
zJSDi#i5#sBIX%s<E#-An%~{S(_f;88@?*+a0-!9f)s5E99G`QRmFMS1+G;6`vd|DY
z42Xwi+C|Qz{Wgm;veGX8oiM4kUH3CP8XI!}(4z$hsC$RMW0<78B2c|5p+l!SPMxVq
z<&64)z;f~%f#f%N9YH)}3OBw?uN^C9`f|JGzkKM5<-7nY77|G56G$0wp3>J6={-9B
zAZTPUdf|5N*WQ1|?BAKD&!*ur3JE7?EX2V#!@$Q%a)6$HWt-pD!Fw00PM@vtsicTO
zf~{&4uAkD#;D0i!`l*16RKW5J@bOU&6`&>+_!uls7C(Ki&*7UZ@Dt97e&eOYyc4D`
zI8@T)h$?%`g}w<X?V9md`#P5gknXG-OqHYEpo!Vj6`Wa9h?nKd_5b^y`PX;fqG(-}
z*qz6~;HM3>vB<OZ&n7s%cYmwqG;%JDJ*3+Qcp4+NPJZpU>*h!!tcg50=`oC2mej$1
ze+jhl*)t!e=aIUU7pA0&1VxEpy+`!LBdTwh7L5GPvOm|LX`&~*J5@{{aETO@J8EBf
z3S~TqLMoNGe<6-xA_{!F?%r7$FS}*^Kv}Td;o&i@*(gQepKV$Us=6mE2uS2HO~jrG
zhRbf0i_qyM(gR|oM#f=!$ZgI9$TJr7heuY@tG_KW&m>;0A`U_D4R{4XJR1Quxm(S%
z@NT^v`ds0*6vNP2t4-^3(u1R6aiP0lZOY#Y=haj|47XO$Fcom=&W$MrcWk{t-3Yej
zxuh9$W@Y6}$E%|7mnBWT=XP@{#2YsdK=MNzs!NZ_5(@7+W=f2zhI|keEFj-hUABFl
z@}qp~hYD@)^(j>|T$Zjqo&AN^aj%x4yKkGWd?w4|3bQb~jLwtCMv?Y>{Sb4Tz1u<p
zt0k8;1)~N2?iwCd5rRiy=W!C<#(dJlLH8}mjQFzZCuz@}J(-l54Bt`#=F*=c{+KJc
z0~k{>Qw3@t;|q4T-g@CnFfAo-lp6DhY?|Q9P%Z7%yOIsw-`eGk-SXA%nS{ct{Q-c(
zg<p*X_pzoci;XPNN_RrUFOkjgFxHLZGL(Oarfbu0=G2-aV10s=b^fgf>W9i_w-C`q
zBG0Sc;jQgi!otnIO6s_wxP2fRVv58ey2R0bM)=YcjkuOeX^+7dRK8wn&Fx1TZv&Fv
zy{*c$<wVqZ$8&c)8g-RymL9ywBt0q^GCw+hFG-a<PUKqxfHuz96Q8!|9@3&jl-unH
zk{aGdLWJK%71_~XM{ti)&)@$z>lmUGc1Z(7$C98VP87n)W79%*-GN|Y<l<cy&;@7i
z_-@9uKJOn5ft{Zg>*5bDO%J_)0dEzM>Yynwj)Anx=t}dty!Q30=jbxT>20B~C(7W+
z9Nyiai<|yUDCb68;itmF##1c-niR`4Y)Z8utJA&>27FcGwuFKR1$g@ph)zV}Ar-)Z
zWRf$iufe%eQWX^K?5$K3qYOzZw`B)4+GOc2i*CcG^9U9=)YdxYwoT1jsYEU-@tiVf
zA?{6WL--hDZ?_Ro*c%}iZ`v}G3K@{o&~!#SusSldXw0WkvTEwH%B>`<w&8SOp#`&T
zvaA}9)2qv$I&8;jWEGoakLo!Nh(YHlgBDFq-G)19j>XO?>r=kvv+IYKkWDp}-Q2=A
z?a{qIpdyd2CwxRil5MeIrl9`pn@zaY_EX(xY((@&Nc0JXERLjgNl(7gbkxWdS#Vfh
zE|zb|el6-mhT<0zPQf#WfvAK&klc+`x`}^Id=pT7Iwi!aRQUx57F&65KI{Ct*sJyS
zE%*(L`6Q=_m83van(3f0jeM>!d`u{P{7!0){PCzD+rs`=_M1g-0N|_-wxnHCkmulD
zo|e`QLFopOt%%4&2*YO~Vqnklg@;RtvEIT;BaVd|sSDG-=vTo@gNaYu!#y4y;U3-4
zd&U%~gGHTRqsZcZQ$i8z!!Bq=vV&-b{D;|OF|VPz>zCB82dgoB10fH$I`YXDIG@#6
z1rVJ-OBbBE9i8wNYtGI(_hv-cKE#Og&rOdDS72|O-pPgjQ7Xa9zc~5cbRQ?t$}P3k
z0;B@MGw2Sp1)3^fR+^ib2_4(r&z^tvHqT&)md+;L2*|1{xt!_h-f8HpIRL3|+gP~j
zO$7w|hqbU0AtW=1zf#gmEv7T%do6b?W$F1W{LV5YNRqzDTz*L3>_4<MgGr1;vHdM4
zd{LMpJO^{bPvi%gxinNoC8WkN+`M>Ad|Jis0K97*R}(Zxgy5>$!SqdcA()|aPFJn?
zMz+UAW~QHTyrJcpYuQf-XuRB0FZD-ny8|aO)M3ThLi~d}?q;IJ8izjHNHe~waYM(U
zjX}of(8fc6&%}W-$fg{iPUks>U>INVP}?gK;_>2RA?Z0igD#y4Qq{%H15?RBh}jst
z`nu=)x)6?3qwq){*U!L$z_%nHqS#OmxKgM~Tf18SmA7cT)p>@ayz~*K9}hA;yskcw
zzbrr1iK@<tadRVkE{ezTaaX4w!6eQlEuC>><#X*y*c(%0Y0|eN2_yVZYN!B+JI&@o
z<-ktTqC5L~_RXo58)zm`?dr+~qa^E(XFTuxcrbGFi5*iD<9)zVrem~kv9MyilbXTv
zFvFSAIXBoLP893dtv9`dq07OAm=WDgaE{Sixl>yOQ)One#Mdh;>~C5O=?5ac>}{>Q
z4&>{>zYYnW^{1*^ndlBSMe-_(vuv<J(rcQO5oov3#o43eox!J-OmEN3)zS+=*hZkV
zZG1OR@(-HXTKU*m95K8!lNaCY{#NFXxjlY8{gGV71!eUdt58p1A?iZQ)in@3jKls^
zt(kv(wcWigm~zfMT5mp8WiD|U0I&v#Mv4#O$2}3O@5j9tQ(`})tTt39Hw5iwpL{wl
zl;`iS0}FdsMTyiM_nBp(U)(&<|B&9SJ}Ry{VII^HlkjE<8KGI2^fE$9Y)}5I;j;&t
zLZ4|$oY3=GSj7iS+L|cjgd{20>~EDQr+mOWt%hWl>j^9_qZ<U!{pB=q#cK44_^(D)
z#|_Gj(IyZ9w5_8%q(icAH!}~c#N7O|`f58`hP8f>nf<i~wB&cNbv;j%0OOwe(aqRG
z=F89ZA?mdskt<5lCz&uNJZ)z!H>JYH`?#R}ruH3<u_xDVZi7FVFau~BxA{Bej9R8b
zwa2#(`0YI=)09Usf_t~X%{<#ekhr+wwkuzhol8Dv3hG)^83j4UPyrp+VN$33Ca8c4
z2sKZnh$Ry3SRzpSEaNB~#PhE(Mm~QU+6BkFQ~JCA4EnqOT-tNG^s18|{I~Yb1gNQH
zO~40PMG#RVf+q0_$|fK?Di9D6B0@j~WsP3I0C|9lTo3{Q5m2@Wh$w<V*+C4uB80H7
zvM<3XOM)PqkRX8w37J@L>djo=)VuYjYTiuM+^+hoPF0_!`}AMk-~ac2{v+Ik+z_5e
zvb2Brl@}=anGzr_5M}hkxopI-;H}+nploL-y&Wo!orM-)Cr75ZKrF%QWtjfuEopu#
zfy2*KW6ONPGHzpkZUwP024En6YRbquZ;eI4_}B&vQFNMB@-M@(ac!WM>G!EZhk_oB
zH!IF-oM@1(CY=9kh_Au|enoHywlI$12$0_+%a_@S6XIC1B5V~+9(?zO9quo&v7XFB
zJ?(E&i<?|PE}#j?>_l*6n32Bsn75W`B*e-|UkJ86o~E`b%(H@Ka|t`>QtSsoYMh`p
zcF~yr9WHQK_daP6?8gOOrq!y|BUg^hl9&t7l~=O#=fwz`nn_TMLhkX*bkXhg`RpB&
z&WNe;YSoN93zg`IzkL2supmh%3lnRFt7VIB0~~&>^o+!KhteE~r|=)TjD5@hz|GrF
zC5g{H-_^Ic8UpD##arUl59%<)1ffI8N$}4}@R^E*f+HX|4Ch=@k*X(;0WZ1zExzF=
zFlH96ASt9R%jH%Wvr&x$1p}_XM`PRDgzthMcwoJkry`~f1cD=E7mt$b{~B%1o6Nbu
zijo=!)5`_2Wa=`1^4v}rG1~gKqWW0K_64;3rD4^NHs$?V7COVVX8Af@7Cd0T?8jC(
zA(wvJ@Lt3}uEyon39-d1ST=c<At4&`<Ele@!6uik4U$s!g)X|{&QI>?<7;vum9M*2
z5B_&3`A3q1I$ut)mYzO>hKt%T=xC|R+{nnz25yu&+|!3h2rvoR%I%>B7&Qh0rwtRH
zuiINa@C&Ig@9m@9vVm7ta;C`d*R#{FZrC>Dck5PmB-J{YrC=KjwiJy*_schsB+sV>
zG~=QpSBt)?UZ;GcWZ}!lPD`lmIoluKG)kBDW7_2uICxM(x;dO79UYz0ior3b_8-%s
zd&cNPkdZ)AAc%o#=^bsONwhk4fiGT03V1%38{`60ZOuX8r7ZV+3Kyu>P?Jp#YY6M8
zmwv6}%aV_7AxK*ILK8xdb~BXW;Son;X@tFA7eA=G*gn^a-%-MpfBrr12*lTaR?>%3
zeBRrlsVtOYsLbL_5uxi4nUUMtbksdU?ty(Q$Hz$)5%}+JKBuXr88y*@=7b_H;3|=j
zhHUk{OHnDf={^M&cQnlk#F^kfSC`gCp7AA0{NnV&T|dFR$2P2Kpl&bQqi^-2Gt4TM
zS{WJn3N5y;s&P$^W^^2MNfqEPAt<q)ZHlu`vr=gHYcGS>BWn*#=?!*4j?xqI^5@9Y
z>T#Nd4?h7KomPN!8>s){2R&2c`M4K7*`FTAZCbRtX!LDCSgtk<sR!B?-UmzuUucdY
zo$K0Z#_Wz+phGT<PfnEGK7`?;`Ql*0hU#qlfr^?~1HYC!*(wGis-ox~=W|ui9UC-J
zeSJFKtsV?Ny<Lb1`M9-xjz=&fi%b|eaG;A;Zr^^Ss*F)b%tzWcPF|G+;;osp^$~L}
zYxtdYC(gt?kZ-Cm-EV7G2S^yN3}nuoUCI)T@#&P$Ws9(;sCbGLwJvo^Ld-e4YTYK>
z->pY%O$rEI8F!z5ohO^v-XJX?^M;W+1&yrZXB!M-+-aP12B+X9aU}vc4cd9k!)BuW
zC>_{H2h3UPt@si|Gj?}i-9k&6P)#OVhha(Wo++LTL0S*G*PQ5aw~pMO1ef=*vV!N=
zYHspj%m-_cDBk$9aHcFg?ZCbX_u#2`l=YC{Gt`L~gP7QeyM`Cv0CBs-{V4h)by07A
zBjV(%=)$pEV2mgo(W=}nW3O+U#iAkaQ+lFh@Y=_gGU^6JqZ)RWJQ-b*z5pE{SsFsw
zM)dU-HQ~AE3iPT1-&eCij}D`QYfxg@kV5Rc&Rfz5C1ZhY%FVq^y$Nw5ni%~<OhE#P
zl|~OiY;FI>&>%PK29d+rj=5|FBJp8Y05qIQJl#r}?hOcQ?HIlB!VIkZvL;9Qq_3-z
zmlwNj2v&*uDwrMGi`>pMu)D)v-aXN}W(et@>~i*!TGDAw%=~^|tn*4j#HX-Og?@c0
z|4r#Re0gd!*kK%g8!zJMG+U>(??ElaB!4=NWz%VL^OsL&E!=Ntpa-@NZZ+a}vH56~
z4ELoxDghkWA=MSz9I83&4?edB`f~5OQPI%KUI;=zjrnmo3M#`qrMS2(MJ0}9&-Pmi
z(3n$+A`NyI-X2O2(YkQ^kz6kGnqhz(t-`F20;$Oh#G923SHv!aAS{RvP*t@RDef^I
zqwe<7Z~@6rm%H(A$Ss?Qofzm)os37BSpnOW73s4D9ndm=B7HspXftRIyOk(+_Ow%+
zo5U@BLl{dwpv70h*%`{mHie04E#zK*W=_XF@K;73?#_5$TI*X`t#>F66{cLuve^W&
zjajCJV++5DPj&~i1jJS?G>;!94f$3ow!KIpPYb`bu3k>JleC$scpg<;<577cxOiJH
z^-WyGSIIA?C4CsE`>&{nRLt8O&v28KlUsXOX0D87+JXmOyjC%n6*_QTEx543P4?q1
zm{u7Gs+1%&+N!Ig*^9ukzHA<2dotEzS3OmwPCo}D+~eO<<i3fhByPX)?SYXu5z&>N
z!Y{3iF9{XJ?4_^QlZ2WwXOAx>mL8dA$LKF~-0Z4eJV^~wV8d6`zZE|j3A`dNqqy1o
zsx~#%kLeWk+P=~Lwr0UO?Na8RHR~~(ssuk|H#QQ(CX4%_sa#-`M74@ZgG6v%WY^_+
zql@NuJBj4lD2|eEf#XTS9v@c1M1VNE-j-#GY7<qr8M!U+g7>j`37TX0Ub3>^G$%?Y
zQ5G(`c&Y(qKwf%|&Sy<b+Ax%<@-b|kUHa=IyH(!$-1R>|gGCgku|A%Q*<4Fxc{0>m
zznL8x(t?T2rMcgmzQF}ND{KP%D}Fm_%Rk%rRvk2db$Zd$6Xbcsw=8)EdDoe<^thV*
zaPqjmvGCKG?v7R;ZMV-$W0yq+9&>@~R<}IuwRfB<!e3#v(LUQjcORw8ctqG5Q$J?-
zTE9i(Rjx(C&)|HT-qyJJ?>{)X`r4#%0>aju#C_;$gea8LYu~xal*g(-8{b|0z5>6o
z^6r&-scp}>^4&(QL3_gP_=Y*ul0aIpSSV<zc%1*WBkz%;Z3;7|g3(SogR=vtP-Ezx
zisErFYx|}yi?(vYjlLYkM1|O7ou-YSxbAO1y0iT<p-933vB%cM$Bv|0<#5(khSa$)
zXwI(=Cvo*xE>ON;{B5w{oaVU;>+``fgO5~#ox7FbIsV?d0kc~tF_sL(1OJJ8X=<@`
zka3n0vlg8@PHFcS@%=gzZ=d+61EOOt7P}%jD6fCnI5&{cnzuGSauu>0J}Gn1?w(Fr
z_|=llD_p=5$U8PF+_4|_^%Ay%(6TJVYI0#2zx0q<${<Ye6nTM0@oscFCaIj#RknL?
z>BFlKD~TLmoaCH~45wOKkLi7u%v%Z@bw);z=#*bK&$P}8#`6&Vfw1TaJ-t^ic}D)P
zoE&mx<tlj{5u!4k^JW?STZ5x=Odv>T_OchWc5waSh+^0CZgf%bZ+%olQH&J>7RBou
z){GwAv@mhd%cM0ea=Jl^KI&+E;JB@jy=HaC+3&!tCeoC>k41E6JZtouAK3`$!3DeD
zQO(HrOnvfAMb)*H)d3Erf8qbMO@yG&NqV$nIWC%CzU|#(1cza<dGuyJ+Z{}zo*Kfg
zoty2NzL4(G&^d4IL-F)+bm-*x6MJ_&@Vv^vx};@hUBH!M<~mnq;M#H@<7;lUsyN#9
zW_rAg+2CI}^!FKD6}BZ*bAdz7CR@bMJPNeLv}_u3fetX8%(=yb?6bs>Q@{I~P(8Gl
zv|{#y`kk`5KtJJ%9Sd<;bxV1b3&`d&zOe6)ALarj{=Hi<UUU*4wmtR7zWyt0CvmV1
zzakqLyVXF3Z73EsvITh%?h?xdTuQlsQX_2p{Lt3nRo#5~ZFg@d!T!+QJd~VVrt!ZL
zdBJ;!N^y7pmime(wW(_aE|46|z;SNyI%L-7Qt9-jMS`z|p<D0J7Im~HScJKMSp2=m
z!CdeaaCY2E77@5pmBM@LHLwjDenX_l7ALhxrm<Yj2WjJ(`I$<%#rt~vTY^?Rv+M@%
zuApucE+DuI%dlPIgyOH?A=#z+ImC4QTzd)Iq!-w>6*KdK3*5aAV-*^b5akj@7S+Y3
zZajnW_jNF8y1pwHF!ke&Y{MBS+ZKw<`w%^x^J#>Tgb^-uu8MVi>G9aRC=h2|lc92q
zck1i-)c*kX{J*>2A1eSub?=zW$QYz_QeM9aZ2Kxx+<Xn3d_tsf`zWzpIGKlPGsSTi
z%kM649JrV{#SpHa2(QA6YRwS{xz3I)l)Dl6UU4S_R$lgw#oe4s-xdTISj8y6_t_{f
z;aMa`|4c?!P;yqFg$aE+e)w_n%ef8bj3J!X+L~Qy)4`s4A(Snkeh7enp0uk3sj3H6
zc*-QlIk_H99t)_NxcM|Y(DiJemTmtk@o1an0Lb1nGN;TG5QATv&fjqtE}wAAlM)cE
zl>s%igfv-{mkmgYd0zAONvzNJOOXuQ%b%Q>FjL`?_1GXLC8)gS(X*PG;rUIHflHps
zHF{pdNhdfcHDYy#`P4g$A#<(vT`G+zw|YS;Szdg|ZE~1{Y6yie_T{+ddr@{iEtnWR
z9&Az6r8o0<EL>p!Sh&J1zINF}$4i$s^^TxrQE|<kp#!#V)34tIkV7yVPdCh>D~kCD
zJ5zJ*=W^Gmub1L;-kE-RKM`a`5&EJh?V98<fHBDN5D%hUe~QYjT>R>P+goKoR=;~a
zJ0#P`FX77Pw7!=GXO%`A_jgv>BhxG=LISGD%U6FW{OSL!@c%BNC;yQ#zPj}Ld8Lvr
zY+V^m6^<IPcw*Ch!QH_~gtIULs@pTvMT+6T+81hnGRu|g%^*Aau?xenlwec>Q`wAe
zcHh<~^ZH5udca%*dwZNo!wVQf(9nZgUv=S?o)x6+SDFq6RTXBTd+N7&Q+u(?x<(Rx
zI@YmXZ7=Q@$y~S)5>NP7?PJlEE~VJ}ne+5E>%W?G|CPd5cDOSx;YEMA*eeZ@Rqbhs
zO+~gs;)mV{Z)4Zwsykt903*swQ{eBwQXd)a;f+;>Hx~%!9WXjq@e-_HOL}fRI(u%P
z_1uPk;a{RCf|UG|TXN}12)QInkN$r$CGr29K>y$VYv#}LXZf@I-zxtGiB7PZ0Db`g
Dqffh?

literal 87344
zcmV(pK=8jGiwFSQ41!+(1Jt_*SX0frFS?KtI?_8qK|zsTr6i&PA_Ag95Rj;#Gy&-W
z0ztY|6_HOAP(W!y1StX`NDW0nihv*`0i;7h4Ulr;x6j`D|G($#bI;lL-1FQ!$vkV;
zvu3?(-Zk&M@9#IW=DvqpzylB0djYCGJ}RCM@3{g0`;(fQnvRy1?4P5qj`rV2HI08c
zXA~IM)YWCxHFY)BwY0U>)OBUm)V0+$)Bstv{}<Zw-=v^`z&rkornr0F^9i`;^B>k@
zRJ(uwe<0^?Wz=N<?fCzbL?5Hi07oyInwtVlOaQ>d_yKf0aMuLk!O+&y5;)DM#161B
z2>{H%f4$Bm0kHmE27t><lK)n|#-#Y~b&d%Bd-=a6rhm-z3E%?2eB{WVKO?a){;ce*
ztSl_7oFEVzJ2xjcHy0-t7Z0z%Q664?UM{YqB1icJg@lELx%ouJM1;fygoK6u2w`Go
z+{41k!OF@Z#KXlS^uL_wUjTk~rdVcCW~Sr75q>6SekOW50A{q2mFe&FZ|y(AXd@el
zor9B$QK9B2aD<7O`3MX1pH?%jMlh}eEc~nj$JCA41a0nsj{6B|JbL<;UG98ki}1Ap
zqP*r^|7Z?Q5m7O5i4!Lk6i=Pj($>+{JF9=;qKT>5CG*SIZ`j(w>~A`_xZb<(=I-Gc
z5Ev93f(Q+ZiH(a-NPL|1EaQ3Ri>$w1X6F|azI$I(TvA%~sk)}N?sI)ZYg>CqC$6iz
zXK-lv$H?f|&v86qc5eRn!s600X>)6PXLpahfAB{yCV-jgU(<hAFMftzM_5>xSwMgE
zVmcD?rvg6<>oIjU0b?7`9Y4Y28jsk8&Od!y*}@^Gd5tK1*MERhL|zMjg7inVzcu@x
zQ!M&FrP;qJ_V0R)1Ki9^jLu``2cQ7;dVZ23@V}npNl7>RdM@3OJZk21>Rrg~llhoD
zngEYA4OIc%z!&^`5p8`4L48ewmYY?s2*EPwK&x>H9q8Y|Q7Kxee=NK8?>~v~G)lMs
zk0cfR5frmxvl`YRJVytJPdrguR6ja^;sO6-`R5kR73`4iGNOo9paZcn*XcmsO8Oz$
z{hv-~{drrm2gBhql_sH3io8S~g$_8OrrXT^vAm0<11~`%1!Fr&O8t}*WmA(YPdc=R
zt2%9~D&}G~v$r`3<t77JO82*amnyx<Pf}vO`X9crdOqo8T|3R2$>EQWahEr`@aJas
z{)7J7i`QCuJ_rr{`kfQR;a~kfRMCT)2mO8V|HJR9n-1V*hv~q=!7Pm?v2D%Wf1)qJ
z<4ncir5^vU{?=c;a%&}S954L&PQRz1%JiL(jg#9s8xtYm|5UyqN(V@@j91_^v43NS
zl@3r4zfhw(G+UcVj#b2tkPX}kamPgHDw7Fvjbyg_n;bGDjU>BWkVtR07W?#9#s%WL
z`bLzl-dG>RX(76-J|XDCCCCZdXyn<HNrdBW=n%!=ZqF5OLit^Fm2jVwZm#tAQ@HH;
z+>_6xU+JWZ_ufb^QAB5%CDH-!)FaSp&pZ~49Qs^pR=UZET)gF9tMU#0*3MBoQhrw#
z=|Ej=2L4+Z5A!VOk)Z`K<mO)l-H~!<$5RIP?z0!B&;gNe*z(JI4I06pCpnYf?wAN$
zn2c=G0sbYshTMz@?~<juOY+g{eYP)7-Jk;k4bRy3go7Kt9!XMs!!yl72f8|ZX)APK
z@nn+<W~c?asqRPz8tA|Q2igM745rvFhS~eZ{1&n?3B}p}2k(=qjbcIv;4XjUs5?aj
zgZWXTV}DEe3IoF}Vrkr{ilPiSIYlu)X+@1w={Y%_biEbMUd>#Nz`;L#*P5IVaV**|
zeBk!LU88So!81CdY#|vwalI2Pau0!JNAi=*(!F*6%DOcyx+)QR`%;%b=hM}`gkjFk
zcb<V*9Vd?-6epxAAM5OWVRx?KUN$&3lBp^Is44F&uCf??+N7$p08WrTu=Sz?<D$Fq
zWIGw($EZ(f?;x4tU&r%eHm*@jW6J*e-rae2KSzD)qkWI^_-{0yO(3K+ACobYw`N0}
zp52aw#nt?fXejXrv^0+Fj}XzPt>*p3yBhaAY_xpT$+ZwI8o-tv%rF~UcBNtD2|93m
zE$gs|4xHDX|FH}apaV~Xd#OWo;H|`#Ivr?pE^?Nfjf`&+vT^2LVimGE_kS!Kqv(K)
zE_ml#bEfrGI32*H;-~?>^S098b8+c5Cc-<ocr!mg;dhu%{iiRmTy{2hy<M8bW-d<z
zq*rFgw4pMnmJ3MlQnKS0hn(r?NH!0_nu<!oqEK?GN2exht>5O;i8t58x9VLC@@k=p
z+Rcx1PZO<I*JgT$?x7Lu*H5AmQ5p9N-_I(kiQ!w#J45+3@|xj6NOiIoo({y&m<9|b
z@m|71)yl_zIZm8ED<9+*K(YFI^JU{F!6^*#{;{Iff-M8$c4qxt0Gt_he<K0@DK8<*
zoCuCD4@~V6AyCjSLqn!Dt)>mO`hKl}TLD6oPLdMNAo?T4W5$;cX-eGbizDO+w6#lU
zceU5*r#5^Ol)-s1M;q{WEwh*RC^By)5T;l0v({t!yITa^77&I6ZXHk<H@rYR*o3|O
zK%f~UyPbWYxEy-#i}G(-gUCekb&^FJ9T3NGyd&4O>7qObH)ek^-!|t;ZxTx47+*%n
zk;GX9w@a8Fu^i!Jk)RN;#3*V49e5CKDQIcqMnn<5ri|zmK&-OLDnQR;^2=V>UR8g3
zn)fN}U2MCa`FTgpXH6$ge654jGJ?#M;ZVr`WuR&ittlyXbYQ)i7VR~-@&!i+gu`TB
zNY3oXG3?84gfV=9|6YE`@Du+qR&OP=!*_O}>&;ujxY0p8GzPp<GB-+-<0ScxyX`)$
zCUd-58fl1pO*={Y-K6&EyWv^i`ge_geZ1PPsBrrXhS&Z{L5bW*&6<C2n}kTg2hZ!}
z6*)x{<{<$FfoY1vjg<$AC$W*#s7Og?q(2GUd}vMn{k>_uOsiIg_hY@dgioHh`pc_<
z+&F(Wu_Yx`@8#FFC>KeyV0|rW!KkY@l9?KHg2Yw_YM&m5EJG8)M}2@7JZ$*bY^${T
zp7!Hl`43H<+_y8L78ao1juRbjzCuSYJ~0-EWjy<I{-Y>B3IKUNOWGfq2yrXUtTr|g
zVpwVV+NncMiV^kg)1AaO5KeLo$JPhTowo`ZAyQxFov*$_;-5fB+@Q%IG!yw4X`(Za
z535l3r9X{?^O`Ll4Nf44k#=u|P3L?NyB!*Ps^!ZWV>`}So&#~XA#Gy%lN7ib-Ex>5
zaW=Ad_8U-JgwL@$xj7%d2KXxK&7FC5J{NI2^#`l3X?1bWV$VBmo13qOVJVT_U@3X>
z9?{WaRGW1ZY3922-pg)YM{MKx7k?Nz<dhlCo4n@S54RRS)nOiF82|d#U$A1b1<hna
zb<dg>>(yKK&+cczg$~3wS$)v$r1FF*=40drITiDx{<GO|(^|&!^6HV6s7r?l6pv<N
z?lNjJhDL_YI&Dv{z^SvA6}+J#kn%gl6L>Q-I>51BL>+@}#LyUMFPwpPg82?-=s;?~
z^J@!PepH9_&4L03*x0*HTLtg6&;bc3-iU!58bMSds<XaFaUtvBK@E8Re<;bOUj@J8
z3GGQsrwLA4AAUqlgDCF*By>TsAoo_=@-Qk>8cb1W$lJS32e5qPfqxLH-DhKF*qHF6
zm=2saQ0D3sJxD9}^;V2EJ0`n!jqiMfS1nqZm`6#`fp37RETbhb+mqwNB=zB0tNg2W
z-x-l@Y%7`(Q;+;(e$z|`(k0Umeoyxlm4<Z;>p0lX?FpK(oDvj$X2)5oKAiETG;Pt^
zrE3wLzB4@z?;Hmu)<}v4N$Xy(t*w8iS7gus><0JwkBaAHQ)Cotsc58&2unj`s$Rc*
z%jrr=A#TZ;4vcy@_bt5K3C+n`wJl12jH=CB;~`0Gi}NM8#<WV@et5aP_^ZN1ZbDF1
z?pf+eupxB~&OrxSlr}zWoT2H9**x2)c%RujHp=~h6b)OH7X<h`534ZZz1dc``Y)gE
zmd#H(&G~od?~BqnGiNRk&_9(NM!j`S$`R)ehMb4iUAmqrm|yg{z-ua(IWctFhzWcP
z_Lw#yFm~baB+^#@m4<z7%)@6|?$Yt|UrzH$R+(x(dhM_OosF+8PedU-Qikl(K2UsP
zB(Txj<VLVSdlkc6Wh#=eN5~PRx{h2?VnV`p1jsR*k9WfUO9K*qC+Yl)PPMyd1E)+|
zdClcxoTq1~js)v;*)1*WBz@tDM=$Dh{0h5wE&z9SI*nEqJqWW#0wMdTm-9U_Y2}L%
zu$jDlv2r>viDGo)yU{L)6cL8=sBwSMOW;jsZQ-ez>eK*%pquMnm2@C3;_wn37`Xbq
znRtr<8s%UwQ?j%zYP@+HIKNLV3{=DV;6!w@2l(K$geZ+GpAO7<Lh<m(LFo1kEmevr
z0~U4tXbsFW9e|%jwPLA-sKsJ;I9Vl`fd-2ITj^Ip|5Fl&nYO)!x3VZo8{=lQr1&^G
z5WU#EFOg1$O;@K=g?YH>0P6s1QyYkoG(pbO0nWqSr*t6y?OW7LgDTCmfF``RCPxQM
zE}UoBK_HxxdJg=*yMe!uA*9-j86{;+n~@?fuTv+~Ld(P>KKzP4KQ}vX#2O$0O43Rf
z8jmQtG+Z7e7G8sSyL+f66|?HKz<_84j8G~P$j7;qfQwZAND_IeJu)_1Rg)$s^~gT@
za(luXpWj%op&g~BlS?WOe2S{-tsAv5DQV5;*^f7!wy!fit@C{@hgD~C2f|PzI)z-d
z)i(V}<=wlbuZ$4cKA!qMfp?@0to}8vfI9bh_)r~PhZRHc;_xADdE5x-m8-S-GOv7#
z+Rq)f>b0&z=AzBl&Cy-x)Mnutl&A-a?*kH6+>g3*XO>6fH`~6;Ny|BYx!Ix@_hg@3
zG9o>Nf69zMSyc|XS6Qy8@Aru@a_V{h)iLn(^TpXI5ns_bnMn&@bk;tZlZ2a^lUMP5
z!mbhV5e08fT4aey7r?C`Md;@cBXSOgffQLKFh8mrxpKU0hbVmByUB`dHyp(0w#?<^
z)s(q-_1z3*AgM%XdNn-JLElqC`}<2hf!1FZT&SSYN-E!7uj!T>R@pJXw$%AOop2`3
z1$fnc(Fe_w&;x$73fhL4MM<uOujf(OhiCsp3A1?e^joxv5G*##_-_LApUa?>Z?qaZ
zFr_sf_%AU5t#xSf(Z&R!u*PwC<0nON<P1_6ZzO?|LEiX5%npeUoVk_pu<MyWCVxf*
zca1I5wwEo#r|V^(*6cS0*YKJjzeIEfot(Z&H(@+)&SyrKG7UdN9KnQifNR}aR6+gn
z`swXC{PXF@UbNaZwMd(-GjFY};2h=+PtgIEJyH?Q98VHmQ=Ofdywd&nE>ShU=n(SC
zt?#@6dRq3jbj(@s3cNf*X=ZwCXvQt-!{D?T`8wW~z478`!-J4dvr{+Q>M4%di#NhQ
z;d1Y8teb5>A{e2=*EcF`wO6>V-AnQ1D|g)&S$1W4A1W1@r{9IE#vwfzU{>0r;41NV
zeUaC$LVER5mbsrygzhm}u;*Hh_0NnMT8;b2=rkuqJb-DzYf%Iv<s!Qu3d-nO$lhNe
zfGGkC__@(kH1gy~4LSy^n!XUdWrBI(Xy<EwEM&5JyH!8^YS{HanCX=OarpG(H7!^-
zO@^8uDNF0a>Qs><4$VDAcFhq}c27xMR!$yubEy8i0DtBB!$Mad6R>u~Eve+wN(CTO
zpzKta+%QZlpRn5Roat_tt|hjtm|r+H@=w9;1=*TrDyh1c&cK($WmV~|NSX`9WRZa(
zeM0}mP=ptpHK}j0(qyhMPwc76!^?A$jyk%U%!_tL9oC0qZ}?2|v4IQ|drMMHOV1*X
z;$Kvj;}kebJwMmffJV=boBbFbM77qC7f51m9T1~W#Cx$ZC)?T{W}H5VIcW^uBgFyX
z2}y<l_-U4j&;wfCokCMn*jg5sxbL3vS>IMn2#i`~sBE;dG{rMe(=6@jvflCTXG3Qw
z?o3;-Q7gZLj2JZ?E90I8eegrs*AP~s{%g-+ve%-%@mX?>HHBSbS6c4k5C(d`Hv3jw
z#VhybV6IDcinxIrOy^{djH4MLRTlMTDO<Je*4YCK1mX7v_=;NpkE4SOCU@M?>(K^H
z7BLDZ)%?0sc(nIy^q7=j)XTe|!w=1u_txO{mEW~`bIh(a#&@><<j!!i!V73icWZA!
z8fW`OBZoy{*{Y9HXp-b`Cf8v@a}A;FZQjaOJz22T=fjSG>YC4B1&SAOI-wdZ>D)H6
z`pYBu`?1EOKF^eHh6pKP7Sijx6@K5<9Ir<Mh<g~9SxZS%V6CzKMuzmo(B8b33le)j
z4OQ<Dnmi)K$r_qo_*8w9SM3s|Qcb-v?f2+F;@nXKS1E1<NIC;dMLC0ss%d74S}$Jc
zjr@3Han{^VDgo%KunN7}20CQGjc2V58NVIj@DZHkhqsbO{o2&b4HvS)VI;+o9N||x
zYDuEyfgV>!%(}mDuL~V0!n)S$Nxxl#qej2v)uLLPk9r=MB|uIgn8c8#<nUl^3zo;a
z+Vh`}y^kuMCp*Q_jy0X@9sL3R1YhI9qsIgAR#~%fp01--Sliv#eG-mbZ$K}(s2p7n
zczSFT92PpR1ujFrj->3*q+e$cqbI)@Fbq(d1(Q$brM|1C=01Q(dt9_6_$Frmc%OX7
zoqcNglYGVW`nvtqr+GUCBiJva+TRS?B(!(YWz)v*mx9%5I)2O#N9(;!HNE*W9RN7G
z(S6NvVGK{U_V&=qnG(=klRb^Eiyg6)YZ1w^S<}@Be6h+vqK{<ESf2*`mFj)@RhhQX
z#ksd4KYvnOskt-hU&6F#l5yD#D``OLI6H6~eXCZFg_9-XW`|t^IzscJPY|aQJ5uYC
z48I8+-0#dIG@Yg7_d%om_CjJ(D%<=y^p2&E2W_D`P)uVKqtS&Q?>_40GLN3%D7)E?
z8SC>GA*imuerD!#GfX3Cha_|ZUY#~+#c3<)l(L37Tg>PblhO`-hobgL3E4>ds;n0E
z?sfAGI-tz=8SPFEVGuA;0?(I^hv8(DBTwZv!$$fQBjaozkK>FUO)j${RNH*P)3pAE
z-ArGCLD6WW^kkIQWF%8l`P1Q57|s@_w>=+_TCfC{K%OLmJ5x_9m0!20`tgjQ9Ufsf
z5Z!Y%J@c9CT!s3hXGUdT+bI#j%fZ*kR<xx3I+_q2=<tm>n}NqAgOURE_TjvU#OR*p
z$+Rnv_Qi0O<tT32$he_I_F~@AKnBrG7+;*Jcq^%0mS>v*M1{?zVyLrD?{Q57w$*D^
z{q|K!o2~0yE9Rz)>&k@!rnAFp^8DLnTK_yzWgn?b>U=a~q>ZLrC>!?>DbNdlc|Hr0
z)MS6CcoK7dYd2?<c5Kq32q{OrH=W2?X*#NXQOzE=(Jj8yrjZL-dc*hDSJiua<))9%
zLej|iiAnh(y^F@n<};kc4UxzZnoP`feT8VPiH$^vtbAPEpo`0q+4?MyT+wNEoZro0
z)F<9167RCI8DS%3opRn5)A4;9r)J+Cp~q|}b<frsj2L~t9FogI%{AmVTKM?^bGD%E
z(Y!-DOO}Ak{8i95xx8!E{+EDkV83V<8iPHFyg>ZbP}g4Wa+&zzsjpC$PJ{MGI}TYE
zh3x%hbLI`*s%Wp-5BzCLE2>E{vOj1imq-XYKoG_>fTL!*2Mo2OK8(kpxX#W{CIhIy
zGQaF+qvM+0YRCz1oxHHBq1#lrP20HR_#J`Hqn}fC_)}cKs$u{yuUD1UE&C2*FKR?5
zZ+;9C*H43zH3JBAU~C9?uvX|3<YkZ;WN|VD|H9xxgB=L{E3_ftx$wcALn!iEEqRn^
z-QuiQ`xJiYWfA}Tc8<WT&rb&Ki0Y|QrLbY9TXK18iCNLr8@zaxwY74}OAAw-uU1!L
zOZvGjqyk2;{?r$Sk<*oMT+{U#XmW+jG=4N`L-HC?r_$d~O7qgWaNU3}-KU#A&PYX4
z_$w56ms&O8n)<dk3I`698JOmEP~6{hV6bjz-SnDz2yvk`c$$mCg1bhppUu>de09%r
z_FHB2l@D3g)IMDnmnio8OOn?wAIKWU>;?_vI@8AwXZJc5*<YEYEvs@Y?w#tf5MpBa
zkZ@Q&*QU_?%DG@RJto91yAe_LMOW?uNn-w8jT2Y+z;K!BIqPPYotFJF*lrkEvxstp
zZ2pF<(elFWQ*y-oU`oXMd76joPRreMpeJw5Jh#IRE(Zz}zE!3J*Sm5|f6t}sl6^_f
zMmt9R-Lt=y5a~d&a$k4#L<;(^bMplX#W=l{benGxn*V_8av|~*#X*SNctL+d8?5g4
z=#AR_`-y3$)+Cki$cr-W40%J2WQ(98@&y70>h|fC(mQg6nq^bZ`qHUYW5YTYzt1iU
zhI)NLtm0l33zodFlY|H(6Akf#&v7RnkJ#s2kbB$n(J-3N&l)Ye{{Lt;;lA$<nnrp*
z^qvkl(18sJ@6e)g<#ZWpDa5d7l@1U~7j9NZdCHs|K5<L;{7`L{uI@8eg`Tsg--yUz
zkIab(4@I`;RjH8$@rUW+h;MYD4x)~DYdO;R;TON$#D)Q_mw&Mo<D=B3Mq7R(qmD0^
zqcK<ur-`pu=ph~8qR0*kP;dKY>;6#S?3&cOH<e-}F&sX+Xd8SzHM9R!=Sa|Ly=?}*
zvBTy5nJ)7~i9z_tu-UpOZ73VL_ll*p3D((|a7iBQZ=Y9ssc=qh4<!2}Qi__5G$X*H
z@<eLVf1(LIEAU?)4<)BTE7zi-)zy8Mas7d|54dw0PdE{@6}FwR@(h+zfKXX5Cx_tq
zQo}ek5U=pq8sA{o3+XyN9s*}we6PB0hRv~R_{D1rrfXCVoG&ZTKCxxgm()+(t@k2$
zGhZ1w1OHkzteoHYhz>MKB}!Y*nYhDpBYb|>#WD!mohmwj_<N53GUgcV1j6x4SW{DH
zQ^n7vGVwFr@-H(atuKw9K{4I%p#u-Nlq2U?X<&q=GpVV|0J-Qr&h;={pk0%Rwuv5T
zRHP{M-{;XUolJl0`J%ks<A60|$}?nl+&e#FY5li-0lZ8dfJM=Pd$jM<{N1A2Y01P1
z%Zv}fNh&_ORR1=p*#($Z-q>`TsICWU*3#FJ^!q(zyCG+`&~33aglW|&>DT5run=tw
z&RZ=R<Mw{)7P;)2f~Ujci_Tl$Ru?><(w3kX?9B*nIMqZ90`^5Gmv`vb44nMp+CAn<
zexezb*xJ<CN@FE%M0<bLXG^$K@K(&Xa^k~-S!zC;Bk5$FJM)6^p(;5Kr|^}WLDGed
z<VV7<4ElL{Z+Gf2NHQP=%O|VuBFG^#s><7FF8f!{=WaPGjNb<hYrQ;;n(PSE76PIG
zfEfsW_Xle{BWIB;+NL=$`Uki~UlWYe2c4xY%ZD=&U8D6>kUy(GTpZ;tt&1=GxaeS6
zM$YQo`)r{sFr?CwZki=kx`&%G`l!+Nn(T21S=Mh(9Kz)Z`;4~BC?Sn1vLB9wxi6QN
z?cGeM1Li2-wyuJV3K3i#bfAW<$Qw;LXHL<b5&LCXM0OTEr94IxKk35`8jUs$jsy8k
zk+(=}aE>vLJT8dzxQXcF9d~5nn@{H+zaPK7o-vQ-{oXF~MP$1lMQSAMq__05b(da=
z-nj7@>NXTe>UHWfV`k1Nh`zLjyQcJ%%tMSmY;~|h&w@qIVn}<5HbqU@R+b+gXvjLZ
ziHYNwnansjyzEW}H1p8FV@O?M6E?$@<r&|U)32xTxz`}~p<TRRIG5e?v9;<pc8)S~
z-i;s8&oL;XF0L7fJTXmyQ*(1wYp3%x+ap;%%idgGdHv!sg6pJ(4iiXz&G0#@%>Bdi
zV5<FQSFvTwq!jcEoU90M-AM805p=*pc_<;wp`y9sBnz&a0zPqfGj~6{o@3f<F6BH3
zm3B3`ce|(U*rTrFSX(ms&<JTjo_<hC9%`-C|Jv2)f|2bA65GD{r0mh!>u^}&1RZFB
zdQ*;fM9NTH$WLdKu2~<P%qdrkGkSiq6*Qc?GuzK>0}Z*3)%&Yy8){mHqv_E`ra4h=
z5s*-f2jNiCg3NHxBNZyz9O<_|9y_Bs6yR+$8HapK2e>r}5&o0jNXJ7n<TaA;<F02D
zhj(w-fz4&E>g}*KG2>yYHpCb_&o9`4b=>jcn!oz+rCwIP3Ov;L=}o!t{F0J1wzA0G
z7(}9Sjp3*E5EKJF&Ez*^ek8Hwp@@w$EZ3F4hFR2pvv%h6Pd&Dp5HIyWN_dN%7g9}J
zNYLtMch`@6TyWE*$iwbvc~f~IS8S5D7$>yP(!<8NLxMlQT)IypQ`)T9))IEPdQ@v7
z2Xeq#|EWHu<BXNwJ-MfLixsI~(?!WnwbMM5ut-tzfZQ-fx|~DwQawie0tWm`eKmwX
zT0LGi#Yb#;{S3zXP}Z|Q=ysN~0@)%!XQ^DndG^v7tIG@Ih<?!9ah%^G#gSOmM%c|z
zi3_F!Zl6yN{+ue_FE6q^;S}x`Z4ezM#s8K87pyTP?b;&Z3eM&<V)Xn4?+~6YZq+p+
zkd+HQ4GY<eVWX))0z`Q)ziv`~yZk*%%O`9h4hbtHPZ(E^1R5;XNb4+xm>8PHEVCO%
z>mPk+JTpM><#5Mp4uP#k0~>ESbl0m*wo5TDYkogW^IAX)BlSrNs}pEvL@&OYuY`O#
zSknt4bmmfOsPY*|6yKH1?R)r$W_(dQm6?Exe>*~f1}i+m-n@<ED4&)pOwNcoka_d;
zo1Zl3sVB6mvtTCq=MFqJSFMSdI+U~a&iU-;%ICW*I+qSU25c9<p%!9p5<3H{$aOPi
zsruZp@T2Z!3l5jUj<hNGgb{zzf!nZwAJQEd*h6z}QsMV5#{EJXAB~Jb#c#M(*(k=u
z>2_u48dhq=TN?qeV*5`z8*cL^{JyaIMPkvb4dZrr0s(^!^H$<5UiAZrIDZ$+_g^@f
z2m%XW^rSrC+iZU-nB_-8a<wA%IKux}UuAx3mdzX5siV9P_2<v6WN9&NX9{o4Z>-hK
zm<RS=e>1*JlO0vTUJ7&!*SQ(9Yp0+@2auwW&SGh<*@VuoIgVFj@bLITC~4s?a`ODe
z?fWeJIM8qFJsuz|t35r9#zD3fObvI8Q$5$SG%}5?_qBc+xgO!jJcX&ISdiOD1?ptU
zHuD+pBNrZ3kfS20qlW{ynBg!bZaQ!jeU9vRsLcrQg=Hab@92Q1=V4tA%<ed0rtVgo
zoTm(?7&M%{a~dG#anQPDq-uvdYi7Mo4+*?joU&h8y&5l>-*qlvAerx+(yueC<=ZRv
z<>j(};QNG>+MA2Ui@*$LHl}PAjw|Dfdpfb?5K)Fm=nRV&v>s4?<aQUqC+qYC8&l@O
zTn|WsMX{RhU^Wbt`nIgc4f9Ir-uo}cxl2@`V4J5e47vB@ndh*AWXDp9@ejWCjZ{_P
zrt;U`enn)TNaaBCxnIurjI4kY)`rOGNomnmaMXCLvn(B8#Akx#;0pHa=e9Pkqaz1*
z=>VrGNHRVwO=!EjSH_EwTaP+4L7GCjkh+0H1nkPw`y`z*km>24BFkZJa^MZ_S$2P>
z9{JI{>hyL+%DocDZctQUs51{+&%JS#$xD-(%)_qSyECc=j*!M%qDg@UU>=VKmCW~!
zCvx-LnBv*x#Mom1K3a3wg4ZI^?{bn?$vn*a+Z(hYd5Ngn2Cn9Nq=M8Wc-`$ng?dD=
zA0hN0kyY)JiR=3OPY~z}`Hd3S?Cx_^ucv25N=1&K`00Sl%{8R(;rX#?MA4T;0*?~G
zIZN`FhN+dXujmVTZq$$!FAGLH0!JKWu!d-?`l3e}S)oJ1?ePb8*21B65!1xn;I6zd
z9R`N@A~%Sk9du4y|CBwrIi=Ki^E>X0;G4jw)~q+JS-=<II^<Cr#{x0AeZ0W4;Y4o9
zjI`6U){WMz_oJY39quHfr*@dI&!F%jw?K6zkGwb97o0|``?;3_mK{0^JPMb-t1O{a
zi|p^Dgng<6_)5_I@zi7t%2(cIEOHnsTz=;2$HTKDJ{$~`$glAzjSi?nqLaa71?y;S
zZ!2jt98HCk*WUZNFPIzy3r^got;;V^X_<W=ZC{Q3Ae_os=u{;uPNtQ8u%3bb+%Q8W
zVU3=P(Sd63Cv?EoV-oZ8gr2~o<wCQe1)5;ap<d)yQ7})<H%EeKdQHZxb#mCgYVk^d
z|BIfnuWCowL;;m^Qi`^m1)wn`4{@Un%)NbM(D2loF$Bu$dIoDOmsl`ui@#U*@QV3f
zMN-pgu}g!<!5^<PmFgG&EOf#`zY6~XgBGU@h%8|`O+{@b9;56R+SC?9q%?9%+dgsg
zI~j$g$2QO1Vj#sp22wP|d_@sQOGf`p2X8%pk*iI8x-R1gts#%%?biBz7J`01tDlP*
z@{Q~Xk5y1ugtH)4E|8P)WNkFxGQX#e0I5MUKDDu+WZvHOVdz}&@w!_gCa1q>d~9KH
z<%cuE@Qg}5aiLw4QCRIRny&kX!Y2^=_RZvv&#@XsI)fBEoD*7&<Q7YndTBhAhyjSQ
ziSpwlOTG@BsN_lBz4`2uQtD2R>Wz0<LYR|wIDtR&KL43adf7bEeAxNQ>J`<C4q!ZL
ze&@N@LQo95^mC-vG3;~Aru^%~#!1pAkGa`GyG^G=9NdfwRc%n3=^wkciU99osqZPw
zfm2QgaXWQM4`zQ^k>^#m^tyK%qWr$k$wcvCR=%fokq=5rx*@BKCl;vv`4|i>Nv*18
zC8@6Fk^DO{zRI<U?y*M+SadL7j$)ej=c6@+pZK$G!^%ghhOkgmBNcA-DyCVi=^maG
z?&D7NUtzLY*=!VC<pGf8#Jt_a5sV<Qpe2_p#`<H#^-DvtL+^&sK9UUf)Oao`Ze~W&
z*?jGLAo}V6MVsK8<Oqk6i`;fY=BqcjzhYGPERGb>4tSiEizysGu+kI-qC(<T%qEXa
zyBEhpY{=8VJC>_0kAB5%@)|F!JQ^cIaQNxu6V*dVSUw+@moB@^gw4m#h&5HAZP)J8
zg|Bxzjm1IpQfo@i3TW{gzvs&Ke@22!wnf{le~x|$6KQsJ9k_1)ymC^dG;*LZ+W#sg
zkGvM4?BT|i=Pljzfq%lWQ|;;bNBm}Qm><FA+~I(sS>f_XHy-@hVdM!uJyUw);G#>+
z@I4qEkcM7f=c8z4#f;6;_<g0nK-b3aLN^6s_LCaI%GU#CMb8Y*&woFVN!p*LNuWA%
z#Q;P<o-Z+QEaS1NR{wT!Gv9N?17E1aG;wb&j#vl3BbB|HZPc>DTjV8gV*SwmNDS?u
z+=_YJ{R5njJkd6F=G=iw<NM_vQ`Gm0*;jPHSV6ev9IKFBWD|O6N_0@A#?oO|8fsNz
zx#eUvVFMZ+{U~iN%L!mF0nZ+S!uD5ibU<b&PZE)aI)Xg*Nc)y6GHdFX;m@UID-ND$
zu{6dYVM@PZupvOk?N&BtSr8kY$kzB+9~aC>;jKFTC!psk!Jzm0OgZhu%#CI~nzTEZ
z+t7eF8Nb3(Rc&>vp<8n=RI*Hc$4H+WU<Le9Y*;4uX^#ACnt9<U5*GRbJE|J#QRvdk
z(s#K8GM76?O55syC=MY-@henc`vylN{T{h9gLStJn0{qM0dvn-JvCXceLV-bW{hKR
z{aAOSmSa!mNc5P&<8sA6^MwbAw=QR0$Uh|wt56^CpS-Dfgn^#>xt8J5hKNWNGWN%t
z!I_h)-0oZc$5Uh!?)+MKyZ+OV_3{D3tJ4g9pewIXr-RBr=;(Z$aBP3#*cz=zNIhSw
zD8>ma0%9HiP^ox5cGoa0NfBLPsFwfquj4Q73ZLtipJ9R@_ktdcJQg`210AMGBL#=^
z+TeVFC6WCE>%?M?Ek4AVv7_d4h9#_@4Kmf>oHYfehJ#vhx;8&$-jXtQJIJ?)=eL)#
zM;d}ukFJktDGNYrrhNL2R7H#EoE78ku|j-%r~|D5cQVY3NtZ#X@AC;wdVEVr+JiSr
zgk&e^9dk`#&bzv>gD!4%LywK(eOE_|&Ks)Y$v!nRP!z$^R%G^CPWlwY11t-jOgVih
z_$@pRG1U1ZV=vHPHX(ZY`c|uTI;t9%pxn{dbg1kZ65N(R>l_gXH8Z;C{!D$gZdu9V
z%XHmnTWUDmnitV9ZBK?1?w+sBndQ=SZ%n6ZK~JZL>cFOJI*~SH-I?@MXPK;vq~`d*
z4FQk!!qZc&5|<v+6br==6VyW4!(p7HgDT7qV<<B&rQOsH3np2!lxPVQBj!_MluE=|
z_*+;)N|a{lKq8nU@;ItRfBinmI$EU$oQPsVwjSigqvf|$t4}Dryn{rQvFwP69nR7m
z_=*i#XTR|gd}Fox1<5`%BMZ~vyMo4k;%nc_`0)2+ORjgx#tb$ki6xFY_&gjgTxzn4
zPExEW5d5camns5wlgQR3Uhvzby%>%f(n10%oGH<Ds#U{51l<16k@@p|I6-WYB1i}1
zgqsi1bl^?)b^}qJ4s<vtyo)D8+_;Edgw9u_<QAQV`83%Ea09kV`9;m^h&jk)%R2kU
z6T^N2yhYjf%hZWHf<o(kM<cJSVG0Cv04aR^=?)Z&YEaMUc`_ZYO2~bTI@G6_L4}ZK
zJ-5iP>p=<a<4(-6XCA5u1@>8?rsUmt{Z?ql$g?wWE`(?Qtzo43J)XE#B|op4y<P?w
zx~AQmcL4{gQ0mj~X=bD8EH;Z(hssym4FxqqGi(MbtQ2Y<K19fhI1we8zeo8i)r=yI
z-y>Wsv#$Qw;qun#RBiSPIo%mEA-}VWPDR?NCerjM#?_>}Rxq1qq%^js+gI>c&x=Yp
zz{jQ|3D@iN|H;B8wghc~vbjUWDEEpTC^vuX<T&CEkbRPE<JIpagKYZl+CtS9FPGR%
zdtA>3wyD0mA7$i*?+;)_$j)>Cr(c@$e%18*&x@h+f6Y9_MMUj-zRv?9Ws!~<YfxdN
z$BEkPwa*COHe7EM-+JRLNCw|;sD^%1eT&bE_&Y~9KJ$+E`>~Ea<?`e91@QdhBj~H3
zR4-JOYAY3PoO~srZPvowcoY5Meo+1O<y()Z=>Rh~=mG2`QjSzNQvq+w{t6c~Gyq%b
z!VFYhtS|ceKhy)OFSWu3wXT8^$x6G`#Q3ISlq=$y68Z?@5&O+E+QaIyzQ(5jKC;-A
z16WO)RtUd*rGBquYR{s-SXhkHMk-d4>_*VZ+l30a*Jgh>cQjQaaud2Ibti*%#02T9
zlYw@h7Enb>5<wkdKX#JZ>A*ODuZH;@S1UxC%M(7E&%jS+Zk!&lw~>MZ`&37G8*&k8
zUlGZ;J(R+WC@p<Nr$vuBv#iw2Yu@rsW)F%G8RSizvPeJbmF3nRsXwf-*YK^=>8;VV
zGn=APg2S^PTN$+Z1c}*%`s$h`7jBB@hoMrr^QFik!p-NC9brEeUJAJ?t_mWSt<P$&
zGq?^7N_K2P@f4Bny(n@k$0@8oIQA(AzIHwLG44u`&kXZ45EU?84eLlXyC@=agm_!m
z*^nf<e%n4>=i%%4m+8dUY%QzIy2Pz<qy*XeJw=9STl?7c;-2N<$_fDzUedMy`Y@RJ
zBQw^Vfo$A-<0M|B_);Qf4S$ih+Tl%kU%{k#ju!hHD2^{qX)U&>*oBE=<^*yw4U40W
z((ZmfScM#?1L#UR5JU(5s(@g#I4BH2d&~M{GfB+GBvf@x`p{$%`UxFVFzidnupO-t
z)frcb@%q;J*6$1Fn<M9=d+R^?Jps=2@Lef<Q~$0q$;zs0TU7+nc7b>@!S4e(8CEr*
zRm|p!Y0kNu8o_3tk}<|>?YUpNN8@D8Cl_4*JOmL$8j{jG8prIFEDJoU%se3#wmFgI
zVlq4R#PD4~k|@s7ZiryLW*n}9F77^D)5Y&pmy6`QJ2ro@G<6H2o(ZDv20UpFAcwbw
zS;q#3dn;TRI<xJq^7Fj$y*o!*xGffo-vE$kjKk+8T~E)hjd*d@M6!6KChdo4xzuiR
z73Jf2rOUh%BAQ>tfcN|pvxcp;KXazuC^3pIn{_e+uXxRrpViBPRi>})K3@KMD0dZ-
zf?I(XXvMOz9A!5=>XqZ7q(G9?7n;5>kBZF$Bbi9OR$k4l<QGHs?!$&7QGc!9mWm)d
zfow8TW?8mRHnCwQb{P~gIuxNq@FtN$+Lm77&DvM&G{?QB!0_YZ@3ue3sj>SQ2eu)9
zde6>7?j1sD16X+o1WVQ?Wm%@TH46v6J1H60I==EGh2YK$oVnrCN<4QLINgKI;l5kC
z?^d9<{QcpG0ocOEM0oM_+yD3rk<pjwG>IL3;<V(p_F{e#G5yk}Z%W>$-<c$xOEu*>
zzpZ)5gZ+GN6*mRaKS%t?@ZsV^VGN{PWgPK)YpkVY$98puBw?-8J@PnSS^LeApPowJ
zb?}=d5R3}q7+){k7^X=A;c*ExMMT&w$#xx%a8F9X_A}8{Qk|{u!uL7XtDwv5vPb>-
z!W{O4N86pXs2M(lA&RLF@(!_Sj(@_7zhb3aO-5KAJSnY)SgMZMgbmOnNh(<cs>mp8
z0R472a3XVBxSc2S&eAgbqVb#z58yHIZwrKW-j~0LmEAfgCUw+R99({K@E5G|eo`5=
z<Q^S>K15Cb`15UI&6(zfb-u%c=2?rEAzb-Cgkgu*dTEg}bfCG6riJpniR%9a-rGTS
zrrrCcl6=D!zGk7ie;7lfp~(aPB<XC{!v<v=ThzY`G~0cbMWMFP0Thhl{`*$zVvUI{
zZJ0sF@;*a1mgzt$c>D!?YY#-F!bt`Horjkck6*%HTNb(dQ0Nke)_P;y!2Kk}_iF#J
z2Zg!w{=Y7==>N;A%V>5){@A|e?wk4s>T4PVsZTA?S_yqWuxgJ8@A&9wR-fH!ckBet
z|6QMk>CHRh9gzv{4d2oex5K4j7L7@Cz!LjN#k=Co?SXHC#jZhe#j??SO$byHM)@6L
zqAl#jql%j*#Y++?Jl;WGYCT;kYG%cdSbtvj7U9jJR=$*}V+^|vL6kxhzyi5@H_75C
z|F$+~Rr{}l9<1Fiz7K`me9t^&3#&T9abx+NL4_T{e$^nQ=|gaM-a8>XR08F~J2Kkj
z(}>nB(^0!Yw4A_%^Uo_)Z5lEu6@}4|%(u5dPyVdg2=A{Q-bFl#97ffEqan=h>nao_
zyks0nHN`FccluZ3L~rw*@%c+Tc_W(6afi0FU#a+S_h^C?!6c|8(q)*-@sUTZqoq(_
zs9V5hXfn0+Vd4ENfH0n##c-kF@Sle*nr%Ud@#&PBdh9eoI{3coiTNSdk5kj?CTGU_
z<v*#n+*LZAG4JdaEXccJwc#2@oiCJ#D;ybBm*3qwmUgL79Cl63Sh_`c%PkRetO%(w
z(u|iRDK~n8d4nD&+fSXP{?2?jc`mZlD4f4v?{?(N4`?BTixnv?E~L=&3CHR>%!_w9
zTqE}@pX{joJ(uZni*cp$(H*zBP_u-cei2%qMEQrkWWQNx0!e57bHeY2s+2TZDbso|
zQSJJ>2(Otu8Fwgupjc2kpXb$OJB^19e|UQtgSUqr<w4dH4chCPc_7DbBK1hfd@pmc
zisX{Z%_rJjj~;ulw*bf_%rDg$?WH!mSz5PBY-82QP$X7*l<!;6+sckqGRPn}_h>Ex
zs%j`?IO?hc2^aQp&GOE(mRr2sLK2O_XcLI2Iz#At&$$$zrs_01@cA!K$*>%xJhh;D
zL(0<<pC@p$ITjP(gRHN9^(+7D)W^z|+(v6XLm#Qp%8>+df{ZLR6)WoH{o8SIPvS&p
z2uF2q_tBWKxVt@2Ci~f4cT!QDxi2dPKKh^lUhh_h9w$((2Iv41_I+n|S{&-`Xw3w_
zfY0Okshflh2}3q|pPV>zO#a~z|99TTuo>4iur?4Hk{+o{ai`|3mRBHQ;dnlQ@qUFY
znbv_@mdmz6W<x1Y3b~WE`h^ff?afCqw_M68&O}tw4AJPL;JEv^WlVXc{KLV<%<h)C
z1;ZygF)<r-fD7u@_?Ey&w!)s6Qhki^R_7D4-f~!KD=qZ;{xtXV7bCVvm-S&Lf-?`b
z49gF{=`i-w>*1n$d$it%^);EF$8Ek`J->gO?>-v#9dqj~;=QV-=|~{;rh3kZuDKd`
zuUp%q9mPXAMIcU3L`n|UnEF)Ry4ktTSA184@3Z#i-kLHo0H5Brrt;0u=<TE<c^3AU
zapBlfNY#-+0nlGI$#;xpZ@b7p!pQ?6&0xdOQu0PS5eX(5l7G~&^wk#^Dcy08Jb0!r
zQ0VqE(@YgtMS&7eW)eqR9Z(6??#MDFt%-Upr6SZrYVZD8g-t0yX#vU-*^Pw~S)VOq
zrO6r8rv!_!K}%EdGqK;h6%^P^l|LV`K4Il=Ev@vPHcTxvlz!@0h;YG)taLvcgn#;K
zRy!-!UpusR4t((+_ie%Yal4;}k(!TYedDGSh?{7h0AJDZI|m83W5DN85{uP&<NbKw
z$E)*cs)D#X2Bv12#clFpU9>Hg{_fI(20=*^gjLg_jmMM}IXUE=g^)v5g6&<7;-yUv
z@taRHWg2fq-qvdoGYW%MlEd2@;K^EDJ3Pk^q|Xl=m~**-zNeK&r13;tIo(>|sd@u3
z7@J=)u$8`+GwY3N56L}B=ASYD3Q;E~pRN5|e<j1^efrMkN2#@Y0&d_IsU{NKjUqtS
z#y3a7d9>gx2)pl&9$^oR9nOPZ1)QP-Hw_CT76Q7V=hOgZ1}Xc=Yy}pk13t}Z)BLJJ
z7NFQ;b;K9Xr4smc?5(bUkAJkqfM8=`qtmuqW7%o%U_WKviTN9|3fVZlZ!r<N$^x_d
zI5lDE>&G1;^2gu2yV|vE%K6dos2Pg{1^*pmIQ@|htRbg)k3Yqe%}J^ebRd4v6ylsy
zoWh~}Jnwo9ur4$GI$mKr4Qs@3*F^}CTc};9dXmrUZ=gg&p3<8u2N7Q{>^*&Yx}6Mp
z<T@2WfL|WTG`sA1>2ZgZ%9))~r5qwEWtDP$hKH-eQr_-`{`*(HC7`EYE}GxH$!1SV
z(RIL6j1Mn*aD<J+<Ibp1T*|ij$Bu2neqnjd8lsIlA`DB<bFSs;$G*!qYa}bsOqEc3
za5`YDeVqVZ)10x6-)ze1ETUF^h^Y)Lm*sj~hYCtGxVnu`ghY)#CWqkD+u_Lv)ep(d
z?QX4MhCLJ2<?OVbEYRtVXJ_|RjN&4=hzm2m(UHQPMQsOCN2i$Ln_q|4DX$!M&bbXd
zG%9Q9w8(z9=FN9Gy4pLSYlj1?75QBUrlY<ToX2A*jYt(LFELwyJ6-$|e_fJ0C9Pe-
zS!^^90mVb(8Sw$0x|?m5y4`py+;}7tW<uiGhBwx|4obV!t;iwI(=LTfeZu9kRyqe!
zLTU_RmC2_M+ZsyoCnaY-b7j^3711sFD@3|QuIXWpf>WvR`C@=63go%jpPI*1y^*v_
z^wi0yO_=3@@)dd6?ZZr+GR@?ky<GSamNv5m|AcG*W;boqM264CZ&L2ONv+-EMa{{1
zf_j#%r*3~<C~*?(^4!Y?h`zCTGC^X7=OJLE<)c7r3FFv04DtlFxW2VPmpS0}46oRB
z%f1<Al!xRsvzI_qM<ABQ+fN!4SM7$S$O(>j|9rkDQ;L9IW#3Fcf`pSmEolA;1c>nU
z-OQ8yFle&X6<zb!r@!I~0()glP1kuhnYQCh0s?q4F?FH>g-0*Gn|oo$SzPeU+z>GG
ze9=rW(ri6Y?n9TsFWh<WsPIzx>$g~&iIvxXW~2ag_LQ$+<e0<TG-gJ(#s&}K=MJ;<
z8&8Qd7jv@jF=Ky=UFCwDAU`5lkNY|`5VP9q=LZc1nfD!<lae%E;#fNPF7y|%K3ee1
zdo>A>LTZ192$Maf1>1J-j{k&sobb)QIo;Sb-HzLo1^^}ptT?eVb7qov$ZX)jouTT9
z8$X3HN>N7UJMx?6bI;6w*C;Yec4{mJ)5M00NT(gl-?njC1gjPZn0(}ZUBUu>ufL5>
zj+CMGWgy*t?4*OqP6}<`q<vX(Wum^gZAs|-HnsD8?&JIum}81t7%~hho@5{tLF-3a
zUXSI-p>Muct&C?j-gf@V*Ly_wA-imF*Zl?0g(YuY%bJ1e6)!GH#M*<k0{c7BCt5p`
z)Jj2bjzmF1=s?^L5_ATyj=VyAJi6`P$M@J!j~Qs5dIkX~$mX|bXL1B7fdMgvykm2v
zMwQX;m{r{!dq1aB7bmzuSoMutPl0ImE~B!p$YYVc&TOQ~UQeA}e}tPe&+n5Z`MOW=
zbO6%>xfU=yqfR5f;msqc9vp)A;QL9lnK9ha?0tG@Npg8P#g>|P#`~4i$Onuce(!$W
zCu{?o!RN7PQ`*A&P}QMgv<O-G@SHPQ@M^fSr+E+$=kj&AXCdZ3tNwl1H_y!B?E$Y3
zwba#$*jvq59W<W5e!8ntRN}BW?mhK3Lf_%1!^N$B!y*nTPcBWtF@S@j^cy{vjQ=%V
zz1Q(AVD}}dyd=-Ou0F$pDDu)@Q~CqHfzmB+D6a7#=+jFpw&Bx})6D|VO4L2&Th#{7
zKWBAZcU&CFI6w1PB>2~i%wfK{D2;Uu?unEn2e=WF@n*3Y(YuMY2Qn{N)8@nk^kw;u
z?YaGG-#g=}zT-12GLg>1BlYqpgE`IR{PDI-aLvfc=1+o@2=eWQ{`d>_$g9N1`bF_O
z2EH$XpF)CbyO!n0bq@3{3_XNi88`<9NA~7P9jCs*axY?))XTAD@t=NpKe>G&RE8yr
zx238AbT0^$y_3BUn|Ys`#}4&Dr8jf0W%LZ}+yNzTlcQSAey?iAN~>e%R6xc@ahDHD
zu?2kB$A_+k=Nj2>S{4fb^S;x~<sC2ws;luHZHW$iPxb$MUzp{`f9G?5mPF#$2e}S!
zmRGz@`#tWe-pySsE`6!R4iwG4>h+rr2q0xhurJ>CS2|D}T+WMwb+1Yg`uE+;f#vTZ
zefmsWyb}q%a0{DrGALs5tXl#^eTmPn@axEU>gSH%nE`R1Ykd-wCIbc%z!#nv9l<yf
zHdZ%8lWKxUs2J7$SW3hLWY-Qm#J(a8vf87+@P1V)O-${JaHMeLfX38H*-TzyEpuzG
zUNU>WP1QIZ7|4GxBeJStOcMwL_yYc3Ske2l<7y@SU~#&yNd`lp12kv_bfo~LKiB>s
z&2!>7r(*u+wEt$4epxn@66XzFDWwDYMKCSYKIb@U)Rq=xLw;&ADHCTR1p5$n?EiLq
zgB|=aM!gWxe;zXsG^sNI8}vPj^K%P5`s}Re_n7^du_zNyn@i#A$d};N3=(KYrzE5-
zvJWNbgHiA@c=?WqO5|u9oO2oVxIK^f(rt&@?Fg*RtJzq)KI6NM5~20$!&Af~ZNl;1
z?v8~RnP9L{g4Ij}n@NPUlV}nIg480~_BwiDA;hW7!1BuSiXeMCTcgcab$mT_spj7G
zzkjq`4O<P3qGFmEfPZ4&L3FjK*kkX(<qr*5tcdy!EL!I5%4a*?nwgDp^loZuIrbRk
z+#ps6VTb&10FLI*Uca%P4k4VAi+Oz0`s{f<wA~;b@XU*8_`1S_ZlC7Ts<?C`%QpVF
zmEWWK1pLZ)lSbRsnja4cjaBvWn}aqDXV>O<(~K_g^61(*$6jx_+O!eSk7cLzL)fUf
z6vJ^?Om$qf$9Za*&FlTF5Z6GB?zHeHZ&&~wK;pl}^=vYnGfaeILVoSY2@AWIA4yUy
z)XXKJTYLGEvpwBMt#=(UFVAc|LGBL5>D^h@U^DYuXOaX15LM5xRO&`svov27V(X$L
zSC^Hx*bP4o!M#^v3Ws0f!5yh|fP)TnwcVqU7=yv`8ubl?+nU7I+20P4_-iTG@OyvW
zd*{vx?$2k{>X%~rG`}e1ivPXwVfXH{(k31{kPK%r)FdTeu@6diVJl##0~ddIKg`lu
zC~iXuERe0t-t)3=dK(b37!PtYI?p3Y2n;}Ew5qf!K}8*oh03jp_zmZwzm2k4O`=;j
z#xI;FnMJAY_zF6;;lH_@r*_u4`3H;LV@d5PdL|ZqWHaPmsGLH7#Rdg_=sek!`)W6j
z9jTaR`N{cGqfgBPne%DAIO_)eoOV{p+?(aIeCZM!9jTtyY!nHbAQhmQbfeU6yqj;2
z100V27L714UN(}PS@=@UHA9W9a{RQ6J%$l)9#v&qS{Hj=%c0<z(f4bs<wWAQ(BKlH
zLC1fg?9IcW4*!4u5s`IB_GK!BWQk;-X+si|5?QArYxZq0cCv)AgtCV$6_f1yGImK>
zk`c4XzRXz0V2q>BIp;dx>pIstzu)<L{+ai@?|D8S&-;G8UUvroRYC@Ki!}MkOX7Ed
ztgr!li%tN4+^&4Y+FVTRux$Y;mbe%js(_ZNs}7OqF`E|17TX*4V&@V85QC!nZVYS)
zs22haixOT@wm3Dv2e(wD&HXxqv2e(U{OnYGr6)feVC=+LKt+;S45%<X`!RrfRe9gP
zet4^|Bh(o<o5bR8*WVyX;tUk{R}*?@M)ruSYFD7HfT-oFx_zln_vOkgxo@qrynmq`
z&?ULGo(SZ!v=03R;nj~<6|_g<V_z-GmRfvcOER7pfAK=c)YCYKLoonVeggnt2CUy;
z*L^hU_nMP!a8p%l)lKs!cxTV&iV14+ZG^230X4&jAC27z?JvXHz{^bR>g0V5eXcL8
zG3H)4285^Y6aw9<^m~s19NHHMEru1UVI0ev(S6s!z7(%_H~Sw@|3DaRbx%^?ZDMQ>
zEWNt}SlJ4CzR381e5LFgW6hZXGteFy`MtRL=v}J?(JIg#EI~7x^s|?-*l9b@?rNRv
zHSP?ed;<l!da2;Tcr9~4JTCqV=Nh&)>`l(u{k<WG<4Er(xL~7b&Q)+}Q^OIq16RXW
z=ar?ur6D)I`aavG$$g;C;zoG8@--UAX<9y%78#tur9_RC8H2pfiCCMmnNYbeM_Kff
zY}NgUphD211~>;<A*G&p$@epbo>A)mgKG2D%bj(<yXm8*UhbolyhmqsmV9`V{P%-_
z#N<rBIcHE>g_*gGnqnwZzNhG-nCr$zssKfIA|0=|jwKJ{|KhlB64m8t79};|MGmJE
z@gWAze~$qj93eh?lXmKjBfVJy@Y2WxnR1!`G;s8ExTQK^zruRGLI>*0ZZ!S*uFNpA
zl^dW{jYOE&PxZq{kAb3wMf7iH;c7<A9^BMTvzi~D<%<^#R4xxu#iRAOj{#FAD>974
zHg?;lM$%@?&a|Rb4%)w-pJ-Iq?i|hTV#j<5)A;h-WuJ+K_P5kvNDr&;$vcYHrt+={
zie&bvpQp@$DVuvuRWlYx1v^GCd`B!oS#ZYIN%*C3$g8rmcp{Hpd0|9i#mx9t7fNJI
zTgbH({^1&@e;|#8c#Y6+J#Jn3fx4c7I|T>HJ_zC#iCVq<=)HW<?##TGPhW|z4hjt;
zwL|I%hXS;V5GS<THe$Ao%eCehkoe{pP$BZK)wBQ}(0>fz1p`-pju31Z>AWHQd~TTJ
z#H*h1QJMCPxm?DkK*}d9U4#}wg7oU%`LWyhst}!beS1&sUT{*<UiC^$Y@7G<c5sW|
z^<vV_z9^7pUv_jC`>#`U5PFc>v85WqsOv3K074j0jmX$O&C+*-;5dXLSI2t?SHntS
z3yCIAu!r2jzBO;IoY$Zj>j?H0h`(JEVL9fG^3^~HqZ~CFnrm8<ovP&j_;S5d|IP(8
zKC^NLWNAK9AdI+Lf#4xyXfRSz4fL5T54_;+>zMqDH&h~~Ya*czXwk@hz}SC%-h7&W
z4&fUNs2JPBL!vD?GCb~C@K7t-SKn{_Zp=pc4j&0dCWSd|vYf;#dI2G$O*l8G#A;_)
z<hk}?J&uo&>Yx5fdLv&GC(e5Us$G>~gpPQ5`gw#^qISjhGMG=|NPLj%j5Bw7raFuB
z;vshS0We(^4wkZ_%5@<LjT@s8(VfnpK6@MrvXo^@94&m~s0xeKQXdJidv9=f)3E4w
zdvd6F{hO!IuOcdn>3^TPO*bSp_Fe4_VW?cP42!=OTq$Q-lW=I<oY<>cJN(8^r<#U*
z&cL^AG!h$s46x_4G#t@TXoZq87u?3D=iB;Ho@@<;HEqCkxPD!TV=*`e2%UughiY^^
z)HP}ZS%UasG`@)H<gTrLK&fFc7Zx5|z}(SS*t*&iXw$Eqq9G_{6pZW%!ziNL&rx9>
zggpKgdH)d8Kuw*`LJm#2C-!Et4Jb+K#3c(<oH`+mW55{&8gH2`ytr4vqcxDgNJ-LL
zc$69OWd26mEA{W{0|3vWaP3WeKRPV=7{J!5$Q*?~^kBOALgZTF*0mAIs-TtzpC7>h
z0FLbhABj?s6}^Am%{!du7`FHR_mu_s#%FjHz>o%YA$N$=C-YfJA-M5t(+lL#)<t_u
z_vrRO#c4xi7D*(o{~ln!jB>(LG1VvEaUo6{aSgJDty4?$wK)c?`ipwZuD5`a5a?g+
zeboM9+C%E}d-<u?S9VYUs`s?`24-k^@7ui!^u1a03Z}fbEF0O>b)FKxI}gA2Gnm?^
zFI(P|mpn|%ZqZZgk$2c4-X>joS8ff1#IDo4N=8H<;7S5G#yEV&ukOz?OhjzL>iPEL
z4<8^H8^=1H+gQk7bRdf(bvoi5>_N({`e%KXMHMd`ROqbkNZ>|$UMnVoSZPv3D`vBX
zscXL#{ZX_U|C<ZalQ8cBKUKeqp`3d*anTc-b~XQXHWcG_1JQ4z?1a8gUurGrraW-T
zmx7+MI$iUwTaR>91Xc5Kw}u?#HR-TYP<KF}T^KO0U7o~MQ&L2ly8e2xcXyBta!GYD
zj8RpQY7MF1>)%Y^-xynJcD;$`RQCGj9p}$V5hwaJLyqcM+Cm=NE}YFhN^>#OUKvB`
zWH&8l15Z?ifg0uX8W#Ee(J!sPPFIfsr=ALIsqQ#NsIuEYyML)n|3L>Za4>zjD+c`*
zoGjO~yBRAkKvuW==)^DJt!9yZ`fM21jsv0@3q4A1mGXMVT!vqV*^*wBtlZ7I->#)E
zEyS5gdus;AIuA6hH$4fpbsP$j&_%3-3L)SmgVES_$tUoSKQ#{3;^h4Y=rg3Xu+3`2
z6OSAwEJihw1G+Wp&SYt!wAD1;-Pqmgf9tjT%vNJAgjW}C7IEl6Q=62&j>-G=9-q_c
zUn^XfUctF^Ke+RtUT%`6MjYj&_A)vJQF1=bJMsq@$N|e)B*=xFEB_i!2%(n|=SkQ;
zPNwDNqta#GOPB*u`QDnaK#~2Zl;Gry8Iv2Jt4kecY0nO?yJWVDQZrvn(ZuA=_-+c^
zNF3o_OHYPEA3k}CYh2#R%=&1t8Whq-FYx1A&*)~o&=`R(NeJTBXL_Vs$&c=_eEdD<
zqXjyw?9cmTSJ!B+<o0Oc{gy{0P-4f6BR!%u>Fg*03Anr9dh*mme@kn{=A*aniWl-G
zx^pn&7<P68Pl$pRW*s)ANB2XKh6xR^V0mhYst3?E{s%z1E}dPcrU0L+(zQBzo*1cX
z|8I=XZvT<5$D;Myq|^~B`+r4$JVSij*gnY2Yy4Dqk`J9HN1&aIP0EH!X=~t}z*C-*
zbmEtN)5XZbEYK|KI=P`IOq`B^o%4f67o6TzBW;H7WNjfV$ZUL9bU52Sol?o=*-C-2
zWF+!?piY<;HEoBVS{ugqjw1x#_jAt2@o!r;pRzjqwPF7z5tr|;!0m5h<GX~l5L5aP
zVvnc+_IS*I2#-53G+7elCJCIQz~MnSG7K3Ilz!Wjs+nI91g7rxHtlTAe7gO{=tK$j
zx;J_B1mxY^yso`yS`wN^axNM4SB7lr`UcDRU#EDjg$52Ex>INHrziY7#M@cnv7MPM
zJst#kY5FB$r|0nDDpTN#)X_rJ2f7%|n1t#@aiURFh)vDZ&O10S27FI#&&F6t<R}w>
zyhk}4mJ87vbG2%`*j)FH!yB#4VsI{z?}awEz8#E{=703c4;j;bv5{JRsC+M<k8CY5
zGC>oJ6}GKAGw3v?*qj{1gby5g``iAnCM!+yURp57*u+-<P=)>zBd{)u1Fg4d3XJXM
z2%RpidH)B5b>2JgU3UI#EQ{+?M<#9|3;il{UL!4))K6MU|A#E<OC(?D_YLAZXJT8u
z6!TxkIf<*N9ZVeq;N#BYVU=NfN!MG)*ZI3_<qmJtC-TLqtenYj>jS{D<i5{$f}k<A
zg|;ipgWZ{cvk3;B>Qc$KUUl_#tjm!xXi6aUQ3d?tf>i5-A-j<xRGN8fefw{Jzq;U+
z--|Bw8oi)u;2JW3v>}3g4LWCj=s|=guwYOM?h{wXF8YbM6^Ze5<20=8L<Ig?*Vis@
zyHK=gwsOC#o@^Wg7)M%l4n=8hM`h2E=G0B%Oa}wpop_}kKnnis@i+gq7udpJ*X-=>
zN^An<L={iWTm4^w0+W@lSO?k$Vw%A$Otj3jno1!GUDUq}PACjEAFDNgZtxPZQzqLo
z_=n_XM*$ucsLtH>@F8WPP4t4NF-IB4fX&8)hE>}O1P|(^E8>hzIo3x7?Gmn7_T4Xt
zt`D<+>NYaG(drgFUw>=ccVq)V0jhht=k}!31bjmS5jsa75m20BoEerq%n8QTj|&sB
z=5BUqV>kxHn^mcVn{3^>Ixv5>N$PIWm$DU0EcGOqFm{7#v(PywYZu<$<t|_P@K-}z
z@jAZ>@ozRiou8#aPw3$pN%Q`&ucJtsIcc>Ya|~D`lkY9f!duW+<7;(njhU-U!x&{D
zff=4b=}Dd2BD}*7^Xl}QqI;V<93O=y9%LrSRN4JjZIwO&CeEzG{(l?cm1GYhBsP@G
zDRR#ucQ{b#;;2y+b8zX7$!|v?yUsmzmY!BCaw&`ho;HQ1bDwx@bXMPZ_1%}x_lqOq
zAlFL7ujUdZBGOEKY^J&)(<?>-R^aR8qZ9d&vejw4K-+J3x}w9+x14&k2GEodO`kSU
z<z4h)m3?d$EYXSZA({8n!B!oK!c(!=s8yM8_t9v6{<BbL&?AL2QbQ3bq0d$JJo(#O
zsfJ|6o74=Po7NN3(%ma;XN<A|r9US)x=lT=lM_Al&fpT(To*ubV9*4_K~a2|%-c;t
zdEr9L9z%JYzHrI^&X-Q~Wt1cqGG#0s0~8kDhUnNr$zha$A9QKs?2$KPaMwQm-@!Hk
z(0xe`&A@P;*5G-iHLtFd&{c=*J_d08zU?`8>9t8<v4vgzc(L=}3$JdR=aBFjYGM)x
zsty7%yaWic4;$OT?!iC1Z}g}KIe$mL>RUO4BmlVwfU?m-$t>OLF4%6E(97QF@BS9>
zkF!3`a2CM8jqZX+5s67dZBWqR9h&P;`bBCfDAvsOQteNgEIVuK?JCyWtbBJNj<N3d
zW6RjMtWa!|y|kb)uq^xpv~H8a;<c?@KT}^@zO%)4W3GrijZ`w0%DgYz+0}R3^RrLP
zqVXQZVKU5d#CW`y+etI5SBy<Y;Ls4-RR8UV5}-Bp!obrrlA@xFL*(Xg9g(|7Wi&|H
zvRS(@)dXkb)0^^J?|ii7%(d$SEZ#jTj=5<=%Emq;)`u+XoMZ`zbNkOw;pH?A!Sy%R
zv+LkvfYNBAJ)|-S*$qhm^H6i}<yG3!+E=?y#iSdZsWx;jz&rKNZbm8t)i1St1dur8
zC~Bn1S;C6A>XhH8*vjoq@CW}v0o=y`VCj~o`4_5Hr>zJ{w<EpK$ucx0`dQM4=Dtz-
zHbW+OK6|bztY2LF2VE2%6yn+=KwhTscs1=ii+sJ9Y>^CP{E;7?4ZyUtM0Wpj&3RW`
zUmwGfW+WGL=Yg_CJPx3hz5nUNAq5bR$g2-KNT}{SistEdyGaD>a>k9u{WPw9iBhm#
zZc_2W$6p8Li4fPE+SbV!Y*ag+#H+|I<C!*dmJ26{9MB%RM06)^1B-l-9L}~}q_0Z{
z^3j^^Ij)S7W+VkZJuZU{1MM_#1IaWa)n;CBFsYzxlM@c@-m{eSHw@(430AC9&|7x*
zIU@cJ0QjzGEHndIsg~%jP*Ba!+JSHWnH6;b-rs^U?f#@0KrE+<+P^9#`s*jm&#mnv
zDI+AcxNGL5f_@-_RXoCrw1S(hWh=><RgYA-@Y4In!$r`aK(9eQidtt!QmDvLAzjd^
zU7b`nkYhAqH{971Sf9=%8Tmj{Qyf^{CB(!##4X6JcXNdk5s#wDQ)5m9S8oz5O=JD{
zkB{os{IowON%b^fmDW(PWXM$-ekvoS5-!HoyG4HMD7}7RTXX5L;h)q416??1ZTrxj
zcKN78zNq1c-~1TD7;J^)aJ^#%>)G#me9KFQrS(O1YOc5(R<2LyglKW?3Wxv_OK0Ah
zl{qc4Spk0Lp^9jL@<QNf$%87}+iM@480*{H>&LHDDJm*ltgUTGD$OPYu!`O-zVfX<
z9Ke(Y2beEP$F%B0)>{s(2FW@PZ<Dg{UFpW8hTP~&#R&9X)=lVBPaNsS6ZL_oveZq2
zeoq)xW_N5!T%>v;R3vnCSh<aYg16hgC!97rjB7f7NgwdyrqJ^qCXgu=)Fmv94r5);
z7bdMo2e~Jec3-HkM@n8!E=eE0Dw)=wW+0tzo6pPfSL&gaL~@&H)|CJ?*Bml7{90IC
zv&3?9TgThhw`KQPpL%Rc_#gP{td<nD^}We>w|Hs1S^G7<tS89|2TNSj)*^G^{y3fG
ze{D>@z*ME=`E^&KTy(2#s}<=>jX%_avlG7Ntd3ljSCzd1-wlzCR!N@#<F`5{BR<cB
zomm#wbo{A0EVIjbedW5vh4iEGS1jdE1DU)z1OMh%5Im+CO<zMN=T>{LtW{FeuH~n(
zUL~2*XK)Ou>ErKL8Z{nmH#TmW)@}Z$6VAi#0o9Qh(4}1;^ASzHjW*#panlw14>Tt&
zx%AA=PLN9Q`Kp4>V?bSORI5SQQ7VePfC(OR#^BnzAIEvM!jy2xW?RXDIm)SdGB&RE
zlbfMo>#Jh`wCSd#+Vu@Fv5=d5uZab(Yz3(4Pu;W)iY!;+1Y8U*KY!AHc$Pz%&ScK>
z!?$`MUGJ;j5V?`#bJ_W*A5pf3<SAli7HQ)r0^<B!8%kKWK)g0OCAG~SOH?kOzc5Jr
z<jo=^vAB8e$!Y;}sH7EK4cCyU>gH(DC{IChjtbjY{bu}bslU%#PUpS>Qt?*t!1X8A
zdPL-3ZqFk0$JSNK^&KRBhV=@k<&@h$nikR@(mI>7o%{p<0Q6GbaI9et9SQkDlVC~b
z>kpkVJq}trx73tx7thw%%Q^mKa!el(k}|;bwXZU9pV44x^8!o#)RrYL;@9W$D8p`c
zdL7W&+39;<1xrLfed54RP{6pvvvS`~epf7LW$rAQbE+F8I&uu)J(Be)fpxH#Pe*L6
zOIJJ=4d^_-m6QT3Tovg9pC8kf$A>^or#>WziDQ^u)0H?rzwNRq+vNX#YEN2h^=S%&
zB#vu$YoozOZQYT4`;co$AH?3O;wesFQHhz@(7*Fg6ZPN>{j3!=EnGXQtDTX29a*kV
zZa~x7>$AEk2XCe_^Mz9LJ|s*eio{J#*qu+rl|OB{(U0liXqRyU@wcjQ)fL*X7O7jT
zW8>XQ`tG~8#qBNyMe0f?hDyRAcyKz;lwYtsKSsT~%e7#R&m>$T-n)m3h4EDTt>!ke
zT0;AUIru70Kl&u%Oc$kKZSGh1JD2kePP5ZK#MAj4Bylm1(r$offBCg(p_>s62bH_^
zIWM?}|JV#w_jQ;(KMl8SI<%Ns&3==uhz4=Gi0e$oG81GlBI-V!4&Ew>ds_O%=J^N?
zYA_y)Icwop5GQ_lvekkR%<`g%w0<aJ`t>}HAf+Lfzxu^7vI$k`s|Z}q7aV^+zLNx=
zxnGUku)^G_F5K%B3t3`Z{7dB_ix9Tu`+Ffm;DQew5K<D{ci+qI@GImtb#oQXwlkvD
zjpaIdg-AA=6*J0pak24r;?dD|()jRFsorptbkY14oDW6TB4kF8($UJRmPfx|UC4->
z)yQ19Z!hWk=Ws{N;8g3Q?l2CT!J)1zrTrQ;k8ycUFrsx9)U6fwu$5rY&qPixmS6g`
zmBMwqM+J_glulOq$?MH6qB}U=`(HP!+=5oDLqd>c#`=tc*12C+JVH6CW~<gDdWCon
z-)D?$4^QWNlL^PCUdjFq-+<PO8Y97qNPP02<)3$6pED4-e39dFL^wb!##?nq@Sdk{
zx@*|B_Rx4*jzLD9yylGgc-3dcr-nIN<@v0*lo$3%5oQl{P7w4Pw%Qz7c9apS><X!Y
zC04z=#;Nyr7s{~GARnEr@g%c$km&U?Mg3c(amzTw#QY8YC#<F;_88#Qak>im@$P{`
zrK2>)x6*r3AM;XL@WLC?y<E=W)Lh1tTdAd&V(XK0iex4z9yUl+IcoI8eXtGEuOW^h
z#>4ojb)9EArGKoJVLYD+&TDn&L>2Q*?ZKK~M4AOmh6|co6$!Jzeqs6cYW(c;ew(o>
z<|UNx>YIE4!BMX0s0s7$LkbEb;vbzq6<a3Ae!ZL0H+FX@uCPW$6u|P(=x)FHFt(v!
zT|@dr+hSSq9&G#*;LFruO;~L0u--_I9yGP8`u8Qzl3j#9hn`~pCRL^&x|OrMfSD#=
z2N3{Ynq81HQoIcg$!j?V1QG0s?1Pu13*NlCihSz{tC?EE5;)cR0<Z#LrImOB>UDPE
z1$z$h8NKjGCVm{Ea9zs4B%Jl}z?ISQThX8H4ciH4@To2Ygw=q$L}1f<xVc!o;Pj;l
z9;07nzEO0rqGVTU%^xmq&_>bJ$Dg_(SDqKWF~2?N*5LntKeY|~eGI^(9)02@cigs3
zkU^c<NvsU+&X!In+)}L+N1Me8P6xkC2k0~~+CCz{esAMr<E)37lfW^4x$!$WWpPn!
zl}*e#nQc`GfPE}oh)7V3qI;Z1*iJGTM2ID~31EsD$Gxrq4y4YL%hp+NK1p_6+0sd$
zi4fJk*FN(@$ABnNx{3MEYY$jP{jv47Q6G#m=Tu0jcx?fHPvFew*LuKeJo1Qan~wKL
z?f}vpzc_;x#uhWa^Rqf~-^bpa$6!DBPyMS&Y8Bw__hO0yp8F{DB#0ZXx-chmlnoYw
zw*=c#Z*{X;6<qH!Dt4UO)4d6?{=riEiTT0MO(uJuxy<$ekLj~e&a-qzY7zm)#xu|m
zNTaGOe>q27ZAMmy{+wflG_{eiJ?xLwSvA?MB8!**mV^K1%!Mn`jE?d))@irq)Tywb
zc0qE&i_*IL@}-3;ndY|P4WZZ52LZInA*vQ7YP`;T;<NTMZc<kK&fkExudVw$^^q1u
z(kRjOd15Jk<mE8{ZMsV5WrA<xVR63d<FReyQ|k{vkJ9*V99%`cIR-eNED-(PFn){C
z_h@>A8!N~;8<2i?ttSO)^YLm#u6^B$hoTfcAEavPZbSEGLJ+h!<Sg;QIAgclqSK3p
zx5%p+T7twH1>lrKu6ZJT7IbbG#R(U=MGe69dr+b660oYM(m?xHuP$A#8A%<&X;y9G
zx7ouu;T~PU1ZCP4a$VbY&GihQI7Nc3j}Qb7+?1UUm1ql^3YUN;LP(}sp9UqZDkbov
zQDasY%Y4I}Tkc%we_3npHhQE2GgdOKd^FJesFsHO{I85DoDMq5X+P_f&rDh~oS)Vc
z-`IAaYHNO~J&m1>!-cWMm~<!XG|N<>%l~I%C5?K7(!P2O;D|requNoTNo$I3c;Go3
zg?M6%%(;DwP`lROkk@~&K!xD_fmRS}{fk2hOI<(1FAsU2+2zYRd`8oVB6O^EkrmNC
zcP^#5@e>xNp)^`&JA#Ey*UOmisFqItKz)5EhtQbZkW`+%lyg*|QuE?nWnSgh-25Gp
z{HD<$VvoMF=w$H(<)Eqf7%*TqPH#qSfAsUwKL(^!mL2)Q7OktPsiR^)mrPA`gZhj9
z>tJN~xBq^yUuN__4*IvDwN@ZGPR=lA2vw#u9^W$Qiadjm9v84`Zg|e|aK-zL)6LyJ
zsuc?z-blX)TuMU79|M>c>ov3b3NXbzY9h8<Frly)QORa<j!mXQ2TI3)_Rx$yqaYr$
z1pPttoIDGIJvN!5^3MO9fZ+?MvM0KEi4+=(Q$!wc=pt^D4xdCQibLM~aULJwA^0pW
zC>M&S53a&|%wLCMY{jAHLI=YHuGq8PqPkF)JeHdUvtO*Q3KlxoZX^V`*EoK1`qf-l
zyL1pT(rlxvq5g<h=IQiK*9iO=5Z5m48pd1aOj98dWlxD^%}GAJ^8n?=vk@XvK?iM`
zk7Tp-c<{Oua3c6{A{F?Ae*JOz+Wj%P4-xJdnMvk-7(j7IiA&RddmM`p!$IPN7h*gd
zoe5%~u4Xo1<_x;iNc!er%2g@kUTe3Jx3{;;XxOSkM+?*XFYK?_kBBQIR8%|D;OX9l
z53Y{a-nq6}J$h%C<`W5(B&qz=UkR*S52o~!tkq(nq_+G@V#oth{eezShV;<_Aa&#8
zWG@v+u_JksrAjbMZDgbIdiyNXMh6bhuo$n@wVtW*Ac0r4NcQR1Mk0naW0iUkpBJaP
z3(_L29!=^%wFGnR)01ko1>v74k<{cc0u<jSf&Z0verVQZD{${JWVUN@x(H;6fPK9)
zcO}CeEI~a6*pLz7ulT(NDZ!7yIo??62Ik;m!UPU{XIxvYEOg)`lcw&9Dg4^(6hHq~
zKC?laZBO|AY_ap**&6a+&)LTjdz*<60pE`3Q05%lCb<CZLftZq-n5#J(vVC)X;0Eu
zybQ2JGaE;!enRL>JtYSwY3ooNf-CIRuSN{`>Ew#?^=tGoEW-pszwcddM)g8i`iZM~
z#v#IUm^&m#%IBMI^B12BXs0&X0qkjA%%0rbz~T-^k7!J0!W11$et!r<YTrJbPJBJ`
zMq1>V$#0y3{!>2vS;@nYa}QWzX%CKao=?_|QtxXsp4cLM?jmfRA1mKBt(}np=+_xw
z5)Y8`uOzKfQZqUIv_;6bjm3dvZ|>AThFjBEe-7kNgM$p|_oG8P3+>O2N)TQ}w1-W{
zfEY)77|@D}?M*gbO{JF^oX3|aM=GJi_53L>3*LkW74$)qF`#Y{nfDnmbF$b1Kl*m<
zep%yl&ysMi$x)#$=|U#5Z5S>rB73g~%RNIS64^ie9itrt#AMNRu9)T~t){UqXJ{;`
z@h@plq32GKL2DU=Mw&Unhzls@^mV46<KS=pzC!KSbFvswQX!(Lb0|h6eWRd~CP2(%
zCx}OJRpZz7#O`(5s`Z50yRWPSnu$j<-Vlj<EgbWJ&-5uk2~dtY1`zv*^rnWbiWT-_
z0F~T$45)IW{i`JYO<s{HYHuWU|C{hCOdUSFh)|#|tISc6=u~v2c=oT%W5A16mTE^^
zNr`&WZ<J~*<tFtLZXh4q39H2>?SyEUzq~Ho=f|YG!!-1<<MGXdsV_8}qYv%EG!CQo
zt7Mz}ydd)+=W2)G2A20*cBjL2-iGU}_N^TQJUF3!@1>}IBoG0T^wNGoi@MX7FgtbF
z_0wK(=et(W6z7Xq9h>rDtJOb^CL5ZcWKmp?3Te7+CoAAFJWb0Hq8)U-fTI5_(d**<
zmM}m_(61@<=5G<aN7QaL$~DCW#!b_EqUB<%(0uqze)TWz0%cARY1z|s^tJs06|qm3
zJbFu)Tn=)2I2z#kLTccKPMGgU2tVHxLjI^Jl$&Nc!E0Dg=7k4aJ8m5VzP9QM{4T}j
zzU1iX8_B=^DSh3<;F;ov<uPE&yL8oRPG?0X?aY|^PrD?*V_iWumK&$R335L*ZvF2U
zRj@xCcT>#+6Z#G?a>H8DSB;nkgd`s--DD|8>HOD0_&=+v-dERW5*D1+*9i}|4CPMW
zd7x|{9S+#EWrk2c{n+_}W%;%r`-IdV-(f#uOSC2Bnu9M;?Yu5zJ@{dmm1DG<cF|`G
zxf#R(xj|S=2y>FRBI@jbfz<0#X@mu-bQ7{D`I^b{&+Ezp5*H2<iS+`~8Bsm$I%Mf}
zoyu9bA%16US)4h7Rtf995@R$#^i|r?PM-cx1?s!(#m-Ic(C<?w)Ed0XpA~<TTk)Ua
zE*rO{XH1wmDw0u?Uo@EVnpj>0_FJrhSVGyHV9Yz0>c-#9xn!vaA6eoUb;fr!bh0$M
z7nAoT(=o2X3IzmCuq)@oKCH>tw{B^3D^h7{;wv|qn_$r=nncqfzed0Q=oG+o_pQ|@
zp*UVl)94V+UtL`yYCq!l?A*<=83P|bZU0~7q|qzidANhp;HSd11z<;MYqfdCMjrUG
zI6oj@RXUxoHA%jtjq;T@g>^lx_^^%$6L$~2WZ^{Ng)gFEWX&HR1cYJ92zBN1&Zzh|
zzzhSSLDt&vTlqKlfc*4db0T#U$AD^tK-nDaCYe!Gsy-#U6IhYOB9Z@5Feuelo;+nu
zZ3w^*2O?}L`|%Zifx7isxJ_W*l-hkyNYIJ_Mp<S1x&foh;`vj^9*-G?#r`lhUledT
z6<S6`K4PA+$j2!A>vyLD1_rZqR62(iQ}o?Pi)bu=-R2qRIZ;w*I<UKd<yNEHZ4+^a
zGGAl%A#=cn-RP{bI%l}tcSP{v#Z?eHUC3!p=}TKnj_nD6kNK5fDZ%{EyQ@|BY~X6L
z;Pz}IV5$rbA^>|g*{D{=Q#;!%eW24&@%e1-=aUy0Ba>UJ{mw^1;ND<TAyAJ5S}Vql
zvdtw~Sy&;3;W2Nuvoo9N{;&x>@3uzyMx-UhUN@eP=hnr<%jKPnKg{1CG>*y;?Buh4
z)keCtIv!e}d%pw*-au}BgbwxVM2vnxSrh_!sly~@4<1(@_*CaqF*+aeb^jLuk|~FS
z{LL*cg&Os6>vNpxvVAyC8Ls__Pz-!Q2suR=C8)l6?Jy^vz<&(5mzyY8{50lPuy>$H
zz&!5^!i9PW3a=_8^}l+Fj7Nl&BQmoeH(t3gpq^&!%UJa%^2Xq0unX?|QHm;%XQ{#V
z0$D8q1?&}hXGW2?oO|@w;%@6_7iQ(+WWDP+w@#1}^Cns`0yCPZZ6weGea(^msy)<A
z>(+S@x2wwcB3mQT<)XMP19$T!O%g1+LtIr3$^(q;;4QQf=u3GfkFsL-_^7CB@18T0
zpUaz~6l$_*`u}#tjB=|H-c6)0TfQ$9wQLsjJ<Id@uw%%n5#FQ65g???{wLF8KqLZg
ze&?6;FR%c0ycf%=YVgI1&a?gyT9e;q`o(90yuk4ND(rzvc4cO4*=Sh#oYgW#NnzjJ
z{lgPNQi#RdaF!-iHqGJ~aGHy1^<mYh2yUk1L(pr;`{azb(WzntIX-_MtMkaSDrhLF
zBiPQ;z3lg8(DWwdO%11;<KbhMEsHOYw-tO}IERDQg`O-<YzS>WuIhMZ1S2pN)A{@R
z6~z#@Yd9O1JDjikC;Q|Iz8(S`0|K0v2YM}HDGk(`U;)gP6I=0!OS5c{8AnSRqo<xl
zT`YV}XA{D2a?Ex)lNIq($#j7cS@uQOAnp8H`83?!5P0I$IWw+<(6Za`*WzSS11>9~
zoo!j#m;`(ExmNwjbAx9O$45>WN9wuFF#vbp3)7OY00z;FNoBn@+%9w;^6A*PYqWLe
zm!({G9vjU^^BxX>DX493pxefpR}1z*QNrcu|Bv|k_CN6z^}pil|GG8dawe*Bc@9+L
z+V+q>WmF4m)A=)e&Ti$~jY9TX^H_K>1x;d3@}bBJTWt!gy!8K2_4?%)vpYKYDGz;P
zrT2uF7l~6nY3ArT;5dz;OoOloOaIo$NN-9_3C8t>)-vqe8Q1sV9hUoSQuyclB7=T~
z>i)CCJAYn}Cemda4OfkBvAPf6R(%sk-iM2r%l6#>i0`v#XirTm(FYsal~Ub9r4ToN
z(aaV(UdqoOsutDzgtwCEOZWC`C<Rn!Jmd^D$8d}$#9(w^^0nUwt83}aMO$g28|R7s
z-a!DbKg0e`nQK3Bk2hf?<6y5M`Y(v!&Zommb-!X`cjDibq+b~@u@-A<H*o9Idbysx
z80K&SPFPYkEkLv(W%H+RTZ2F#=t@$-?8$?kI-gD-YN!l2pq~UAHfcKBV;t<<p8^T=
zF@X1)G^w{24a)Gav%ku?(JRj#);=3iv1m;zTCo08w#0YN#FGDirkv++!ipU5^y;1(
z*TqONFu_1{=P^&U^cEjkn#OxcR^kY$ed+rLGs91mTjnCMn7=~et&tF*3k61%3Fs{c
z^MnrCYBq02x>(K@I|(Ch(6&Z=Iouj%pDTCJP8A+yw~GvtkT~fGgp?O3J-E!-i@mXH
zz$#1Jo=B%_O<limSRrT@^knDb|1qAKP*aEhAL5x7y-NT_iVdBF)v-iW{rNER3*|wu
zdNNAx*T|UnK#zupf+zQ4LcQoeENkfBAa&B{4&+Z=MVf@b079OyDP=J2apBt|CDB}N
zkyZDoPv%Q?8S}zPxdQ$*`d+Tuu2&x2=}NovEaSz$Mpw>&t~uEP*|?}OGZXs>6J3jM
z|7kYeUZkC4?0Y~Jp{P?ULavoc_?3XO#K&JbG{0y71RVn^H7G1~AuKRJ=&bk!C`OHV
z3_xS2r|$arlsacQ6Fh_r1cz>)Qh3(zT$A5i+xOd#{6Cl0s!X3RO7XQQ(ZmkrhkZ!U
z&Y+|>Gl(bIQY8%(OuQoB#0Xo>r?cLlS>^JnYk?K5BDm4huFd@vh4=iGq|H9hyaLBa
za;8Skuj}bvi{;kasgaYr&p#HOu<eabz*7cm6p}8q1Uq{T`*npf{f=Qv@nZJ~>xsCu
z=_KFWsZupr8!gT}EB8cd8B0rkbc<?E*1tJf*&h=sc~oj$pdjy${btgIYdQ<+VX$eS
zNyBPE(N~WuaXJFM^%Z+%KdWf(L`V+OTgk19Xs|X}e{CN%eVeFT&=U$YXXR>p`C*}f
zC0t6(SZ9AneyqKPgpIF72|E44)V~g6{!nJEu(TnM^)4`3)4v5gPjW5FTZ2>EZGF>R
zjo#r?>Ax}lueu?^RFPhB?i_Wjm-E6c2u}YQ`1hmi4Q-9X`v@hnUY9db49qG*Rh+Io
z11@M|Q`w=(jZ5lg2#MV;J|*<To@dGRFA$@Do3xpPcvebNC3(cuI%_*3hOfW>*z1sh
zaqD2p@&jy0<OiKnQeL&pJuxEQ6<+!;(2309`NrJELWST#&n{J#-}r5#=NbAWHFnNc
z=?qaMry#8@m@wL<10*2B&Cy7XAfM?%O`8eXDEk$qmK*)U;S9FKKc}=eK)=(wg<sSH
zi&p0@{_wj?Fm+vJ|M5zXG4sXDpY0flb79?i)G7Rxux>T7vub%3{Ruo_++Y8FO8k=}
zJsCzAj-_|@A)UUB$0m2xk{1%#Jkr%Co|^OwK<CuEAahJg^ee%}@5?QhudB?>{n2%H
zbbn>wnAa`o!fdD_4iGPG7d$Ge0$+CKcA`s?{bP%qv)fZWi=pJ}@j=!1Xp6&FYcf;(
z)Y-0%KJVPdWUaGxh#>Us9LbvW?b{@`?7ng5W=RgS6|cm1IsQ0Srf>#~x}b&&V9zLk
z;Fxd>a5l1BM1wem-GE(Wjo3$I-+_U#<YfSEOUtugLW%|EM71Cx(FJF;!PnYOjoX2R
z7bw=(_KYF%GaU$qaAi+sBOe%$u7SAdc5&R0v>5N}rqFkq2km(EkQ?^&Ys{c@hsKFU
zS72tlLQX;5t_5&64h{0YG86B4z?61zC^7l{&j^ms%4c>|x3NiKj&9~%mh*F0saiqJ
z%wIx-D;d6D@p$8HDvM`bWQ<vi%yV6Gkzwp_KO6e98HQ99AkB3+(Rs;jHGw`$dsyjg
zbVUquZWj~Se?Qsv*Z+!yRlEdj=Ozd3`FMt^tAG9EdC*dJUFq3{KZ_EaG{XYCwqGAm
zb>1I)uGQmEXQ&-SE#PRsKxz#~+Y#CCIKCL@uIyaR4Wi9@+OG)DvgXzCy_u686F`qv
z9Rr+Z>XTE2k_hTf6s|G}RH^1-a$V@yhRER)cef7)>kBs}z5BeYO6<9VBjr}B-r;{|
zx8EmE^`YV*wFN!Fn<wmFGu+F>FFj0_?vZ71HM(~8W-<#PlBUF=Y{24h>K*gmdJb8o
z=Q;%}MAgT!cu=2&@hEiERK7@xyKeUGRN$Z}iYe(g{!XICy#qAGhsuCQO_$+q({TkH
zo$Zgm#w#i&oX(`rM=U@uX=a`WXkgx(dKR^Qc97~lR7ZTroJ-%E4@7kWsnVO>sE8O<
zt2`AtZ->a;K3CK^_&eRno^){wqW)0tpmt(Am(WaQh@vhZAwS$a1L1y@1s0(HIIz#0
z(;y|is_-^{8C2r6ulB>MZ#Wtqt^|H~FP9_A_LD9k)ES<0lUn?5yKO>#p|xWoRF*#K
zak0mJ!RDYK)->g#=+4zGKp%H%_(e(P19UfCW(?tmUx(QIs5_%-bdi2Rr|xfyy8-te
zXdr6fd{m^|pZ`?w1?f=IKTda*J7@jSshj*T#FF5YcduD9pDpF?`GcuNeRR^LO^WPM
zg*%;X&e~8BttiZ@3S+jmZu0mFO?y%MC#1{T+h7Z6!5Q@SzM;{hoF7D#f$9kF>X6K5
zh#+;`fFvHT@aeM+d$2Fw)O)MJUYC9ea^w@vigK&t^A8+^-SHDPkE_M9{sC8hw(Tuj
zuCLY_@zQ$s;3?fc7r-K!aysyn0l{O2<*_UXLsKLrU|y;m4d*?3ldMT9MEbEDep712
zo&W*k3bxkq^-EQx5LzJz*Z6kiS>Y!>*IbU+HzfZ8+Y(1X7U`uhF@yt2peqk7XJr1k
zuG&e0r>kUICNSfMi0PdIUpql*#<}(Y0dP=O!|loU;!l3B(p2$zid`77Qslp}vdi`5
zXJsx}19YI?(1%dTEf)|Ub?PSRdbS7c<|j*M{@e(uiT7#+x?)cb0Y-3jYB4#OsK?tG
zDhOj*IxDVv4B(bg(e!$itbAsk_osDOt(-@9*jH#Qej>|~+*cMtW`U_Y<d~%<H5cf<
zmOk&T)RXAX*W)WLP4M8chHn#UL*=Qg_*o~HpDHJEvvNx+7TUIPTLS0)a=WoYTdbws
zWs+)$e{%9!lI0G~h+JA3{CPROArIGG@a(RVS6XrTnt#kwuk=xf-Sm?g=lWDjj8(6z
z>)GgJLGFtelfz|jzz=}}eVgGw;21c~0^dT$hlwqJTCiFngXCYZf#-Yt1c?FaqN2$%
zjGvm3T`+()OGx?6XU>IA^(LFCZgva{?g>pR8~p(L^nr!e2Ao>tE-XI(N)AyV$!*BS
zCbC_8TyoG#&enf`VdJK{y3jC$%n_&`2|jO4{{=fs)4(4CL~WN3{L_p2++p|XQ`UYd
zyh%-+u=9{S75d;Xlq|P*>S9kq+<P89KGUZROEQ2|ep7otuEaAho~7u&M`c+X9!elg
zI9g8Z$`oF@xFDJ_=03W0A6Q}Xr+;)JBZe4uW*MwSs`~KW)$=pPZC+Puu*(?VwKjC)
zJmq4zQs{Y_APL(0>A?Xd$t_r=aBEi(65v@Z%5!;3zfePba_jqta`-B~0J%U7CL#NR
zkV$tVb+e3GUTK&ccJ}?$$_q!;Pcnty)HZ`fh00LJ==e|;xNEPl51O+4avfo@D2BCr
z0L|jr@I3RdB>z?gBPYz3CPjt(yA3j5)JE}6_&a;F4fY#D{&GKAV@F8cr1K!{F3WdF
z%_x#4Kteuh`AVkbeWbb)_zr7qSHTzJbFE+0Y#AgL<(H&JO*9cfw#%4|Qq&8+zey8n
z_2~*riw!9TBFG1wDf(mO)PinJ>7;A@<GZs8Q)XX1_FwpMdjiWxFNgGT!rk8=%Ddz<
zw`G}94T$WcY=5rKVRW`@Da!fk8iQh*Uz>J%0YS*#f@oouK1HrP6^rN+@{6_IGw}Q|
z=;nnT+zTj+cp=aH922#L0E(nb)RTa{@f?|7L5lih53tb?$7zEnJI`k02L@(SZoiZ7
zaE6|vU29B66BZ*kd>g;=AZ`|;3LK6AR~XcH)t?6*Mz(4Gr7<4ffh!OO71j_2CHEOT
zr22E*sry!-AZeiuhBdh4112o@tbJP7n^$If=c7{e`3B5WRxrr;4u=p+G<r_9DF;Rk
zlQ=h4SHy1UPzGvF;TmFr8k@g2!C+}dq=kadOPY2oCWd~MPceb|-d|Jvt0fAbpF%zi
zNe0KVJ_Pq(q8tNMhEW3FXOl}?$4>yZf0Q%y`y4|fSuL)_*#?<zS9U8bEJTCpv-9%G
z6H;-$F~g2)OZiJTho+{C^9?b=22!b!A7zBD7%TyVMi4cJQgGIemj%fO9qSB)LpvXw
z2ge*Z3ukljuMcml<6^75P8SA!p0<eEFstpEq9~2~8wrPwS*{t9#wj}F#r0sn7sO4B
zGJRW@^f^(%H|&+A{iu6tPbksm3z|hDjb&5gZ}DN&1`>c!CS`S3a$XB#fqUF1eSh0H
z(*DjRMJ@%ry)NXx>oBM@;yYpf{hqP=R#5)OdzS|;S5<tla1BCGgp=~x-`i=fWnxyO
z(qsjGW<)TBxJ>d%fUi&|Ph1E8Gr3YRIh4@_MM4rp0RySd8^a%9qKm_s;&)s;na{l>
zz8$sfJ<5YP=iTlXmz<yr(eC|(MujRm-FQ8#aD!iD9<+t|Gru8tA&x1sL^;!X{jk3M
z!BIQJ!Kx#H&RI*1#XUU;>m6l8u5$`gelW=Yxf{?{9ERX&EY#Tl+9vsO_W|m%^48S~
z15;ba;-9u|hnh5pYU&{=3{6+5F^zN`Xl<E!{PHM&nQPi^w-28&DrEr;+Q~k+TPK;V
zotf?b?Ef<JDepfspZ-UWa3f5Z=1l^jXP2ws`P~@2GUsRLqA~D&lA03j#z~Y<UP;*h
zc?@u=bnW~0jOt9nT95_pYF*S;DXx#fElo@z7EULj){8t!fQ}H5qSI6fXnJ}X)%)VO
z;^CzmH$l-!JB2`3GOP#42^XFjB{R1?E4ie6dSWFRyHVwh4Ob;%3fAM_ub89CsJv*)
zMWiZ2Zpf?2;#l(Y58->$$e>A5*;*876FmiSn|N8(`nck(&_><hx0#$K8Kqb{6Yy8t
z>-%G$Ypt%;j<2n+$x9rjSly}a>(G4wC<GR?A~6C4*uh|}$EzUwx2Nq+Eo?h)(r!Z%
zgquo2`@zCZtwy)|nim>~BX>pLsme9n?Ku6YF@!yJ-S!ypFH75XfkC9lsOcRs{(k-F
za6yZi&o4UW&c8r5jkrw&2=G|U3~O)w_n8NJY~_XaX;1*eT)Q(F(u3ta;myC;Xt3bO
zr7XM?XlRmV-C2dtw~KnOAs)>?N<v98`VhkUlgw?&ax^gY5t|GB^7Ds#X6o0opD%J4
zdb4vK-OzrP^Y>i45JGsIE)?>S;5=hNmLI>`{!77o>|W2`$n>;x1&I8zdL0jZST_oY
zkRbgD0!KJqG!_o7P;zJr+Z>o3m1Znl0(^db44~5g9d9JjRx?|iC4GLmNhSh0skiTv
zq64az`LF)|@OOCjTeYvVh!7Z^92J3S?VvPJ*|4>suGMv3=vxbTF*HNfh;8ROnsd!U
zM$PX*=U3htPj~+WN`hRHL7C6H>v^M^M{y@BrM$x_3;<3g@Em23q20#-vLadGwWoGd
z^v}Fk_(#apfq8~4@ZaUyhA}_V)H=`M6&Lyi@`tPw#nl|6N)tBTc96F^<m*v_fIu|~
zrn5I@8Y9nie>pE}1dR?OJ=9?nGlQH)1paiS>X6w0XU!%gJk~zdjDLF-dQ0|v!P__e
z2EuKErjPVrI+P%5oj$L#^jJa@Y6t4%1s>J>y0#pcywQ!9j;AL_M@{{zb$xWBd_&=4
zO+#kj-@E7I4F|6z@|oNil*|V(wgu|<Au%0z{Z4j~c!V?2NDbN+AHaEuHF|GLyvuD|
zagZo^T5JH7jCe$ZzQ&SFWX%o6YS;DgpiUuck(SkOmp&2Bc!PCK^2*U)z9G^q@4u|@
zHZgzhrfZXCM2LrYY+~xsPrjIE;XJx$ihiZn-?5yl>^cj8+gr;Dedp2hTEvAESZ|_7
zV^6}xW8ADt_ng3Z@z{=$rs8FG_*U?I&DFo};w7u+(+sK-S~2O10qTwlj-LOgzW1xt
z)T{r$``)WZsmFlMY{ecdvkRS{%vpqW!jv%25qP`0=9Tlml`xyFutjQ`j}+8Qt#ej|
z_K07prOJ{&uFq2gB$VNhes5W&Uarkd48D5n<D%#>;Ny}Hj~VpyY94tHy~%}mKx)Ir
zA|97mxmxP{=*gUtO947O)G!l!cUPji@&Z`T`?=>ZcgwX9R_B1zN1xTqYV>{Fl&_>3
zvA;6lPJH!H{e0w^i>0XVI?^~uKV&)xUqWRiQvF8He5&mrMN-F=yTpy}p=J7hgfP6G
zeVaRF4y2uvCTvDYxO#v>zk#I%On`H?@aOYbtE;>F5-&`S0dB(069kIms404~^k}I7
z{u9PFd<;0POuqDa6Zi_^fQ%Qe>tu9lYjC~Wb-p04=OL4j;NUS})m*ERnnP(jDmGtJ
zQ+C4EoG9TX^^BeU(CJTuLTh4A)*-i94|Yob^0iz#=VFAp>i8ves-u3kpz9Ts0RgT*
zkzU8ixQ6i^5#8;WbR~jouIGFGZ`>(&=hv<4yp921%ht${E(qsJc1OI&_cpRbFK^!M
z9?Ne@Z6)vGJ!Gw#D;zZFmn3&E%&1Q2v}@-{Fr59owd7yp0>02X*iPmur~9oB^JWec
z|2_=FG>$$>9|=10o~Rfp`n~7b2({ZB4I7XtO6c540s>D`kqgPn=BC(4?M&KThZ{>P
zrjdf;XTYx-J<cLn$!cBe4#Zv<ivs1w`T`cNw?1rT9WgaDh&$kjgv3;D<sSpO!?u*B
z!&<|3cx|$f+CC*TbyDejf|yI##lG)4J-`kLU(=hjZ*dHp-<gg4O{ma<Z((t5z%wUy
z`*K~`<J6z3ypmZA;*N?!0qE7ndppSh`4ka#kvH6Md`<X`@2>4f1z^rW((uiuiQT4-
zT7+w52d2(-UnJOFvf&M|XfaK*9(tn?z86hra5)ClUJF9?p!in=LWQPO%?z$FRcl#x
zpF0MaOK`E<GbEc5<#WIEKf+^^Pv}+U(+nI&<g2)zrk3%HaCk@F!;|Hv0t@Ggr=aIO
z;zH%9%z<8O15vSm7KFd!<@^JEsf{r6M=CkSF*8%o{aic4<&qtu^cQU+k{Ww>2Mb5q
z90Q^y@aL|=(|ghtH+SO(g5`C9=GdWXq1B<Nh;2}2b3qiWhVqUyWKh>;{Fi1xbp6w|
z(G-{!GP0L>Wxjk3YPQ)20L;@th*Q51c7(f#0J6XD<ZdV9lO>6Tn(?gi(}5i5WDTMD
zOzjxz5vj)Xhu;8QmU?#9s)wJ4_;{>)LFeHFKu|f1_q3R13Ij;FDm1z%TpQXg&Pmlh
zyhXh})1w48H1KQOcYi2o#vXS`a8{(%%F1(KQ9BZN+d{3NIxOZ&mE~%8^D!Wf2J)LU
z&I`Q<{N`~2hO4{;07gK$zv<XI;B-#^SDGHFviJ{G(KPi@Pmp{wnL#?9D8?dr05H~s
z4DUF`w%#nGqV9a1+{r0v8%JH_y{R1{23-L@2cD1Q1i&G*%O1(LJWTKa!iTDkQ(NcB
zrwhKJf3!=!q)UJ&HmQ~nF$CxsAi;tb0iCr%F^2)?+#L3%--$p-JlO~034r|#EaeC3
zd!F(P9EA>*BE`&k4u?Iz$GAJ-LGy2sX;-JoQ<IA3p6C;jgb&!?S_qFmc*gpMXMFgr
ztVgNbxVzi?)(CuKymU;T@V}cl_E;(dr%rs*u_IPoy0$6punS{3T#d|e22SVZI!YB`
z3=YPv;tTS__&3%NT7*0gbUxpNg8nN@Q)f4RTr2Zah6^FL@-;@~CUE8-Lx*Afl&@5k
z#|2TD<1bQnbSyy&=4A26_8_u+>Nw(dh2>G|XGG3BZY%Iw52L8!$SC7fwzgX7N_JJ#
z3%4S)_QCs9%hUCvZv{nPvD9D9rS!R$@_k@y9o{(N_;6arqCl5Ebl^>Ue$>&f45w_W
zOjDVMu-ymu2D~q18WH?qo<IKpOJ)zIuo)y2?d-_IyI>t_Bdw87hQyYFdZ9CaU{9}j
zvf>8DaC@iPS3bKK47uMQlgpGnea^(<&3`t%$tW71ewQ+g{@o1AeQo3mH62f>ajg7p
zjconsa5qoC-c}8=J77Njm;kGvF!Ko=gz^2NF}i3&yuYcfX|QK>b#)C0;<l54_pURJ
z`gtT<awU;gZ;|hpe8bEps%3{3@O-P;ul`gn9>t+N3oZ#q+=B=^W3chh$iO6F*g4mh
z4uS>VNS348a`E(kqwLMYp??4N;SrJSB)bucvX$)HBt;ZbA%sdok`M-CM)qAu3K_d3
zCVRHAjHSrFOx7{R7KRzicw=TfKHvMkkNf+5j^FQij^~g0YmQ^)dSAzNz0UJ|U$65z
zVVFICm#@^m<ZVt2=8c?YWE9@*xqz~(XR+PQLUZV_w70GN3(3>#dK=T6f0i{gQAKF+
zBvIGE>!x@mM&SN>N_=qn_`b87Zicx^>W)n8?wFf(TE<YoBRu6rhPD#NisM<h;_7#d
z*B2DV{_I_|UP*`jRH3dp{rmBWB^yP}b@-O`p*i!&%EHC+PqeacVa+x64`P;;u-*$E
z2HD;@;(-}k$LzM05FglN>N3}S!aZJXG^mwfZkcY2bvXjDB8mJ*pc_XZ3<MB|0J~oE
z|M>US^GQdb<Q)$F?LF!RM71@I?+C<x1e${$fe53Bc9+AOh#?4RfX<Rc*L@VZJgB;z
z-t8&4iabMg6llYp?7!#Mn%ZGo04w`kCzAhjLtjX2t8}KKhkApSgH{C;gCRFsNT{f9
zHO((GgY`r-`#Rg4wxS;JCbvnQjAY$y`=ZASKnO|g3Xr>|kk$tA+%GSznU@alkE)&z
zii?s?t`E<Do8q<Fv6sVTM@FtU;=#O77L!4hmBbiCUI%k}=@k1T0}1WAD}stbl^ua8
z`YajCXhEvic+<K0hiGw_&4Ur@H9dQXz0BE$jAs$&0fwY&ua%+a>_U&J=$#{wE9yXo
zE_BQ?$j3#cw4bi?0(_>)fJ8Z@^KCK!_eKos#Gm5CZphbez2V^DO>OypmvbXr{5`|&
zQela$>tt}}?(T0ZKfe{{jDW05hW{{ceMjXraT&5Y(WP|SRgdq(G6u))v2f*uF!_4w
z3%!aCPQCkoSG;N0!JlecM5wp0R6e|zcIQ0@^SSy9Q_dH1`j`YV?IJ#S^<E-?D(1XT
zbs3EXGPjTkH-}z+JJt_$o|<8NG@EOg()$~HU#4Oz!m^gpKG(#QIv^8qn8`Yfwk}(H
z?a+!AI<{-ms3kgCa++QAm&~koPr2QfbGHt2tdiDu$%&3JwjASW<=qr65m!%&y=~Vu
z{!j0T+t&|xHM<Q5p4KlR@K<ShEh|Xic;ztulQ4jiBhN;W>u!^)9nXxk1v(^sb@N*$
zcHe2VIZHVy8uN6U6(-x=+Yj2v3E`@7MWs`@$aMrn2ZFP$92(Sy7HAnMlF!o*d{D|5
z`mN~kgQV-)VS}-_V~84{#2$h25O^e&3rr0R(cM~_6qyQ;I4+5yH!qP*a?{?fp3>oQ
zI8Oji$h9A2=lLQ}9rGr}4!Pg{fpYKT^vEYbqHnI}V^t04e2_^wddv|h5<${IkXIr9
z?`O#2QUoyb12H(ZH)luSl5Uem2|E`FLGHo#)OgaL>RDWR%ooqNr@!i-+6VJLv!yTZ
z-L;-#Xdrd5?H;E-I3Za5&=wjt|DhD$cNQ!L^*0^dA-}ObJ5sz2`*lK;d4EduZ1<gQ
zEdMr#+eNj!M(&hGmpb?e4!ne>#q5?{?=dvzUQ^(TT)6c^?@a;gHO8uTX(`&eAF2_}
z2ve*A<VY@UXqL(ZsiYet=W5>-U6G#hQHc_py}l9r2fG02P|#bfc073-mhJ@&By=6z
zfxq-(j@fwdtF2%8hPh^|?yWxf<apZGo>DXWnBtRc1;-WCC*yw%z2MJ?s*MaicK8GY
z4Cl3@Z{DcS1`D{Ut-^1mypDXFKPWVxM4vc)@YVR9;H0sb#Tn;Mi6HR;Y|jx$RDRbD
zGFOlugTMXp%@K(25tMhDb>obnPExh;QdL4h`pGA3*m!cQ)Q}V{yIjXea)5vkSk0Dv
zxp63OzA#m?!!`iER<rTsh53YRmpU~`t6FNnZ~YBi2H2lVViQQTf3AI)Mp!?1z5NGy
z2i3zh7UAj&5k_BvVvTESNCFs_>MG31ahQ=znvf7gxyV|B;<8-FTn;$lQ4v((Gb;z@
z3k$wpmYG019s(1d<IT8s$1TMPL<3?+%7$QRc4KLrb{ey_t;%<-0`b{sNvi%nU6N|v
zbQaZ-VIC#$GaQxsQWg>mF7tO_rG0{n09?I6VD>5d85zin(WuFSfiJjs#k4`WQCF1c
zrf_t{7fS|9Ll7hAwC{iG7aIo|vSHd>ALr)7aa7(&^^_$}p`#lV$S|D(1$i5{elw2+
z^TTvWNf>W;VcMS<&4&t3Z5fM4pnBmwo%rmq63HfO41~SO^w4naL$(s|nLWcgA_t*<
zixivE7zs^*M<uT^cv7q*LB&TPersab7ep^EoX%??b698FW<?DpZ+tZ(^K8+i@JVcc
z9oPqnD#2&F6tqoi?33-<FSC^B{*p;Cm(I!<tgVggnUxEIrks+zo7~A8@xY=L#0e5l
zQvmTV|8oN>?U?S{i|k|bv048&-+v#liLZ-*%dGSjLhCS-{?XOjSq5M1zU8v6{=C(r
zerTe<SDq5f3`_rtXuC>3+xg`e0x#y&c?9Bm^)S)`PjpT@lXS3Jeq4S@Uu2Fzp}r5@
z*^}3h!tjf0Y6jY6#aMlIE{koc*Fj`Q`ytyP<2|Kp`vE4v`jCx!7BKMPUrL>u^0@sV
zdc)Mf&MD|pcIS;xH>4^oS!SCzVlRn!dW(3z&WpWx>e5ou5$K$q`N3*Sf5})Mn0~z$
zxjTmV`~WWfz~3m>;wJx^c=W%8_%aZ#{+G%9^M9M(-TU;0UY@aVusB8r%6n4OIx(ls
z7uBb*Q2{e?%is^)Pfs6y2I2`)ymc~(QA0@j#Q9+E!t*ElWS#{!KXs?8la|BDv^vz^
zK(#=GtAi19AR;*G3w3Q7!U1EW{gXDfV{hH9C!0E_^!Ak@pre_4VzOgMJk(9OOz6^U
ze;(ZB*qzPON1(Qg2$$wltiB0s*5vFQyC9tV<nJY8k;I4BhR#u~>-*hvErM+p&9B7z
zg}?G`@a#k}x~og#%*n{8&2p|=Wi_{?J%2sQSd$yr#Q_&s)+fsC1Y7X-30il!tNy&G
zt7teR^j|{EUBiBKcO;pEqruEiw4&Dc$#D<Ab940KA%eFQ*JT(1{`Pr24RE7O@^<-E
zgZzn9(N0HP%uX=>wskvra%pL~`!)$f07n)8E@XYR0zWJZjpd&7yQ0ZE$=gnE>%+wP
z|B&?3z6pEzxs;j}yy(tgKh}Gx2DdT(Jhtvu+XIpY>~^0Q{&{zM&scf~<d3>Cl82=H
z0Fbo8*C|TvNcL&|nus=wQhTm>sRml$MMhAg(<gFQ;TPBwGO^`&TXQ_BsVRBvOElBF
zaRqVa+MdQ-;zGoVFgglLwU|RqVDO7s*2)*s3v7%QY$iIlO<!YDcmk_<y&vvLr^wZ&
zRGVjY_nDP{=EU6}lh@HLG!I7jFe8g6yButm9w&E%)yss)Eh+H<$PPr!*eG3ylu$oJ
z7wML@H|S8W-4AO2cIu&JgEvEK+ut9k8MQY^2B49WKU!9E3_0?25da|)Q@zRFvG}|Z
zjFIjRWe@{24s-RV;c3HfG8NcNYe#n09=?3(@y>{&=@gHY#UMz_syhRA3|Oa;U5Ls)
z;&>ft?Kc?_UOah_9p=X}*){Y#x$?{Owf@V>S9@xeb*u2^T}kY)r^6hx5dm>O-iwr}
z5!ifQA|@ROLy`Ks{Wtx6=W+Zk&krH|t84QurySM8ntT|PjlN7eKmFlp{N8uqdh4@L
ze`e9UEY(Af5P*%6)I8|BLKPyvXoWg<ge`<UkUFbQ$a#GP@@3sUtRI+xC<IwY>)7N$
z&3cJ}<0z+sDJMz0sqh%{!3rU&><QX#8oJO$w{V(gda^&^%=ez$<iU7;iCre^-&!7h
zxb-OVDL4O=+mLny_cWiIkJMGfvuSa^EA-86i$+bW0Xr1^51qBHQB7R9cC5+$t2i&t
zsOo9KVHL=0<Lg)tPz`%Q%TZx#=ArWE0qPj*9!7NI#ShD%1#Xs0ts#a^Ox%}U8UNyH
z3@A^0mW?Rhjno;;^Jy2&zw!PkB{GtzPa0`G9Nz30$L`%urg$jPE^cY_ezJsYUPgU-
z!xipi3=}$(2=mL)zP$;Gx9&oN=SGFztYwLw`abtwvps@|`<mD&vRp@qbicKQt6hu@
zBOILdYwwjUMD4hIdN}&fs~Hv%q(K~y4$Kdx1K0;qTn5s=0tx5=R`Evf96B$!@TL#R
zWqvQI%sIX52*i5?s#x#2Dkg`q2zXqsw|46UcmRqjxb$z`;3Z{)pY~_wHzz#rW_FJs
z_=RjAfu7!enQC5w`;Ji9yZrhVqmL+K2qf<N5;_F%Ayr)moQr~yx9f1el5cf)a)gTX
zmMafr39A9+?Xd?hpn-1|7WqpSOuiL;8^}U9iCu$Trqv=`m7=T@oy>V_TH>mkmrva7
zOz0Z8-!R9B+f%^u@72Je^a(vNG7ARB4aiZ#Nv0)@WPuQ8sRxyJ!h#Dm^<G>N(!Q(m
z`G<{tONjL}<TP9s2qG9^X85|`XPfPad{%XQ?KB}S!-Q^9!{tku3g#wxQ+2aa`vH^q
z`nkl~SlROXkN<Vo=DTA7QUcf36Qrfqcyif<Db^Fju8-({uQTjOxoRT)n&tJq3yybw
z3_Jpt7bKrcC`eA@(SQD21H@U7L7hwOgpj5YwF)gU>$e}l&H?wO3<{?ay$bg&rTLT(
zzNQ&6`NdxC{&>^PjI&_W`ku`BQ}tI@aLiQ52vwKVi^=9{s;L{5uP#%>nq4s$gihb&
zEqaZZ>GQHht?!*p4Lm@;RqY*MU1eHmO?+0HI`Mux$>~Rs_4gr1^GkPAs*K(h%dQ8g
zN1|1sEqKKWMydg+qpsnZjTKf!b({OdmHLqoO<l_2MN}c|B-ySV!S@Y4f^U9>z1KQT
zlJm%h?B%~yrrqi;^r5reot1d|@WkmK5_|imPS9=2eUfz?%CMvm0}0#R@l;m0eH|fD
zN_iT#IM6`U*+sDl2?&39bs(O=0-f!y#9kpw&7VypHMZ^1CC8<w++Vavo4^AWAbzUH
zst4Jd0%dNxPOt~2NaC}`HPH%iS*Y0<+pi#oEzg_;!`m9Gua(Vt-z-LWiHibpx$rTt
zh<qXWY`$kcaNFQP`W+G_SF0c+qSyb)=dw`F2Yjt}nT7gZf0ikN4<NaGJ^9@C5CS@0
z0@Me#4M(7k@?ULo#kM?K5iGfvUTYJjv+YY)gs*`>3?bUqnN$~mkPj0Z&O5xWSB#oX
zXn-ep{Kj&cF-W+)SdJ%ThwycRJYeQPG5%OAWDBL+U^l`ow!YgL{l+Rgg(n0d;bA2b
z*0I2Ce0mJTH35$8Lh=!DW~5PtC09DPbi`D@xL1f7R42YJul&p!><S=O2CeI;pc(+*
ze|>>QpgkZ4YwkrC%$TR&J$!v_nI*-h9;`peEEB6Eln3-;{DnCyiuD5u`oJGNK3zKv
zYESE)k*W}ss)7A=D*-Yo*T$T_k7fU3pMT{%;Q?npOpvb7Fl79?u(J|a&iR05YFQ(u
zg?;$_M+LercB6k@E!OqI-CvSCp#|bOv7oq{<Lf#`g|@4K_S5PI!NaWQ>fd6dP+z2|
zu?^QGLY28*><8J9>S9`+%{*Q15a)>`bU@fQq?K2-7@lssP4qiQXHj$fcLd>Yy9ePc
z<nb7?2<F4&>|0p+IA2mzlZwZTSYRpV+!07pmT`H2dvf4V26p#&9uwOyir^73F03V!
zHwx45)ay!HiV3uc87_9SI-1$qw;|j)`l^3<{qw&2HXpO~J!L_jQEs1%-{i!Yd?2ZP
zvgi}4K-m?0S=pZT&n2dAg&(wGt;qFjx}^kM6u9~&lF8E-5oPs+E^RwCbE-+*r|36p
zYQX&2m!k2>3Ax|>8PcRcU=Qp^sxd9UiH+1&P&?}Dj2=sWTj-WQ@IztZ7RxEbtYj03
zstfdzOO*&p+})%j5T-Kmjp?b4`ea?{cYN`Q^-m#x`2XTK=nPOW-nJk^JdS7dj8BOJ
z{9rN3r-?RwAYnng23Ds&M#Mm>BQrhN07Ebf^wkA*wJ0mULwn2@3DL*=OzOUj$R>9T
zot?@yswcVH6>NC%n~I3d=>fTm)@g6s60YK;m?#Hi$6Um}HHHg`8lUz^;1`f4TD(p1
zM%YqZpFFw8la+Wnt&mkCe-SmchOcZ{M+O~%PAz_^PHk8BYZxoID{<ip-+O<dflDbr
z7qZPpvX;=xfDK_yDHg%eq(tD*99pmGQ^VbjKljs7`EsXu?md%t0ZiQPzjvF32RZ^(
z>8W2@>uiH*O&{tmLuOKB?v)q+2{j@(YSUYBJc?)#^#+-1y^Y{rY%ITVew50G@P6G8
z5gt+C`e0}Jlz%H-NylYDXh6)Q&(x#hq1%i`)#jvZUQ>Sce>p4BzO=d{kQ>nQd^3Fb
z2=t66n9SE<O8bQWeBu1|g^b0Yj_QvWz6HJg8`{YSd|_NhhEImV>}V};9jbnHM}u0J
zhbLY{v-{FUrVgz*x8uWaZ|z&?CYptV_VJSA0)ybESLkB(lJ<dfwfWzsnWvqb-!sbe
zQSGihY}?=#-(9!d0RrJY8FAC()!uB6oc}l~PPhI7+@{~0Qra`8hr6ePsj90U|Kx|F
zjwMNWp#DD^%)ybIPv}o2eV7!H%?<YK^vAqH=1w{nrp2^Zy-g`azN|e3nj;{Amcwh0
z+-&Co5j^oDStjXKWrP&Qp!lGvcw<BCwHb@mI%iTFBIN9`Tq!RiD@CKYVqgJ?^v1Cb
zRxh4cC7gkrQ{P!dU+dX|T7f4duqC$11j!ATyOYrp`!T9=*pQ>>V|`CZQz)kHE|=!>
z{iv4zZY$>hX%DrfxPN+f|LnLoMqw46h8?o$KOK}TTk^^aNs4g{JXT1-U{XGvM{e_B
z%Vs{Ee`3hy@D$H_-v9M$qwgU>vRN)vN$c4P1kQ*md%Ukr4#$-aSB^jtes3(+CB%DV
zgpQYB+TJ(Ey`f_|3BmKCcjM>>5WM~8GeM_VDr343?dOBn1ZX>m_9Wm2s`=>=2%Q92
zyWA9SXZX**F4;K(u{+o~>dSqkNJVNKkGjzJ{rv}zKeyeB9X(z`(|`Uy<4<*#DIvf_
zK79bk=8F(q12>=Tk*u!bh-|EXx$mVk8fv`$wY$a#QsKCc{3>P&*M_;66Opl#^JV==
zw>QPTBwFdkt@*^O?Z0a32^T-N1O<5Nk;Oxi)i~-I#&!pWrp|#dxUd_cqoF3>M|7|J
zE=D9VVUO6)Ww~wS<Cs@7VWMWO=PzP2TJp2@c+DJh?0+?U?-8{Rl>h@u0f2x+QH@^K
z9%P-~auZ&^NKY-A=Gpog&TZq(yz&df@F(KQ))8pEknG<1A^SX8rVSk1kH9|zAT;}(
zD99v5G{FBfuIt*fUf;*7=6&dvYwmama1jw94R{5RmMO7=Wk1Jt1ioUNO7|fb8%AWF
z7789-73`+L)N{UM1*9c@({(X)+gS}}b#NhP&g)_$J<4<-_+51eg>kMlH+I;^2P9{`
zkIO=Y(7=mjo1(DW!!Uul{CUmucBKLNR#gci%!^shCk}b~gV;0n2ih_=qp1e-3fI?e
zMEE*_U*ywd9=t>3yAFz<xTq;Y($tv)<?uDspA%Uf2;#9Z9G%}qU~zwTmAIfucW9aG
zFI7x03{fo!M14&2XSBoZm_MRIr+%Hf+P<p`{q<zkkptmEG_ob1f$IP$?~Ef*>=c)K
z5wGxoso3?1!*3Y+J`3^_`b9UOKpHRr?w53aT~+_(`|Xa+W=-?;YXvE+FRy~rd{(`K
zkY<=pT+IN+@pR8PX^m@MZk=dn&nxx%5#==15W!rY;i0k){qdtmIrhb*#~;cIxn*<U
z5kioDE<}f)H1>u~VD5#3%=_H7mG-OO>H3{uQon;P`L6<5U_LR|%U$N6PY$z2fu7o#
z2|0DpjSL*c>l$?-MPV!?purt2`@QDJosF7AbFJ6Ggd`uX4(pG`hYw%qUG@vZw_uum
zI^~TmTN=!(Y%lK4{OqneqdEm#r91;X<{l?G<Kygl)0T<09}gdV+7;RQikLaktk3iG
z=8e12`KP~L2JM_-*xHLhbHJVedoKKSNve)qP)9@FfVkp6Hf5va?;C>4ZrwsQ0n3t2
zxDI4h6OsXDDiE!spQGN%#n>tQt+Mo&SYHUZV<>8}hdFT3JUDI7{@lMp6uvOV07m6d
zaEIQR*zH!gVuM08#<nLX+MgmeeukVV+g;0w56d8~Eg!2-dcS%DXQNC!niq$Ppx)AR
z+o_1CXc03xvl4y~JWbd@X4KHn(1*{$<;JGkvz015S)fbIJim(V7mtPE!-u;AcMBFm
z)jo(0ve=W;@#5*^)V2}Vy(7@do)=QBizc9K`<sPbSAWG?C+RwaC}f>R_OxtnX##u+
z6%=!^m9WSDf4{V{@aEDlbg}L-sqa;6_%1-IS=s{7l5Uj1BT&=@`fu`VB#fqfp!Vlk
z+d#_lAZ6fsuuvt#+Sth>5EVL&z^c(Lmv9X9s~EBiX5?n@sSzEULiLH7V1+j8*t>q|
zrvpyk(m}8?iGR{dMI1k}I(WCQCGCO#^&oNGV|c)d3;xPYoEM0FVDr@X(hA{HD@k9h
z6BbApB_IDj-(k0KalD?u-}^xK9J*9}vSyR1W8%jMIRBBIC|`|H5!@n5&@vg`nfJZS
zP&MC8`2Lr_jrrztN<;_=M@b#zyH8S&CT=c0!1)y3X_s3w`}5(1KSQ=Ud&zoNvlW^j
zs5s<;T?KN#)kQ0O+n@@}_bpjsVOJm%7q%hj9_IF7JBCE_zDLtw7Q_n_gQOjxO-c25
z-uaLR9QWqqhVxYNR`sy_2JTVV4(crQQbnL#cDNUJ&<Z@tT9<%|b`S!b@KXsvoejHD
zQE#Anw59yl-@CJWSRq%%dz8%|yW=yHnCKJY=MhyeTEIJEoNeju=$M@L;k^aE)0ZJe
z`~p!&pvOvVtF}~!b3n;Fj+goro8R>dW*mI3&detw&dZ~o&g}gI<WZ8yS)KhA<NLR~
z$h;Fg$=403TgvsH;|k|?gx;SU(X7Q%{eu`{?adUEPF7xKx+ruKPG(}+8cp*o&=Z6y
zllIz<Kum7ge5B@g6~pp;FD`0l8_Fn1)seM%jrFLHA6x4ex%t^%KzcqYw$*>)^eR-t
z!{h10`52>{mUlsngO*<@vQG&5{w^m?%Ja8Q%jPxwHvLcIu42>3zlXnmGoO^2-`x!R
zwhoiQ-%v;EkXg6nCE`w&4U$|u@A!m5e@c^zn%EI!>)gLj&|IH>Qk}_)h2(*544x}|
z!xneDPSWwr3(}TkJP$pwxX5SJ|0U;%=0K%r$a7Sbo-}FZL!2+Z(G_1AVmD=P2VY%M
zWei(VlJ^<tduJ@2X!-~MlOlWPHp3)GoSyiWK1%i~S!MZ9pS7#0LllzWDp@sseFWmC
zYrn`ykDkJBwyy|0B!_>x-F%IG%FUg6r!SO^>z5ge!WK5mH``*q%(3eTR8^et26qa|
z)l~`Iu$%q{7aYw|zd8}`?A-&mh_7YPPc$K9c(x`9LCxF6_0VH_3TmarFtdht$%uJ5
zN5$Fmns?7hh4+x_L+*l3U+ag519BwrZNS(?cVPrwKi^12=a$X<!<K%rrI?hP>ntY8
zd>5;WZ4%0b5=_;E>8Wm|lcMw5U&4qo4cT9VyEs5@e{gm${F3c$rf>SYU)o57ifU=>
zvTH~dqdqGkoQdvu+H^*GJ4DC<`pQrncXKpr`e9G7CR3tzW&QG(f@)e3^(JX)NdYpy
z8M?8f(EW6+;ohh90=<G0UTa%b;_#Imu6$XPA{OE{5NUh%t#~<Ip`t0$^0yVmeEE^)
z&K0gt=7)Hr68`1x`R-0sJfRl|!du`um>nDE!Q3Iz&g(<np^)=l-4@rxx93~Vt~hWE
zKSPt5i`@$Ufb>(guI2tX%n^Th$e!xRR~P@PIw>ZwIJ(894{Y5$EG!dj+qW+G*IR>W
zN%f~x*t-@vmve@!wtEOR`Y$c1)wpu#p>+!ofy&~$P$<RrW-}s5M&b>Iw`DzcPmV(m
zYf=?C+%qYnearggYr8I974g{o5WV*dTSo1O6A+?#F_Ig)Ie&7tUX=H#C!1UMdWEf;
z*^p$sFJt6g)^41%@a|?bxJo@7W<*+Y<Xq9RL`I<$1}0Swryf#%@wX?XeWq_S4yG7)
z>t6fj8RvLHRa=Qiy+bLIuPA|_;k{=ibz<s)!?*Hg^|}_OITB9Q7km<~<AyxzQT)CB
zZ8b!|ZQs<$pkd!R;Vnj)u~!kQ<801SAt8jUYU!N92~Idd%{yIRHU;1b3yxF0m`4(G
zX6Sl?v{wq3A^&;}9$amX;dtO+12ZLQTn0w2*>7Ql7hasjXy-eF142{Zj+R@OPY0^E
zBVrD$J8>s%piw4|^DHWtMr$1=A2Y8^{mlxS<_;QU`VX6}eN>0zE(NmDvd!O(9mYe#
zz-uK7Si00{GIG+B;`yDHLsfF21pU>~CFt^c*1GDQtL^D?(P#rzQm0r?n{nz}KIxe|
zROKI$Eh!dy`7|@GL5F3|h`1Xx7yN1pk08pq>hKUAf+9yc?}ez8IkTL&FDv5Y)TYqx
zqPaEj%hA*AAfdd>=_E1Xw|cGrDe1`hS$%2mb3*LjG$az|(Lv-*OcFcHV(|Qsj!mwO
z`@y5EeSr#=KYzLwbS;{NZb^-!y=!SvzAcgJmG&@xpl^O4a`0*Z^3T>S&u{83kkwD`
z!KG2U!R{TUZ?o+Gt|na<{MZs_u=6~Y3PfRAYMbME?QW8L+pSB6wKJPf^S@_j81nGy
zcJRsG|Ei9w66jWnbz7j`rxmCkmR;$R8Kz!*neozxarCND&BGv-8*t11?`fHUsr?K0
zl1*ealDIYp4ylAhRe$h|2J^ZgswM!VYo1yNogJGA#(+nJ#WtVhaH-^ZM@Bdkw71KV
z7fguiz8zh|6|K%%Wjy$FX#_*^sgbhp@NCn0M(pMNy?$#48QN0Cg-HrI+W|Gee4l#3
znR#KXJP8>fyTM-86ZSUEw6xn~*LH^lsG4q@TX|cSFO`Xf#>%2@Eu8*OkFIZEQh3XP
zedQ78&2il}Z2FBLVN0<8rRDg|t;gl7ITY6{M$7Xmzt_a~s&=gNOIeJjI`I5G`*~bd
zzT?uDe?_!w0-FS-Xgx(}>~gE)sXt=K#xa=95lE2|GZ>xOC~k=VMF(%jmzj6WUZRz$
zY9|P^S3Uw<@Zr2F)(3=(#=G*bo#GL14gy76?xyX4-!`%2QZ-0KbPWFR)Vln+g$SVl
zk;X>f=NhZQpcPOKJ6Aa5<PKL@gOu&x<jgV?!ZLcSocWa^vP7ElXvd;;PdXJh^uG?$
z|39Dqv)fen2-K$S!C6IKqkJLX>?yTxiQ_X$__&y&?egOhrv}rRO)BGv%tj9Yp|kD>
z91FZ|;+4aDz$M~u|9kEvAaqvda6k-mX9M%Ll~*4Jbk{?m*12NI<Xxo!xxFVh&sYlm
z>r@z;M6AO<q`DkR!)!*LJ$Q^3Z)moVinvSpO9_|GW)_z44%SMnA&$*C)}i8ZN!IKB
z<fn8url*7(>81`|9~vb8a7LA69)ZTjLm&{9&pr<1WBtT~r`_odqJ<Sqx_Q}6q}&$F
zyWzr=)YJ(li?YA=ax@06wuAZZ8^oCouJ0s)+}Ig~SOo;Umw`{{mla2#khwRV+H;Pr
zVD?ocgAT(Hh*k03tD0YLw2e153etw>dZkU)#omW@obA^9J6o5uKH*;A+;nArCJ0k*
zcD>sRedP$mAYx24_3rTsl$tJ%Di!G8ZE<nghZhw`K%Ohebcjn-A%xc&TH+R;Yp2ZP
z^P=PVLKx&(h9ZJ5;=z&G{3OvQFbJLoF@EWxJk^?K1tvmfJv%b?Gp~~dx&}B_Da$v>
zh6(f6!}njW*3$1^X<L2ZQe29C*&+&8g5F`6M}=+Uat^V2V)ffmqlV}w-D45%<{tf`
zwk9uzYCoIj$)XsCk3b&gQNz=;blW5wQqADu&}uEmF+Zwx9#}dVXdR`e3<z{KGN;bZ
zNNyM%evhj*Hm6lR71=2><E#9iSgCTp`MNd)JM78u$)ZM=xF2s16$?Eh>T4nEbzZB*
z5(h5KjSaGlOWNS*Y%SWS4g4_rTmfC4bma+uajXY-(>O-MGRs%DAqCMe5N|!?AWjxf
zpYIdLSq`CcMyuB%X81cBBdvB1cm~6M>b{sECe#+Xsg}(&SzA?IG=BV<F~4M0$eamm
z;HU*e&AG(takvZU$pYGMObZg09HDx|Ub|_|B}sFZ!Dl@EB47oOD>V>s{ss-|;HEw#
zF$L+E7on(k*7>x3weK>|JqTVpt>t@8AFNy+tc1sQ%z)={AyQ=KB$<uHn3SOBkDt}P
zSLzcvOmdhD)liKhAt-ML_2gOQ)r@nYMSpnwb8ytuF3hIKPm$JcofFz`9iEUO7%k*F
zo?wTH&gQF0A+$}FY*CH!TABac=G7Y#T+!5|NXHWqX=Q_-rY#;++duE^(6THqEbgyU
zvzC>PeX!Dc1d{rb*j0|j76)74BhFa&I-WTKx#9$=#y}dRZD3-ni^H=k`GhKxclz<)
z17R0;-swnRnXFKT{p*d3)|b+pOmv17xqbYM8mNXq%N#1)_5}{w-7)DOHTU|Cd%$Jh
z+llB$PKwJD`Y0p{GFQ;Tv8gB%)}Ad&E>5()nxuYXC?@X9+LG4yH5$%j_Xrev1R|ND
zqTp;z=Dbv+6hyq9UgmPc7}e|!1y7Tz^X_@zr|c<JI~%sfan+=9Nd9T%$T#A;{Je-$
zT}_Nr0!xl1h^xIRfUt!7@y*|egP(kLyTKEXMMYK%cmgJ9p<}L5z?Oj3ii|$1!o>IC
z!ZjnaR%{tehW3@tPj&yG?Rgq#?^?U()%D4@VR9uK|MZ=lp02gjEV>N^0&s_?slou6
zX?RVKyJlE*{5%1>bMmbE(65Q!oD=sTfAfA6iu&H5EqXPnh<7#flPRV1j<LA!;AnXj
zRIFCQKnF5(uZqnbEMml`O8L-1t(Z|Xmhc&^kZ=}!=LEK0Ip|D^Rw*?#ztb3b1Tt%i
zaj9#leNFBaVK0^AJSDM<vQ2<mS;t_syo=U*-&<I2ZW4DIFK#L>-41%l{Y$=5^pdQQ
zG*<u9{SFA9%0PI7Wp{LqLG$B;4{Xjabvrt5-NLR0X1DT9A`f+7S4W4J4(AbxEdc}t
zU!9Ibi)y#=0*;EW$Oe<Q+4yTLpC`7GA}iX+ne$rE?aq_(FxDI)uxVZM<K$aEm0ePG
z9`iFDMu-wM*Hk@uflNxx*oKPI>0ds0Tm)q+ATaXB&wC!X#hjvfS}8l4@LaC`r^^HP
zKY2L)%lg<t@Qj0>1Mu_ONCv7*=FjiZ>*G3)AY<2}zP7h+nVn+to+h@6tYOh-V3MR`
zgurnsrsYo2$t0}xaEMC-VpBViJxfVHTPwv+hi@)e0FR4Oc%yrwQ(ooJt6fQcJ+QH2
z@#EfwYTpD;a}ouqKradQEVj9E^?c`Ospt#ciD>?gHv(=#p(W2HK+48MN64|Vs`*@0
zQF1#Ec9pc0OPG1~OeiNIuPW}#bid)PpF5Vi8CBRhmyV=(b{Qa_aCW%*lct_9>F=qJ
z^z{^PPF-q$3-m6FJ>dv6S=^?_-l*jq(J?EJ{qsKQ&fv(vQ!4cvVNo>k+WDwG_xOe(
zxEd^|E=A0O2p0fM7EPM-ls+R=Ty)g0>vG-+M&EZBI98k4JirrI3DRtodvU!z%3A=o
z#YHK2S#jVZ9@=mK+Rs~~@RQ0ue0G+J@})T$k~u!%#J}`cM#1Y{TOB^vCU-v?&7X&K
znTO8FBv94CNjeblELMdlF5J;No*|3f*Ah(3I53F?ZR3s?nmCn*Y^8y{K8{!atvOQO
z;^VCgH`qg$CT}`KI5?7}D6dE|9TVcKhbF>)<!1P=f1lY+zoKt*tJUv4HVxRMO5sHH
zSm#zIvmj@hR0SYUM!C|z7-q*VEH*N1L(I7EvJ6`u&x|md!GN&#5oU@nt-`l}-GCrd
zs}pzcJP<Wf)CwE%V`U6W?zAE)5gq4nXB-jF)}9mE9L^bTEsXe%UM3>+&$UuT-1O4c
zSJp1c<+B(HS>1nU^d*N&08k}U9C=o^0J-k7PXFBE)1zA*_R_i)(R<7M&bzDCe91iv
zE}jAen(`*)3Gfle{YE?B0!h62Qs+E}GDGLL0%}#xy0WqOVtT?_k~PprGPntJ78ur@
zubv4USRUEUd=f*<@@|Uf{j~flxQ!)a+mV@S-88bxO6Q}xS4Vi3&#R4RG_{cbmR|SH
z_wkC-&5Y0H96xLo&n1Z68?mb%+GS&l-dHcsy=tIiBqymhd`d6e`ig^bp<e!>9&iAO
z3MBqc7w2p7M`AMKmp5KLX<*j9n#JxXE&2r5Z8*>oP0dSL39uZo6ge$hE=2%$nX|)K
zNvgP5s!gDW%F3xk4TAm1qaL4H*>vFJ+#!fn8gy&j<EDMqU$ynbVVirWW{t(15*VKQ
zX`E)cX9}Z{`6xTbig;=F+XeE|wS$b>feI-6Y_rEimrf^b)~8E-dGf~*h><=3;cs~b
z5UCbqmj^9?8|edq&n=sE7~I=5nkyxD{&)E60uL$g6hps)08aP_V0x$lQwD^daG>5&
z`3qJQgd^0G?V=h}<6(PTpTKmcuBO-@G!Ij;L5y#vp1dO<`#dv<Z^o^KMNii8@6WG9
zzXo9<Xh-K3hNWUawGh7v$S$`L&F|Sm9%VX^bji{I;^&96c{Qa{dEUv+ZQbHO;D4y6
zOyY`!OaA35gqC6sT%vcd{T`}z8XHar5wEZ5a2gf*eDi&<TTzN>-BRKJkVfRhWaiSG
z)mq!}Pa40iKM1htnD01C3y+|X=nKb1MrRis8JISZ(DM8rN3=-Cr8`)0o~9peZ-VcG
zw+1-PKXhlklFvy_QrcugRw0PmEL}FL5H=wFR(czPZwEZHCN9)$m4Fpwb~W%G%1YFf
z`|<3ZW#({hT!N?XO25hbN2<2@Gp>2D8=CFSM*rKl=U{Iut^#0ls3Ih>i2j1UfUl2A
z6%tcl`RI;pXn=fbd_wF*UFuu)o<jil(KQUrbC7t9ZZs}DDoRmxfZOqWTD%VX))tzR
zN1*gdTPyf}2rKGX8nHo(DV1XiySTH3-H~)5eVNSS71_lLqei6Oi-L2zGk@PT-~Q~$
zK;_5MF99iB;#7MJ>5^?X_?iC-_)@;OI>UOu|Jn?)p~Vmc0<kkpHc1U%`fF>M9Z@vp
zcHsNX%1Z60ROBqvspKy!AhAm`|7&n!xCx_`)5nH!&?<|6jxt+LtRi?6Rl2X>>?eOK
z5T8E2)%q^q<(-iPV_1Bmdl8^}c+(AWrskMlb2_>S%<_>Vvl}<wnBI9pA%8Gzon|(=
z%Ox?F=$?{&DZxgz-5Nc$wn9NjeV)Hq;_YZR)k8zoP?PNPn!52vpe__(hQ0_8+FJhL
zo*aR~oeAc(w9?;)X6r{F=C&$9`u0889RPa-I=wcp6hf7wCD3w@nYU^06b(gmu%7U%
zWMBU?MLz=FL7Du(7HhP71)3H#P5Ds|(=H9hyIk0btJ-o2_yWx|9<K4wzTGtHQMRja
z^OJ{y__gy{Uq5)1j)pKu@9-Bpo}oIEi{prpms~6}c2LgSy+k#(XqDb`C0)gqQ7<w5
zoYp+A9}-_8(`TFG=F^Ecx!Q<!Z4|gPY4mu;JpSm;42G`UR>=CAvw!;S6*+-(>aQ&N
z1Fg@?2WwuxU$Rl|!k@qY24@7Hlv&P1367Qsx#oKKu6d6AopG<&cyHZ*A3omD67fpq
z5G6)c9B%^SGygJqx${0^OOmgf*>*Zrq^Bv7GA+s&yMb1t`o!Uxqwr<-I4=4Es?FEs
z1bgiFFMVBocuT}-Lr_ekKM-ruG3@1y?o;?Y;vt<e<3Kp*lQXd%l4};9p<;7aeK{5v
z(VyK^$41FhcGn{oGF(nY=chc(28kC{Qr+g2fEzmCa`zUOLr049do0ogSM_^quy*E;
zM*o#8(Ln3|BhXp6vOAck<ymGD@84`Lms+gCt8?g^qsK(_dB?ZjrIfou0Tx>!TzX^K
za{QvYmr`wet~A$d*Y&Jf553q?^NF{OTSK&CsY5&Eab*vIB$L|Fo!&2-FBR|qj-2Ov
zpB%&;loaS!`qj=g!+t40NGa9DjcwZW!tfN&gDO#waVty6RR_64321tUR(E@XaJMC?
zbS1++5_!VyAk(1$V{yMxUJ4{ge%mIq8Dw1qvpfQ&JefZN9WvP;3;XmkIjaRsPqo&C
zv5}RAG9f_|q)qsRoo3WFk?uUe1mJD}arjbu1P0qODSUIti=+EJ`PF2n-J-9RptDr+
zWS#pmV#1(P*fX@{S(%9axZ=&C%*W}e16PFF<30s)F6@AOrB9b|5$y>aL8cz=7|2Oq
znhWNcKqTF+f<0A&em>f8G3u~j^tOOWMw0ee!H(%9J-u@Ga$(H9Le|k)d;TMkg3Zon
ze!{FUEngnyOLDOz<09X<ZEjRYFQ{vs|2=!-D><Y;v%Q7%x%KUU8K>7Dm?V(uM)tuk
zCC!&==>1F|N<Cka`up;&Z<Zr|%&+yA8KqeGhlIap)k|)yBY&;;5p1*e10441dJ&Bk
ze})`v=D>;XfbDJv?csC7!z22n{Qax$l{3^TldDYuL7=;hV1*_UL$}Ur45fOEdfzul
z9fOOQ#3j5DF&35?>o@{ACt5k6?|uB!Qm*mJ@LJzv^7{ZMMb8*}*V)fR_rj%)wUNZK
zUH7HBx}X5Cp@l;b$%pVfij2I~_+#X0{YZ_jo}<*=Vb2lg&Vnow=7+Q>_~|Jo9tRb~
zE7r}z(dy&Ui7uHJgLlz`0#t}<{+S+4H@K>`<(bNtq)$$IRWCL?D~Zd!ou1*y+fXwz
zT<2*?yC4%?l=IcmsCb}aSC>BxZ~qj*4Y-($DZZ2*E?4q!`Vl0Qsl?#SQIW_ZrJ-YC
zUuH9rX=vhg&+e*KtZaeB|FPc|_KBS8=sDp@`3WG%<&WCtTKMdaK<zT~pRgNkKbC~1
z4>&67h9{!n2t%jj3klB>(uP8uoi&-n!-e<j=1~|0E`mUaD)JOG5;rist8wR*waC0z
z&#mws=f{Fz3kR`7Ir@lj-TJS2ZAb`MY}0C31LBM45;pUjz)e02kF<G+7lK;K6rcLj
zUbQ=%HkO3!zWW+kZ3p>HJDK;6yOmp?<=st%R{>A{70AtqXw#>fz_no3fY+hAtGCXY
ze=s;&$Cg4oW{<TzC+?|B<vz9v(b#WF+Ki<jC$}lT2NBO_#1;5VqM^m}AHFkbfg{$t
zKg*d=jpSDglsg**8JB$8k3f7%DfAvN$QKzoKDn2ugMg|g#iB1~5;MmBK0X4u-1Dye
zX(Z@ywQlhz$l1cwMbkUKicNnUnKHcx<lU0GMYx%$!eDvkv8)*A%z<n<<Ep9WzvmGL
zrUiJ0AAuI0R31<JzYb4K8fB?6151z&hX{rHH%E4R5mFWvwWAV8pls{A-%9r{OEyh)
zASwsiaeT~I0mJ@&gP*|~&E+HO74KR;b_Uk)*Dr!~z!NNVo+X7jWHkLE2~_}b-Od%C
zI@6uJZd`DTId+s-5|qW9>=M=F>;aw$dr-UO{A9z@8WR5rl_#6SmCe;X5bMBuPREYq
zM}K3nJp;fA#7nvqSr}Joyd@D#83DY4H0s8ljlooXvs7#g!%h_AdU0<zya4N_ad-uu
zrqrQEH!LlhSbdEcko(DNN&U!kO%N^&RLl(sdI+n>s(ij~_?O+<ZHX>xhTpfgaWI-U
zpD@CE#?)1#`Rl%?H5eEDq)fqM!<FdTkJ72iy12;J<AKY}BVu*bzIuN*Ri8X}@bbQ3
zH&%sV$CQ7&b+E;Cfcb;FIXBe3`)qG}CxT^SNs4}g8~(}-g3r^F0v?<sBii~NNFIT{
zH4ZKv-s1}XD#)-K2qMBS!v`|BI-kegH<ljMv38HF#=g|ENlxF`Y!hj1?UqRF@Ik6o
z0weR_NF|w2w5r=L)wF3(cenZIr`rlj+$Dk^ok~{B7?zhFIVYLt!6(5Ssu%mvx&B`O
zzZJ1E&eEr=4-WXmNf2JSx+bzYW-|^h(lm^<9;6z(`0#K%S9D5_=>}i$mt*{?!6eFK
z`sxC#r04|;1xD7fpykX`Rk6>u6<#IB$2rNd6JGZ}Nczp5s6S8ywVq#&I)*Od?Q6q|
zaIC_6$@|{n0r#j=56{sh3Q)S7(4R|aQ=tAeofWH!VsZ`a9Hk2x=O?c(zSAoB!Ttbz
zQzm&;^Ow*!&u|ANJr*D;|3&<vLU_9f()=RbN@j&~^K0y`Q|z1N^>OTBvreM=O*5St
z^8i|<p3)kfozxXI=g5#hK$huFM@$GSx3}sxWU;yokcjLcPrO1SU>YoPC{O=Q7o~lm
zpPhg@-n`j3cK$@LlkJ<1$+RJ`N_=1h)4RCTnqcGSkSp|lVRNct3n1&+u_@w!LKae0
zMjn$WDc>$jHDp0D2&3`#C6r{JL2KMtZ|v9A2f@iTg5LT85oatOJ;xt!er8A<RVp8P
zh!I9c!3>Jpj-l~AhljtjcVqqswN<X-FH-NWNraw+2f_I2#t_fACgtwYvTN3UZ8q_j
zNoTaYRh7&Cz?0Iakgslv;rhFM?lA<G1Am9@M3OAQ%+sh=BHR3jc(h{-TGahRLrqzG
z{P#%Xdl&g;ZhsU!H1z3mf1xUNu|!Ji-TgSD8tec>K*z(A%;@lgG-{hfL4@u2@va@@
zWsAMte{8;aMIwaRyEW^HW+OR?5?V7tJz0R{>_=2L89d=%P-fh=oiceFhBQ3)T~|{#
z5|-LFiHL$hTwHt!nN{bT*q>LEYBKgy2QJS2dd<R?#V+^mR_kFC*onu+_)vp-1VXJ<
zU`#%zYwwG9k`CMaan*)pMYc0jrg&sT?IT;acb`*mUpdve1E2X)E4?wRj@(e}@vb_l
zqwVhbWmOmvr7{JrowB*W$S;;3<J|`^4%qHvFj`Ta$k>ecXJ^H0qVVX%m7jHWG3d44
zHn60VT(_Bk5WaN0iQ`ZPP{*j_CL1;<VV?6W+6$wq%iC`~Erfm^fzWk4YYq-lgnfjg
zRAqX*kJ6z&usPpbQ`eq!vL_>mx40u)Li_>SsI!2EMt3RrL1`dlJ!e*C^vSs1CE)83
zs9QK?t5!}*-<e1E`qw!EZ6<6~_<-X-nmbk()vi`^ak@G1{}U1utcAT@k+d=OKOr%P
z;|c#x{m<NGQ6ius#m@riP83Jpl@AShZ)?sc`}GPQfu7#PD9zo3HX*Avli++;A5mq=
z1(j|+w~t{@9fee5oDP5a*kz?S(H8wKFGMf!pG#1-@d?)z!1Kiso>Yz!^`5#swd|gI
zT4~O|x;@jOyOPb|E^!``3-RQfkDCsqVf>S#?IgX+*{?2DkV_;UAN3j@?pNOZa7g6{
z#Qk^c&x;Rg2MO=<dF^IyR2pJt9`ZN{py0ngG^piYs3P@tWJHD3`1}eL61LM!(qIo?
zV9#EB?`I1lC&dc;)!^l1^+c(6f4I{?ssH3^#-I3bnWtU87zKJ9>dl~&hLK0}(sC83
z;G7v(x@Le;v`XGNtGDbeuT1NW`5U~iZ1XZq7uR<lf#Q$H(iDJ4a`fd=ACe*xE8OLk
z4cx!|?fZD)ec1N~a`l3XloLq#BrHgN%e3lu^1`WH0Uy2^qiedQwfvt2iv>`-G^y+d
zqf_+<GI$98B($pC7T&@sQPWVz_uWr0?yxl*d69{a=%@@IfHQ;Z{E4h3R2#Am*d4ez
z7t}QZlPh?X<LWC?)0VP0O6t)wTJZUZMJlvyBSc}I$To0=f6`8)ThEKF+u~_Xb(=4n
zw4brz&YhhL)#8=g@u$23Zo89|<{U$txIB2fCe1R>_@BQf?%d0<-2$FX%c856%W$t=
zzhdwE%%ZUO!hmsJtX!YgKexJ{Kx@!iTAr^Rfq2}kkRoW&lJ7v(+(Y_==(}@}$qyC_
zy`GXu=<5QPw7$z{ZHU0T3nVT}gy_$OnQCk2Oz->8`FA#7C8W3aF?MC#-NKLN{H|TB
z9N#XHlQ*Ja*{@DN-Yefgo6$b1bOKU`7O+4-?)1KvLNgV9VMyT}$GfTNO9J}qrROUd
z=xrzq@<8WjqetVowj&T%zP*8+=ZCfXYo1S=`!^z@P!i&o%^Ck*#G0jK)zQo)A53P(
zs_9*zI>xnIUvbqEtD|hjmD3-p0PrU~Qs)Y~^%K3?x@8zMh<YbrjP{aHsxLtf1>Ayp
z_^C^$KFRwo8dAZT8D%=da+5J7ASu$3wR$tkYS`?Vhq*(;K1%g=;<Mx=X~jTA1mAS_
zl^vIVKQ#Q|75nP+{mnX0K*p3s+(j^B$KqxH%j^F!b6LqG%GB4AtXfn*z$e6!uSvwU
zCc{oIV7~URkU@zRs6<d=@0<I5UpJEszXtbBMc&1fbf;y{KWgJW;0b!a`{SaedHUkU
zpJ0N-OZFpB>j^8RE?Gh)ts2hz$PKVJnIC_+LGg9H_KpMhHg$_qN`?=Bx#?%1ah45$
z!F=<jZvC+f>U^#a5;2p4(z=9aR(hC-IgtNj;lFyk@PUTfw}f6)w4NTB7h2j?g;}^h
zcyn>e#?bI|*X65ShOd8|{2`s{@W#msH8wX#Sf93=oL#L+52D*;Un;z(;UBR$ez;So
zy1BMEy>Ltgx05E3gurap4~$72BM`TMGxQ<c38`uIgzHvvIObcg5e9bRdNd@}vlwQJ
z1|mo;9g}`ZC9BEuwu#7ZhF!CdK+2~}cPEQQ^qF`?vaEz?AJG!-5WXfpLJJ=hjj<oE
zo?6Vjo5aDyZI^pevi4@ulBS}u$ETlrGN-vq1)qG;fXwu%?bN)_Y}nVP;T`xzubdgR
zTlOd)_kgtAMzqRP?;n9urvTF<(Bobt?ZkX0d8w!OF0GV)8Uz0TOhB{0LVAjCZ!wUh
z9+xx%-n|H*k+eH!)ydrvF!woKKa3&yLS)j3BamCk6=k&mCaLZ1%Mz$_4e{2g^r2ZE
zEX<CKn9GhRahEHGX1HHGYj0<E&q}U(a`xq8qQ-1}w`ic%z2ySD01Lvyx3;di><Gm2
zl$>tSq3t@V`nUL!>)ya*>}=061wn4dT~d)N)-IwChq6Bl+GLB}&zz6n^MUt*Pe`L>
z`h*K}dJzMcE18aQ;RN=}8j=C6M76$c&o-}=SnY#9<!qyPZm?`Z(UX#ubX7Jt7%I)!
zKk_!|7BEH<F$F3}Xk5EmRRY(cW?uT8Y?bL+qVW5>))HxxCc8VjrGeF7$0%v<J*e7>
zs1!5PeFuueTiY!1qj9Uf^I}0gxU{-$b+lSfGE9_=X*_9>=irT4i^C4UIgdc^p|eMz
zKNt3nKqk^<rJHYCL}0=AUsP3sXSA07piXFn%EyQ1upWqC)vc%Yr9Tvj^*>p{#v750
zFgEn?5$O2`z@^O5PGP6kS}nzX;%thvvUcV7=`&fcgqS#^)=ZkUetewQe2i7qvkacv
zLlgV*mhaVQ{-{p*K*K5Ri#S%9Q{O>OApR&SE!u-S9rrJYD;<Hp-^9^l0_PAeTu}%f
zx}Y2Hir(e%onXi7RT1Zg@0)QfwKJ+2qzXTKoJfqfE&{MOksa9_*f*Ij35uTUGBq)U
z(h@ES@6Sr`#b>|f2oy|w@yBMXbJilzcOqksuh=B(!Z*S*x)qh{6PoL|!qeGkQEz6<
zyM*7^cEuLqMa({&fA{F;tA(vWW`g#2pD7`t-K+rFvm15R)i!|C{gx^-m&En!74nJw
z(<NgES+#_Z4+p@C8qdHcLqjdjhjE(|a4a4X?JMHyD=TiCas)D;k+<kwZi;Tbd>z9P
zdRZ&WPu1=Q;6mQ7mP4?=aL^?!ea`FTuGOl|beC>O4(G_>`1wV|Q{?T>wUmHz0paNl
z1#R!H>009n_uqR%#Koo8BM}d;zqfTeUse``^9ZyfPq)&-b2IGa=AQ{{8Temy@2_f-
zi!OLtaK(Z)?DO{TqfHD_ENS91TotNFZ12vE8dtgT?pRLP`29cRy$4uRU9%`0q!&S&
zG$Be60TGd21VI520qGqCL^?=sA(7reKtPb96zN1H0qGqC1f)x-AyT9!)Bq_LzjMC#
zeBZg}{P*5-@ALf6{r^Sg*_*vr)|%OC&8(T7Su@411(luHOsF+JM~T#>on#%8l+?RZ
zLWnD~uI_u!`()VriU<2U7Ng5a1-_}dhfooGIhe<}fLYmDH^!X&Z4_K#gKFA75nZo0
zuqNada9u0$ObqTbl4Ndp7gJh1ANu}x(-L99S@}^<kvzMn(@ZP1bmmD*PD9?r`L@zT
zXBH7a-+p0>UYQ4@U4vIYPVN;G?g{!*-X{VkQwz$=2~<w)Ja$-ZBn12D$OiBbx&FmW
zl<D^!&~kMeXt*}DoqxA;E9b1#<cB0aK&5%i(>=}o-sVEJLjxbp*q0rVon8<MNd$xe
zIlx!CeF{PHn7%k4q)F<p`UVV(*t)rg+zlG_<+CFi9i6dsD=8;=qxlad14HaJCJmz7
z1sVDHlld{^s<wg4UWVjO>p~vE-NhWsfo|^0=fWPXKv}SrHt4*m-O}g0Z4)46pBt$H
z6uik}_wJ0rb@|uD_6#AmTU1Pt_<)a;2nP(yn>tilg=yGnGuUqfLPEHtiLTi)$56CZ
zHN=6OYue87+UYJ~xu;$!PmdHGX<q{_r62oZ?^oe?WBP=uEPq<&Ihv+FYdEJ%KDYM>
zJK;sPW%C^Q-g=VvL?-F2y~0+A%&lgdJ)Ej!NuX8+&V8%0;daK-a>!*CJ?qBJ|2z@5
zrA=>2#3B=R0V&s_fV#{pYN89tk@n2A2omo}m0e{IPHv@Fxqkds@3Ula)o40M0qKdb
zKYncU1tdM=Mi2)nM&LD~6?Fx(%FKEg=A-QfW%a@i+z;cMxjeDL4h>PqofwhjfF4~6
z>u}cL>NIPMh&T86edt%$E6R7K+rkR#jdJD>gdGll?XlT+)(`>9)j;~)xQOm{dIEo7
z$}^F-;d5$9>Y>yJ2@IXuJ~=Zx<mbkOA5I8Jcnhg5TH-;jd%6|o_L9?NM#!lb3l}wa
zq6TGAGW>9rWx;Wb`C1e8`FJA|LIQb_LI=7!S)-^G8WMCN+X7}?_pZ7oJ2a+#>kbM@
z2zq`fC=k0`d(EUtPwn}f$MzwTYQnPCSVfIR|MoPx0GB9jvj@bJ&6JyDVs#1XTw*6F
zgmBl%$}cHhAc8$kZG%Nm%_B%f?Vt5^ikA(@Bw)|yQ~9=Z`-<Mgpn&>;d#DMz%Rx2?
zkQdOXrXzUI5<2PT`&X>*v`TgP*->9e>xckusAHMfU%{I{HK;nSQk0_StR@vT%ey~1
z$+r>;xF>7pMy5y3i|AA{LqEKDE#j$#e-FOyqGGyHIL3fku<R?q?yfAAg!Ze2B#_W9
z)CVn@Wjz8N5dmb-3HKd$g2H^zs8TUOI)|$v-DBl9`|ZGW9{b_&;IJRpW)|RRlWjRr
zJ3Yir7omV`o&Gsz(iqX(^ooN7^(I|OVuoVvM!#w`HgwDSHfl(ar~f0H?>jv;R&w%x
zbqIpn+n0d{Agp$ik<kObTyZb<Yg^p7IxR7|Wqlzt4JHE+w4n`;U_`q;5y0$R>u{vY
zVr9NzeC5OZ_yB*Q#xsQiVIj!1XOsg<0IbSN9;A!FJ>~1pb7Fu#|2daH8@rpzv6Crf
zT3?r@&S*%N98nUI67IXi1X4C~7hK-9+ot21OVZ@ey->SiPr(Uc)>G*AF1`-nWsnfE
zAWL2;!?iRmk2+pJ6~a_Jqidt0OZBnBg9s4YS8O!#9;;@MK+p67<LePE`W7JJ^QRrs
ze1yp)pU%DbZ6W~ag>xbT3NMByBlg0$YWDnh4#e|zUga<MKF-8AcdOa2?_n(=oXQ&s
zM8K+@GP;Ps2JA~Y<0S%4O;d0p8=NAMzm>p5KvX}bh!EROFy{N)?wWt@;etnA|KD4Q
z8eau+$1wNivyK;Hbi*J>rU~1oMS*h{%b+jr&nx?NXidIfu^95Zr$*-~S>s=!PAdU?
zKxvLwz<*FAFTp*UfDKES91ItG<BQxFKG#|U1SwPA9d@47Xmt@t(I?@^{X;9>8C{ff
zn&kJy`?R^WIPGA`&uP`HA{*W#O`j0;rRxlR$}ybnF^BCWme}{YP2;d(jb4O9i^!{7
z#e7@A?<orPvEO4-2bKf+cSOM<L_nuHPAa4dJRD#?oy)$2wJ@%mv7OD0;Ch$fg`diP
zcP<2NQsTjm46Vq##>uTzOGTKv2K%)YQP96i>#L!h^ysyXEi4<&;M6aGCC&c0V2eSw
zzdP!IeQ+V@o--l<$726%!uP%oN`2Z-nRHHIs>hMa4R>;&xoaWb^?}asEcGxT3`Aap
zQy61@I<|u=!`}Oj8&Dtug610Wl+Q)hM`N-gR!VU!x~P;S(YeuaMbXl^K%GS$D(1`4
z&r7FXG`s_96ON8VLw*W78?1-qzvR^(dMh#s5+DM2ce2h)i2zlp)v15Zgkt(9^NAj^
zO=iP&OpO)E^X4A{Tl@tkv@4?#2$@9i4Xj0<NrTLoOFs9~@#7DvWMb{mcQjX(BWQNZ
z%HHzHunE*P?YoxWi_f*>ulXY#9LWDQh|C{gfw%H5*^EIChpLxewubGj{(LoO^(2?4
zpQ2@Md1-rMnFt7HYnV9`oP!=|!S^q8RKskCZ2NBtY%XPhH`FLW(tMGrvZ~g%_X{gM
zgG4OjR3)4TtmiJ8HqZQzTqB_`_Al5j-<b#i#Vd`n;SO}qf0>BY9^8j#i~YhumU(2<
zLIh`5{(Tj_ZBReHxIG<11iTJhVBIcoOdtf6;`eWzE7}|kEAORh)E+SG2;d4*aHM@3
z&=-GQ{<o&?r@VtW{}1};KG%c@qKE(hrb%S>uj|Kw?dv&^CEgVKH6TLkR2hh+eD{xw
zPk92Ji#Man(WL?lRmJ%Jyg2AloRE2^8jEqv^?yHXhJWI+puC6x%ehm-qkq^Gqqji4
z9eBi|XJZ-YI0%Fv7gU2du?fq{2<5KnG#jfxRTbM6E_U7FfRe{OG7Up(kMH?J@HRYx
zwkeXMs+m_jBP1cLfgXhWx(}zS(>BL9CAy2~BmJLy)8wm=@u&s);iKv+?R^|VJXsg=
zzPr;+A6)m%V81dr7<fHgg|fyKwZo-{DaA26_L@iS#(lC0oICcL^3Yz`eNVY329gnZ
zF43oXTJ(G9)dQ0Fh7p3a1pVHQ{ec4d$f%Ly%wFx)5*0xvkPFoi%{POiUC|P|@_U=|
zo#?Ng>1{~3Q!<Y<5#Z1vH9zwvfSkn`+`4jEq6=|<vASwgx=AO?nIm)M_{<@i{h8lO
z|LeDSihoc<fxQDqd-Vu&!20C(Jn1rXU>4Ag^byyx(OoIg=eglAMHTn%Fz_+TQ8FdY
zVW-Z`u+Zc7sUu(%u&%wp6&XF;5KG`6W$q`tn|jdep37wV!ZF${@1w~{#tsocp+t}M
zS=Q*@QRq@)ufO}#JbZsx@ZQ2xuQq1}e>aBb08#+pKHW|3$+X;h4bk8VkqnOjNj<et
zk~OY2>I*)NU&7D)?WY&3pwe*2FS`aJz>f$xWVX)whjekb3gZuwKmXms^WOn!3^sw#
z){mbd*nK_QR=VH_Xzkw`c~_&<mPL(YCjzW=Z?<DNvR6FmO#{2Tb3R2?bliar>r7iH
zlvxz@@@CkcGFx+9Xc1^^b^yBB3Ip68Kv|^ov?GJon$g3<-S+1OPyL;hR0D6nZ`h%{
zVHU0I`}6E-i{SA`{Ch5xQFO)%kZv;t35jdtgupiB59#zWI_Gb^N%m)wRovmvy<%oA
z?t35R$^Bzv@|S;z#LlJ&N_Vz_KttkC<MJClhe|30NnmRo>e)EMuKaY#%Y5u0Xjf;y
zyY9-QSnsxD^CgUqgzQQXnaQZ+bmCFK)%aV*B-V`QFToR6mFp-w^B!8MHYf>CO4uTI
z?q1&W3q;ZQ``yZ>U+fSw1{A>Bce2n2$9Ci2gI9@w<A(V4XSWqaKHbeE{GQvrRHS6i
zjB^je|L6zcdeI3;w&XT>kVy=xwEwa3l3pYbs|LEN#qKMc;PApvd4C4L6EraY^LfSl
zu_EE|)wtZX^2)jvd@_N05yXfUcW|gq7fbNVE-;%B^q=4HlcrYpSY;8;e|(!g8NPeL
zj=EH3YHGfUK{A5>%lVf7En(IF8j!uDDoFv*<*X8};dIcWy?GScu3b4<$2Hm4c3dV-
znY!axE)CVI8Fbi=h!6XYe>Tlc{Yoy#13&2RfF_=?E^TkEv~yuYIx@kmpU14v{Ksy4
zm^gHIU;1Vd80VkKZn-1d56&3SZz?s&)aEbzbr>*K9b#eCfaw3R>^)%r%Ba$yROpFB
zDa%BByw93(8u-9tam!pp$$jk5h;@HD_1tBXApOJLvxO1;-U8!0K+{$%dWU9ic2FIZ
zku|8~8W#RQZ_*EHjAg(asACV1TU!fCBU_r2u-UfZGG~(yLG|N4$r?to&*@19-#qBm
zzU@vk@hh_CWOlO!+Wu|xyR3=q;D#<^!p9+Y$&x*_uMLH3i7!@J&ZN_V&6<lQHlP`b
zH_l(V02$p^>(B*Vc-<Zz!RQsW0&RHvCE7qX5{I<aLK3aKj+;x!t#7Qq3&`7@<{0A7
z7ORnYS`XXMkR;!XrJ#>4f<9#5PEnxCTL)sYz(B?~J_Z`Vm>$^*?{?`*b?A)xG_+~K
zPqP12reEenxe2T4VEz}&AKvk;jk)%M0&L^PtD&Pdg)LBIN<R>5y2RXe9sBT{`COU6
zs1Un+?ufuG!9DIs)Y)zKjbwYkKu@q2GnwXKrL_A({ty9n5GP2NiDYVM!w%SH)}NJQ
z_VeDz-k#W#Y1s;iNMUtz?OD%CkEQH&i0w;di?ftE{GQ-_C@>}+VSUojSq%kew#7go
zD|vic)39NlJ^lKIDZhIwBLl0KoR5V9gj*le6sP|HOI6y?fNU_MNutW`OH@l1p!;bD
z1{1eys$s2?F10tN+%eutKlnvrzrJ)&q&^kTciq8rZf+6*8numWb@G1OKQ#D{U#i>)
zSN#ABiMb3*yC=U=+qse{QlRBMP@RK2JK32$^jQREYDagjT$^ldm6CB*nQhcGyH&sQ
zqIJDw7wS9SiHYe2(gjkQp?A}vE)d&EMHQdp`}cr#LxKo_`GkgB#bWmaZ$DUvQZ!YZ
zS{!!}0ck_$ijbd!%7>xV!Jd5u2oY_pW|gxNH*Cz0;VQP@SpV4kYGz~V8^?A3br)Rq
zH_F}GJ`yR$zP>xqdVjqv;l;g$goEapk1b!lKk-p#jgoNa)bL6P2>3Y+$f=)<sokn<
z5cFtxd^9MsW`AzCNw_j|Sg9Nbp_PJ={LXnh<VjyOlMryJk#NfR0pKUg|C+iJ`T(=u
zpN>99X+$V8mTf$D2H#TGau3;gNBxSLM>EDz?&gg=6y6CHkB~($YnFRV`n#HZajt}B
z512&N1xxYi7^ZtCmR>ET7E-4<Sbh5fCCJ)h;JttUW&AQh24`A_6-3V%Vcb`tbA*A~
z8()}C)mKbjRfK)^noN83ie%+8&gWe9w+$_UW5fC;#yZc-zH;kI@3ANSYy;wsb-=Se
zR{lHwyCm;kQ|f2}<}rrOjY5O-_3atQr_Ju|izy1UX>05V-wNb!IpMTdN!0=T>CPgH
z1Cn#_EaTu_)`Le&a6&th&$TS@LVhvCDsm%3;;t(16ewCz$y^i45$Ak&cF5oGoO-31
z2uLmfZns^=DU2)HdfEbGVLR%_iwNoDy#ZZeLq}#xs`F?z*?ze#vnmhnE+e_BBG(1U
zspBS>jYHvA^`5$n8ke^{g=>DSz7^n*YFbOo^=SIE+mv6c7AVsNLHHP8;i$v$D)2Se
z97yCYZ)S)xt948=ZS0R%oljN3_OAinLBIT<LDd`6(Y!q4@0$)?4-~70M0!_QGl5+6
zX0SXcThZR&LVix^%ckOgehu*KPx!l}e!P$8s=?$xXkbPfzJj%=$&0n!3mGN-$V@MF
zQvU9ZjA#{}=sYz5XcV>dQ=?11AnpamvAfmB46A~on1{eXA_@;_Yvlxx<bkM+q&Gav
zlg7pc_0ZR`5l^w%%PBxpjs3nw-if5vzWWw~iH?{z1=XPnYTmlEPds+G)M2QS(d-&U
zm2>aFRPtI3TemWGIi=rDURB*=d0#WWRL#9_q>EpQ@G(caaZT|0YRu9~Lu{MS0BM~4
zSY&j+c2eWF{jn6N;r?iKsFm8&SH5<8u1PaxXW2?u(U}R$q4aa=XhMc_;cw%idM*Wa
z<69GG((>Kl^nq%ZTQv*ycWG`FzLo(o69Igz?dMP;;6u(~(}C{4v!4IjOZ&%QSm(C=
zckfgF`W=)3!-fw{IfvqUy^el@{xS1r=pAOPIH4(D8?th@FkYmp1F0m6qepA>i&8mT
zOa2gr#~LU-+Ijo6L9oTUh6DV18TH1s3pP3%g3{>Zzk*}I9#xn+jbSKbKvtKQ5MAvH
z>cU9M^HkEkoU^X@t`ZtQFuyZrTJ-nR3NcQ?e7$h}ju*O8@77}(PwsDGqxt|FJejfI
zNjC&Bv+{+@lMCcCp~zM|*xmv|tA&}x@{xmfYUOhS=sTD@q>iLmM?M=G{G3`5=HWw|
z9N)+Kp*8g6r=p_z+E)f>B^UW;*3ON#>aNuZT^D8^75*42%Rkegr{>=$9u6}~4iFUd
zznE^%!vERj^`CNONc;(M;`cu=>Ha5xgu5RXGDP@IFv2iwUuYpWZ(y|W!;VD2<MFno
z47ybL<dVm0ZGIO^kt{Z=16XI0NZ@f{Ll`2aVGS#bA<iYPFAjF5B^QU;ZNufd^5tK)
zxLI+gmkY9nOyrz4aEW!`8DhM!mUk>-ORNhMODq`p3ydfYvT-%gO%HJI)w@3l-`i;S
zpWVn&h%3JC{oO-gK=}hsBnF?!YpU$jj^N-UF*rBBijaRX@Je@6J4CsIA*Ly%3mQS-
zLVv23s_rZL)Xnds^<sNW+}Wyce6SeZUB>$T!q5?`#wlRyFb2Jx8Pf$y1uK=`cHN5L
zsR+|4bNxpjQy(a|0DT*R%k>NDyJ4{SG_dtf>O$=BUAXZ2{&D$vz%C>NUwChVu%La$
z6LO&x@6>)tKsv^IAPvub7kTsUTq9|A>F4Pc{NZ7zyePR87v`?)IT|%?3TnvWl-vqt
z#uifhz#4u969K-#CTZYr)|ar^K1ea9zyf4L{NaN1>QQ->m}HXw)eyOGevh<hwU;J5
zuJn|A9Ev96?c~aitrlHi=_#cff#*F!EA2d+jG?*OWFew%A5kZ-!YM6D@+$k(<?KN7
zOy#lN$m<_gJ`UYHgue}Th00hX=FbQ>G8u$MeV%sJ2ut|gkax*hTVKm%apy-37IE)=
z-%Zo9Vq-hMm(&j1+qqO4aeVF*@w>^zaYs-Xh^SkpvkmiW`fAtOA<DzB0TG_T>}~I$
zc7Q1Qats?%IUcqEd^T-odURjRA}$A^|KV~+h1|MJt~(j`Pd9PaB!=x;45dD{aalJO
zEQrZSiS?_N-X1FdXdKS&KEq=u9b3D+oJ}oSrx3$2Af=ucXz(${cd1cwZhU&(W6NUu
zJU~v!ZoeWseq=Q%T)+FPcgXyvE5mlKP4kNySjP5l+L8uHn+SOO8Yf5ueC8e{B(y*)
zD%VF549kLvI9Cj?tKf<f4>m%345a(9Q=mr5(4JO(u91xIAda7mb)%b%(Hez(2RBue
zE9tM%Chvn~)`}iZe~2y4y?($!LP}%7|4J5WK+ZAYXDbWelP*v%@SMad66(1)vR2_`
zMhS0)LgX~8od$Hgn~mW)X?5BKlHi}feKUw-$~lJyAq8FyNl?COJFX;e5g5_|q~FFl
zV07*)`<}hs&fKzhQ&(*%P46ZRHw0{w2-I3>^rwV3t@i?fIFT92%4(&>iLR#h<ETyR
z$6NfVr*G<cX%cDeZl3pQ@D)4-#t{K^>xnZcL7XT<G!O5DWWkjM^XmK}`|7?OckkD?
zMt&Ld-QZg$8N9^)BL1L4zen9qfT~$OZowSa5@%&Is&m|UWcE@<s4}wSS?^Se<gIDr
z^qd%2hB1mW`D&I(LR<3aqr=8!BB1^$5y0U|U@^7Asit(98^`)N8FMSBNoab%++P*8
zn$PHXw&Y7--My$86XpTZMN3KH#k^e?nR7&4#+?Xf752lOhqMQHd)9uOB|cM0803R4
z8d}-A_M6%DxF+e7(-}zL;`cMAj>r2-eQD}KZyK!`K+HDJS{TJ5Ie}Laj3Q09CrLna
zQTrIN?;Q@}&1sa}56VlU%Jtdb4a**Qe{H<)kzVJI7{2NIlH(XiswG?*ktb)^Jyo_!
zqgjTq9##7e72wacY-+BAza;2j!%(uyF3Pb$GLXQFwwEbhZuz&TS8TEw4X^OpZx7ve
zTvV%Af(v7d^zl&eRc8X%d#xC~1f8J6z8TW@0$&%z?dDVzYB}vm_sA+CXkFcC<0ki%
z1dt~~o@mC--0xt?20nVXBJb}@=$pb++$gCX2uyIB2=K%-YmeV?moft>thCZY;Lhef
zv|qbU$tnRV!cWS-$&p=>_;#<%s0YaxSGV>yLp?Kjfhly-Z6^D-orGmW3^tj=lKb=I
z4gCU%q*F&D%|ZR-y^H}>tq`?*d(?$M0*@`;M(Z>dVPmYY_E+OVocU$E8=Dg~IIL}a
z)S^Tt#W&PlN0goRs5X@-@G*)EA6%O9qU!?eD|;DRCj=!d27qUpiF_IQ4^%5BYFU5t
zT*q=_PFC_0^RHt1dw}!^ths(=sC0}Q5#ZLMc75l$y&6Zy$T(#gU&JKg2e8g)G)0NR
z3CI9qyjf7vU2HWsz;OJwrhHZ7TD<qSM6I;t^kf|!7Z;?Xah{nv)0A4s$PU6J*jMYw
zzQfV<X*=8NA{)ksnNm}l!gpO5pWG4Ff3IZ%i4rW-hHa<#Pu#Yhket&}aGK@{3Djpt
zsl2YJ0FeG9F)nAZeQ?MEKXPA~$I<J1d+7dICV5r=Zs~q(c_>i4S#<kb2ZKR;PQ3H%
zJcBS15b0=01Qc*Ugq4vTQ_AaoHiNCk@rQ(VgD<j#YM%8BNS8zfWHhlo1Bg!CEI2@k
zty)*A0d2}>DW{6&_n}yIKiyMI)_xIzG0%Via==wiITdXkDO^8;h2$F%0aPR9M1ayg
zg5h871Xr_!f=gcjjZ=IE+yCBY%HP|88WpBMC$}Ru$998EaD4KmIPB1=-oLeDv<cHY
zI?H<#YJPu&vMX2spV>%PhZtaaE^QV=8ew(BV#4ZD18dA4)1k@va-n*Cmi(iC{YM}N
zy!JaFII%t&LHUliey6DP$=W|Yg1yqjyj!>UVREIa5?K<C|M7Ehy+9@Pf1i@k|7;wT
zE{1`jQloLz=I@<q{JkZUiZP9uO@EamrxV4>zaIXN78y^0PjRj+-|SOW1{?GSs^vxU
z>u(yZ6!E+2#DCiTqh9uT{8#qc{{=BQ=C?>k33MRV9|Vd(e<DEbhX<Ut*<$WgiW;bw
z%H5M(R(vTo!sCipc<_OBm>Me!lJnWK@Gs_kP__hU3iP(lO%H?2>einZs5t)hzSblI
z=Y`L05yub#=j?&;dqyx$v-BZOhptBgCb!I9%qOLu!~5=PD9%*qnNAhyv;+zFZR~ZR
zFlH}#Wm~v4Xi{5q7Ocz&bOY$pDBpInKp)*7zFLx>$E$8!v6$;mW7Q~EOw-X6Ow$xh
zH=oqAxRxQT16$QE^3YyrsC3_*+x_EB_Wd72u>F4lL{NLAAob}&A|N~gWHu&c!@N`)
zDLG2p8NiblR3`fw=6UBEtWerhlwsu8XVN_x#hi1Bc3JpRIHyd10aj?VN0hdw+D9gO
zN76#;d_jpeCdxD1Vzxr<hAOr8F7&mAcs!mVqRiF{h!F2nGQF+%-bHcP-1jKn-F3{R
zVXD}B#lmpH3UykvXF~<jM?3a#)Z9QfB*4U&hTSA2gVjqf&pna{bxF(OTt}2{I597R
zCxG;DNo)f!dKnJSv*{28+q?S+iS3aodB;gR%I(kb*ytrIJ;#?TQ3q-qJ7CAqf_+Pu
zC412stLt%g<9pTHFT7ZLZjXi2WL|fmi}yN-@ELC&O=;gxW0>6|bY?A`39q%f-Fz4m
ziA)%(9GW%$7@M{p8I)eiA*2QkmY>rUH9D+&q*I(BUt5+LFQ!vIxL9cz8nR-;x{*&e
zUOb6YsWNpkcY8tEGFg+h$SU~Z&M%VFnE{*n(LTql2Uz6_kRoP6sh})ydU|m4q?@zN
zt}!v9NificzV%*l>W*bDnC`-ca1)!W>_>PeljT6e+PLP_Tx~zVYTjsiW9fET5?3?z
zgJ0Z!{_7G0@Z60s)W(RFLD@tuilC)paC}W3MJvrBq8b|TRKgH{=gGr+eVR0#k1D+S
z`2gIrhhGEu2KRAu)G>X=={<uy6IrS9uKDlZKA+1tZ8Q&9{o-%^l6^4U*R@P6N>k~i
zzWxyB22OSf&dX1*y?^{MfIKOqiQ{HtW2$|!6t`J+EFdK3)bn@(yIQgd{1T8X1&waM
z;<oE2Lhbf;Q;B6db*+BEk12KS!Rga-zOrvOzd{k|x)22Lc<92MduY_un9`A%<BE+i
z`BMM3_qd<OW73ZP6`9kA<OiwWt}iD$5@<-zxp7Kpu?uaW0z2B##1Ceg^!4#%oj3PQ
zPPIX_U_K@r&Oc=J;fmgRyRK-T621Slj{4oIu6VVWyY;hNyAQc6mOHG{zw*{S;>tcj
zLv^wzb90f}HW{v)Q`@b~Hun`r>)RM?EH@!kIOd=bFVTgCz8D9|8Ro_ye*Z993GrWP
zn*8~VpG-kc*wBi=NT;jT3h3d*CYvEnOU~Zr1rNgBTcCevtqmi1mIz?;0CAj)<7m+O
zE+<OB@fI6<2OfC2S-3-rZu}z2y|Qb$ryXyU(5kKM@yj^9m8G9H23U6=WUKJ>p5|oQ
z&+7a}`u73HKi6sZpNume;C>3|4$T0(61rY#_3N8%{;0k+ojVztamSC9L-s(|X!<yQ
zdZ2kb*~RbFc_Q2eSH95fiFX0#ViS6rG$IaN^K7iA++f-_e+uOj0eWAx0(pnXxr)76
zy?aQ~C&NrR_xe0z2y7uN(;y#YwZPtr=FZh~lZME7_4&&F#=fKbvEQW7GB{DZJ&Fy-
zl+Y)Wf4w<NZ|X24);A97=ss?FJ0KIvC9YWKmdcVoAd)@+wZs=F+d`u9#j)hbJe$6z
zYxzPL<v7WNkgbK6XS?=Z9GL<y9>+dswAZir%FW_<Uu3wt>nmLmgK@o&wetj2Ub>W3
zWWVp;n&pVJF}PX8WBPN?AXKo<hRM@9(;4VsF|QWnc@6y}y;C?w)BgG3rn(B*D;4hF
zFZmz+dgY&kog+{OmAf#4_%NkWdtwf)F!X5r9oQD1&_FVGnSY8(OSAn{QfRu&S+??m
z6sB>_1fP&SJ#>0{%AF3c$;<KoU1ZK}cym)tPDBBQH>ks<q(PRmtz|I46!4|3;F!20
zyg^R@2Aj4?)-(m>Z_2R0(xFO`8b?FRMBz1X1pSE9!xSJ-!j6C!la&#>)ol%I?@)G_
zk_riNeiaO=Stexc%ix%0v?t*yvwZzmIi<@awavGzCU3wd5kQ&kG&qbzU^cbc`8tZP
z#9N6CJ-Y3OR_;33AOe^ajXr=LVT#k9oYlkZ8YczvU;6WlrLR$a8=+^v7H+!M{&d9q
zHji?w56hIM?BthH^#aAJ=`SJb;>HcebF-p4@_quGnx{j?hvs#MOv+cC(y(5*2RS48
zXEYCHL@TPtxs2jeL~mp%sJ^Ba&86cEh{ov^1rDi4n4k$z`x<8%u9o+B>D4>e(_LBJ
zzuXXhr9w#=P6Z8zEWejQ&{v~UQou~;^|_l^my;{HpPZZCx%Jca`93!qP`*cIkpI|d
zR0;v^2spSHnSkEWF!|O_2BfGS{J!YA$BWPJfp-r9K{}!9QU|jQ)DQ0jZW?Z0SRxe$
zb;cffk7^LjF3-H<)0)le9<OlewYCYexkJZleiO>hc3Rw#lNj?rTR2`E^aLGR?rR?e
zoTF1DpdU7aM;L|^y$oAyF%LW28LY`LCJDY`mnBi+cWTQYI=ujHpC(FRP5KJzJmYoa
zq=<kEiB#CuAY)vs;b;5d_UoE05!r^^U`oagdgVClHZZ;bEJ_&ROwOnXY)35&FD16!
z+-mUA4n$-`^?sMG^$@w{6~(2rL5Y8(nn(^ckGbSC0>_uAsTd@pbgmuvIr2EGi4+%I
zE)kr01Rv949;~UZ%F!(;YKGizE~q+FueDvdEPeT*R$KjK)g3mFjB#33?c=tJ&BuHj
zuJ2p0z02J4Ge(WgnEn`1ci4OA=S>OiN{d3X`4=fS?Q9s^jM|JlmodE{hI1gu2CW;T
z?Dl??9K<^%MOzRBzCH1)N&6;7z;NP-3gyf-#Dx1f%gC|Its?lmSNAor&M~UpsQ{66
z9L*yizh#78i&+UNjnPQnw+M7-%voCws<C%)tLP02SHg^zXnW9ftxkDKe6Q4DDW51c
z3W+oH%59Xh<kHlBg27gYdfHtT8dTcUKJecARNf5jA_621-)l_6{&?PJo@Hi0+{=H@
z(6(_k%VnCD*ckcoGwxot13krNBS#~7XeSj-R)!xr2XntRkm^h-mU~H4QSl*7>k5Fh
z)eOxR2J?&t(~pBhkz2Z-H<}DDElywV9sF=)Kfs@+shFl&3Fb%FC!x3{&K0c3tB&#%
zzom0&VXrzR6j7+(>Ii=bP}d+MwRzXSnc7<ec2v_E(Ci}uB#UP%bQ>3VJZwTjKv05-
z5(zfw+#HKSL3jGm)BN!179e`%c;w7o?D5xBBxe=yWdWPP2Za|yz56NICXR!eL%^{M
zt9NBQ@<Mz1h+{P(F$}cZ08jvUtni^w8}~k4TIKD{ukylg1ysc{Ts;>J&0BvRPPNXC
zs%aXPecC-F0>%w4?2LAK`K}h)V+`gFbeRa4eb}GV7E!UdkpqYMf1<_9BX`ylWL}le
z@m?9q;J?Nb8GnK}?IzF&5&<_oG0cQUGa}%(8|&e`IMw9n=PNl)e&wGEY{1w3fV5yq
ztU)(~0?gTv)Dl$W9K2Y9Tu?&3x0&XRUi-bcK>HXpL$~@~PVv`TtBFCTs77~#w}lJi
zT=)G-SODgmw0a4>03Cx$-Yg{s`bODN`KJa_BM#!YM7ybU_Xk%Gffqxc<Q@=EX=C->
zF;<i29V#;Vl2j5+;j*t@?qSs^RxrP$OyyhwXNWFYIBb~CXP<_1%lCeg6JkiCAKQ@e
zbPH~TmVMuUT3_WA*wWOJTW0jc<=GQa9c@8c&E}ic&H{dppZc%Tm*~z5i3&KZ!;g;Y
zlp^tO4>fVh=-|E))QT^J?S)18Pi&#cg#XPUBB1hs?u(=2Lf`HM0qHACVIMJ%(ckkW
z(ffg^1w)()E}v<;mKOE{_D1VP4%(4fy-6m^dpM?=J(C|b)<b43Qy!8-7GL8SsB=*C
zn;PnjF`eA>3@&F}NGOiE8+ipQQu15=B!*27!+I3yvlJLPdcB2BCoJc2f%0`@WnmN<
z$yaHe5Q*Fv^2lC0Om-Wuru>mme;Q-6SKV7gi(O%3Rx|h9moMt<!6Sx9DiN80O8+nE
z)}F@7#~05x7GxiQDRVtLF0Hw&+!tV6Z?E5yhT8rS?sYBX=eEkMI(aDLif4i~CEB=}
zB5MW6frY3{t^;}<$~WIwXkQm3bx{cS*%m?C{H(N`1e1a&@Fj4OxRH{oKoisUn?VDf
zFw0I-z;M3K`ZJwS*J^yS@(UmU_Y`}>0YwDFnG0i4eLVLhT-chTt0j{+lK6y$ZKEgP
z53d5HO%s;^3V|sCLq+dUK6DNmoLBPXy$o00zs-{pDW2EY6ltr5wRA<9dEXM)dtz`Z
zAP5=e_iLwU@mt}xA<$gL6m*|h-<aG3$=Menv>i&cL-;fOw+CFb)EMpJb=KsbNqjw<
znIPQ8j7H<gpJy%J#86})=LFpod@8N`T{)D(il0l^&=UFzXbZ98=l&pjbSQkacekL_
z=?$&WonQAJXO2wo(K)18#STWyQa-^H(BOiwf*2W8OiGK-xr(*5t9~!|Sc&sT2|Ryg
zWMqKjwOuW)I5jMyFQAs}GhZ{X>+J!kbBK0WUh)+-;QXt#&swCRJJ+n7tB~|UT1agl
ztUN^H{`^kcdmAtUY&ANiOpEPS!L+?=%8K!8Rj`@Jj!$U4v#an<^KdaTksw8uuY~TA
zKRS*Rj6HliT_xH!7f>p7sJgPal(MABv-7Rs?rAJ(>ZsIdDsnVwiCU5T+yW;v1#vuY
zp1aJ}GgDrOE$CjO{!r(2JMP-OLd|&&j?XQm?0quQp`<Io&K((40gh$V8r|1=e+I|*
zusd_6DK84%-6gh?bYqq-F_JsqRlUa%OL6y=Myf$rxos`e*@EVH(YAw`kLnZ~|62!Y
z)#J7%g%Xk8g_}FYUNVWt9*WKPff8D)YoW79vG_I`41V%>ksT8qwxX5bmHlyCO7>2^
z_{42KI}MbTHaGzMQ<((BeJ&L^v^>+J938BW)iB}_s-N(%|D4fd!uWOJK*)}AFgb>O
zp8rN&FF1ClZ{5t8nKGu^{xhq9u+e1Ir?j`0+&TX4QWegnGluRu#d(?rIaB<H2koao
ze(=1tG=?p*O#nQ4wTY*_!f6^&ny^ye9;4fmM%ElV>_O*UahK^1NA&c!UuQE@=6wyN
zfdM-2pG|r0_zqF2|Mr01G%ZXFt`i4<Z?-91h~+RMV89Tct-S3YLj-78m^?hU!U>{H
z;t70MnO++<0`KoLjc!pMj!Rh8WA)cI3%LI8^j|2_!?TSnUq?F9PaQ7J{&e%Vd8=P7
zNPcVDy^r1}-AKTxPbBDm212WnTC=Y>#{_0yI-=2zH9?8JR`e^&Q|`8wLoda8cHDON
zt+rlVv!pL|ky9>lu@2npHiyJl;lwXkLwK&Ip_XIE*BNT3HGty!@|)UyyPa>J1Sk^B
ziTi|ywm6?lmPW&p43?i+DfFKym5^bUHl$^YKliiG>l~*V(RxrBy=6{KL`fZ<?BI)S
z0$o+H7MQKak!`ihc~1~pNZ`=g?DwHYQz9S$%*>N6<!`i{!w3zJ$S~Jp-Qi=O3Xhz6
zTr%VrYA<uE_|U0)i^=l08Jx$#anAQ#sed-7udWmu@T6(-$YRS~LK~Ji>=hJrub1wY
z-0j^Sa?(aY)~o3&_XVxTu21_8>yPS<ttX5mJcE7JbpMrdzQvnX5~vsvf68}^#l9ve
zVlB`_z{Mher{{zSV3iUvbze>44~-E2HjWwd{*g`?Ap)Kpc+PG0<uG*PwU$phxpvG?
zx|M0Cr{#{;CNgrIt3LAHvhDhndR~em^APUghW~O<1S*cOo63EitlCuYUYz_b7HqsQ
zFz`e?#kktXY=w_&z?GH9@xVX+ln6*YVTb{XV<~$PL(ZIR$%-=7E<<_Koc`6GUb6j`
zeaRLnC2oPGee#rHhrN>^jT+}R-~>V~(rPoyYwpntw;k=>2lKC9Nu<Z)ny~|@f}azn
zSldwdWv-BRTI_^&gAC`@@={alhw=;RGLLu+P9p9C>eAG`sraO32xAxY!gUjY=%=2f
z*anlg1J&9Ub1Hsbi)PQ$R+xoOJjJ|DZnmy>T$}ow_M}?a&!s}!=+j3@g!>bDIRAKP
z)fBw-J2KAO<IeE@K&8Sq)Da(`tm@t-fK@%|^~PE+^_ue&1QL{9CS5I}wZ2KAV|%&9
zYg8xW#}IEf*SB?=54B2NZ9({gcDub82$hH8eH1%vEAQb7i@)1R6Mpq@Let0V9+;q-
zxJn|Bdx_;5kEWc|p*GktXMAYG?tZ?NzR~r}8a=(1^nfoq)>~thTMHL<n{euS&X<tj
zpwj{}+}%2C?YXbZ;w?<lXwN7kNCZtDU3%wzk<%jQOIXv}>y+O!AE+qaTUYqf`<jNu
zu>>o2vD(BgT~c;QfP^V2VceFk!0rb;4b^=x<C%q*7A;g45@FJUl)vTfZKP5GWB_$Q
z*4S>OO;{Us3r08LNOA%;=LKhukox{kz%EX{IrkOiNB-B;Kr!%O8wJL*zaZI!2!Iu&
zNY4DIDuGG!FpLR|X0v`I3);zlWmG_hKM?s$;_rBfkV`pEwn^$)L8D8;60ZqqWXc%1
zHL?f}mI?_58D<c;4m^_0Kd-u43^zlywn|^Xjt-Qj3*$vF>0|%dtK$D{z*Zsya3vBe
zL;ya%zWwY@SKCx!>o!!9FdQv>Eq4*|zO=PENZn+12?>WSh0;|#M5u`z>Fx$7iC###
z6v#K5Wj1{B@zNQS9_ECguX28$2>5yLoCr{}oIx0xG>5OGRM<M%wkF~8<PZ0TUC&=k
z?F^NQ9#7-PScQBJwlKG#O*%H!DUmIqAr*w)c6dWP_|}pWXL237vfqEKrX;jPWnC%D
z|2^25^3(G000&d<i%$&M;AtT2P#I?3i2i_;c{ml*y(80meAAQBMf%gCNb{QKLNAqv
zsW7z`Q2!n!dAm*-RExBa-G#NxzE)PuIj)49Tq%HMrD)(PY7l?KE!S9GGVfWQ1d^=i
zE4CLAT0FBcu0kAbVNJ39#~0>&T-z1!7!L*hV~fHE>%bz7@$CW)3{Uo_Pv?>KN-u93
zaX(;c_$>bEF<I)bf_lhKX(0}R&xaEM3?Qi=N5ckBhvEIPQtzAj$m{cHMmZi2F@H}I
zdwW;5Z$)D6@Kb!h=t|3-tETeUc_6$x>Puwfyg{PF_qRoZyDM|6&XvyJi!<`S6wjC3
zdNMzUpKjXT2FabPgMb*vL?Xa3<zhCYxf3=Cx3LiS(#_AmQ_CG2W4?vu@+MnFJmf&g
z@x9`Zr65!i0p#-NsJ%!*82Tg%r0a)~Hhn5hU=9^WJg)iHLEpW1-Baaj3X%+ST*5l6
z8+X@l!=QUDP}!~giGI{4q=TG~tky!7WDtBJWAkV-_F0CKiH@<e0DPi;q`#*friD}c
zdNCV30xPywiZj>y0-hm8{m5ZLR~n4lB8HN}F)epw>>K&&U4QG&b`!MskXiTyc~3&R
zhINu!knH@yR(43oIV<gSo=w2x8mB(Rc<He_R@}Jisrg&`gZI2!)6I`s;W&nb_T_BI
zJ^<TK1lYGO5ng8;`z~$1AOd_RHfk&nBnh+MUDmGU<{poqz|$;d{kL6pct!~B?X3-G
z1vyJ!lw&~JD@#~Glimej=bFKaNUJeH!D(4GP4Cwn2cPb7{@soHKaDQzAlKt`EH^p^
z86AbJ$A(mNV{CL|zO{|JD@smmHZ;7(egEP<`h9`*9qaGc0O(kS$d0(Bv&P-Cd2v%S
zZ{zYEcg~h-C2MOlC5=%oUzmA7orzpq&5US%<$FKe#flT-v#k76%w(n*U!|l}ng~lr
zLRnpDT`Sv?ZpZIi^z7~n|MEkmcTpcHz7DK{Vug58p*VhQ(g&O#I{4l~V5vq6+WONQ
z^tfd0-Kc0ev1tB~j*PSGvk0%m<6)4+pG;Z>*N6Zo<z#Cn3}xQ}?L#y)0>|Keat6*W
zp1-|>FbU!@ET(>OUC7AuT{*y=_CTDcMz4pZHfZ%Jp{evqhWrzWPpe?x^8H&~Vv7M@
z$vOK@^ZCC<m`im^-oGv0h0ShEoJxh{FQJ={!L{gs?M7R4KI%%pbi|#!GA%Z_{1xf*
zhbcDs8hmJI#>E7MAS?*%@4I|Jyx4^gTbvwY9;Nb1&CV|-W|u5lk^~34(+?!Bs0X@9
zj~4@f42q!cwy2lTY156TG?ZGq6u}z#D@4x6GKw=MbDVoWHFM&vZmHh&e>ID%+FM*7
zg499#mE&w+8hw{NdyJc;`lH*Z5ay?x-kZm~BCSu>Qx=Q-=7yNV4Q(~_f1)6^45zXw
z2q>YilL#nbgCE0Lw!AL0J!FgSwl$J9j*{gs==vxnGBCIFwz_g6WV=e)UeLqYs@w+c
zGE`3lOm)TIg2v11#)I!b2KCx#AwTQ<h=9VqtsR$J&=ARC2&^3GyzhNZ<C1^US|ViQ
z;2A$X<9;OZ47RGg*07w0zjX0^vL7Pr4A@%*xa{^G-LqB43q(Gfj2)VkXV|ybNzf@N
zWWAr&k$~@j%>-R%5y$XtJm8?<pLxbc6x?C6uYEDby62oyVt8g7zcK69+b6SnePhQa
zXFVT9mXNFOGwP>L8nirYb+YOV;aE;^y*8*pS%v@rF}Gle03v{_w+Wq}Rd?>TDPxlJ
zDkHp7m$eTanNNY{xEILvCD`Nx=t7W($NaXwioTkM6l23m^m>6TCq(Th3c9cSPIhQ-
zRHK4&Y}Y&x<l_@0`AmHt9)IFxX%_vxG}bJcIX`!Eu!^Z{1t~O_?c?3%1k0EzJZy4W
zFOBjo;xP*uYiezh2{FgxMn~I!8gwGEdQ&1HFwba^VZm(EOsv$_)_qNNuI-t7U0qL~
zzwuY#iGeFDQ|`FBo(lxh03#tBxcgA2d|ota$g|XPc2=r~vx-99371Ls0yn2}UQGV$
zTG$!6nxYBYv+u5U#tPKKs<4kzRHGSPuTlh1cT9>*v}IdGtEc+uBqEJU_5=k(AQH{m
zp-Wz;r{|o3{Z`nBb3fd33@{eP1*NlbF=uVTILmaq9Uu41h2}n1kc(CC@xSfT&<(w|
z@F$)(K`4+2=(bk)A>|CdicK+ew@|}KzjtwrrU27<quM&+H5MuEs%E^>*l=Lk!bDv&
z3a+h;v@t2y$&)iMa9JD>pjN-IO5E?16ge+>XL8Rrhy@Mh5d7;G8VIQvBH+T3aIxVc
zjJI=x6kO2-Pg*A0n4QB)FqE<IYFS<zZ>so9^AR0-5zXr_9Z1xD6h(*tAyZ5iK2Dqn
zP<V2#6MxV?eOrpnoL!N(RL9pr^N#D#Vp@9c>d@YhmOHDffU7HbP+KX;s=k#qCs<4Y
z$N}lpR{o`j)wqzdQ(1(B%(mgecr+%cbo$Ho9n!SL-cgIakqHFCd|?5XV7Oy!<k_UP
z>rN=`T|z%oJ{2yg6T~O~AObeFT$Yuc4@z*-_(sWgAjl3qJQLT(GEgM&Da+d}X*DmZ
z<JxIzZut#dA3{aB#Tc-II#i861Jxz8*KHutyD^f^6P4O<B{L~w%i%stXquMm`Habg
z1+#Gqe64vw;6$1TC<Z?#%q?+C?j^O?%{M{1Qph$n(I=X%4ee8)r)Q4Gp|4L{%07mf
z3Z*6tmV8!vMoD%prA;b@`nNI{0nip987G~-qD{i9IYS4?yrQpmqHh%Qm33okr2sY>
zt0c1#{4y2T38CFs<ZEm`Ku+>cIX=!Airg<r0gYAqrSq-*5@R26F>8?sT;Bgq#uwvT
zD%a;zIsN3$!P_*;fo}Dx{5bV9IQtV;R_7<j#_*QC(o<JR2X>D@jX?rgKx)_po{y-<
z32Nw&$LU?u!aZ*qv=z7d?ezJ0ulY0P9FNWgAlDNmfw4ruWw1P!>rp`uA|`w_&`+Cp
zN=sKNJv}X+M37|7;PvAJ(&tewNA0tM%lz$iGhHI?wEE53T0MxU?>V0uLOE{NS!*Tz
z6ij1Pt<jNy<<hOR7B(NWiR#8VzC-eHjK%v-`suypbOAeS4|-b}&Dhhhuzh|j;9GK1
zTvgbnDVB`^UT7N2#b>|-98fw1ozmULKLZwYVZ`STJ;{2)|MTfQ+xtvPM?un8{B&>b
zsmtJ4Xn<@g=ahk=pb#HdKNM;C&$ds;UXgJxdG+;FG5g_ceBp+v(``qEEiGb94yPVs
zw6Q~I#rO<0y#w7G@A|RzUzW92jn+-|Be4_tE&Ht#78ZW4ZF{@8CkRhmL^Wh94RS-z
zHD|aq#@FhhaeMke?dOKH%^>l-+QLkGEWqW66)jZbI^g84BTOS6{1f*lw2rH8j4~TR
zVF~vLyTt>qbT+qaRun^WaXrfKn^yd#*qQgRf#z{1-2&Li2<U^mSV6<1LASAU7>37a
z-560g79Ohg{p#9rpDdEcQk%~qGBPEDUiJ`LS_U1js}USHh<5?b9>nm=IhWYXN9RZ}
zynxnxncfenOO_Pw(*d2e9^zzXeg7odUekE>Uv&8VzZpn~vw;79mY2?sK5ou09DKy=
zJv{EXymD~-Ki|jQyLY7}C9nTE%1BH7?RfXzzx;nuq-3NnE{aRYNQg^HN!=Bfxqeq%
zO5(0K;QHPF^Qipy#m~pr*82h#Cl?0~9|w>BwBALvmoNVx+W8k^uK#uXzoZa<5|;s2
zo~Uc814u{!0FsM801*kWQwwyypjKNOAbe4Y3P4W6^k=W(zuYI`1d#vj8~}Ji!u8j=
zKFO_rt8<C@-_HNFko+SM=Kv1@q?a!J`CJsTiw`*!IXM{_ISnNx1r;3)9UUzVE$wA`
zrYo1}8R=<huUx&t$jrjZ%1Xz;#(tHBor#5&<xe9dq!-^IBc~=Or)IfKdzs~bxez-5
zj8uTZOLC+nynsuLB&3Wa#2x_f0!MO^zm>o6zjT2k1tk?V4edn*_!Yn<5>nDjWTbzv
zzPKBDaUVd&NX~RaT!n&J&z6$ci{)Ne>PIR*)!I&0{V_DZgq?RdHO*Bvb`DN~n}WB5
zge9e<Wn|^#A3jo3*LbY?#K6$V*u>P#-2R2bOGhVX7aw0g|A4@t;E2em=$P2J__Xx5
z8JSt{vU3WHic3C~mX%k0se{4m05w3$zZ<?bHg$FP^!D|C8yFm)n4FrP`8kU~uAqLc
zuB~tU-o)%593CB?U{BBf&_x0uCHZ^#59wmOpz9JD87Ud%AG$~`1^lVNNJf4`oPtS3
zkJ8qQnfG286^m->$J$P6J_&s^tDW~4%~gI$qyXj*X@8^ae~z&5{}N??5%zC%Apmrw
zBo~86$_P*f;0^x!eXObv$kJac`~W-#q`jgHM(>y$=Vb0#PTt0GEDtSlI#1sVVs@#Y
z@#-Yi#%Xap4+!i3E{nUX!;pf5l!1Jvc5!@2uEyIf9Bymeu040JadqR2r-$eL9XB)@
zmbN6_xwO%r5-18vJ6WUOJM&B?1sduN@E7!?1@CJ+2+chTc-AdnnPbk!ql*yhd)7Cu
ziB5R-=$TnIJoZBT@b6UXzHWzAL;FSTO(zO}l#H!S6lw-YjgNe})T=|K_akg8cP0$<
z5F6u%2`7Zos^)9iP(C;p^J5>Go@Dn+mcWl1Ed72RqCBfKcv%KTneZopo#@KgP6nUK
zYl;J%d+xL8=83jEPHv$NfA)E^|G&G`-nf2Avx9pBzz@)rC2a*EQA7Z66L{??q@4<@
zzN;lhVIdb|j-jKZK>vA1To=Hdea5(RG77w~3ZG%uD;KaDT{7Ngf%j&=%g@9<mz9aZ
zR#(6Ox;^?M#j}0d2_gL+seSdbvTnG@W$E$z<(%lUqBKqAIR2wwIfFj+v3uRNt)=bv
z+wYFS+-16QZgYYQ$c#DNi$nCKka`;YX%HsOtpTZnlOE;R02_9(^&c>yUvRW9!<Q4i
z#2t5)7*mw;{Ly{FJ2GqxhJFi0jyt;=rI^`9A9ifLDV|qdc6aNV<9U8*`vjnEaPR^H
z5ew5y+Q2Lq^#}k>zo)?@;Z6b)ccHD4XS}`W)wsU)_EpTC$s`H@>l}Efecv4CKm<I0
z?}?|IxnTv_ZYBZ*pnW#S)EDOck6|O_*lvyzvFsxU7CkG$=>M7*sv$*tNqs%y?qoH6
z&1Zb^?-jrE2MMwViY>cy`7-t#CB86b|M_+`7K4jva^pD_*=t>utp~Jfjz0q<w5B}w
zwyWSGT0?UbhRyX2dDdDlZ8Ml|$*F2wzxEO)bTg-kb@!qd@u+_Y+VDRI`b+@+T$07t
zm;T13Gb2387{W))s7W(ACDP;97qmm&J3r`^j#r)x7z+y-ui1i+JWHSp+cV{eAvw5p
zDH6mNAdnRyn%i9+7)=@x2#H6`HYb0_Qyrh<UnUhfx3P>ppE$(wex7jWL!?4!;^p|$
zv1z*2npLF=PRTfFE=0(|+`(lo^$CzrVQ1-Yw3_WFga(|?nw7;{QxL<S-;!>l{6FMC
z-B`$#-L|_nqaU3N23(gnfDXGw90)`NeQh74aWe#emv+1A!-P9893J<)A<n<p=`Ntk
z|LDH%%0?S-lmIX@Q~88Yb9H6v^kmw3{Np)^!I+w6Tjzf_4ZK-r2>&!yrr(~(dbM|*
z78vWPv5p7+^9HzkX8Kh5|B`Yd;7`g4U0Z85pL1QS&8Dbz+>J%DNkxkyi^1ns)*@eb
zR`O={f^7TDR8Rq6=i=2uRd><RfWrUA-g`zh^?mK4TR>51A}CFW3W$J$N|PD|=~4vg
zB`TdrlVXHGklqncP^wZzgwR5V&_NN9E}<qMy(9=BkdPewoqNwX|Nnd6`<^lGxMQ3T
z=R-bZWM%I;*IIi$&olR2b2V{Mc|qOiT`d_&3SS!y=sZ>{ENlyqLm}QYD$6_#z_}h}
z(tvld>B@ut{|l|1iQZJxOG8@$V;APp_@7i68jyRx7%@M~M!l9uaz^mdfT7EyXuJm_
z4dBqM%GoJjO7x!njHJV}Z9blrAeB1IE3370toUd3rbuAIUeaAnlCFDF(yc0Rz;ns0
zWmLwE=Whw7|4_%EPel;!R*y%=J*V(v5lq&kk*=z#z#3g{_vCRtD0~54b@$2(e`fE^
zq?}tMrv+U!H+3Std7$Lu49U*83jZF}G=db;R;uMXJ88l7+ed6ojoA@(3S9)haa0P|
z3&551`eZ3~@v7Gn9Si(%cbPiElQ{(8Utj}g_#2bV51nFsa$*oD&DfL<t|40ur3(_7
z4dGR%VzisEe*_b_S?1j@*@JZO!j|rj*KBu;inoZr;~z8;QQLF+ky%F8dzT!<g|C;m
zpm5uBgG2mrh-!~0xZ%WM<ZBh4H312H^IIg-Iv8V!3H<n}BLkLs@hz*ThE$v{t{VP`
z{BWhgpiLJMtzZ(@P?ja!X*z7tCz2%1Jp9}j@vgkryC<lJf%kws@AnR?2#TNq&j_Gy
z$%`aLt0DVr7PWhN{^CNlzJB0_@P!-q89-_B&(w8seIT|KM8()r(Z>!Pk95E`d)sGv
zoN08K1Oom#A9CYF1<~H#J~KPa$o$Kftb_5PmqVDHnAM@eUH{dVB_?*0w!zZE&BFL=
zyX!t1S-Z+#JTS%<$4frZy%aYX$(t6@#P#{dn~pMF1n01-5V#Q<>&y3vQBdA4Aq!S!
z+}$EZ2m{qzdwn4>eB@CC|9SJxX9^7(DN{|MKk9#<)-c}LSgaZ9Hy~@mA^5~jM93W&
zq8-zp7Ues~p^U1@DYVe3wD!1@=db{h4p7b9qGyar>9ov>45k6Dh{#RknjZ$vgyp%R
zKC_s6ef9}ysy>S6?!uZf)JaZkC{cOTG1(sq`&PYCT97(UYAgwNsW*1fU!Q(CVh|+0
zwT47Cmbm<C#UnadUbg3eg6^zRdGI#Naj1Z1)p3t&;&or=Bc!83?9=r*lRKC~^w|mm
z65G&mx8Y`0<5%m3%HZ-y=UT+G(cO$<hIoeey<+pgcTZ1G1%Y28Q6%O(DpzffM(ZH)
z-J8Lq3V+CJHIP7yf*m#FBIxkSmvVj2c<%3JM&XAp?p9W1gI05k&oe*C3GMn`EB%wP
z*E}(^pS^c1^UMAJz3|w7@d*n)BLckq2)sEVG|#&CC-Q%V87Uzaj?@;ikuR~J(=*>3
z=KGv(t=nD+o}UHbYHHjUs=%tsEi}L)&l%$pE^RJ-q;mSJ^pqAcAKIRqpe;ldN|{K8
ziVk1|(JT?fYzmRmp)piFch<Sw<q*zp?oyMqsWlO<ekoa-qAh|Hni!lrN4+uwe}c|Y
z{aKVe(e_iNZ$?83g*wG=7e@oUNB0k(aNt0lipV&wKP#4;JC-B`#>s%s-F`J6JHC45
zZ24!?fQNKs@$1Jnl>Wn^Y>IiA0df9j17lKyWaJ+j5PZ?p@c6!#n9g#m&PF3?Zs76m
zBO{OUV`{@Jjyy_3ETTw#)Hd0<L!L6#1{yH=lBx@dLvL`n(ST3s!`aKH<VX;yweQz+
z8w{BL@2v)+$=}~R*wI?Q<jf>h-)g9k9B@|n-R0=Tr185yFFJV#fAke1Uif1cr}Q+C
ztnQk!;(zInIYqYmZO>;{uk*=J)}r^%ls_?U%Mnc4Sv^vPYI4gux&>~v9Oa_4ca>RR
zeGTHX)Ug|txf`u56KVY8w!qKO{R~n`(}>TM4alIjqM|lod|-ybOYQvT<3~SFXk-*(
zRt^K3<KvtBYW!9NW4O~?a}$EkR@u&o=_FT9&u32m-F{O4{-3QUi4;o`${QLe64oSq
z^&9fo<$}4umEDy=@AWGi%?;@)H{q7`sFsv=+w)2G6Ele6?Ur4CAE)}gH$>YRJ<n_q
z{-`$c20ThmTd*P-8m+&2?aiK_^0%)U-tdZF)%CviG~g5k?1ml0&|@>txhHG78EKaZ
zE9R%GIZqz9L)Ne3(2+SiUT9_-uq8qR^4%b=W{FYSCN2m|{_3pxjI-Bs>-+!5KCT$j
zGV$vXN+O7F?P99HnZ{~|knIh*^zhDS=1UFF-D+NYf4KC*%K2_@8C07zLN;DEE%9=j
zUw<1PEV!w_Ui%zGw8tQH{;b>CX}}ZC2gR%`6i3I)zo@HSNo-g_4kQXte}Va_7c-0(
zm`*oId3awB;p;cnxMf)#nR~Dqbv%LS3Q4uZ9@4_Rjg{6gs_yw$l-tu77mVkKsZ~a$
z!hzQni8(`Y&?Ywu;RL2I^sSduwF7%rRM6N`!7kHleOAL?QL$IsS_TpBDLqL)H$636
zS;X72K|jlI4I@4`r3~t0<~EZ0p-z|d-?-2GK4o}XZ+i2RQ>IzBAA;+6Z7W&#QVL@G
zU`Wi+V3#}E=>!eeq1maw=J=KPaW<0o2q)hvu$F$&xm+Ilc&Pt#<C{!|=*9MW5#XeV
z(Cg9BukBe|#X2+q%hA=ghLx-%&smeav0x5YQN&jmL}Lv32B<Z|z>Rf2`gGh$77srv
zg42Jc+#~59ibU)Z0^pLR!dAYicZ1`FPL|uegV7fTz1v|eA=^=23^lp9V3}gAT$D^+
za^<kbLJI<OH}eMPcJij%U`lTo8HXET)SJ|?7gJ|?LKjbKcw$sb5OYGy_%0_U2T=(#
z9q41KKyl$i!s9iKM~i7q>Ha>hJ-;xJ*Gh7Eu76y|1q6n_`nvpChFDpsU)pghTfLxm
ziJYF+DZu(u$%^Q)+0Ujo9`s{sp8B&o0U_FK22(VkPlDu>9^2*}*>ejfyEMQ8?lNPq
zt~u2Bl+R}s{gX~tqtb2m)|mtER*{MVe9vYyVWOLQdCUjGYTkuo(bRcAHP=(%y!88~
z@FmVvt(~aWkXu^v41O|26RVin@1`{1G__WLKq+wxtomFDT(D|3F;QdbmX-CJEIu=B
z)fz3_alQQmH*-CGu)de>Q|ytXF1n?qa>VVXq{pP~*3#(<LiG;Z#)2o(E_<FC5XZ&N
z#eq?#InPJ><_|sq-_AL=KoN7-wU5LHpCjBMv4`QvXo>R+I^jxk-IZy6+8)=kCItMo
zO!B@lu(aBTFiMX<HPAImqtb!P?gvH(_N=O3IM|;@V4%bVQcAwnS_>b3x1li2vvIyf
z?@aQUv@s#LIGd-Q?t%eElbB3;m2d6Zav49}5cI)4LVFurp04%u(OA~V70vqAJhBex
z&lZAo24a)Aqr0VYO!v^r6foHsjvzq7=Ly#mU5Z>47zjKIqotcyozF6FDelkPF<jcc
z92#e)NtOzlrd}Z|_vp!CYh&hX&`sk;9#;ddr-CY^O}9N?CgNCQUnhjI`R{q2ZE-Mt
zTt0ej+D;Vh_f6x>cO#EnZTL*`SpHo*(eFC4u17=rH;)z2%l}%R{2x8?pX!lgHgM~o
zgJ97_8yb*2u^>rim_n>L)}LLc0S<EqH-qlbfO5TBY6Mg#Ad`Cz3m+^wM!8j`fBsdd
zD|Vmbo8VWkk2eOEPuyG6-ngFnI7B2_(|qBcMb-e_>(P{7X8pFe2;uvi7rq#LTJi3;
z!()bx1wIvb>S%ve$bK3{3<EXIb(x)(Z!u5C#5A&T!`jKjm@$!`yYfpFBsUk)M8i>>
zu+TWa!L7~!Rkm3VT-hJJ5jUSemD%p@Sw$;FANDS0UDPAZdnb9_GoX2^M53%L0W&%E
zG2!|%n=dg6fLMQkPqCOzpLdA2_pzEXUHV5-*@6jB+*p6<ZuWWn+t_O@>(sQs{k)Mp
z<-xC!mBck?5{lruG=6As2BxE)0`a8*x-I0P)uHU=aka?+Uzv6oqJublsbc5yj8H=N
z{{>=WzxJ<Hy7lk$JRAN!78&^$>H0udS%$SX?9*|^4cht2Hi<8w`Z3}Ul1(4T1zLbR
z1`qDeMIiP$Xuz;65+(lsfPT@@RB>v9K)B@V<GF0z{pp_*gA^&oq@q9qpD7v-3E5-Y
zI1(5U#j$bb0dFJ;R;@G4VfIo!mMJ($XTJe1qW;x9gl<-*FvES;v>9j4#?FSTsde8G
z2#=gWrlXx8lQf_$2gMcFa%GC5(Olm&Oku-`$8+eApe*Z{DNCzPGl$LZ8YauH60Cj*
z6?ZTvh1=K(m<$SkVwu?~?Ia<KVNTeb&cSe7Fey+jQt2hz>yw+MbD=kVW9{Ipp06#}
zCUyDunW`6`RKY*`CwdRlfcZiM;0O`1A;=cYH(PL4V>F|775f6-dwNAo@fCXix~&A#
z*^r<CPbDW0_w{A(z=iPN+s)ipzEbMF;t*erKq?3Q@MW)gUA9RlVjf<z0|L12XA9vy
zVxD^-C6!o;XO2)!3gd+Bkfk$1!nrXT*KU=)x{_c%f3uc%#708DJn_Z$CL`I76hKTP
zS>V8tkTXiRKf}gB?3S`QzPM)BZau~O!?{5$ohjC0DOUq`4L@!fYG|E>1%kU$K~-tp
z$WtV;T75Zbq`(3uG17*~DhvH=H3RiKbqccE2$Ne<Phor^!Ky7BTvkMUOgb|^Hzn7O
zBtRBe3&4qR?Q5UB*(g`f+v>7rVN))}#kvt498?^~n7a*{c=XR4WURz{6HgyyQ|=eE
z<apzAI?CJoYJXG&V}7V4cbc?OTwmpJ;@*9)a+6A{q5&Zp)PWb6-u+p{j0cV3Y6eAX
z{x^5c&ZSU<9|%c#4Keu&pm31tiA}bGph@bfNbGj-t|A$HdS(tYvj#p-(e*g`@U*at
zIahDoO?Eul@ATs*5*G#?g<U#6uYs$;T{Iv%hZWBA#hPM>Uyh*gfMa~7Sd1X38Iu2j
z{^-b$86AaiFVGxQEq#gHOY&`MFPb?>ks^nm>q10fQYcOYr>GmlJGgD*rk52Xi=ND>
zZ>$6v?sLGy%{l7Nv^zP=+i?~&clk|eg0M6o-rs3so&1oZM(A(Ogm6%J7V5s*E|}nb
ze<&mrefs9cp4gW>V-;|(gQZ@d^ZI;J43Ts^nTkMFcIK$g9QKxh6+dmLZ2adlL!|F-
zckXD3FvP`$4qj-+AxR?38)V!RTa_NZM{{^ae`TxMB;X9r{fEFs<2SFC1u-o`1kElm
zC*?XK3dG_g$=_0oZD#B82Ssd9MdrR(Pp@2*`@Xd5ki%yWeG#HO`TfJ`1O1=Knc{1+
zIZ;r~q}3c=75P;JDzOt35nZ^N0%L5=W+yPH2&VM}Ny2tNY(La#wfxm1$~x3*@v^@4
zJ9FaOQ;9^2qt9?T@@3MAARgGZPCA6mMegN!3%D9SAB;(N%;T(53b9H#wH)(=&&7!Q
zO=vmKA$=<dZ;%+A&YHu&IT(J0KrDd8-k@G4sQ0;g&n7o%=_Y3;yoh1c`9x47o*)Sv
zsxL=LL)r0>)+5gX-oq`h8QmlO5o00&Wmxw6aW(huGBVfqYf8I8SqJv{18KlXxudLH
zRQ&EX-!v?Q=x~iiQD?}bp#D8!C;ZlD<3SB^$R)TZDvel8=!sLg`rSt|e7G#pC4a@1
zo8VaB;TDMG3m9#X{H$utJy!EdCqAY9Px!YH5<f1>j@V6z3{erFs1uY2!<FP;V%OtM
zvAd4E@L9QZ@kzJQgZB&?c8NcnTbYAi0DL-u=sap0qG~p>=w^Ivy9cZGX>OKR?(PMv
ziF?VZh#nc+4G^MLCVAu4L?~ju5X|JZ+b8Z3MO2t*4}?DJ9#1d(Ak&pmzHlV4#Nd+*
z*1U3`{!axCK@!br*Q{&^#ZoM{yKLOpD~g<X^*X&o=@kpJoreVD5FGKpF3nH83mOO0
zH*t|K!fZj&=rb~B;C!Xuimj}4=Ml9^7>{TpizMM-E3JN1`)#*x!C#7kwXV{DE^yS~
z!V%x&00sp_$0qL?_p{e`5(<Yz-8K21o%DaXKCv~X;8k2TQE?|FdS|W!5|=^>=&;-l
z_hn(TX&w?B+iKFq@hi4j+RaQl<8XSb43$^t#9{5r^h#UnZTxZ@`#64($?@Z!3IkgX
zOFU~Bk~tk~q4!+HRC6d>rga{vbnc$ur3CTKL6Qht2N`mj`puSORDi+Fs_U5z;&I(|
zdh0&VjmNd+Q5qlp+5Aa%^E9BNQ8J7M;2*~#&bm04ubK)7HbcdI$mREMR*&TCZsZ!~
ztwKKI8Kw);M*@_D5#L3wiOG7V?I)4hAa`MQ3(Z?uWJ8K0o(9A=8QT?*_&Q8^AC3ze
zK?BU)Y+*089(IZ<W>i)5JU1KKBY-1<*hsp#GzuqZL1rS>M;CmqE#JjocglS2(GB9P
zeQK{#cl2x1C!DYAF1afoAm3xZb1Uo;wt1U>4K)wzYm+T?2+dz-ragw_S_a{2ULjty
z1j-N*6alyRM4He}e2gbn6|H0J!@k(ArfGXKv5zvU-tQT&ltIJ>{eAZkmG@J|fQJ;I
zH{l>wTrlfeuI{sV^z*6{v-ecY0aN=;P^&>48dZ=JasxZG+bq&yn!cIv4$5uiED`=T
z!9oxu7BW50?8Dtk*R((H6fBc-jzT}uHL+3bLbzzaj%nVUi}Rp7DrDMJ)sDTLsV(6c
z7C+4#lJEV;J|{Qw^iwXCtH%t&n_On~{jD$=<677<?3GxTl^*ME02Tkxu&JVfYTO!6
zeK`WDM)Uijr_nqfsDLZ;a~nAVE-EMgz|~#&hsI1fj!I9=v~+`{5v3{djVc~*aSa|^
z6jeO3W8GOo$rxt?>M;FNp;a)pjjOict1}h^49byVGDmzpL3~Hq2*-FnoDd6Rl+SiF
z#^%d*bIfM?q<u#3j|xTcr>AY}+v6q*;Mefr$QDKFSPs808stp070Kz~@=Iq;DqKHz
z^T$Qm7ON_S2lHk2PaFND`kw4S6)5r%pCWvr6GjD^7~99AR#{)VGhaE3$q8*=J1~vX
zY4n5FxPKpPTt2$GSJMD`XbV^UiBgucid7|CO#AFQZt2o;%fhBhsX~L08e@7-`}D8x
z$+u>zY~{%k)Zqq4G+CL%Xhq6c*gQMZl|F6UC|mpr`t;_IRk58>5=$`ti;BtSx~(1F
z7&{P(M-uao@}=3B+f)7Tz=FOn@o^^@{p{oDC48h_nEs)f06by5@Qy4H^dpCvVnIHS
zaXnlPlWw-N8sCuoT{kA_mTL7mrcc^svA4C_%7lyW^m2A#qnLfY3eyP-)6D5h_<=iy
ziOOGwi)B&%H#4*w>yVE}*YvAJOZgX@YC_QqwK>JxbEub_R~R=Auwb?s%&apSgXu;>
z9WL<}^y<}ooj33-uGhWcs$sQN>?`t<EE|qo&AvW-)7OUYQT5nvW5Jw9z*B`mhQ3S(
zer;18#y8ME0cH_wT!=Omgwv6x1*&#*8%t1+$WD_p{F;2Tt|=RQ{WimfD6gn(*!4l3
zBjMrQ41P9)kPf}7+L4VXHg7PP&I)zjL$S{QTLH{k?Jf3B<^{2HdgJky__qzZli$=`
z8g};Vhf2l9IBWFT{r<Q^ejmI_UCE6h(Kr1|?Tp{Z7znSM%1!8g(3zibO_yAl{;vc7
z@wy2Y&a~_r;-pgXqErIg$o7$B8KRfVkOm-Mfp<9|0&sSc3s%n?>-akGGc3TMsVQCL
zj)C;qfkmx(`y=4%vczLw>}RseRK*%ESF{p4w)EP>!x5S0^8?+Jbn476hn{L4?R`_$
zR!V*W?WSI#s5Ow<5GvuH41|%P9|O}G;+&wa%nCg`)aBys-B#tII4TE8wy*7Qvr~X)
z<Vv>qnZ<+$HHP2YWiN*&e4NODolu2|tDglx4r;-QN&QdKQ9~O#ca{uOXEnP1!|OX`
z&c_)WoR8G$|HsCK{~(9n{JUhTE~;;fr*i!`tc<x{meO(7zA~R(DOsV1m+r@wCnsxu
zc6*Mk0omj`o;uia$9X0xzS>nM^Ab-UL*QwlAn6m!)c;y0*)07h;XJ~O6!9hg(MYMe
zrETHLSd{KT@HZZ%Qza1mz$_CD$br%T3U~i8N|`n5gst8@k494H2eQojw*pu$EX@6T
zqenD=d2=3pKtvx)gSa`0iASGtCHa_Y@p!AD_x7_p)M5J+soD@4u#EYe-Fr$3L3oVl
zP=rZR1ChswVkbufE@<A=@7tyU!Iyq+9|bO#PyAQ0MFWzP$4MG6h@t^hW=9%;EBV&K
ziRtpdI@;gk-zuCvvcNCW83dhl5LW!$EMaZ!=Ho&7Pc&HdN^Z5bu)Xffv2e~u56@5q
z5Z}5fQZ%40hnj>6EQVa5DA9nFc#;zh@B$wRx{{5Wlne-vy`Q_H@YBareo5vm-UEZ~
z%D!Gc+@y#FpS9Ms3&g!Ds!1kf=H+{RbToOwI^3Gn$iux*rr9(xI9L6(n6VQQR;jgq
z$>mkhj}1@G(_BL7!|6upv1{K_s;!jB9T(9YN?f0;Toeo<DOa4HKCd(3;#fSa5?I_5
znn;D~lDLo*XTpsjk%&1p{C2!sp7}_Z0PBSN4PVatKc%DJ%m61x-vB_cI20?=t~)pN
z3j*8kX(%gxihsIQfWj|_R<9S|JXOl6Kh!+4P&gs><LiYJ+--|?l4}A&$E?AHav{B~
z0B`&5Gew4=bQzTv)g<@Dx}d9ZkH;+O-Ncm_`S!v090s5@1+Fg_wNb~D9axT+r&vj~
zUL(<yHD@sy;>d(-aVVYayHM*Lryhg7KGu!-?4ay-6Ywh}Nt{7&g{xdTW2Dk8Un@eR
z(ZpV$&eGE1_N685%w-<n&8>Q#Zm1R3tUDgs&ik@6@hsfBJ`7fiP5-c`Y&3qe!qwmV
zfkViSOe<d&XGmz?iJvtLh*Mw(aw2A~g9zcEDq2HkSJE-1`;k-u%xs@qk@=)tKSQm6
zmAIZ%*~GWYz%Rk9lKkfI>&5N+_L45L3lfnm3!urtPV_k?i1z}7MX!j&(KS*df!d~9
z-@j-yT`u4GonA0=@B8n@6cZ1GQzC}Tx`uqQics5S*5EFSYOZN<vs9<9f4g+${Uspq
z)gIE1kVD}>U}xSXjj5dLL*OHOOiLR2%!iOg$voUI1nWaI9a69Q$I%5W!ixrvS-qKF
zZqT5u6u~K>x~r0q_FAH7NM`2RDNN<s?aChRxIba<vXyT2@@4Qa{TO~MF8AYWkDO7u
z>1Z<^iS3D|0x-uwx%QR|c*l6B8zIoXZ`<wy0mW>5OCP>D$&2vZ<+)2Qvn+?MoQv6N
zQhSL@j7-l*Fw5{AFxS2yaddJ_S<c9>*@s2O9dTYCl-U3=gP-z2Vh+`d;%I>L>Nea1
zccguC80J#8R5M(C!S}!o7a?LZ@WpeWb1`9&3kqRJJmD8<Rd5@OK{&;hpd(f8RvM@%
zONdVv4x_R@=2T60h?nzGS3EU_z2kPIe2eA2kNl=K#TurMA7+b}1oeaY;ZILj6A)c<
z=Sh(|O1gz2Q33<ASnqvd_6JPJ8clZ?Qhlg}A8JcSa-^O_;pIA}*iJnx<!o^KF(NH@
zvi_Rdm1~-c_oq8+zst`*OD!N<s(d2Ltf<K0iHXESK3KwzB<pIBK$;Tth1L)^-zQ8$
z!t}z6&panzm+iNK@cIyjb<^AH5I%Qg%S{q-lOe6PA$*r$VOhV}<-+hVCJA{zDJ3rW
z14j1q%FkNiDMwG@I~9q@*DJ>z+66u>>JfoHheDii9zH+hjM0p~RyXqOoBR&$1Bl;$
zOPR4iWt;erl(N9DI=g_=jFkb!Tr5}UZP1KrQujK_s=E{+OAw>)<=1BgF30SF;vkhm
z=qky`J+^3Y)$cB1jPzqSHFNxhu8j!=N2F_Tb-T=)M)kXfT^8EAfK&S1`Qs5%Ht5Vg
zO>!xsh?zSQin_S8szsW4bWg)8gu6ydLX+q~@1Oih35~m;{dAd2&i85mSnMOq=}!bF
ztVRhzyuASuNio4n(oI>od}O(%8!)-CG(vQ1t4TWho`AZ0$+Mo3u5JcQgdcsX4v?uD
zYPkz5oYS2Oi&%ijJ&c|-G7BpQJeBY6B0nnKrU4A&B&8b{mwSkoiCv(J`t%T3WT)l3
zZ>cT{HNC^?S|dW6hG*!3Dm$4aG&l9sY*5EVFN!wF0`bKfE>W2+LjVVFZ!I=PjE_xU
zX{>NzQn+*R<eQYb7y4ZEAQz4N=H1QNW4vHm%rt8lOjVJl+#I6{kP_NtG8$<BKip_c
z1T|7^qXroh9xY5?TZ2ek?@+mL!E=KS`9!Ke0I|tI7KDWopOOUNfq9hs1i85U^6;A!
zy8_pd8{;o)Ql5!AR7cYH-sU=KCiMJLrYH-jd()^JF=b3djAERZhb3g16~DhfXZP~0
z&duJvKK2;0Cb74ns`@CU<(4wx(HVSU0q@r8l1$ZCNW#wh{jaXntvtgO7q%eLsy_65
z#@t2e(g332uus5|>g<Ra{&J4icwY9C+{vpioqr2gM8yf8_6|QVcls?u;A#U!`=KxX
z4Sh9u#Jiob5wRegTv+|1)u^;?AL4uk4%)r}C8ZbPorw1dMVlY+hjZNxm;6P-r8fn2
zYkuYhYHli09(h5}-e`Fhp-#tkTa~dx?mMFK^*cO$7kw9`X1{$Bqy}NO{bk!xFuM}<
z$+R0cV;`Xv%n#V#<$ukygf_$<NPb<qZ2-UNKoU&hocTl*3Ys)aCCSYrC|4l@Ete@e
zR16tH5l*d<otw;Ym?T1~G^<`_G;PmJ@R<qhtvxnd^11C>{<Gw)QDw1!g)=Hq+QH3l
zTy(}_6|?M<Tlkd*lnCfy=YstX<+_grR|~jdjKc^;NX>C7>dVu1LOs>8+nKuKq4LBl
zTeeJj{l^W2od*v|QmY03w={i%<wRJJfQ?eP&l%|`N%tEIZ&HN(_7mE0#&ruoQs!cX
z(9t9H08Uu!o!qJT2V);gudC{ePTT<9$Q{E~8`a7luJ!ul&K%BCUshVGqzhO4p`yqj
zXE45udXXg6VVRlEnlcYzw;;Xdg(9l^;6JMxOR~P9@?CghgYRwBwF;OK6c*Y&sUs~C
z866<5Led$+c|~!py=c_BxVBG!5WT>&RL*P@Ggwcr)#bQzjI$z}I!jG<_`x<BaIUM4
zU=}V7=3s&!2%3+4YTWrV;qkr*6z(to^1NA*rBTm-ajdb`{dIX!?e9`?d5>gKM(d90
zF6?5r7k|gp^etx50GoXk<)c^%Z~AsOBoVXp0avokyP^<h>Tg8qCYYodHSN5gt`4|o
z6!nAd3+0mzpw)&fXkmo2(|~X(JV=}5afCXmATn9SdTfF#5BMmPLV>fU#BQY@c3uOD
zj?EkzLq0Tsbv5<S@7+_kLKotCC7<qHQ|>&7x;k)T_kv&f04bI|eIg^WQID*wJT?1P
z{_2!A)NxAt`pjnTS?T7t^>bZ*RPxI?hmyIVq8~d!T}gUmF;c)hf$$bTyCMhuY)(p^
zZ50=8pVq!3AKd$b@x2+}LH7#<vqlPdNM@TAruc9KZh!-mbWu}2;=ZMU;Wt`Fr=nRF
zKZ(6(V5Xl^UjLNa)a33R|3mb9wwSB-fQbtCTsDG3PdtC$OHYI7*PyN%cct%pjj3%8
za<K7lW*V%p?SD#YILYr#z1I+V6!7hDW$}YpV&E}3SHvO@e~VD#{SRMS(||V;pJ;&X
z4r1k6%d?{##9{&oO0CzX0eQCNrTFYJ8;aQ~Ra0+a@PNHZod!Hv&pF~8!v@jOfF#M|
z??`hR@GhVz5xl5D;5rH|L+l!xg_D@551P5C&306iupo}HVv72T$0K#l?*BPW@82`L
zCNArr^n+ABzOzOH&cj`hmYq8*bM_%!9)39aYo>i-Ik?vK#7Ar5Isw^1GDs&6skQX$
zpQ%G!Ig78Nm)+)mZ}HVE^Yj!3OlafIn<Z1nXuys~(D7Pb#*SDat&f*{L%7m_135G|
z4VXBM-i7;|NGw3~N(@r#il`xCV<yAsMXqD&Yi&#e?q6_iX?_m7ug+=wPi7S|jmkv>
zdOxH0{(z}*u7@El5E6t2T+Kbc)DIePc)KJO*#pu%<Ws{U1WR*cL^=ODXn_Ai)tAHO
z9Cgxw-T0DC9TF0|Ov&`<0F$F>z$%{^<{cJ?h&i?}@{{iTef106*@BAqgw*bQ_7=d_
zi3-stJ`5eKD6=!tImjV)4;2vvkn{Z4Fb3C_j(nl!`2(_w7&pdGD_R@S)fhYoqej&+
zYX=`5wa|beCd=XR{J=*XB#U>)V6{Q8NsOWa_KEbrT63q2@ViFNmZhO|l^EOmg5o__
zvrVGJNHpNN&niTFq8)u+g>wzeL~-~+F(tVWSUa$ONI2(u4Q5{BC+eA0jc%ADL_*(S
zn{yed`on-HFX$(j=NO!|2g4G33_`QHV}8Ty3rM~<_7$2|rtfr0<ec{ECUyV5nO#@9
z9A3<`lXrJKF3($#m^kdNa+hLvRJ7G3R7jlCz?;-Y*h9VRJ!H5i%ktxI-5PlP@WBhU
zb9eg#hse^fZ4FWip0Sme8tTR0o#T~@DHN37vDbPh6y40jZ=tnk`G+kGaYjiVZ_r8%
zx`^*8vHEqd|3=EgspW<b^{)yLIKR8!HvZ;;3*5#H(n04ciO3RC4;IqJ#zs*Zl6lpC
zJYT{n?)QH>H7HfTshP)1c_o2t8G-BkGM>qg7u?2|Zk1(b9p-!#3TH?LE+s>EG;xjE
z3)^f#Lvx(FOs`}l&-5(J+&N?jrjsn)9pSan139WClN-)O)*+2Y;0S^)u_4BT2|9Nv
zh|k4V;0@=fPfpk$ZQ&!zUE!pHIy)g9|MS5iCm7{rc5}igC$PuE2D{eNuAq|U0mV)!
zdQ@~hF@rnaVV!SX_S>j<HhIH5B{Mx!_M%Z!L2milxXw3$Z7h;~wdEF`m{I0Dk+jCE
zU*N$%Q!pwbdPD!yyU0;?W7osx)Yk>JuKw>vO%;4ICB3|TJ+XojGABPwCIY}25e!-x
zr6<=7s_@szVQG<sfjg*Y8{ZY8@Q}^}^GS^2CG@0|CP&yek-kgx$=X8c?&(V2Wap_^
z5Qt)rx6PZcvp?@DaE@dm6@?5*Lb(?u{b*MT*0Eze28Mj^Bhw<VaC~63{HsJEyu&cb
z`xqfusCQg%YeVOwRc_mml0B4lm~0ewOhd$rPpo`yy_D^_zz^MPmp|`3t~F`Heah7M
z&Ha`8$n0YZ3Jb8oKct>(9=_>_A00<Lag%#u9>(VvW~j0Sa0@a#VGj{sYF6uloJZKj
zLJ`a)vjs>;W8&GXo5Q<oF@LnuEaCFltLZagmASio5u8t0LmIreE!0IdZJ3Kq^;dIz
z;m-)ON%$VDS*(%?Qjb|d|6Spr7dYD!8Fs~PK18?TG4Mhcke4|?ljU8~K=g1EGH6{z
zg2koK)+N0xYGkHt$nIreL%?JAQ{A386-0yg&-C3K781A<f@+@qfCI<ma4JbH<S@Iu
zWb*Zxp93nWo#cB16ZL*}Lcf;JT)Q3mt6XK30LCIBgKohD4e>?M8|qt0u#XpvYqYFi
zBx`hRHDuqXu-?Dia93L=Gqbv-z1T)n+(+dtlJ2NU$*q7SNYo{z9;Qy(k5xzS>he94
zQ`^|SMHvqh*1h1D~Rh<q>O5ud}2cJxbZzXkUp>EWP?PtN~>>!-qGps-**BemxT
zDWUs|i}RqTqZBa7?hD+LpdU$4!y!7|L~!BVZ#S=n7m%DU<EbAcd6I9Pk<z)_ffnyE
z`l5N`7D`yf*u^v^@%AE`qlI;PT|nR`MFjDMg|AZugm+K5*;2lpSOKv-+1b?y^tRUK
zaae=9*rr0ea%wz+-GexE?LNUgHe@#gar~--CVwK(L__;?lZ8VlA=ub}MLDCjD9Z{B
zKy7{ST|gw*vXXqEUcz64?lz71SgqzhE346PN?vfewCrlf-*$OoXs$M~)s5cpC>bT+
z1kT5-@}_l^+%6^gS!t<eTtJ;zjQ3Poc#v>*S4dP|OmxD_<fdCl*0j4X4Jf~u<xoKF
zMxVi!Pc6#ErPPdxlPCEX(-IF?1_y(h*XN^e+?1Xpi9au&)JRgDy{Y}dZpA)%(pGIx
z;)g-K%nG7Xb2AL=Xb_S*dz$bvlwwf&D__$X@3s|g@4;bM)o?`dD{G}B=k@lbbaVR|
zS}h|ZWNk&Oj~Y}2ph1#$9ys)QQo@||s9P?)w90zAV&}xJIk8Ofh5Us>NMuslL`6jn
z$+Zu!sUo49s`B}$AvZ>RC$#DM=ik}ad1aCx#Pv>{lR8mfuFqK9Dk9LG|B>#)jba7^
zp@fR=;3~z;lk~+R<wX<U!PUq_iY{)B-vGa^IV>ZNWHT3%T94Fm;NpT_-7*j52v@zu
z<~{f{B%>zRy(fQ~Iss<>Sf461E>SivC%SC8I`WN=y$#cLOS2A_M*|{SNYG>aWYMJo
zC(S?<b<cEg7Yn50(55D%kRmj3W5ex@CNsOz{YRP$hx;Wt_y+iK%Lk2}p#2?5!q>o(
zP11(|HRON14hTOn$AVb7QjD1Y^0$ginHrVLAR4if1`DN9)v3~m_Em!9(fRFjh`C;L
zcN!`XeU2D;4ACrs_z7wVZyNP%8JH9U#v)sIX~0i4BO1_V(nAB-ts^NCImhC1NP@t4
zM2Y=-PG|TC4ss|wdzl8HlMwL&V?>G?h95zZ+3$cHo_2+h#Oi1O?)rZ;@dfB#y~+Dm
z3j@ri2yD*B^XlI`$)oa;vB)C^8A-|oYXsp!@4vcImk5GCA%Nqy1@R7r@d5M2Tc!PD
zn`6^4PGpJA_pu(T{Ylxsq{lR;2q5`g8qnbpYs*bS70t|&91pgiR=-sPf1vN|?X+jU
z#j`Nnc=jNVcw^+|u6fI4bIg^5*}2Ct(nc>lukIlW*6p{8%YOfVVT%?2zp+J**Z1#U
z0KJAY6;lExgh#4Z9m?@;nHe(5&s2El_iKaqg?!zEK_4AcQYvqkm`dmMU&c$m_5~;U
zFE;pP-KY^@sY-Dy`l{JsEkMKzEGCrxq~B#BIz>?A3Snw$RYSW0cQ>S`v^qVfI&X0z
zjQ)EW@N^|PY_n85FMoDo-ek|yLX3uB#O#&eyu;)de``)Ejc54^<$g`fDK;{3=ojop
zEL=gXeL(!yXnA^6B23C)OQ48RuLfR`!9J$}lbcw29}wTc01e3S^LN{Xe0+DNvKCt5
z^CW6IY^Ss96uy37X2vUfOhdCT2arHJe+c>^)9VINo#!HG7_Ga5Bb~OH0(#lTn-z(r
zgA}QWxH|yYDfaRu8qfz@A7B69fZCQ|f5eJx3gc&Go(yX=q%?#ke{s7+7kJ`wVtdON
zl8u96cI2ac7DA^Kv`^*Qo)RI0DB_69SMiGnGyo&u_=D0|(d`w7YtcU{EFxui5YisF
zS!~!5oQRKAf0AxZC(gVYaRbQ>qWvffac=v_A>J?xtJv<wa}OlaZ?oq-nKeAGy0spc
z?6n1bKYGwH&oZ$|f4TOU(tVk2tJc4R+27jsqXAXF<0p^lcAiWgS%3C^5sZvAnfXz2
zvYN{>tn&Nv*{GT|ZIxDfh^kpX0!0i{dW1ZthH=@3qX^V<N~_k?l4aO{rjip1s!K}R
zdtiWddK=UaqLZG({5D%i>uQuFX-UVDqkBRr{D?oRhrS_-8dd#lGpR0TD-_$f+u9CP
z0~rZSL}}7Zf=EZ@Um(h&YrUYNPK<wgA%3QW2rG;>c^h(~N6hoiX<VHLNyIF1Q|Ufg
zZl#5_e##EFO~-{N*5|8?rH`BE`)Y`R!++B8aSPr3CZ-p7Ps>pI@`-QPa*bZ3yqI<o
z`kFXM0|5A=akLW|5^H;cl=4Lnr#Kkl7y7t-ynxSXT>68cQ*edQl7sl~+k0a2mn$}C
zfFpQP@Yu85RWx_dU6&fF9AP_ZUQc!8;h6EjNVT~9Dbmf4HSoguuLZGmY9N~*vx)@W
z2Wi9#(g(5@DYEUi`)&Y&S$#fBfO_qRW!9q5^bhu`kc)arTPtdGccdWBLEpd^r9I+u
zs+L3c`mfsG=_pbwTD$nFGu7-(cSS!v(nPFv;j)8H&8_x^68-$*`?_}A8e%Uh!U-Nk
z5J~&^Ubn4CW(IhtD7Wq-xz21Zl%C3A_tWnR)%RZpBf@v&@3yspq+B4?6H!!N$_Y|L
z%bnLU9?>~$zSr<SdSm?7Y9D@2H9MiH=x|>A+%5f{%gn)GHRSeYbqQa_)KpaIjm<ol
z!nzObLRueF9qn2HI&GDA2Wks9C5=em@XHGm-3XRneO^yLle+tChHmnoc2n)qxgPOE
zA$U#riNcc7vpDJHiwBCg$Kx!mX9mlubU_`q9Qc+rw&)g77h(>FikNFs^j@G-=WXFd
zznxp(YHOm~67DX1Gc<YeY9(K6lC(}fOXBMBw_DFXF}y+Lp*$Qh9L$3yc6gR1h?pE0
z`NT1#UTZa-&fn~1<YAmq;e)IH47v&9ctAizc$m@vwXx0ACCt;f`8`u0tE@<@?^lST
z36hUA@qpxw<+tjj3XWbKu2vSC*(%#wzH2_Dsj`?2Jo+XCoVjc!JhSEI00o_xp^i{R
z37Q*o*!pVS>(7HEU}@F{^gS+O)U`gdJJrTI9a*1a;x7-VZdotV=M<>0Qacb#LEO%t
zPzW}27!Bx}u=_FAWQ>b_+|t;RxwB24!O0As;40_2@uCNdWc#(2eVSx|-(P4cC>Uf3
z196BD2J(tFf25}C?MBF^IFq{!JLg$%9oz*b)|O&R&0&dJm191Qy}m56M9#XJ#u`sT
zKObN+7Rhk#{;39nRA%NYvLNrH7__AK;x@*ZC_hxqat!ad$Tl0<iB>R&5}1BY^J47j
z+RE`~_gk!VGkoY|XFrv<($|$4crCTuA<dOiJn@moO5`j>rb&S<sztn{am^XtGxYUo
zobqlA><9<%EsA^NsrDqX9PEsYQxIv*2@SeVNNDutJU404Tk})P%y*{m%DL8<X(`1g
zc8qZ;tuhB>eC@p)qG8j~JNQ%Fu0l|>gHs^(C@3J=t3dbjR{E(o$$nzJKN6--GQ};>
z#Yknk^lO%X(U5plPh9RW_|-T+*uKrCS!hmT)EE~5%M6qtEqXpfY_(pw2&bDo(-z8P
zzCH-BinoJZB`0wmM8}aWssfyQBoE);9cARf0W;QzNa~II9L?*r1FGsbTy-Pv0cWQ<
zCbsS_YUf!9I`}-{vCzNI_~}%O?@<f&nmLS(=uB{r)0u%KcALH+IXxIQUdkpet-_N9
zZONvpox9xpy^N`Myj*&`{ZdSwdLh*><2R(i7`FN20T+4;j;M~^6f|mTcp4tlj-8c#
zf5h+w!As^M?vY}j&Vwcyae9k8c`X(Bx}PRBzA{<s!3s~Fmilu}$Z3J}DADx*jAISQ
z4o3KciCpp7e2J{i*>sg7G8f;R7uITx3g5DbI^i{dyFDP@Cc}1i9dj`j+lY+zXX^}L
zJ2hve=Y~IoaILe?{j#vprVq?iZU}Z;>}GiOa9;A`bo0AU{YIOZCC=Z@n4Sh|54a)_
zeGE_>`h?m<xGqldTKW`OuKnVn9@+BRJD2AwgZw|$!O~>}DgJ5p#7BuaE&LGJTj~s&
zZ5Pa~Q)_z-KNwwF&9^>kQ{p$i|FGn)Kted*ocuPV2vUAA&*4kZu~Jwn?|IX3=iZcg
z?Y7;^1Y5*<E}Fl=4>=&oyIyYnmvY*!8~DH`Z~|$LohRbIkC%T}`4pO18kB<lky)km
zP?BKy6}@B3Y^lQ@<b0HkygCNP!fAlZ%-<4T?G4nyKRFw7u4F{UZ+Atx%>}>PnO>Q=
zX1^H9$88nSqp~?%6N1$zmGA){v)cEy^NGqOxo3;kucX`tGrDIQ552Qm_8>$=(@~y1
zS?-rzsvE;*2ds|$;%t|05Gl@QG70kB+M><xXEk0^N|pLt;{mfBoW0)r=X_GqnyV{N
zQkc>4^u*}Jlg+=Bu+PnV8e4{2F9kxEryvLNInm&~g+jzi9z?@zzz@CS4L*-p?|FDk
zt2K^Ep_D=nz2e&S+@_J>e@UQdRS~8E-`bBKzh{*+VAW~@J<*TgLy?*k2^frDgjM|#
ze%g;XUbAAU8=byM?Ed)jZmCO2O;Wy$Q%O)Sm_R0oP}>PKVCBo<ou){lI|))kQNk}3
z2jUpxWt#n0Al4RY-!0*F=KD@dPZY90%scU@G}|T*@%KG=Gyz^@`iz)p$~ioPe2;dT
zI}rRU49biUD-*~f-@yaAHPGI+2sH${BssZH)!L3gufNSXlv(GUJG1UE*A<kr=rgNy
z%xU>LAw<^tV@LIDx#lGg&hJ7F_Yx1x8eo5@4?NfrzZ8~ffVvC}5{zEd9YU;+gMXWA
zVwEj(I&CBNI3GRyC+GV9-D86Lh{V;}kkixr!bVX))r#%Z<hLs0aY1YuTL6shg{JAs
ziNjMa@`{4IC_5dcVKzDvmoaI-Ot;8oxbNxwl)S8~e<xg~ZQJ9-me$$~0KPZp2=nOh
zcM6kMR0e5QC3NUj$7!-Osh!Z=6Ix7uC6xRsuOo<WMCr<>$(2LpOg~<?{AYLFaNK9X
z#Uer}--J^7T_%il#Ci`7-wB&q51+lB`?>d@G5dF>>AP;Sh(;kO8FO)n^)SexiUQE{
zuWa+nDrASL>bP%3k0nJ63SwKMborP@2LH3Os*eUR(SXIrki&x<8lWK)cpoB3kv#6#
zM~L+$#1Z#cpV>lU-Vw_uJUVH7NS!n0?6)x)-I~$oyL#vQQ6B7TES1CE;IZlB6`WOD
zgrDux<^T7e1y{FUqv_pLI9)~{kcSPmv8a=bhGX15+rQOw8o3un?lSBGUS`ORqhDJd
z`Z==5CnP?+><CsPOJ;AkuLRa;Xz0uGC{myL*n(V<pez+^aF4NgNc|PdoaxJxoR75V
zniz?1juq28d?FR?f!<Y}K${JqQ7R=KpGYIvhyuT^n>Ut5%dS1SqbgkPc=wRra+oUg
zPnZ^quI>p70+IME5^=|Z;k-MQ2{ySvzC()C%Ggg2xz3#cHDtrQyJsuA{M$P7MB>FN
z(jb)3Ku{7Ua1hbs+tqw?Z&u4;kCd*<oE|)ByZ+>q?7(nXT<A7Lm-_dF<7yfZ!>beY
zlLnl-ab-fu16MCpH-u|>ByGu>Sy?&N@w_PfSxHmxsqLH!$;LG#h;kQ??lNGpfg!pM
zSrVhFq3^_n3n*9B7wul8d@tYlu14Q`c|zS1pQZ1>;BfXu-19})_Uon#A1R9X!Yu4I
zv&;CQX{3WdAJppU&UF!?<&yK-!qGy1qlO1n#Nc808N5`tnSkuipxZVSW<pu@gS1C3
zUM#9CCa-CLmF$OzKUPW}0CP%as!;8HLc#XNYi~jcwx#5iN@E_0LmP4)rlY%jQ@X+D
zOS__(d%or^^H4;!KLG5{{c0q7j5J-CZ)A&Bxe+3Hj$%oGv#%YNq5VU&-I{*0rq&#Q
z>Jwz1<X^j^xvzS10~uW;_Ndwe(b}FRD%$L)qKO}j+XZ2u7AQQjOA_<alu(+Y71weu
z?LOqJ+UIkvxqT?JO(5yb>#9sUZe*QLJa5OnVK?b!*#RaN*<s<3+2Pq+N$R|DVqX#f
z`ZzN$LfX7XNQ(+dVY?$p=I15~D*7g>$es>2guj=1=Jxw(rx2B}b6Q{qwgeSPk_cWA
zmlm?^0fLYs=Wn`#&${r&cQdDb^ZDKo*!f|;E`I;q<lx&o+CrmPXuF)gEWhguznAq~
zT_$*g4K(gZ6%v`lza7N1?%#xVX~Y+PC@gF|)&k&hiA<BG)Tb0p`q#n0XBA!>7?@Z<
zaCnF8L?-Uj04@}Zf=PW1-i?~7q-^hCtEL=fLRPyjKcLkn&v0IR6ET@bw8o=1R<YNg
z*1VQU<gt~^DT5W_U)45*k1+RkoAQOd5^?pRFEg)@13M2+X0(H<BSVWud>f^!CO)cN
zOTy`zOa>NOv)U!gYw$TgKmVb_Zj?@5xk>(@o@<X3bc#A)-PF`=vX$mk?2@uN;a5Jr
zx_=JUR8!f_D|*!d)B7DJcK>q1dt@ZV4hLZg>dU^`gkNqy){SPSB!8sTH<6I}p|md9
z@#or3TDf9#j*E-MiVfK>#GNTH!d${Jc;=iYsbB#;6vFHFq_<;Lt`Z)RUIi2%KN1pE
z>Vm@k^R4_>4Ld*2_v*ZU4S5A;J<4ffCo7RuraGuBLmw+l?h{Mzzmb`tygw+&ws!cH
z{c7F^08aYiO4_xB`Sxz*>F8_`Rj!aYib#CKFhUkG2L6apxWAAX>m#Z%<W#tpIydQu
zc^<qlkod4Y-1FW6{@xV>LzX~29Qw=(RUZGF8j4)~>55UNIErT|zMD>#@E)wYd`|On
zu*T^xVATFbM?S?G@4FnU1ZMDO>q4-$V-jBDtT@?cUJZ#lgqU*wx$1fL0{nH;8->t6
zDkTKP$46hAZsVm|d1W?QKr|pcgJC~gsHyT<rIop*$f5o1?Ahn9^Na`S8J@<Qg4p$?
z7c<>FI!#=(`=Rx1YjYQUXh5)kSPMG|O16aht0X<sVL3s$)pEl|o>9R1<;l|t(&SHa
z=kGE$`wwnRVG|?K9Dk1!WfG-|&cNLXWBEart_@XD38`_XuQFYdoK&;lgKR&Es|lJR
zLGe}X5XPpPQ0(9-=Zm%iLz|;wQ<D$4UeWW-wCpAXG@kFNm-%C`*?|`u?6Bo-A^pK0
zcC*mq%t8&<(#$SuUD0!FJ1u9r|I|}Rz}%5J=xI5i$>24DJUzPPsj*We!sji(MmBIZ
zge@EkQuX=uJqzhTsO1Qv`m)#Cx)82Z)9^@Nw~wHLz}IA7lEh#Sq*A0ySGU^exsQ0f
z?U~aDdFewe-|u94dSASwcwTX$6J4DX<L*xJnwN|d;H^$SfJ>c9S~%gvF5uRcurs2;
z)?{Q)7Df6W)zAQ_2i^KyW&c*vya(rM_SK1&D;O4W-RjB))1)WwPk7z<erM?F1ACSx
z=G(wRrc<<Ev8Zyqvxf1bFq5g_8F%<TUL5DtZ7{ijWyryYSdu)<@lMejxf2@&6J?gu
zq!&v|oUd9;82cl=9qeqq_Y~_PzxIh<^~b7PmE-}jK=G?evaPX0(`%Ykkr?;k`RRk?
zt$~M?EU!<@)G~@dIfh{LZ30)1^7ooKS_L@RoUr_K<4mu%f2#_{Tpzuheovv|tg2>?
zZKxNh5Pi1g;tH4%&gF2d)~wz?-|Su$PB~>2Z7`dvHj}sr08aq%NXY@hs27s`?WoUu
zO6<Fo<%a6yhM?{2qYsCLih_N05K*71D6zW3Z<bk@$5;1^-laEd4oj+!Sp~JkB)nQc
zMQ9f$J&TZ$*ik%bVt7Ye<Re{)GiEjmr+kM+R~wBQlO_jS{yineIUjhV(~!(|If2b}
zc#R0Yy_hDcT#Y%B{MD%Lv_`!$+yo|qH}&)f^~etGmR5n4*sFh*pKnIXvDXi<a=!3{
zmHZBVQqLD9#Jr<<a5c7&_533vsAlbZ)RKzqQ6`*)K;K!*ORae7b6C)RRrdzh$b(B)
zHzDuLSpmAnO~FnD)0T-)-O-IbK?l$AG}U3O@Xj?zGvB5NG%l{V?ZPKjmy(Z}!ur-#
zra?|IG@#=$T;`bH1QpPLP^(0$L?X$aEdssEHj2hWz5WVg)T4)?T?p(OmA~<4@Zb1z
zY0vS}t4X=T*a10YT!WBg{>GKpF`qxOU7!IG_x=W#?;YtLV|OixVgthWf>;nMpkl=Q
z_yP@xMY`q&E16Mw=&{Hn`aMYs?=*togZOtVn!m&V1jxUcGQNL|u}<I<6lKC$Xu5vN
zl`+A%o(o;Sf25td?EiMQQ*`Z)e4AiB^5K6^@x8vDvvYD3ygqAi#4zy#QngJ<nqxss
zh=O;a6C(}Fl-Is&hx;kVR8Qs)*9Mse2)*op?0q*l;@u1P2-|zHjYAvkD0m1UdgF1D
zWX59^v{Hy<!gEog{Uwk5dnhF{O0Q_ZRk<ki2ER8A$W3pOY=!Iyub}tVK|5as@!6#Y
zI7v1CnCpeNK4zXh)B2ghG;d?DFk3H?Rit|}GW7Z4WbryXeuA;NiSfU+cP2nhZCe7p
z$S8t{f(V+#CnyL4GNS?kks%@kP*7%l5W>(5iU>lO1r!1i5K#n!G74gt6(I@pDD!BP
zAwiHyNRU8;gmkR#`rYsSuK())s{gyI`t7P+b*s)D>YTG{-?jHSci&p+lG^}_TRkf~
zCCR=#AMDBhhbd#f%0Ec+4v>k$3(xoUFRzDzd(UxR@hV3&X+pe^;j~oPmsHqX)zYiu
z;%+F`#ncidPYwfKar?{uhTnlPy?D)qMBFi7ZG+OAlvxlE;QCuMwz-Y_!TXT|)?yqJ
zF|x<u9KicHl-%$iyUjVyoDHnWE3;62Y#>*Pkn_9D?Q99HePAc9pW)K6<Slz;MCp@N
z<$$`G=19Fsk!H6U2iUI$Fy+q3WZl-g7xTB9F@*#^rf?O*s?a<vOm%TmX;=(oc;&_r
zG5uipvMc7&^nre^W@lpMrfcocKSIerixd$2Sf%RPI&camZcDebwI*jPJ1-CLudu(T
z0~X<-<MI?b!gW!~G&n{B%D+eus~!A>*kAbWQ9;GvYYQ18mmltVS=YC08jE_gsykEb
z?M>1n8g+M0>W3eaZ6u0b$_#43#K*3ed{c5KeJAB&D<{s2C?7B$NNOIZN(Rtv^IzF}
zkivRctYJ+}&GM?D3CE6~?huhZ{!m<>CpB1{hHUK{Z>NZ~Idp?AUv=T}d@eJ@24>n@
zLO?6Ih$0djs8v;#PK#`e>}-&HBk#wMjc>(?n)yMJ!%y_k6kt&?ClV<*F|W%XRh(^}
zt0(O#qsu=3QFt8e7dS8KODetOW7b>|PSR6gux9X(O|aDHZ4D~&9<Jc<L59QQRI?cD
z4>#ZQ<nrvgcwSRn2^(+~NzQa>^NS`azPgE+fe1So<pyI6v0rM->tiqY;YEINe1XtO
zHtn^EY#t<tF+KX%KRH1y63NxEv9G;_4%RenXjAl0i(gR!xXW<zjAz@z%=3(N%KiGQ
zpv~C&!!z1L-QW|{<ietbu~`+gTJgirfNGZoAlWWHaQUOQk-;MRMQ`5c$LMXdHfPoT
z9WcYyiYD<u!_4Qf;m`}U2^S|q2gQKdIS+KoL~_XpGCPJ*TogYHlwVJUX*XC^m#7=i
zN|3Ih!Q-k*?y<hqgxt0A#;a`3Cb>0$VCQ%9;lZDFb}w)UW^|bj1qBawQ!4E`j@MMs
z2>2owyQb;uqCk=*eZC=P!FdC_mvH7n!UNgnDx*U-wgf=LU~MpG!FVNCFu}J=wty+X
zm?2|H5@bTgjEIm^Ud^UeRG?e0(1rvMzBY+id{Zc$($Of%BlVV+F$0Mua5HsBPzVa^
zqV73Z8M=%Iqe{7idf0-u8>a$WS%4{HvkhAYZ;99+Ojv5o<g3eJYSPTfJ#(egVJ?<K
zh`KYq2+P<*sW4ez3kz6Lz1lVx%5<pS1<9Fz2F8eir5rvug$SKVLRt><K0}^K&`n5;
ziPpRP7C`S44j}1_65`(8!DD1=sQigLAe0~#-lotaWv644%b>X2C-uflVKq*zWD|x2
z;~MvtJsDq-w1x~4&GjHmed=bbGXKJJIcm)-zt{62kB)hV)**%Rz{L^oyKYIs<qddt
zDYW!8_a&nR)KEG{>AW~1Ba<2i-`V}0raIQ56*3mZbSPlT;qec<gCJ3K{P{N0Y+q1h
zTj#j{3losS%es7pvwp7fUS7=hVQ4k-8*g5CpUZB#uI*js>i(&=4L$I{bhne2#ENE1
zO3sh_LS5IAV?IZQ%MIvA1a3<%U@J3PK=zZc+gJey$9aPC!3Xsu!=hO<!>Y^h<}aU(
z%@F>o-h(?wcABvJm|T=<8e*l8i~|LCN^~c-gsY7Nf-Y`|f4%3fUoyNV28Qco(m$=n
zL8R#CPA;!Xknt0F^8@BQ6#7hxK%=eoyQ9eh>eja($rR9U=mp78s!aMx;JU(KtVzX4
zRpL?@+zkH!SyNw?j!5trN7zZictk&6?ZLhsYu$$LML~uMQXUm1ub4)RSl=DUpt|W3
z$xA^%yKYP5trQvK^NwgYky|=?P=;(!tDlII6NHOt1Qk+WDsX>hO2s?~RPa95ll`H*
z-mkh=`zRV2sZh?a+7@RTFpP>PmhK2o_XM>DCDtspOdcZ+`&FN8e~~^m%m2=@b~VdZ
z)M~Ekd0c6oNA;P|(p`Pzx9F;GqF;^5`cV+XugHgF)I0xYnCa^2odXOLSK2CN$pb4~
zf3koPKIpC-THNR+{b?Umy+Rx!pUOAhrlqOY2ajNU-9E+iq-`dydn!qse-45pl0J}R
zz6&U(?DqfuK;H)s@6Jl+mQ=u&g$ts@sGAK$zUG|y(<>?E$7h)dI!m25yK9%vl0)Q}
zur-zMrB6nKugOZC-0pi_pOFzjcZ_>u*JO8F?bSt%a{7S{%L%KR<N%kRh*%VJOgO-s
z%m&7ZWQ&AMa18539-+|3yJT^{qd<Yx$?>sUpd=!{$H%ou0RU~=*SbnlY9?#7xa<nH
z_BpjELUD)^8&e3FWyNWxNW%n|&ov@-$5x(u7cr)$t!N5l*#xHMKAp|c{fh5=qXQ39
zpfSanj87L6wl`83o;2mQ?<Plw)uBQQnTUI{{%pXr$|@+Z>dpxp?)j#7D&mXRXO}%a
z#XXPvRiw?iL_4uo9@maNoIb5%!2fivr?bsh!|lt;#8rX8$85me;+ChK#-4K}*lUb-
z$`@P6{u5Lwj~E*R@~3P+%Xi*b#T&7(3mD(#cXe)ohmK6IzcFl@0yEX7F(12|z;cz;
z`uA>9m5GXwruUbBtik-(-oI8Ux9PoDxnI96<Ur(IzexLfqPRLV5hA`)I?4UUf%C}m
zcDcE8q27*~L-T{@kQ3g$Ri%?4#_nw`24(Gx8+A2~jtsNPJx`fD;~r=?zPIBlu0+Hh
ze!#}r*OsVMV{dFDMeI5lvJgPPh+O}b4OA`}d>?vsQSGAj=3=PS&?CiAryhA&exQ$5
z(EQF>lsOIlAaLqlrg9<yJjsx!*LxRCk~#tf{Jza4*`++{1Z$cKC9a7M$?9A+C<w;2
z6>dz9UI*`oO-miIy{B0bb-ir+8XGVN3QvvmcOHU%yAn}_YhC4IG&?g4UV2EaWaFkd
zhP*@}`7}A65>-g=uGlYD{_r~3LL}c0Bf8)$#j4fNru!I=abAT@nxmsG-s!(?Uuc`>
zP2wQ@1Af6X+S;#Qa)SI{S@~nt)$3!M@G!;M{I{##cN!fO5`x9~=C67|>W4NTj-Kp(
z-s4>odZ(YPCy273LE|`mi!h<aH!n>c@iJ`7jGb+ippH8j96oKsXQx)%Y5W72S9399
z9%SGVv}a8Li=$g%y_itMU8U?I&x|ME6_s4u7@eSS>R0a1y997Ltkg$)R?+d?vhD95
z!&x-5?GrbPm<SM&d~P^m<KlextaX-0W7ndkFUix_!M=+Xi0|9;!1FpS!Z|Z1*BVof
zTIgDvgK5ZoN~*inrs!bXm*ep=VT*h1=wDZGmD`nE%La}*8Sdb}a466m)w-?620B61
zG1e^(WE<nd&i&?VLY447;+n}%>UYdz0|U5gwhZ`Hr5%NJHXvO<`^vmOd5jH|1@`SE
zcu|R55gi#n*Y#gVbm52Ev1`)7i93yB(5;gtO-x=6ggYm)0q1fyAm0Sty*RvcY+b8J
zc2~5&JoFFK&BJMF6{`O$J1;n6D36Z*OX(|))@N+s*g#q+4a4&1bjYOLx!Un<s|Z&s
zO{>AaJ?=zZr~v)Yi0}uGBL$#qpgi<iE*`jBlg=6SI_MS!yCqO!gOOMsqcB`ehbWUd
zMLF`A<@?&)JG>U0u*{!xSMeT0Ho&_tf@ZVA3dg$NCE8{L*e7)UzV%8(vvzR%PQu&^
zHV}Ou$|%+&!Yf5e%xX)G+&BT_Z`+{cEFD)iU=+Za*~SYHrVYfU@MHW){^wC#DvH0@
zsV33&rN?8Rl3<KwUAE#W&Z%#bGX4wLi~j2Rzz7~FvS-h9cJ>gdi}dDgaQio@(v};b
zv@-(5yT|bz{AnCin=3_Mu8dyYI(#{2hQ{A86;*>3RA0d13Y;8TNzpMyUg)zyYcKmI
z&^H&dc7*`C76}R;e77pgI1!2d-&2uQJS{ia%#b>pH1as@<-(Rz_Ao|$W5c$*`ABaA
zAJPU;ISRnOOxsqAE2)H3c}k_B9bHePO$60U-F%uC>}uSvZZoisKhbVJC~jvIn_poB
z2*Ga57VU|K$tK_OBn8E*XG4t4!Odos6@#Keo;Q4aQyPi_(nTZ1xYJUS=c+t%AL}Nh
zhg8-*dRA9AvbasubuLuAK`m@N>j)Fih*{rbI`iIa*i^k^pJEfzt$~+3CM!JbHa)^Z
zHim;~2lHKvyhwYWzM2|89cot6tv&a6B8unGM3meut`6xGhbvdMwU2vCBhf8g;e$4A
zvv1x8jfJ7Mo^F}MSCw+%_GT2=Efj2!->f9%zc>2&VJgIg#P?NO(lym%5T%>%Asj+-
ze~K)qUj7z%+edLwTBm0-FD%D5Ao<#t%>I|JjO9li4s})Axn!D8g$31&tzQ3W@aO-#
z!T+ZUp8Z?F`1;Ck{Yo{J-?BQM%pW&s_Qa~i8ey+5z*-s=Cma~=CMK|;9ZU7U3(FN6
zOu%~vB9=y=>7mGEx`GMS<i3q>j{Dib2EbG`;`St+g5}Wzdqa+>f760hdsY#5Uu!-R
zQd6Ak-P^Fshujyjs--W|uW6a+)&AmsiIlZ<SQ74EwT?xVzmjejV9K%E-2chx{wtNM
z;#e0t`Ncq#&}&tJb&Xk(?UPKol#hKeJ_fF7HFqPK07{UaDaYLzL4KrnfHPOwK5QV0
zb3p4_$BHmQ&8Y=R-gyfLEf=-|i~lW-#7inVyCaj83YSSW_Zav)OA`O{6!QP}UlV`I
TpYo^t-;#d^B!&Vl0CoWYl`KXr

diff --git a/doc/pub/week41/ipynb/week41.ipynb b/doc/pub/week41/ipynb/week41.ipynb
index e0af85976..4ed69faf1 100644
--- a/doc/pub/week41/ipynb/week41.ipynb
+++ b/doc/pub/week41/ipynb/week41.ipynb
@@ -4,28 +4,28 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "<!-- dom:TITLE: Week 41 Tensor flow and Deep Learning, Convolutional Neural Networks -->\n",
-    "# Week 41 Tensor flow and Deep Learning, Convolutional Neural Networks\n",
+    "<!-- dom:TITLE: Week 41 Constructing a Neural Network code, Tensor flow and start Convolutional Neural Networks -->\n",
+    "# Week 41 Constructing a Neural Network code, Tensor flow and start Convolutional Neural Networks\n",
     "<!-- dom:AUTHOR: Morten Hjorth-Jensen at Department of Physics, University of Oslo & Department of Physics and Astronomy and National Superconducting Cyclotron Laboratory, Michigan State University -->\n",
     "<!-- Author: -->  \n",
     "**Morten Hjorth-Jensen**, Department of Physics, University of Oslo and Department of Physics and Astronomy and National Superconducting Cyclotron Laboratory, Michigan State University\n",
     "\n",
-    "Date: **Oct 10, 2020**\n",
+    "Date: **Oct 9, 2021**\n",
     "\n",
-    "Copyright 1999-2020, Morten Hjorth-Jensen. Released under CC Attribution-NonCommercial 4.0 license\n",
+    "Copyright 1999-2021, Morten Hjorth-Jensen. Released under CC Attribution-NonCommercial 4.0 license\n",
     "\n",
     "\n",
     "\n",
     "\n",
     "## Plan for week 41\n",
     "\n",
-    "* Thursday: Building our own Feed-forward Neural Network. [Video of Lecture](https://www.uio.no/studier/emner/matnat/fys/FYS-STK4155/h20/forelesningsvideoer/LectureOctober8.mp4?vrtx=view-as-webpage)\n",
+    "* Thursday: Building our own Feed-forward Neural Network. \n",
     "\n",
-    "* Friday: Playing around with our own Feed-forward Neural Network and introduction to TensorFlow. Start convolutional Neural Networks (CNN). [Video of Lecture](https://www.uio.no/studier/emner/matnat/fys/FYS-STK4155/h20/forelesningsvideoer/LectureOcotber9.mp4?vrtx=view-as-webpage)\n",
+    "* Friday: Playing around with our own Feed-forward Neural Network and introduction to TensorFlow. Start convolutional Neural Networks (CNN). \n",
     "\n",
     "Reading suggestions for both days: [Aurelien Geron's chapters 10-11](https://github.com/CompPhysics/MachineLearning/blob/master/doc/T\\\n",
-    "extbooks/TensorflowML.pdf) and Hastie et al chapter 11.\n",
-    "\n",
+    "extbooks/TensorflowML.pdf).\n",
+    "For amore in depth discussion on  neural networks we recommend Goodfellow et al chapters 6 and 7. For CNNs, see Goodfellow et al chapter 9. chapter 11 and 12 on practicalities and applications\n",
     "\n",
     "## Setting up the Back propagation algorithm\n",
     "\n",
@@ -490,8 +490,11 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 1,
-   "metadata": {},
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
    "outputs": [],
    "source": [
     "%matplotlib inline\n",
@@ -559,8 +562,11 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 2,
-   "metadata": {},
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
    "outputs": [],
    "source": [
     "from sklearn.model_selection import train_test_split\n",
@@ -681,8 +687,11 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 3,
-   "metadata": {},
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
    "outputs": [],
    "source": [
     "# building our neural network\n",
@@ -758,8 +767,11 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 4,
-   "metadata": {},
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
    "outputs": [],
    "source": [
     "# setup the feed-forward pass, subscript h = hidden layer\n",
@@ -923,8 +935,11 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 5,
-   "metadata": {},
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
    "outputs": [],
    "source": [
     "# to categorical turns our integer vector into a onehot representation\n",
@@ -1022,8 +1037,11 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 6,
-   "metadata": {},
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
    "outputs": [],
    "source": [
     "class NeuralNetwork:\n",
@@ -1144,8 +1162,11 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 7,
-   "metadata": {},
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
    "outputs": [],
    "source": [
     "epochs = 100\n",
@@ -1178,8 +1199,11 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 8,
-   "metadata": {},
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
    "outputs": [],
    "source": [
     "eta_vals = np.logspace(-5, 1, 7)\n",
@@ -1213,8 +1237,11 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 9,
-   "metadata": {},
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
    "outputs": [],
    "source": [
     "# visual representation of grid search\n",
@@ -1273,8 +1300,11 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 10,
-   "metadata": {},
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
    "outputs": [],
    "source": [
     "from sklearn.neural_network import MLPClassifier\n",
@@ -1304,8 +1334,11 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 11,
-   "metadata": {},
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
    "outputs": [],
    "source": [
     "# optional\n",
@@ -1387,8 +1420,11 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 12,
-   "metadata": {},
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
    "outputs": [],
    "source": [
     "pip3 install tensorflow"
@@ -1404,8 +1440,11 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 13,
-   "metadata": {},
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
    "outputs": [],
    "source": [
     "conda create -n tf tensorflow\n",
@@ -1421,8 +1460,11 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 14,
-   "metadata": {},
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
    "outputs": [],
    "source": [
     "conda create -n tf-gpu tensorflow-gpu\n",
@@ -1442,8 +1484,11 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 15,
-   "metadata": {},
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
    "outputs": [],
    "source": [
     "conda install keras"
@@ -1464,8 +1509,11 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 16,
-   "metadata": {},
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
    "outputs": [],
    "source": [
     "# import necessary packages\n",
@@ -1515,8 +1563,11 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 17,
-   "metadata": {},
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
    "outputs": [],
    "source": [
     "from tensorflow.keras.layers import Input\n",
@@ -1540,8 +1591,11 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 18,
-   "metadata": {},
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
    "outputs": [],
    "source": [
     "\n",
@@ -1566,8 +1620,11 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 19,
-   "metadata": {},
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
    "outputs": [],
    "source": [
     "DNN_keras = np.zeros((len(eta_vals), len(lmbd_vals)), dtype=object)\n",
@@ -1589,8 +1646,11 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 20,
-   "metadata": {},
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
    "outputs": [],
    "source": [
     "# optional\n",
@@ -1635,8 +1695,11 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 21,
-   "metadata": {},
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
    "outputs": [],
    "source": [
     "\n",
@@ -2155,11 +2218,7 @@
     "\n",
     "<!-- dom:FIGURE: [figslides/nn.jpeg, width=500 frac=0.6]  A regular 3-layer Neural Network. -->\n",
     "<!-- begin figure -->\n",
-    "\n",
-    "<p>A regular 3-layer Neural Network.</p>\n",
-    "<img src=\"figslides/nn.jpeg\" width=500>\n",
-    "\n",
-    "<!-- end figure -->\n",
+    "<img src=\"figslides/nn.jpeg\" width=500><p style=\"font-size: 0.9em\"><i>Figure 1: A regular 3-layer Neural Network.</i></p><!-- end figure -->\n",
     "\n",
     "\n",
     "## 3D volumes of neurons\n",
@@ -2190,11 +2249,7 @@
     "\n",
     "<!-- dom:FIGURE: [figslides/cnn.jpeg, width=500 frac=0.6]  A CNN arranges its neurons in three dimensions (width, height, depth), as visualized in one of the layers. Every layer of a CNN transforms the 3D input volume to a 3D output volume of neuron activations. In this example, the red input layer holds the image, so its width and height would be the dimensions of the image, and the depth would be 3 (Red, Green, Blue channels). -->\n",
     "<!-- begin figure -->\n",
-    "\n",
-    "<p>A CNN arranges its neurons in three dimensions (width, height, depth), as visualized in one of the layers. Every layer of a CNN transforms the 3D input volume to a 3D output volume of neuron activations. In this example, the red input layer holds the image, so its width and height would be the dimensions of the image, and the depth would be 3 (Red, Green, Blue channels).</p>\n",
-    "<img src=\"figslides/cnn.jpeg\" width=500>\n",
-    "\n",
-    "<!-- end figure -->\n",
+    "<img src=\"figslides/cnn.jpeg\" width=500><p style=\"font-size: 0.9em\"><i>Figure 1: A CNN arranges its neurons in three dimensions (width, height, depth), as visualized in one of the layers. Every layer of a CNN transforms the 3D input volume to a 3D output volume of neuron activations. In this example, the red input layer holds the image, so its width and height would be the dimensions of the image, and the depth would be 3 (Red, Green, Blue channels).</i></p><!-- end figure -->\n",
     "\n",
     "\n",
     "\n",
@@ -2355,8 +2410,11 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 22,
-   "metadata": {},
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
    "outputs": [],
    "source": [
     "# import necessary packages\n",
@@ -2410,8 +2468,11 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 23,
-   "metadata": {},
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
    "outputs": [],
    "source": [
     "from tensorflow.keras import datasets, layers, models\n",
@@ -2448,8 +2509,11 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 24,
-   "metadata": {},
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
    "outputs": [],
    "source": [
     "def create_convolutional_neural_network_keras(input_shape, receptive_field,\n",
@@ -2489,8 +2553,11 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 25,
-   "metadata": {},
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
    "outputs": [],
    "source": [
     "CNN_keras = np.zeros((len(eta_vals), len(lmbd_vals)), dtype=object)\n",
@@ -2520,8 +2587,11 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 26,
-   "metadata": {},
+   "execution_count": null,
+   "metadata": {
+    "collapsed": false,
+    "editable": true
+   },
    "outputs": [],
    "source": [
     "# visual representation of grid search\n",
@@ -2568,25 +2638,7 @@
    ]
   }
  ],
- "metadata": {
-  "kernelspec": {
-   "display_name": "Python 3",
-   "language": "python",
-   "name": "python3"
-  },
-  "language_info": {
-   "codemirror_mode": {
-    "name": "ipython",
-    "version": 3
-   },
-   "file_extension": ".py",
-   "mimetype": "text/x-python",
-   "name": "python",
-   "nbconvert_exporter": "python",
-   "pygments_lexer": "ipython3",
-   "version": "3.8.5"
-  }
- },
+ "metadata": {},
  "nbformat": 4,
  "nbformat_minor": 4
 }
diff --git a/doc/src/week41/week41.do.txt b/doc/src/week41/week41.do.txt
index 6d3f70ab0..0f6eed14c 100644
--- a/doc/src/week41/week41.do.txt
+++ b/doc/src/week41/week41.do.txt
@@ -1,4 +1,4 @@
-TITLE: Week 41 Tensor flow and Deep Learning, Convolutional Neural Networks
+TITLE: Week 41 Constructing a Neural Network code, Tensor flow and start Convolutional Neural Networks
 AUTHOR: Morten Hjorth-Jensen {copyright, 1999-present|CC BY-NC} at Department of Physics, University of Oslo & Department of Physics and Astronomy and National Superconducting Cyclotron Laboratory, Michigan State University
 DATE: today
 
@@ -6,12 +6,12 @@ DATE: today
 !split
 ===== Plan for week 41 =====
 
-* Thursday: Building our own Feed-forward Neural Network. "Video of Lecture":"https://www.uio.no/studier/emner/matnat/fys/FYS-STK4155/h20/forelesningsvideoer/LectureOctober8.mp4?vrtx=view-as-webpage"
-* Friday: Playing around with our own Feed-forward Neural Network and introduction to TensorFlow. Start convolutional Neural Networks (CNN). "Video of Lecture":"https://www.uio.no/studier/emner/matnat/fys/FYS-STK4155/h20/forelesningsvideoer/LectureOcotber9.mp4?vrtx=view-as-webpage"
+* Thursday: Building our own Feed-forward Neural Network. 
+* Friday: Playing around with our own Feed-forward Neural Network and introduction to TensorFlow. Start convolutional Neural Networks (CNN). 
 
 Reading suggestions for both days: "Aurelien Geron's chapters 10-11":"https://github.com/CompPhysics/MachineLearning/blob/master/doc/T\
-extbooks/TensorflowML.pdf" and Hastie et al chapter 11.
-
+extbooks/TensorflowML.pdf".
+For amore in depth discussion on  neural networks we recommend Goodfellow et al chapters 6 and 7. For CNNs, see Goodfellow et al chapter 9. chapter 11 and 12 on practicalities and applications
 
 !split
 ===== Setting up the Back propagation algorithm =====
