From a32635c40089e1d82de3709cbc7e04a7050308ac Mon Sep 17 00:00:00 2001 From: Soubeer Koley Date: Mon, 27 May 2024 18:48:59 +0530 Subject: [PATCH 01/37] first --- contrib/machine-learning/Random_Forest.md | 175 ++++++++++++++++++++++ 1 file changed, 175 insertions(+) create mode 100644 contrib/machine-learning/Random_Forest.md diff --git a/contrib/machine-learning/Random_Forest.md b/contrib/machine-learning/Random_Forest.md new file mode 100644 index 0000000..3506a34 --- /dev/null +++ b/contrib/machine-learning/Random_Forest.md @@ -0,0 +1,175 @@ +Random Forest +Random Forest is a versatile machine learning algorithm capable of performing both regression and classification tasks. It is an ensemble method that operates by constructing a multitude of decision trees during training and outputting the average prediction of the individual trees (for regression) or the mode of the classes (for classification). + +Table of Contents +Introduction +How Random Forest Works +Advantages and Disadvantages +Hyperparameters +Code Examples +Classification Example +Feature Importance +Hyperparameter Tuning +Regression Example +Conclusion +References +Introduction +Random Forest is an ensemble learning method used for classification and regression tasks. It is built from multiple decision trees and combines their outputs to improve the model's accuracy and control over-fitting. + +How Random Forest Works +Bootstrap Sampling: +Random subsets of the training dataset are created with replacement. Each subset is used to train an individual tree. +Decision Trees: +Multiple decision trees are trained on these subsets. +Feature Selection: +At each split in the decision tree, a random selection of features is chosen. This randomness helps create diverse trees. +Voting/Averaging: +For classification, the mode of the classes predicted by individual trees is taken (majority vote). +For regression, the average of the outputs of the individual trees is taken. +Detailed Working Mechanism +Step 1: Bootstrap Sampling: Each tree is trained on a random sample of the original data, drawn with replacement (bootstrap sample). This means some data points may appear multiple times in a sample while others may not appear at all. +Step 2: Tree Construction: Each node in the tree is split using the best split among a random subset of the features. This process adds an additional layer of randomness, contributing to the robustness of the model. +Step 3: Aggregation: For classification tasks, the final prediction is based on the majority vote from all the trees. For regression tasks, the final prediction is the average of all the tree predictions. +Advantages and Disadvantages +Advantages +Robustness: Reduces overfitting and generalizes well due to the law of large numbers. +Accuracy: Often provides high accuracy because of the ensemble method. +Versatility: Can be used for both classification and regression tasks. +Handles Missing Values: Can handle missing data better than many other algorithms. +Feature Importance: Provides estimates of feature importance, which can be valuable for understanding the model. +Disadvantages +Complexity: More complex than individual decision trees, making interpretation difficult. +Computational Cost: Requires more computational resources due to multiple trees. +Training Time: Can be slow to train compared to simpler models, especially with large datasets. +Hyperparameters +Key Hyperparameters +n_estimators: The number of trees in the forest. +max_features: The number of features to consider when looking for the best split. +max_depth: The maximum depth of the tree. +min_samples_split: The minimum number of samples required to split an internal node. +min_samples_leaf: The minimum number of samples required to be at a leaf node. +bootstrap: Whether bootstrap samples are used when building trees. If False, the whole dataset is used to build each tree. +Tuning Hyperparameters +Hyperparameter tuning can significantly improve the performance of a Random Forest model. Common techniques include Grid Search and Random Search. + +Code Examples +Classification Example +Below is a simple example of using Random Forest for a classification task with the Iris dataset. + +python +Copy code +import numpy as np +import pandas as pd +from sklearn.datasets import load_iris +from sklearn.ensemble import RandomForestClassifier +from sklearn.model_selection import train_test_split +from sklearn.metrics import accuracy_score, classification_report + +# Load dataset +iris = load_iris() +X, y = iris.data, iris.target + +# Split dataset +X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42) + +# Initialize Random Forest model +clf = RandomForestClassifier(n_estimators=100, random_state=42) + +# Train the model +clf.fit(X_train, y_train) + +# Make predictions +y_pred = clf.predict(X_test) + +# Evaluate the model +accuracy = accuracy_score(y_test, y_pred) +print(f"Accuracy: {accuracy * 100:.2f}%") +print("Classification Report:\n", classification_report(y_test, y_pred)) +Feature Importance +Random Forest provides a way to measure the importance of each feature in making predictions. + +python +Copy code +import matplotlib.pyplot as plt + +# Get feature importances +importances = clf.feature_importances_ +indices = np.argsort(importances)[::-1] + +# Print feature ranking +print("Feature ranking:") +for f in range(X.shape[1]): + print(f"{f + 1}. Feature {indices[f]} ({importances[indices[f]]})") + +# Plot the feature importances +plt.figure() +plt.title("Feature importances") +plt.bar(range(X.shape[1]), importances[indices], align='center') +plt.xticks(range(X.shape[1]), indices) +plt.xlim([-1, X.shape[1]]) +plt.show() +Hyperparameter Tuning +Using Grid Search for hyperparameter tuning. + +python +Copy code +from sklearn.model_selection import GridSearchCV + +# Define the parameter grid +param_grid = { + 'n_estimators': [100, 200, 300], + 'max_features': ['auto', 'sqrt', 'log2'], + 'max_depth': [4, 6, 8, 10, 12], + 'criterion': ['gini', 'entropy'] +} + +# Initialize the Grid Search model +grid_search = GridSearchCV(estimator=clf, param_grid=param_grid, cv=3, n_jobs=-1, verbose=2) + +# Fit the model +grid_search.fit(X_train, y_train) + +# Print the best parameters +print("Best parameters found: ", grid_search.best_params_) +Regression Example +Below is a simple example of using Random Forest for a regression task with the Boston housing dataset. + +python +Copy code +import numpy as np +import pandas as pd +from sklearn.datasets import load_boston +from sklearn.ensemble import RandomForestRegressor +from sklearn.model_selection import train_test_split +from sklearn.metrics import mean_squared_error, r2_score + +# Load dataset +boston = load_boston() +X, y = boston.data, boston.target + +# Split dataset +X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42) + +# Initialize Random Forest model +regr = RandomForestRegressor(n_estimators=100, random_state=42) + +# Train the model +regr.fit(X_train, y_train) + +# Make predictions +y_pred = regr.predict(X_test) + +# Evaluate the model +mse = mean_squared_error(y_test, y_pred) +r2 = r2_score(y_test, y_pred) +print(f"Mean Squared Error: {mse:.2f}") +print(f"R^2 Score: {r2:.2f}") +Conclusion +Random Forest is a powerful and flexible machine learning algorithm that can handle both classification and regression tasks. Its ability to create an ensemble of decision trees leads to robust and accurate models. However, it is important to be mindful of the computational cost associated with training multiple trees. + +References +Scikit-learn Random Forest Documentation +Wikipedia: Random Forest +Machine Learning Mastery: Introduction to Random Forest +Kaggle: Random Forest Guide +Towards Data Science: Understanding Random Forests \ No newline at end of file From 9e1a20de440dc20bd465e20fe2b0709241ac3595 Mon Sep 17 00:00:00 2001 From: Soubeer Koley Date: Mon, 27 May 2024 19:19:52 +0530 Subject: [PATCH 02/37] second --- contrib/machine-learning/Random_Forest.md | 97 ++++++++++++----------- 1 file changed, 52 insertions(+), 45 deletions(-) diff --git a/contrib/machine-learning/Random_Forest.md b/contrib/machine-learning/Random_Forest.md index 3506a34..8c62255 100644 --- a/contrib/machine-learning/Random_Forest.md +++ b/contrib/machine-learning/Random_Forest.md @@ -1,7 +1,8 @@ -Random Forest +# Random Forest + Random Forest is a versatile machine learning algorithm capable of performing both regression and classification tasks. It is an ensemble method that operates by constructing a multitude of decision trees during training and outputting the average prediction of the individual trees (for regression) or the mode of the classes (for classification). -Table of Contents +## Table of Contents Introduction How Random Forest Works Advantages and Disadvantages @@ -13,51 +14,54 @@ Hyperparameter Tuning Regression Example Conclusion References -Introduction + +## Introduction Random Forest is an ensemble learning method used for classification and regression tasks. It is built from multiple decision trees and combines their outputs to improve the model's accuracy and control over-fitting. -How Random Forest Works -Bootstrap Sampling: -Random subsets of the training dataset are created with replacement. Each subset is used to train an individual tree. -Decision Trees: -Multiple decision trees are trained on these subsets. -Feature Selection: -At each split in the decision tree, a random selection of features is chosen. This randomness helps create diverse trees. -Voting/Averaging: +## How Random Forest Works +### 1. Bootstrap Sampling: +* Random subsets of the training dataset are created with replacement. Each subset is used to train an individual tree. +### 2. Decision Trees: +* Multiple decision trees are trained on these subsets. +### 3. Feature Selection: +* At each split in the decision tree, a random selection of features is chosen. This randomness helps create diverse trees. +### 4. Voting/Averaging: For classification, the mode of the classes predicted by individual trees is taken (majority vote). For regression, the average of the outputs of the individual trees is taken. -Detailed Working Mechanism -Step 1: Bootstrap Sampling: Each tree is trained on a random sample of the original data, drawn with replacement (bootstrap sample). This means some data points may appear multiple times in a sample while others may not appear at all. -Step 2: Tree Construction: Each node in the tree is split using the best split among a random subset of the features. This process adds an additional layer of randomness, contributing to the robustness of the model. -Step 3: Aggregation: For classification tasks, the final prediction is based on the majority vote from all the trees. For regression tasks, the final prediction is the average of all the tree predictions. -Advantages and Disadvantages -Advantages -Robustness: Reduces overfitting and generalizes well due to the law of large numbers. -Accuracy: Often provides high accuracy because of the ensemble method. -Versatility: Can be used for both classification and regression tasks. -Handles Missing Values: Can handle missing data better than many other algorithms. -Feature Importance: Provides estimates of feature importance, which can be valuable for understanding the model. -Disadvantages -Complexity: More complex than individual decision trees, making interpretation difficult. -Computational Cost: Requires more computational resources due to multiple trees. -Training Time: Can be slow to train compared to simpler models, especially with large datasets. -Hyperparameters -Key Hyperparameters -n_estimators: The number of trees in the forest. -max_features: The number of features to consider when looking for the best split. -max_depth: The maximum depth of the tree. -min_samples_split: The minimum number of samples required to split an internal node. -min_samples_leaf: The minimum number of samples required to be at a leaf node. -bootstrap: Whether bootstrap samples are used when building trees. If False, the whole dataset is used to build each tree. -Tuning Hyperparameters +### Detailed Working Mechanism +* #### Step 1: Bootstrap Sampling: + Each tree is trained on a random sample of the original data, drawn with replacement (bootstrap sample). This means some data points may appear multiple times in a sample while others may not appear at all. +* #### Step 2: Tree Construction: + Each node in the tree is split using the best split among a random subset of the features. This process adds an additional layer of randomness, contributing to the robustness of the model. +#### Step 3: Aggregation: + For classification tasks, the final prediction is based on the majority vote from all the trees. For regression tasks, the final prediction is the average of all the tree predictions. +### Advantages and Disadvantages +#### Advantages +* Robustness: Reduces overfitting and generalizes well due to the law of large numbers. +* Accuracy: Often provides high accuracy because of the ensemble method. +* Versatility: Can be used for both classification and regression tasks. +* Handles Missing Values: Can handle missing data better than many other algorithms. +* Feature Importance: Provides estimates of feature importance, which can be valuable for understanding the model. +#### Disadvantages +* Complexity: More complex than individual decision trees, making interpretation difficult. +* Computational Cost: Requires more computational resources due to multiple trees. +* Training Time: Can be slow to train compared to simpler models, especially with large datasets. +### Hyperparameters +#### Key Hyperparameters +* n_estimators: The number of trees in the forest. +* max_features: The number of features to consider when looking for the best split. +* max_depth: The maximum depth of the tree. +* min_samples_split: The minimum number of samples required to split an internal node. +* min_samples_leaf: The minimum number of samples required to be at a leaf node. +* bootstrap: Whether bootstrap samples are used when building trees. If False, the whole dataset is used to build each tree. +##### Tuning Hyperparameters Hyperparameter tuning can significantly improve the performance of a Random Forest model. Common techniques include Grid Search and Random Search. -Code Examples -Classification Example +### Code Examples +#### Classification Example Below is a simple example of using Random Forest for a classification task with the Iris dataset. -python -Copy code +''' import numpy as np import pandas as pd from sklearn.datasets import load_iris @@ -85,11 +89,14 @@ y_pred = clf.predict(X_test) accuracy = accuracy_score(y_test, y_pred) print(f"Accuracy: {accuracy * 100:.2f}%") print("Classification Report:\n", classification_report(y_test, y_pred)) -Feature Importance + +''' + +#### Feature Importance Random Forest provides a way to measure the importance of each feature in making predictions. -python -Copy code + +''' import matplotlib.pyplot as plt # Get feature importances @@ -108,11 +115,11 @@ plt.bar(range(X.shape[1]), importances[indices], align='center') plt.xticks(range(X.shape[1]), indices) plt.xlim([-1, X.shape[1]]) plt.show() -Hyperparameter Tuning +''' +#### Hyperparameter Tuning Using Grid Search for hyperparameter tuning. -python -Copy code +''' from sklearn.model_selection import GridSearchCV # Define the parameter grid From d138b0e6252f0cf28be7e014f8625352c4e90dd1 Mon Sep 17 00:00:00 2001 From: Soubeer Koley Date: Mon, 27 May 2024 19:23:51 +0530 Subject: [PATCH 03/37] third --- contrib/machine-learning/Random_Forest.md | 22 ++++++++++++---------- 1 file changed, 12 insertions(+), 10 deletions(-) diff --git a/contrib/machine-learning/Random_Forest.md b/contrib/machine-learning/Random_Forest.md index 8c62255..a3672d0 100644 --- a/contrib/machine-learning/Random_Forest.md +++ b/contrib/machine-learning/Random_Forest.md @@ -61,7 +61,7 @@ Hyperparameter tuning can significantly improve the performance of a Random Fore #### Classification Example Below is a simple example of using Random Forest for a classification task with the Iris dataset. -''' +``` import numpy as np import pandas as pd from sklearn.datasets import load_iris @@ -69,6 +69,7 @@ from sklearn.ensemble import RandomForestClassifier from sklearn.model_selection import train_test_split from sklearn.metrics import accuracy_score, classification_report + # Load dataset iris = load_iris() X, y = iris.data, iris.target @@ -90,13 +91,13 @@ accuracy = accuracy_score(y_test, y_pred) print(f"Accuracy: {accuracy * 100:.2f}%") print("Classification Report:\n", classification_report(y_test, y_pred)) -''' +``` #### Feature Importance Random Forest provides a way to measure the importance of each feature in making predictions. -''' +``` import matplotlib.pyplot as plt # Get feature importances @@ -115,11 +116,11 @@ plt.bar(range(X.shape[1]), importances[indices], align='center') plt.xticks(range(X.shape[1]), indices) plt.xlim([-1, X.shape[1]]) plt.show() -''' +``` #### Hyperparameter Tuning Using Grid Search for hyperparameter tuning. -''' +``` from sklearn.model_selection import GridSearchCV # Define the parameter grid @@ -138,11 +139,11 @@ grid_search.fit(X_train, y_train) # Print the best parameters print("Best parameters found: ", grid_search.best_params_) -Regression Example +``` +#### Regression Example Below is a simple example of using Random Forest for a regression task with the Boston housing dataset. -python -Copy code +``` import numpy as np import pandas as pd from sklearn.datasets import load_boston @@ -171,10 +172,11 @@ mse = mean_squared_error(y_test, y_pred) r2 = r2_score(y_test, y_pred) print(f"Mean Squared Error: {mse:.2f}") print(f"R^2 Score: {r2:.2f}") -Conclusion +``` +### Conclusion Random Forest is a powerful and flexible machine learning algorithm that can handle both classification and regression tasks. Its ability to create an ensemble of decision trees leads to robust and accurate models. However, it is important to be mindful of the computational cost associated with training multiple trees. -References +### References Scikit-learn Random Forest Documentation Wikipedia: Random Forest Machine Learning Mastery: Introduction to Random Forest From 834aade79acb60eb6cfc8dae1731ac93f794c2b6 Mon Sep 17 00:00:00 2001 From: Soubeer Koley Date: Mon, 27 May 2024 19:41:22 +0530 Subject: [PATCH 04/37] third --- contrib/machine-learning/Random_Forest.md | 36 +++++++++++++++-------- 1 file changed, 24 insertions(+), 12 deletions(-) diff --git a/contrib/machine-learning/Random_Forest.md b/contrib/machine-learning/Random_Forest.md index a3672d0..d2d7f5c 100644 --- a/contrib/machine-learning/Random_Forest.md +++ b/contrib/machine-learning/Random_Forest.md @@ -2,18 +2,30 @@ Random Forest is a versatile machine learning algorithm capable of performing both regression and classification tasks. It is an ensemble method that operates by constructing a multitude of decision trees during training and outputting the average prediction of the individual trees (for regression) or the mode of the classes (for classification). -## Table of Contents -Introduction -How Random Forest Works -Advantages and Disadvantages -Hyperparameters -Code Examples -Classification Example -Feature Importance -Hyperparameter Tuning -Regression Example -Conclusion -References + +- [Random Forest](#random-forest) + - [Introduction](#introduction) + - [How Random Forest Works](#how-random-forest-works) + - [1. Bootstrap Sampling:](#1-bootstrap-sampling) + - [2. Decision Trees:](#2-decision-trees) + - [3. Feature Selection:](#3-feature-selection) + - [4. Voting/Averaging:](#4-votingaveraging) + - [Detailed Working Mechanism](#detailed-working-mechanism) + - [Step 3: Aggregation:](#step-3-aggregation) + - [Advantages and Disadvantages](#advantages-and-disadvantages) + - [Advantages](#advantages) + - [Disadvantages](#disadvantages) + - [Hyperparameters](#hyperparameters) + - [Key Hyperparameters](#key-hyperparameters) + - [Tuning Hyperparameters](#tuning-hyperparameters) + - [Code Examples](#code-examples) + - [Classification Example](#classification-example) + - [Feature Importance](#feature-importance) + - [Hyperparameter Tuning](#hyperparameter-tuning) + - [Regression Example](#regression-example) + - [Conclusion](#conclusion) + - [References](#references) + ## Introduction Random Forest is an ensemble learning method used for classification and regression tasks. It is built from multiple decision trees and combines their outputs to improve the model's accuracy and control over-fitting. From 0185122fb9ae4aada0feff6cd45f1e688644ad1c Mon Sep 17 00:00:00 2001 From: Soubeer Koley Date: Mon, 27 May 2024 19:48:33 +0530 Subject: [PATCH 05/37] fourth --- .vscode/settings.json | 2 ++ 1 file changed, 2 insertions(+) create mode 100644 .vscode/settings.json diff --git a/.vscode/settings.json b/.vscode/settings.json new file mode 100644 index 0000000..7a73a41 --- /dev/null +++ b/.vscode/settings.json @@ -0,0 +1,2 @@ +{ +} \ No newline at end of file From 91a2b6d4b81bc46098d6c5007a45b0d0df3b572e Mon Sep 17 00:00:00 2001 From: Soubeer Koley Date: Mon, 27 May 2024 21:22:06 +0530 Subject: [PATCH 06/37] Added Random Forest --- contrib/machine-learning/Random_Forest.md | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/contrib/machine-learning/Random_Forest.md b/contrib/machine-learning/Random_Forest.md index d2d7f5c..59c44ef 100644 --- a/contrib/machine-learning/Random_Forest.md +++ b/contrib/machine-learning/Random_Forest.md @@ -23,8 +23,8 @@ Random Forest is a versatile machine learning algorithm capable of performing bo - [Feature Importance](#feature-importance) - [Hyperparameter Tuning](#hyperparameter-tuning) - [Regression Example](#regression-example) - - [Conclusion](#conclusion) - - [References](#references) + - [Conclusion](#conclusion) + - [References](#references) ## Introduction @@ -185,10 +185,10 @@ r2 = r2_score(y_test, y_pred) print(f"Mean Squared Error: {mse:.2f}") print(f"R^2 Score: {r2:.2f}") ``` -### Conclusion +## Conclusion Random Forest is a powerful and flexible machine learning algorithm that can handle both classification and regression tasks. Its ability to create an ensemble of decision trees leads to robust and accurate models. However, it is important to be mindful of the computational cost associated with training multiple trees. -### References +## References Scikit-learn Random Forest Documentation Wikipedia: Random Forest Machine Learning Mastery: Introduction to Random Forest From 3ba006417ca0b66db7c42647398a94b05c4bf99c Mon Sep 17 00:00:00 2001 From: iABn0rma1 Date: Thu, 30 May 2024 17:30:59 +0530 Subject: [PATCH 07/37] understanding cnn from scratch --- contrib/machine-learning/IntroToCNNs.md | 453 ++++++++++++++++++++++++ contrib/machine-learning/index.md | 1 + 2 files changed, 454 insertions(+) create mode 100644 contrib/machine-learning/IntroToCNNs.md diff --git a/contrib/machine-learning/IntroToCNNs.md b/contrib/machine-learning/IntroToCNNs.md new file mode 100644 index 0000000..5f81ace --- /dev/null +++ b/contrib/machine-learning/IntroToCNNs.md @@ -0,0 +1,453 @@ +# Understanding Convolutional Neural Networks (CNN) + +## Table of Contents +
+Click to expand + +- [Introduction](#introduction) +- [CNN Architecture](#cnn-architecture) + -
+ Convolutional Layer + + - [Input Shape](#input-shape) + - [Stride](#strides) + - [Padding](#padding) + - [Filter](#filters) + - [Output](#output) + +
+ +- [Pooling Layer](#pooling-layer) +- [Flatten Layer](#flatten-layer) +- [Dropout Layer](#dropout-layer) + +- [Implementation](#implementation) + +
+ +## Introduction + +Convolutional Neural Networks (CNNs) are a specialized type of artificial neural network designed primarily for processing structured grid data like images. CNNs are particularly powerful for tasks involving image recognition, classification, and computer vision. They have revolutionized these fields, outperforming traditional neural networks by leveraging their unique architecture to capture spatial hierarchies in images. + +### Why CNNs are Superior to Traditional Neural Networks + +1. **Localized Receptive Fields**: + - CNNs use convolutional layers that apply filters to local regions of the input image. This localized connectivity ensures that the network learns spatial hierarchies and patterns, such as edges and textures, which are essential for image recognition tasks. + +2. **Parameter Sharing**: + - In CNNs, the same filter (set of weights) is used across different parts of the input, significantly reducing the number of parameters compared to fully connected layers in traditional neural networks. This not only lowers the computational cost but also mitigates the risk of overfitting. + +3. **Translation Invariance**: + - Due to the shared weights and pooling operations, CNNs are inherently invariant to translations of the input image. This means that they can recognize objects even when they appear in different locations within the image. + +4. **Hierarchical Feature Learning**: + - CNNs automatically learn a hierarchy of features from low-level features like edges to high-level features like shapes and objects. Traditional neural networks, on the other hand, require manual feature extraction which is less effective and more time-consuming. + +### Use Cases of CNNs + +- **Image Classification**: Identifying objects within an image (e.g., classifying a picture as containing a cat or a dog). +- **Object Detection**: Detecting and locating objects within an image (e.g., finding faces in a photo). +- **Image Segmentation**: Partitioning an image into segments or regions (e.g., dividing an image into different objects and background). +- **Medical Imaging**: Analyzing medical scans like MRI, CT, and X-rays for diagnosis. + +> This guide will walk you through the fundamentals of CNNs and their implementation in Python. We'll build a simple CNN from scratch, explaining each component to help you understand how CNNs process images and extract features. + +### Let's start by understanding the basic architecture of CNNs. + +## CNN Architecture + +Convolution layers, pooling layers, and fully connected layers are just a few of the many building blocks that CNNs use to automatically and adaptively learn spatial hierarchies of information through backpropagation. + +### Convolutional Layer +The convolutional layer is the core building block of a CNN. The layer's parameters consist of a set of learnable filters (or kernels), which have a small receptive field but extend through the full depth of the input volume. + +#### Input Shape +The dimensions of the input image, including the number of channels (e.g., 3 for RGB images & 1 for Grayscale images). +
+ + + + + + + + + + + +
1 and 0
10111
10101
10101
10101
10101
10101
10111
+ + + + + + + + + + + +
9
01110
01010
01110
00010
00010
00010
00010
+
+ +- The input matrix represents a simplified binary image of handwritten digits, +where '1' denotes the presence of ink and '0' represents the absence of ink. + +- The first matrix shows the represnetation of 1 and 0, which can be depicted as a vertical line and a closed loop. + +- The second matrix represents 9, combining the loop and line. + +
+ +#### Strides +The step size with which the filter moves across the input image. +
+ + + + + + + + + + + +
3
011 + 10
010 + 10
01110
00010
000 + 10
000 + 10
00010
+ + + + + + + + + + + +
2
01 + 110
01 + 010
01110
00 + 010
00 + 010
00010
00010
+
+ +- This visualization will help you understand how the filter (kernel) moves acroos the input matrix with stride values of 3 and 2. + +- A stride of 1 means the filter moves one step at a time, ensuring it covers the entire input matrix. + +- However, with larger strides (like 3 or 2 in this example), the filter may not cover all elements, potentially missing some information. + +- While this might seem like a drawback, higher strides are often used to reduce computational cost and decrease the output size, which can be beneficial in speeding up the training process and preventing overfitting. + +
+ +#### Padding +Determines whether the output size is the same as the input size ('same') or reduced ('valid'). +
+ + + + + + + + + + + + + +
padding='same'
0000000
0011100
0010100
0011100
0000100
0000100
0000100
0000100
0000000
+ + + + + + + + + + + +
padding='valid'
01110
01010
01110
00010
00010
00010
00010
+
+ +- `Same` padding is preferred in earlier layers to preserve spatial and edge information, as it can help the network learn more detailed features. + +- Choose `valid` padding when focusing on the central input region or requiring specific output dimensions. + +
+ +#### Filters +Small matrices that slide over the input data to extract features. +
+ + + + + + + +
closed loop
111
101
111
+ + + + + + + +
vertical line
010
010
010
+ + + + + + + +
both diagonals
101
010
101
+
+ +- The first filter aims to detect closed loops within the input image, being highly relevant for recognizing digits with circular or oval shapes, such as '0', '6', '8', or '9'. + +- The next filter helps in detecting vertical lines, crucial for identifying digits like '1', '4', '7', and parts of other digits that contain vertical strokes. + +- The last filter shows how to detect diagonal lines in the input image, useful for identifying the slashes present in digits like '1', '7', or parts of '4' and '9'. + +
+ +#### Output +A set of feature maps that represent the presence of different features in the input. +
+ + + + + + + + + +
('valid', 1)
404
25-3
25-3
032
032
+ + + + + + + + + + + +
('same', 1)
22422
34843
22533
12533
00323
00323
00212
+ + + + + + + +
('valid', 2)
44
23
02
+
+ +- With no padding and a stride of 1, the 3x3 filter moves one step at a time across the 7x5 input matrix. The filter can only move within the original boundaries of the input, resulting in a smaller 5x3 output matrix. This configuration is useful when you want to reduce the spatial dimensions of the feature map while preserving the exact spatial relationships between features. + +- By adding zero padding to the input matrix, it is expanded to 9x7, allowing the 3x3 filter to "fit" fully on the edges and corners. With a stride of 1, the filter still moves one step at a time, but now the output matrix is the same size (7x5) as the original input. Same padding is often preferred in early layers of a CNN to preserve spatial information and avoid rapid feature map shrinkage. + +- Without padding, the 3x3 filter operates within the original input matrix boundaries, but now it moves two steps at a time (stride 2). This significantly reduces the output matrix size to 3x2. Larger strides are employed to decrease computational cost and the output size, which can be beneficial in speeding up the training process and preventing overfitting. However, they might miss some finer details due to the larger jumps. + +
+ +### Pooling Layer +Pooling layers reduce the dimensionality of each feature map while retaining the most critical information. The most common form of pooling is max pooling. +- **Input Shape:** The dimensions of the feature map from the convolutional layer. +- **Pooling Size:** The size of the pooling window (e.g., 2x2). +- **Strides:** The step size for the pooling operation. +- **Output:** A reduced feature map highlighting the most important features. +
+ + + + + + + + + + +
((2,2), 1)
4884
4884
2553
2553
0333
0333
+ + + + + + + +
((3,3) 2)
88
55
33
+
+ +- The high values (8) indicate that the "closed loop" filter found a strong match in those regions. + +- First matrix of size 6x4 represents a downsampled version of the input. + +- While the second matrix with 3x2, resulting in more aggressive downsampling. + + +
+ +### Flatten Layer +The flatten layer converts the 2D matrix data to a 1D vector, which can be fed into a fully connected (dense) layer. +- **Input Shape:** The 2D feature maps from the previous layer. +- **Output:** A 1D vector that represents the same data in a flattened format. +
+ + + + + + + + + + +
After max pooling (with kernel size = 3 and stride = 1)
888888555555333
+
+ +
+ +### Dropout Layer +Dropout is a regularization technique to prevent overfitting in neural networks by randomly setting a fraction of input units to zero at each update during training time. +- **Input Shape:** The data from the previous layer. +- **Dropout Rate:** The fraction of units to drop (e.g., 0.5 for 50% dropout). +- **Output:** The same shape as the input, with some units set to zero. +
+ + + + + +
dropout rate = 0.3
880800 505055 033
+
+ +- The updated 0 values represents the dropped units. + +## Implementation + +Below is the implementation of a simple CNN in Python. Each function within the `CNN` class corresponds to a layer in the network. + +```python +import numpy as np + +class CNN: + def __init__(self): + pass + + def convLayer(self, input_shape, channels, strides, padding, filter_size): + height, width = input_shape + input_shape_with_channels = (height, width, channels) + print("Input Shape (with channels):", input_shape_with_channels) + + # Generate random input and filter matrices + input_matrix = np.random.randint(0, 10, size=input_shape_with_channels) + filter_matrix = np.random.randint(0, 5, size=(filter_size[0], filter_size[1], channels)) + + print("\nInput Matrix:\n", input_matrix[:, :, 0]) + print("\nFilter Matrix:\n", filter_matrix[:, :, 0]) + + padding = padding.lower() + + if padding == 'same': + # Calculate padding needed for each dimension + pad_height = filter_size[0] // 2 + pad_width = filter_size[1] // 2 + + # Apply padding to the input matrix + input_matrix = np.pad(input_matrix, ((pad_height, pad_height), (pad_width, pad_width), (0, 0)), mode='constant') + + # Adjust height and width to consider the padding + height += 2 * pad_height + width += 2 * pad_width + + elif padding == 'valid': + pass + + else: + return "Invalid Padding!!" + + # Output dimensions + conv_height = (height - filter_size[0]) // strides[0] + 1 + conv_width = (width - filter_size[1]) // strides[1] + 1 + + output_matrix = np.zeros((conv_height, conv_width, channels)) + + # Convolution Operation + for i in range(0, height - filter_size[0] + 1, strides[0]): + for j in range(0, width - filter_size[1] + 1, strides[1]): + receptive_field = input_matrix[i:i + filter_size[0], j:j + filter_size[1], :] + output_matrix[i // strides[0], j // strides[1], :] = np.sum(receptive_field * filter_matrix, axis=(0, 1, 2)) + + return output_matrix + + def maxPooling(self, input_matrix, pool_size=(2, 2), strides_pooling=(2, 2)): + input_height, input_width, input_channels = input_matrix.shape + pool_height, pool_width = pool_size + stride_height, stride_width = strides_pooling + + # Calculate output dimensions + pooled_height = (input_height - pool_height) // stride_height + 1 + pooled_width = (input_width - pool_width) // stride_width + 1 + + # Initialize output + pooled_matrix = np.zeros((pooled_height, pooled_width, input_channels)) + + # Perform max pooling + for c in range(input_channels): + for i in range(0, input_height - pool_height + 1, stride_height): + for j in range(0, input_width - pool_width + 1, stride_width): + patch = input_matrix[i:i + pool_height, j:j + pool_width, c] + pooled_matrix[i // stride_height, j // stride_width, c] = np.max(patch) + + return pooled_matrix + + def flatten(self, input_matrix): + return input_matrix.flatten() + + def dropout(self, input_matrix, dropout_rate=0.5): + assert 0 <= dropout_rate < 1, "Dropout rate must be in [0, 1)." + dropout_mask = np.random.binomial(1, 1 - dropout_rate, size=input_matrix.shape) + return input_matrix * dropout_mask +``` + +Run the below command to generate output, based on random input and filter matrices. + +```python +input_shape = (5, 5) +channels = 1 +strides = (1, 1) +padding = 'valid' +filter_size = (3, 3) + +cnn_model = CNN() + +conv_output = cnn_model.convLayer(input_shape, channels, strides, padding, filter_size) +print("\nConvolution Output:\n", conv_output[:, :, 0]) + +pool_size = (2, 2) +strides_pooling = (1, 1) + +maxPool_output = cnn_model.maxPooling(conv_output, pool_size, strides_pooling) +print("\nMax Pooling Output:\n", maxPool_output[:, :, 0]) + +flattened_output = cnn_model.flatten(maxPool_output) +print("\nFlattened Output:\n", flattened_output) + +dropout_output = cnn_model.dropout(flattened_output, dropout_rate=0.3) +print("\nDropout Output:\n", dropout_output) +``` + +Feel free to play around with the parameters! \ No newline at end of file diff --git a/contrib/machine-learning/index.md b/contrib/machine-learning/index.md index 94ca1e2..ea8b5cb 100644 --- a/contrib/machine-learning/index.md +++ b/contrib/machine-learning/index.md @@ -9,3 +9,4 @@ - [TensorFlow.md](tensorFlow.md) - [PyTorch.md](pytorch.md) - [Types of optimizers](Types_of_optimizers.md) +- [Introduction To Convolutional Neural Networks (CNNs)](IntroToCNNs.md) From f89ffbd58c0082f8533064aa864483799a9c1751 Mon Sep 17 00:00:00 2001 From: Aman Barthwal Date: Thu, 30 May 2024 17:41:16 +0530 Subject: [PATCH 08/37] Update IntroToCNNs.md --- contrib/machine-learning/IntroToCNNs.md | 8 ++++++-- 1 file changed, 6 insertions(+), 2 deletions(-) diff --git a/contrib/machine-learning/IntroToCNNs.md b/contrib/machine-learning/IntroToCNNs.md index 5f81ace..7c0812f 100644 --- a/contrib/machine-learning/IntroToCNNs.md +++ b/contrib/machine-learning/IntroToCNNs.md @@ -63,7 +63,8 @@ The convolutional layer is the core building block of a CNN. The layer's paramet #### Input Shape The dimensions of the input image, including the number of channels (e.g., 3 for RGB images & 1 for Grayscale images). -
+
+

@@ -76,6 +77,8 @@ The dimensions of the input image, including the number of channels (e.g., 3 for
1 and 0
10111
+

+

@@ -88,6 +91,7 @@ The dimensions of the input image, including the number of channels (e.g., 3 for
9
00010
+

- The input matrix represents a simplified binary image of handwritten digits, @@ -450,4 +454,4 @@ dropout_output = cnn_model.dropout(flattened_output, dropout_rate=0.3) print("\nDropout Output:\n", dropout_output) ``` -Feel free to play around with the parameters! \ No newline at end of file +Feel free to play around with the parameters! From d9ab12136329bf485af34a2dbf31fa8e933a9ebe Mon Sep 17 00:00:00 2001 From: Aman Barthwal Date: Thu, 30 May 2024 17:42:51 +0530 Subject: [PATCH 09/37] Update IntroToCNNs.md --- contrib/machine-learning/IntroToCNNs.md | 16 ++++++++++++++++ 1 file changed, 16 insertions(+) diff --git a/contrib/machine-learning/IntroToCNNs.md b/contrib/machine-learning/IntroToCNNs.md index 7c0812f..802cfdb 100644 --- a/contrib/machine-learning/IntroToCNNs.md +++ b/contrib/machine-learning/IntroToCNNs.md @@ -25,6 +25,22 @@ + + +
Table 1 Heading 1 Table 1 Heading 2
+ +|Table 1| Middle | Table 2| +|--|--|--| +|a| not b|and c | + + + +|b|1|2|3| +|--|--|--|--| +|a|s|d|f| + +
+ ## Introduction Convolutional Neural Networks (CNNs) are a specialized type of artificial neural network designed primarily for processing structured grid data like images. CNNs are particularly powerful for tasks involving image recognition, classification, and computer vision. They have revolutionized these fields, outperforming traditional neural networks by leveraging their unique architecture to capture spatial hierarchies in images. From 1d01704f40d48be974079d84a7511d281caab497 Mon Sep 17 00:00:00 2001 From: iABn0rma1 Date: Thu, 30 May 2024 23:59:27 +0530 Subject: [PATCH 10/37] replaced tables with images --- contrib/machine-learning/IntroToCNNs.md | 230 ++---------------- .../machine-learning/assets/cnn-dropout.png | Bin 0 -> 28288 bytes .../machine-learning/assets/cnn-filters.png | Bin 0 -> 38260 bytes .../machine-learning/assets/cnn-flattened.png | Bin 0 -> 33305 bytes .../assets/cnn-input_shape.png | Bin 0 -> 54890 bytes .../machine-learning/assets/cnn-ouputs.png | Bin 0 -> 94050 bytes .../machine-learning/assets/cnn-padding.png | Bin 0 -> 98306 bytes .../machine-learning/assets/cnn-pooling.png | Bin 0 -> 79460 bytes .../machine-learning/assets/cnn-strides.png | Bin 0 -> 75501 bytes 9 files changed, 19 insertions(+), 211 deletions(-) create mode 100644 contrib/machine-learning/assets/cnn-dropout.png create mode 100644 contrib/machine-learning/assets/cnn-filters.png create mode 100644 contrib/machine-learning/assets/cnn-flattened.png create mode 100644 contrib/machine-learning/assets/cnn-input_shape.png create mode 100644 contrib/machine-learning/assets/cnn-ouputs.png create mode 100644 contrib/machine-learning/assets/cnn-padding.png create mode 100644 contrib/machine-learning/assets/cnn-pooling.png create mode 100644 contrib/machine-learning/assets/cnn-strides.png diff --git a/contrib/machine-learning/IntroToCNNs.md b/contrib/machine-learning/IntroToCNNs.md index 802cfdb..f832980 100644 --- a/contrib/machine-learning/IntroToCNNs.md +++ b/contrib/machine-learning/IntroToCNNs.md @@ -25,22 +25,6 @@ - - -
Table 1 Heading 1 Table 1 Heading 2
- -|Table 1| Middle | Table 2| -|--|--|--| -|a| not b|and c | - - - -|b|1|2|3| -|--|--|--|--| -|a|s|d|f| - -
- ## Introduction Convolutional Neural Networks (CNNs) are a specialized type of artificial neural network designed primarily for processing structured grid data like images. CNNs are particularly powerful for tasks involving image recognition, classification, and computer vision. They have revolutionized these fields, outperforming traditional neural networks by leveraging their unique architecture to capture spatial hierarchies in images. @@ -79,35 +63,8 @@ The convolutional layer is the core building block of a CNN. The layer's paramet #### Input Shape The dimensions of the input image, including the number of channels (e.g., 3 for RGB images & 1 for Grayscale images). -
-

- - - - - - - - - - - -
1 and 0
10111
10101
10101
10101
10101
10101
10111
-

-

- - - - - - - - - - - -
9
01110
01010
01110
00010
00010
00010
00010
-

+
+
- The input matrix represents a simplified binary image of handwritten digits, @@ -121,39 +78,8 @@ where '1' denotes the presence of ink and '0' represents the absence of ink. #### Strides The step size with which the filter moves across the input image. -
- - - - - - - - - - - -
3
011 - 10
010 - 10
01110
00010
000 - 10
000 - 10
00010
- - - - - - - - - - - -
2
01 - 110
01 - 010
01110
00 - 010
00 - 010
00010
00010
+
+
- This visualization will help you understand how the filter (kernel) moves acroos the input matrix with stride values of 3 and 2. @@ -168,33 +94,8 @@ The step size with which the filter moves across the input image. #### Padding Determines whether the output size is the same as the input size ('same') or reduced ('valid'). -
- - - - - - - - - - - - - -
padding='same'
0000000
0011100
0010100
0011100
0000100
0000100
0000100
0000100
0000000
- - - - - - - - - - - -
padding='valid'
01110
01010
01110
00010
00010
00010
00010
+
+
- `Same` padding is preferred in earlier layers to preserve spatial and edge information, as it can help the network learn more detailed features. @@ -205,31 +106,8 @@ Determines whether the output size is the same as the input size ('same') or red #### Filters Small matrices that slide over the input data to extract features. -
- - - - - - - -
closed loop
111
101
111
- - - - - - - -
vertical line
010
010
010
- - - - - - - -
both diagonals
101
010
101
+
+
- The first filter aims to detect closed loops within the input image, being highly relevant for recognizing digits with circular or oval shapes, such as '0', '6', '8', or '9'. @@ -242,37 +120,8 @@ Small matrices that slide over the input data to extract features. #### Output A set of feature maps that represent the presence of different features in the input. -
- - - - - - - - - -
('valid', 1)
404
25-3
25-3
032
032
- - - - - - - - - - - -
('same', 1)
22422
34843
22533
12533
00323
00323
00212
- - - - - - - -
('valid', 2)
44
23
02
+
+
- With no padding and a stride of 1, the 3x3 filter moves one step at a time across the 7x5 input matrix. The filter can only move within the original boundaries of the input, resulting in a smaller 5x3 output matrix. This configuration is useful when you want to reduce the spatial dimensions of the feature map while preserving the exact spatial relationships between features. @@ -289,69 +138,29 @@ Pooling layers reduce the dimensionality of each feature map while retaining the - **Pooling Size:** The size of the pooling window (e.g., 2x2). - **Strides:** The step size for the pooling operation. - **Output:** A reduced feature map highlighting the most important features. -
- - - - - - - - - - -
((2,2), 1)
4884
4884
2553
2553
0333
0333
- - - - - - - -
((3,3) 2)
88
55
33
+
+
- The high values (8) indicate that the "closed loop" filter found a strong match in those regions. - - First matrix of size 6x4 represents a downsampled version of the input. - - While the second matrix with 3x2, resulting in more aggressive downsampling. - -
- ### Flatten Layer The flatten layer converts the 2D matrix data to a 1D vector, which can be fed into a fully connected (dense) layer. - **Input Shape:** The 2D feature maps from the previous layer. - **Output:** A 1D vector that represents the same data in a flattened format. -
- - - - - - - - - - -
After max pooling (with kernel size = 3 and stride = 1)
888888555555333
+
+
-
- ### Dropout Layer Dropout is a regularization technique to prevent overfitting in neural networks by randomly setting a fraction of input units to zero at each update during training time. - **Input Shape:** The data from the previous layer. - **Dropout Rate:** The fraction of units to drop (e.g., 0.5 for 50% dropout). - **Output:** The same shape as the input, with some units set to zero. -
- - - - - -
dropout rate = 0.3
880800 505055 033
+
+
- The updated 0 values represents the dropped units. @@ -402,9 +211,8 @@ class CNN: # Output dimensions conv_height = (height - filter_size[0]) // strides[0] + 1 conv_width = (width - filter_size[1]) // strides[1] + 1 - output_matrix = np.zeros((conv_height, conv_width, channels)) - + # Convolution Operation for i in range(0, height - filter_size[0] + 1, strides[0]): for j in range(0, width - filter_size[1] + 1, strides[1]): @@ -443,7 +251,7 @@ class CNN: return input_matrix * dropout_mask ``` -Run the below command to generate output, based on random input and filter matrices. +Run the below command to generate output with random input and filter matrices, depending on the given size. ```python input_shape = (5, 5) @@ -470,4 +278,4 @@ dropout_output = cnn_model.dropout(flattened_output, dropout_rate=0.3) print("\nDropout Output:\n", dropout_output) ``` -Feel free to play around with the parameters! +Feel free to play around with the parameters! \ No newline at end of file diff --git a/contrib/machine-learning/assets/cnn-dropout.png b/contrib/machine-learning/assets/cnn-dropout.png new file mode 100644 index 0000000000000000000000000000000000000000..9cb18f9567de980472ceb8cbdd7aaa1130ba7231 GIT binary patch literal 28288 zcmeFXby!tz+6KDmZUJeeyE|1%q?B$XHr)-IR6;~RK)OUqx|sos)_IlpBpZlqoH{^|?3{(Tu6!ER5|lOY#_s|BycxvK+lvwNrY zxdg}o=$pfP`p9435<$N{on)0hgX!b!W|UVZqy*_bhcodzy!60zb#{jL1An`Nc(}n) z9-rrB=b1bBFXzLS6X@dABxTrhgwC>d^uAf)`1vJGub zKq4gnmw-aZ1%vg9u^@Ie=y~>$`v^K{j9nsCt9j6mcpHi>9Ew222_lyN=N?Rf{}21%31{?K5`$8 z16RI)vc%NCuwjI&O~?IO$LJ%Rjb~hu&O2wJ3}iIMs;8Lj07W397zPdD=}9&L`;vThls21_-# zr-s}XnbbjJGdS$##h~q%wI3_7W`S4H=ZmI?t)p+I-zGCMfRWqNjBZX|xj*Z96wMD~+D=*hQWI#ZNgD;8?YX_Z%&4dw$o{ zgIh4dKQfUZ8y3IC5Yh6WJ4v=msZVKQxH%3~SvxRqPM}Gk!v~wJb59?JQkpond*v9) z6PgiffPO6!uI=n7d|s*d->mf6Sighu8$7zdzi(ZvEm;Nap83r_47m-&B8Zb&f|7;X z;$+djlCm{nFX_C5+vkCM2g~BenjC>m89M22(R)36V8YeY4mQDBLf43B3lH?ZeIzo33ce62d0JL9-eBi&w~GhLngO8I1x3 zDA|zFj-=jyL8c*)l~#^Km;BuIg{Vs;^NE(fDrM3aViR^pWWkVRnVHDhE{RMMk1(&7 zoL_^szjgfJ6(R+vTl0{H$#iBAZJJbIr$n&5=pD9nrq_;0m1O@uIgELVFcS4amAJPhTP$qNc~ck)C=j@JayVLYaPJg88=eyOY)C|iHw|vL->>J#&!i&VGrg2KEm$q+w*%V;HW;Eff{irLw@Ip+w@uiRSz5H;d=Gz8_w@+Ze#9&4+T2_c-$cXHf{P_2oH{L&v z##pv>Ed?vXJrg}6uB8+6&mX=Gym3?B$U_@5%h$-8%r|)JoBLw8WaOQd{1C$! z!{}78gKUOO^mmKG?%`h+10w}vBAFg4KS6jc&pgP zAxGM)DjlM9o^C;2lWRtYhhOc2>lJ%&&iKx>0L7< zoY0JJLz0r%75ixBJt$oige-UItm*+QNe3_bWS>< zY_4^S-H%g5rDm&79q?9N+n}LX9}S3o`G4}gh#gA)5h=h{N_XU~{BGn^W;kbSXCJ8( z=QM}b*SYwYuVv+@6G!;@%y+j-QX@FVIu7ff%*jyC9@Dt0BMZ9UocdFp>7G{3~qGs&CR?{J4-yXPUh%7@-j%27w}*MsZqP%_`5`a);-}A~ai+8J+FcMo270 z9O@A?{kfK#e86OIjt4CvhA-cvs(tsWTzDg(Z(1x)h##_lAT=Px{Apo8E^(4)(fvO8 zD#91#DyvEps;TYe+yABd;yapcn}lWMSQ5d%-Q3 zy!;;gWFouH?z1@M)G;zH5p1H;qI?gU4@WJl-t;u|o{FZ)UUMR)SNpk~IYpIi6%{F% z;@|CrwFY9LykY?v*x7j^(EgN<{YID6+ zZ%UbcG!*qQ8a&sDFJ{_zA}438dfRXf=>)v4HToXiVeAYZ)Rw_0F_ zygU7Nja4u1>Q>`M>izVxvmr<)sE5k{pWX1|}GvUA`LJxM^PCZMN> zUeaXwA`gtH-$lxqurHM~J)~8ixPi8p9w72qxAnCs&8e)1mk4LP9WdAmNR$hB->79U zC;ZD3S9BTKCEYQno}Tiu@?+d4eHU@M$G#Y3{gPkzK5f<)#vIFwqu7kp1*j-Kfi-1K z6%;^Bz%w!k9)=i%06f6}mlzDm|2#{>Fo58G{|*ZR1zUjN|0bgd+#f$*fa|f%Uw62u zAP^Gp7cOwQe}nzIG~&`XxW7Mx=74vg7b+66vcO%%*bxl2buzbec6>b<4E%s%|4PdV z1R|h)ykKONX%2z@CoNPpoi!EYg^cZN*bGhVjKFN}HujJGfJEGdfJYm!vmvFsjkT?l zkh>_=ZwVpb`SE9VD$3s?&Q_vSnhI|yCF~r*lzeO)Y#dZ#sFakHB911eLdufTe`^l> zB}!%P>})T@&hF;s#^%P&X6IAK$QZvT?BgPv1aO zk;h+!-dMPUt+gaAYyixFF~qpII7EKS|J$2?jQGD=YW|}oCqLgmHT_?&{`W6UcBa{jIPeQK{;O+I{ z|L+d|>RS*MdC9SD?8zTS`7H#K?DofL00Pg$VPR`XIeDf2gV1x(2HAgD4CR*$KMELs_q zPxajz=WyGfYZ={~%Qeue*}w&8EM`j7 zfp#c@Ro(6;5qBdpb@a*Sr-DLWf;+uKP@T!4a!B=1txC>N8F;~rSr`iY2SYe4MDyB# znQN0(Y4LlZ0j9h@-N#oHU5=Me2+JnC+z~zq>`ZkRBK2}8l&zGB#r|&Ekdv~XuTAT)(;)9g98!xE}$|s1IKT2SlD^`~@PZC4E&x>8A9xg0C zOuy4vyc%$8@xR%};nQ^6L)_u5I7xdTC$NRC79CETrSYq>Zoy1)?Uwe28z!;xTa~1y z{K1;O38En{C7z601z7p6#_(8* zJd7hWV0+3_F_H()w_nfg47Sy)=TR+OX*|NC3z92p;3JzUGYX7qAz;-GQc)COI@QY7 z=q6A4)9`%o@b;BQq)zpr`Nd7NUWdzv@{oR3znjjc1+2+MFP(Mr`&rnUr(&25w{Z!S zyzko_;4AVq=*9?p^-X%~n!U`sD}a#IK4zI?!8!B+uImU-VLbC>1NQU!}SG{;c0wn zy7!H%;T0G%DIBn{J6#@~sz=@OD?$uYCC4W!0NW;NXOrBM&MB$Vre1SdH%5KML+tLf zyrpRBgL(HaFNXwSH`}K%$q#pi6m&4sq0X96eYuNQa6v6F4~F0uMK-36Hlv|Mtf`jW z=X=rmimbf@LU*G|t$hgw1U)*mCl0K&CO6K)1M~P2Uz2R1)ZXfWue*DOgDIl=U?zE? zO<6)LS`GO>&k($ARieyzUFQNg&Z3oG?n>d$uL|SAeP6pi#oUY41k%k*7O`u0%a13K z^Mb$p`x;}^HN{9P7E>*>X4?rAB!0*T7W1!^lZ$#%r1tOt#V#> zs;Uqa=kjFct=H5E%}1v{t9RYft-hiVU;XmsNLZtRG_}va+Lnt@f8Md3-4>l%I;?0` zUL0%AZLnlZddIAXc;;hgr|YLUfzNA#IVqY9pALQs-mznrZ}hQ!ZI5vp6rgq9n&VV4 z{qTptKKA~329iB{)3tQXMY>G%l9_43q~}fnzU!20KF8+v4)@J#xEg~+L7Im9b)7(= z&?*bOaFN~M<@H$~l17FtkMdQ_*`O5HsiHR*Cxd1I^l3hP4ae6>3D|{io@s3@R9a2h z9n80&_MS1au@%G(W{p$na+Vl;gnrg{U#rA!&0nEKloIy!y(yV@t43%2{!x$y>^3}^#JHyg zs=`WBw_?>e)H1=H4~LzP4k$*qch0jYM@rLW4ZlSUvKtj3kTN0^6O>xueCTOaOoS=y z{?J!+$+F*gkbbzXb3nJYwqdsqo?XH)oB^A=$wJ^=*!-! zooSVu2993qYx3RxYQ8tificyi9=;3_H-U89R99@rD{UjG(zcft!BL;lev%sUS)V zJ>LkcWF?EouYRxkB7pzl+SMVZM3d|D!Uq0&>5T8eeCv9W7w_qO*IVP3hj|B$c&@dK zzS09CdQuSx2ixdrR9$YZrQ1o_0^;yg$ZyyGQYW%74E-d8Wqt zHEsZjpJj{419BYhE#hLCuISCIV_W4g@{FE&JiplbM1c?WhF?<=J9F`=d@obX6CN<=Vt?^I5|dWj3q|f?7iU1=s%GW;;XP7o zE4ytCtqQ@1zLNnFS`|KV63$n5t7)l9X|fjV<5LSb!KCq9wEW^5_UL~|GBq_3h-v*o zAC`SyjLceIpB}$4-s9C!3i8^;m5v}IG<04u&HA}GXIg&B!Vd`I;unnOd!4_6=5aVNl#u@OAQtd{;j|f{=pLtbm zWB-9VLn82xu=-1*^Oebk+HYcACqB7k)@2nG4%8j z(H#qR4mD#dYac0Sq2*C-%xn&PH82%ws?9aeU(h;LL6cL;3s;rp0vmJEG}Rwkm#JITCg)YX-0R3U@OPknBU_$XAhz=- z+ivH%;FET>?YEV;W!^2aon%*PI)Vl|PH0QSp0+p)i2BiV3Mj@MVREqf35KmSzpSUr zz?Os$2^Ntaz(!z&^}Bj=BDqlpbiSWpi4JAD%jcp%niFQQt=) zn?>8erTZ$O>}tB^^4K8A0gp~qq(Uk**02fc%Gx)sSsL~7PdxnmXx{leZ7V?5f;*vU zv&NMdXW&hZPJaA6yCD%|8;voEg|t^~2ILlDy+yUpJ`;;(+j8H{0bye24(ox$wpR4^ zF707pOq^qDx(diAa=?Ihd$`(HH>{@cVt10Cix;qPoXO8~mm$5jZ}u?Dmh$*JXzivB z-UG?ChPyRK&oEhuAecTJUXr-Y3QUEuwg#Ms8-ezfGobzMoHUFG(JSX{=#;Qs5-VT- z*Y>c*ITr)d&EAPZonZF?%eZvArXr~_CVBuY8JVD zM)WgH<9Mxxujh0u3;gUmN}@L#9MnXp&ZQ<|9d1$Srg9SJQxhniq&Vu@hP56R>Uw_y zjOEdAg#q{6`ZYqV2bAp5gOY9*VVh7^b2R9Sg$6_`tKY!Ik-7bFBhM|r!Rk9z{}e`e zVcwg}GW}74qNDknI{;5ou&CQSBWvE1=Dq4kTf=%JtBX}yKIB{(@70$L+4BFb=NfA>! z_PRB_%5IWP*L{H~!4s_xrwOfzT!f~LJh;8L3>xjR6(xwk`lrHX;x{5E4KU!J-L+=P zq;<~ws_XooeN&l)17``xKv2A17YxNnaa;a&*j-_fz;~1Cj1CI-6a>yx^@OADhJp2Qm#M0*(@b)q5zCwGM#;sg(aaieu z!9JfQo)fSGE=Y#d@tvkw;dm$(bAiV8mJ1L^g26W|DnC>jlr0vj$f{=PcFR`JEh=Bb z=#u(yVM=J#^`(jmZ^0053|(g<=_YG1V`@AVKWVH21iV{qQpXRzA6|VXN|3YN1{I}m zoUbq!2iNa=b2xdskm%JrqcF2IS!?>#E-e3)Nktyb@zt(ozH2ZY3Jlz{gf!2)cFr&N zS1&Mpv7AocN*6=QGC+fOYhM;P-Cmm^zaKgZ?u%bX2RBbyVv42<$>B6TFu)Phi zKf-TfVMG1PtqRey0(Qj|7#y?j((#!+D%AsyYZ9M74`peaM69BqH0u4V4@Akm++^x! zP^wNAxkV}cimkA(_+#FYeg;&iiW@V`PXY**6JHQM;bTUycd)Idc-U-ynNp){u{k{N zb9hIiInSlS9uf9>d9bd7;JL0&_Qj3KW11LmdZE=^ANl zM4r4pGAlPKd&fZOL6qvTA867w9WlDBU<`teUR0q8+cw!`hfJ4m&`dh}a1}2m%U@l2 zD`AnvTBk9V|B;UMzv=igT>xnu2wOxeB~C7Kw@Oz)>Yitkx3O<$@CsKCj~RyDO@DUNZ-+h70jLf=uqQl6x2|e$VLCSje|Xvs>QDmQcnY)JSGa z;iZ(#V*3@*esbK(SSZqy3m}43tR&_a6Nd}!1_uH(IAMnp@7F%KC6ukj?%^pa zE{v|MMQ3VEExo5N76%~&rVCvdyW*J7n)9@@U-E2*n+Cb?MomwA_% z8uhx4?UjbRmq!WPS)$a&?6p%mMedB4siD0)I=;(mlP&WXqF<-yf&8xp(1P6hFDo?$ z->XV~mDZOp%5V=hzNe+B>6)bcsca5-REr(-r1~)50E-Q7#=Z3ns10`$2FCH>;uaT9 zo@-&{y1(ItnRAdKBT(0}2rP4IdykVL9n9?GU_9NZFAB*#sT=Q?x|pn9MEejdf^GU0 zy?3g9a%6cdja&s84IpLd(Y{`N5OwsM9J!6~ty-fl?ADJ{Xg}PqeqP+ud->iRJVxVL ziZqCr$XE1k%?6gTE~@3v(Gko6A2>~Ff!erteJ!*TP*99#I+<7Vf??~~B*tp=f_I7# z8-C+S%E6PgqJ^_ZKQy=(Z4zE`@_HRP6eG=fH?q4V+8cGFX07Vy^4`V0`c#ol?eWTe zVDrWuR#33PoW0z9(P_C=+zfFkOpnEs_Xd-o%b#D$G)ALJaQ)6op$J*G8A~#?u~H zt0j?e(a%{Hbmnp4svdVe=S{C`AtThKw&6c>Rm4&5*y>oy#ci#j?WTqoDNS*hQbfTC zVao?*Y2K;1)s(Gzro?4`vL^yAe`%DZ6;X9!OU~HN+8=dp2LyET$xk15mNZqXSj<6$ zB)4l}4}|xkyC2G`jDISV3DH0!O@pSA@|;7UMT+8AQ<=dk)h$KWur->`&X2dIGa#2M zq3ltRL{))oTiSC(A$B64G6mdE`(xO+6+EeHP;rqoKIP<)g&&L7N#U)j?Y{8^sa3AH@m?XYcrsSWvqh@OkOC9S zxFei-mI7}4X#rb-oEPEhn;%&cVjVkSPB`jJiwr1Q##7o2B99w^%cwAOuCrd;@*BL9 zi%~-O6oiw;*bT36ClnuegbfVl2q7)0mq5^H8P~LZ_J{4`O3EQ~n7I-xnEuUDXy>9i)RdHWd*Xl=GdHUklhY~G7@|`>pVYn%{ zVGtD*5kveDx_q7Lqg-b-@{7?zQ<(U!^)M}JZoCckczucNN(Q_J(~ z$L#%{=tZzb)YbT>xae?UCYC1agO%@}cFN5LQ3jh%Uw=F)S3_M{`4Z1e^RsoK{(w}7 zZ9LtG1V}You>>i}dhELHu;R|dcH;Y@t1fJXgLyR!vWy@ufgZX#dO8&0Pl?EZ4FnFy zU-5m(ro$d#QXwDK z9&Ye@MFZ#Q-r;7nQqF~%*Zk2{r`tpD%|XoR_KoG`z{;J&oszEy=iEbmCs+H;ZmAsc zcQ9&nm_X;Oag(o5!{LZkg4MWaCxcqoE${`aNU8jy)CE&Y}e^T)e#R8BLe<}6@6`=xh z_^{ZlgkFRMddYJmGg z+W%Y9@d)22Omc+;Qok5l<0;2hr6QF@|HmiibF7HuqBB0F!5J<|Z< zc~LSkb;%*hWrp=Kgzf|)bV;k)>r6BRO6s!W( z(nY0ge9|u(KrUcf7Fm@QI7IUJ2L8LBc#q|V=%I`jFp0(98IRcgOg{MA#)%H6&u+^X zm9D8-2VfiERrid0_z{~Iz14~Wgs&%dS*X)J0{)Qtk$;Zh9tNN?S7dOzw18}@9!P> zuiJlI$$UT!L0jti?pM}>exl~0R1MdMUR|s*?lS0oaMJ9&SafzbE~w`{2vx&i>VJ0T z&T?xhD;g{P7Td2Gi}r9gHO1E)=NDE$v%g&*W+pAqYA|jyn=iwQV~KA*C`$%Aab}!y ztxW{cK6evgdK2t}MlnGO1fgDclc5R|;wrBLnz5ew`QAM{ZRrX1#gN0&hV!qo(Vk!q zK!jh~j-d~HK`8($-}642#`Qz|*LG<_UFlse0a^1mJwV@Q=%Cwta@JBgGHs)1A?FHc zGu@fFtdcW!bJOz;z$l!Dx7;>94;wK8do9|mBWyd?j_n!bI+~0xR`KXHlxAZ<`eemGyQDx{3%b1 z{ZZZzq-&x|699r|fuRF?cECh)iE10gc~)f73n4EVxNMq3{C&qWXkBw9b6V zGRaxDLNH-3HOPc!^ntnmo+u?KM?%IE2F4CP{;P+?UIP^!cmM~I52`;6O8i8L9Rj-N z#m$zaX=$Q+nf)rPjg8t6RSbuZs6jH4yi?AWk=s>{h&`1=0#<=ZX2X71>{-#q=3*YA@EfM2u?$+#;TNEX zJPDauv`0iACknshy!ncqaw!ZH&1Lr^MWZdVPteE?|6Sa|8x6Wh_B`+r> zyPo7mPJwCUf6}+meTZXlge&dt`tb@Gh8+M_p97K$|FDg~9rK)$3kKFc@@*tGEOrU7 z+$|y``JPkr5WVf;sMJkLKi~26j6H;VwTRfMEKJr?*+-b(MgJS3z(oB($(UqY2=EO! z@08auDo;Mee-)J3Eri`2A+cn7|C|Q)a`II?ZY1(UB60#1FK+$k0LDwnDU(B7DO*ZD zA|ia0XG?*g`*Ac=_LwCD=(}uOWJ$W`Aht(EgJ5HU<%+KQ$tFDVN>qCUj|^DFZ*Dbh z598L4rPK1|KEh=EZ-_ps9@}p;of_)r#}9C}A>CCUCz@;HBUH^#&r?D{3j3oZ=z$nT zg$j4s0PmSfrnFI9BnLtoauT}!_b!Y!I_eCnXYdUGc|CvbX5l^(P8HU_3m&;y>QT|p zn{feEnvrrajz040FJ|AXJ#y<(W6CiKp;yh(4+w)?UatlMwq*8dw|u*Z{MU z@;Yg0HEM~R-H%*k*hql2)P}Dmscx;POg@dh+U3H{Ucxhs{wepQ)_=<6k?Veh0D0Nt zp{nlz!X><>#7+Rg)cjU7vL>+H)6Z~#B>=qoHxt}d4vwA5d?a3p?7t%VsQTw1#o+K$ zSl9(*4#tz984oG>o@DBCbr{zz3|{;fl(;arRy6Vh{wSN^)V9FX%mz)*rb3pa^79P~V8jVIEDrg?s7zhj z`zrpuQh)*?``cfLM$x~yt`7hUeRr;m@<=!)|K}!vVNpFQn(`eAph|8`z!DyL^{*!Q zU#Iq8r}qCFr^Xx=p08dQReQ4DO}}n1nymn?->ZrSj+D!NK*aboMwJ&PM`@#NnetUF z5vHW(MTi$|R@rOft_kI5h0M$=n4yxtMuiyj3ktp>pb>N>3p#qIxO0;&DXU3ZKVQ8#nW#mJlnWZI zXaSo^%V;l+4kCeu@6}0`R474NKi}-D;KFJ) zHfzfWDPW$sexbbbD`sViE7>G=QE>(yT7A4LF5%hzi9Hr~NL@#r6!f&Tv^!oMNr%Xp zYI0%GWf#QOOoLx}x^*{;#YPglwYOAjrNe({vW4O6pHA{Imit%Jr+F#c)_WIBv7ekp z6zSFX1QKFK+6cSqK4i2tO`68_$~W&_MRgpT%1RgqE-uC^8|~W5;XOU5a*f&5t`$0} z%Swj>QUfLf`=8hztNr}cE*%TWo7t8F)$1tTz?1U;EQ689Cy)b|npH!-3_#$@f+w2!n(!Uwzzr*ghwv zt_#@K&%U<6h?q)IjWK1dROm_NCcHA>_*q>S?z*0vw5Z|5+^@j!d^COQrK*%vAbR5x ztjFD%?*!x9X*#guR*BrNAqlKpZoLt)F_Snsg+m)%3$&{akx1=Hhs&9@vTBqzhX6Ss zI2Q2*R_%U~z>21oiR~HZbT4~7i=!v*HT4%$R^EP(-H$(;uFFZ{c;_FQ;kZouY+Tges;e`# zvQw}jk&R^xID2X8!dgs3fs%7dtIdZsMsn;HZ`;x8M@%Bfk-Ce-XQg6`HJ=TNoP72= z{26>TLF&+k)1?wNaXZ@Q~NqG!smLh}GC! zKpCg^vdChTZX{o$xUVw#%5*l8J~f^WjgYNR02z`pFLvsfS(`e5Lk>VA3mO=WZ&*|E zjY6Whb;{UZ;7}hx=mXFu0Om4{nvLf9Z_yfGi7EO*sk(~B* zI*ZP(=00LaL`uYJf<4Ox@4h@UQsnIj&uA%P$FXO^ zn!&}7sx&7R#DueKWKXJ=`h43=6QoDw)ssc#SA!^ zH*r;{Y-S{O`kJy|+cMz<$p0@Vjl)G1#U5@?K4VQ;=CiITI<2~#f%oIs4cInP-BN+W z0Jd}ry|JuoA)OS_3jBBXrk=%^0X>$qUtC#Hn#YBBwl6;T+}-rnADB+&?YWG-<#OI^ zCs8eNTzkFcayne*yv3g|7!I<{|~{=uJn{d@Mik$cm=_WZM(9;bUm5VPd#PJ29D!Tt}%si%_FOb~DzJ(=v?B ztLQ{A^_bG@Q*k6+ek{?hhOK1A+hn6n=4ynw;R+xZ`w2;1Ekri^I`4hQR}CurASJS7 zFanLtXI>hBOI1IgeV9C+chm9}v{-=$B1q^BJA>s9H&}~F<<{g*OjMw_d9bwmet*Z9 zv%YtGKHJf?Uw@G5n+7==mZ9i9P!!?3P`4#<%HJ*RB)__7U=V|;c=8OiVL>{BGM%S3 z4Yg#e+k7e50Y76YgqQ*;x;;s=GZ7Qo{y`0%8cOG-UDd@w$<&i$jV|lX>gOgoo=L9= z^=MfS_ypz+&l_Jidf){YW`5shtWoJ#aZF*F9E<0(ps&>Z=(*(I?YI-qNpxGNQyrC) z>QyGqxNBORlj7=mU*4n$F-ob=Ii{dF3`DF?wu7~+P6X-h3rxb-&nA9zQ|Ls;d*a&; z-?>*>SlS8G!6w)O)O+Jc8T|a;**N7)L?3R^5~}ZRE~4-6uGhp~w;KTE+e03B;e*09 za`Ta1?nZstb-O|R;wjpoD1DN3nJiNyb)^f}p0;KD$rNXZ@A$I80lB#Tej~)Iqrbwc zI#A?(%DA#Fd3(}=z5Wz;Yr5R5(Bi6D9Hs&Y4K`zB;E5k8;=V7k8_WB+%;{6csyphG z1kY5>mcjn>w76Ssr~rE?C*s+k*Kj`^7m*V#Xv}wCF@pk;|KTi`a&w zNy;1QZ4E1po}X44V~KpS6y~+qyf!liYui>u-QTKj>P78*PGi5CR2CL><`*~85u$gq zWta|%Tb(ahPu$*!g7R||W~?w=uYic^Q9zR~gNEl`UlN|9AkK3xFgyj`ulas!HJv$g zTv~f_%Zdq9zTp(m+F_@;Bk0~g9+8OvLWd;d>rztw^eM>O@>Dvt6Q>KscID$qaXwCvqWKR4tKRwD+Rs@Zo5@0#C9@+e#u2f&LjE$646nf}xlom+S^l44`+-YaA+koJXe0N7qB|5$lBiO;Ow1%pH$ zpO$p*GQ*=)wX0Lw`>`)USc90?l16yH|{K>46nEbQ4+Jc>TI+ucq)Sb?$|b%XV9|gVxrE zJMlg36@%IdyaI=b;gXQ`>|h7Cg&5i41`nqM&buvQyRM=PWHHnLKT}q%_amc}kej{g zwa?sxJ-YB!YWKhB#Un%6NU2hU-C}vICb~}88z}kbomG4us#&YdnN`Sr;=`y#w^Uy7 zK01xe-4jHxBg7)KVK*)@$BcOL=(MlbG`HR?@7&+YjO`eUhwz4%l$cF&&G|>8SFzqP z*l*fxbU)p~&bcJ%eBn4pXDusEzlWC424Zbl(tFciR#HxE!Y{YbyZbOv9(1>O44X(u z-J>n%F+Nh##QWkXtqe=oMMMe`=G*^$a>ZaTdE=COpf4iF;LbAb76W`0t;QZMRu_lP z#UI6Mo2*CfdN&w-4cIe4&k|rCNd~~PXP5syVp4^uVCTZGi}j=gMVspRs^fnC#jZ&M z_5s(jnTK=4=H9-t0loS%{vS zly58(i6@Ew43)%w^&u91Nd&V!C)G#02zp*8-^{A9mz`wGw6?Zp#1H8w6W%>Et~I)} zzdDNWO%`rnY(~!1lEa8Di|9XgSx?d6=hu0HL~`PxJztn5*HUazPyni z&Fo_n=5RkmI8IC8p8RQ@Hqf=#WFcO1!*#khDx*lk(ad5z^e&}B)+gg9+W{Y+V2)+( zBwRmN@lC>-{*I6>$4xsc?@h<_s%Clcs;w!zO!x|3bCW6W(24;?DwmXWL?h>M^NeMF z;ji^1+ppnd!qNkHo-cnp!=G%xX3}FAaq-q5_v(6fMUu;Zo?`0jlk7Yq>$py2X3fHw zM(F<>$UHbUem0q6$h0Qs&=J0Q`}SFDkRBBpA?uTRv(@eKMFThiAQ*7-YnLOLDg0^? zj^leYYSIs`_M#jfWph3^Q`%On%*-)BLDQ|&j|W6K+*jqX#B|_;q#SE<4yD;CV--9< zlEz5ToG{&;gxXe8k2aBd<1$&;a`?oIPPUEmc{hDpm*1#j9FYEw%~;$7bku%81jj1m z^S*@@A9@#8osgs4#G#Zkv-;-bRLAY)iLPF!0v+f=O%1e<9DN^0B5Ri$Dx=w=$r;RV>v%<6-N7#yay{|v_&EMDZh~G@JId)E!TE_bP zY}!65Cf)_ML~#68>y-!H*4`ta{p!l_Tx*2o_MNS=VTQQwF$o`(m80~r##Ap{YuuyM z>b3wyh3xtD8oQI-6$uEYqOHlXkRBv4_X=-!Te_1JGKZL($Y(>PVc=bPdra8$#wFtn zkxb2*>ah|#;c?#>Ok;;bN?|@(@7Kxc0GO;B3#35I8by$thi-Z&jWMBbu%6Mbth_)d z9fTjF(?A&G=22Hm|{s*;#^KEYKSt)KE9|6^kyG7)n?S z8*yvZ*cRlaY<8`tmf-%nvKS_Mh3kN!4&xf@Sj`y1pBSHBwZ#go^PWdk9 z_Z~HY^bm3+L+4P-Gv!gXdFn*%cEK=%P#aGh1|G%_k7gS;a~X|y4ae);sqA$ z11?;&n;7>ltNfBQQi_jWaJ%MtpBs3>%Dki-SiJ((4~%D~G&S<6m|UBE`i8wC2nD$8VA=Ce1^9!yKJll{+dCAGGmy+!j1!6ov25EynX= zrsF)d7IM}{$eo8?aNQpFW)`qOjP3SCd|Ro%{3hko59b!Siz!nw>xuo_dxmkM!s%{y zTayQ>r6iK->guL^`@92bgPZO$unMg$Pz=WHE6UVfZKR*x`!xmC(R>5dwkEpQb;KmXkaKHp-xPWS zucJU9E!!_U84#X~9}x5KYO(W)Hh6CZWJysWcpa0Exztgy=WTva=-VlpCaFtS+==Rs zdWgR3K>B;0tY7IElWhO;K%4z`!DVuu%cmv43(1yL^mU!fPOR@>=nC$0xm)J@${N0> zm994?4xt3BZTe~6lb4G??HKiNcksc>btiaG{Gj0$q8}hfsqjcm4$L&KW7*nmER7AT zx~2C&9d8{-#JH?3{DpUR%hwhke1xN`EXuQwFR3G5m#_C9dw9Ing*DGLRzCNiHnUV7 zC%4a{1Mxdj!_;g;nkzXW%_)QA_U*iFmX))z8V7hj)M?)5TI;*9I#qE<`%MOkJQm8V zYUgX>`o4Mt!Uwmv9uT}Xzx%tBgXr6rtq3c^VG0BJGjK)0({Q$nu;+d|dPjw~1;_O? zazC>70QdftnC+BMh=E={p22OwMT0JB-uEPMBTc0g9v#?K_f<&JGl~xMU*Ps%??toVQ0x$<|Y+qQ2jT2Lt}kth^K$k?*nx!tly zSq5XWjGYk~H)6REd}-Mogwe^s!11+{Mg)ZGc1tvqToz;`|Nr4_20bd!_5 z>pMQ2;=Jo*9;57L){9c8yk z)J&#Z6p^T~z~M&Lh#HPI);vpyeB!KPH{andhqrQY(ctj3Xp9+WSK3n9Sy#N=1?8Fk9L8&^tieCwv{igR21eSz*j5?)-j`~qXw*K2 z!&%$FoL~aBLcSiUcUGbnOsaBm(+^1d{gfl7K^U?7ZEq;6BJLlWxn}S>X_9K?<36t_ zyV-T1Dp@IVz-u{6nin;CA&s!5L)nX-6LGpwJx7W-<0N1)VnssPd)`gbnvTEE%L$cZ zr%F3%I;<6uHDtMi2|Sp}IZe55ZyCW>h~;DWPN-NOBhQW)D{W0SYU}vzl4J2b@~^s0 zeKXu14v8$qX{XRL?r16FX$RF$gXO1A=UVl&tqfVX8swPi{k$;EyQHw(+EnAVcWt&l zX{wua8Cyqk;ARaA)fkE(kwTx51etvYiGGV5T;%qM;b^<`ZS2wdFONH0k;t6Yl6HiF zfq~g-Q|iDs>zCDOdwqjLsQK6UQ-a^nQHS0V*|<+uG~E09D4j$c;nAt`q zlmkG}YE_8yZYtD6Q&QcsaDU8dyXWxL^Pw6lulBQy$cm{he-v0>4!OBXLg~UYsUGOG zvULBJGppp0Wt-mDXqU9SH<<7QkdmV(2kBDNQirCt9wSSfkm$>Qbw1jl)(}LaWGWmT z-Y|CFmfJ(>S;S=9N8x2Pb4NbLF8_Yth8WJ9UucSRJP zZj8CaJzf5rtgo~4%K`Q%Mcsn($OX5xlTHp6h88EWbX3FNgD`L~Bu7~svO}(XX|_dG z5*B@gpk5w{Gsche!8cBkJOpmi+Ss?ef-ao;ox! zlh-QIbDsD8YhqvF&Q*N~fa?(5x>@NfnDr6z9OxPDmOwUWA%+Br`tlTkIHIj#*(HF02s7<=UG(ynRv zfv``$GYhr%`i{1dff%&iD+iPv?Quec3|F4GGpV-Hl+Lnr2)x>P60`6~OT5N>pF!@l zi{h3^czK&Q;ZtEf!l1TAJ9YgJIXGLz ze|*u?hJjk|eC&O$76DG)lZkZu2a6iSx|#SB$@i3?jB znIJ|`dW3IeQ)p9+6U-mXXggxh9{(No9j=v;h#9*Lw_opo%&Ypvu<8vPByK%BHneyI$|^Bh4GA`*deE&&s-@&#IquY9?vMGd;`5$ z9oAV~it1Tc{nW<$4L6ONy z_DO(!pm2V!xM>1uA0ONtnMg~jr$=I4=Fs~FfS2WIvz1Az+Q2H#hjV-JgLz`9Yl}q%30rm%y^N8h|W&bn2U~1T7 zp(t7Xc;}Jtgw9o?81k?t_L>7fd~;A0J>n;G+2Pg`*Z_H(D_w?g+y-lt@f3UQ-b3el zzb?HFZztJPF5fAMs$<9yzpU(M*Dap-9qz6}>BBB482cx}$gVgT+cKOZ;8nrlIv1lV zqwbd`{5mJ$UkBVSbj+1oH6Nb+Sb_3aKH!T3#C^$3I-ICZ+MjxXpo=QS>nXkdPvzL8 zwD#D}`+G0Be4#lO6Wpy)6-h{854{GMW%X(7f=z-S+&2ABtb_CFe7?~1f}Nta$^%5Y zKVJBu!nY!P(8r6qK?$NTexYRWW@b^7q#o*8#`2Zjr5cp2awT9#?3SSL+MQj$J7E?N zir42n?2O@E(VB!h~Gaz)2xk!0o?8^-G+ zTTk+GF+`OQ*-AQpZ7!ivcLx%J*VHZo4p>^J!}-k!L}EL2JMl2IvJ{; zWi!{O4xKx@z$j`QeeMEUS}ywPA;2auxTUvF*DTixW&IJA0K<>>C@&^B-2hFb-l?~Y z$oHAQfwv(MMjfaJHTeon9aEkXLqB1<+|d{@hwAP)>8c(y&RmQ+S;*N`s*;UIzOBI) z(SKV{E=25)PtQ0=Sde;yew{iORa%={q|}hMGf;2S_m!KN6{9rF_sZ`IuIG`EZ9==a zh{(vKhi+SXa?i-%f=gEjHMgG}Ovw4TYWy)IZ!Tp}fon$G4>ypATg&sFA&Z$NLNF?=D-Lj&HHSSsACt~jw!|J@#x6>&g$pO#|0Da#TxZ*;GaTus|rczm56KG5#z$H6Go6OgV|cuPgw)sfkkt~j?`fw~9$kWa;nr6bj6$Ktv#T#e=0iI*di>YJb>H!+C>#i3_RtB8UudECQ!&|2Q zfUr~oX^yAbCrs6rF|8Bc9}chPpRX`??iEKRmRZI@W(}ZBhlK(#VyUn^xo<_!uBg`Z zY66$0@h>Khg=+wp+HUl=)&`tl@nV5XCUq!_?7mB;a+c;VhzPnd{rJHgf!FB~&4P-I zUM4)y*<%njRlb8jt8~;p(6*ue;FY6YnqnY|UdE?KtMc)JC|aouNs7>bh37Xpar_F* zUxE26Fn{fszjn-DJLa#W{Qv1u9yrP9=5Q+1UzLvGSO6UpA00i5%76d(_JsW{2ew;3 z`1b$&We zWJVh-anJwZKmU(Q6}o7N9|QCEbpw4as?gzhEe%fo6O8}1W_c@d@aNaP6$6s%egW;r z5C8GWe0NDN+wyZGrny)knD+_R0FyI6v*x2~J^fRo02M`$SsTY(y?pp*Pmm6%&QCM` z{JMGSz;N!kpxSSL1N*~=89;Z=`lm*83>!B=I4Zu&7XEoSf?z!QD;WPP;2;>O%F5Xv Wl5qCa5HSBmr=h0%JMPBahyMbWFn6v1 literal 0 HcmV?d00001 diff --git a/contrib/machine-learning/assets/cnn-filters.png b/contrib/machine-learning/assets/cnn-filters.png new file mode 100644 index 0000000000000000000000000000000000000000..463ca60003a60f2637d95e9dd0caa92db550ae97 GIT binary patch literal 38260 zcmeFZgx<^Zoq+XS`rH_xrl;xbEw|dJBE7CQo{u?m7_>5vihrtR@i=i6Ifu zg_x^X2t6c8=1W9GH>E%_GS3xdWLTcNI$MJrtcZvdLKE~z^tHOFGmJD9Z7*EPeLH=L zg5yH&+d*Olz@L#Zm#(sPNI$R03BL|Ck^LM54SVD%y&97KtYW|Y&p#`9s8@SZZx4Ey z9ZvhsOdhUF*_^p8PJZxNC(3g9lRfrFRtnLNuH`5^WZBv)&N0Sjp5kk4uf2u%62;htj9h}87>AG(z1G`g8}=m%ah*0- zYLSh4d09-8mZj~+f{09L$qy0L(EUdaa~7hv>WHNCH@${$5{&|6G7LTsyrx@$-4YMK ze4meqP9?<8&5ivUi`u&hYLo&?BDQMSf3W)?)f$!cFgl4ZB*9JNwrkcS_4lo*dja|L z0ZEe#$k*id`k)WPgZhVQFQ=WX)UPL5(6vdknK1F=^1FFQV1?Y=V|1UM|h&Aw+WH*acwW5iO1QO#+KF+VuGP|;H%qVt}rtg^HJo#;K+ zX$Esd^PeJP`VxH1JPLWh44&@jD7}tQ8tcuk*$JDoz0#LvUrT?!^hAH;^MjbnvW}9Z zk{fDBs+(TY*Ppz;B~qLmNTC?U#Mj&r_{Rpj>>n}0B0SPVoR==g;KjNgmq)Rf3>jt7 z#*$|Ik#^(jQ?lkH4!KYg7~7xhF%+k*;^ex(2w@Uk2p+yfGO>P>{t3lnIicCuBEA=j zH}7AelnMPJD?WK^za6%D)oV9+?6S@GjEma`PU>Dtw(Ju3joICAIjP^iqZkn8DzNNaEe(doJ+KMcNn)fW4+Nj@*Js= z#_`DNYI~;n@wS48k4N3pI}m6My=-SA6{U1`h8X6Hf%;9YVR=_AG2F;wB)g zSI3ExCtOM??AxPyHfLA7v-M3bPMhD^w3J~Sk~$%kFn~PVPP0#MN{2r>-g>9Cux|VL z!-EeG?^szb@*_9GSu9=KeG6Wy(Ad!E5&f884_V0|b3Zty3TavS;Di5tG3aDtP5ENi4^;_uB=up20l-;j? zNN?SH7VDMktH}rA2C-vAFxQLXb^=S~gxH#5-gI+V@->7Q%S}H8%C-FAFL$Ogl&i1N z{7B>W1uHr$4yzk^p{}9Biiav=*{J7!g`1;XE5>%BHBsIv-qA;Ls;d3Q{aV~|s<}e) zWyX%tDAAf=J7y3{z=tkaDC%Jj_eaN^^sK3T*vApfxO;&2b;-<@x27^u&jO_JS>K+6 zHJ6HSjM|jw6(dT_Gy;mA4t*MaVXrdyX!Oy@M7fJ%j(l7;sH}VF2k6^y>8M05M5{_Y z{G;F>oJyQ}_7&q~W!r`(s0WVrYb=g|d%kP^sDf%{e%_$VZ(3*h7x`*=)%e9cLd1l` zjMKT&HPcPR9K~Rbs77dGXrq<~r$_XP=kiPMac^T711!(`c*kkkkIr1;t)#4oQT2Rg3vIyNYoY(4*eCYk~?4iNK5|yYA?x{~y{8XRi zm#9Ri0&P~FAp01sMXTN+?X4}T?H=-@Eyu0*kdqw}Uz)po>kj>H6kqPWKV_cqtm;{L zUeSVW@7=_jL~!XNU9J+95m%{a*@ONM1Y1yck;$f(Fmm^?>f7rQ*UUX^iEFVa-$2$ zN-y8y>hbDMBNPl<=y*lHR{OaOxmMB?+%yXN21B?sdvXI)ow)JD8CwT`z7-|!pdth3r*da$tijcqA^ zk%`hiysDi$PUWF$`Y4Y}%xz_{hd!#Nk_~!vJX0ybKOzV0`w)RrTV+p-kbuhI%Pp4^ zE;HvpP(w=T*cW@*VFivs;rWuBH+fDR#enHjFh~$GqJf|3o8`cyz>NoWR*(fyY#~(#*AOlQN`$;>EKjWtTVfm@@a8fn5hgqZlU6&7D$jAl%-n#NZLyh{m z&7L7ntRA0ULaBF?_uR_171WA;{5hg)+&x}CUKmhX@K_k+-M5oh6)#fq_ zpVhrYy6BV?Mmh1s@$8yP6fS+<({%byeM!Z)3a1e#*i@d+XC`FEl8@H@^aM>yIYi1n zVCq82MIDm6B+4tXN1$C8=?7BA$nMCs;ICwY!VqCbiB)$mG({C_g1kpGkEEQW5MCd@ z`E}lp{Q>1^;M99Lmt>)6nrW znw2_x*h?NXq1Q>!ljs!y#MrKR71B`Tr+r+(>940|O4{tbNVkt_$HpSi&JQN*J=`Yy zqm2GEGO2TOf;VctTH2lx-M_l8gtMOp_`KN3IwEh9@&wm|8T@DuKJ98u<}^hgMw>}= z`)uX8ooOA7RNCYZ*fknAM!A!K&HRCTkfj^l;RBNG0d}VbyTf}6f$$T&-5TQv`cUbp z7{5Cwe8ssI=HZ8)q?y{EHNiCvtSbX)eJ9RhcQ;oHa@7Zb>W6__T0ocE91 zj*reUw;Y9FSxkjO;5CANkqHhE5}$Of#_!lWmu*%EGNxAq6P zV@ewT%tQxVuqF59Ew^FtAb1psR{Dz8s;Wef3GJ&ymoHEfT_m(F5Ppa*&=Fn!*(M@V zxnBV{ZdcXKSh=_#HLrM63n??K=G>Ku>#lPAY z&LN4OYRM=n65h2eT&=8}+-#lQKS*6L?8x)d(%IaK$IJ26`8Y%pUSfo%qm}zh7B5E!CpR%K$@@Pc#0c&4Zoqw(pCIn` zlK1shpR>p~yIQdb@$mBS-UnW1VPTPQwX_z~l$HC9PWVsqzOB3aD=`273n;Q&?|S4vlGktxG&9}J=`Vl z-#?$|AFp5QwDJP|b0#OZ-`gT=5ODqkz{kT2_*ZD|AnSjFcK+lS+RuIcnoi<;FfkoB zD_0q3M@K6sci=xMF7b1w|MBpj>->fI9OPx?pf3v|V7d|Z1mqJDdGZ_We?R({s0RN; z72xIlJLcb>{0;Mb4`P~DZq5!K=QPxD0=WYT6aK62|Ba>pPc$H(fB=B+H`?F(|2K}o zzrp!?|Nq8OcLfp1`0{*tz`y4Bd*5$(3BdXJ|BWAhG272x0=EILO91}Cxxnky6~a73 zMAAfxvQM?WF04*nt7p*pz}#lmoDA8fZUgSsi@FJW=Zn?1iLnofNSxq%UxGRtj$0o z>WUj1S4P*3o#1}c#coGs{YsiwZ?Q0dXpdWlb)nhO}O{Xf7Suneux1pfPR z&Zn5V9b>!pFnQ`fk3|>={2vzd-+>4R^dyoAHEleofL zgv;nHn^iutaveY%=|FExE*r?HFflOjMgYxaSUHGG^M~ zRP8LWUwFniTM+T-_kFGo`<$aLk>J#Lxaa=m&?b*BM_$Lolfq(DuFcD&0S8^xdIx&< z>&8AR{XEz1oLds$qNQ^Rp73wVSE=^Ajg~wc9l_)qIh8j?>4Sn~v2%zS2aD#ZAC4uI zWpm|1st5u4YSj^ecVPrFWyxPo5gS6~=;3DnDN5`vEj9}Yn|MsAVh%CmbE*I3$p+9+&1fTf8A89Q_Jo*2isafQi)<2ioe8_ z!}SXlQ~A%qYClESpf~ZU(fWH6zJExvz*^z~31OJB;<*3mQ4Pcm=IH2Y7f8Cxu;OTc z;d+w7FB#)t9tiF5Os!9$_dD{g(1khD``st`J;lm(uk0G0@W_AeXXQro(%_p)NX1^E zx65cDg$}l?@oT=0(X64nGAL&Uy=~_y)TqBDu)+rfILoDVYk3Rhg!zFMczN zk-JHs_)ulyp=uY@ud2Us`jWG?lS>5B%~aXQdUwKqEBC#bUU`wz0^KK<-IXy+dy=W* z4gLb*X)k_Ky6R_(Fx=4<^65&n%|!B^kEA$0MGzZ&OfTlHfbyOFtjUCBe!yU&GRDU;?IL9qzDW@P4k@Kfm$dma0qw1>N!k*L%DPhTR~1AJ6g>D$>o%>IoHcE*XUR~{n%05#TVlv za~4EKO?L+jesvx2n@vq_g$(>SG+>?x2MloS-f6cv%5`M_v8hc%rh&4OISNSxm<^Rp zKyw<~Jo?hb0kC)z`8L21Tm&-Qrg<_kblkXYxbw8bxRSpL2+uQ!5Po@bP<`r&PMm9o zG8Axb>Mgj;`2t{-P?P)6hy$nyAtn|4G=>;%=70s(w+Bd zb!uAG55xB&u0)e>2n)ENP1toc)V5x7F#@(x9l#j@Yu(! zbQEMjc&7Z9Ofc{60<4@vpBFaDlAu45Fvh5f;Fcv0bz$<4*)TYt+3)VXYjDMIY5>$v z)GRaYGC7GF8UR}wytwVSOU)Fp^1tp{9Hvi;#lZuXA8V{{thH4sQm z6B3nDeu5U(ba&8B9PnS6tUiW{>pAr#l^?bPMkMzoW6C?fQml&#qMwk% zj9}oque<`&Xy7nA!;u4elkkfG`13{stT1>g)!$`=`*2S`q_IU=zLf+6SXtT8f~TZVKjQ5^==E;+Gi(rz5nbO>AvZYM5~`G8hrro?0QwDe@%sQ?57h` z>d7{$w`&Op8zBY8%w1w%^H*E!F^Y(-7JgH2QDvTQkIEDM>$|-rWOtm(B7oU@9lHi}fW-xzkJW6TuKG<{7LHqfAfPyvLy4FrElH zUK?49s;%D>v`0u$(5Wji#IbSOD^Here&6>`&iHoIPF$CBKk~;CbLGsNS13rq8g?2o zR=8)>uu|sU=^S7NITFaizjjBb!~2Pwjuf}Aa;^Jp#14{xEeApu z_wz(9vo?x1eFVrEV@l&KnlJ_q`zJdzC7#iiE~8bZ!~&iX_1KT40ennMp+oA8Mss3- zV@yIGuq(uBB_YIWy`Qq8CbW(gxe#-)V8I!qbaJP&eszm;#Cg4?Zs(Hy*`nADigKyb z)otW5FJ~2Sz@&IDWd7twnEfyyU>7U2dD6&;$ZyzQEU?MUG*znX@JQ_Ka%f9^Lq1X7 zY;J0AH+l9(zhJLUY~5H#5YK8L}FxY33d=S8y)(q}{G+ z?<)v?QsbL^=DX1`2s7ar`m(bY&V$or*IV0ksz0oCUs}gOrzUYnf4;jhQddP*(y&df zz2=6OiOwZq@~*Zsvf+fEp27?YruZE_k9VDMN^tmoix3*?7;-NPzWt_Gsk^HOF@JGK z9vN2}Te?~}dc~M=4IP3or>(Osp*@4zTR9%Bt*>)TVRW7Gay4c!3;CEkiqla_H>tcp z)DGihFDHh4n&p16WGPLES}|&Cf=ay(;wrEgx>mFK@+O8^i(fYM?b1HrP_aopInlHD z0AA0EPekKNCx*|=Vf2`u*7aIElr&($M(-?pO2cG@Q7~f*wiY$B2xuSqIs^Z))fjcC zCU}ElzwIjE87Fpbv4&AsyagZ|VPZr_0Y)&5i2brXKYb@&FLlCSsS>e!0LU5Oby6zf zmPE>H`#__{S@Ty);oL{t{zh@OxaH)8^7at+gq(8i7TtzPQ_5P`Oeoik$H1w9`N1Z1 zHD#U358}f5!YZ`QSwerr0K(Q}@F|Cczh`|Nvo}3RTU9r=5#k^43c0ZMaHm%-L24eJ zs*u8YZmteW5KF$djU*4{EQft`lCDJah!=p4aav%WH_1)AIH}1Q?sS?;@oy-s3d(M> zz8d>YJkJ|^AUg<#P>KilIQsdGYAk9gJwiTW$fQefBv$X6;QnRlZZoI$ky=Wd#iaOG(+nQ;}Rdc$f3nkR1tM$7c6 zRP)5=+?nHe-I?j# z*nf}JGh9R7bMobJqfAA6jLm{tQ82U5Nwgb(nUqSzc*PP^Q{H?OBtYzOI48DUa5q0y zKj7`6X_%Mj+`1QC;~vLC;o+&E@H3&uI47q&HLO29nRcp>A;D?CuBRBkbw6+zD9=v$ zfsu3L0zxO1eoX%Mrbdo=)3y0Jj$QZ~e0@~l)$HfuqmFZbbi&%mhTfbmJefM%^g|;e zez?n(VqLIUR?2aQZ#*%iY1(c1DdsIXQ#$FsIGay{*%}Y#Ff8_ph3p4<=Z3O=;ARV0 zb*g04Scp)nS~fz;AW4Kq;rbok%t6&G(vQ- zq;Q{;iuHjdqY1ZYlaXm3T*le<*Y(6j!H~B$ecXJY;kU)})A6L48sgPmWN1~j%+pa( z2#E=S^J<*W)Wa^BHMEUX$TRud(ur9rV8WbWtkAN?wi9g{-?={Du4;>E{Pd303|yeU zCxZnSxwXYvRry`px>{7nU9-cuV&UbAIXpR<6^AR~;xk*HIUA$V>Pw=}i(S)$Wb`a> z*G)e0Tbbn7bSIV%eJ!*?&6t(IhI73!@O{m_S{G;k;g|bdNKLpZAcN-J-r~r8o7QbT zGIHu1^EO`d1gokGsd*Ro0`x#yH+PvgpF4`;(?&*sdEYX*R>U9Xl~mFnJq|7kuje!m zqxG+^qD)tiXNGuTics>Ehi3aoA0^Q8cdr@Y7C#MdzSPF-N?(i{jb$s!(WIeBxfe0@ zN;&n^6A9|v>2?EV)K0oS$J8L*7V`s+JL;{YKJOcNKPYI}(^B z7Q1e{4JP%*mr~(U@$1=Jc9^C24k8Qp6EE)h=z)Ocu2b{YJ*b9PtWn`ZruR*wh6#3` zFHq{Y-PdtdR+IC+tP8(U{IXHg1MEfL)zE@swO8tixe{j=*9W8!1eSI`Llz``tE#9l znPi~xGg+Y20zWgewhAq{k>l+ATqwX=PuUo(5};vnWaWm)&}?k@)GYbFKFJK?2=RCm z!1FENz-Zsn;ejS%_ahm-n80cKl2qLOeES!Rl}Z#4-ct~j1Kf;J@xUYjAx`5nO-oFN zi(YfPok;C_=?QW*V4Edm%une^d%eE4?alpQhWY@|HdK6Fzma1%TX|gd4Q2Nk;lim= zbw;c6*_8aWW%}<8vWAY+wzkx^rK|GY^=^=z+pXss3}a^$nvpn;=@R?&P;KI&qc8}) z#gl#(AwFTB851GwwWN2PXh3l-(ZLGT{nY<*;HnMwq#;eMkEEs;g zRBqKhCOM2lHXAirGCT1lt<0P(&{VO!dNr*`y;R7>SFZ0Uiec?sQXKbkoSfjxE4W3_ ztmD3jW!dl_W|xSZt!3QWHf+c8%*1xv$QRC`^o5uL+#h#HSi>~TLt;=*yhjT)3toys z=*pdiA!zsM4+Z)7ll67oHKtI+vUZvT$Q;xXAy;cQ7W!ECKEzP1vSU&6{@G}BUcf;T zzAy#Az4$KV5b(f|bAyZWJem>6bDg0c-@>stQJNMmu$W>N@39&$j%4b}2Ng1?J|Ke) zjBoHj?1-8>9E)QHdY;~r6ukeMB(RN44K_NH6J3;6TR&;=^zA19k*X5lkqN#E-ay; zA=gXGkW|$-*b6|6t_H#B(S1@)<_pD6Y=&m3;_$6cc?eO{1m~C!?qSd)-w_Yf8+)KTo4^k%7qH2K0X6Pu6 zPY65F`ii8=?-bc&aKdZ)2juoKwW42`g&l0qy>Tzm{Ib#jiAT4utskU^&j^7!2DY89&X z^(v8lCR5pDfktn6pLb3F$0+}q?~q`tg>bY@;AdPDZt~$>rl%$k4C$x*6P8TjlK}PH z@|i@X?$N9Ic1x`xmcHs9jk*~#Pk!q9WqaZw4Jt7jp0p&W& zET3OA#iTeFMOg+LVNb3Ew%|IG7wj>SBEG?E>j)-jC&;;gKQ|#!qJB7Y<&HQC+B6ihl7Zs5(4gz9UMN`Dj)HV}fMR zT%hd!=DMT~F*<495SJ=6b_5^p3XiKkfq%00J@bQ)<>d z9<_77C07J4jRe8HJ!pj-6vipgo@9{|U}K@Br;bH5tK&F}^xv8n@M_{I$>;+^Eg^Pn zqpZFAY+tHbI&V0EnbJMG$aMih5%(7DfOLXgu^%+h449F3*?GX$u?qB;YAJ zl?)v|-}wG!dWPg(-+e#wT=op#e1RV>deDz%@G}!+8E7w5zt*&frqhZXo2ZmChx1;l ziTpMB1aBZkYjDjm@3RFh=t2%a|F&@9t{St(VVcRIsGatyr7moO+v#jYttjb}0{}8^ z8Znx_ovy(Ckp`E9N16Bac<;M2 zFy)N2J&K98R41y_oAdcLG62*&0*`qJ=h7-JACNO7%)teJ2Jus5vgEyAOdB`y80bSn zQyqn}Tx;(2?*5s2#_0aIJ>8a1#l;cg*4=p2@#!;Pv_s#W@dV0rTdC&N0?eY(dKWml zp;IV7KC?rA+InJ^>)GU&4QH83h=NbfXcfkmDU3F%p*B7f<>-Dqmcc%bWrOt|>yS<9 zP1iOrLff`H|FRCgU7iEIMHL9qa2e}+6K|h6Ow6KOm!BT%Iii+DisM8|VPvJnN&+ty z`q^sS^vdX;ILnYR3h(=IHoXL7+>WBqY%~-Lqi^z#n4T;*wb4t1n>Oo15pwvJQI)%z zKxpQ#Q2ghh8}ZH|#o)JJ0@4avwraKpt9OCh!uiqVTMYD-3?ph|oP4OBjzn)i^@}9q z>){OVBk4v$D7E5G5%{J@r*rv6jcJyV+qxa`yDKnUU}CMH!@75m&z`Q2vqFRc)WW&_ zJ0I68iM6Q&%8@rG-vfh(;Q{mLBGm5q{YjlHBjR92U+L6Sf*aqx%Cp8^QY)`Qu`R5V7#W!( z=97{&z2Tg&)2Go> z-V0P^5$8~dKvL_JtWI+3)%dP|?>K9Tk6!v25hi$>B7spsLnekxi+fvYw$3?*u*-=(_qh5%l4*YX!##crn-2IZdPRi>vlEA$=H^?{S!<#A;rEFC&wFR>3{buchc}Qhjw4mwOG#sj z-^%(I4gEM$S?x<)ER?nt%9>K`Pe0vr;l}Tc0qVrvl^oN}VW#heq1;i1;=HbxRCa?< zCE&@(De6|uQ8EpN?KkbdrMpU!+t&=}O>; zxcB26*uET>v615U^L;jB)H+j&;l@hTsk{NQpx6+M>BQbt#0k z(L&I)-FlCkE>C$h6T-RdWUA0+D9*)8w)%KgGH(aKbldx7N4P=2PI%tgGVP*hhGiKy zzo~DYXPn3RQYU18URJ#VHw+j&lSCGTXUB3@;2KTv*%mG2MCyF&JGEcR|+=o?;c z<%Z^c$3G@_Lo>pagBe7)N$IpDOO-C^0AQ0s4JJ-#f_ZP{XP1=YzgaDB6`CcrYTr9C zL|Inz@K@Mg8hpvz_RMb|$vKtO?=zzzTKz4qJcm>6E(d&nvmgT$XB+tu67as=B33&* zb$2ePxc;RxcGX6nqvKnCW#OvA<4pLIZ(QeHZ7wdQJ!G*V8qEGXq1E^nt&-`pLDuMk zcgRW0&GjOiaSsL6DSIWcrFr%O8na@Y#NELFX|bgUt&$9`5ZYIpb13#JD9<+S^y=tR zfyX!;D+AA)pITMmophTN-;WjfO8VBvy^SPyO-m=W4SDukcVwB*<0L`}(_N&yM<+NK zAzv_f;aDefk%C-z^03KAXv?8!l-A1f^yLn$24V70B>s)5pboB_^2AJ*Z@75VTW<9(E^KYVXGgXpfMRos`;> z?`-A9V@HL2}zK3BuO+IHaPK^tD#mZqnASe5}>-)`i^1otl^TcCyqr%JX z;_LN(sjTfLCsIe41Weg#>*F3m9yp&Mj4I8W8`{4(WT;*!@N=>SJ=Q-OP7)Ah2oD8cPkE&oRsz@c9^jo_@ys7+PCyaH|GH;k#5;UHco6Wofa*yrrvSVkg2S-WsHx#XM97t>KY*(22No5^AUp6$rH|nr7C53ci)s=$lVs?UhmNxiJcFhe+JTasKePt z9^=zJCtCWym{-?l8oQ`YC*5_iNgj)Oj982`nYbm)F`kIwH-(~hw)zY_=7)PeW|4oR z{8GafByolk#I+;^#7yyNbj^?AU)~QB~+?h(7x2P8U_#W}B6Je8PGU0C5t?^?kQ1 zM?L=@QX-!(UIFGhdtuDYV+>v3YBU*GgOk8ISA?;Xo4f8Wd(i_OZ1>)N#v&Pk*i3J3 zcmUjQ?;DN`RaV`ozW3wvm;rIkUiaX-FTS#RJ_?;@0!Hz*%i=cUtshts(n%m~9=_%& zLhAGRZm#2l6#Ooh##%;t^~3$Eq2IosWX*1(H4fc7gQLYCo(}8-xS4Br+}gwLnUI>_QVAyD8zRJ``mn~VPLL5Uw(!buKf5?E^+ZonMZ_Q&iqb zp5hhq`DX6{mi+Pg6bppV$@I0Wn(R;GUyL0?2b>=JY!%rkcSeYmPP;^V4cr6C1YLnj z3{e|Zr#H^YOTivD8EmXL)YV|ZGf%Zqe5rHUoC_B040UjLN*wK{c>c2e@bP{!Lby}4^BZ<)$5bX^ykP%A)# zl>Su@C(VD|*zFXcO#ZZO^j5hw-%FT(vPHC5rJY#0+&+cilUn=S%~9B%0tEn}R(D7E zyX0Bh2q@LzaVTVa36b97AhSZj3Li}*$byWsnP*AyL8p=TY%~3#j#5_5HlX&~cAK|X z2$}q}48OB?#jVen*V&{FA!D`BQh`QTlj}F=DwX=uo-Tv`^3xF|zlnt1uyt|#2Ea_c zX}#7>h;m*i?`5WiKr5<}5Xf_1v;V11vcQ+Ue}s4?Pz)P@YsC0V!le$z+QP&IOr7iebGqkQYb~p zXyWo}aheuaqDA3%RHgYsk7nn0)@aFCU90YKlq(^WA0I@@;~d?+%=#K+G@kF(m<{7Q zLgQv8xR=sgl;~0jI$DOlztR&T_-_xeT@G<_L4Hqz{euLak1|38bJM$R4pHuAjKIJ|4n}Zo8%8HdDajbTh z4thF6S4w?<$S@cB#4n3SP?*eJfypJj;vQ%b16OjBAN-YUZWkF@=a*9fo)gtX}bIYc58ui`DM8HXulr^!xDQYegoH-B_6}< zS4UFkKaXGC(-(s>YO|Lx|^=)zdS4U%VeUj4EK1yE|aNKgyyMdg5{=w zRdRb2f?~DUjqJw5xhS+c&n~;?IE@D(=3XUhjMgs~FF-$}zq5W3IntV*U|WZJyV~bK zS9^G4JUWYVB=8i8KXMQDbBqmm@ zX7JPdx%&*B&r!_}yA1&tLuaE1dFIv@HKbqN4k1^k7;U-jF$UzLJC^-pZKIV;aUHnx zxv_QZnctk^sDeXB95EqJFPr+^&m!}Yi{9jBCsV@SH?bRNPI*Et-0fDolPx_s4{2QG zlP1ox+O~!$u`S8GkBRnX;k&d)2W|mdp{t=N51AQVCX4f22xH~v)#^B48YxHRC+(#k zi_xlLU)Z$k&1hp$X8lqGCxyNr#w|cIW%7(1Irxw-TG}yGR2|&2h?V5PS~5u5S+4aS zdb6SdQp?OVwMcK{qirstlkPi+eD;h!++ay^lAy17eWI57$t#fBBx>LAcPXKeG0S4= z4BkXbd4xloi7lverL^tV*MPl(OeMtBLCI|}3ZP%lOaIYwN-J++UoLGE=kOqY-+EKG zZ&TZ7rJXmuU2f=_$C)UzdXkNcjN>F-4kW&Swh)}=jjrvY>@UbXiRGG3e(&!x&yA^p zY-6LUD+vXs`SVeJW?@4nkCh`nZ_$_?s5A=E>#)tCQIxxpBKBym=@U5q0I#EJ#CO`B zRN}dA7ooSCpOWB>&~Ga9lqAzQ|GQUo%5)2zUG3QE?@^mK8dC2wb@x>2cPdmT&?9ps zR>Fw6KsD`?DyXeK)5OC>sBLwpxWIt*iWg#3a)g6DBj&zZu^t^#esp<)X9Ybrkvm;8 z<|HnXs>bTX(=3si>M>>e0LwUU{nZP_cO-5>Hrn&#Rn8PfdLHu-_t^akdSY9ZS+ zmQ)>ZjTAf~o6VQkX6wTYT}u~_sll>Lg0& zdh@9oM}6qWZ|0MDXrf3*bLktNpfy^cgFfnIlcMVOZEJl%9~6KRAub@Ki!i(ms3AdT z+v8{SeutcRLyzUF#rPut&{ysR!tQppcvfJrE4lBF9oaSqKD6?)*JInGgMMVpQ^G4LB@nIy!i^Wk4bzrU+*&#M^pE}P3XWuBa@6!(T}n5ju&avbGO zj*(;VNBw+;4cmOkk%mqiEjY(7@mzjwlpqw!P4ckkO6}Jb`FkUflrlL+GS=8A(K-ie z9H;6=(K@Bplm8`zU6YxhO5OLFaphBjj+v2tzm!X0%OjQ{6phiC;6I9v&#mphM$URl z%!kL#elf~l=+)C^borkB3_~dJf}(Vro`@=cRYO|Tq56OO-I=$RXV{@%pZy*G9|OIX zj=tJ<(zul2VkeYo4{nM(&eKwpxc9-NT0W)y={dEsu3c(8?w+`$qkH+6Vf#1v3f*k7FTUk>|+NciuS{Df{Og8--6ejD5b5aazP^2zCB1z~*rl_auO2{0gi%8}?p(9{6&B3plq{drKHFr^ zyw#=HXqb4+J%vJGoW1Sq0ys_l)K69SW$ignsT*|vl1Klzjel>@-6z;nx8z9vwR!d% zGAn_gca4f_OgR6J`hScA(sPY^HRsL0(aXPZ2bBn#zHAf)P56)6If$Uch29Q5{XdiZ zGOB;!+OiRB3tc-$#D6wtWC*LL|GLcbchvu3F`Tcg`aQv#;jF2(|IZf4J;LfqU%KD= zk3=<8B$xvUhUl7~MO*)lW`nl~s}JH8yZY|{elkZH;bXigrLr}{|7>N25LPck%lP&` z5>=Oya6ozEo@M`$5(4MO%)JP*-xXT_`ZT3EuPId5YT7sbm%aYWUcWAB;QwFtdQla` z(Ee1Y=yRB`UcL=3B~}ye4_eV^%D-@t#9{bIL^Lpia-(Tlt1=4hK+MmS{9!t|?m;Bl zUj^maGv2Q&xqDCg1=Te!gUSc6gJEURVxIx1FPRUN>6f3+O+t#2t#L6VhOx4oVHs^~ zFn)mqXDZjAyd}ldcQdcr9a&kupW&0=9ml@w|12$*DDdmm0~Ib6t)k^C1o>KZ!FF2H z>%wWkJLj(_#19PJ!e$-jt&D=)a?t9y46RIy~nG%nxI3lKUE9I51aODK@x(*MNkxBV5bCUuG=B;!`b4g=3W z*_gjt;Uk8pTFNKS-1z9*beY&sQThatWH?e2f7#yy>xh|w@9he(Dp%?zR#av-(o11nEN7gq*diYv;+ccRcNilY|0Z1T5YqRQR5;Hj z(+_Cky7`Rov9tLVoDJ&*;xcxE!rDbO-+@km({*jF5c=xUcr(cabLTi)?pMcCJ_|XR zpz;}mSdqCCMh%*(Nn`^(s2)(hhz%Z`?S5o3aeceZ6X@qd5;lgZK9F-ow${(6YJ1 z_URt6KSEfx-zV(R#vrMBe(D*v4zw_=^Q^Q*jEZGYhFH(U&(ZtrkVZ*9`oW4+z%DZ( zG8n)mg6^2{LnS*(8vOaD}go$rTgp|9Qtq( z;y8xH!2hqk?*M8l|Nhko0-~a#pi~>6BB0U)1Y)5RKxq=HAT<<`CM7`DwSj<&fPxf7 zK%@!MOM*&~-U&SlQbGtp5^6~DzQJAH-M2sIz5o1X-pm`FVZ4*v+DC1xroQF>Ocx3k3<)ukSgMOmvWCMnUpdH>bvCIN{t>Bk!174U+Es7+lZQfm%LS0#i=FQXEl zpqFp8#h7+Q1qb<~T~QlQy*uZg_9?mzzc!t&(90?gB3PBsF=*W4t{hgEF#?0O=xw=p znDE-(W4vetAKI`|02E2BniOE|tcC4(SpC8jr-@S}7s$}irt`y_qXZtS>kB;0GY3eH z59_E_ui0CxmW{1SCG@o{&<(DWkr;3tjk8x>)dDw%K4hs0QFD>-8Qy^!3_&{5i|ry+ zjztJ#U}&qdYakC7+K zXYry4N)$MM`*BnC*@QH1#>ABOfN7+-$9yz2c_$Ej$N0`$-uvI!HZ)Rt!%E1?zC$6G zR@dwuGc39KRd5bWKJ2;irK5g6JgRj$(Rb*ib)MzvnfCMZFs`d!YdIFqd*=me&RqfY z-f}*+XBdjo5JJj?Ojd9y$yn!F-c>NOyZrh3%CMzh953xKG)0XiNVHh57LoWN#4u-U7d5D=ey?RO?YOu7{x)cOmo-LU{%Yz7t>Vg zT_d(#*D$hAWPxD1M(|_>9m(pSI_ibmxC-lqDkhs2c+!n3h0eFYl20BYD%eK_R1B!6 zWCD$Z4WBtyvJ(QpRvoX&)~2ZTe%#b@1^v@`uDHdJE31?>Bo9cZ2yRTim($gob0ZB+ zD74M0aG-@CseZsmff56!oT=&DjMY~%-pvmnJ#b?VZ)>pROD&bDHz28Ek-C#>^%|YI zcV-PL8?(S10%ge)A@A1K1P1kEK9LK@vXy%vjf_)9IjhXX*zM^8G=f_uIQnW_fs9IT{Hb&+EP;1 z=*p^LSF!d&$S5Y=jc*SkwE9sRA!8%o2^%P%g6)Mn8iLXe-t^V^ae^w>d<0<3R2QIv z>P;>$)vfju#B9zv3^Z64J}8ir)6tsy=ub~D1quNp4jK?jPT>;s$O}fr41E;( z;X4Kx+YsYrS5|tnEiunLvd#9MzGj2A?|E&^)ID%o9|eLq#z@a=5bc_MxtuV<@gQy; zIc6oEVKW8ix5&p2p*`*FyoRj#SRV;&fqFS-ptxf4y`S%H?8v25Nt zRmq<|?XenrGl2l``#@2zFn_>mP0|uag-bIqU)`_0X>=Y zgxQIw0qVs}v>NKkdQ6p9T$Maf@uCqKgmu<$@P)9*rG;J~q<8zd|A;7jSc((QDBu(W z7iWUXM&6IshIci_h?C>$BKJj00fPl-fJwa>{-R|Z^F%pNuP1=0J&vR0+tFZo3;BA? z@Vb^~&N8Eqn6U`rnmxehN7!

C}&V{e*Tl_n=cCi*Am#{)R^RmrYse{u_3`mpby ztaZuyS|nK6P^LPZDN`5y)sp;uX*gn#UG2GX+(Kp`$?|KM+B&rcLBQn*!`;pW6uE$4 z#)S*s$d5?&xCh|xHR~7nyeDDsREv;%*77lg-V8M2w*7pkmg#x357;)1yuQsYt zylTiY$%l7_4Z!|IiV)Fm%CRc z(~_+{?ZNzyo{ewizHKkUir1Ilxs&g0|yaSKI=0@Y=+{s2Nw z=gMTFA$&1peT7r2)Ss&1`EhF|N1rP?wj%H^ge`{B zZ=b)jT}!oeO3AtV)kQa}x8ImRVP$%{=ful5Ls%Lix3^2`?4~#}=(J0XbW=$hn^R}O ztz@JQ*avVf`lelIppakbn~u?|OBg(PfVw-ZV1feGyCp_dN0|1J!?{VK#MLX!+)Y-J zMmA-fq%OtC<@G~Trhp-SQIfKX-9jN?rI54ws?li1FheU^O#d~qtT1@UQ&$2_2emoK#+*bfvvSca5@Gzl%n`ix_g)#9zhdB-d8xQN;@%rp{;O}1!{*Ny-@g2Ap+rVceBVF$w}agpz|%PGmX`i+Uw+3%WbOg{C0gcp zm+zna+kt@qP?cW{W3I|ChJ6>1UlR6Rwf~+@5SksN25?VdZ zi5c9a;iLJ(D+5FxDF{`0c(5d?+)o3F(F7=fxzWauuJD2D)5FIhNltFDu?Ad!`@ug^ z<}J)@@{{yS7k_$Iyc#e$X7F>$-*C16V8o8x2PWrAytdj8XZO2E{Wx&q0Vc=6nPcIf zr2OsLDJOu9W(JHu2fxjj>Gw2|?w_aM-!s$i>6u19Pro}blLfyR#&qU?G3-xYKINB$ zeP_jgS0!Cdp2WX)s(^@(SUgQCL#S)hR?{*yh+wpyK>IA(YQKRDN<=62C9ieB(UxZl z;t~6#MTu|Qx(mVFO|^=x3nJ|jsC`_`x*pBukLui1MctI_3GP+cRu*=OC4gs;C2p2x zR1h>5dW^-`a%P*o6>SMv=%I{;OQoeX#ims)4N1e2&SduSrh$-D#sc8m*DkRg4+TKx zhKQG8{D=3`2lp4BtJHc?w4h)iqQ`9+=_N!qz9>IEBPZF^);VD^Dl-wf4I$=A)n&S!1iv2Wb?> z$Yt#Uo&~<{gbk>=Ez7#D*%_&ObO5$tTNF3o3;V>6t2}`uy!XHkE9?@-;;tD6EcQ=! ziSSxKF6NWZoeh@TwZFXtU~0$;q@`s&h+QI)m2lmSU#(9L<|q}#IuRpHqKSc|n5cAq zbg*M|kG2oH>$Cq`FkOE3

N^zQrED94om(Ms%*l@yMQaGXk)r!wvw?Zp}&Ao}dkg zhKZ0Wm5*K{`e02(%*6DbJW#nd)8RbRmyf5j3G4a|c{<_)TfN#;ll7Eo&E6k;Of#OJ zkGR|u?KaxS5)T7>c9nRcQgsyMaA0ne(-Pn@JhVIp$eM-jaK<)8DnB_k-5QKi#`jlP z52Y*xa-$r0N~fn0UF0^5!E4*Y7a1#Aj3;^uD|Y@ds*EyKSn_E`7e2o}`G^~Ju>=#{ zdYV>P*zZ7pHm=FDMeJ|r{H?5tEdZVO0s&NDYwyDZQ}8qsZv3)?r%`?*{$fZXf$>!M z>hwrpny>{WQO2eFRcA4$3a!Drrw99zL6*1HI~js)e)_i47;BNxs!Yvh0^U2Os9;6V zxVqdoIY4M~?nxlegNKwK$QL*5z=$jLXDG0Ya-gYs$TLXteH{_AgjB8W7w;e)&*u_J zmXC!&M5vbu#)fg8H0?;^4N;8%P(mSYjg_-g0#hy<62%Hp+jaTOWDDhklvQFbJ;PkY zEWG>Lv~~(Fz;_oR1%@&)Abf=N2psF$1i{iqS9gCvsqX`D-Rj~kTyu2xcvm7NpSz0} zGgmh6vf2?P)wT|EE-~c^6JdF!2oOE+0gmI|!C0u0b0dWvh7U$Nzb9fDR3g57MyuDj zX`~TR(zP*~Y|qIOCb>hupf!HMC3xRtqIn3t=gNoMXZDMu8nVvZ_>v+MFnTv=7y{yGaS%uQ45E$mhCsT=ZKbztmWjFt=hq7$jrdf9 z6N0L)e+H?jV<~kQ8Bei#QWgkYR|cty2M(g>XKizadSt z&==CJD^-7D%lN5l?GE@FjDjK zz z1EYe>KC4tJ5Y6Ws+I3&bt#~C=hIlA2<3n3g?CmK2_*hJS!M?0aJr+65b#xt9@>g^{ zLig|JdL8MMq~O; z#5m!iwn?f2YhG zP-f0;z8VR{lvHSMnp@NU&hnzmBc!d0i54h6-P$v?a~*fD{|oL;jsufklK?bmz=D#v z<946{{`>7q8ju7{usw+W<~INn{Z!E^n(8-dwc~PuH%$h*Y4cXiLmKy|=CbYTjDp+D z7=oD}2_zgm_o47x>9Nc8tl zgj(+e`WF-HEeK)}O|3O79K#A~X?jL-W1)te*_Llb^KkJ+ggA19EGH9VEy;!#9Ru37 z4!49!0;QUgctr=ly)5-6eLk=ioNMjm*s%1>*Tz+}y`HJ;l_*Q0jSvZu{d7d;gx;SQ zzrwPaBHZq5i3_F@gOcCfENNz%2d3+aJ`AoZRg{i608Qy({R5y6Y>mBXp4hy=?W6@3 zMn|q9TojrTwdF2j)#-zYg~{a3;M!Im5Vf>NLvG&5CSAC^U^1!y%4AmBd#XJtJKk=+ z6AWdw2Z(CckRBKzh!Wbf_e?1mweh+{QQ36dNf|xq-DuASL8=2I%FLJnV)ox-zzK06 za4H%Y?Ahhyf;w>i_Ll|oLMKm{>@hU&4G&fi&c`;9d24dl5Q7R7?A2XM#I(_G>wFF7}pp094f@6#m z5$sO~p(*E4L{a@#>ty9XU$lkCm9L|<&f|?SD>`=d_Z$A?)eBBgXb`~pn%ODR%jqib z-?F)E0nk}Q$-Zb-#U zrI!`PF$pPYom($3Gws1GtlcfK>Y>mqj=KzaR;>seWQHvzQ0Y8J98@dC-E@EC;v$u;6+!ujaWOM7tk^ARBy zmy2D^0x0?QKVO`d3L!;14t(mJ9Ho5%LpFV9_Z9-vc= z(XLN{%>`8pgbxvfPQX_t5d@iv3kX7Ck+_HfJIYXBK_PEaB+e@KMe6ecW^qa4QiFa5 z80{_PCUolRHAO;ml6bWZum&(O;eK%#CYkci<|afdwT{-Db9cdImuz9fb=HIbNB$k2jzWvIkAC zkvrS+N>82S$O_hiLoJ=(FA*&$6Z6;^a+Fa7T+&t<8(Cqy_yQ*$)L40w|3s zywiuN5oyoQ-$}=;g{$g>P{%P+o@bWU$oQIQeOP_S+$QJuH5_;po3J*f8BNHxj7D-a zHujo(#I&lCo^=*54uP~d*pWcql}Xkg6kDLr-Rmv*9gKXDxmC?$?g~1=-z{}I(kW#8 zE#CCFX>Tgg@8z*o`t9F&QQC0?6N<{$ppa-6>@R&Xz4Ps3B;-)~ZSRP!*8n3)NIi7>nCeaHtLrT`myy z*&LI+yu1uAOoj}{P4GPQHD7+3SB;i-tW)+2Iizx9lxhd^6v{jWlSzi7jpCAoH9!BT zLD&5cDxfsa%CfZC4awB^-TI-`dK@=kLr3NOTy|`nRNS`frTuOQ3-B9JqHE895^Q((%PI4OQv2^gXCaP-E3n4rAK#>=9^#RJ3#fue%uPoSf47@xM%wnBz&udh zO^?K%f9@K2iv4tY+yPcuir6RFeAVagFvEn=Py7i7szR|f&znB=V6DO(ll#GKT?)DR*xsOxB*tc3Y2aoOqg;fKkfEn$H|{LU z!l#!~R;IB1v<)LSWLwE;&TknIrTw2|J=xyOfXo_dD@$DzKd*+c&(+N_$H1m)1Rgc* zk-QX8u+18P{qHa}%y z`7h~+^(g$RSV~8{<<%|!(gyY8&iji()e*ml`o$@hUor)19DZpZ*m#So-*mG+NrvV^ zfHgje?d*a5)5P>S0|#P8CQ->j^qTtMt*6G>77!)LPfyTe4zmw4s%-?jXzGZhHa6ow zrXu{ioqAvZO_e=>&irWx;G-cpscp#Uf1}~A83E|IjS;=SFQ@<9)aiqh&UhIzi)a3B zbpHe`nUBfADg);pO^E;NXa77Mbq1V7Zie;z^NYV-`S9*`C)iyNm!I8@WgeW=A}#-m zQNI}VOQQaZs<$`xn!6!M6os`O+}oyFaO6ei(FuV-gTk#I{jjK+E9{STs=cZ#{9ifD zvG6FP4_jIG3mG~T*A|+2@K@eD1BS(#aTJ!6)Q`qlg@MF^y1BWzfN$Tkz2G%JCJ2M$ zD8W5*bPtW%ndES&hK9y**lsa@^_)?7~UKX zx9xp?rv?-|@eAKzB<~*F#xC;+c(@Ud(~G=0gBb4^ccmbr*@3RxQd z+K<7=h~5{r4vA*I?dm<^h1l6cTTWx0Ym9Q7_d`?|{C6*@P@M`VqK)%nRD758!$lE! zN`0CsXIai*`44M0_2pKtT4RK)8G|HWS+kOgRDO;fufT+gaZDIjazH>f=V>s+2pffJ7IqIFULVN(2CEjl0r#YdRbZo4lh)VSWw^i(`j5azp-+T z6`Om6DQD>7eN>s3O#)8+w7u1>JxbWjbB(=y34o&kh7ExU>WI1(W#6S}PezaG#~ZBL zVkb@jCZZJj#XV4qJ%q7hOsSYj6-9f5^P=)bNCnKMD;?%JQZG65x*=dFKVZ+v{-0KEKA5zuW z@{$lQsH=(;R{eV6%-S4Rc_d!dy7UAyIRM68`*#dXOPOpl5fX}nNG_?m#bh?g=Y*=ngi;c9ZJrlz^!q_wg0`JXy!`@6FZAWDZmm7%3wW}{Q9^=Cum zgq9~6VG#tm@yqCVjvadsqdeLwo(z9zW{&G%bRxx~1(O73F0Kt5e zp~hnUU<GXl&usy*oZ? z0s5bHULv%z4=p9GT%>Z}D{7ki{9$aBkb5*fdpK(8Ggsp7wYX2NVb7PW3io7&Z15sq&OZAy3{I|EWFn@a+?mF&)d6mEp;EwAv=T+b^)qv%bzMkY?gp6?|gp+1qly{35q- z9*Ia{<;BaimO2o#L#q+Ctmkrx);J9E$Ej<1d?-Yr!(%U^{%vpT+ff2MV#tZo z)5Ll1a`_dllq*bPMYTEmBdZBS{;>!w?)YLK$+ByV+wYkmo1ntmK8JN;gIB*8){Y{C z>qr<}aP5K{8#8Fl$A4`lEf1Du0<(Iku?I1Sc#h62)80ni z1%}P;Ll(cZIqEvcR^4hZq>bKUqqR-mri_}ClQg|%#ffSsm&Oof6M#|9?gX`*8R2Y- zl=kyso*p;Nd`htN?xFB-p4vf2!GXct;+;o|fK5w_)bOkT%?9I*m-B7yzp~Q=QxWwG zV!`tmODbk=G`yC!6;!`QGVR4x%i0)4Y^LpW?+h7J7rfmG@R4*eqe@zDHNh^m$W2$> zSUjdWlJiC`sq|^!7qJu0)4!jdOxC-9Gjc#4Git0GTAjj~4X=9r#BQKFv$nM4UClji zyDuMi)_N33niX6~Es{Sns4W4OIdkmdkmQzZtD=A`jVFq3)cS?0#euB8kOrwZ9=lBt zme&UWM!~0G>L1hU^ZHrvLhmdY<8n2(BmQD5I!ZS6&Es+%ND`rNS6A%r`N=_gs-r}F8>%N&iB051t7-`>W*I?&<^Ndp^FUSY=(S9$iryuo={ zsDy;XN$_qrqKb?8;LFj)ZT!WRLc9650qoETsKGYDyjmY%;lIkfdIap}>xac8ryTWG z6Rkfz{Ra=Bfq{V+iC|b+n5Xs4oq?|XjvsDgsmyx;uUB*bo7~(EIF*xyW~tsSxFlaW z`(z|JZoMLZzZ5`){KwPa5SCTmb|)h4uSNH_jcmL$G)`LAPe_c}j>+Bp{NWC-xt!a8 z5l|B*=vM{Rpq;ErB{vmP%TA{+ZTjNCyxm85b(vo8*Q221-ik#A9(XqY+kA8;iVrno zfybM#N{oY-RUTj5ryO8(Qf?jGzs^xuNMW38HmRpvOl>OK(8D}?18VpPFun4&0Y@M# zk2old?W$(ZqCesW0~o5qt~lro8XrLqIbRf<+}qph^1_T6GiV<>5#T+164}L1GcjYt zaR_u#$G;WUUlsjiA@;WVu}&xy1K=CluKpk8&eiQ(?(dUMe=-+i+RHPB-n>A)FQ`+kz9-+ce?BMI zBQT9#s_I{0w_2={%Z$&UBal{~`Bwb}0B(b5gy{U&YWtTEHlb_KlMFIvu+{xQet^OL zW=j?d+!}S0OXU^Q{DC+L&ngOw-yJOqJHEv~_+9LnZ+^bJL5M}NbfE9B#O-xT_lE$P zSaEa$%{8tPDIMiScMw0G;1Q>QoMRIkL8^)p-Z$c*s=PWC6&B#w62Ire9WJR{V7dX3 z3M|_bq%)3ux#Lq$v>jt`_2yX%>#Hny=dZl^P&|~^=kyKJI>mD>7tHb11`X$?Al$Pu zb3ObW?0TSQvk11hLTAHp2pFQCQ(GErK!wr?plUlG&&Bxk1lWA$^NjC2{TWG;2rFY$ zaMd&#XN3_KZVr;x;N9r@BHw-T?ydq~b3YK6h$i#we$La_292jCc*@pS<$^ z_goh__o3aBt+dM^2^NqwJ%M8kFOSCMN)#vhOla7G-iBJA^|jsKegv!@AK0WokBR7s ztv?z)z;kdf_pNx1CE&`HVkMCZavqV(cbZgqqt@m+PR|dMB0*+3aGv&v!iFv5`&9-JPBU z3F~{|lJpCebMM0~ja4Y^A1+c$h~rNdfEogD#k@Ls`)uq-isRp* z@BYic&PK#!^iU)WLIgbAGnJ~Bkh~?}x7848>$81JL$rwHPxnYspd)eNGCnC#54`%S zG@FbYS9X7k0{HX@V>~HkwPcTzi%Xb}B>IRu`fJlkyE=-S$834G zIU)3Br0Z{%JX2FscR|NaeOppsZWbO+wXyw%f``0xi1TDy-q7l^ncco5o3&DOyJb)w zdJ%qx-265lY>iJ;+~6}Mq-G}sg1;Qne8Gm7rBZYjl2qX!^2x{3Y4m!ov`j0qW%S8N z+QpxES$9zX)8!sK1(`Yr)X{G*0{;)EqA0*z{vzrZr-0M{OQs2~gZqT#^9A zJ-7vV)9aS!`M&FsKkv_XAIHv~?7e0tduGi!Yn^K);+^VSA_6J`3=9k+1$kL@3=G^Q zbQz9^i~bCvw{*q8cpzmhBlAu{Muz^Kvx9}TojC@Ed_=PLeI1Pfifny#1uIPKyw3~R zkJvEtKL5g!=l&EOhmFVF`|4fQ_b7sL1KEzy^2itNuXe+;-c+6UeEL)kpx7M*KmO%r zbh+TW2){&1S>C#=!_$3`7`aZLSmRf7GcZ;okMpaZV-E8Uuqmrk(qkCBx@Q)6aN$Mj z?&^viXr6oHbb-$T3RVJA z;m_s*iEVYP)5m}5TxJ?CfXvkhQcbD4C6;UZUxe2^zojRAvHM)ED82d@cD{s7H*}>V z6H_8Q3#YZiF+7_7A;pHQ+>eV|*Y`Y>CXKIu6@BvsiFaR-a6ewh=_#@kRAlWb_WgRg zrT$RACuc>0+<&c&w!fNemi=ZX{dKt_od);t7qw4m1XboiFL+5!+`_Yts#iKc2!1Ev z7m2?osIFTo?ynSX8nPURkuRwg2Ox_1FqgWzPBcsjOJiG>D?Vtu9=>>I?q5kEsrE=J zlh9rFg$3(_O}yh?*$2aK=pQ^#AEntLjU-#v6s2Rj#H<>u68?}tR#x3N5+d^4dEuGK zOzS7%DIIZMI)Hq>WHw+RHb%QQg3{u!^IP)a($FjHrNCDm*nBz@9ZYdJvi1^05(lbt zWDh)F5%2|)2p6XXJyM9IauN3-%%51}9^G__5o<|)5xTGCb=SUee-`YMzQ#tl|tnR#y^`+GBEsfJkS7&A_Xl#lXCq zG(tF)P4jQeMqcG!zuK_G3B4GUPJxY|-MNYvSG9`%JbM=6F3h zCa7b1Hy^hKX>zaffZDr&ZF1SAaXVA;%D6HAgJG%(3?lV)%s>E5ZUA$?Gv{|Rp zN47mNgNIAo-rsZ#?k$**ADYR~j7eP+iR*f?9B0~QwPdxwKpll>tRbyB(wWj($j!~x z`R5Lz=*^sad<%`0DJ?0rF;*8T*LHSPzAU!{ZPxj3tlwY;LJx0mZ@U(oDpxRePXl3h zzdS}$aHME#FfzruQxyo4o&nlPmh@lW+ZVX^5sNdBD>IhR@%};w9d-~83C^-KUmC`D z=;s|EK|$ctkAN4r`5&L^5V3dQRz9`6-_jxZ3nL~lap^%L9w2yU2>_JHTEf}HL3UVT zOUvWY7}Mj1vzdel(*y7b58r%>!(*aQkW-I;Ap501?rFbx-ecV$O?q$~l^M4)o=8OI z+xd7{zf2yrSG4bI-lWj&+}(@nEllB`&O^w5}%E2~~?wk-Fz znX!i_IP+nc%0v@g!#r=lgg8~c;KWht%bVpa#pdWLt`aT9m)t1Pxj!B3*Bv2z&vH4& z5hrXJUzjfE3$&ARh5W+3mDb^B#WL}s;Wc%TDi4D&Nh=<(^T}&mV}h?~@8o6`h2@2b z&ed6e)5Ob9GR~7oEEy`ZSUf5uqa>Soz42qK~dsKP^p|Cb4y(C0GkVfoJW)&6@{E}{VPbz*DUHNcxdBB$-Mp^Vg{!oWp2^;JkKR7$%y(~OzOa^#pC=S2xNC`jowuGxm0#2?TvSjLn8lf;o@F3vFA8a% zZ!T|+Xx4D!aEsk`-!%4~_69_j6 zi)8@@cZrx2^!@a(k4@6SKmBCjzZQ6U9V35xcTI#!$hd`i6R#0-J#BW_q+y)wUwAQX z)jVeLUF*A~DcK5%X~d!4ibk&li#N+Hi!Muva!k4_m`~YH|z zoaMasypi)O8aR#Y!<}dIVUI$+9+DB4(7I2=ulKuiI+nWmZsu(~{;jz>xpd;(xOR~A zdnhc2=5}OmM5NCv>KO58p0QKzakLfH;e5|R2vU8^*`CSI?4W;+k@{_$vzsG%bGZj( zt=ExyufC&fZm*;ztYs213W0TvGfsUUTwzfPBOlQl7Tr)dcyN%p=hc1uJ+(!;M!`m*ZZ0$L{)l(4 ztvol6eu06efE@j9!GJnJi1FBdK z{nS5>N9M)wcJ&QEbK#xi(M^J4n#-K*Cpuu6A|HI$hQ#tC^>*$%cg!r?&|9z7i-7RB@dYdFvUI%Qk+e?icj@(JcOl~RpQcUpyV#KQBVGg;wY9{KXRuuA` ze=+ZAtI=O@WwWzgJ<>gb1I_^{z@5*8rX61D6E&oOZkknE%S?TC&r+yjGcsu8#8Rfe z)m6{$XRWv!E|52tRE<`FCO{BGzE213T=s^K#`(f5Wm|=plx+Y4v*g}~`%mvHZO30( zpF)VziD;q+qW8i&AHEdw5~2~`b@hZjs$ooi>lO=;kdqL+)44nBT=8RNV)a(F$n=Ga zS6%KG@fMcXb=THrJ(Tj>S3XGBu$;4u%{dnFSZ$rTl%dq2qyjV>l{&cDN?f$SkoRG5 z*fzHp(57|QtEtFO6II3GudQlG)at!Xbw1y-6F&oUV1hTgxxhza^grnvG`N7;n>8Fa zju9f~c;|}ywhP=oA5U_xh+CxGJsLfp`8~O)Jbe%U-V%EmYa~A4bCmCLt8q0^ZCL=d zX$CgOxZd|L@|QgG+8`W=f=cuR*xcxzj-RaswO`-a?9oiXE)}ne?@pJ6a2@I)ZhkO0 zCE|R^V7~>5RFZt+J9`^{dbnGdrv{Z&yL^46emk{qJJE39aDL^2LNSZ8sU30z{){Xr zym#PDjd5az@#Mf)j;2KXj*Z}(cnv$rg_@3+oaSQ>jP0d6A7$d}mL~j;Y_5X~oYR-R zm?SFq=@-ay8Q(sf4ywsm2C{L>dOnnK>Ls9><8%Xk>rm{b@z=n^LSl*FX|cUcbe1;)LqG(cdH(tXy3kMY*{>JUjp%`~U}MOKx5f5fN@4K5jlfF0=%fixn?hrB?-j2 z|2=7v1eRkKXvY@gnYFB%Ci;rjvcC=x`tt?){BuR0N-DXr^k}(P7z(n|nx2@uSvYX= z9$4rG&OKS2EdZPT<0^eS-b(To!uKTcaUk-Rp6>nb7%1qCvij>+ROC0`eO_#<`V`sJq9M;D-0}>!2hmz^g0(C6Ttn<8vB0` z`e!ScnApM~?EksFKccJg7+64B$8H!fBzo;YUEYm z3hjUUvi}kad_{{--?8?HMMe658k>KJW5)RW??PB!Oo14wZ_@;D{+IDaw}~yh|Gx|U zzd`(;_3{7uAZj&X2$^pS93RgfDy-@|ERNe%N?+z(yFNXpt^na0h3fZaXX<>U}@2UrArp%J>HTz$et)l@VV60B&TjQ)6W8%4s5+(hQRL<}9W#)yQwc_jJ!g`mA}U+!i3bGnpB%(FZf5{@c`4 z7Ch2f!-DUfi+}9BKF}a7>>?KF*sa+BDO?uu`RBZZ!5w!H77N#x3k+^$7&DjSgrq16 z&g-t$gCkpuEVpUizGq%m3_0X@rPkO0VZEI0_2O;mPs4bX#qPiXA^@tnzRxR7Q3Jq1-vp^WjBHu)j=sROaj8 zE4LAI$}H*ueXw8~>Xy}|FZ@13D*X21yGqGkeZ;Euij0`7jxyFD4K z^zZ0Ajf+*H83-10Xxn=EG0h{cRBx)d07XtyZF1XwcLP14^A1ku_W%tZiPg`KW_s?% zuP+AN@gv*rzIv_IHDHx|wXdJ=%pAMDNIIRx40_EPH-cd*Xijh=S|ib}CV4j~ezKVu zu?(|ssS`QvHYJ6Xm@w-vO!$LA*}iAYwm9Md;W7^Hmj20VEz!fz7?i9D?R$18_JFs_ z2vRJdjo{qwm=)LDZ!v`uE%D2jrM!^i~Ca`9{;*AOA^C-`rK{`Y>> zqx>0I_Hb|W@{*7Q59xd%KJ8Wu#5>wfDy6taE#^>1%}pV3TN|FC1IUj9C-+mfAU`_YxY zJ3|NT(U2f^qijQ2b+X-+&4;W9Simg>wH#jlO5Yn~`&!DkEjbDI`g>Fz|f<%lX92($G4hb9@$@0t{);TPhR;0kR?WrPrQ8!^JxEcQ6VE=_T zR*7<;26nXMy^81uNqt+Oh7qmx$;6mwOIG`UlP${i-+ehTtJ=jnW6!3}K2js|tK;(x zAfk1{0@OwuDPqXzB$>ddrH`gWOn+v%5ajoT^G#Rn?bQzWxnD~d_)U%P`AzID#$r^& zPW;Ps*|HB=)xRaKQ(A^E@W~hpLoq0SRw)Wrd--b(M9c=a8k@2;RS~)Yb#3+_edQ(F zkEdr>E@YVvhUKV1Pkm!pp*vXdsuf6ON zo=}7p&#L&f7#Xu`&eFGWRNtKW(4`gT1c)RQUN}W&t*|WeK3W34F0F{bu6{&f0Zq5X z5f)N~H(MkslFxJxsd$avT_4A8@o$Q*JgluKztEjK7&ZU(lro8}#V^Fe?D)F$K)TxQ z)%ihY)0l7p*iL<{v|+zlx4zPyBGGoM#G42YuGoSdBT-y7*W^}0uQVOP9pjBa&~yiW zXjN{?1A@ornL5mm0i*TE8Fx(3?Wd6SGLJ=}tC6hHae*uh*6A6zXG{9Y&MNj9mym<; zsN`N=w;;#gjF%u2mdZiwMjtxwVc*NshNrp()US*x4`Mrt-)eDFJ$6hqY1yqq0i>s| zb}B4yZ&7rN+R08zO$A`VhbCP`OSo%8>CS6|O(WPl)zjdYw)N@a=EdnCgtivkiczD* zT}1~?X6n-4f_U)jz(@DI0M9^b2yd0n)v_`-fNx2;b22MGmH%H2&vx8C#7s8ws3 zn`NQ1FhD^;Vd+z8d0pJ_ zWNJk3ZX}Q`o>?`K$95$X#D_&J&%)>C*Rz zaP2h}_K=X_|6pz&OohZ*_k(hX_RH@H>8~sU_O*$s2Mj^Db~^a)vrM2 zkvuZTs8^MzeI(7;MzexKTxuUe+LZg4n!~vrl>v-DLoa3DKILQ~hucpbBZto*)VpHi zX$ea=tqS>y1zr=q3g=*0O!nA?1i3%I=BS55^QlHz*m=G~dv+9?Mga~|w#TaY!}ri4 z({_hscF-pb4SMmg&YeB+)XM6@$QxIlQ7y{aEb7F^W6g?b6y zY}5Wyu_OVCh0WgLKDuu-cxv!NGHIuaCuEVA_wF+F@Qhk?#LjMUE1G(mctM3db>HAw z{R+U1%bWYoW{#^?E3wXYijRnTlE}X#NiL8hbVJGa1huNik;!%Iaf!rB1-E1?1p4fa zcnlL(a8h=i`kri#kD>IZ>{H$T=KZ|lCDCEV%EWc_IitxqkvdYL65?zBLQVJk$IoMs zI|)L_cgnWoWzfe*MN4x(Oz*!dPzJBhk&W0iou;hUCdNjY`J)iT>tn1r(K=I~aURQ9 zz{&)L`s!_Mds_0bbQ;x$(%Yfchd)AFbk{3`&l~sy3ezguZvYAH#vud+wuMpADe3CI zXO*%Q$D?)*w4SR9$B1(&Ic%xA5BbjKF1Nac*snk1_V0Tw#rN;7n~T2xf!vH8jXfBrb_Ogf9{4VTMrXU={C6BSt3#R^;Su- z+ctqtsWcvy=PKaO7~gz_%o*vs5sqjQqH^WaIA!en=|a(p!t8(n&TxDcQGdprw!KNR zSOu^4AH$YL{`Ry1+n0Gko)TjE4o&GImly69aS$Pkk?gW%>G|6CGA;v`mH^nG|5r>; zqV6z-n_NqFW$xDN46hNrpN(6OeQse6Ep=dWoe7)JcJLX@_Izp1?O5{SqGk!Pv>uYu zu;oeIemVMz$z$ho`qWNd_T3S_qhGTdA0wU5ysw0}@;HBcHaTIe)YUhBz0+?vn8*(_|iaaYM*uL={ejk=z)-cPd&JBwONT8@Ay7K=>LnJmBtIGjn zI*U{hta_k@Z%uQ({yANF5%dP$XEi2wQpQ^J7DW3KTxBmZUOdx}e|0ktM$RLLS+!Pm zNnkC$8n5A)sD1OZS+_!_=lkOT3C459i}HY9{zL6|ha2H|#GF(JtPnOWdB>xt2A+3a zv0!NXHH*{yRZNz7Kn>grDH9}|HEXQ>(5|LIg0>twsTn^I!=Yz0 zwPnp1DYcQB+0{UXYiL3FqJkP%6)$i1Ku5P12?a6 zj4XW?Af5Y7S$xO2QW=yz?@QlTZ=3cq^1IdI3AZuI`lkSM#N=UTO(G}eOTxQ#v3(V) zRf3(K)M~7j>*Zt8dq>dL1I2{oZ`=J6X`iu}brXnM*Ly#n`Nmc4s-d^jhs&O8Cc0;v z8qjoN$*W2G-WLT+lut_mYXl1460pB`Cg99S5KEd^=-bu}7co6?RAJ7vUOJkie2Y__ z$=6`c*~cD~U!XQQB=@RcQ?9eP`<~EVvxcqDU6k?BlCO?x3VD`TW{!Z1fmSGEnNIG{ zIoG1Lx`$(dP1HH6oeTLQz2*sZrVWz848e1#JxG|yZmwWN*U<9Gbj!YyovnG;UOkcQ zDSbLz>9>moxXJ_H>0l)I1nO%&E^nEwct z=%Y#gAvmgsT(B(r<(D!L@p?$h9oFPVkEM7$z+7g72|zS$`P?r55c>4Q2h@Z%rU+Ro z8olO9cW&J8`uq6()!_B$^G{KbNq9}Bll0y zvL~7f({iGBp6t;Wg#&z~$JHvN&tWwF^4~nmk*{+$up{QH8>0$W=4n8z6}5^ez{8zQ zzoVt^YI5BW$DEc)SsH{A2#rNO7rxAdj?99BYk(lzolP@Vvg;IwDQ9X*jqzsqn-($$ zf%si@6f6<@@7{{&_X{cxf5nobZr# z=$G~v%)2dwu*x&9>5(~+6fb2-ZL9hZ9X`^NS&G<83TR?NxULz zlH{nss9OIt0>o)lYrh!mSLdE_oi1WcFRQnDq&akIut~P$+n4%oy23KGgI`(@+EDt0 zE*W>_2sffIpg%kqs1{M9~2Q*2L^6Ly4jt(z2r+5DP4RGp!oLs{o@PitMU zjQi_y-y8HO#IGeKm<%Xqii8OHKn+pM^6Kb_!Y90h6xYW7*SetQZzfUuG(8VpOz9D>C_FTK=%i5i zd@MhPMxJ=NRfi*uS?N}Z{f`-jdqcW^VZQ6>^4*@>AGlXxSf=L))U^Y+npB6ly;N!q zu0uCMV9wu>R-`P7sd_WfYMuD8;}(W&Q*(|gNVAlA=8X)a{mG zhj@jM _pbkH-fS=}@8**(x!%2OfVRSW+1FE+ zC3+^R-qyWr*{_um7h7{j|G?n95ip3VTPC4Sn%%M1V;9*9(l6ld2-5eq>1UeR=4B|>KJYWc79pmG|?!!8Uk}b3+ z4{8r)36DD8+Aw0q|Fn;inDH>K$AW83+i&Z-V}+_0Z0L8|(V-wmehv)_&c{pVE00lz z$vzH7k%W#tSH4>fXHnk`iLYksy3Y+91hIhCU`JD~ezhu!RTaOqf#82rUjHHr9uoy6 z13Dy*x@*>DsDp=8Ml$TEbVS(H_8*zgv828N+4vk_YJHg-8Br`nTcme5``4sT-QSdN zFomNRS#AT|s{%}pXJc6_C7E!3umaTy+B=rd{!HBSK5;fT&r_<)ZkTN*LoPTeBbm0> zNr}0mD3ddkYJT>47ZXib*K-jC9JUF(8`eMrXr98g-)2>zrbGf9bG|=DfcjA9wRdmm$&PU>m5PzBu^wO^ zp4Z49DfNBTM=HepygnH5c2?Sfxv9|a{ARPG%~mBbqGA4t|B!)&ukxlNLbX^q%@-^4 z^YT!sympJ$NN(EaJzIxnMupF4m%#vw@{_TCdI}niny9C8-4lC>f>{I{%TNYL9wCLi z9s9CW=w@hmef>2l!w^0ekqd23TPl4?`RllPY772inWMJw|AzC$SFp= zMSMnqE-UuuvuTzt_B3PFiJqlO?|bYXwBrmOtIYpG$spr-F0q9Xbj}|7B^J@SX)OH@ zn)Xwes$kJk#WkDWrx;=bFQ{Oe$Z;1{DYi%7wJ6%W{aMt~sj^?2q?cf99KYgO+Ae<- zuf;1;QG;YcI9xp(+^#W%%u6e`b_eY#qS<4#ntZMnJUd@tK&f60_6j5|{fNL)OwV?R z7&}xjx-!i29_I~$ZRb8hZ7|wS-Bk7!!^oGW>o=6_d>K4|Ru7*xD=oI{|GEd8T}&QZ z+yl`}KmS`OwCB~wd;-Vnr0oR{`SD!1LLaTUUV}~g(gw9aiB?_YmPJ{@^1bciBnQm$ zyI<4^%SCR@tG*f)puKJpo3%Qj#h2WzM|w=i)fE_yQ+`sqVxWuZl`Oku({HjRpcB2_ z`u1&C0`8vs>_Ts=Xt{w7laAq(Pwx8%TtcVS6cPJNg}4{~cYf|9N=Rh;UDbX_y&&2Z zocNMn^XbCx{Z2h8zpo=Qye?(1&wx0@zEKJPMV1)wV_AN2dY~T{?}4kAzN~dHsEf5K}xJv6)ft31e^8D9uE$Y zIxYD!qnR_T&QmPgxb`{=LS~utX zTx)H??~ak+4-s|cj@a6S-H?;O)MaMS#IP9Jq?MvFw2 zt)O4Nb5R0k)_z-n**K`jBatMx(J-27l2Eq#a(cVoCRrt6A;Q>vG1s*K8L}`!EPmZ6 z>Dw&)mOX+V?&{ zZds2HR=uU6=*B;$=#Oi%8e2T`5taz=n(fg075cidc+2Y@e4;U&?W@tJfIlFCY&yG{ zc5GjZq@Jl+jH{e8A0nN|U%`n#9)GAixZv9tuA^&e`a6qM=YV-nKK{20W%+q-DVpFs zEGcrMHtT{ZSrK0gRN%}Qjd|ql0@9z$39l*@Sc~!LCC49@D*N7V%kNL{@I;)z*AGc9N_Z(OU(91^XZUZWBd%9wE39q8@larQ+QY4>;H?VnxrOdCyP6`%BZt zZ5X0d9p6#=WRvpI*+hh=6Q&M@toMEESCClpC3WWH#Qu86-#G<$zcz1{o5eM5lv_Wr zz@#mcPw^ew&coFnI2)K2mnjc$cO<8Mz42kfmTMr&(TD!pGjBtl`!*Er^#1|==!zj5 zA8UBWMfMtu^gdI!!fry8q=$3Yk$O7U61@6%GiQ=f01D#BUp4tTc&L;)SvHO*+IwX$ z&mpBW?5q|LxUz*wy*T8?TWtFaB@W9&HZB_{3*ts#c$9@&0r}8@ebPk#_jlLt_u^3_wp)jba7QB~PsKMEv?rCA5!pb7xBQ z`&^q>TjjGHzt05PCUDldkS{&vSE1P3N`O1tbcWv!FpOl?u1pV|{c$MxRCxN)>c~kB z6@ftfw7^Y;Pf7#q0=ii}OI8u#|`&~$$yN|E$$c_O)K_-Mrh z?6Ux{o0K>u<1J!`Bn-ssYVWBQa#1@vpRZ{ICK7&i4L;a+^4lrj;myG)lpZ#AC*jw% zIjmcU9-}AViL>(!Q~a<>b*Pn9886&5I(k58Yi0nH;+CGq*s#6tlHP1QM#Hm&u^-KBQm*V82FKU4<2K)f2 zYTh;!vRIWyIePm;unFY89!Ir^3!HsDQi&VSWbF;#UZ}}WBFJruyu0Wjr@G@@uezKc zLkl|eKdk3k8@n1p-6pNs14?v*e^-;Frr?jPIIDmN0ZT^t)g*W;-ehneV}LOv%L z(=0W451dB8G?hw~Cw!GGyl(N_u>V8g1Y{Bdr9XG<^?S%Dwi^mBbKl_?c1G6CgQ#s! zuW~NOpJr<3a8>VXwcj~4MGZ^1BSajn&l96rpBA901Y8PNbUq4UP4#x7-Tka7D_~$E z4`+~YwyzTa5p?D0wOEhdg6v9yAmvBoLTyhJzn+@?1Jnoy*DntZ8xxBZ7(Kpt_FrGP zOdd_KCDgJL)&Wut(11+nvChyj;8qgzG$2f|TjV&2tc;_B!nh_PA3D!k#UGUnpy5__ zwoWU~KxSnWmvRT7F};DJ@$H3svM6+;_kka%*mK$Q!JUBILslPRwbLB?EsNJClwcH!l3pY0tB5G?T6Dm=`N8wE zX<+-an52=v&}%u7j)4dktqg5I`b;63#ZqC%o(aE1yS*tK6FEA{qM_Q>IKmKq?)Kt_ zsPSGNkQG4WZdjg{`d|Lqzk}LACXAKi*up(BUOR|$?KEvzFfk8mzfM1oCTGcMR@(jx zT%u_}NSG$}^fwvyWUr|Cj1yR~sjbOPKS?!Ei5$#s9-K`hiMhXBp86ITU6kSQZirm5 zvw$DpkK zzA{YgjUMjV$a?g99R&A;Nvx+jt&r)G+kIhhYX=U5D1hYd>R*3bFfdjJwt5|Jyai6y zTTM@%EwUNMe@u^PAh^ti{*`x9tjjo_BDjnNY;s;3UXny43|=lcU6L%By=RW^Of~de zbE76c&guscRodoC{N+JSC|FKt$44nueQoFyT8!b%@xM(bw+|RqOo}_;z{> z+SzpymD1C;F(Qszsx1SKXTJTgRrYuttUuiV_-Y7jhyy)HrQGHrTLrFUVvh?o0sII}DooYWjo~ax9jrl>0_v~{dzit@ zxl$|QqI|B6&<_iaru*r~k)6+}3f%KR-y@ec*}CNQyMtY-1u_gN#33;~WNa8?#@S)i5y3+S8D2wls; zoeI^lB|A?W{jE`c#ksUSFP$~&?r8Z&y$z=k?aWGq>#t-d+P>caPjT(!)`vYKd2}Q2 zqR7)`NF$VAI*hThjzw*?+Rv9-w0p^6trl z$tqBeR6X`9?7_cMh0y6G@uI&&9UPA*$?SI__6-CkU2JlhU_p50r!Bt^tjLd54jmQV zb;WX@wA1*Ac#44qd}yAp?Wel-EM!D4R6fvAgPJ3h&gRXbr}xoZ>TnA(;K{(~hd__l>#HJXZMSRRPbv|gvVWZ|4laX%yo3`B-zJ)u2e}lwF8At5_A`N z?)ZMQhBhVF$8BbaRKfKo zKY_bP8@tEc?Yxn(n*|plJ?Q=LJP8c-9Aqj*0WaCLthlK;1a4d2Jn%lDd6sUMyOOLG z=mK@yuH8DzrlW)(oA#;Ul}00Oeqa65$$bj4xB;BtJ6h(rJCkVhj_KGjPaOMgK#1pN zOQmFO@*|~pY6M(y8cqD#7#vNg^U0S^O3!FLRr2bRjZdjWF#Ze7!=kTLM|-8CHgSU4 zOmaQ zXPEaH4Ha9V?HB!yBB`}bwte-nzVA+qJ2CxwyaR7^$_H3Gtk)s3>5aBUec9d%qf2*8 z&Uiu;1BuW!t2f_b8BlAuA3&l*$@)?WcZhRm!t+;XDER96+#^B)MWjH5ai(iGn6NTW z4dRciTVd(>B?7rQj3`ATAv>C;mBbOI2+!6Ei?7sTe zN<%`oT17$_%lHUwI~)JD7_W9V=lG(7sp|KutGNVN$#W%%$FaGv3;Q1B#;Egrm>>nM zl5yM*H8vhndUAaskPd_yx7bI2ekl}Tx+py}r{gsy*?RJ+Cb~56=r51gKuOSV&Ctd7 z28?VI(kQ&OD1crghAZ#ZhqG(}wTn@rb*cK1HkwioLGEa8Mnj44OY*iAK77Y& zs;Bzj>_lSQX zw6*E5Jf-t!PvgcMwPY)Ei8HU`EJduE=HH-KG^yI-RMqpA+pPvwMigvalU1%_AuEwZ zMUYr`;fQmo;NE!+MKubU*gyE!7VSHeW;I)G=H2E;v4eK(g=hf697miAjk{tO%}bZ= zFtjCvR@pyZRjO7X*kmys9Rhx44U*dVy2P_n6!>4nye$5pWDA-CDO*#UWGE=7$=pT9 zW~_B>eyJZeEFZq<^wcw6vqSH1!)_6lq;HwafvI|wbSq3#Ev7})4XMj^WAgX@m6dT7>x!_e$u~9=z z)9$hhS+G~b;&E5TMIsirdF`?XCZ6A9q@fx|4!etwVB`L z{d^yWGn8AW!jYNGj&;?qXIwc{wW=6$fl}<`pD-*-PrhAzKp|T!9aTxY_S^1^`6VeV z)t3uD$DWpL7AGN55{rB06WTvR@-E0I13r|ZGYx-YwCVSp->^Bb+CJEZh}RpUBM(@C zrAmHC1)8@hXQI5+$|R~v^HLK_nca0X>r5w}NBa(MdW3+{UX}n!4!qTM2c5{B8NGICohM(z z0>*78?i`M2R~4Mm|0-~cq6gt;a$~CY(fA@p6nM8>0~8*VYkL%ly-|7H+@~6fPSksG zZwmS*a02xX+2a|h7BlXjGql_Yp0ryGk;EJo5*d`JP==lheQ$3-R>gyoKN?VYvpz%r z^-^M5HZN@Vd^Y+Km6jt;qb3>Y^vMQCtJY$5Emo#KH{!UXE1pG7v^>iia#J5K{$C}^ z%F!dyc}PsF7A*VzJN;b``F!@;l;r%|Fe!;J&bAgrc*P|3F-JEF*u*h)-@ZiWEbxR2 z+MZ?7_bS0vl5tbW`BhWAoU14)iC04v_q)e+JCfyqcu-#8)#qUH2+Y(09P=5&dkMH) z<_5VJZ>VN(w!z5Z(z%+hi6~B#&WvHT**kA^{_&U3*`A{7{aw3D2h$d4`X~58Z{J@R z#!8Y$DjAB+P63@6>9|im05sBl2^>Wi zxI7jt{jV=6%hod$kH559aYY&Wt?czzt-e-g&?-h}5F^tl-$M2gG|dmDRCA`3b=V~1 zo~>IhmpMz$SSsxd^;$(Zz1i<1SxU*y0>VBN*H+Q&=Iqy*MYjkHpZ_rgs<%q=wfFB( zy|6pWY#(dQan+_&yRp`YB7+(W67JmGI4Wj${tMsZSwwIdqLGZ6XIgJp2N znY3Bg_u|lU_-&4Vwx+?iuM1HtrMaS01gnR6`1l!V?|uxho@@e7RSlp7bCQ&^q@3Jn z-@4x{L3xJN)=KxXH<5EBzWUSe(PJenrLzcn}I&#WCd;f{_4HwzlEqp8FiCO6Z8tC%;me^>H%d)E4k=0L8X#1OVxO`ippVP2))<;LEJ_YKR)R2(q)mXPb z*Qdu5QuKs#$(0Me6yd&yQ$oMZ*J{N#PR=Z%Mlx`;9IE}JqB2zWULis@vS}M%CMuPP z$%&^;wk>-Z7{4j~MfIX6kd-kDV}p%$K=kZKy~RwHd-&=|O-i71YV9w*o7IG6KwoFG zQYV&BcEfM20b$bIVPb(U{t1=os5a= zIfp1y5C1Sa`Y*Fy1F^eecTbO3(QP$!S)ut!?`QTlln7sjhFe^W1k(8pZN(pwmBGea zI=_i<%V6%G>(<=+`IA*X-$wJB?aUB&HNayipS^nic#YzHqJ}TobM-7B*zZgrjJDCJ z7_s0B5i+x~SS-4lB z5}6(!HB_i+uUV$v_dqt6t7W?9IJzGN@DeT^|&Lc({f;X3!?r z{?6~50A9n)4ofz9v?TCOY%>F_-uqL}dxSZNqc^?%E!W`c*)IAY5y3h}ho1h+erx8G zZ%UKYn^3TSrW_e$K8IVl>p{5!f6r~JDg=y;yN=UeR{;zu|*|7?~hJE@LmzNX$Pv; zi(SVt327F=MT=F7nJB9xKWsMpCK6U9#u$JMR!d+G{ikQw-_A=*1Djo2dDL6~#DlO? zERGflqv!V zp(|1Z0!r^l7ecR65>ynWgGiGAN|hF*gkDqxgixe+g0v7?=)K&H=bZO`Lp|sGym#E| zzwD8nm9@3vZA$>sji9A?|c~Wl;!aLh>ny)K9;*-UAYWZSyAF_ctw=LPe6r?53~kP6g$2C`s)o zziRYayB?b`<~=0*NGTp2XI`0gzp&oX)K?6Wk~*Re$gM!EjO#lT5Oxo=y=%P=Jg-d6 zz){m>NyJk>-snvVI?xBFBf~f4<8awmsX$z$H6)Oah#QC=4_4YDk(}QESTHsqMs^$n zkw$Q&?E3B2vxdCs&aKaa>jn2&-XX0h#b@Xi)-z$D*+6)$_$^Y>*)*-P-AI&rIs6X! zZ%W;-az*J8sk?UXQM05x1K4JS@Q6`-;1G9e9%CQX*rnR-PtCJuC0!=7v{B!WN{cjT zjTV5?2k+3+C|Z~ykb)3|V5_bf%dl?)(o2`qN93geUF1?GIVZhx=}?VAiL&2t)Tt{( z-JW8D`8nZq?pt=lCiAr3Xs#k<6f3x_2tDdiLMMyZ8w$_Uy-@KrzveDic|vftp0;1{ zVoqtBX9fjK4xflK1DV11O)^lRFVNc|7HOUfBG<0!+pl;Lq`g+pb?0=jRxi`qY4lT) z6!5i@Q8Fm5p`%-eujRaL2l5};<+3t!xaL$}nU~L}`|})3=*nPzFNfULXPvydt0r&M z9(VCPF}j^flWj1X1ai8x^3q-1zRgB0anI<+%j+oD_^kQU!Hgl6r6Ym`S{;AYpTryD zFM6w!Nb`n>}_pAkM_WbmM@A%$V`5dyJr3kvSd1|i;$v#mPEa+_U?&@tkByHGA~JSWkm zPBGX;0Z{=mUc&P)ZnSjYTfI1SJsxNHG?l;A>fP*-O;>F;tpC}iLr}Hl7n0}(V+Rx0 z`C26G70lUW3P_)9t(XoQ3M@3OMv-i(UOTCfn6x`qOdB#b;iInqFtR_KJlvpHk5vy1i&J4OMHJJe=jf z>qh47nL}FPSU!AJTrW61W|933IoHhXTv3v?3tV(UE-9Y(D}a_GGSqbO`7!?wqWTOj zXh%Vy94AM&K!paZUKHA2%E`*utG6h-8Wq~4y(|wW!3dO(cO$*lTp?y<{6^N_=d-WpQttP7Bg78j=)%exS`@mJsb`|Z6&^noE{ ziK?wXXqj7O+#>@xAHDv0_bnu3OBy0X!v!k(9C8O?0FS$_^+>(n>Q-(7LZrY&-#n1& zDGo>mx899YS%nEqJ4)-hoxX;#ZgecTUu(aT94{AFm?7@`Nq7-_A62l>n`@mr;$7bP zRKl#et&C;tg~%dWBSoS;q?RdBk!rr6_o!Dyd}N&9KHEWc=YVRFKwNuFM=2b0gUwXy zT@59dvHF3h3e7gOo?5qeujEo{;fGF@lGhLTzy$MGun~mUQ@LF1S#wC-Z=%TTANtR@ zG=4KMmbWKPB(3Nb8!Aa7l_{gA;3&v^nq<}G>0iz9il5#8u(Te1PJ*+!LQNu$+}Ur7 z?Lu&mw{dlYRuz$FXXNG#DWLZ;s>S60@atd0OOzayoY z=>HZmRhmxf|2e|=UJ$)YSGU7Vj2y3?Pe^cSwP$9|pj~m?j?KpQY+$4;RAK{?o$63F z$Y0C%Ekgx!b6QbrS+t?I!knb%`_q!o&^QAcxyKmd3kp9P{Mc=Y*KtBO^r$b}|R^g;M;w24X-j`RJQL7}EU$5orY{XCA-2@Jm|CdqoLu(9HsI?;$Q+W|((; zjx^QsCd*B!3qngkevNqBJ#oG`;QC3c^*{R6^)Hj2^d9)w^)dx#m5eMMps*|$-PLcN00}bymRSkw+&#NB`j?NwJjR?=S*qW*RRhB84UWJAwHk=#BmI7st zCPn#M0syPtU~;QH=?cAv81ozu1Nsp_Ob+*C^7deUS4b<+hvz2tP9`*+`jH1Dvna3b zKL{jzhDs$5CJxwM(*If-L#B3$2=!R3+`D%|Q;@aEU$PQjJRxmOZet+~B`UJN&SLBFmh_VuSPX2jU-w3OWA>XkKh9w?v$o ze+~3qs!Ok4;%=J06ZQFLGr9K{V6GwW7%Its$?yGS+E1Qdd3;LmhRLVkzmqk;kllZ_ z0ML>E24^v+fFk;(-T9|VqPGvuuz#de({s7{^Td9-@f2~R9I&pJe4}B%_8I&%p=*}_ z6CmrML!S9XZ2c7K2Re}CfF>r>X~ti>XMdUyMG#=zpFGHT{;|S8z572u{%1-5Ea@LB z{SO=DWNZJ~_4<+*JWWC`p zvfziCA>IzTQ7uAwynE=8y#(nzN=>Z9q$Il2C&Z~mPi>6vv(4zgNjpQsF8BBp5i#j+ zUTEvRH5uV@FL*#i%goXn+G@w@5#^QIa4Z>RK@4_-$6na*#1wLh!#f0OOpI=jDy5t`ZrkH0+I5>Q}N z9-*UWA{J+c*wugVgp49Y8ST!?9`DYphXRfP010qI-_RD;Zu$BjtTuPm!T0qY(K?nz zjr4qaZ_{>li7F3t)2fC(ZK9nt`yAeOwAlDKaQ^Sj`&NZmD@!dHjP);BV_FZw_`cIDSo5MV@R6b*yWSr2>)vhIYCT(3?%s4 zq4`dV-*PZUkc}D7_V{u(6g0#c0m*XC?|#=1mc$>gCMT#G7gc$tdDByHMP)-5XBGTP zIR4!Rz6YG*Kc6$i%T=4`Fzs00bQ8lbVzfV!<3&8x6du}ePS&D~Z`3ZEWh$W#ZtE8y z*{}W)Vj#I-KfoE^M~{N;q6*QjCy(#MreDw z5vOYChQ!{SzE!jWqhrZEnJ4v!hXmK}&lwtJk5(fSpBnisp{KrboQs5l$2_tebzf^d zK(uF)S1_`M9uroZ)6MD;t26Q8LI)y2AC{|Frq(cm>(lLFH@FPrFc-Ki1#jOaAm+!i zHKbO*q$UYuHhvbE{DR>5;x%$uJap*5uP!GTTGCl=`oQ412G}qwk>*Eco`mF~5zM7sHw!~l~QDarsvZfi*P^a2WP|Z?l?>Zj6DbV&$MpK;H-At zZf2i{rbt1&M(uEfjV;OL3h6jAr%J*GtE8UUL|aKmh$Nu~R>n8%+RT!69L&?Pb>x*F z{@A@3)Pa7lAP9SB3`CvHU{l7ShSSCLt$hU?wdYq}a`-iw9j(RZr-g%;X^k9;6!=Fi zzCW53>ij}NEtTMCWEn0t=`D|X{}=V^0#PlBCmXY))(guREFsmE%g>r^;c$6mF17UKM^ zvS`RA$r`PZEFp&Q+rtanv<*6piRn>CdiERe7=g6T5moXlvCe`*G-P*0%f`J%7O8!P z`=%>#$g#PW9tP7#SI$8;zAfhT36u>=kth2tkLoTRLMXplcf2K|$L$%iBO-JpdECG7 z-?|?uy`{fax&t1gDRUeFp=xL6ZPHchaarKvgjSvp$3$aOySUUs(*6=$bkP!MM3l*D zSURT2C5emID&4)9W-C{lH(tkFAmiS^!0o|2%^YLh=jg2G5vP)DOlpIs+D%WGBlPLKlUxI9$(_uH8@5D;$rV>}I3vk?md@ zMknR$z)+PSax+&2YxrI>Q0wAnJKP*%K{j2ie=G&_hPP{#R1_m)8i(}KXTR&2vB(D< zpfopTlblT}gYy|Tua#60AMbVsOlHgO`yA|Jk5Sqln=9sBMV8I;Ks`n78F8Mx&xXqH z+zih6Eyq^X&8i#fKXEg-1V_ra-=iKEsy`^!94TsMy;!fpIBX-fDK-CoXFWAF-(lD~J~@gr2)#${%cJY)6v_GzdLi-qf@OF9)(*-%`J!hD`{ zFC5$y^X|S);h|3d*%Ymkee9a16uI1$?zf%WmpL(iR_%D|jzrl~U`sNVQmDpdlULiV z4*(*a^a7txZU*jax!QAkhIIY6DPl3V2i*%qYr7uiSyiM4F-52=2W{Eqs(!{Fw-+0C zd+sIQWth)eJ7X;7@zB__@0*GokQ#A#l`8q>FQ+~1`MuOfw-+m>BB0AK`6aSx0Fl}w zs!U$V*z>3(wR4lFq;O;xY5$8+8NVjSeUJysvt z(=sUlG~LefHgpHi@5LM>BfZ_ctSn*cU(&SW}{3LAL%cswzI!sJT)^EI4druEii(qOrFw#KsAb8`NB!vaD6 zWyRGAB+~mFX(bZuhToHLo)$i&b%Quh-&`_T1RQE`e9Z9?A*F&(VBPDh-c*;3V^Lb}=$`X6H)u}Q+R;{<#m8n{W?8pex+BdZQ?x(YE~>lmLrj(_d-B4DFSsnP6c zZ{X1qdSLQ!8>>NHV#d`nWmvQ~pXYZIr2_eyu8*@~4Vk$nFNleLA(iF=Z8M;sM7}?T*F9YBHGluU?Y0+2D6Z zbygZS>fCx@+L0-EEzjj+TkcoeR5%EL)eo)Id-yE^aG00tqA=VNIF~aRI1~)<2{M-$ zj9qSljz)?xWuYTZwS6C>OFr{IN;alPK4#8Kt-}nIGG&p(6fyR=GW0a9db}M76|CcA z^__c{uQ+wnhBh-P8q=L=6+!F|BZTQ-HpMN~~VO&zQ1QuzD z4ypBgl|UnTbSQtZvKrvRx<8pRWuf~GMbsg?)B5i1VobU|0*Q}}XOs3wFP^+=R&E9s^(YIa!OzNvV>-~zl*?0Q>mpfV} z-J_I9%g4Qpkly3FB__Qge5gwdrY;DNxo_rWSj*<&GY)4wOooHqK8Mhg66O4sW8HCUo-E1@m4SxpY}Jg#~Y56P;^>-*p$5%ZodX zltB^M(i|ZDfWsgiQ&s+z%>gr|dVrw$#b8ydF2lv@02AaNHt`yQbSKS*yCS1s``REd zKL}t1{*!0J%k!?gGl5C@VzuiQU!$(92E5{#R+|C5(c0{6wIay;^E2_4w}DgXkuvsc z1$dJ~rOibHynS9vvcolFIYCN|TY^NuQ-9VExs8m;RBkXPbKSda+cb8!Jcg?0Hl+aV z0*~xHhtV9ZQfL)Yty+j##KvL9E;k4c8hmhAVX^YSq+ay#>j}^9jJrkag2_w@-6)gl zL5~=yMeXK*D&wq7M0P5I+wV5Yycs}gcEm;l8;IXm;0ZvGNbi2VPFmS^ws^wA*o;zL zE}{=(9g)DrZ$I&|hnLDlB&ZAkpys3fB_=z0WX=JxorX^Ru;X?jvEx{k4id44Uu`(v z69Tt)W@vTLts4o7ee9bPM~b^zFb}Q2pLPlbGxk-q-NbwW_7**!{aug7{_+B<2QS?! zDim+s+BDke2JH662|m= z{N|do?3@Dw5IH`)5i*7g+UETi?3m=-OI&KeyERmt00VTHMn(j2kj^H?%f%wQ*h@q* z*}_wgAHOjsp69B=RNHbJO2!J`D4Odjch<$(Bpr%(NKJ*cdZ##8-xRS~9Gq?r6)7H< z=qAX3$HZ@!r}c2qNfD6z;zST#p?HV|z5eS3*pnTIoBZevWcyGWq9S~=$ z-(Q4RPpCIjNuChAPvlIm`R%{AU0W%yKNL78u@j|lpOg8@HAl9+t^~AL-VPf1(!#SD zNiSe1n!jeK`D9n%ge$tRXL+v^(B1=&q%Y^oBzFrgZUOFWF6FPCMANx!Xp-*bwxdP& zj=QGKw&z(8JLa)Ygkm}y3I@PkYegBI!bhrQr%15#2`v$-0oUKC$mFP936o{HbO>;n zh<7cIQCit*1&i^Z1g!XEP^(W;ZI{}20RJCRxSjpDDaO#GNjr(;S73BIYXLCymhA!X zWU}{o@Cid+C@3H2$3RoBV%G=iJF3Kuy~jK)cRsu;869XHrpz$;9O1WI*}79^hFG;; zti-WGk8ZdE^ri8ec|$RKr>yfmhq57?q2fdfGJt38;S8#cXLAgU4N!%RLpYNX<5>Yi z)SArUbdP?1pGEZ6Y?RDy0KfsbEnk?#{B?H2lk3}tVhMn{V%=DOwHbQl_3|$lRPXHV ze*)S{M%+(Rg#g%ool@d3Wn^J4y~J8%hAtu1Er$q2-uzv9d-P)V;nqfemJHliDSlcQ zcRAaGxdIT6HEw!}RRx|&{jRw8BkApvjy9uxZnuE-pz-hM%L;hEJ-9xw+Hd5nqhf0Q zd}b0)NS+Jd!a-7+u$`AM6EqHb9{=M>Vrxg-)E13KY-ssUgU;Z<_joGdC5nN!4Ps zXE6A|$DAU^wKD}d%)nzuNVPfA3e|3CdMuS`=(fGhsD2^D=gnoQAIfm?Qd`9aFHUwvc^l}CNo9OZ}dcJ~E3t(Xoq*Y5L z&Tc%bp|Ow4N0#GVTU&SrG*+&*15EL#^(``=?Na|GN0FMijB!C&?Ocu-baS zo0oq9jP;nRS^YFyq7f}~HXig@Os!tp@qq)nPnRmxbV2!n=?^i^{LG}wHcZyjN7!iQ~Hi&H#@?=`$meoUHk{U z$sT@e4+`BL+s3-CUC-7Al|L-HCxxsD_(aCwVDG&EyT%?b_66mpgDTV^N8MlTXkok+ z|KUv#;(+OP_3mtvMeI2iM*7SoCpVWJRH`yS|8#* ze7{&x2U6}m&NJ>0IM<^Q6=4M5o%j?>9U}wTT#*lOSbGxeGj*{uDfQ~o{bgMJjzWK6 z!<84z%==5#gqmG*+^U9H;Vi&`sm_fGwmfpa$X#QQld~}H*L+@ydEBAkG@F~U)czut zcYm(B%pmuT^J~nT7`@ETMcZQx*+@5g-7OAkSz3I2K~@$|sxNPZRIzQ@E_KVkCsmRw ztpvx<%vWYRghwNV`S|dKtTuVB#y*ZPmZ!&GHJ1kMcPGqsMau}MsN;7SsSj5_aW`Lv z71J=~US0XtK;6tIRL{)jzhmz3;b^aDv)>48)CovIfy+Y6$*nR~<4$q8Sxm_>Vlyav zYc~U(@i+9#dxl!exw_o(Ix__Z-!Jm0hef9=ay5k4u(0u1(gM^?p?zEm-U9_UKW8w} z4i-8*dIS2HS(r6RrWnb35l?l^x%N=kA;}}Mu3>8g%+xFNiU3pu?ipCKVa<{#(-J=d4c1po*OB?sAcXAEYpqYS)vvZFHh3TmS9ej zU);)yd4t7fpI(yn*!HNOGtz6=xx^7{9AJ`P)WrA$>`xSSjr}pu6Ebxoo-;`Nv$Wi& zk}^p6vk$B5fr7bMc)(sy>K6ms6obU=`bj{N?=Vl(Jz0R6zG{q{LjJ(io@LoG^<|J_ zMT`gF6JEM^LTux7GN=mi*wzhceChVMUt_cqvoHo;kYD@p<#~?DNmaoDZRAFc$!@uF za|Da+<*&hq8X>^R=k>*oK)5s%fG0*(0O_6_pb<->MN-mAusrE3tz(UiK-b_%rs4u1(! zwPEq9tzlUEnsoNLOEdt}Y%dFE5s`?~w+IG=dlB)8N8(nRcy{*UD>kmJmc29+AQ)!h z(%rlgHf14oxY6Tkd{eLwplgP~FyKI`5EzPed~vERy&~$VfIt)C!Fd@%JU{0wJWD$Hw%jfl^ohsglHberP7~nO-fX zLn*m04_QmdzQxvoXcOgGbwC8?HFD6ru9nH_??ZU+N};oYx&~a` zJd7h{j!6o!h=AJVss+mxIz6j}Ff-)+*EWhLTyyaOvJMR{DAmc+Eoy0EYF11Jg^)4m zf9Nh;0`yvx$Mzaj86V@ON6ki`R{&bznX#v@2(h^kg-KMNcWVSlBP-xCFiqljWxvel z6|yEdRo%^xuE~N2Wm{f8k={eN5g9z|)^v{T00wWhMHM4medo0(e0{)Phc)--p#5iU2 z)ZmRwjq*$%LpZw6axx=$buK-!alKO(`vD^1p%1A%zcR19EfT8)M6^?k+Z;LFAb{5P zsb7HkxCxg>f1Menh?_b*38XESA0G*ue@-%x!xIXPoGPvPi|4MccCH?Wm2MZ6mZt~V z>4tkyjnvTd+?d}QnL@gEv=cuZ%ko?tu5{0TBifXCEaNk+5n#V@==q#3E^+~of5WAA zF}2NV5YpLgV@8r6v8GEN5+85^<)d%HbG~W!VfV+k(@rR6Ct|B4%d4nP4qyZ{vbbEc zN8g{%1$4Kp&L$o2H|FQ?g0TMU%FVa#GIc^{061Nnu~NSZNZ^ZTx4LGCK}Sv7U2T93 zRm}M<3J>jkfyPFIdXfH`0RUzb^FD(6!dY&(JV;;mc|(D3mOZ+U1U-0AYA_nf0KH6C zISw(%K-76+tm{$~VrTaZU;7_zwpt%$Da7tfh6({qlyHl+>6Z4O58rN6Dh$+aN61r4 zXZ#4cPmJCQw3z;SAIm7WcM9{7C|hwj37LZ=UUxsStDOY@Wwzs&w*ostfnckQ-*CZ| zXY`ciFAK_td40DQFRKbvS~b`gasd+W75A)UQ0(}oC~$hTgd?*?>TrUqp=%=@@6_N! zn$-JQ{unodI%J#O9aQ0uEDnSRh;CN~&&FuT1oIVE@hXNh(rINQ#@@f*XnIX3tajQ^ zZ)$sJA2y}v{NdjFk0(TvfeTCqNKjb#q;hi*3|j|xNDABcmUA0;B+(y~Qv<-)51q*! zkJZVq*{`mP0|TdW-K;!7SQ_C@8n~!zD|kw{RQF{hEzb^3Ps>Y^a7!fo`Gw* zwkbZ`!&l1ao6Cl^^-KjC!iaxYgfi<4F!0sAa>p_zJY~#Qm2*Tq@lT zL098T&gz`5uB_|7^~#3h4sP0ky{|=3D_b2}^$@{ls#@28*qqUeeezAxWA1HHmiihP zp1{UE7FriQy;@L?jqq9;Zme4H*!VKCA?JCMfXt7%)2^7{24Pb6EDl|OC3D(?KEK8c z$>j!`WT_=ya8^s80di%AT*GvF`tR0^H1>@0bLvQCb|wN(5zY}Y}q_p1>bqN25x8>NCfJ`^Uf;Z&f zXQ&@PCKqr14{x7(!b~Rmapq;twcq4;|4xoGPB4@Fees(YZnWP{#dBo+$0Yy5$L@dq zEB*ev%>17P`RQl+vmk#Ktdz_07P&M6O z&V>H7>i=Wpw=w`$^x2Nzf&Kzjf5X&0O#qX)pkW_=x5xiE*1Q&geehXm?{A>wUmyE3 noj;{BQ#Y-n<3loa?gM{&dFrl$`W@8m2T%VO^*`22 literal 0 HcmV?d00001 diff --git a/contrib/machine-learning/assets/cnn-input_shape.png b/contrib/machine-learning/assets/cnn-input_shape.png new file mode 100644 index 0000000000000000000000000000000000000000..34379f1db595776f40f5598108a896f5d973f8a1 GIT binary patch literal 54890 zcmeEuXIPWZwl63MiU>APx}bnkBE2_h(iNmPl@faIB>@Y)s0c`?(mP6k&?C~6D!mhW z=m?=Dgyg>b%ijB(v(No>pZjUUlPAN>ta+!bS>?CZOXL%6Rq9I&mq5`CK zs3jpecT9N!s9{q$Y9S%HEN8E%_(V-nk?o0xtF6704GGEP$V5X5BfXz=8K$~wcBEvv zZ)VA^-Y3m{^XuGW(5u*YWR&+h#elsFn`Rh^pt5=o$bemnN zH-33rVrPR;)7Tw3Fu`+WIwf$2B+LC3*ZZX`7|Bu;ChzMV(jI}I_ce9t*+|Ug&s&H5 zKK8%nlvr(d@r{2F$veA%%}Q}E$R%}ev9(bxiz`N!EQxDH9&tM@SV_<#NaXW({RS_S41*LijKB4VFswJx zN=1{i3y?5qMg)6$a$RK8emhP#`Is$f13n1(_47E@b~1}IK3O0l(evpIk1TGT53Q+( z(ELSc@-)*#2#uqWeahf3BW&7>Sr;3fOUYIYZ88g0Ke!{R?-1Coac|yHDN3pQMV2Sy z(AK~BEsazrBAvYPn_EOI+Z8&rl1kxmmDkgU^e<~4{VK{1a*=Ms(SmNQkhd2(iK=t8 z7YBVlSkt{?+Mcv6M&U>a2R8o*121pDum13FQ{J_R>2deXnxy zyS`}7B_WCT=S6jmOQj(i5ms*(-jU?V8ocADj?Gm=2}OJ<4pCFYER(Z(NJIzVJ+v^)8n5Yu8tC{SVj6Dm#1M zO5E|7WqLW*_)2`lNLqlE|8bsd2LI3aIKz%edfVOa*@?Sz-ST8}A@bkI1dWEi-F-){ zDjruc}3{2sCztKKjk2^yr)6l z1EyKG-cx5hN~8L&aHTQ%zDguT)4f+*@2;MpMW zCZ49<=eD72&wP1;dF&nNjZnDhUk>ew7aXl>3G%!U2s!M>-Z?op9!7ME_Czx)t0PA^^u7mNjqoy==`?u`RC_&L-^9-soW@Lzp;{q zKBOgIP!Rk?lGFcY!|YOi$nEF++!ykm-!`It@a@8v+fEb>-(-K0#DyfxU9P3%f4wos zZzhvIN4`qF^9@X<@R*YM1>1#)`!C;$v++|>?JB=|M|qb{O-1+pWu=%O?{5E)&b?tA z`iw309fLKu0uN&86O#Z_u9{p zCHcKO%J2PhgLDP#`RuvY#MdqrCG3Zms)*iec=!6}eQSZb2s4#g1zDBmM-UZ)=0L8| z(?;&6H%M(5Z0_4osiSUN17>#qmb ziT*6|h)12r(DCc&l`@Qp`Q%+^$1OJJ(8Hju-pPV0R-uPs6odPhi# zN}8qfrt7AgOFBz7)lb%!*GJatdGmP3uluaN2pA18Yhr533pm+#Sq)}*Dg8!9N!)ti zUHu)yu*Nag;++c6FlYc(1~n%pkPfl^V4Hegr|`NZSaJH-?86SXsQG}_p)eUm@2~)x zWg5QQ^{%VT9K%0mxkv5l2W)c;a%8QpFVb2??wTy>b;xiAa1uC;IZHI-QoK?HHG{R3 z^Gh`2Q)R*Hf)hPVwi58S6OOjlbPk+CQ`V!lhZEBs(k+cYf)LnXs$%kP_L-N7%5de1 zyrN~h?%PS#Nxr4r2D~MjB_amkvRTt9BQ9gm7nd*Fl>>HwJJBoA=^u@mxjM=5UFAls9mDQq`|R5t-8B#AyTR7K ze@t*Nsy2IgyWOlW_ibA4aAVx6`BpZuRnidNFx=GFG}SuDF_P1@$f*&2v)80Y60P?8 z^6#`Q|29lca)V*>m}~{Ik+})po6{{y}Mn7lf70Z0r;vE*(cf z49jX12b2mLN;DvXouqu|0=Z6XLOer+=bnEHjH3s0qLtX_Ze26FHcpqz{E_zA)qS>c zX>pEP(OR)KKI?!Vkxw4k`K|8#N;R|M&oph_qQ+3B-X6WCO%+d7K?%L_g7prtve>vQ zj7=WYUIzAhtybGt0SDVZ<0V!s(Ne)}tU@s|7bMW42&XqpI1 z2B=tzs6x`Ow8Hn|{9ieDgLR(|M&-r{w08C|c?wKCH2ygK>5;mc<^*g|Ow?|3{cHN- zbjQv327y^y{O~TC>z4M96gRKz4yD<))1o)Sxx>swmVQqg9V{#v#L6C&CHK!)%p+-! zsSGaGzm@%d^E-G=c8zDSK`=+f;)2Tjx|)XcFzW4FS;Ul9@W7h{&*^zxg!3g<{v^^WJ4QIQcP z8WDHDF9vhnm&%{XhE;Wu}73q!W z8TjihOI^JkWsV!Bb||K%r`AFKW)6*;{&hvc&rZJbKn%4ls2c-T7>*{}H{Op;x!#?w z_4b_ZjWc~^YL4)9LDuWJp)oTON0dkE+m5rKz~}o}IGP4IAKzMErr=x0Uk;v5=QPA) z<1M9s2JYo~67+CGmEin-hkCR6I4=rcONi{DKbq=ibiYhH)Zx_lVDNA`6p1G~Y%vc_ zVbyWP#DfK~3$8Uy-oaDT^fO0u=GzVZI~uaLg2oB&4|X>Ta&`J;b+C_cx`dH!$05XT z*CU+g$;mzG`#QTkAuUn)1?PVUFp}(Blid0pq{3VxO}u|8TN?g=_E^WrU**{iUy}7X zVxT4szM<~ow+z1D$K(eh9i+5c6l}BCvp7_*jE2Ht3ugC0N`6L_tqcO%3?gv+}U9aq+Zs^@5D2%>fk`-5wizl915dJNuBT>E8VfxI@$S z&y2i`v@|8HT%Gw}Si8Qo;rDZP1A3E?Nc%|wMQ0nY7i@mcPA;C3elqNT^pFI~XVoBf zwm-UfIm)mbX+2?6boH=d6Xk!%|Bzkw5*r(vw1>5=q^^?6U*y1FGVFFXc)jrBbMfT(bCAEsQL^#0^00UFvUhc1I~(`K zOIL3%8Fu!w2mRyc&w1MT+5huNE}nla3s@lNtOX>%{}A-=$h_=r{}-~emOsh zd(*!qHU1~5nCO3@{Fj!0p*&lIq^^yptCRQH9_qW;d&vq)gZ^Fhf2T6~Cz-5(urNsA zFS38F|L-)${|n8(*8g`J9S?h88(*AFPxkL;{A=A`{iQ)?=l?JK@F%nVQ44UJ>?LW? zKR8$R(yY#9XA%;55;Y}-XMUucGZ!0~^wYTsiUw@f98NDR&%e=nA}>!)uB?2+##`~* z&CEDVrkLw&AO<=mE(`XDp9ZhTOx4bJa7M93M0CY%o-~KaKYn!X&U?zcM4b0#^UTP| zkq>ISIk*ACC;Qn^!h3xfv2NfyQqd>DSlSzA%_bj0LVAvjlJ?&eesP_}DpQv2V# z$Y6$)n}wrV|IfD^9S(45X?} zfG`L2xRrJA)mxncavgq6$I((XAzMwIDc1%qSi*8si=ANAl|NbH3CA1(5!Tuka(fv9b+d;_VtIy02UK?ZN?G_h9t)!s$RVbCFjp;MFmead`H6Dc2R*a4 zpuw-hDz$jtzAqo{SmTeLIm)Mk?>(d0{bdwVVv-99`Z)+5q9+_MuGOJ4D~yj<7j-`7 zBb)`ld5_5)EO^otW`Rpz;<)JhHh26poR zY=z@1+hyvW39q63c|9Gqt7f8qNw6GxH=EwM~&ty5wf8Xa#UH%*j1Yp>eB*Ge#%#YN4HXI7k_YLLd%g8@2Ht%$kRK4@d78ZABb=gT>MW)Yk$3Sru0&)^K}xKm`7jJN zXK+yWyfHYG-qI|+`ScV%G;=hIdI-(AK|lTW%Q$a zLKkV-el^QmQq*=eK$k_v3bx?uNQx$${<5*`7)NMcKfea?Fm~rgv%nBvmnX$(ZdZ>T9>rkk1hz9E&|CE-@xOKepNJ~Q<`e>H;M6accGHxk6 z+aD{uf!T+5ufH~h+xGQ1=r+kH!@asQ17)4ZY(sd;LGCh6=W5(bY%U!T#cgvevJ(+|SJ%^RlGocxJNZw%+vLOV9 z+}&y%y5b#3Z%JurL$k0DiZ0eiRv0(0rnKAdF67xq_o9b;#1yiU&kd;ln#CNI-v7Qi zA>^2^HGFA_wXk#WyhXLn=;jh$_q%h=Y_ZQ|iWlO$$^ouLLLaIo_I!3eVSo zBe?0L>mKrT%^GD&1P4q?20m#%Ffh&Un)2|L$gh$-c^O6FzkFxbE)&}c$|)qRd_J$Y z_PNx-V|aS}jw7c4Oq-D(Y15rqQWX2p3dfnZe>JP&h(Iya4MKj{YPGfRpYi4E;p)OH zCFms9oP?Lw71o)$q%4FIOX-Rwstmge_(_RD2%O1FT%9PuOtNPMXTr$R@o$A^3eV07-&{Sh@_)hU-RF zY5pi_UN{!ZnoT2%e_X#@BQ-=g*)oTx<52--tYM5VSu>4(TbmpY4^5z^JM5FweR^O4 zQ>9hvY4IBbmT->4E~2z9!Ao|XF-a+4`CZ7&VO|&ko{yf;118w zs3$s~*>_x2xlU4~P0Iixjw|GNFxF$Z)!;h`nJRI>$wz1HFi-SOI?_n9A&c$*w7#=m z1DAhnCcc4a?yFlz-K` zgrj+j*{R)_)jgQ^vrAzW9e@5PK;J-UDR&WNeEbVvK!3_^?!9%OU*i0Zu!|vboJu=N zZon8_N=ww}XlmmnaE1H%92SOSW-q-i;p>pZw3Nozsi7d>&dEp zOCo;gzO2>p>5e7uRx2MuwI0VgVrxy>a2@68H0fI*qp=oEpr1l6Rq_mKAvzdO4*c6B zug*zn#L?lTCW`e%XVao{ z9-ESrT-XU~ka#fG5%nMzO+5OhNK=5sr9ydoXXOIr11Bl8%kM@uf=)(U;fn!^2A*JL z8wlhsixX?`($zKU@TQt;*2n8&{p#`3hkmm5_P7!h-o*yeXn!Mk7{5}yF{lG-t`d|} znGN2Y8oXK|S7mGA&yH)U!;cx}+4a}0iiaysIH>c_8TRE|uUT$1Idv+)6OZ3QdC2#p zYOtv#<-NQWe0-D6$>~D{vXFqMl{M(MJG(oQoHM{o^PZb}hwb_%(6|MQp0qHK#_iE6 zd`6yZ5DhX)gH%YC0_7y-q!2q5Jv{MFAIP<5p@V^yBm>2RSj!C7%uL+zZF_{dPjL^! zAGhGWJ)Up+-GRIJ+A{LzxrpuuQ$ORKkXxc_Sz}C3Fh_>zoG_ExZC^3N{w{}~=`cXF zoAx%^au!>1+i?ar*e~EtFx`;E1BCi1)0gH3vRGP2D7*L~lnbA{X6MQI6Dwn2dr%j!$(RNu*+bq*RBXxu+20`3w1uk`h zd5&df=FFbgutiY7z(V~&aetkWIJDnigtYto%xp4-^y?bxDOl9oH}Be%*HhF zN;=2#$+AHP->`$I+%f@y+=g=09w23jG$Is!C!6gn#N)O2u^vYIg*D{^(sFa(*{`tVAy!u;5aI`k$0Z}T`50xF8KCb6k;SfQU(X;@)z_OYlTHv~ z)L@m6tGRRe>mlH{^|eIft)b!A7tY(R^*A%0g4l!R-5J&!7bMwxsf>SbIHGVC214MI zy$6Z?jX&L{91o^rVuN-zF+;IpD1!-$ocEn@6YwsGAj?7Wi8ZR+Ss4w1G}-2(o$Bo+ zm*4p2fr1ZG>y}!_tb6Y9r^T3LTjJIcy2J8}5F^v#7biNFcy%8PCKATrth}9!-y4H& zjujJNtW*AXPaww)P=YQ_CsRVGcf!l3H-GU{=0OfdSD1O+E`q@YK|kE+@UeLS53aNg z#&q6yz$769yEc|z2M8cn^UKp_x&Cyqei@SfrrEd-E6BoCbA86?@RR76^Q~a7%nlj(# zfhz;(C&J!0Me^9)M9kbu0vMn2tPO6?I~>r)8xds!zgaxHo+W91;5Y0OE&l|vs^q`t z-{IBhKFVFIpH3)tbm11cZzNNZxLZ@L@HPNbzwWe(`!%Muz=U7a7)>YKkqSp=I$F?_WWXzM23wVkrWuca7Y_c!$oN*KVS#F|fEbG|ZF zOpUW;%L!Y4+*-erz5p2iGj_%7tV*#qV`k9BFD+$zICqz7skS2vNJPE^Msc7q+Z(fy z!ucn3G<^bnfM~ zR;7KgL}=%O@%}0 zwKm&-`x$85!G#vFh^7j0F(-<+->jy1X)OHH$I&36EFgl={*tiWu%(+a6S?Pu7I^ z7(rs5h9C@p4 zoLmgyXrW=gqm)jcu>1@!-2AKSox1tNDopQL@b)3T=nXfuvl4#m7B$4JH`iyer&`chGTDuBPkI2RNk(+l{hHipB2AmDTbRZ4(B$#bLcc$V6-3tMx02m{wZ&Csb#@fZUyutvDBwSChz(8 zQBQ>CR8P9;dO}H~OF@@jABcEcZXtVXvCh=l*yKj=wbN5MXx%m`X0Y_lp*G0U>}s3z;_HHsgNa0h)5IQc7Hu_4wmsaNv~(@iN!P)H7q#uDEprXkYR zHMDEIjVwz;BZIITdA1$JL->gTS%dq52l)3Li$;mvdfJSnl}G7*W&!%6P@~IFw^=Ry z)O^w$`@@f0ybipPJR(}g7Kl~2{#3R!L+{2jtQ>ZLCQZ%8O^^dEGslSAMY28 z_PEWFR;aUzFCVK97Ge{3;z8$3cF!r+RkS<0))w)0*~+w+83+9alzI@0kyErsaXQ`sw@@`Lj?|@B==L8K2IXJlpM(p;AUm(*T2s zE{%Z$|4)i%KKIDE1D&$skeedHsk0d0x@9Vsq^y;ZJx^o&tA1i0 zTWZ{m3mLy)9GmX@3|P1!J~+c~T53>^fPy-TZp7tAGh|?4Rz#rLJ1O&>9@9-uH9xZHILwcZ`k2C3g*W*>#=d0s1Nu0ZO2|0!1FH=-!(}zy7uz7 zRm;9=9TB<2VLnYuebh0N4+j?{%nCgd;t>oUnh#HO0PCY~VmOd{B9ay^KWN-z>0L&IOurU%pj@b&I( z@!1_{&XPMzo5sl2mEG@tjriMTEQ#T@NAZEU;FX>CGH_h$7p)Uj4gNoY>oj@Ph1x4U z=N_pxZvU#<67}2=9xXLC9p^O?R=Cw$q)+2GItu;T7N0a9_Symtg?*6O@1ap9EpIor zRO6sNL9rds7)qzg>+|Ekq zQ;y<#6;*#8kZ%?;)x%;WQ%4&6tanQBD{t{6^6;B?^+xefhaZ_DhP3>xvDU};VJ;JK z`dHOTq&CR)#t(-)|L=KqwhB2hQ#NBIc2NRyPMPE;IV%f##&h1KVd6wgbuIg`hbJ9x zJXD|!g~H3MBodVM@~(64n%&UCpe=VOnm9LdYz+c@COp?x zN|AZ%oMhdm@d%D=Vqx}+i9HWJuWAkCaU!%Lec-tR){}>U*I2A!z{d(A-BwC?W%_mD zt?swGIPe38g*$7d#zN6_dF{~~udM*<#&=Y8c;@X|3Ap3@Ps-Ug#s^QS?!fB10o|?x z?GHWzSC4J$pih^;#<7*B8ojaxa=&(-`4=NhS!fX1StbMY&_L>hJYDK?yHSuslZaj` zf0wwKbL^gM%nD}W)fJT-{~yOK&xDnR5;tZ|nRFv$UuCma2Jv<5;by9sqxX`Jta#<3 zPjFLNH-6jKFe$f0r3j53a2R}4HIwTAA|k&-DRMVl8}ssgbeqk-(hO)i;HQG%y@OJ- zKFQ}-Ththp#QWx$;*oOH92GqXrq5JWGA<81A&;=p@3k4W*to`wj z9V7N2JeB`}&eMk_cC*75Q1-3P{epj;n$b7F4oTLYajgeD=l*PwYe&cljTUl+x1Sr< zP2ehe+fc@zx%Chf$Fg5N(b;TofG-k~ib}2C^wJ)+dEA^BH}4UXtP&n*>{08@$8iQL zzYLwv9K2aEbWk^7X1V{`!{bn;4T^>9B;riET!kHGlnzih&TqP-;Gy3lU-tZ$3rZa2 zmXq7P1{k*mN< z4BAlmN|(=&4{ukdoMFdHNlW4fWN^-j|>+qD@J|8OZIB3GVWtP7rK^Z`zxcov5o(#OwG~5KToudkQr0+%}6}v!x}2iL-tiGp3cYKmbP(WWjI>_^*2bMO@x6VR3^D8L4@jyv?2Vxo#4(CN=_*{`SIe8*zY)LJ)gcH%gI?Id>S^Z6esUH*6=wfn(t1PI2yP<%PHm@;O69{6UK^8yGLSpP;Xui zbJ{;Ec^EaUHT9L7$Za%baSugdYJDsCuv)O^lRy){|2|^a&ZwsG5I4$g&V}wrSt0gO z1Wabq>}n$b4Z{)J>yEK22w%0)+%?jQNipZI{*Elzpa7ZREow2FYJJcH-VPwTo8dF# z-&uvxo=Pqc1vP%Do4(#zu8scVxx{Ad>thg;KtIJ`litI!f$_y|ZrrBgTJRPBC_ zW#uwmYJpBLe!t4L7gnmA1Xk~N%wA6S>+r&q`3~DKtx^9W42?tkl!7Ep^XsQYQ$+cK z*-Aq9+be>EOhs>0MZN*zQqSuJ1F48-aAf6deYnr=up75?P#yjhO$gLtE3JGK%@GeW zGOF#Umn{yfxgBZS4-;`hid<(T^LaUwoz*Udf_O>pMU1sCB!$yXxMpJWpjWy5f6uE9 zo;S|4rm@xXQg+T?&%fE7B8UI6xI7S-T=2nNv{Kl$q3@_4<`Ik* zP)){lJz$Ml&G{nwP5~s#wO; z^^N5n%OHW0x&}eVNgtDYC)v$EqHrHdN}s>^(|EDd0K_M4<_T)OfI*QT`*j`Vxoooy z?Y_1l8yyn`$s8-!mjd9&(VT!{%WY^$_w=@(G$ImKunU5zn2+VPJt?OFGDi0lz0_X( z5t8e>w67e(%T70$xO;7mh911TWmw; zK9{C6n|Hz0Gm&-clU}N$M_m9vCoPa3PM6j22H&SPbUv;rrF%vO2h1`^;K6~2f2Yz@ z)%=l2hLQ3};Z(YP0l*MH8YQn?2*n@u%Odx~eoXTrejHjwJIU2NHnq*%!x6|Q$-jA1fAu3bwN%X{p!1DRgUW$gD`LCNpIkcY+05n`KHZ+> z^TJe58AU&Ky1jBk%kmlPm>=WOn4!hgXwA!I${5$`cDl79jQ0hp?}%C89+eNy9U4x`vG=&yPP+2l(6i7#xvcBgNhu-j z4A>l`;t4mwjf)1>U@{5&35k~t!KYI6P0#8&8_jv_W6t8?@?bJRquy)?1VE^?1qD;- zcV#RKpLc=$2;ty`)&c|DorW=d=|mahgCPNrzIZ4AI(KS!JLvcF_CjrU2lo>)&0!YC z81X>mfr`i!E(V%RYzcjaxIPMa^!B3*UXMfv9t zO6!@K$x^R^?P}$Smra740RT@`8=Iv|_T8jmOfmB{Qri!LRfybB43gjo6L)`U&x+BC zPY-NSUK7g1Kfdpb2RVG)tjB04uI>F$F+TaW%=#<+fsVwEc5sL6O1QY{f+{495y(2x zm(;#W*+T;uK1RKzof$^&Ab}&^BAA!#vpq#DQ83Pk@U|%@Au+zLS9hPDl-ET9S>>xnzQ)LHx-3d zO}%F9eThC3_MP>~;w(SA_D?+_Y2MHA+k@3?uRKxR<>c;-k^-CXF&1wCEUKnMnY$-2 zgjN+~lv9T&KX%sz=hrb7ba!) zDv2>c0YF{Tv7{a|E)Y0?u_}OO?Z(~YO_8utF-d8MVJ{-bZ-R z7f{GaENsa`?XS0x-OW9RQ{ajlr?583k(Qjt^jf3LxXJU~CBpa^+g!p5EAril&v@~E zULs9>d(@EF#9ik1BeV=#w4t@ZEj1DXYP|5nt2fK~HfA4Vr@-iH^YS3OTNq>GTHW#l}FkB+fbGpcOL+#2n=` z5Td~=J@LcR+!e+WVppqwjFO9oIfN;>@yAWvLpHMaeri@T!ENrqQy(13i7gyLF z$_;3U1i+LXvlhEM;JZ+UKqa2e$e1?O3bLG-4oj=uYvEQsoC)l&T0ES}3)0IKvldhL zo%Yw6d{9z`?OAAltEC#}k&w6Z$^+mdWiT;i?oGa{H^Q@TEg%|)@Q=R6 zur7SY=o=@zQf?_T76Rvs*@!^9T3=KIYmJgw1rrqp@N*(#r~D;9-CsnR+`|%lvQd?k z7SCwXxpllD>9<(RX^IP->1iCTB4GD-=OpEnUxoS3db$?bszCfTJEz1)6oqY! zrAw)Fl^J__uRJYBF!}qv*eeXJv0j6}Z7CB6#mQ}d`hL*iVT(+LikBSk`kRb80k4Dk zwU!xL(8&eF`vq^Ip7`b))I?D>kC-hB+pP)WtdQW%WDT6wI0E+=DkZBaFA&tAID(c67 zTY`=p;WOzTu1h78gUS~9Be3;e{&a|-a=&u`j{t|}tcu;0Yis?PNx;)5PITJPTtWen z@k_565|#8hTqX7*VBemMWWxr$Te9uGl?;4QD2i-dm(^s*W3q<`Up+tr0Xyf0-cgQw ze0-TQjs~~{Avv?+NCS&I4tc@2>)>2IL?Mkw;?R=ZBI$9_j?C9*8l?foTwQwUmRub< zoC?huaomz5b757}% z`@&VSXcGbKXKuDsorFx6!)a;Wk|JD&Z?g;%_hra^w(022Vwlzth$kD>9& zJ2c60hUuIg5E*6e5&)SQ1Gw322jF|C6`%p!g;&yI2g;c2=cTc5`D}?40eBK=(kq*cI5P!W&;MaY`ThpeZZ2yt&ZY>c!kTz#x2w$GuaMkQ=KIqL` zDR`Og=f=bw3LG9%UB{MLfleg#=4l4gLC09Zs#S5y{3is9G47=#8c#KZ;c+Mta)pgB zNfFXJZ-_F|YeDjEgns>;zdG8>2q_s5p3Ag7wK-kNw#+SbzHw^5d~$J;o9g9R%zJQt zf}qmy3wsAv7cD<=fNaPc?$l`3Dfhq?st4N{EN=9=N;VD)Tlhl`z+7F|JH>Gy?$gA^ z`48aP;R*2ou5_rs|M{6r8%+ddR}+wSS^D7k&3;m8n8imq2Pb9us(=XCMG70ON6kWk zBg|ZfpOvkIdQj5b!`K)Q43I>i}YEs~Slx8l?CAIj35;$M1svP(Qi$daR z!rEvls5&UH_q{EqcO_Sh$>(vv%x8+kzTEnl%G>ERe3L2=Ah$QOANfNdGvBU$;(&c7 z%s9)-z{h3t%Z%vS+C!m2NMXApzfZ1Lnk+%;oS_`5iyb5deMC zqE~K~P4zThoItsdK}%9Urb`zQb_!D(yX@J;@vOw9>uaw{^G)zg%i(U)1oO_^Ro-Lb z=_YD)*fz?_h$ZUKQ{6}QCiMZj^Fm%e8^+!;@K=cjNCeNe5$I^{{Ppurj9jBm;jy!lOJ#e@+L>> z>Va^9X|QZt?Fs76`{R?fCHULVX^io8M02nn6bg4C_%#vJKUV*`)>)48pL{%q>HsUK z58za$K^DiD?gxAni9?+}3->T$%!GO`w?mKYKMd-6t}hBcFIWRSMw_-`qSLHYGlPAM z@BcdR7(H|$lU%stx!*ZxcRYZV;}cmlY{Ku;WPGrw?zhM@?&mknVCi=_QhII&SY&*c z`SmNNB6sa-T;}zn2N!sMiSTWxf`=yH>I!M^XOBUyQcMP=exHYTMk{Ul6ImGg`(Pq5 zb7o#i5-+~HRS$L5T@J1>&G?))$r_zCl3!3_BEr}tjP4hS3G4!Qho9A@%f-+J}5s`SGjg7 z*e04yF^2-(yCP+2pnUdjGZl6n8)$@7NN zj!l&bLVWlgi!Q&-NG=Vn>m9y5M#40=gA>}1OwuUasb>PVlP05BX}Des{-Ga2>43E2 z^7Ed=RCUm603qC^DSW#Gq^7+%7&>*O3ioZxb>CqIU6h6U*_$MICtL;TxgdI{Z-AaM zo+WB#ot)dy4{;Pu;~=k9Nht89q&1QwBaESmUvE5ou;WZ->O1-3WSA#vmhG<0T~(_W4kp7fC+Db&dIoD zyEV#lQvggU$ww<v-AD|D-SXtfzXy>IG*!dTW~vx^=AoYDeVNWKSO7u1 z!%0tz?hWI=oX5ZHcFO%n0LTOycK&1OWVLDM0ov6M^7}d7I0_wYibe-ie$7Mwy(O=< z4H_+0>E$>1gBnVI>WaU#I$6hSzz^MZyYcaC=#9*87aZX)PN%k3j*e5>HHNambYrDo z&f?~=Vv~io85eo54_32x{Jz)EqB&U6nfal8Gq@7#8kBt2*rg$bW{@$a% z3sn!YBEK87=W_k`hJSWlEeatq)WJHZ{euzyM2|`LCjVdUG%Pg31pxp^~9iK;o23#%(gmdvmyv z`)b%T1^12Z;_<1-W7VN%z?9ZGx2=O)v%Al8W=+H-hMXjP{jj~|Gd#CmhAjKPsZegV zZzJrVmj0c3{Y#5Bohjn~jPgH@{6G8h4}tJs3jf)c|FRwbxLN-{_vQC2^3mV_3*i6B z{eD=~zW?Fq@$->l{k6V^?FY`MBNR;UIJMKpb3Tl3W>JNy2w)rKWJgDQpMQ^2k9BR%3e4~^*Ty@G?V ziiN*qqguc-{AdNe5{IY| zdNy9&sub$hwR12x*u{)h7 zOS5cUGa4-q!IYRQYyjf%xiv5g4o0;BV^Y*k`oTxjGN5U+^%x0Fk5H}afLWD=Eu%%n=M$s*QL z4XX>G2IOW)lbZ?6H%ZZkJSKxR59aU=`kDcNsIKClBjkDuIaYDaoixeFInsdU@C>t^ zgNMKrx3I>6kDvcKTRb)Ml)Hq_vc6#+@FpMBEJ^D#PUHE(`pDXTd{8N8?ZHY;oF)a+ z7pSP~)B4TI?x6%K`uu7~J>36k?=7IJ?6$UH-GUMZN(v$<(h>$Cxs^u{6;M%-Mv;CyT5EvDuDaO8G*@d!^7C%tMw+Zf$+rbF32X&pXD)jGglsbJg%oqyW8auS{;qw z7`pR2hPEKRYu^3sRfN{MoyW9Dj@q23cMfmOIKAc`yWX$G~mdWWV_KTtmHj_^4 zA4)IKB_xk-!7CM9zB`eMS=R!Z7D%?Fps zGwLM)JiY8{^LI4WWKPX~g3Rcn_Z^zD#$%1b89apq@+iNzGhV2Oz@;3e9iK|RR(;}( zI{(%hY@7K}J^Ad}s<^O8H&SDVT~QiHoG!ox0~(JsjiDS_R5iLO-2++*PIxAp3D(m5 zDQF6R`{Z%lUog?-5HcAnBKdKU8_MYcjWDn#+QwsW=) zo?+(TRMBm-9z!3?I-8y%bBE4%Vtd|vNgIs}ynQ#;2Vb0kj3C8Fs*o8wR-9E-?Rd0h46 zaoec@@JzDoJjo%}?Sxz74zrB+;VI+#&VBhF2AaI*AO99~UiHKu=u((Cq7>0yHtd&W zme%>EH|0_w89s}y!3ERZ=3=Wtx9HHr}+1eCI5pAQCV}o~GA^mQPc|gkKnX-o~bK(kPNj{Cy)%|>DHf(_y zX+(nIks0eJC9IZ;M6O{fpy)v& z#n3evr#xUy(St&d1JG(j0|{l#x(x=&Z=NmEv@FPJX56{eRpis_rX4E#d@_>n;N_pD zDNO)6bABnbuCo&2oqdSTy34k*5>6y$OEQOCVqH$gjWQx{~l0UE|G&5dfsQH=82iH~EjSnuoCcN89?K}!^PN9`SJIJ?G zq^I!%_(|Qt`e#dh3Cn^i7Lz=^kMHe~zTn3xMX6@0duQN%0eiZ2H?xSZR;mBxp9CvS z2xEU1ZuevCgge_;#_pMU6|uLzZ*O{)mX{JI9>_Qd2+y>NMJ%Q+jIm!?DBi<-V0p;A zyFk=8)L@N+zB+o^z2N=!ia%c$c>o2~C&b_B&y7rTwmB0fy6N6DOtWH4=yD*f)h(c{ z^eS7tEHI=}5b42N|2l3)N&mgg+(M5?;NpCLnx+;Bm4{P*jn222N?88U3aq{246ylm z1OUhlw;~X{RcC}7s)47QcQw^H3Y(q2hU7drjHz8+@1&6vfx$aonmZ=OTxs7cRQ0QQ zYe2l5GI3?cowQ+-SGY-F>UF+IE zAmAHU2r`L3g4rO`jfU|@FnxK$HV`t2X6hGe!+WY=?`G+4s6*HGPT*yzlDTH5G&@En6QML*rKtv8CS za`m0yIr2#wA~#?1c=ePdt+!h)?f$9tjSfo#fH)pksP(alhN1QHEV#>p9pPtpv7uus zF>r>*J|(ek>cx=SK6okhe9yLR5adte%BBKO=W!@WYI7Gw@PP|}!>ktvHgx{;xbiQ7 zf8Y7uz_>l0=AR|_zt&{bugoiazHt;!E>_WJD5F2FB2v{p zd#i5yrw1DE6&k42G03@C;-LLW;n7cbPimJ2eQ0r*hf`srvD0*I+#~{i$dBDiSE<3m z8K%RaOc|Hou=nyoI0PtUPtViSHH{`3WwjL4j8ZvQK;)QndBDrwc(Ko`DFhP+|d#7iGa8IPOLc3h(Dtx4sPf!a$O5 zbwPzs%pm|e#3=If#X$_U{MbE8CX0%6o?|oh5;fRS3!%BxllLqa_q9^@Xm|L?tN4@( zFE5y=p3m#wfvJj?kE#?MdmqqR9q1)+oYpvdsV?JeZ+tPQj0~r;e|4NxFIIG#PdCSz zu=m@Z3>=-M)c%4uFlT;YZu{#;Yp^0hI`*o!;HRjQyNnZBjbO%M_+jqJOD+?OBA46} zq69nUs$`cYO*`V$`nzrn7fJ7FpM(&QJ?i}Yb56A^p~de(W)&({Elmzuru2^qgRGYD zdoeo>9#!~jq;EWQUSD9;dHSlp$l@EB$jY)AowP+y!3*PA3tiao>0ndSfRlxqFYn$8 zX9XB4(*MZq(_kX_M`A?ox2tup!Emk}>qP*I5HOCZg_x!w_}Lv4ep?Fu#KZMQ zf*CMI0Y1H5T->i~AZBdyn)l}t7K~0L=SfzzGqKQYNG}VNhk#T4B-QAgHc&oeF^fj zRPt>#Jfb-M^o}{$(krus>JikvJBWcSrFhL&Adr~8S73^TOLMy##@;kjh|OueTI(RZLPysS?(9Ph zbLi5vW&U0xexc0iPVdS`y(jJ{ktorzVDc_d$WLiE7$m&`)Zt70wSU|CdpCHM&-(Q2yJxz;Z5U^@Hw^lN8F-Z~X6G#SD zy6=2EK(mu`5f5^FrSA8N2>SAA6uVr^>w#QX*(kYu8jp*?$De6cy=~z;uO$N8ei|nX)8Z!0>QP05vvuD7 zOQFHzEtuWVy#>J)Ubs9>%??w=!I5s^YW!)G2G={wl7J%dniyJY?3XVd7ej6Yv5fun|${KsSc$S`E0=qflsz*sigpsduG$Cz`)YX zJU{znr|he>uFv-wi&@J3%oG2%Qa`?s$;Sf0gw$g|KKY7`Q!7U^^fbO5&T6dzuum)H zuJcatKS4c=*;R{X1r!7m$}0%kaiL!zXyz(lO8uXB*ACIj;Eesf8|b&9Xu%!zk%a4m z49wQ#Kl|p^J-mEXb>K{5_08LITzp1fKG7`pji6MsdmpcpaLR0TuH4n2I1LU%+m*8* z@cgn`(v9?qAsIvhIundDWkkq-D~q>pDOH0Q?+)iGtY*A9tcH0{0noXgNy93Z&;$T| zv%Q8AEK)(}tjWX&F21_QRPnS4Inwf)$G3aO@=b8U4!SG6*GJtVH?BWenxgn?yVN|stR9TmiMCVd29 zn4}=tkk}wXD7m|8ifOjZpM$eg_VUAx3Y+HalvP$>mRa>Qq-@3C%yT!dcM}cpjGASidrfueYr5^bOtO`i($Jb zCom?A+>jJA5$m|DIQNNJH29$1$crm=oVW#dGN|!JSb!4@o*nh-(hh_wPG}jV@f^cf ztBpq^jTf<-!T7-=Rl%(UG=2?fT=(U>=%|a@$3wYaOmRznk+RZaH&vj+gZzVni4HTz zXP&k-?{!~!@V;&6;z+6$vR^U_Em>N$XgYp^;L+L!#|nnccI6lrakow1eE48NR>KZs z#q=l<7Ug>iy2im&hh0POX|cB5b3CeE7ngnX%Hz~%pa#0?0oR9|X&)#96H6xlYW3If zRF#=BpGZ%IJYKlu@Cr{KPK_cPaDUj~DLC|)N$Sdbvm@7Puuq(i8iC>K-4cf1J*XqV zfgaItV1^kBykOab>c>L2U+HXHhF>Kobbp4aOYoX5QrEq{OjLsvKi+M12KjE@=@bW> zr@~_f)K9qK$w*@^E7JL}3t_~hy`&4=TF~rucd*R#Rumw9>;z+-nf!A3% z<){l&P&xO@8#Gw;kAD2nzarsPesl$rtUp5Qe+~hcAeimC`g1(E;kiGx z+}ETDg8n0mKeE`cG=8j_|2cNSO}O=!kjmH!4CmmxT9}sG5bs+nHJ{s=)4!VS+Oz}q z0#YU}N&Oon$AkfpGQ-_2cmH8FKV-K4%Sm{q>e17azwu)o;($&y9w5Sp%?tMVUp}b= z&qVKM)cwtX^F06&N;ySFwDE55r{>BZ0dRNXe8q1D{D=mCFcUKtvcFw?{(Vqk`{5a_ zXW_pYFk}y)lsmdFu@L+;gbi~Mh~BVE=H`v~#o?kVb! z-=D;gk)HiH<7EA+{P1-m6zU13h8-ozWEBhHX|k=mE}_-94!Ov(LR z@wc6{g8Q_VPIQDeS#}4V^2@9dnU7gKqjN*5u;pv2u7d$_=iI5S|UKK}UqL06Lha`k- z@)*Vvk1VDBa)sOGx(GXtB2xAtbaC0F38beF3}M+-o=-(K?Pn{>2WI{DBFoX=nJFy@ z&+^&CZ_M=dBbc@3nHz_G)6*abFY#6Qo=vUG39SjUW+Npc{!LFh5+FQUM*Gbq2d%Tg ztgk)4bp3ZrMi7Jo@jPL5^mTT6KB*oe3tykj;Xvw6*Wx4Vil$e-UTOzGsPHAv@ z4wQb$*{>aRkm5^dOO0*(%xxlm+}r(r4yr!m(V?$u)by+zCgn3?e79hhC`Rmthi4`$ z`-EHOSrr$P%Q9V8F z2z;J7bPlcv z{0-b~%N|<{B;>~h6s8=!kUdL&p*jUr&d<_1>rj3s`oPw*Y15#7frkrk^*W2|k{8$t zQg<_ihHdEd>fBQ+uu%t_ZY2XtzA7aTX#81`WQ&gbjXb%^DTu9{A@xtlB2FjtPct(BO zrP=P_46(3nfaEQxSkHFJGY$K*s5aSE?rO`KR4W!wS!$?pDaNL-35^e!dDj@Gc2$o| zdeFcN5j-F4lzhbjDp0y8U;!%V+j5~s8Lsn{?F^CNMgkQwOl#5`B$hocydy+OiRd%d z_b-7?lE-SNMjGoc(dUKYwo?y%%L^fVbr^xi@*Wwt={N@Je_Sw*0=x%o)|vC0xW z3>uBwZsf&2;A2`+O%@zZ5k9JtW**DIp<;o? z6ziPP;sI7c!!vCT9LT<_7ahUvbdrwPmG$%tP2mZ(G?-OH3vQjQ3Jy%KCw;7j4COic zt*9fbv4!uBueQVU(^>kSsb)Q%%kC*2H^w8S>Uegcng|vKd$r>mJcXwi+B&_Zi&SL{ z?D}FX>LPF=P_8SldSKTs)Gp|M*Cpj5j@E@DDIF(X_-H`i3$7W1!WU{G{&e!HESO{D z$z~Qb+JW3j=YAw=ESlAW&y*m1Zrl&2%q(%`Q<_Elu#kXWQ=0M6sh*2V^(l+<(**%V zDLi|v=bmdP1jnh`->x?NcHQjryhcaM0|$%wsgd_Q(KHqo9@!^>%Xe}gr71WAQ&d`} z0hmL>RRQ2k&?h_f*czVM_JI{_lIfYD3dlUJw((n})52}i^T|p0R~Bb4V4?@fSfV=W zLYFS4Sm(OFyEXdMlMm?p~Am79%2ZM28?{l^;xJ&D!R`$V5X)x;EPNua4%-fug z`^waalzC+I8?bKKwcmA;zJ7gwZ*ZXfS$@zdvv_s}1_l@#PD39=5ezD(pG>JYuXo599ms_U*f#qH`lD@X2!0l6dtCdr$uM`va=GVEE6#2-%bx z?N+*_&w`oQX`EQ!n0J$(n?&4?*sB;R`pcNJ4=FSSZF_=0#x%v4=!E7yq~z=G7J&X< z!bDz@2Ckmjo^gU77f2;s4azYAYeFp{ZJ6pmWZX{3e2bN&vfl4s-+kSPR&-tS;&L_9}a)+9Xw6IaUNrsCTEOp^_;znuEsXD#?s;DZZ=dJmdW;)`VnwE+MldOg$wB4zWo zIel$h+IXCCPZO?GRpz9tt7}fTkt+H29n(^(Y9DlNq)a@0xTn7ZAEtf>XvuvpX{T0n zfK~=?l9DC9o|}4+{M={{woaC>muD+Vmq^(ruiLoAcRn{x#mM>ndy*{2+&wuRGWxsv zONCAhG^O(`EbZCjR-?8oku=aZT+ zd6xbTOv1rv%PTS_XS-e8if=@}UKu-C66oRlx&Uu2gU;?+MZqbH&7lMJ)Dj)^w-yFmH)D*c?c~Am3n{kcE?LwVb(-=ANOtO z{O57iQ$j%1a(=C5?60PcAwmKwdIH~Hjbcr}K$1bp&&%CeBP0GwgjdlpYt-p?cq=~Y zmu&$_1_G+&ihtA(Sl8c=afP8^*5m|t8Gj+y^{IW}f`BT^C*)V7So<0|2?SJVlB2(2 z82vGsU+M@J&Hq-D@mIm{{9;9-DN#Ma(=MDtro_m0#5GbbwrAR5uN?R{HR74=@@j5n z?VB(4Nm8&xwJ(GI|77axHV<(wk+MCl$iwwb?VnBmn+&Wihhxbao31Sw3XAci#mrIU z4*%B+yG&jBl|rY*sGlO|xwi{StID9r>R#M-2%iA3-fe}wXP>HHDO zFZQVal|s>~F)Ch6?PzOj8)-~3@P4@5S)X_VD(5ez?v0LnlwDqa$=aa6^PO2w+&lfc zc;o#2QEvE9MO(|B;;E<8ID7DQIZftL?AyN%FIw>9D!ks)?xVj%3#a0~Plxk-K4ZyV zDFzD{V&T)I&y3yEJI=)|YD7J7c%8MWta_w$(3)Oo_M5Etbev)$+fm`M&vE%Cq!?k_ zDc8`23GKEJ+hphI7Dv%-&;J9stDq%)!QDu?XMd-B@T-=<%z3*CnrSj^sosSvt$p`) zE{Y39X>T#sXvyWU#Dx4fS-MMuNBKh{x7sq3%z{dGM_F2i=ywr#;5Kfbf6^mO4=U#Qs1xtrXn=KrnTQFH9 zfHOYio*X>fCEwy2$3?-3n>&0_lB1%>J3ObZvLxKvfz^2OWJG{WmQhQV@vtw)jfU|u zzTr+<6fJ?!Nb7J8ym0b3lm{N|>kc#|pp08qLOY$h+3|Tst%{!?Kto*%fQl zNTO9{dGXdcY6C$1>Tg!O*#>ghaXqaYqL!|9vlHO*1@^&U?qBHH4f1ePPqR8gj}vIG zIOXP4#=M4ijq8QANvjMz&z|wq&LLGwb`5k~V91LNIeRGcfrNxaZfx?RP(yA0F3N*) zjRDpg?ZX!7@M5Lhiiiu$^O>32GS)L*Ubm)oNOTxM7mS^Y>*0qij3w&k97Hwrp%CDR zh+35$rN?PajT(w*V`#lD#OZ<42<`XHnR1fefe-P=ns%gR%LZGP#^H(Ia!O=jUe7A>suL9)u z#(xPmywA+7vmwyg3F6P#=k^(aGH8dc732*fke+aPZPc3(05XERPERz?FJ*ECY78}| zcHj;hYSd=v%34YxojcmjJs(KY@LD{_KO62HUn5HsDxjxg*^8ACvAO=9$5!kf8j<@p zGCxxQ2~0_;!*;<(nsaHeVTYk%38Iee;*ml-NHLhaKyRJBMj!bifC#4kET85S-W6fo zLYKg)r-!R!NgaE>D)d*on3bK<9;?Y8Y{7|EU(JtwkX>0O6gX%;*(aM|`6@p^teK8x zS$ZP6A=%9sk0Oh58T%41o`&m=7$G=O2$M^-$6ttJ&>Ju^~j_3c|{dA4~b zn}QS?oww`AX7_mWzA)EQqOU^M!~Zq3OW)mTApq5``2x_VOVamv_KwVk6Fw?VU75Cf zj$K$3#19!44t6 zT&aC9gqi|hEBs)7voODL8tNCR0HYap1TUb**l{oRDiVWu>(6fB3 zShTq*DY(hg1)!&g zSb#n>UDw{dN|hRgraNeAY_`u{>-oOw81>=3DeVL+OCYb1^3}fAF)K(|&D6hvg!L`s zb^2Yr&q|NG*+(6?1>=um4F+>Qb|SA9kL~a~Liykpz^9M{m?XAm1Kp7UL z?AdEcDYGkEFWL?)bSB>-uJhB;3dWLaO3J6b^J7u!3f*8C)KkHG+}^hUE5-+_uA z5nIbul#mNRoXksntB}Keb^uOml80UuW7D+Jjf-&afnZ(qA*&5%1RXmL4X@y5!v7uo z%z{qEp-OmL`;eML-Xsseo}+kIgm78~9mkJvMP1vMGUi^=mc=pcrjp)FprLWdZrPR2 zxgfNbx;ofKt2W)rXl9Y_;BG8#$xFYmM%B!M-?Z@Qs+geG+(MsEhRrHiCj=H}og}{d zv_s)te^{}RaqkI9?z28_pN+^SKcVqRD(nnuyt{_&L^@@MDiQpbsL$Bf8QVcrFVXHj zerlJgj@@vlzN6kc!d((DmiF4_k#|=1a_sfJ4z|mgZR~0yqDQa4Q?GaRI@O)q2AeO{ zvRTK6X}DcmyGJ>RluNi`!Vw84u9Z>KkBDw9vwBVF%PnBdhA~?2@X00zIyfQ#{5)G4!nbvTVWu8zfmrR|Jaw1 z!H2=ISA&tOn@|bZ!xYPAtLk4X?CFs|vA^aaiLnFu3!R;)N%9dptkFPmZR!G>SZbEp zXY7#1@VMl(<&ZuHoOASEk-8c)c1q0UJt)>7HMixVbdf#xyVh~ZU!2s#0dnzi zcr^IZJWY zLu3sje|66VR_fwaT1F&d}C~X;Vf|?!smQ;Kp8ab*v4*hGxaaBM< zbg8*Z-mXKSj2_o_V6l0ismHrHC5m6~JQ8ox-TEF%XM2?0aPp%m$NL$sBe)=`F;=J9oR51ZEwJ^&-Yw;~P05Gl$co24Dm?3#pNyc?Sfc zdw}T4MUQjm<7q#=b=G2q9~LT8WD=Kq^bHb1l9$yN9CQ?FTyDKbt`^ZXc`y6(=W{6w zMbUfM4egg2&2s7!L#c36a@L(sBywret=!;R-7urq?S{HKy?)xnQ-|6B9yx$vu6@m( zXp0Y0v#0tXq82|MUA9eU7D{xxfY{ZX-xyJ{;^uhG?cW>ER;>`kL_&{{mxdTJ~Tb8Li3HCiAsj4Ja+aSuH3ds8i&Oy;0Q1%Lu%7_8W%d~SPdtS z67@<62?$U_sjr^8d|tzpk$3IYpJ+%Ou?@jijR$=>RRrpW2Ze2OavklCq)XgGvx6?! z4-H#E0h&JX65(Jl3jy*Ruk(6Pp6Q*E^Xs7dg$8V6dEox_l!3;;E2lrTgh2}1Xq z=d(tQ45aML`m-@l^y6gZVv7I(OWW%|(_IbphFi&oSuvw6*hGlX-?hD*RKTSVElEL4 zS$-bofBy4)6U;yDvYE^pf{K72B4C5iS%?DKp=6+=$qquY(9rY-QUGg&nKK1Q&rKpo z?a@6ir<slbg`wc1s1u>mi-bG*p(|N~X0PkjN=y4t|yyQa|%ep+EkD zqy1Q6u6t}94K*G$^h+I<@(i#CpTP1h2Jo>SUEu;#$0xGRqP(yBS{F5@NBw{p9~gOn zm{nbcsRKdm^P$Imvt3*@g-^)d8wSq^dX*|Ba|YR#p7O@=8ZIv_zK=X^D1{f|&|Rk= z6)hhckVwN*ePC>SOYS1x5lrY%p_eg0v)QOL}E5?QjlRqXuz z;Vjz`C8uIPE~7$|iLVmcz}Rw2NZO8h4tT+dUTdrp^pk8l)c+UE?}2X?GCyElg|w6t zlmG&3Ml;+=`B-gsmq)oGh__n&jyX`dAndNP;Y_)QZXw*cYnP^8h^UcjwPCeC-?VyY zTif!&Z6=`YvOC|`f@upX}Z~#t3ScG?Jh7OfzzT2Yv|4!Bg2*Wl0*V1P@d73 zmzn8tx6WMLI!i8FQ;SFNiR>J5+2H=1HJ|CyUX05_3N7~ivqt?!w`l8o?lX9P9;!lU zDZgN5#f|SAi!S}@g|nI^Utz9&+PktE+`YNP885M}@WuLV{yo0fzm(Pk=Ag$MM$yModVa!^bPp%ql!lWv83 z4okf@ZIY`lf=92~8A_4f`j6qZ50xSZJw=yN^re!Q$2?=~bu?ggw}s3n-SQKeu1f$T zN3fn>u)cnUTCL!CU5}n1INnR*8tr}{xmvjh4lh|u-FbQJ)NbVr;!Q>3Hix)>J@r z?vo3?t70lg_5*BZ2~=Q!xwQj3s~~~Llwj-v#V4}ItE4UTMGRRP976o}Yv`lM_xVG` ziIfh?WEZ`-bdij)qG)e++ljBDE)PUymlYg8%I7Awq}}P$#3I`oDKsuqyT*@t2{V!C zid}3Si=JX7AxAr$?c#wqdF{mqfN#M)=RTqAxAJY<+}w1lpVgc#(FmIO1=Wn#?!QVk zTSND{{0;!-+~XaeR~gKeINmR&(Xl*nSb2}|gfUknkG2$4GaV-8sneZqI6zCgt05Uj z5#re9Qh2lGQ(Mx_yazX?^wkFRBl=dZmxRE9SKonu;w5)hsP3xHaF`-Et3mP_3;Vde zhFX4Um;C;s$zU-D3p(3SnY3gxZN06bGKw@2#t#1_zL^7|k{_HFYpjW8uXVIG2w+4g z&$TTGELgQn+MtiF7ytWvhzSlC5MP9LHld;Kw({v$ga}MH@FkuSMV0&yRW+Cm2^q#6ObF!mTJ(zWHnQmsFxd~kg%^gHrD45B>9l3HoRvTXiQ~Csx1eZ z{c3MM5X?h4U0-1&~C||67TY*1*xpc(Zp8;1KsH6uSuEKG)NX zR!edaLddt))gn_f9c@cH=wfErIIwhh;^N{tUGl2-*G>AzdqPO7;}$>7BqrxDcnMKA z*JcVn@^CTkkP_XwZQHgRl@Wduxz#c6U&0aAeS72{zpjP8+_1^nr@?CW2hqjM=rAr2 z`~2aH_UY$ltCC;zKZFd&ECSgp4<}}iyfL0KKLJP-k+QJVn{MynhGeeBbj=I3Qw!rI zf;MjG9jHA*GwCB^#n_9@$1Z22%|_DXxb)+EJ&YHVvBZ9ACfBUH>w1{5O{T@MZT&QS z&UFd2r2C(WE5Trr{LTB&02plbjY|a;q|P*61RL;#!3efjq>2F^cF2E-79c6l6!X$U zaK*e^D4Cb++~@+ePKqq0Efe=r*YlB2)v;~ebj#;VqO~udGiv|(^{a8efoK_Tl1Z+M z8+kHQv;T~3(@;jo)I0Ouib4mY;X2ZTOC1b#!v7!9x28QD_4L{Um3HiYbjvI_{`NHl z^zz7t+)2y9I}GU-i0*?zYvBWv z-}0kM75!A^`n_ci%ITLhm(!Oca^I2{(Fkflm|0@*ar;wFINh^cg1dS2MKLMpxAL4a z%OyyZ1e7ciaH*3KqvE?USIf8@Q>8t_?_2bLYOyh|PP1rhG%J0+kZ3YLJ?zEfW1@F> zWa31bmY_=J+_#vf*YCR@PZ#9lNEiF22HQ0B&XBWlwxVBr&z&3ZVTcU^;%CwiIQ?QzdLkPAtOA9 zu;dCcCV4s|cyN5C!?PIpI1;oLSN%mDRi{c~0h(ts=DBNX5CD`e1nh&APlP`C7n0_8 z;V>y;fRaf88mqgSC|SmZd&w=YNai^HsUu0_hU0oltO8m%bQ<9V_0_FBq?mY4-Lco_vA0ojoi2<6(P zyHu`dAN+iAiG(f>Pg;F#%BYjwU!Hrb?BBm1;W7Rv+$?r(J$1R_#7w}&6hd8yDUtMF zK-l4iqw0da*$R0B$Xp@ACdwXjqFZMfNpFH<^H0lXLeod0S1uU7Ts+2$aVI#c{DE7k z@Ux?mBSU|=h{OStzbvtvBpLoIj?+=a4=hS$g2Owt2DLoI=8vjor2Oa7k``)n*0)Kw z8ZHG}x_9qhTQ+7}ol>zMNAZ6sF~*UvKI{pT88a0v2&0miE9;p`h@h1;f7i2Xo902T zU(U?f`l*e2lz;&pSTzy;*_&5UKKuAPCMx462Wm&})Mn0?AWOYE*0GQ?Rw04u%kvY= z|1sPa6+AU+yf9jzsh76sXLV~(j0RTsp6SM9#`KBQvXJ1_G}(U>wL+j?axU#h7Xink zR7nw0D#OAnbQ*1@u?aP#9>9-$HybGL^6{g+cMT&Mv{Hsd+8xD*a$v!=;5cqLJ=a@{MhCDw}z z&*#eKh_k5U7OSgdxonXn$Pi>EIfhqO8@~pPvl$KU@KZ~a!Mxrw#?FTQ2ml`o0DOSH zi(mAA>3WcgU+7{&WCrK$gB{^qBaZBH&MnQ3Ml(fzWqXJzj`}CHX?V^Wr@)4rzud?_ zC`?IUzIMOk0h_7ERpo~O`mc_fU;!qh!nB6&ao{=ua~27G#t|}P2e)o?)HKUclx=su zzi_iX7JEwXuyM1E+Ur8;^p@`L-*sxU786Gnh`BCOq%-YKCTEx4nV>y?Q=9$T%9Ti%%-mcD_t-6>$c^EoB%R+q*Z4cLd{WSB(Eje*Tm*R#Gj787a0tY`cSWsz& zC(Gh8UhG(CNUJSOaDOyUK5{ zd7YcbNXp^du_Un-{rv)bb3$JMLYbsXrPiN;tjXw)M1D4wKN9)>vqTcS>pfPHae&;d z!5z-q?eLZPn=4uIGEmIH$as~zy_wfRW55~U-=zLcS$t>(CBLC0#4p?i84KiPL*B+O z{5Pzid^Ez9(>HA|VzgVyT?`(xTJ^TGm^2_S$JhGjk%IUmc0(jWl z^ciEGStO(sA*rb z43@AgXA;WG@J2=vIj+;%N#(YEQbNAlI-s)ekF2-1+b$OBTzgS%i$E%y@ z1DO}s*bWqOh&wP+R-89DhkT+Cb3(s_-i}b$y6azXxORl(0jHP)BXup2QE`tq(T=Dx zUtc7tR;M;WH#i&=zDme#9dQj@1UodcQ?DE%faxH822F4B#_((FytxjD##l=P?rq3X9_2mO?Hd8;3qw0Yg-mFGD1gJIP~=gmCFXNQ~sdvbkaW1wsQue zk(;0q*6K1z@oCK=~@0iN=QW zXotjoV@R=gcXC((*JWPO0-myadgSeOKv83&zEUG0Mr~e|-71Z0cI@Ry`t+i$8gF-F zs#$^K*~Wax>5Uu+*bZfDmZm07)EN{UoPT0GCdsS`C5l8Ws?!E4)I#mX-rdxja@~&E z4*1P>7EW>|dK}TKlb5yuG`S^@5q#`lfJ+@Y9QWeh!c)9sz;wE(q5<{y#tRX)WjHcz zoO0CXqpUhH4{+n(V=Ve}QhS0<%<6Sa4Y#P3e|CKny^>{VVE?bPEH8xtB1KYpsS;Kp zlo*2}0f`i>Go26H4+6ESVl6}+5Cc`5x+<1Vu!<5Kj(?SXtaJl^jEw*7i#|?-KX&5o zzfn*OxQRrcPJTNH^gxeKQ1YJOFOc_%pn zdSaL#)x`J)%)aL#IGOFKzvE;!AUY^rSO#grS{0)Q5xr7^Uae&r4gw=(2ge6@m}6cb zwMiOuR4Jug$h87&`H1p8LWhkZg&FIpd&&a+AWN{jM7UPldPcM)EmeBM(GBB?cn$wA zyJ%s3?0#oHcpbdP#coJ~WIMbHl0Cso*oiu}0+O{!6zon)cqmJ4*8Qb0?$M)1tKio9 znTwzk5dt=BUK^GGu-~|_0OGNKcL1IcLZ1U#TlAyjT$qD|U#R`7OuPTi9Q$zxiaCs| znk$vpW$`;}uI|M?gqfbbP5p2S<~aTpf%?(asQZN2I7;+&6h6%mJOM-UEsuR@SEUuE7;Zc?)Y5pX1w%rC67|eb~FN~ zJt*!$scOfXE@U+nWL(4=9bbdMlRqX4d(pV$KP*L66>_@aCI^7W^Q$?<^7}PaQH(@E zCUd%V`hu1j2Q7ykPS^@3zPIRCi@%^(9YcudR24~){b3#SSeQ19OVk_Eh5@x*vqq>l z&%U~fR5hR35I$ouuGgj8Q$Xja%-b=D%V5x>tFOg5I;QkwanwbW*-s7pH2K4T9kZVK zN>u{>awU%r^b5cUa0ge@ZYK{1evu5@v4V&NPRd&~lMiLg*VOet$cL&6wF@QDIjyZ( zh5ZjHIClD}nvUME&)6_#C-h##@4wJ_qy>Ke`J?``VAn|5DQw6(f2$tg5zJ_Sq(Rj%!>*gP<33StILVY?Cvw}e9E|r zM*a{H86yb&^3!}>Sf12xQskb2CpurjkCeDuNuIoBN%7nZ?Sy!4 zn=#hc$#&3wLgq$$**C{(wUH|0!Q{A*J9T;+&!%LDq3uJDw(A1!9*~`)d>`6&{tC5( zC(!hHTDS?(Jv0m{Z%&+^=~SR%0O`@ymqa>lu`I)Bi~Z zCuA>Dc8%gl3rwchzrvX&t z35|CfwLMOeVeDs*vSBnqsIqy-9+%(DWh?5?R!3u(A`&F~V#7B#HmXicm6G&?c(Xr4 ztnt_h!+zD3Ws~nCd%99~cA|&!##Vpq^sB!(gI-=7|LZq;sfFnpbl%;3ob*T3TZw!Q z5p6mVr)&ytV*>`t-;DFn)@|~ycV7OD8%=V;=+UJ9$~@O*L_kQDaz^w1Z)(rK0}E^T zj8rb&q#yLT>}+y2F|th`*n;6uZ><;S+~oU+jLv|#F71E$8vp`-strXi9ucejn>;nbBa@So&a>A&L6sNelRV*EaC literal 0 HcmV?d00001 diff --git a/contrib/machine-learning/assets/cnn-ouputs.png b/contrib/machine-learning/assets/cnn-ouputs.png new file mode 100644 index 0000000000000000000000000000000000000000..279722659207e9add1d851556189994dae0e1ee4 GIT binary patch literal 94050 zcmeFZbyS<%_AVTX(^6=Q6k4RXY@o&6-HSV=SaA!k6|7h(1&X`7Yk;=6J0(D%xCfU& z;J)m$+ud`|z5Cq%zVVIm%NQYVUR_yh%{ia>%x4j*sw{(xLy7|efpFzyCDlM6j0F$~ zl@bdB*faV$I}8N6BVr{Xp(-aKL8Z{4ULP5)Z2SdBZ zh?4zo_?9f&n}`o+ShSs@sufvbIHiV?Z9%2)8Qes-LsFzFPCMSbsbnSE?oA>Xb~QeS z`OMFr!$mBwTvlfjyy2j9r#JM`%jt=r<@bj<6;Dw5pY<>*su5Fy3`J4R0`|^4@4JD) zXaVNwNN-PfRLX;s!YqTWG~{*7E9`fkKA;b;LdMWMFAYNQKmpqOw4*+t5NFs+`4-7Y z6LfyHhm`HNS9m2tip-Hp`yQirK;vu@sk&`L0i>G{e1R}@s%Icl z#SmW?7kX?;<-qSmP+7{jt*TMK;hwW33uroBRQ$7$SeNGn&gl#)AHOA?_~$P9$Im{V z4#2a~u}T;n);Ui$f!Uj@;KZAfwhJv*cQb_4Jh`I0&#?VOx*(x)7%fN0x_xM=Eg3~9 zBn7>>%`qf`@-ESuq;&pSHTd~c;+OT}!v&c>_JZvQd^Unr^o{~sZh88SLZ6J|4Yj-a z9cfE)gnlb6eYJ+53_uoHQ)ZnL}IY+7Jt>D%C_@Jb+Yao#Ci8Ji+%cJ41;r_P=Jv?l1I-qAdi7wxO}>jJ45xK$o-=!WG(6o_=u0 zLllQS0H3$;^Xq$Z@5!GvcfJL;G{^6Rxo2e|4(r`pS{b4k2s(AA4 zuVQzb;~AwxF(I^X=s(;;eiOjc5)S9V)bhM;UBmnizw?Ow9+NcpLUh5i7jk!~Fz!o) zevuTI#Xs$U>|=Qx2T!0|{zyeVJhNBvkbgxlNM%3|#yG##E=DCo?jbV|Apx zyf~C~^>(cjBlIlw?~h`itp4KTg5m9VGIajr;@0<|Yuiwk_7hXfikR=7+gOvJ&)=Pj z@7Z)j_U*gu1^ zhTd%%;N%97ykKR($az7cgUi&0QBGou+0-UH42lehS-4Y=#rk$@fz?1LWdVI19o}Y% zCMJtTW6XZTrEG4Q5(B#^C|^ymHH8(F z)0;GyJCZJoCBrT&C4DXzk}QyRX8C>Gb^32*c6_a;M z4)qM7GHQxt{pGPZHH zk@*sT!|NE7G0JYrnHLRWZ@>6T%nrkzb~?WQ;q`4SNJzpp$O~@;kA*b-jaH!5SnQGH%^O=&WCfg`z z)3ha28*1zfJq`mATK5&Y0t9$z?eu3y>v z^rJ60jqGY*dVsIXGwg8g9+dj4G(p5G(tT(LED)}?;cQ3aYqr-tP5nrw#o5)7u(8w~ zveN1JXs5QVWO}ElDY$7IG6W}e#rw*TvEt!qnL z0XHIV`U~>_mjKR%7j52=#Flhxl9WUb@9W(EPLxgd318#h5#@J5Ug~=8dY*O`Gq3JY zRqj%S02BxfO`rmeN53!O+ChsfXj=Eijp0UH`OjJU@^3 zE?jC*iupZkP$6-ObKc`B`8>)W`#h&s2co0rRjCVn4EdoAWZH!f1Roi>q;?ABj_f$8^MsUte9>u-^0f#r@t9<}K}K+Ofr6 zO8x+lS(WIpkf))+L|?!>?Ef4>!r65CW1UPz<~F{amXH-}s!8gV-x z$@BO?8K53E>fN~r>#fbof%XCPq%?cES?3d;Ne3={+`2e>WTR4ZiK+MYNgQcZVj`Ke zK-^?zO(k?UWy#G5Mp$1|F<4KZSiKZYlE=vGJcnDMHo6B*9eJ zj7C@;LvRyt$s&3pc7ng&<>K+=ArssNd(7OcqK=hujbas&7UI6vx!(V}@zD^alTvdEU&bty}CN(u88lh;$DKf<+No~+999&a`V)=1hEb=DQlx~v4g9P&{@+A z9CK!NW|Pg+z`A+cv!TFOtK>M}bJssBdbu+GK4 zwNc%1?Qo9o6zf!e*9OMs{o*Jcf!8GB=3ejq*!SUC`SJ7Ftfr{*C}Y7M?}HqdD|N(J zrDg7rb)!LJBpB1(*iZPxbM1Cd*pN_%zco_#c=TlDb?fD|^$ywC%(*2g9Kj z;_5pyOFVbFV7S{f1XmD#==1$5`gngkFI#0uSmj(Cp>{Q~Ycp22=WvQ}xwxPeWK`K_ z4*2puHxG5s>k;V44D@i%N1Cii@R|`PQ?QB&|4c>4Q(A+-9kjV{?X8G++0=mDmddhs zhJMV|iGr_$NeO$9PAzkH@^w|>q5&hDq{rRw50AK5cyKS1G6h|(@K1+WJ`^>+Nt^LU zd4=cmw9N9cE09sVH`kG~P*MUh0ozz0G!#-0I8L-qQ7M6AATf0bIXU1{-PGCK z+}`Dt1DM`JC>7X&?I^430s;}y-n>!d)M)mA`%hVE=zw*U6!}dZ>{w0A9A28Ude}MM z+y^A+!4GWOnS)IzJ?w1lUHCnOsQx^IAK1Rx%|=D}=P6(tAu1guRZ0m5XLCw!)~Bpb zsf2MTDJcb=%`EuUB&Gkl9QY4$@m;QF>f4@}I#oSrK!47yPSorVp`s=yBAN=b=LAINp{##f4tDpbe3v{$F zjv(8wK@-NoZ2E2sjN@Y~Nfiy?6PRT;FO-}4#PF}ro9%-bY}6q_5J(gxCn=`kfwDd4 znXEEF(X~JH{nU4;CFbKFJw#Llv6q#e}2~TE!;DuE34%+8Q#!$>IrwhI%`RB$y83HB0$3z1)==IpO6qy0u%$u z$9!0~uqXpS|Mx$UAe6^=xBl`o{;g~09TIlg*O#6p+)Vq#yjE^x7n3y84L0#EA8wr=$~8^0Ecv%dRbg7W0jx$?A$7JPyGuPv*G^-R?WyI}t%SWn%na1= z5<`hZ9-I4x^eYi|Gw!_sCe$TSs@aZWo(DfJN^0x1=vsKBEQV4XT;X82p<1b8fvW!6 zuJ6kCebbCa43#0=b&^8*;0Vv#L_PJPF(px?gt6jA51~`mSV%df4J1I`?r@Ii~^y|o;$7n@!Q>b0nd$nL3w$yX6xy{nU@i>e&mYhYD-Z_ ze~f#p{PnoIy__mbCz=vR|a&>jQLeCLf1_O14lknT-|cuUYXM7 zc<7$13KcjQgn&vFqWli#{jUSpM{|{!zac~(0i6Oo<s=r%oJ~8C5AuN0Qj=#BeDPZJchaYdwl}dEl`Ufn8~~;k-pzH}4fQUSIfrLK^*V z=lwbxh@tTWyZ5I;e!5{AvDefxk=?li`sPckER_!Pem%*2G0WYlN(XCw%(_)(k%#oq z3YU1>@AbOC{9_3ocy_qlON)N67CXDjCG|d^gWM7h^L5@n+gD8sCm9WkgpD|lhp+2a zxj~9`>x@Qo<(LC;R^t^8)?&sB)PoTM$Vl>7B<5W`yN&*5?4M0LG~D3R%#B0|`k9%T z;;*H|F&s%hxxjxeIQW>05o8#(OCY3~vN+7y-c^wcS=c&7aMNz#Q5rN1Ae zUUs?*+tP!!9j^J-<5itUJa%zM8)FJ1Ujvij8GimhP0Ua_&l~>H*<3yl3ic;6Tqip9 zt9SK{Xm=m*7KER)ATiD65V(WKmQROVLSIS^_=?ozrs;Xg2yXh3-rYowRf5mu*Wf|w zpdc;jnZ8m^h*^+rQ6+SI6&a_gt+OjBVxQqW-{W z6`WXYc?x@*Oz}hInD6+SyKbIog30)Y9gbZTi2&>!DYLli+$&gdDBw`3$KBS4O)SZ3 z8~oYW?S4$f(7ISyc`9%;T#h~78x4jgWf{wH0 zz9Ckt*kY$(iR-=!yDf)q4T}djS~I~E{;aK+S9y6X^3r!UX^nj^-d=LM)IJt`NOoyO zkGpRFYFu=4s>FQvrDJHB|M19FGM7xQ4>98CE%6cesYpI)fn&X%lY8g|RuBeMu?47`PNH`nC4A zcglvEg)i(!+N@W%8xA7QlPX-MSnK$_F!As+Bx*E$27i`e7VE)E}W_OJ4VG&@~VwAZp8R%_ItBaa^X5dpG7yXcS61ap+m3+0GFmsT(JL zfL9XrwEI|ijNs$7wsWj+gQ9S@CYK7j2D(41i^ZJ|q_>pqNOu7Soacjf%uG z|EDVuI(Tj5?AE1h-Y`S599$(<*MH#9E&9N`uN!$fYgJ<<Rzk7+U<({IfCUgHhOh70_*qy@F5-By0nvmxwq^tm zJ;H&}QW*WWQ*};dL0`?rx(m6Y41gZg4=zw7KrWrow3=$WihzhH%SU z^`p;98`h?#G*|h&To%0t*L5A5-1Wa*W?Go7^=cMD1~SFe6VzyvYh^JUuSVXK=*|8y zQKWe{#T&SH;xJmG6N{_0R=mDLq}{-(q zV&yd9;78ZkEC4GJz;jvEguu|rvrn>Uj!r^l*G;(>R&Bg{4I`DNYlHRVotyEhD9Os0 z4Kv}6t_m$ap2?OfTUD$553Xee>*f=T1gdSQ;|=w=$7$7uG*m&v{BmgAoR{Np?UrUw zahb68Q{%3_QoR~Fuwn|kPF~5sSr`G3Zhu@fTuS`g!syaP5?h{H6c-y+v2OSFeq!1v z7CM1f5)L++UESqLWE?6j(VwvrCd7!~nDLTqTJA}i(G^gf(tZ2kBfs1#E|v_x ztLV-z`t0WY49Ha9UhJ@Hp^qrZrE@u1&dpM&5tZih{9F7RM^04(8j)8u>0sQ#RKqk! zH}#Dx3F)9fvXX%#KW)u|WufZkgHE)inpta`>E?uWS#A~zl?$d9%P&7AI9I<}=FUQQ zMb^d^W%(OS`GPq$({@u*81L1T(w3sknp@!X-QCek+PHEP5L+&Qs32Q=V+m6ji6C~3 zzPTV)naJ(IM(2wYGJIbnp5drS}p&?ByMMYLd2;9e1TN-8_-Yqr);}Jlw~atxRHq|JXLAsxlPRGq(XL5s6G+LtPZ_=`-aRUO-W$h_DXtE0Li1a6^*#ss$lpH zZ5!MHU@h@C!)1XKJtq}n6m5XNfSko(s5pf4BTFY z-`B|Z(9Y71JI@*6&wnekM)z%xVV{Cbmyh6%&H22;$QW)T5@Pb979HJ@G!yK%?AMdV zR#r9h`YopU_wM*_;)QzJ#Wy(2v z`KpVY&&~?KNo`+FeyWUkiL<~NZmm-m$AF~wiSD#I?zghrkD85)n8DPVq+P2bVFgyG)GK9UCDtA$@eEc|xj9uv znKM^i{jJLo#G7;#s2KG#tD)YSv#a$0fHY`q1_+9;7I7CQOXeT0ZRjGqM!b|CXjuYj#yU%p z(zWoJOy=FqxjCG;snU5$YGcwZpJ2{AEIO2}1WX!jdMYe{%TPoym>$i^Ru66xzjhSG zC1%6Mb30D(y8Ln>Dbu(i;oFmdsczK$9$_3?iw01S3! zaGlG+`@Aw|0BSyx@mv)&wYX}b*3?;WPk*;%%>VO~G=HJF68$W6#rXQ7RBm z#5HNioXst6mc}3`T9sQ;oq=9yIJWQZ5jUSn2Fw!oXf~7jp0}$27nRD)n_KnaB<#G@ z*vPF8tLyaV5~Np~uh{&taTeGudubs3pR%kJdAC@^sJ*~sEM!(2ak6JBsRvfy6ZHw% zn#R$-yVvJoLn6Ial=jO!S8*4>5#NkZg8~t0&=B@9bm3Q%`qg;{`i(w0x+0FrWo4fi zj#)Vl3N-SCgl*b32KUcPJHW^}0TEjz_xvtnfc6OY*w>t^e$Nw71(R*-9ga3N%OU~h zJgMj=!$2h+4Xb8lJ&-I8gwxG1Ca_yzCBY_|jDc954pMFv@J-!V43~8qA|}|}6iFS9 zPUTmdWFbvddG@2B&OmcT-V&)~7q)yl?|rxSXZQe6nrtuNPUOyaopyfu7*VQw9;RQL zDqoCD1WLOvE2&En(9)i)T=Tdu%3LmWqXU(-+?1XzYQD( zZFPuZ%`inncSN98`%CuCp5InlKlFPgGFLx}7P2FVl6mD=x|QT9!IUEG^a5_L?D&9F zDJK%@nLQ>n=|oM|=80FTe%l_}+>x+rKL9wo#ZX7U*8r%_isR|!{#b)=2!>#!b z!_qNU>3A&``?$>sCIQ?s2-isXCtXrx?({*a@$tZ^NKgE{j{~F(WsZxhNZ(E`f!DK; zYf<$r3WU54XdH~8Ur5E2)F zScT_nVlCQN8RkO<#N!D*uto{tdmtrJ{d+uDAr4+}q#(ycZmAN{{9?Eovkv9toGVh; zunVpzh021IxrIoJoIS!=dA375Z^^_Ez-)U}@iy}6m?-7V!SO!aBS|vvVrPFmOG|%v zzMCnwpR=q)j~7XGj;vEw{IAyj<(u)R0sT6?__6b7WwrWYTCJG`CIH4t$917jjqdTgG1W!R=Ti7qcBIWk^ z=M!9qnpfl=aIUY*-&m-9(~5L~E~K!vh(mKBX71u$M6Zv%sEXwdzk9IDFV!w*ea=kQ`zQuJtu|s)#u+}k zmWZD8t*O24*2n~&fGw!ai|_g4=}Yb7S`cew3a#BknHzIb7oZb+~p> zY`9&00i4VT9|S{Mk$f}787Bo3IpTbW-Hm%fSfc>FL_sNv;WJ+Og6dDWA|4tc&B{WAS>c%j$VGa1u)L6 z*HL73Q@BS)N&Zbh@;+K`B2>h8{{};^;r7oyz(8*l8&XhJ&AZ+cZH?Ch`q;fJ;@{Gh zMj%~LYkg}V@+)1zLVKBn!O9x{kh8{USoY4TPB0$Dvmf1{y3J;3shkf#P`!h4wT>_y zN&3g*-(Cos?qtQLQ)D33Rjpf>x>RK0uZMbCkQX(Z1cuBdJ+8<7q|WmVN`p3YP_Asx zDmtwsEh>^Wr0^kIg5XL_xy!4!%G3ZVc+O~4T+Si1bcMt$_R9*|SQl$T3si5!n5W_p zTL0=Fr;~!De*E?g{@fGIUX4k=XPT)V^;rbXG_8kjK)NuNXy5p44Iq9tL<+r<|aV{D!$L(!CvEh*|5xAJgq}~$9U5(Q$M+wX8LruVyiM4)A+-&K$oL^m0RVD}_%8j6juoY3 zybX}AD@|QB)IZCLp^`6V+j}x?SUhjCd$y_ZM(E74Z7UKMY++bQ_EE%vyeCcNx{sOm zJz;cg(NvI9U>u{m^oGr(?HJeBOWJkJAA$`o2{mElw_T5cyq`%akN+sX{tBz)0zFXZ zYRPI&;Qr8s9}Q}K`g8hN)-p;S*s!d#*$@$bNvUb>R)SPK0}0n3K1|AnJiF)dMq#@a zwKz$~-?;6eFC#Xw^{AO%k#;{Sb1(yP>B;>?HNfTU+W6;bFc{N0o~A!nZ}dwG%D=i# z&}L}Te?uVoYfm?u7u_M>&zwl*kh!}tv4fq*qIM-V#^5zIe`q8+YWALJrea?rBF9{Oq<+ z83)kRfTEb^y5r?kKvSb0VMA7j5^Su~Q@@p`v3C9-w#dAraKd>ZU}u{=f55IzwuUJ- z**ZJ50I>Dz%clMWiz$eO?{Q@Q|04_-9x)#CKbvkOHC)Q?8jI2uL2Tg-L;A5QO}n-k zp)#G5kEqhPA@nSg%E7c*R3xzIfArXY;h--Ya)RcN-(+*ii?mMgnx9wS4 zWKcNgS_nBIA^(q!2#$#F0A$Wnmv-CY>c;5=bHW@g!CK()%pdY)SDmvwEJnjvypT(= zZ!!QeI&(?UQ^;dC+g9A=7BP{V>i26o5`{;M$pZ70=1|-a@?eY{kKuM2_uReCF*gEnxo<6)@$#C zqYthCK4eSP9lo83&yn4-Ej%@J(c-fSRI3P_ke zI_KPWp)8~xc5}6J=MH_7I{@AOWCZ^^ixr#@kpAZc|@HDPBB^^J^2EU(-4COs6v~fK zFpZ}SymeVVb^_Ip>F1Zf)`M?p_$xi)`Q!y;_%?A?=|{Vm6n$SiPCkV+#`u5IxN*Fx z+At4+ZI@R~`n%0bS_NJ!g!U6&xHt==ISTXu?`wN3puRF%nUhlIo!B;NOzxfjHo>h@ zHlOcR1e@>S`4?Xr^WIMxe!HEgB1oz<-}4B4T#ty4!b9M=?deKA_{O+5Qt7GZWxG=m zBq0G-@!Ha>OU{C&0pRJX5ZOvelZv=dA+V4bN-~c|g&TJ)p~btAb?y1;#7Bw4wFTG? zNPwpq6Fxs!j=sAy^BJEtxT-I$YK`CK{wC)qm^0FKNoW&siDxw|%Kj6XT_sBI%jLMHNA77>p=ax_nc4R3DB2!~LRt0YIc z2@yxq;I%&dq$o{`)eMxk7PXO3;-slrpi%MW!RYm6AAHrJLv4p#Yw)P}ool41ges^w z*f)P^POZ_*Y3ojwk>9uQ6GPLQ1`{Wmw1PvjHmQRPhaYE5#wyp(!4ZNERx^cx>-?{O zSCrgO0)o}cCTat`U%^Te-N@VKvUIb-Pd(VeT=mJPr&LQd9DAxrFF{)V#x7&$$7BWP~?`AC-SUf_Ynr>>f~r5vKRp!<2yPqO#!D&N|nkBnPP*4rsiEl`EpE z;))|_b3w`%R+I%(p3B|$ENNo=jcazSa?=!TQl(g;kMQ*R6o;Z=@2bxGIZO2G?82OI zB>X|doQCP-3jGfy$Xq{Dnd8u>3wW|$dgnq|!JxWcMVX7QePvrY?M2COkaAip3^h1d zX{Esb;yp1NRnCAro$IEC#iv=dLmqpn@0ObrJ8}TU)ig$4V7cjo>FBi&5G=tco}{@x z2Mw3-IuRijs@-{sqG+6-{SL&cP*I>%dT$nu73F7ful^w#xUlXEwY2^dS=%&)DPr%V z*WM-Nn_&(u@eZxmjJ1X(r>^hocMZw0lbBdDZt>#gA?}dgd&o3dplfIk*}&LBG{*w; z$O9Nsb+!Lu#`}!d!M(Yu1jtmuFRm+===jt((IL~&Vr|!njpM>gxWjyfI;Skue<9$y z^kG zF&6VZf|LO;&$hQRMc>`p7Z+T5Y|9PwVcrxeCyS+l$hY&m3rCG_a$kh^9d9!HsCBQ# zas}eM>yumcvcA_64^bYWIifId{w*7#(F1&tq_<7|(_cPl1N*YcHjqPTc~{8$CFJb! zbX6Zv2LemicnWOgCj-T*CqPBT^gB2R=C|v%@sT$B*lyrkuH^XX1Ezp6(=Gx6H}3*acWp_4DrWrj7OZ^+hv2{{`>LVSSN-5~|jb|2AB zjQ;V?NJmMK(OYQ>tKpH(@RRoi+p{(L2wYE?VlW9Alz_#c8=KG+Dl=oM#|BrtIGf)Z zE{!m7v(nAe_20WPgK9^{v>FR@%%w+m0_kn*(K=6iqL$h>Hp4!a{B?~;mE&~^qJUJ! z>O*pw0TJY){#up&e8o5RQ9u5YmnB6-JzR96Hcel7y_CAd|{DNLvZjzX3PxU;;ZFPCg;8t!E>VnDR!&7Z*>jK zgM@-h;ke_~^Hb%H!qXy)W2VvaD=v1QqoPWCB*y9eY_t8B${$Fqc_A6pTq4&90ZVY4 zAKP4}IyDhT+)Z5p`@pUHyVHZYy5uPt{yCdSUC!~z0Y$iHcU7xjFy`?txtB4|{mrJ_ zdEyi`mySZNN|Plz`eK&n#rdCjvUr%a7}%Ej`P~paU-kJ#M&)pb8{22Vu-fA*t8_m1 ztFPtoDDQiu!}G)WV(S9iwM}Kw9=Xa&YXXBp*@t=nm)x)uX948pc<5S7!@T!-8r$-V?+H@3zh~#y9RRmlJn} z1PR)YM27=npoDx>=~F$=h_p7k!zlh#?;nJg?@Tl_a7bqJyqm8p#4bw%Ab-nt@I!9! z#BB2VZ<4__Ddrn>6360Z2#3@Pz?cV0$9ag9EBof#q|>fC?+riM;9B6bfsO?8!&q_@ zi0}-#Q4B(;M(o4Mc;vvi9R3Z3y|HR7wu~HKIl^pVm6|uwfe}V!h?I<T4tmy%1#BD}raa9%`mzm^W@ZcV+X5Nx?*x@* z+l;D`ju$;xx(MBfkO;o3aBV`JQZshc2$ugUJSU)_Rdte1bc;9m$@6P?d+39MgRKX* zWL~Cjd}U+Nmga}~t|c_?8QmxTd_E<3zK}JPopJ6Bwwu#aV7SeuOlAe2$~ns*uY{{2V5lhY2Awz9@i)0Ecv1G)GiHN;yJcXRTtFUBoi zZCa;Ra$XNr0Kx?JC!D$$6DRV!THi2Zoc5|wHEk~^1Jx-ACMP#-B+kyDk}1|wz?^n95N@iYCyfD=OdpwRiicQZ`2rIU-(Y~mKgt%6#Q*& z|GxD<^5MUTmj96t|Nlmn{<@vH+_O*X#>iY>uk?QU2vE5Cc1|V5e@Rze%M1!KfoyEV zT7fAR?uSTw%z8*I4$)D}10`={zg3Sf)y~7|>0~;7UazaA!OFBiJ1=*F zXi8o!?Tzye0+)RsQeZ+oKRe`~3$h1zANE&g40T0^2-U8p9B|K~)+54SWAvAQ`&a~` zUii3`vgY#3V~46zK0&R|l8Ndqo%1eHNfGkal8fQ5`SUTj3n#lXY$Ahantq6R2j|81 z(1#+s3AcFXtlGoMdZ4ab5*OMG##*~J^K=@PsQ>OmZ~6_W5hufi)5v}vkp-a`ezRLz z9iZEWQa8EnxJ}|H#B-OJ$)*{-CuXKq96j8gX ztRHl=L1Wkm9;_udK6L@GVgik3p+=?o7Wd7Re=Oi;08C?AVB&+JQDg#iq{5zIO;L3N zO(n=Fhwlu3YByIAVMMGn=Z7n?1dN&e>lKXb-n%{y?i+wYdetYVn7b}hyV0~udCfz= zpCCz>!!qs`AU;KL5rpi1exJCm=dsTMutRQ-!PmHtc=`_Mcm>=pb&7+-SOSu+z6~9m zPt6BLJ|HBd6>1WG!slX!S5xu&XI%bETp{XA2;hWL_tc??pGPEhL}k6z74LOibK8aunnE8vMb}jD2-IR;j+SIqx_zfri8D?c80Zh`fUwWAEvrzmR z4gWLmbT}4AhK7McO5()IWRksO$&U9|V5Rdg0ZsljU!8op>BhTEbbyCN% zmPC8kri~zJwd+JZt(hBk8X)^CuiH?H=JZEKj9??~yWUs#9J`_@;tSIJ7+bHNjqTiD zvN^GxUM|&QNq+tYri}flO)7zF&$`?Hnk>fH?|RHz&&sAWM?UAhR#NMzL$~a7g^Oi= zgHH6gd@DViNaUD>zFJWal6Rb)yn?_&z^!dsvHFysTTm+LmM z`1n@#Tln&@@T2EMCJH3ob5s$z`e5*KQZGyal$W(Ed~u+^2OCQp-dOfr?%X*6Nc z8Q>00e5=1aYTKEv(#8R+5^)7;(8>#e}CeG6Ti1!uGzXVK^eXP6|?ib zL80YrRba$h{cCa8THG0iBZc3k=lAXgfVoiUjmI^@91^1VouU47DE<%dc1&JYaYtSdHb6_ak^oP4K4pDbGR;`OCkzAB0p*v_yn46OuxO<90s(~KKX_l$my z)&}H9F`0q4`Li~;9j58(l4>TZ+lkr|l2sc|$<4&uB$WM6{Yme;H8&}}@~!o~z~db= zFP$$R*}3lp)L!7eXQ9~NuW%rFkNBmQP|0>bR$F8z5gGQ`pRH?`aKVl^j&`rI*P*U@ zfIf!-K+8_GONXt=bbEcNEL$Y*$n`3=l{I6(Iz)RdMzY&ob~|W&24}Hx7eCQ|y6u~M z%eiUE$LOa+r$j*uihN8BxV75IWB{LYK+O0U^$Xv9V$wALXef(Yc`0-{y_6ecc~ywU zSQntwFbQ^vA|oL3yns(Psd(W&Ty+yclF}Ncd8pzn4_Vf*8Mdf(hLc&=Sam#V!}YPR zwws{iRoBKn_b-G}*nb4J&6Y_@GN)(wCJ2_~}v0;gE7ht~tQA3b#kI${vgO&F<_r z3;SFP-8ZiZDb|5Vw_L7z7is3Dj~6i1+<=;d#Ia3BB0|gT47Q^~IQjga>H>De1W(B~ zfO#L$6&bY7{w<^X(>;r_-wP}>UqT1_lb!S4>rY`t3L8eY4tOhRN@1z zBdSZl3;or)=C(!JzU2XIcyGm0*P>4)9q!hRJY5_)L>n-6&kcx7_xR)dscS*IdDlof zKpWW$-8sNnS%Ed_>Dt+^MC`+R-*jBpYpzY)6sN{?$xI78hLS#`)|+-k%B@)<^wM%A zD>4t(QBU>$502jc#?kLY9(w-m=n1~pct>3#y8cKI@XxTFub`gfkOFGqmW_f;p3ZAv zL@Ll{A`V6yljlDk4w2}Cc;N6rKty*jc}3V88+);957aqgqfI+|H@UWeg#eovjzBUP z161+!;a0RjMcc$q40ok3_Aa`AC|`hp&sq8Uv1ea`G_Kc0#jgdx*}S3f zqn}{X5d|G6&rNGT_@DC|VVyo+I^lr&=>ZLP_~2M5%Pu#p$LD1cJj{#NO8 zV9;1R?>9cY0OMg}Pz%;v{Knq+}hDmf}qtR`c63 z7jTg(3)wy;pz_nfG{m`kP3cO)8 z#4Sf3TJJ!Q2T5)C`TO-rdGhI%N{ir}EdMh3HD7%ufLI)c&SVV(0=CT6WEH{j5$KeD zD9cTchL8+!wKauRAw`4TfOQpeMaVi4%f1VBks4_AB8FolvKV{)WM`%gT9C+Map8O?J?05wc?AVR#{ap8_F)~%Aw4JR3@ zSMWitzkyW%Ev3_CuUc2j*>uHU6!AB)^Vb~jQ39$b4RcPyY(EDP-{keQ%UWOiY3@6+ zTAwg$msLRR8}_{#awjL0^@pKWq>8Ta%Mk30NC5F*(z=&E2I}pE z%zBdK{^f67ZEprt`Qk1apENKlw&#>alPI^JJV+*F(gA9=ANEQ@V^)SUB(y|IJW%WL zgcB`tmy)fHds3L=1;;BO!0N6EtNBK|;5#JynU6)BW;dt)tYMc$PQ-nP1p?e5@gLko zaNYRH)f4?3y%px-e_K@d=Om@!0mzeOx32T%Uzx@`Km*_LXBDa|e=CKM%}7!{o`JPM zJ&!$WIj(nfh573stY{|YBZrJzp3o~&S$9lBwo;AIm&yULswlG}*Q_zx^FA{|Icr?< z2Wtb=jB2@#w;BQ2PF~y*xbjv#nSgswfC*bSux_u}kjA$WthQgNu(B6Tow+9P(g;GE z;@RqcN-sgZ&O4eww3w!f5a0;Y!~AhuJpbi zJ&7dBKyrqA<-+m#+_rV{2bwanesC_x<%5)mjsf;-dEUGak{5AU>yDPZdjt@5eNh~K z5rVf4@ z&JtNRZ(;JowZJUkKX5ut70R$-Hm(hKQIWv?kA{9wya93>5o@NufE@OdbamOykg|qD z;-C_n)l@eBmS#FLYf3=)KWSHKgqK(|0YfgB>XtKjnZ>%vMpFeYbBu|)c^;tvYLBub zvdd@I1#rT}X$Q4AO+GX-j~Xrh_Bg9_`@+U+;106kV>KK$PXS z-H)HBQR{X4cKJ({9`bdIj)DDJ>OW3dCo*^%27{wQex73m{&~ld&KQ>dQZMB6i=q>F z^qsJ|{x-jU4#Ub4M=28ENtu2A+D=aYn-jT^#fK98N(_=tx}1cI3=dYiCg$ME?-ZGK z@Bzg87*iokEC@435V84ebqY3DXA&i=uWhkY#iuKj+-^*B3qqym`brH661DxFa(2BW z6@u8(s~=bL^QUW^)ILJwe5pntL2Ph1@uo#)%Wy#XkzBKKi1R#cWjYOlFU?^ zszO;P{z6}3+2=)^j~)(uu0J~Tk`p*hnfP6+IQ^@@aJ6K$B*xqo&141KtcNFWi&~c~hPbpr~m`}2B zq^tayqWy6t4vygR4Tp$v8bLp<3A7Av9{hv#cyIcRtF;k7VC?z)8%py6z%=0T{_=k& z9W>wG3gUZM{RN+Vk>^>74OSE2OsrEo_@BDgAB*lpPH;}b;!%J9FMoa=`7eiIkY08$ ziE{k{zW=$>KmSrXNls6`3#>5wNKbwooPT{{9U$2r@<^S$`S(Zh>upf%YgPaUk>rq> z9Q%)h=$E_v^8)KVKn^Jl?M^!K@4xm(wag|-))EI*wOQQ!dBK0Ztl$5k1QguFRA|Dl zyWwAc_tIhVLE;d%8S+2h#1uG9=d14>PM)6bOf7Q59-3|c|4b>*fAzuRDk#vM$8e~` z>Bkx5sdAIwZ88vLi|PWx88g6}>iw=9$M+E@r}Fe4nZfUOdg%eU^W)xAyGE4_y}kT0dUW~L_uFAO=&Q9d-dNo9l$_?b-+g1htmg& z`YOswIla0I>}H_se-}N3E{nNeo#Xu;(Jd};lLC{jC{LDGTz9cTXT%uqBvX-X5gN|b z`|P3j#)}2(O6_s&YD_9X97=X6I5INIb4&i;m!SL{I0yw64LST(%9;LO(Aeug{s1cN zq0UrAu#u(<^zvoB)^nl&p~u}P1k{;2!*AGBVg-<4KBQdnJE7Sh9{#h!`D1Yc(k?p; z?<@abwD-l>FnE8Wn{;B_CQH(RWl-@ zCdDSQf`f;LhwdR6a#DDiFngrHheS*q;Z8Sgi%<|UZ(}3(eB0uMD&y!pu_hs5ULmVg zRA(O_%RR{=bqVm3j}r2k;@cbpeM{sY&xzUgU78iQcdA~QYX2==?=LbhTqo+RZg{dF z)0AVBi{4Brh9hk^c~)d>qxdEnc#KL4#@nL@@fpNoRQ-3Skw>dF9}MtL71<9|h*vtY zS;qXQd_P2zWDqhQT@}4&cy)q@i!VMqru{j?%n-xT{ivp$>WSfgT7II$rB}QkLjWcZ>AsJtErS?m}{GdaI6tE^M-2z_j$OAnCS&|d2oBH!F=MzFO*iphwULdk>y_> zJV@_fquT$Ol|XdI81+(5+V9Bi9c&*kJ4;@-z0>g5%zMl=o!W1V6DLoND=#o8VMpfP&tyO z2O;<6Fmile8XCtdkpHOEXJctfc9&b8xH&knw-2=u5vH)qq+qu6K}F;}>Q$_w?1L z2Pez)ljZbSzVa5fcTCfx;)nW;;`cEOMLC{ zy1QTNR%*SJrd)wKp3_~`;(y*$ha7}0P%&AYGT$Pt3(qjWT(Qjbf_K^u*Gf65W}e>0 z`_*|QnhbTs2xpem=2kGA(Ylm_GSMKG3;;6#``fK^(Yb>uem9X)NqGiZv(^<8UuWS1 zfGTbFZSkHl78HBs3=GaZP0O_mbbT=by4P7|`iAjm@j%YAXmxp#I-?phj_r-~g0tc; zfj*8y0#}Xv1M?kG)Znd_IDU2_&N;Iu;|#MT&|==D6}TNfk6yP(_LZTV$u+3a=B!=` zATtASQYKlhuA*Eni{a%A(Dh_#zkt$wbuN#~Y)X!q1?VaJulasY*YCt%Og}L{C@V7B z=`*$yE*Fj4eKSGc=})INIxJkT#2lMwRR*3u9BK(7z4}u#3oNS1UN!W?Bb`1d(uaVz zeqSEnrO5oa28?gor`z`rV{`48IVPDqCu;H3Bae?exzs6nV}@2{XL(l6*GMkOTCC9 zC5x`MeI9=x<%`nn&R6(u`eb--EQugJwetb5H1q6zYIvEb90yR77@%hN^6;~8 z(C_8Y6ZbN?qfqxR0sg6+iqu)Ro)4n$!t=zWIQakG@%@ph170r8C4V%lHsqYJwsyWQ zz4`r&`*$n|g?@3l4!s;Ju=1s?+3|tOQuw5n@i&R$T*VTyT@$<%h(3mGdmE0i_N3Y* z$*w;N`Zl0Z2~(v^I^>qFawaZG(n&(g%zm);vs`=@w!5UTJ6{WJl`ISmV3=4A=;87z zv+W|qcl4G#-P%-SJPd7qeKmE}Yn_{muWKF9e5G990>ALHR3a_0-1r_J{C}p3_Xn!@o4Ws^ikF@>wK|38e~_0mty(y3hb9OgbdYp1h4sb^A{YFpS7!W9 zir(++^qpsPi(JImbae*rJP$kJod-niMlQ@nROUNnz#$?V`}q}sNGS9JZT&he%0|s5 z+Pky4%yysLx_Ss1d19nvy($B<{W%0OfkxG>LuyxiYO1eG(td+WjID72J+zww;t@adlWM+5Rl))uV9rm(Mk^Z40Tee+5Q zXeXX9Z#AgAdf1s?ss}qL^&k=YL7qz16Xw;0UztgSO4}4m#-R7m_h@q48CB=?jHIL{ zL2Isc&o>xh|BbLf6 zosHS6SgzWr#mLn3zIOBW`w&08sNv;#1Ky#Q>FTTk6h&5?rtwRI$-SP&YVGfMqc{3DJ{!n0z3p=~|i>>v+MB zy@%03FARDLJM};=o~dt`0#ABRHCe&fW+{fxNOfL#mLfwXuD8MfuD4Q(z@vk19+Ix;A2=vpQ{$evK&sOW>l|i2E;VzY2-B6GOgJzI}&Fg|guXjAy(r3fG%9B5KDH_Qb=uXw{HEk2A6xT}=XU7>uV zE>HG|$}+#LN4P4lPF&xS$d#_L@zy;o6a#cDgA6s3c7l+y65T33DUoBxM)Zb;RIM=* zqK{j#uq|rg@s)f+X~YNZ+UxVxJKD(LVgW3*{2tAR+2=+uCeVfrJ|tQxH(-pDGTSX3 z1szem$C~0YVG#Bz?1YTd`1|`;6{7M&C&2mZ)&wAN2^)U3`eKAt4{bu10kQY~W6XD^51G!_ZZV2jD}9DA6D9-|C=Di6Pn&OpIkT*a10d5X3F)7C0kjnoE&RcsU7; zHt6JvkwNwXaoYC^$y120*ps%YNm;hefMA~_YS~`;7OdHlD*mWR6;uH;2F%qmyc`8k zn){ZVAO6?VescVh5&(;z0O&6;}Uca&wmSqt^!k4a(r<|VE=bpTP@$=Z z>m4FIuGzKB;M@KpZf2}{K^3GtT5o4wNcwfwZ*l-+8dEkD>72riCqU_USoB9vc~5G1RxZZ5y^6 z+Q;W|hZ4Wsw?6YUlnPjV~+*%9- zmMda?7h`ptyg;vd2m0@il2q}I9isj!FD#n(HsAjQzW;`4&<9_8a=;-ls1CRj+Lwm> zV82^SYng$>unEwu3Mw}gOFVGf0D(wT3E$1Q(-?d+?t+h(($*?I=@9T^K|ikjLwjCX z*4yXkwm~oQMDTK&wU^@Y&b6;kGEmi^bnQGXFdPM}`(DLlPlyXYvHqUI(XkpbywXsz zvprI4Zdf@T>7Yib>N)TPMUfHJ{C}Gl=dvDo`+p+ergR&dYo}`vls=R z%AkzzK;w}hW^b9%n?vhv054mRE}oA^%2;JE^mN9MMW0x=)=Rm8%K z6W!{7n2padoSJvY+m+gKn94Jp|DJ@&(~z;foihQtZ!{v-?M(G#?$Ge&2)m6uELSFb z$pyTsh9Jk;@WnjIs{@8n9jnm_n7yqHn5ui-7tdCvdSLr?#rFP(9?4i`vg=n&{6293 zj>|G_pfJ#dE;(|Y^4m?vtiLCkN|WnR8e*QFZp2;D?bfF{8x*DSM?CQH z5~y&VST(9Ft^qm>;q>WN?J8i`pcmXG{os9RYF72t;SeK3+i;`Mn-ui81-)-;vo3Eu z)fCLCH|07+aVpRXVBlo*dF1pYE4yvs-qlaPDVV-)V&qoUiPJT^N-|oW)ntOkd_#qt z$}p-~>_6~ihpR_H1fw4mad4zOk)VfS4meY?_x-b?s)SIzCojMsPq6k0ip*PYbVL)(NcxE+PiYmQs z`b8bs=%VeNseGcZyYWJ6qCnluuH@Qs3+~woiRCl2@0Cw_BClNaHYS^r1hc$m|IL?FtE%tjF+fFY&QuJ;zi z2naS_tv3oNUMPYiY$sU1T!`Vpx+7uO7rdmeHXf$uVZyNl9apTlzP1clKt*Wmew zDe@BIpDcg-zL8G}b1mg@QVdS44w4@$%M-^q*B2DnEH9+t(Cd^agER!UEX0;bkfsH0 zOy5h9?uru}Uy8?N=?DLO5`Xy%C^|q;NebId04&Sc=salm=N#zxPUl{IKwwvE2%EGU zpHy2X{C^RU>sI*&BC`z~`_r9ZgVNXSB99%seprM_+HF1xxWgvoPT#vXOmrowvb%xw z)N1-$E{2Ww9JS3^Ue(HtDfwsnE@FWCIx*uQh~#A zr}E9d5%5`2-TLix@^32g=6t_CEq$OO*JP>Ahtn3^Y}o@v$IC-D!(i`x~#~B$uz@KbC$?)l^jXXz5k1NpMyDfu;y{I@wB0T_f$&~LA`m0!g z6-uzT4v7)}l#ze&nazO|o3~%wy=Ps>d=oQ*UPE`>XPI$Fz826OnxVljtO739(Iq}z z);IsrZ((yd{aaWPVN`4XV4!>NpJH*o4)EG2BRi<2?SQw=owI_Q7*9y%mJOi0O^X0l z<4ws_*WcRL<$Va6d$8s3DCHM`1hCS=1P$^zX$NbC!-jYs_6 z*Vp2MfnHh3N?BORQd}d;^rH+oAC2`%y)hM>x32^19jlCwD{%{cTekpZk;$((lZ69t z%J@J9u~zc(=(BYtKU!55z|ydWnrFalN8*PaqjXj5dm$@1a-ee|suk!)ZKy#u19^}G zS2F`nG{14hCZBVb*jazlCu0{S8nrSWX*Wn>oKIW7&n;&jz3ZZO9sayaKd)|H&#L=v z-kfM^{>6zs;wQiHb|h{M@wGosJ3Xqzx^u{5UNp(0QHBWoFOSSlQH`v&zhiHFb1yhC z*B3DM;)3RF1fLy4B=l-#8Up`916h~ph#n1I5a7@&jjL_Fz(n=_hld}J;{A3fU@OSS zS#H$47M|Llq`0mHs=v(AGTBY8kq#if+uD=Zt+>7pCKb@$=`VVAxroE=eOg)x{mvyb zjVKyB8eNrJcb=xa*1Ue<W_AQx@Fkq{8f9h4K674EVF6|Jul`M{?%o{ zgy{#E^0&s4lKMX5qa&+_aHK~m3xZ!Djsx5XmJ+;y6$ZZhou=P-^ zo0em~8q*#UHHh4Mn6p=%L(`Ix_2zDE7pzD*Ke*>-PLbg zH}`DF=2bz}29Eb#h5APUpJm>S?h&U=*6a|BFHUQ1VJ&BR8d0CZ-t7Xf61I2}S_%bj zmlK6SQU_Vgd$_>7l(YPs{FcdKNBHg+_{W>OsNB7h1{eL-^{7?gDAkBa(u{Eiqf3w( zqrdej#)kmFElEBU`ew37q}N%(OKf9GmAbNza)?uynZUGAz2LmFWCK1Y z8KpxU%pf(m_eNfj%ebt&2OQq{#?6sdNl9A`s+&tw5ZS7^x>=df|HMB`FP{SZrc6(- z&C=6|Q6ym$7+{>h8`Jv2kqFbCgu_@+E%ho!ps9O}E#4x|KgxUH5k|;W) zZJF1LNI>4+Jum6m4Mjo|^E7(VA^zR>{N3(YQe-yPEOl7+o9>gV8gGfw+L9rL7`n;{ z($!1CANQA!^=fNoYP(uMFYmZgefj?K1bp9qiJE%YFsqT1_36K@(A8ZdrW23dZh0lY7e&pfdU7B=b<1c5d&8xrsqdBgElJO9#Uh`^j=o9>f~50!L?df0 z#3`|$&5(xnw~T)>f%w&MYjuv3jJ1ougiyo@LWlhnt2fD-^RHC1@j!F#)oXgZbT6f$ zu9;5w*ky;%n?$PVlGmWYz7q0n1z^aR#>(#`H$7KN#0Ghlxtdys+GGQN%Hy(x<_rGl z<3QWO+J@?vL6uhUQqAi zbJ-Uo6@!3h(6qpK?o%3E#--PjTrocT%TJn@Kx10w?QQ?7v}@L6%b`pfUIIL{w&@tj zlnY_FEzK>JJ*K>TuC@20?jw9;Kfmc_xI~MTs4typ>IoH z&(&hR*687gZvs*SA0YYjZatf;Og?D}iid>`*VX= z2UrF@J?N!49UxO-%xTLp6S?M8)wOS~4mus1(*f=iHX=sfuOCd;+53pF;t<{LDjnK> zr9QD?n@_1gIsNRqG|{8mdrJrRkXQ-B$%RL*qpi-&8tL4~6Q@+}<%`D#=?${_ZdLG! z*%VZ?Mu*E><_M;D756CHK^T(;2bJ1qFz#EM2OZ$5Y$;l9>YLur&r5mtp&^Q0rFK0A zJP-J$tdP9#)^W8aajFO*wX#GaPWDW;T)tH42S7~2w0=9yZZ^2iyw0Hnybp5N&{i9; z8;aH)Nx~Xu1@s>87z~hH_4;VOE_amWszx9+Zim7x7mDOq{Wd6_J4V$9oy|Q_W0O+@ z0$ai|?=>{4tbLx(TB!LMxiEgAN-S#v-Rqo!G9bUv>o=9C+c(7Z+3ey%K)pY&~q_Fot(5-7{a+CnEyL)ah_VI=kqCyU;5ttgVbTOR`YqS=dHXr?WcexwcsTY`Th+ z2;K1c?UJyg6*%Bqr>k}65DRVcBHQ9QX37%4HG3(fK*qdsPFs1jc}0lX4b(z3BJY=l7+G+638LYU@i+EBndj9_rUGBNX_$txD}v zJX%>R)u9&pf}K58^*ypRM53JsxK14n_DiCtL7n8)$-S!8kBCh&Urh;n)b_Y{IG6wp zLE>1sY}%XjVIH+4+9-L55e&^y!8`-wp*ynIP_^3YWnuJeou$REdG7Ji#dS#j zt+)K+V?e>N?MzC8hYL@mgKk~(`zYtMdO+CI4nDQ5TWlV=TeE9%`&Gl)XYQ4d*UBR= z5Wqeseyt3#_Eg;QpacEHr%lap|Ga$nkr9PVwa)`I4X;{Ksa=<4R19HGO-xX)43F-} zko=S0S9K4Ebc7b3zEse<9T=*EbH_0HFrALPZ8Se9+_@LQHa#x#U!|5Y;DjA#pHKIk z!Rh3b+@_xqacS5;)>z7ayQXQp4Twmw?%yi$Eep$2+Amrqr+M zQ4g~BPFB_?^yEdpnrMA<8qse`L?~^plag;{O1RQ9Z}#sr=go7L37!s+be__m-EByE zS(sODyd&kL5}Xm&)Z-lgmm=dwdyOq(0I*IpNlYWod#E*r@hYQatf6{`n8D>`=ihy=gX zGfN^D%h{c(p#SRW85{$ouDDNHTjBXjF~-nD*8#W2tz7;@(MS-<8!P1{2@-%KapIg% zo1U~kIzH=c3kFrD#^*#70usm+EiNAiD38T<|W$XO#tl49EF) z?JB&tHsiX!Y#Tb?2&`0PPY6jq6Zib*{r-hK_7PmiN6xI;i5hCoW{AxbpcjAfA_Bx< zUVdopQk0Nhx#i2XU~f_illc3oq5YIexlz^f;VjaeT6G)BzK~tTB>DsaL#ZsS>`D>4 zp@_9lqut82%pC|>UXU2Q2F$WcxZv@YAG`E#nPI>~K;wHHaHqW;ydRNd9?ZvS%a<=D z*(HHi95}NnKLA?sc>_s7$pm4Gj+Up*S~r1EoFn*5jJdd2Q5k#cdNZQwvd!Rb>iy=& zQK0azYOxTak#M-8N6xcS*5{)EVW%ZBbxRec9Cp4x7bXlwCjd{Sqs_6RDnkFziT{$X zvXve^Ri9qQ)v@#MdBcyzeEI-T0&}X9&j0Ps{`&F%eB0mOy#KRq|Nqz_Y(0k@s)#pR z5=v)f%YptTa(i>xb6`}zAGrP44i=)(I|CmcT(XI6LzJnc3oVdXCpHSlIwpE^d_#fA zqcLlcskm-aJvQIK>;W2Bx`n&UUyc_tD}k;26e#`Md5|U*pu!21Cgbnh0F+H0#Dhdy zMUEd9C1={k9a93iiMUIldHJ*UuZZhknzhXEMQ0GZ)bbAbmhWL%F%PmaE$<6yW#fEM zhR8dmV2CPEkhhNF#>cWC5`Bz|T(QsyYfzQL9kc+7)GR&ajQ{Pz$UuC&+Gb~wR3A%U0V{%D= z4AG#L{wt69?I8zrUyR7vUk=2CO-)y6RnL`RJXj7moV?XUR*zeBm%jz-@nG=nE;tLc zh3}TToB*UgRwM_goIQSotxzG477^60P`a`*sLCGN0bA<`aO^xB^z@bO5CH6&$Pz1b z3pggoJ-8$}X7uV-SE$#zzuZ-}(X*_p@a8=9J9F_3k*cl$SXKkCSl_tE_ zDK_6Z%IKQnFizjbV8B5yKXGKvr428k0S2~<6RTKFCzCunJT6|mINrdbp#sh}xR}_l zi~KiEFhxK&8@Ys6)Z#a)RIS}zKiuy(axEwJ)^W%M(GP~~Yt|TttqnC<{G5y3{hAL^ zLogzorWBTZ;{5V+n{cJpM5qXRky{gJYsnDE96bbaTH-eEUlHea6Q}IDKtzKaIc3Qwx+odR=9%&e^1yI!rTogGOb#ji@`I zR7HI!J0fMPR()y{fm&SH9JS_RgWcw|a`7?L7zER#%;vs?Pc(I=BHUoN6ul}Gas8xy ziuDa&zBGaNtohTeV1huQ?_^u#jd?p*qNwsWEO#}f!BgG+`DNQHD636;90UvDCBCuA z(k^`kN}Mjg#Vh?_$Q9Zk<>5V#na;y-DW<<4Yrh7&Yu2ZZs8Arsi4#K2A?H*q@k51y z_iKx)p9=e<{nY}?JXYwx01|*m%>JMg;%H+Q)#==aJNx&S*IO};NoA?`^k-U#H-lCK zTGRkQ;Ww){GnjI?g)`gV*ySFb^3Ljip+Q01?W3MR_1GkR1{O!`fh1y7D>@T4(DzFs z^Y{Hkc|f!B+QWV2ktUo8r|fbY^VD-0HM1T|l4as6Kw318oKX0v{c^NnJWEkZpQISR zrvw^F#=Z8(jgtAYb#ZFJ>Izw@lQcs6=KVBed34^C_Kwol<1beTTKJ4rQ0Gohm++V; zG=X$YPs`vos}x^6u*H|__9C)r7xV1`WDA+Ebk6ef_hK34V=w;yj#$)jQEgiR}l6~d70PL;p-Y3PG za2KigP*CixnveGo?DUkEQl91Ba|I`3Gu+%xlphpHoLsLFK8NB@EtdD-c5mGHBmkF^ zm@Q_ObLvhIa3+Pp-RP&#%=K!B^%}Dj4t?w=hLaQjgOtLd9!> z3ZSa3Q+OHi8+(z}(PbacX{rl3Bmn<8&v1#7}g zI0J%-F^DWn^Oh)%GgoKRT>VCkH4+W+($l4HxkU-TX~T!JlfQiV(ud7OHL_GfF3Rv< zWDO%Zxt$MgGyYeXWx9a2-WlXNRC|zq{v^sZc>mUoFkHQ|X_JtmgpF`n4JODO= ztW#G1(1!oW`F~&WN8}Zksu_O)YkWAH(9kFVMaXoe=$cdv`ypi0V;; z_Ez@HG%cbdkE@=8x4@UE3rJud;~&Ecea`$GweXiq`Xl&UdqoT4PNuS7t}d6!SHDY{ zChCn$xw_G}yGrid$n?|PC&6q?d>oz$cRpzU9_`Ds2mG3W zDBN-M>GWDf>$BTbmH!ll|7Fe6=mH83IoPMctJa$;shG@abh&H;<3B7-cbd#&JKTb# z?PC~wKIpmDQFO>Mw#iv)j2qE9P$9E>Exz*}YcAGsTPvyH12op+wRXbTPJ00B8~neo^>vndt0TqMYd9K) zQp}dd=R$zxC^ephNrW-pu)|hj^-RNhfLN*1UJY)tt+2?D950BJUrhn@d?DfQ?l7SWHok)NLBxKSOIf3s0Y zHl9Fh@Xk}rqfr)b@DlW1TY%4^$U<;R9&d}pT8D=Wa0NoVv>tbSao;Mn?Y+q;p0i+e z+=8Q!%r5d{2yRBG%sN=;XE_pj5dYMJ|0RP^K6}Ii5SaMaJ*+Ihxr9^>>o!L!pio!!1?4tQyI}2r35m`T(>$aHo|_FnsSf~Kcv(y-&fw_mmdvI`aTGLDkkw*X318V0LGk@&xb1m*(gc9IVT`Qsq!_ zod(Jx>9oVV&-Vd!TMBVf;(}>M^(LZqW^FFI*HTV{{U6pPzoxTn2ai9M@z=_J7987@ znd}$;sLUF3KhkTyvRzXB^ZO|hRV+GFL)vbhikT{tQUfghPTr4p&+dRbv`VoaukZ%| z7eZVbD0B_Z=z8x9KfoN_-SLuYE zXO^L$?hQ0R>ErK63$dU`@I1*Nji=QGkq04G6$M^U57&`kW2%vp@vehk<8As1tcrI^ z6_VqHErif~+Em9dpgSeINMiMi#Y23UMQ8-3(=xt)WnxrXNwi-aJkwhS!nV}%(j%7= zXS*}4am9!%bp1a~&3<{{5yuXIP?-f+zmUaV8=Jc!Y-uPwSO)xTfbyF!CZ7)~?jaW* zU)m$L<&uzN;Oa|OeuKGybXS1#+b<>iCact?9V1_C2_pGek!`$Yubv$ux`M}-^1B_f zB*L@QrqUBW*lopb(GhPEI!6wY#gNP%zI)a0hx*H6rnl?>$m+6=|2jr3&j#!|i6vG5 zeO60Bi_Tw}ZLL6A!1#b^x9&(4) z1ZbR1er*dn{J#hUzvpRP2R|{^_q$gL4DIgk1SO?H=V>!?^#;Z{zS(n9etsfu5G%0Z zvC(uIm^2|lKk+{Bl(5PDcveDh{t4HSkLsbsi|nbUqz=OL)A;ktm<-O!*UQ_4dMw|TOu&TdBE8#Zx zhHOMyX50UGXCaz~9gQhVx7-r>GnPOBHzyyYbCro zd8WI2FV61~4yF}A$9AX|_BV)E*CzcmLc7Mcd2E}D(qOs9Cy(#`K4_i1dT*XIQH4Tw z%2o73vmvtdWDj2QPq+5V0cZSjUMmRJT~zTS$xd&y3RI~k47;SRde#PQ9yZ#1vW`B5 zq^NaddREy8T4?DH63dE=Ki{71W*>G}OTZ2Bbwi?gIvPn#!;K+OcOOaVpE>ijWqKg8 z-bG#RhkSo`BPbp9_8r2Nfbh);rW4e=)DMxS{~^BS)*0!+Ja6_&9gZuWvfw7x=o+G6 zO(P20=cWb&ZbK-&eRMw!_aSi<@f+Vwr@W4goPTw|ZgH3M1@A}rlG+u9!+&n>v}-`5 zOSf@%Z&rV6YjDA=7lV_6=x06l0K%G+>hPb8?UA~(8|L#e_^oN$4$Mr+6tcvg!>r|X ztQMC>b5y#@MAfFe1WEKM$@(k3W}1U>W@u>N!NpK4XOfri!^PAg+~x(8&Z&A2iO~n& zcS7m4Ql!5pT1uV#6Gc%n27ovBT)-{{NlQ!14&E2T6l&(8+N~myxMpW8OPsfd^ByMd z6Y;0^KDHj%(@j)LVWjs+Y4@|Dgtls6>6Z1=+t9j1!El-Nv?Mv70jWISno1$Jw##&5 zr|QK|zP&;iuJfxK8X7(Oa-a&cLUq7_B}{_1y3Ub?BWr){ozbyJKS|30beur4|A}SV z&qq2qP9U0EerITNV9PZxx-1v!_6u0DS6ZI}p;rm=nlG*uyQub5b;q>Gz;J>G*Xlk? zfSL6zis$-BDbI(S)is;Nr7saw#w01d1v?TK?_wHFhmp90K9ZCMrZn?(yoMehEs=hJ=*`LCgrPF;sf9=@KN>-2kprhkBV+#aROtv%~vF(XSrQlv< z%hFgQSsS+8tVo;QxFXo}H8e3NAIOZ~$SpuWRnrVG;hw(gWc_(iuyFtfg!J*u)|#K& z;E2K9N0U0m`aU8<8z!5YK*am`V7RfxHrRo6F%S(y=;pfgVCyS+ZM`a1kMdv_ zuSssJt`bSgB-V#72HpK`oM3d2^4@>#{hvLBGzs#e6O}41{J9@Cj#!p==G?OBFECE+ zsqwBTTJ8xE9Cg0zGkU=vDibj}z5MkB67Y+L3q2p?JXFX~+qns94HXLdApOg3E4$CH zUtM<|VXC6?`tu;x3+l&oyUoj|u7eqCUQ5-;42~%3Uoqcv- zu_i)&h!o4OF90j4(g&X2>~Ro;QEk$C{A&Tj7O9BS4vr*_(!&c-&4Ys8GPU&U#y_oj z{I9i;EgTq7kyU@95K8{_e>{E1u=hTO{P9=pLvG%@DPd=O$%q2?We^{!7|s$inLE%_ zW;d8iTeG%IJOnwaKY;iS&Q^pdt-+|k}wQFeD6Eo8Hf z>CFR9(9O0m6A)i|LNa7;Z1ppf>mmiJ0;xMxBGCk>0ZG8-aE(QxvQni>SUL=RBtOHO95kvCNa@Xs9Gs$pl1U-~OAXASiVmjxvk7v@;jC0kL670+#?bSNW-uE{yf{-fBM=;7nz2z;4JMt8FGSA>N@BVJJ%|GR-=;C zx_-gGN}IP|n3t4*z`V>p3{T`HiqK$QwsF_rO`FFUfkz%y#!HB~aAf zSy7qkR7((x3uO}5&16{mYSXnn`Fbh8Dtd+nH|F6t#v4=ayTREJntJLQn;AG*o)I^9 zw`VoUsCq}-G#xD5tpb$=nE}&L=%@>f!+SE-W=sm5A5pI+}1-@sI5Xi zS!dl#t*l<@U3AAP0Kd~q;A~r(Nz{}6?zNOGLuoLYOt}CZC5$>>s$II50+PzK+u)p8 zlyx;^&D|R@rtz7MSm7Yz(CQ!V=S_DDu7w7lu6JGNfsjGjp!wxi!MF?ijT~lx)iB2) z2b}L-s<-4U6KbsL&z8~|utw?sSffi0IwUXC<}MtKYlF!!D>qeMe)4-0tyYB%qirrV zintAndQzf$)yDakQ192%ffW0)?t%Wfx5iEtt8Wyjq6pp!pavaJmy>NTS>c{y?J*3Y zu-^*^Fbf3Ww_e2+eAp^e8eml8J9%`NG$=>0HP zJWklc?M|`)d`fOi|0S=#7i`s?-2RzzftA^TktW(JCF$eGC=HZ-D~EnM14gtQP*0NJ zj>)UN-GPuyd@s!^J7Z|oH?nn;Ig#x?gTdbRE5W0IPM?y5AqW|=%SW?~96Ie~ z`H06!W+}c*lgCMN@xo?+SD!lZnxV7UY}T#6$gau&vFpUGS^n1GUV(qw@r{Y%-Kedm zPLmEv7vKk?$myKd*Ya`tM#qU016&s`emctc36Kx(rc6rhn=8;gtc#KENaBt*1tTMr z>3Df`i%m~fj$4}0jR|#~WaRsDR*zZQ*I;$Q1J?YwW0t5wivDWf7-m@HGbPYk^=Nn7 z>DIe_QS4q~#P15AmA1oD1?FjK7sV45QAqn8&K0@2DBDS99^LBqEx%f=&`P1Q^NzVhf zgfq=R!>|{}q|I^KkM7DsTpsKD1#q#nO){MJ(P*g>F zUS#rf=m1*LGJ0b8M>Y{~pR>Sf#oWq!to>!}Y@x~RJ3F4Pv_9U-@~3d)OHUI;q5L+O z++u-Af;cWBLmc}?Mu0CKauCEX)>F0L*CytUT!MjaXzB!D;I%6SIqwm<7d>5HE!cdx zAIk)_9NDIkEeP-3(<|1l-Pz^Rg?>5}xl*^<>!Y_C_l4XQj};eRqVg+2X~}we`5N03 zlUJ84inbP95fN>2Q~dl1cWMSiORQtSlrx3eEl-4DshnuNJK(_rI3$ILTbRJe;A}2R zFbpkt(E6H2I2>@$M-n3Xo$G|ntyyec=LasPdG~8N4f6A=y2lS%cOhg+C!)sk3RGY& z<;A((6TB=vfLc11|0lXAAgACNWmWYY{++*+M{;z0S_d z896^T{N{?stL?Hm>dZU)Q5{6r2x6npWyZNm6=1=d4lwOr((>?OT5MGs8VhUb(Rr|3 zq82Zv<-0YZqoe<{TI@ZtZFf?7(_={!I1YL*X2M0@+W2ajkg56}uXZ^KcYQ5LuFVTV zu3KPgAM{a{wXZ#}C*is=?A}Rxj{_+&i;ny0kqk{uS*zq#ekV3ex{p@m@#A}&woUWPb@VSsCXyZD{( z;r+5&F4l*sCPm(^@NN;u1Sbuh>1E?~&-ZiCE!+pq!QGv$iJW{weu>r5vtl3R*_D{x8r8)&AJMTYNoAtwLYXV@H;=enA)*NgGGPG zwXC`3N9*guNkIl)7l|!z-+i<>LiAkc78aT`h?tNIf;`eGbK&+VY4y=Y-UVYfpJ?1D ztbf$vqc(5HT!8^pQx1)J1MQVG4Y_EJI(`_8%=c@)LBtUN*c^6#wrB4QtpGwF(m%iW zX%FFay(Tmf*|I?`9~0Qwxjz3;9?)?=A7kF9I20n>D5zi9<(vfkM25JO3W1*p(f;G7 zTOC*9(koPr<}ql7CCi14@fG08d`o5S!&+9r;sPaZ3jc^Fh9@>I4B%L+c(j)rM-NUe zzLvPfH}o;J2DOtF@Mz@4tvU+zz*&&`h^2~$S5gygXD?qRz7X)Tctk<36Or2ZY}BHC z*LP>gbU5`Ya0^Sdz!K)|s8)g#BjG^c8jBc`5OU#igNvc1iDRJD7agr^nAX?(ja);yQiH4|IF}z`Rl3Cwaiv8yDv&O$aLenofN-W{=9}B z`{y}1NXOP!;3(3W_jOhnbXF$(G~^k*_%ljqZz8Z>y2h~4LZ8{~t3Oi6; zY{zSZ+M-GvZzswvu69LCr4OS#5W#L6o0#<8CP2C;T3%z`6mw`|qCep?)HqX?oC=|! zcTXZ)imcBYtKsOUjha-s20mBQ1~drzAD^SumtEAm@g3@i(Wv8#L^feP> zcX0yhnSx1s8KP5vd7|AWmzhk0DHhOPe9vMy_J6VWmSIt@ZQC#)k|HVy0uoAtbV?h5 zNJ&eMNJ%5zVStEq*PwKF*Puvu4js}tbPe&H;|kY$KJRCHx9$D&{;1T4C4PZ7qapO`W`KU#NvvdRNQ_k1f|Xx%*l& zJ8^w)sy)Y>yOzAr3tE7PXT4~RrpjG2U6j&qyD?rzH|4nYPgxQ6q5Q0u@>L_Xom?Jc|96Q{T9giq4G7i)*XF@54vm#usm z$cj|W?EA5L<5$yYyv6eMp*0_$oRTtAgEbGyx_kLf42b!zCiZg>fX!WXYI_@s2P|gR z))%{ue_Whj1v5Iy7ftQ3`Aa;y2{p@mj;L^Yk91MXKD!22NKLyz)r;J9&t;P=rw3qG za?3_@+hJbIO#TPP1PpF$9F&wimd&>Thi7+yVC-yCKcpP60?3o?-A?Z?gQ0g0BYg%f zhOLE8Zysh>!-pcq$4d)OmYoi$j?8*n^IN*KRMIYV&7TT;?tr@Bko^a9f`a?Ii`wR&tAy2DHerO~Qn>&D7Dgm9{YQhWB&2ww%fM*+=}Nf z&eqgetH0{xtdK?9WuJ-?B1jFvZLmM#qRdij{8ajpAC|C5wAEoMI-)n%VRu)>#c2=I zd@ru1?&Wc@@ka6Pmrw{Aa0hH@dK-4M)$8>WYR_c+n zj2Zb9u9K9VPq^5~2bWJ?F=7le5QM(MfD zlCHSSS(tF09lvTXwG+NL zQ@K=xE!U7P`F3dAc!S@msu)Q0oJ?$36@Po^(M`TQDk**Uz;korgNP*{8S#GZ4SnY* zVN9Kq1g4Yoo(}ctp^S7#9DTqZze-TLz-(^KQQ$JG>6?3$rl!%)j=V%1s!xlZvp}#? z0=kd#-X75ak=8g(N$X<{tpbAibA^WuchFlTLARBfYmLn{1F2o(nW8CfVkWIz(0d0iJfb= zM9*n8tIW!lUQ$(7)bX5j)jjf3O6+?lbbV8*B#$;|idvv;Adc3}aTkBxis2dFI@_>) zv=2&O1au69m$vXPpNLJ1zIxnIeA5f)ix4ddPF6N!exTGve;I*WW$(e`^dxP7v3_vr zK5ON6d^)!c<|nAYAbJ9HfL*9;1`P1|3EM&?jHO07u_V?7b>q9oW1q#KoNF6ObA}tP ziH_Uurhi5k>WyyHauKlMTu~Sh7}*?ozyi^nGDH5^ka0aQmvHh()B=kocY4Ir^WhP} z(4Ol$pjZy&ItG@)QUWKNcJEKiX&y;7M7%G0&*_Ptpx~2DyND5h)w~2fJ{GOEnw8*L z?J1xezQ{XlZU=}8(TsHzL8ADfCWsDB-#l|dTN3zD^kkMb6-h*hohVCja@CdXRbJ(@ z%QpVUn&YpeEUpsXD$<<)_^=<)IqKo!NBb4IPE&OQXY0}NS^N%`-@ZqE19Fm}BmcwU z=zDNRd;heC^U0L^9D%LM9F}_gG;gF&$}3Uq_@^(YnMD2?V_@r$n1F#)?+}nMNQ^x?p@#IkNl3aKY z>3s;+;TjMhGA<+sec(ir-9U|#B*ZBqDnF0UELY%km*VRZsq&;9+%kDB@K`>{#kyu( zQo6EIRuJQ->V1zf=1@}3@bdwlY8XfP1{|_!GMpwXuH0=+ta)6VmIja(>fT=l6jxOf z9NQA^hhK62;%65m!=S3YgS2&XdLxd#zp*0MxJ24e_G;|AT%P9iVXctmvw zsWO7Pb~S&HD)$@mhwbBg_SV;8xt|Hw32jN5ZwMj1Q%}~G8_%$>j%1XGxrv;`kQGHu z1H2bZ$$S&d3Y8AbW+b@N}j#x<^M%E*;C-LIJ1O<({>DP#Gspdjf; zpM{8kGG?JCOqV=atudU2n4Qsi%jPLUky5`6qCzo5J%%k-xA#&Ws7&cD?}=Yj59WU1 zpBzY2zPuaYRAQxOu`o)Ce5;x*5To&dxH|Y@6jKL=)YGm z%1WtIj}5h^2A+3`c(bh6-fxI5PHf@Xr9=$R4AZRvZu*YUE4{qT4~_!OfO z&}d~nl0scw6CmDmRHHZXE~q%JUcpm<98If)2`E|}rfH7@SpsgAkM(n=PS^mN7h08p zw}1N;OG-;@UyhfHbb;?(Z`t*pxamKnMm&uYj$L$&J=rH0US76wCE$P=pL2bqLWoPD z^7Y}?%7!w+$pO*o0&!xnz#M-q@O*vGU?nKSktYcq{wllw-2nt5-#0LS#9^_EWgNt7 z^-%dNIGuvlvb-R7Zwc$&yMb3cRwPZ#ni|m9!(&f&7V3N4g8DvAXC^l=GnGL=5ocw$ zKgSld>+`fE;ZMyS4IY*FPKeaBomc>jajEHOCy)mQ;R)>Cxx5(kM6%cE)uZ}a7u47? z<>JRm=B~|vtmAq@K%Ja|UDw2ck{eM6aw8Wy8h9$o%D&7peWG;q z@y)Ov42`{zT7kN#HWCKUP7 zuMW%r0(v9tq&(ic6m8>$o|LTKyQg)pM~#N#d;zT55#ri9#Vl(2hp~ai%fAQZQPrc~ zDx<$@+W$JuVRtzlf3zfKh_ALF_;(D6gkf;yKc->o#xS zX}V-XaOf(Ee5V2@07xS3N%wHv@>SQ9rVQW#M4hK6Q8?l0#xtqaZtlbcmS63ua0@xrV5MO(XOs(I*ZlYvzY$r+!L~KuY*u z&OH4{&IJ33j$4mGvgY;~3$1ptzrK8bj6)R|7{uqG^Z^()@1R}u@*_W^2G+*M($>`h z!rpG(at~ZiAOC=R*W^p_Rjri|Zw^3gwBhk?uV+F&o8yT*@lrtSx-6lhw!^0ViN)An zS6|$M$tF=2j@yVZWR3D1m_lpjzBV3$0Z9@P_b&?!s!j z#?uV9P1~`DH}M1+c_~mxtMj*hv3a{G$rDHiB0UgCR$?AUkq;dWw=A z`Em63lJUCG}S= z9EWIA0c4rZBxpUKpX)EOAS_kzMZwQOZQwKrGnju${&^T16A zK1bS6_2_2L*K{!a+n9{Y_ESccKdz;A%ocy)(Pf*l42gR8{^Rt@y6ld(Kso5w2Tvch&o9wlNBmaK+nxZeqgIwdJpl*S|h1A^pus_B_- z^>>rMGA%TUeurg4uCULkLQvhvO(4OCLvRn2X~q;U$<%d+Gb+ZJDj#A|v6vx~V?B3l zciGKw(_qtD#wV0y@`Bn8as%XKH|fEorUFbXo16HuTap>QCVCtWvyCJ2UZ@)UQ|AV$sTx|cqp8`Og}x^ zEr9AC0H%GjY~@2VCtPt6_FsUsf2}$FXO+6T(=i-HoEz}6<&&3+_8tvH>K^<16+2xO z(=85}CNPET@Dk+Q1C=IJ&CDhn@fq&^sOiMPOlEwzOBBA859k{d7k$LX9>>(111)Yf zvKemk?RZ8W$Y63|8TGt_pvR@4**>S9Wh!NS99f4v(5CeYx-Gg|F==q||~NJxSMgp`FeieulthX*lb}#!G4a}*6~#6#5r{ai#`>}K*jK>^ zD5cB@BZ8x0@I0q{zB?48!rlYs`|(jc^@kz1F)FFO^zNy^Vy@e_3CUZ)NUOE=t$Gy( zt^vggV(^;=Ow;r*xAJkE84Fyo-KfyRrp6aLrh{ z03$4WmHPgTad_+jj=5z?X5G|v-Jxvt>xolc1-VDN9LF!d&41Pv%0{G3i9XKw^eJyy zH6o?Oiw;1|Kjj4sK!LiK%jqnVz5=&nf}RU~qI1;FDBs0;B2BT!DvZ_Ss#7>H8-VO~ z=JY-O;sYn2#Y7=A7O$+xM85nPl0)k>UXClp8w38te%f0>w#uc0zXwMY6t$(fgW~u6 zEk;U5Jt|GKrPGHi%di2T+klWL%*VHoeV-|oBCV(r0vmUck^hljg5m}!|eHaV7o(3wPc z$@I8gZ8*Yv)ARcErH5%K{9JXJIc8p(;o-(q4s3bvonv30I%v36VZsM;tXZ?s`3s90_AD(XZUi`Z5*);1RXr(-1F)0kYB)INE}u%16mo!I z2)S(??z3X!wso<=vGc8g9-|`7w_UnW_xT6RppRP9R(XLEQ)2m2=N`00w`Hwo>$8|>;#u|e{3O0Jmp;>L0v2O;!CDu88&d|zSV-V zdpOr1!%MnSwxWePDp@k$MnE@WBRE4L8bZRc@9qTerbIPXq2&!vGpDjDq4A(ubsSwH z0`UR`u2j8^I8bbOx2lULRFtECq5vIqlj!CMs^MboL>Li0$T6XX2>(SpQ z?gHqPV2uu`%r~-c8`l+0RN47^nt78)o6 zlvUtyRg&g3<}yb#9SAg!BZ9-rjOX)eM5fPh+mbj6YF5G~uI>xp#Cvz~L)Ue!uBg}w zzBNpsEc^7w}jygRnjQ2%p3%m<8$SO5*% zHgm@Y_qQ8LPfVYg!==YpZkTvI{~P$*N9NFFuHJ(*{(Vu1Xe9_AX`gaB-!fT;bCavD*v%M7K^) zw3BIPN|%X~CzG6??;)TOsvzF&ADGws3n%bDq9PwbOg*xLa?K1&r`$R%e)GG4Fd6LN zWjprz^az)X(=evYF-|X9L1v@^5Js(rsRkdSnQ+mqiQCgH0;hOL>F@Fd9|lUcK&fOC z-$5V3P4fJ7-1c~Hmm|&pxMtB8%RyzHHyP~7(dGI6|MU#c@M{mh(jA-2Z$)@rzkc1e zao-_rp-aH`gtgJN9nUfvHJgKV&HFApsfMi~civBq9*^GvW3#@J{MHBH(1iTBte&q$Yd)MG z4SetofQz(^>3a!KdJ);iC(U3itDfzl*J7t;EIAtOM(u!tVC~eKY)na?>BVao30QR_y zltxz?%?yVG+7jX+>wkE(zbYvHH^9XsrojFM8vfNS`o~xQ{onp^qfqPe-*x-z1pT=% z{|}xJ30snpI!5K;TwO%fcf5smpo;4vYPp?fF;b~nzA>`0;)w#s^2r)2YTM>N-k9WY zyPIN&s0VjznX4_9wg!;ow87YEMsTOL>-I*c-5F6l+=DdPXa%a32_**DW1-I8h)BTB{OCy}fH;*Hil{a*dIrMzoLgqSa1 zcDKHH4{J@R+C?$vl+$)1Z7)Qf=#S%ZXS%PmqvV*5+Dx{0I#J2?dG8qtS*LKJa&byq zhkYeU>{G$h8@D<1sLs~QIx`DvqePWrZ z-cW1MG-Cs@7D0zqc=H{MJz_&cUy_eS?L)t{Kb19X(0R09W?zw=X_Wrh|QwYLw6 zes9#?L1qvR6!So2%B7x36@8*&`hp1s4YAwQ*5Zxy-CuJCGBH)e8K2!j_p;=w-d1Ti zN?8LCMhu+{QxO{#wZ~GX6Q7|a)_^k$Fin*z@W8*hnVCPWVFKVG{bz%J>W49VWd?9W z>9Ma;?M%pbCydgAX|*V5GJ6@IdIO}3TS!9Cc95reU#UN@l`7Xa&*DwV2Fn+(hPfERiD?h@$<{x=aNe)aG zylKA)idkA`@ce0}@C>=#bq;`mLN&;SqL>Y^+CuKUEz4Vi&#ePu8i{$_chOKK;|RDb zPuI&&%dPo8#YcoTetc<^Uf&$f@c5L`r`C5)E`>rl+KZO{kPLY872_DtRd7&>nHtsB z)1GVYZk@#Oqb0_X4~;F$9xl2A!CD88948fdd$!u|Q5eTNT35Z!_bY0}t-_XgT{%Rd zE}eMO>y)5BqAGT5Q4RY3V2}bk$bJq>yUsSTWAB?l2S~18zr%6htOZacE2Tw0XXM)G zp@u_V{s$@GpO*TXq7UGSw*E-#m|wH@3IUnZcURTEZyHYm&EQV}b}&KAlXf~GF+UJh zbw+O1tDqFP0K3#?<9eUz4Tr!IfX{FlD*;1VMQg;_VeFd5>7P%>N($sVj8VBHXb=K0 ztVJLu=fJQU68M%wE1eh z9gn%JOBb}EWB`(P*6QgmR9(KtrIJT}gtBsSpB#ez1XOG#D8B8gf9vZyHn^20*5fi} zUpmoIP`6@=(kbSxDGL3QZQXDm9Yb&m^FN%G!UoWU)VibSPM8bHh$^+5oLH>~m`ZNj z&Ub=@W$?GV+HdcbzjfGCpI)l*&qo&kKvBhtzd`YN)0LXd3a8Iebr4f04g1n)x4pzV z4o>fZ&ub+sC5E%fO^(se#q9u)=diq{i=2LWlQFoA<`hA>SZsoa~x*!K& za9OJ(6tNV-+Xj-h0JQkGUWwlS&?{L2a(CU@h5sSM-M;o5G#FzIhp+zMNS542Pq4F1 z3UrcIGquL<=e8Sn-&L_ z8`J@*N@1@XPx^~BiI!q6xV-~5&Z%9t#F#bGBkbf4nt`0fLI%XMEN6HLAQxG0c3>+u zuB<#p&7FGk9{9Ka4OaB|ut9^KGlG0DBZIH$i^mj7xv8*1D)!pir&piVDxW=IR!$9> zuzz%^Cg3RVvWcSbiaj?4s5Y*2wdMezP{a>OeC?D~M}&WvaMzfb)(^ZHIRz#F#^t^< z^EwB1=I8rPSIEI;*aRiO#izS?KqjMOd$wlT%ZVR?N{qh0% zMgsQVpY)JhyApc1iqTeY#zk3`CCV!RmRsN1tBaN zxipCgbFccZWeX-_$)-f6{h9SSr13wk35x_;i`)MQc>ekp_lC34$at|G+^UBQqzpU* zvg#o_#`CJN{q*zzR?lEFX8k5jE{w|nOA%57ltH?! z*Zb#Iq*I@L#eg3Csc+y4Cf8q`(Q|cp_gG5O<-{J)rHPq!S_k+KYc6x!1vyz(*C5PT zhYG631~?KT|6r^4a6)DYruWFiK;peNdhBu0Z^UZ&V?GP2`#*$Dz`*0H{|N8@dKd7r z5*U}$<>K|z@9MaJO|f5X0zg)+ixqp{YQ5(#WX#auhsT>mK?fgMy~hhp15g01rLd30 z^$)AXw+l16^7knBw0a}rfHa7Tnj>OJ@#V{xz+o&I5KeMjhmwR6#8v958{-?)d=6RUm%#dcauGOZcb8&))w}>7<)) zrv75vB$nTx9FfKctS^w0>~%_-Ty{zVFxe<+s`Xb+vqY%Wcm{%={&`;Q_r(n`D(|$+ zI1$BsN(j17i7<%gaJ<=>1c9y{JkOmb1Nt_nfHJtA zswmQ;k^M}MYJKo4P-F3}6Mf2WJ!-6xrwCS$h>ZTNZ_w8Iz|*mgeyp-%KGoc3!?@?) zt~am&+OAmqd1Q}<^mAGS;3KY-a8#~<(rGe$iJg?3Z+Wu?1zQUl!vwv%F{t2Sz(2S6 zFHbLV3loS4(XR1ZMC=|uLQvbK1AsW z)T1~jqe~Z5i^hR^@o7>W5-m9S)g{668?GGvmi8+&loayUa70yC>PXR7Zn;VeedJwv z^F%FLE7YP&n^Ps0;q+*SN81LLS1bFpG_Om!$b`{_${6@cvv#yz zYiy@sYjA+z+gDYiG?C)?bIos`BUi4)3VZoKe*BKY)J*d7BW+$&nzX0HB{EK>ujdx0 zxouBX3g!<^UNcNZ{I#i3kUTpR}Juc?H@-sqb-WQ1e5Mc?)hby?|v6-90xYDcu%_k5wy_lek%=7I@$R$W3a~Q2QsIVBX zYGH6$M%L|T99!J@Gbw_4&w=^yzm30&ZXD@PD;671raz}l;m$d3*5bE(bvxsY(jzH_ z5Fqw@%XFA^)l~52%(gn3!cf|xw6M!+sY97Ufwnk$#?`T{1^&_9NkRqc0`B2FHC-0i zzH4sKYlhuuZCF&nq5G2D@GF;>cnxev_G!0~`|m;_mq}ZK$;sDTHa}rJM@?4#QG|wG z={!-zOYqyHW{sO!)j4cA_QQVrcxOJUdqe9<@?(FCAkAEYaA!eZ{TpmPG}c%%R_43q zSzpL|5=2ex^oA*zRnmKXHQo8NY}m}u6exGQ4&99AG{#dzaZPlgMs-?|A8#Bf-0Vn~ zhM0KNK2h*XaLAGia2tF(6dvp~U2Q~1+IjJDDnkl`^{4gpMMD>S`{q9g;(Q47CV!|b zy6v#sy8UW5PO}||5C@PLx@vzUZnu^xcB#wg$}6|+a=m}hA+D&Nbp;(AJxq-PtJtd` z-*)Z5lI;HZBewmdS>oM=uDqThW0ja8-#Rq?M%S|WE_l1v!bcsoMk&uK%M5~o!SmBXmT%h z$3BYR-H>OF5l^CTn=5EvYwn6?kIcacntOvie%qV_XKZy+O1aosDc;@UXSnki@0#~+ zpYZIDCfNL5Y%5nm05Nl}d~IB}G@pvR4harEet5}JciZvqmz@c_9+{$-l9Hm;J#V}- zE>~i?XhrN?lW3U6%!k0(?~HN8ap=`Dt&}1Q!KBnyS&q78>#nW%-37On5?u!flJ>=i zgSCCTF%N@m)hxOS{(|n^(GuA=n7AU^*V9%oYQ02}QZpxOnu<>!(XiuZy3ATl)Y1!F zoNEsf88L@3hF@73Fuh2NHf#$Ma#+=O+H8Fv5Ix^LSrbNY$KY|bJX5jbzFG)Htg17j zJu>G;_8A+8{OVxX?gD9wXc1}r4>cEo(yCcin$5O3#<6M4fz{pVG?{QA{-*E#wJg2a zz9Xgg%r>jJ8nIvPp3#@nZU4jyzo3^u>&9!;&HbeX$zJ`MjkLFUh_LHz;H=Nt zZMJ6aknvc(0ctq+v{dj&?$Nx(+KYxmSg9?kar5!F>gK9-2Wi!4UJb^)22nJ-|It#* zxobH{Uu2T>MAM7{7Q;C8B?cdjIx2Y|76jNhUY$EJ$9_)-!m5Uz%Zr^Cs-EIEkG*}& zo*r|jrJ6QYdrq+^pmZ^|px|sHM#<@eZ{635%6o2PLFCQDBv-W)eGriGxmkVzaPqmz zA!jizBfVey)73g|?G|1_Hvv0+pH)%}^V$b%jHG+d=>VGW^~k;bfK}%3X4wrbpR2n? z11{+GP6-vZ5xe7v`e-Xr*Kx|?_LIuH@A2ERZ_3iU?930h<<>AYAmNggOn$G=0x`ZZ z00VW65YuY*)*CIM!ter;f8u9Xb2ZV@><*eq zqkeXvA0hSANvy#6!K?wPnw)xF$)^#{YWS_XQ3bH6G=eglz7f*Xug^@=&az`KqicYN zuVQI=jYBhoz*s$j*Q=2F;{>}~FsWrn04+U(HZVR9vviHHy(7Up6J1?=G!yb6uxz$7MoQjWe3nQI{u zB%R17Eb7i3Oot#>FYzo@;0y?c0m)35LvqrH;0sIO1cCu>d=G9QUM#VMTJKJh$4^QpWG4lKUt z>&*#jV&w z<=sRJ;&$hbLLY_QBYX7n_<=x+zsfv0F=Mq)IA|AmDbd`$`!GpG?d(W$GC&QC$=V#v zykq|~ik(r*(L=NBWmW{UZo8G<4Fi%HH%4O$jl)`{FU6<(aQd^YLu4JJu8>oywyLP5 z)vUo7QIeVIWcGV^PKytd?sU!f8ju3u7Y=OB93%q`e)UFwPHObn-;WvZC4}S=m$KDrI=bQo(o{#`fgbVv z?stIDDrdRo0~!W!|G3&hg1{+xOd2D4Xb{WqkzP-02^5P4g5Bu#zj-f{$@nXQ?YU(- zG1ix5?lf9v9<0|4ykN(00Y4CVWwvaes6a@Y_EqOymw@u>=VJ0@T8H})v;gNwiFCH( zfJ3ZTm4C68By_IRpe@ZpW5n#rJL4$Ho-=1&mK!e6W%Iv5e-~{@-^A-)zCpBAvD=}* zNmhl~sLmTdLWhx04<#Xy*o0K8E{mOU>_gL8i=5)Z*PmhNymh8Vg#)7mwDKH+|S`rIJ?<`U-7)a!2M@~W$+?$6@*t(mXSh6bCEuxdS)3Otek3ffG; z)Z$?Fad}53*IeN7G!Qwt>*i<`O(zrfy_`j5Y3XbYxw&+*=4NO1OU&!&5EG}trT4rS zlzu8T@o@RGQ$m6bF`)oK=f z)Tup+hMMerf8GE2Tzt3XzDP)7`pp?~!?FB~Szz?vdAs(fX$Mc)9`6|4cQ`5CrX_XA zj1x9D+~Sl;5G*314{f^Ja2O={O}GFg%Y%y%=MwRt5BCs1@*lXMl^D&3BIF@?+R_U)&a6cI< zH#Le-KxB*`@Gdmo z>QJJ!%ssYvpUF7?Jb`rPY^^qu(bo2C|J_}qU-hk{6x7+@A8Yvf$Y_IfZNI-gnw7O9 z>R$0N?N#uD2gui4K4oid){I_oZO&IOyjDi)6B~yIadKH>T@yMgE2 zt`9J!!N93x-rD{8w!hL8-(|Dv$_)|{<6YHImTH9!iCB@>wPN;_@oKR@Z0enA$Gg1F z9Avc0Od0Q(^o*)TfW94=36wukC#J zvDiYUV0u!2oe6baDzB+>M^qODy>@vnKRuCm$t>prDsx<(@#)XEH21qYtk@7pcJp3W z9RKT>q{Kzk!I-MgSOBI@7@hC%*@^wiV^K|eKzG2U;?hz$b?j$Km0yeEc3*x0jyn>= z9Wmmxt(xuXrB0Yjd*_bH{#_zWx)A<6t*1RylXT!_-vI5Gmf(xKjc=q^3!+@H>!%fk z94xFy=lQVEr(j$S&i7;23Zi|Oss2uPXcR!eZC9er4yn2eTs{wj#Dz`!<1FnA$5Lh7 z21!~j-G?dZi1x2<{q`LX5pd~h$IbcP2G^^VE7S^6V`UhWizXTOK~NK0GHuESq9myq zAdA)yT-xzIoXFJ&6ilM59lh89q2WO8@^&FHSn=sBNUk==B`g9y7QFKpnW#yIBjvdw z9^~vmgC#2I>S-=xnnL2!`tcIkK|&18b^NXKGQ^H`s_ogCz+l$Ta^24)xd9Hm_5?9v z6e&7d>`$epNmke$x(6=YE_&8$DyI<^htfauk_PeG7;Q|k8fg%b*FAs>}rCBCg}a0vul-jWN1gMg?~=Og*z z69T92H!d!}kq!ctSoIZxymJM$SRQk!Pl5F=yXg@u$m$}mx$2%Yo7soW{H!?7F)Oh& z>6}qyaUCKYV7cx#$&!h?uD~lQ?AQ87%f89k5KjwFAq>~_JA|E9X)c`{Iy}$XFI@0& z-S73wws-dSciK4*)K2W#7qHsLHUVNfe^&5jT+lIWtm1Zp1qIOM?JM!>vuF~uZ0DUd z9XCeE4b~3Nosvo|CKaX&d>e&am%TF??`dB9z|z3S^hea`-HKx7FTKTHI6Z!5`D#n6 z|Ffs4-@PW;sL;YfVOOZ0XpIa4DtdcSU{?WkLXc|NLm_rc=e?|}BVTks(ONw3PL(Mk zv7K3Zd&01LOG|YfwgV2uk2ckO7<4f8-Y=u)MGB0bF`;L@V`2g!frl}(Y&4s|d7(;` ztt6mt!aO_2)mUWu?n**kM95AdGb2TF${d+ReIW?gC_T`V;VIA4NS8}wpzew+`%NG6CZVUek$QeNP_5 z{Mp*s6!1tM24?>_7t~m-#{XIL&? zGuI5Jfw`NH)Yu^#&#hzt0UV!z12G%|-w?d<%fy`Hf}CC{2^=2k2QAxXi|)LNVhJSo z54GQ5-le`m&620ZE6^hgm%B6VRioXXE@LncauW@3u&a`rxxi}_ zV60}9XCnq9fV|(Dd}}0YMy^izV*vGd#VgY#%8LSRa643os`HJoMLDI3EzXIE_2q>M_n<35;iO4}{n`?@v`mHU;Emwn2Z8C+5C^S#3cDbLk`dFqy1jI(>4 zZfEp;1kltc@b1J|jo>2%QMUJ)LF)R^MtH7Y;c*6(s-1hTxz!1oCr3kPrAwdJ@Cg zCtcfD225^4E&q(K{9H!-y-Pq8mJoTD8UyD$qt8reiqY{$lb)1p-@|x*hcGT_Yj^IW zim`8pL2f(OwG1sgD{+e-$4d*{J-tvWzx7crof8?ZAPd&hMJk)qFl)OwAe0;?R_FS( z@iSH!=bcA5qcerG#b|Jv=oP_h=$NNEFQ$W$M)R3P%^oohc7rOfDTPa;fkK2VDBI;a z6>X-Y))3?TVvANfxDAIb#2rkBn==)VZ3C3VsD?N`!{r(Duse^iWVs8du7Owu{c#=+ zurRMv|M9pNH-UwLd9=i~Tter&kbBaqdA$D}R|W)!K&$%^u!*ta7|n0%)Bb>>zLrH5 zWNc45$$%oo$4o}woy9g`8piSbLWQbL9S;{rf;14N>lVvk@^)q zx)wI{eAw$8aY2e-OEryJszg4%{~Zh%g9=M(4*>0s}YzhGoOJzciHp zdYFIz){HRtSba%4!>{x5yRV4A<>z_8IPmy)Km7G?|2=BI)&ld|e~;R~ZpD9(+P{p$ zf34cTJjZ{n+COdme;w$*+|_>_=)b)7&sZ&5=yX+Qaw4U=p4((0j3Smqt;@o9~y#ZDw3`%SRO_jj^H#_wGJyw!NWoR*!WZadyq{sU zpq7(aGvcU>24Vctikd)t^i1MR{&E@+2YQMahwW-&G)<8rjD2j-J4i&sqT-hq+jOC%z!}> z#JoS@Jjcz1+!lB>sQT~`$MZBP2%B2*yUwJ;P*4&UI#EzY-E;AqGgF;4lzs(H9#tlL z`3OdQlD6+MD!0{>r8r`_O=}19Z!U9E3#WkT0e_@DH0(H^de5oKC>W>Iqu+qq_3BwZ#51H;$B2&|Q97SM5M&JteZUJWG4L(tti%x1?tzx33dn+Oe9NWXftZK3dpc0;$141w)XQ2SQ5cTka-~| z5BvY)T^^Y>XVZsKsG0`9+OS(r}vyYfBTrXESAj|c%HYchK zeGcek23;;WZcNdF!pZQWo@?@w+gg`^;r2I05gRPxSw)Qfx8xl_&Te0Yiu;xKV3uyj z>`IO}4jssQ^{E{1zEFH|<}2Ut)j(qaN1jcEvnL9ZkDD#zdR~QYotihX49$aDRox7!p0nNsde}ZOy3^)F=7DOxSoy9E^FhzV-IFXIUj+dps7VS~O@zdit+oxB!!5EIK#%w<76#Szncc7x`=$is=mTB*--=}UXB`ewUSY@tC@$-y&hkI## zk)G=u7gZaT{1_eLH>MJ5Hd?u9$E;KN?9`lz?IRHem~F)et!2|LF9mgw@0%|B%R?>B zH}d6Hlu<>Fn&T&^ctKwLkyuLj<5S#)l$#VfR$07wMnK0nBcVAXC0CD;TZ z2d!-E;UbInk7)sAp_r&PCrH41@$1*J zpz7e&c8j_9xgQ}F*)HSFM{@||2NVs4l$fNg3SV^IDcs1BKjDnzqG||orhjopcnpkO zM^Pj8-WRkvHCj`-glo=w%b~;^n$0)|LDUqxy)wI`>|wY>v`Z3gP|`dZJ}dYTa##tuxn6ondb$Hd+SxqmQIWJ?^Yz=2WHll3(|(@3KR7o+Yy$eCuTRuz z=j+t70DM7Figfey)%W*4Bc1k!bDQvpzBD-E!vp+-8x*{8#+Kvp_s1@Yk?bobPo@N5nsASk1hGS#BI0^j* zuc^X72W=W=7Q*!`0HXYv*o$?cDg)yTj7_tsBRYuZ{l&= zif`NP!3J}?`|-|>VZwz&0R*}onWG`u#bt23A0HQ8=uyM(IZC2{syNx8PM*eKf;bjr zDKFQ~w&zveHX^qNiY$_g7gH$^`Li7vE7b5-*QWrRiy066(uW5#D+3w*nQb*FrkAL6 z1z_burp1)YTxrlNZ*@MAYET2dDZjnxVh$wZg~3DEMN)ayu7#3n@fVTqJ41^EVj`Ea zPbw~Jm0K~oUDPOw2LcXvTb_2sJSfKWDXbl@j~e}W$E94aQ}qh@M4~|jLwx+`dUdxM^whkbyQ zq(ty`wd4D?0Wc^!;KKY_GknR)>9J@L%jcHw`TXmf%Z6uTLnkmKZbOrL*eQ;Pws+5X zNnmH(TeJj1H_ztWQEof+dD5h8$Gz4UE)|+Y)KGM_vvCaij#>sU^NUO;W0??%RP&`) z2;=?mBoQ1z+Cre|q6yPaNf32Yp!RIDl}ne4HQdZyQBMAz6Gf*7Di1!+96ulX)?beu zNvl>^`H%Lai1Su2Q@YHY6gRi?trV48cZ(oj1rCgFqli0|tTxNN`lb`P@89|P$pSZR zIB>#&GO(^ipzSiW;EZW;`KWf(wa|za6fuSy&9#S#IJ2MXu|gE0JgcAG$YjC~V8&$E zt)?r0sg0)RX%-s}PvI}5ea~rTop)n65&t)1;1fPspUXBQyf~tgRgiPR>1Mv z#dA)Pf_K2y&1F#1@Y+*{2`R~f_P60$bo~7BQVH_2WA+DD47d7Ybs_szvjl}EJnOau zO&&VU;#8?PEhR+sgRys&!BnEd11LG7@d}G5FbZ~L9`H{zFhOULL!45O22^pQ*5KA7 z#9tN#VlUcsm^G8%$^V@8E8$Lr@;{;<3*71_c&e9{+WW+*MVJEP%R`o+9W?s-`pq7d zX}LmWmyCA4Z=V$EM)B$;f>Y+gYKtoHL9g8sbk58@ds(NadTeWCG#Vz!FHUy{Z3)+? zU4lm0yN3et&IHsA-J(L`k>bOPS7UZ&dHp|pxW@!wP|46&ECAvWPj6QV!=0%1(7sA7 zcK6E{()p<~H!vVM=hnq_Yf-C0v!!px)B$)SMKRhE1b+4xUp@SAOqDZ2fron8X@_)- zIj8l0hfB1osLQTdts{1FLv+;W%(yLd>T z%lsfvDikO#DK*kliP=7WBmF6g2$snE(QKeP?FzNf$cQn$BgRp4cfbV7O6Fhy5sHva zQRK2;hD^9txy2OIXt7Zla}Zhr0X79r<-LZbE+ZioN)SfLPM!fI_orKYT7cqhIr!O^ zGgcfy%dAozfr@huFsa1cndqNJ!xxjBKO?A)JjMUN_P#nQ>aG1(5Ja&95J`hB5v9uj zLvoP z^WFQ|Pk)~0*;DNqB_*k`A-6vpEnxNQJLisRh+0>@Ii(s?ST1_s-eMpZw1GJ_-&QWa z@3ZkZIHM8rmQFB`2wZ;Io~of@|&31Q~Pyhl;|cQ=iIu=7(KJ zI&L4o^dcX44zbE*VwRb7662Am?_|=1;XVDDMyqW)kOe|E8fI8jIzz7wV0K@v^(^l0CvUWD4?ULIZR7TUzI6}asbe&Sq8PW&7i6L(`QdCSF;T@e{M5NqE|?~|F;OhRm;5S*TV z`3+YixI6E*GPD94WQOmrh~6X_()Ac)orw^!zva3J+EDcLP5W7I zS9zWL%8+sNq%|_z7=0f_z?0E^dHUg7Cf@TJ?{m#U8`;4`RjCW#F0P7BPvser&gyb{ z(Z!$HkP1V_DwHtB%y9&|vX9jT^5(`MiY1<d?O@M!Kr_^*8V(wqveX&tNbm19>yu#zrMbon|K-C<%=kFDYJW!X=B@a>`_Mr$Y) zoTgD+qGS7?Vvym0s!DAcNP3AP?HWq#K<&u$DZYYE4uAxOFjb6t|#9u?F{FM%{SGF z*Ram;&!GFg=eBq{92u(^)CAOD;XYK0{O#Rx^-xJr2?kZRNpOt|c$=1#Z&nC^^IS@k z(L+ZPi1zIx5KyM{gTJnnoE;#`KQp!_ICS&jy^Me@y4(q@2pD7u<9HYou1FZ^_(R(*=9YN z*Y%E?3=%J(u-dbOuEhExg5)mx2FEN{)(dF3dvhh zR3262;VUWDdb3*COKk^^igFwi$%b1+3e{)^RYy78xf{Hwkn5UJ*<0aSTEfz)VLuRV zsc`TFpHg$pg8 z#3P$;d`>0V#wsiKseRW~XDy=`4dhEgd+cCdQR&nfn@-kF-tUshL-YyQ!sS9 z`%NeuFpUD3#fv^ZAkP)s$s6LMoTgrRUxt=GA!>a^xO(c6q`TxpD!;I}qMI&lYiT~* zMyY|f7bb?z-^8qoRCVGGLBHNuEB%fn6Ru35Imb~|>>=d5>i@fxlk&DgQ)>lC;P4>`2G8B$t7N_EuJ ze$Y|C$bd`zR%%AoxfavrNXu_a)rDflK0f|8NKxC%^~rsq+NTB3PR=eXjRHbX)U(Y_ zhS|*1-fW^9cJb_M<2F zk4#=9yEW69X1Z#csN)XXk(NHeC^>{X(H$cLCflG)r8Q3$E#%H%mw1a=Jl>yJL~BTy zSx0z2a!>J>Vezmb#pBR~D(BNIm8B9JjLO=^aotbLkGf*Gw3grdh=XS1U{%|al9hR9 zwKoiDxhAO(*N}yMIp!KI+?0y$;Alxg3*Ezc2k5p|+l-YqTZ)wIk8B7Asu_k%W;Y49 zd?{z2LK%}vO=w&XGfyG`@+1lUkq|NYGa*_?ymtS&_x)c!oMHLAPbEC6=PzIV?P)8L zGecH_fLOo%l@qKAl0*nWiL68a@X+W%*0Ql@g6Dt#D~$KT-HA0V-3(9v!>eje-gB=X zes*J*D*t_Jp9P6X^GCKDnE&BbeiuLQ3o0Le3DNey|BAb$C7@sBW(eqm`t3uEGLl5x zZ^{q-_5q-dMgaZZukeNa{#V>3eZa<#y*d5cwSd2^JOnh>gc1Kgy^0oWoc0*Y?{);O z;Qi@)?i@|cX8-i6KaBJHBmcuV|G-^;80Qb;{LKXX(Kx@?{C`Z40Lp{qQjWag|D$cSHum1bI=Lh8C%r z%Xcycw`Qb@a?RUZG%|pX^L>;=oG{{fR$#vNEhU#OsN>HnQ1{3AHO}s@bq(uzy_qjW z+!imNQs>KDDZ}7;rk=aVEVU;?FBP@rYgHx4rs)nFwlh2h{NFRg%h?ZD-(Uaygm^jf zwUAOPD{20zg>?=jijqqEDyO!Kv5R+2ntIOr0sFpd2O?diQ?p0jU)#F&=yjQdwj0y2 zY53C@RDJenabEq!&j)-3{K~!&gV9W0Gq8MK3tHEQE<0Q*F!Qk zMm#50g@79-SwJ3cH9fWTy}Y!Ud!ffD(knn&JxMbtbgOU_HFdI{x-Bcs2cXvyBio1J zbRjbi!(9hxQW=c!%d6dnGij9bR)dm6sG~2hc!9%WJ5+q&fVi(iUr(o}8 z(P#V;xbB!QC`|#v-zg=Y)-EO191uS`?>f6AQJGO4^~np?XSdW{)TP~$)F#a9+&Swo zTuG(Z@2|+R6LO*d;_(3n33c(YlZZX$>Yd-1_J*{@J zI^vmCM^+H3)~C;{$Xu^sk|w!)L+v~*I-5zccoEm-TpPSXREk$d&(frYx3~^?BjBj=sLs_@zLe9g=eFHXNG-$sC~FJ@W!tScx)+M$p^}| zOC*NPk#rFP0~grVwQa7?VtM9IDoPbVL`SJIm^Oh6c zMb_WUveuNBx{L^c*=%J@XO4>d;rN4-hdivRo=;6jIPkCkg zkBgs4h;);-`w#*{8K8(Jq!_t88g$NB%@*r%=Ji`SqD>}AsVW=u`UB_c zMK8|#Xuz686fT1BDm1L@Ihv%YzsYg&vqy$?2l7prd4_G~8Pfch(O=1|&<7q;h~8^b zy{uJonh))@vl!z$l9o5<6Ue1eb!?z;Ji_0xxGX*l79z@b#rlnVLj-a0H06*+o-Y_t zcrccWQBh7q)rA;+sKyzU_3Tj*IK?^sq*c-Fngz^gR8*vQ#O8iDN{MgD2P;)CVen=* zFGwq)fF0Wa*^q1Eoyl~<73n%r3>xMI2HxUFN$xQDZOTnG32h4r7&I+Wpo z6fRSBXHN&JM5*k2L7hiZa$WEd+e#FiYkzp|O8;5nqo?u-9&biaIEJ^Np}_!F`bf<9 z6@~I!QyI;ZVP?y1)65dU<+v9Jk`bO38e^ze8uiqN^EmVk$9c~CiVTki3)`C9b(o0Q z0Anv?gBt{%zBwXZJ-~MALlwiKd2t9eHW>r;T8Q(DO{M&Fc}%oy8)Fc7f1jqeX7r_RIf>0Mcufv+bJbY*FVaxY4Fs{0vyi0 z`C49$Rf&xrfK;G%H48vPBH+ zg$=D`dqyqUT&FEA->@Q}C3kK_R%4vex+L6jT?k4ln>oKZmQuk0C*0%As5m8r2|vMl zYsjzsK*eDm)Y-YdJYVBBeAwv=D=z-`1vF{GEHY&kilFSwCx0MvTSaVGt|rwc)o`8j zSvuz=#Mfu+pK+FMAJ+vpJj&8*)0ra0oGX~hx$6cIMTL#cuMaPp=LdZF>p^gXgIMMG zJibLSWD@!*du%J&h)mOIb1_(g(3d{fo8#Fa5Xz4>Yx{`ueJrmt-E-)3=~mUFJYm3|YUaqJg0TUJB2C$MZfNQUhFGb!gE%Jg5cQ zkz- zj%If2nvWz|`r>5nyz(D!7!CW82}~jfPV5*(xMT^?N>16;4%s$QUe~VKzT@u$}39mdb(y##12bBY2w`!hVu_4ZmjUZYwaLdV`(GCca$dH8#uL%+8~Fzs=rH?+#vl-wo9LoT-^C?c)nI1uSKICHyom!b%F-}Nof(h%@NFE(aN zqbj-dq`NlF@_7Z&=O%PE=JGiCn%_8M)`GH{-=(ZH+CcFaHMBR2woz1EmgeP;;g;9E zTN%*6QC(`?Sc3e@m@aLx7_7K+*CsRYa-!J@p&#fjvLOB0`)c#tE?xvaNIEOT^GHw zE+3WnMkTdT<-)u(V_$KLG)>k1Qe5M=o8cgN{+Sa9z@04MHMR#Ae`0P%ON?##FDGu8?Wsm3~ zmgvL*IcB$=pvki!&98Fo5Z)?Y;qijB$drU`S@TFLpWBCU5yod2dA*Tn9i5M_YWjN? z@Mt{ql~+D4$kQp2=QnU@iO@N92uG`x*x_~CC?8un+9Az|EZg2(iK*}GQ*y;Dm-gTX z>;{++C0ljHdVGXukDE)Jt7m(xGk4ysO0XU{elFNPb=Oh02lk|O_H%#URiDl18tRp# zPl&nPz6CGqW)TURbc6c9r7~A+)dA2nk3Fh@tE9Fh>_5%~+#P}$XWnckP*r~k11W)9 zY+|{fk_3f;1DjQ88bk<7ZATfxjJH5cmpD#(TpJOhGOqKWbDe77Fqx|1RNTcCG701| zl>@c}tBaj?X6lly_A#xX%ky~`pn{+8aR*L}X9}`}A92pooY-plUA4-<^W}}vej<<8 zzpzoXYn#<~;EAQ*#7*37{tGlJwz8eqkjK}x2aMk1hCqmg*#10+^1=%>)HAJYD4vzH zTrInq$nEl&oKAJV`Qo`i)hDkQn7{JD5i~;6Rr#x#93T0c^%YpBMTaz4d=((yksJ7q z5#2N=Zi9I5(5n~1fJ3IsjGwBM?#ERFqu2g@SO_ESTPR*=J0PCWD7#6nDPi?U)2e4t z-y;N;N0$yqPlBe`4!$D%X~@79Z}rZ|5<=-js%{?N<_A$;5QtpA-}7iNXK`Aa4_nc! zOUs=-C9;sM$BEj|{ZJr3wlH5LVb~ZN@v=cBR!(?VPagq`c#pnS09$&!Cglg|X-z_e z-h$8d7jG_I@6FR;KQ@^MPt{WL3cF%pBFZ~o8y@bGqa;#MetdAEVDO#l0GZ3odJCBG z5{s>mc}KDmt!>4y z<(9Do(mEWX=~ajkN9^v+YlQ3EG^(ejVOQG72MLWTkl08QkW5gRsaUM%_F3zyhlkZU z?SC}64@FV96e#arhGl0|#A$DqIFbV8&F-*5lWqeB7ch2Q(;tbI%~(MU%$$2*gbr>I zs95ha*8uVJ7|Y^>5&jgF=2gxs&4cQT0gn9r=S)*imaNZGv(Um*`~11zV)7q*@RV+6 zPUpsVRWvF!Cv5}!ZZ(qqwC4iG zIIoFRe%-7wly@2X*Ly;Ghfd5KT*6L7`ktTAIssD*zNhzST*Fd^o|n-_+)6UrVxUmE z^z{~io?-eVz#83QV^rjbi&rH%$HKT^9InFT4ScoOf!t1JmPIrHF94EhZ1|n8WEsnM zI4Y%SEGB7S@)kjL6Z;!O_*j^gwO|XyIBmQ8AfwWunlPq2SF1iT5L&W@0d zdQDkOCxF|N>U;;&3iZmWGa9bc-%7CUe0PjaSr#kL!77X|EB3*P>_hd^qK zG+&f8^mo4j{pA^UfX60=JO1uBpr2ocfYj?Pw|=``K(Tw^RX4^OF8+Q~jD8SwdHg-| zZ`TXn^$M6r3U(v<-*0N4hbT}`j`yzq69HKt0oUf9(8@pUd=$cUUqrwACjx!}uXW<3G$EsL_l&10D(Sg68)#0I|9}8hjo5Kp+Bti8^Y|0+aK2Xqjmm^`uU@E z{y}a3F=>S@*p_u@>Oz%fc$BE(CPEhx1hj$88tpc>+ui6`aB zi6c;DYb-oX&j!E-M+FAYlyjr9o1_d;aW+$xRF#op&h0`9EFPq!+y!P1Ud2xyHAIy@ zHf_W}P%QTk5VTR73miHt15trmp>wT(mX45KA!<3^##O6yYU@3i^Ww1uQVVN`rdaL? z6FC}8an|#GB-VUqXWN;o){ZgRV86<%Z$5lJsF<{x(m|+mam8|F=6#@fy z8t5TMw|z9KD*`p_pbV#~>JY#v8OYWoIQ#me4LtL)^BFCUm?%zBPG`5ZYh9I#95VX8 zeCUtPYYaS6Dau?cJ{`BAv@ScNfSmor1~?a^yIm@SpAg6yz;-PVyb$FsnkIgJhvW{} z2xo+Wkjw*jj<55mwa&g4*X9!c+rDC3&r(Y|)oVAOlTqn7Id<@cS~T#Rqk7^kyCU+q zy65t}NI}9wUQ=zYw80c49Gl#-(sIWvvX|m{LH^VXo9@}<>xy&1n%DJ?KSgSKkRsF8 zx^#Hx-G=OQbc+Qnb!RjC?3*cFRzHwsg;>D5fC+!q zLeoWb&404hU?uN7DLN$9WYEfsYW{b``p25F zv(^Wy2e8}BG{)6*MU{y&$h2$IMp3SmVYWTYQ}9JRy6QDeY7~@cTg&OHUgSX3-Qdhx zZC(0i-CQX3xjtH-__)L}PH=M~!3bkO{uPPZ*3AZCwEhw)2!JIYo`3}NFc3`hr#3=* zH{-||SYYh^2vi6H4QV_OG+qWd+b0%=Cjnff9g_R3%^SH-<#~zFaxg}>KR+S!!?7v& zy(9Ej)0+}$@S_P!Yh78HdI)MusaaGX!q%rtjhZ-0NsQ*%>Ljn_K+XNy=h(R{#Un+N z&!NK);gO&1q%>)!gdzUK*;Ise8c6k3I_wM|aWwW8zmM;rLf(u@B*HRkbhdi2Q>Coy|f>cCC>_ZHD9WOK6_;& z{3$5e=}GH*j$-iOkkO=5@!puHDZKgSh|m7x-2~j`<-q_T@oeNZjSni)$QuDmy0naW zZ9Zx>_ZJ9g>RQtfmiqt{#-1cmmbe#NwSXbN_K$!lS**@>hzx>Cdb8N9=52!`g4^lB zG4sWy*kD292X+m9Utf$2y*QR=UEH)xvGI6$bTU`35hlr<)Bcfnq;5E4Ah++9XG1x0 z$-y&h&r~|ht;m!IxbwDh)|54>(gb|%T$NsI$hu3a9ja+;9Z*0`ZH;?!gS78Q?Xcpc z>|6vIA{-!dAuuSmQ|+(9%Mp$0Lvc$Oy;3)_>g9R@W4vM%J!qV)RcqFEUVkh1TI&qx)NL}h zzww!u__*lCJqQrgg&;cPze03ZnY}WAbUfuJw#u@}dk-r_ZARCF3~p>2GmO0?WhQSm z{sNSj=Lf84ex>&irxTImShN7_{X!z(MJ=^OfjkeRh?9^{ARNQJaWsFWnTD2Z)%tM0 z@YlLxz7^V+V}V-ySMm%UHv?337Yl9zU5RPUHPKnMgs&e%UW31OPIWdOwG zlafesj)M_tma4kxqXG_jaDC4Hu@?iuiobO#A>vduz3Zch`&2Fm`YD2RPC2j7-k)%Y z2w=Jl%Kaw?@^rRmbYS{asGm?&^&Jj`s@e%5Hy3=wyAkf_AHSIgCJby_V>?gaVCivwkfx2ww}WJfvwpiu{ZN zpk9ok(n-XYZ3jh=l<&Orqo^&}&k7!3lwXx>LijmMq>C_hUh5r>DtQ4PueDAsfuH7| z0@5;Bke0a%*Ke-53oaFh%Yg2ed)8_K}#MSZcbo#w?zhuPyf z(ypKo!59CAK60+PUwyFyUVPTPH#^C;*k}1<{sNxt;>y0|$thC#ZE z&Q5kq#=Dr`uK$aM)x0}YTL?<4#Bp>d7K&FaT5~{qza`#rLOMk`?c*+u6G?Vy-0;)t zg-kBTuX+J@w&(p_pnYBgK>u8{{t=JXLp2<7Mrw!;;NY{fwCwwtm=*DtY32YNQvKP( zsN?PSSdfmiVLz~bk#izz6D-xBuP5#C(Znm}`E%bu+7;djG?M6*biKbN1;#@KsON<38Db8zUi9f zc$0Un1e=|}!AG|s*F#w{vQw28bLb2=@QC^m4-DCXx0-+~u+Y~O-|jOi&3SrqDdqZp zjNgmnccK-e#Kn$DEKh>?WX!DZqa63iX5f1UsyR*Rd%akutZ_*(4;m=kMY9fTnWkOC zH1%1$qfp*>lDXJmed52hvyo6zZgkPWq>--cjp8Dg+k!SRp{2XMgit05+4QpeC<~4S>npWH5{G-sV@%#li53_ zmStFpa9+KOA`LoJ)TSU(p{!L1VXh-sUp((1oPe$^ig4>?vz%>_0eB$$&7r0Y`|8hM zZ+@)I1=T6E-(dvRqU5sz26gQYy8Zr3CLLVzDLN@>z31uO7R8$vk`FBLAtis}N}uwO zsH$JO1s3cEQC0+E5)I(gfr!TmL_DizDHn^}Qkb?uBUu^tB7jER`b$j50F!XQnh&dG z5Nn<{$!Q1yhLYC0|D90t<2nHwv^cIJ-IKefD=v+gr9Nctnp| zpb`}CEc+x)psU8_UG-ED`?^#!oT}h!9V$IA!JVZHiAr!kskJmRe}vq!PL`z#)lvNV>gxx>{JTQk%`FacBWkK?dKhFMiZUOo45S4b#AdI!Zc@HFNi=DZI zu;4)0!N$mk@n?*O(t&Dq72nP@4S_F;Q!&6~JZ^z#8Vpf&Z!5Vno(to9&v?nZ1`rrp=bek*7+q-NQDe#ew9CZ?Fw;#a>G_mvf{I5{2^ z>-ua@Op~SXAx_q3wFc#Sqeo+O@!O;trt~}oqo*vFkUR~P7?X7qsEcj$AQU9a{S1-= z1mb39w&c51Y0 z|Dpr=we;NCH?%JI){>$qE-bD_MZF`Ep*y`Y;i_?7?LWG#-W4qr2(S$qzR#hGl(@P+ z)0c3=hBND|M>Rd+A+7VatnAxdb1L5w=Q#>!{6&^gqYss~3vZR!rcG-U+c&Z0?>kE| z9m-JgGsrmo7=Ob2-vF~+qae8g%40ZMvVo6m1kQPB6E~c;3l5jg=7wHPx=9DBCovt09F8f?&r@m^-<2+X;u-d zc;_=CQnz%*bAEL-Sqbzyv;_-U8|FH9OHEXxTv0fF19R(a6-3$duFfq0dLEZpEK=u% zT2SqJ=8B9Ci*MDu=;?dZDSVv_1Sg?ipxlK#CJeFgycAhaEe9+4j4Wo;(&>Utxk`a3 z`ls}1L!|CTHdq%miBG8X++N!`k3lx$xsC~$N!po>Q|lm0mwjto4p*jZla4n>MY$e( zJkFE$e-{n;5o>)f2K82V(shiY_yhqQDbRCUWWJkAS_^KTc3%QFo7J6K&o4`OfhF+s z_dni&&KPjI_-h?YB}(0pxF8>@Nas#2Vl7Ih(+(RS(ABjP5@Iww4-dMnfty>k`f`~( zc}_rdd%lbO4uMfSH)DR(YYI9(9JX6)wAtwdnDkE;mHbS*yBbd!dTUKD8tfzf*7G~Z z-+z5^pp)9RJZk$Y3h$(B*8ozKZ$VpbPScVw$IYRvWTX8)^kU4Q0+(HF@n0=NjQ6}i zLf2Z)H?m}8z#d)yS`9Y9Pz~Bgb?)A_hOqKW5Wr}07>NE1$l0eVCSKiHj&Bv@rmb8n zN`Gs;r^c)(fd=?GPycN3%hL7=kN(AotrdUFhMq5iZ`y9a?!B-0mZDt)=sNKl1Tfon zzA(!)r-(-b!|Y$1gkmp;sZHW>Yu@KyG`|Ba(4Phhxh8QZ^MAB0Eug-L*whROF1qiO zMjgj!f#vx{1CmMgCw~W7h(;={reH;;0e;7aQ!ZMiaam{fs`Y+JhST~q9jtiCkiWIZ z2!G`Ot($7Zefz;2XQ7z?1-ic}Ea_(NHwJjjh` z{7Q@dWPb5o=1-w_o~MJM*6jFOLyW@dWsV6JHzhi+PEvrHRzJEnzvo&JUv2nJJ%Xxy zeRsC#XzO_hFF>BwsUcE~#YoI~uIC~*wcUMC?<$-yaLA+Tpy=YzO{k0Yz2nNbK>`=B zMoVwrFnB4c%zgg}9`x+mBldq`FbV*JSw#R0CJaI>fAYKNE{7=|#xqrvWi`Dmhub{` zP{H_P_KKr){NHrrK8w`dx9{UD5IcY*PvWmCsICwYuzbsXS!I)CS_xf)36MLV` zM@@%357HJF9d9PsDT6z%bU|OP?n)!S`Q5Xil9%!O4~kI%ikU(@CLeDz2Y5`4o3{4f zdF;*!h{w3G|1)5WAH=#U1!`GO&DK5H@8e|@xoOJ(V7w0&0{hyr*1O0OR?(#ckD$Ba z)V|h|HnVPfv`fW<_+L#kzX0~>hZ2X{*)IRJkq4s;8r$;!(#ZposXz38AkhEb?*Ovm zrHs#5j^F0?jP6(~V*OMCsoq^{T!v)>E(P(<&sy(B>0Z}7t&z}^DxK%nvYN#`8w<_+ zSAhmcS4h*C2Lr^`?_`Z=$_&&nbH+qbWNlP%J~(FhM54-&zoNMJMIUi~ZTL=9Z8wAc z^d~M&z z@$2Yhj!N@nN%NpBH)17TqhfW6FH=uJcqw0Sv*PqLEz)0doae(9UAqLcypzP&Zo9BF z%yi0Y8Ql8ZPiVy+%IJ#xO8a-TjLCPLirbmm;z~<+`h#R(p^X}#3YY|k1ED^*K<&ZwR7{^xqMK?!B|tLYgOMVIvLEH%tB9KA|* z0@bO6H_MO5%p5Dav=bw(O(Qb@oeSVXkDqve5|CR_JICHalfkBXLDRb+q1YPqz$yQr z=c6C`|ECX)zH8(7(5VW6UNw{28x{4`4N@DuS)l9a6d2+h>}~fGIyAY3U#{o5$TtrT zxOx4j?j2B0Kkz7oA8OXOoOhfsQt1LW>pBdfob_ISs{f+iHaAP@enQzzS+r-f1P#o(fL>$Y9rlMB@w zMLIfZLX`rwXUWF6{5F;QxQlFkd$K01I`vS< zYJjO0n7ZL0=80nmIXM@R0)t-KxY;mQyT|c7VDPUg60xK9>2Oo?_PI}N2c4l=ZvBv2 zo@Ye^zt3x`$M8b^U(3FJdawkVb^;;ce<4cl#)@F!dHEl z@Hm3P;M{7P!aF@*adsj5d(_1CJ|!Xtj?#bR%>6GT@w`-0ER5B6Kmyg zZ=)PZx9GYhK%r7xvFqA3s5{bTrqI>3e7De{XBQ^F@*`we0U?8t23^mZ2G=ml9*yG; z1A$s3htT<7paYfL-#~{v^nF4W0Zd?Ld97n-nOdri?|Y=YFMx?(a_C&*RIHY1eG2TmUPdO3HXQ3HT`>%`+~m4vxWTjMUspq#Kifl>(U0l z+Y$KFKcxBXVg4b_|D&Yw0L7U)C{apo^)ql>pGd}lU7r+g9Vew$ADH9(wuh8-gEBTL z$b*7Ixgxp}?p^6gCF2QZ<#lQGRe_cT+Qu*eWYC0j%>fUCD3?=XV~R3YjZc9a%^RFy zXKk~Eh}Rd!6APfB*1KZ$nfKtqp>sE>(!k}hKfZbFPt4%DYI!$YlxaTF76wLYGl#T+ z|8J}hI#&*5HF+(pPeo|j-xDE1<*X#n%}O*oGgT2bLW4nRrAglS`XVnFfKgM=3(t2P=4P?p26eVHI{OuU!{;2@Q&o4gB=@OIi~q2qy`uJH8H2n!;J(TRr-sR6mf3G?*J-!T4u8;?jIf z(^q^T(8GkVAh-Y(IF7L)x3(Uq=n3 z8uYf6^f0lf6>lqt;vy2V4jg~yh-_%q=|kDhYzzPm&bn7MnVqU)WU%Imw}MPP@O2dhNss!|BIz!N!Z!hN8Vb9jYQZF zHX?D^_=Jjl-k20>h^-6hJ1&BmT!9Q>4sAzUQYJDw#tZH#xx6)4s=I!{$T?M$6wo+wmRdOTOSuwLxkDYoQq ziu7$3#}!WxNi>r@3)5$FYQb~W>G}O)>(3P=B+r;SpVj|pWCz7^b{<{Y&R`~%5VzVm z##fBCXdzd_N_%stcLt`37178W=dqpMh{(q^={%yhD?UVo3X@E#mOITN|1PX+CkWT) zyihVSOy;JwI0fU;JS&5p4DgazuX(YFOuEZ0$ja)6ducZBA09wBCbaQ+W5*hPoBPnv zsrTm?x=XLcU-_`NE)r;6-}U8wfeW8LmZdt+d)6cKuau6`*>5{-&-W}puW#OIUgtXy z5y1d!y`+ne(s|#{@@|^~Ofc(?7ngRk>ubi;E-!E3$QYz2nj%J~Ep^-irsE@b=8NQ_ z>rW1R{DPY=;8^>NMjC$$Tc0RaJ(llL#iXyjdt!FZ6d2>kD^Z%oC@M-AH_ zrnCG(#?G5MZ##QvIfz^99GwWmM=uEn0#Y{$wh$36lE___S6JIwxV(7Lw@NDLNOP$; zR^!czcI0UTE)9z`t1D}Rn)`wd4=D<5#tJUdr9)ceUsC3+srD&c0qU<>seVF`n}*szgc`TX#$0br{?q}%NI&TVf zY4M3G2)dOoR8y2>+#=hIk698>45dBIV!mNB0i*!~q#`yZ-%I(td!Lne7CbLJM_|ui zG7}>r>hH|L8vb&!JPw?g9KgO*O&{vm_U9oCA0L_h`rEsZRkzVQQz}M#}VR5tpR$@Yu;UGhl}tyW)H164Y_5% zg-&xEG~urte)$5B{pDBLCOPukQf_rz5UD%0J46Qe8SN3PwH+Iz3<3BjrMHqZ4Bk>hKpjsjHbBpehY=;gG-AU@}1h;hTe#c7W(V|yXNmp9T zQQE5KL`(OOc|h}7hg=Z1w@<`z*(!#OwUP|bFm!3+Sn}b;G>g`ID&pIVF@i>u>0nYC z;iJkdsBjf`>9edxvUJh5`+~Q+(RN8x;p_})i-x(P-I#AVNZ95nLuuhey<(G6Bql$3A3B--sO5T(QIu4v zKMwTgpaMkkqyAf!N_y}!AcZOi=&S>zVaK7jbPY4fw)dhQVCoD?-ze{?f%Fe$!!q=A zP|-eq`ZFI{RpN(>{GlhICPW#F!PdC?WxXJ6tNWUl&7S5sazCOP(in&pM|zc*xNL8) zi+q%cv|d-^9mynsA(>abAR24Vp|>no&+Jin_VB@>O!Q5Ty^yW@_T##CV@u3S9yl&d z*8eSsjJFq`0)tq3-<4vwY@HC|qbKUiraSKugs@~mxD-= z(rApL*ZcSHw_}cpra#=au^LA6rf#G9@u1Jbcuda6^gN(Bsj^4XI*jOS&@X?CP<9g2Qtd|LNkzDJ{g-rs*1byL zw_=m(CzBG6dQpuR=d1Y!3t<8$og9K13%le`66Y5#5aaUZUq|i{_qziyjot}C%6nH* z+>?Dg94%!zxQR|#7Eup)?O%QxN5{~$y)ix7f4z}pqq8B2KMS-s)8Xs{89S9z(;Cfd z`!y02y)PEhM*@2k0Z448P1BqU7ofD<6WG7IS{1|`%BgSf$5dCq7~ofw5t?cSaU!$n zu9_~x*pUi1mRnhds9Yqu!coPI&J)S6+pNAA=FN+l3Knf;_Zceqa=)QdcP$GzcgSUH zYFY5}!wFSf^Y)5jbE8jZT3$*^3A0ppsU3t${ID+`$$+7g>6`Yf&ABg<^VsY`H|HnY z24A1Ea&UM8q}@B$-2NS??GX6ZVX?`yG_X8AxCeUfl5C0+XwxEz+1>{~3m`v*xESGIoBZ30_IY># zopq)2@)&3g`=vYo^Fyo#NE6KC6McRu^S?jHs0XO$b?t7!o|OLm!ynT8HWvOvnm?rZ gcY*nzCQyrnNKI8nvA1Qs`@ny)QVO>aHy%I#U+P#`Bme*a literal 0 HcmV?d00001 diff --git a/contrib/machine-learning/assets/cnn-padding.png b/contrib/machine-learning/assets/cnn-padding.png new file mode 100644 index 0000000000000000000000000000000000000000..a441b2b96e0ec075fb0c2e2e81cf21bc33ed4388 GIT binary patch literal 98306 zcmeFZcQ{;I8#hc42~koIorvCvZV(c^_g=z`=wZ}R1}P$HbWxKi(K}HFA^PaOCpx1u zME|yPPR^6(yl>8TUEjaoGuJh#o0Bu)3-o0UiY&1_lO!g8U;*42-Kt z3=FIf*RBFBLjk+M;fAD*jEuU1j0~N+ixb4g9*lt@9~`TPtFP5Tnrft}@Dvj}kD`>8~o#k}}{W;w&mUVjBoW#;T>`ajD*WUTY zWK7WO6r8%oXRpKP2uYV8$$t7#?yA8<_T;m4U(N>~NAa)6#N2n5a9VQg1r?cFa(&*T zS2YQZTGHkfNckP}3k^+)wGX<;%H9j7$mLNG^hKo$>gEB?2VbWCH7)Z0>gbDe{Lp*sqvWLhYtx zGQPV33xxT`%C)`LM>mj<>2BQ6?4d^8dUJbDTa@^Pl?}#4Fa|P5-aKm-Bhr-!6~fi^x@cI&9oxD=$#;`YR`A=~9A0yU8~3i> zk_rCuNNkGuum!Pm%@Z9ojAQ*h6>IN@qnfAUQ)cmd#>}q7j%s&E6yR@M9<2Buf5|n@ zI7`5vqOh&`j`^8LJE-$BjpH>cXT6o6u#<=Hjeh9z`F_LbjU+)yFlGSy{l> zTGF1J?8&=%yH!0T@hXHPBlt$jzxY6}dix#p9?_p%8GChMAI$mnz|y)TddzG6+6YF% zm~;M3yO!|Ioo{Qf5BerpGfzl%tYoPBB~J;&4ZN84lI>DzQyN%Kc3)~OZar;Gyr0NK z0=8P>AKwn4vvO(i$+l1?vnJESSePYSM4?nd=V}91D*Tq0&M^IZch1kxn`f(w=P}mN z{*xDd9zAh5lGL^s$zosQ74T!IxEhGR8A)SpKEN`+!r{-E9D)A~cczgRJAj86XYL_i z0>;PQ7bs)AEPrxyE|#mA=H&VWY>iio$?b7#8zuTM!u_Ma-S~Ws>pAKhmoX^i8_o*O zR--lcL-}je7Iar%vp#t#OviN%f9G*vx3bLASZ#)WZk0fsw&$wd{piP$)Nnyq9 za!n*SS#IL(WV=iTrB|4bG;hqSwe;3c0-{vl_x2B{!sOaMlCM~m6Q@LQJ?!eYb!9e; zNPWcpVYHucA7>&6Q<-3}8Ob9T`qH_Wg#$;aEn|!$ADg{7w?s7my2NsxKJLhK-><>K zSk`ml&v1g~t_#zqh4wa)&6E6aucLALQZz@jV|q#%pvrSkn7HnmanoJts}^|g6VzqL z6ouu52@W-xzf-@JA7Ypw27NPCW`f+zzD;&}^rpvcHR8$Oh+yCQw>tb6q;}~azxB-U z(d4z^v|(NqUd79a-Vex^6=bZ9eBQxo#asQ_Sa#;2glxUEpX|AEe}=wB9gD^tOfUtQ z6^x%5>=HbnaHKr^VM5*a)6Njbnz5}&ML6soEaF&JMWx%gTZ>axB|}iIz}O*TLZl+d zmeywCfj32vU^vr9&i4)A(p|hB4>q%-U~>xE3)g^ zm5dM;>=~L&+;_0spmPW~@Y(2|$S$Yl=Xr(my!jjd5^p7sD!-`PYf(W_;}nh*%@h++ z2T?@LL``8$aE+E5yIaJX`-%l@1ZIq&LS({D_8nJzDV~VG06h}6>W{3U*LzhrN}Ib? z#685_KU)AdxroLbq-&>}G_QX6{EM&5RNo9w>$5lCVaOqpTs^cdt8*6v-q@&Jgy?n zBcDZ=BTqR`K-aop#%NNX*?_y)v3Rqzzv}zMor$4|sj|$;<8jr>f=UELuH14QJXugy zYacODJ8nIpH(}-?MLk9Rh!90J5p?sF7vXK9JR0}Gw@dBr9MAIId{#1+?|j!@82YyF z@_BJP=6hFA8ufYic(+KKSIFM<%?XAk**jrRDRw4WW?pV-uDUqf_qEz?A7`MHt9NmG zMp9Ggfmm#PM!8YhSTMekR~uA2gy=y`HV-fif9#xRQVJsJHbjapD{SA`PTuhPy7w`@ zRH_`6W}$DrOnr`0#)W z-3x1bF?UjL<`b|#)c?UZ^G5G*GHa&gM|7ljZ|UC}Bh8?WA=bXRPd6qm%ea zWbVA4lPw6Xp>?Ln4M_#%9p^EKnM#mWR8Ss&Hy|YVbbYNPWqzuaB%+pg268&I zv&?)~wLS5f>xb4yGhfegUJPXnQRkR>YSh`9no$T9uocAjelPmoK>Pz=_j=7si6)XJ z>s5(W_JLYHA3j}(tMxHB4t(%h_yb;lUVdLbfx%YU$I>eB;qv{(^&ZCMtRJh?U17ZlDh?bM&sP2K{)j|u{g9@_f^Lw zb?kCIZBY+SY(la?>^Hbh9YiHkBoSV(#zU+5X?m=Bryk%Z#tPHJkN9o<@gIUk>aH~kN|(@<%gFJA+lVCwJS!! z=?l_1P#c5Q_?FWaGC%4ayQ88rrGfZzPB*g}KXT_p&+)KDzpkpHK0}Sqb32{=>?l26 zz1VzMEo^RW4_pYQ98n)_A90P5ixDcUFs`-Vh<}58gNntBt<4L~CDlQgG3$GP5f{fb z1g(^cLOv5L6P|Wj?ODgRsI}7WuiZFFX^sk0E(by*P69>*RdJ4IH7a!lmfq_JaTF2n z-cien#f^NZD4p0$nRhpxA^DtF(o^C%=!lrk^lqdXPhIxbI-D6Jt1t4Bw1ZD$=3(jJ zlH)3^y*;);BM1@+sKYwKHiDW61%$kWsKwV^JtuFLF~rKbMR0*+L4p_h7duV!zRdTT zVXBa1pDFQ@qs<)N?81t#<>e`alD?bD+lgA%W&G*xfiE1U?JwAI-`QH6ejMkX?SQ~K^VJ6<;y_*R= z*E$|7wa)6btud|%cg6KE^OHF6TE_1P=>@gGZO;tQ0|$!%4W}2j8`Oi7M~cU}7w9>m zt4@^&H{Z!AvgyNbCY!arTS^jlea6n;qIcG_Gt_z|)Q+T&HP44P?FOs1oeqzoCnt>J ztZF;#{$Ji?Wn*o_C^7b}Fz#;q$WrHtU$EkR5HDjR{-LJtC98eM17q#mg|{-%X>IlO z##GMjA2?`%R!m|QT)LUt=?rp&BLQXa=8RdnA9)gv-Q5@96e2iH`XCNHCqC@ujLfSE zOq+ybJ|*(uDYB+=13Zd1V0{IMiV6lBaDEK~8%Ucf#c;T68K*B`SXYs^$G(Q`0Ez%^-RC=^X{w2bgZA}SRa6E7!S2%6cm7? zmZb|A>J+Xx`rj#virO)_2ucQ5Lmya^SMCa(V*h z@^pB1ISz)nrzmjh0Cu&Y^K`Ixgo=8C?)`CxC~$t+%zcmUk6T>rK=<@j)ahiLT)=dK zTs&Mn_ayM>=;*{R`4<9!l zA181JC)CT))xwk05z6rAAV0@>1cq9=*gSK!adM=)9M|HBlbb8(-o48Q{r&m#J;9zf z|9O%l^si|F6Xd?U!p+OY!~OTzKv(h0R#9~uPq4kgBO3>RX23HfMEH5c|G5AExbmMT z{?b$bKRtPc9{g9=UoQQ#s}2@P`zJ!lQ@Uo|x;? zxCyt1XWDk)m4jkRMxxgOKWIKG#y^HY#Qdaw8R_^+IckHy!pWiG=M25I&ErAng=wen$W z!^l!I_=Of!C~ z$8lKq85tFdKGqs^>i?ek>2$HLAp^_zD>uZfroJ+B+8I7@V)7uU>Cn_2dAaI@V1^@8 zoW^~=q4gBnWF`{99z>K3)n3(0rGz{*)Iy!c;jxK9v~XELr*YfrDSkDv$EsqBkCe#q zNyriaWtb><=Fo*Ho+_O9z#=p=6zBKb@X2HUMhIB^9{d4QY{)}(Y3)bmOX3fJPOZt9zuV<>rGHFRPxwk1JwvOas%p4F(!zpO;*W3g zP!V0oHDt1r%@zLTFcGZZF}IY}b>ajZo1$mf5Ha^n&49$MyO3bbZ1)1Xijd6@%h~lz zN^u3n#tUGw6lO&DCWO`U{oED1k-~UpX0;NkblvWEeC`XNI6l(f%c9pLm~B_eOllA1 z{ru`z^${=u1x4itafchT-0ZrAZ@Z1E6m6TY4%{m+NNj#zSLb6;=7?C=ZJR|{M_7Ta zm11M=Fu$1%aG>{%Y$bHM_`+{zp$`+|b|Gvi)k~B)-su+4o1ng@ z<9fSj4@vz4S|1d~u%JyGAEMaE%YM4j#67pTq}Y!pCs%oGsE{+L+>Uc+X9x2HoBh)Y zC^pTEuGW34l4<~@nXXu6KhBmnk{^wQOL+GX9J17XWkN=h0VQM~Gi*1S(TKa6QufAZ) zr+IkhvBb^aZmid=m~vOjLVZuHz9z_VMEa^pJglK{PN?JElVr=7$QYqMEirXU7%2}n zhPHrZSHR;XPaopqoB5tRKi%)zDz)C4F;BV+78MrTh`GJ^V+IsLK<$^b7hiYS&m2|I zboFAjohwLTHaa{yI6Ac~f54*6hEdNNa4BJDfD^9_h_?r0tRFi`Svh*@PNIs40z##p`ULdCRq( z$wliRbZ^w+?i_njQrjSZ)s$!0O`0B`3x@G3rv|Wzk8xeHs}J`uB%5M!_MFE^;$(Ny z|3uvk)xo?kxM|si=|f6L)qck}yc+*pH*>^V!AZAFmY{l}3A(VYO6_|fDtyvssn!)& zFo_Jj9f4vbONPFxM&9_{M#s2Pa{rsWd6u_Q80=YKONmL+NL0N|^!hBkQFF64{dr*C zu?a_tD9xOa{fm1=4M&Bi0`T{lbr~<`y=fW9QarN2x2YqM2}kv(A|3&mQ}Bm}%AI?w z9hsxJ_(pP7aE9yg2L{~XN{TeuwJpv z3NE|BCR1J9=R{qEkMWtQyhDvW3x}nP3=3eA2=Oz{@w|z=&kBby+wl+8`G$zr(qO_@ zgf^qCEM@#Wdt1|bwYr?-{7+tcZSIF9fd`=6I4_i!uKMu{i5_C1%p_;%_u{OqSe8GZ$ISH;@}?ZEGP?(NfDBS)q2f5y&{3$Of!)jGu}_k2*;@x%;1xT z#XM*kQ`lmvm=kaK_p^BFL+#>5q_>ZlgqyyI>vln0u^O(!Jd-LLZw6ZOFelez*&zF{ z!RrfctgkZ!`}`$fsCF^(_hVd$8U zUdgVfMAT8CdS8YrHzvctTMhf3VAu^FknZ3FdbQNN|4BvXv90Kt=vPNMH(vJ+O-uVQ z9SfybFr4=nuAa*>k0U&zby)=Ejxvz8pT30S zhMp}>f9YxDpqF?!k;hxRHeA526vwHjdjV6k^h@nU@zK&ep*89AK zdPRIg^RQTLc=i>3Cs3g8+Rk{Y=hEY@++*~b0xcS^8|H?(`*31lksUk(7#9xvflj^~ zHQ9=IND2Y_r{%$fS)^xW_tPb20d4)0|peI^whY_N(|A@#O?~O5x6?F;J%u6xc>f(u<_pS_a!eU@jzq#o3$@4&BGVW z4n|cE`rKX%IJ7ETo^sCW_Pw7-xOw==IUzTMkSf*dJIVYYw?duFv#{bV4ka^8 z^Q?UlR_tiedd$%SA@8-7oiC=IM{eVh_5I#=_XZ0aIe2+BHtRQ-Y^GR-g)3>C?B}9; zaj@UE(Ll?MUA~OB3f7!A(|WReLni=l97S2QTGWcKBv?6E^v-7|mb_FIN9R1+38TWV zS+?K7pg$)SPZBj$sI#^yd2tZL>lnZFs{VU!mCu@;Q)g2HSxBYLfUwRxK}U|ZPkNMs zYIe5vPewsH#^xjMeZ|N>=DG)RMkSy40SYt?h5 zA)%#nYm)1m593BJR>MYfX1pUL_W~F7VPAxJ@by>K12!N3*hM*Ar80?7%koCA@=hC8 z$n&?RO_q720aldU{#}T=q<%?USc)@Y10}ktnN@%r>$nm{rKEh z?WIZmVE=S@)J5(7m5=ff-?Q7e_EPpe;aG);rmFV4vm56@G-cQyYL>sPoLn`kwEV4^ z3V5k_IMR0$ycYU*KHgVAAMd1@!citi(9J7KJDFsu>k+UldD$>fHGZVySGcr?J+m6~BC z=>0t{PSoH#fmr$~qHkGS4>Jr^3(~G)8$JqG>ES3Jjt(J!X^M(lfNF zvN(%T>#u%O##Klm0^M)S`kcbiHW`aVT;ggSGJjt<+xmf>i&xv5GNRBaXh(Y1S7YNZ=Aw$fB+3^asK$~s=&p1m(${2QA1h&9{gqUCc%S~QIXx{I-IC_=V_nUdHR1oNV{@Kx0h>ToBJQDc^ z7XqsFi1mw4rd=Px1Rs;6eraX7)TmF>XiOqv%7#Gdd5|!F6*g}*94MhX?aC;JM3HIL zikREeSzX8a@3yn|&=akTYa|IpMx8!bW#nwwNqeK>UB|l|;`@y}mg|kX>&=zFk1~7> zlxn}i51JWkCb~gnOJVO*k~$rsNw4@lZ{^^V9-Tx{>y^pRHqgGa5O#4iV20)os`Ba* zQ46T@owbe+6n(W*Z!m4#6pc(0UKLTQb^_*FUz6R7w&ln%9~{yfPB6USsLU^A;89Xk zjFM{$E2Yxo<}j((O!Z#X`SbR3Y9ft!M-zA~n0Bivk1gX}t@&Gz{vS8=%(KiIM8Ff2 zYZL{Rz6NZ|Rq1VYl(5l)HG$yaZUOfxtuI&jXZ^dpOU*Iui%%H!i$^wChMU)Sn`j_k zU<$Bajkw-|0z*L|)9XK`iOQbS$dE~#j`5dzp>(@oXLwBy&}`Bd ztoqNsQzop^sMcghZw*y{Tzl{40H3`FJ@4qy3e}T;hhy6s-kATHm``77{zS3W+|r`L zaa?-y^WxQ6ws(7K5_xM4j)>}f>n6a6fs%cKlX9HOEuD&}k#d`GReO!>Z0!pI>SEh! z6hy+oGYFbEHgkW!xSxrWFcbEC+M^y#NNT+1dwcG7T@cF-bUPh$Am3*GR+R{mUGc)loR;(YXY;OrB_{V`7qR#w(l_EW%2co-TI!tXw4 znZ8=m=aUV{Zv)jWXq3jC^@_kVE(tT(!OB`~2rq3uZX6(J=zXKbk7Fuu4L1xc76NX{ ze4Y1rK;07xnAI!1r9RAJSITU4^4|D$r(S3qN_}ke^`09ic$kzzFM@Kla)bUA8SHV8Kx!8tQ*A<5J2+KtT{UZ$9LOThu) zC}pZ{sG+wfLDS$n6K-nb)CsHGTR6kf10K5@B?mj%W}Tz67b~@fJ)pe@B8y)wIl&zN z5=E0b!j`5;SjAArwWuHLF{xerB9|IdcPc*nqd(^}eT5O#xuTNdv^;J)Jq~g(KoEeJV9USX=uj^3T>zuj+DI#|G)2T9T-vB5MEv z8kH+SG_xAlT>Se$ArJt4@qp(x8j7INS|8F<<>@8-yJq$ z7ZEXdsd5pU22Zjyn5}o9tafV@fQ$pNgN2+9EiCD(dS+^F>$LvSF~=_-TGou6-ko?_ z!(7EQk4=06)^e73VH${?9$Y?;bLlPHy+I+K!I!sFlc{*J3}e}JLNS;N48GgZp!05; zX&{yfsf5jG+Qe@6G%AnMkA1eSsT=K|UTo(|VU(0KJ3k!c-tCP7*9e>M5+JYWifV6v zd!bNzHVfyJBRXK~FbB1}PKxtoJ-Y#_-ON$Pg|xivT2A)L-51#r9t5`{55q=vfITfM zsYH}F^BRal)VCo2C1UG;>zcETw<8ijJio|04-9L`tQy2L`6!!{J31iPk0=UMIS()>9$hBc zT;_$LzqYiI2N)KQVy$g7Q706bmZ>s<4mFFFIJFFbxR+D?g~IvCDj6nvu*hirnY@wB zz^AvoBhyasQ$ddrY;H?^0dJcDHCBek@hMbPsbJbhm*qFckPV)4?194DxX&&ws4 zI}@{0Z>g(d{$8G3d%?=Xe!X6GV}fsYM(jR4{g*NOy|@BdC$8!FoH{MTvMj!ZgMcN+ zJZ<@XZ5)|<7mdoa`_=KL(V1O(32ofH%Ac|LO-9X&ayU=|FBo{bwC6KG=g2z@V;puM zwEOY!i|0;={%bM&r6z~kz_d0lHi@ZEA?uUb+?mSJgr9<@LGt@!aOK}CK00C+Vn9{U z*r+!Y+D|(AW~MsDQVbO_kD83D;ZWw(x9qlzUfLX2!xa{5H4rs!VY~&W&nw;@5(@U9 zrVjwg5#u45%vfDrJ@dg|?d)AEV2V-BTY2eS#YgVzYZUb)u(InF?KqNd8x$Do^dpi^ z+w`C^=2`3~i$WLCYV|)v(`RP>h(y$SY`v|G>IVLBa)m}0HEA-y1TNqs-1eQ6L?GnH->ciZ3z$%IcTr_MW74m;s z8J2)Z90=g{Qy_e1w2NYl_787$d0L7N3D-g8s`mq-I@}2h{;yco8KLJ1d8TcLnd{x{ z#rdO%Aebvix(9@Ma0D1oSb;f!ohvP3h+|$q_-bEeZP=Z*{Z-iUu-oy<%VcVTCtL3*9(h_7L7gdeS*C&e?8=Tyqcxnm9OQbz14X2g^F7Ju?Rf{p?5GAT5l3oYU<0ef z#SNv`1UU^VrT5vH&|_w+ltSd+6_a#17yH6DGL@EP!0#x69A=stIBREoC#7ssM}{;> zb6>6}Qyn?<7$2RtrD3%`*c*9xa>!=e^-dtyJEU7pY?BovJH5ad%Wv(1l2itF*qmnJblp^EF2e$65`L{V_{e`*pVMd=11q521ev z5K7U#mXgORv+$UOXg|CdYXH1IftevL1r-N6mEkcZXJ_Y!xz?Nm^D(P>0}kVr(<}5N z<<6EiW9g}$;xwpNVB-PIQ6Om0>ZwYZsJU~a6c&bvxQou!NHoM=jk1$PC9u|`eOH5`X}J+$y-2{e0s zs(SL81?m%VN#UA`P98+XO^w9kF|M8MA;@jJ{e9BlJjU7%Mpe$Yd?1c2>pcR5n}-k0 zR$PT||DN+Q@n10Vup+5rG`*ng+aSKoDah2BWD;x{q>e#mQ&E)!digv?slC4taYzvZ(dH8~G4MZ%3fRoh$|!=x(vSU}{l zeya$w`O{@Nn2iPh-WVj7qPvH44SBD0oXu=?%P)bGw&%D`8kgePM(lt>m{aSEe@kjT zdJeqHgYqXhh4g?@0Zd#0gL)U8>VwUp9b)ol8zhSpL1Dj-76V~qAs}y9UF}9yI9Y#o zvWOs(eA}?Z$%0l64h|luwe}D}^eSj(Dk|nomi?RLr6Xp&5{?bQ!NIxIX;Eor=AuKN zOjQgk<@w_iog*3n+XcKcuviL2WTW1esPP?a=*6-}PE^<_08%c)F=b+)(v3aG;A4$9 zn~KW6cvTpfN!V1^k@mTiK@WeQSiQ`}G&*Bm;cp8T`PaAh_fG;SKQ$#fMDV-(%^y$s z-(bHOUjL)nuMGM>*7zT5{6@C_k2U_s8vnm%jY+IBWOcTPsg3F{%z$y_{Y`n1-H^WI z*Jb|d4KDf>&@m-`76{Y!HUiHMP*kieEIdV-eu$cZqdvkV z7cYozp&+0>0g)!?vXLOI=b6sHPNX|ze7d<+ z-qEzMI$tIni0AUfTY{Iqx1|BFluARcL*$_9+d);4e$A!0m!;)K29>NRbd0-GkI|1R z<)rh<#b-2X$;$6|S~Xk$M$oBWu5hVh3H2{VJ}*Es(&u44jBPPeHt!#K@m@oAa><(O zmg?9XG6fiP_~k$TlT8A40gA`YhJt>b6*(|qf{iWL16dqJe8$c$X>(n7-YaEM1&Xyo z*O7_`mYjb*D#i?0Ss-W-@XN{~FIQGs3;LdyP2G@1!!)iG8zLhc8k?LVUdjhs7K=Y6wc*&#o${D#n#V!ZK~zQa_m*jf>pQbd|igT!%GEBRh774Y5_8UR0fZ3tI=WbJxVA@JM!z=IL1awlOXI zAlBz>nkeoJcUDxYLM_*7ebR|lO=Hwlhy0Gv(d_e@`Y+P$ol{!i>&VW5%KB0DL~pHG z+{17m^-DmWjQd|%7HA^*PpOhF4j>>G`=1#7YyepCi83HNS}BsdIIE;u>wWp2AQ|G$ zF7rsPpgA5d%O$wBe1$@R*pKJ+Gdu-0x2u=lbcObHm0I=(63*5iAK)IX#?OL#XAoZm zc{^R7xlgb%ckUp(-vxwUv^I1fDb1X$-0sk|NV;&fif_)?QfTY=s?ckoTk&*8d6U95 zI#Zc@h1d>Ahjo>buw#v^));#owd}T;?JY)B?{T7VP2sVw;u$FhOBWLo)K^uuy&30I zWaNwjBTeNz;-Nk3?EFJVWZWi>*ICd3ZY%ER zUxHioi(ltq9XQL$7F_(#f9}&5Ibh?88`_8bYydE~p#U(iMgKXpdXcmbWGf32{!(gI zD_FqjJIFNNaN(ah015J&~2u;DcU)WYJbQ-|rWX^RRF*~wit zff`i6!obl+j&-vC7r+jle`S|U1PGk--1M$_1dMYmfKf1xfl}=@$y~M}Irp>ZF$vmZgN?R|(4oNZ5}Az2JBQI(>89krcP5g0L0y zd6JMBU%%5M53jA}2t)|@oJwrhG)JfGRW)KWqiK(}I(gPE|2xgBN@vrApQxk%iv592WWpl_h&KLzY<^ znxOSMt&g$NkNIzWG!_^y>@(8-tXApc|~KpPtdEe5gMFir&|^s0kV= zoW_NZ9j0!7HU%xt?e+4Jy7}`-$p3iX#y(i|AiqF5y#h4|1`-aoN+=+GfFiD?9&Qm< zjia!%TYtek>=O2CUZe-A!|A^El!(x{b+-I3P&d83pdrQQt2*e6Y0aWu3;wOglSie0~r# zzT2PQdx<@fX#JidmP_M)#5gjW(}*;RFi(Sb8?h82j>5sQ2lenD(`(=m-0VvPKy-9A z6^A->I8P_KgkkIEw0B>jw_;j4KscMpF<&P`uH?T_x9L%U^oqr{^7uu1eJ}u|*E^$l zyE2*d&zI1e#lN97svn;R+VDdt=ADjG^@40IWv3Vf?6k?Ijbslw2GVAv@@3I$d_L_P zniXqbO!<53g;nB;j2FG{3iyq#x@c<1@1I3TcHz>(AqGWlAK7doG~&oj*qD} z*zUIILE3B3l%e2RIFNY0qW}Urup*rvVuSWfJKc-qnqBSg#Wb!(TA_scrqk0I~``abfwDtSSIm zCk$gyB}M>Qkqza*WLw!3lRSDO4@u8JWOfPzwhk4^9Lx(8Z8AcmFK1xF0P zTdclU#IEp*ZK-rUF8kZZ7z64GAXxW;egVqT83N|YhYD~tS1ooRE`-L*^^`?H6ZGLU zdvq42cb;6^sOhu%a#?XJb(#}AgEJkX*x>He4cG;`XJxPr(T<*;gF->clVsPrWT>o} zrHDRJIp(M-Yl>*4P}O12so*lV*WNTyT}W&8|0~jk=B~v3bdQm#qmYGRAqY&($!E>f zRqwE2D8Vdq?IH$#mL&L=-ebysysdV-{dR9B6N&DzYJI@CIC`xE3pXD#cB?Y8TgYxe z9{@YEI`CVb0WkWZA6C&rb(_q3!T<|){5aX--z+dMau@giaw!)tcfQEZi3z%``RXn451_9>(!Lfk$t|_FiPkk5(57 zUK?RY3h|trMO5#$ie2nXKPE&C^?OcdWEMZA*N+eY7BZrZA(|a_Ufioh~; zk!r&gC=h{;bJIiDKB}Sd`V3ew+6SovZN&^cWq_b}dm{bQ4P3uN$5W;iZAt$(H_;J; zfxV_Nl5QNnaRtDD{DobC6`ETLqu2dn@HDu1DrY~M2K=bQlLm}XD%XoSJT=hbBN0@& zR_L&dTv)VngL9?qjCb~fPiEZa9@OnB$@|HfQve)HceYuKmTlB*xVJRI!(uHDK-xnSsgtpDs8kJt*Y>-Xje5JOUxV@3XG5uLJD(Y^gN z!0Dg|oR4nJbznyL_M#rzpz^{Er%RS!SO72|_YqykgMT4bh(u=f_|5^ebKOC!SpD$Z z5l*9a>Z&Po);%uqqx{^98uZNLdH&>p#CqmU+KW#>@{k90hnB3UHBU?-Z2F+WAj?`+ zyfEJQSr2&}U+pHVulipS<7MgJYyAHvF>oNqW6hiODioQ5H-x*(SDiv=aJxKe=5|9e zPBi4?i+7qdH8n%^eoz;hKEF#wX1$)(CN*gJ{z%n28CIG+vwZ&}%a|R`Us&d04>C&$ zeIgD3z9w_bF}Ba`ZRseoLz(Rd;Mi&Eb#?(TzP&KChV} z5kF{i@^)|i({Z(+2+JYQr=ld_r1;{z9ujTH1KTvgCxSPv`bH z+PEDoMy$;`^k&c0Cm@&VGistZ04(o-9pBk)Du+|F&C$kw0wJK5gcF`eR9Bfz3&P5M}nL-CmsiS zmsUu%C}ros`_k*+oLt2LXgPU69hH(oQmkTtZ#$gJ3K%EssHIl-i_I}s%VFPnI6*xw zYvE|J$zhB{Jas@~%3;M2L#<~`&DNsUh0f;i)xNZY+%xl&#mELTm5cMSPm^T*gYwGc{O|1k485^b2NY+|V4| zOhuAvbgSL{f8a~L!9N#Yh1BqUY1J8A1d4ySQ7u3{7Uw!G%a6T%iz9j4ha~4Bqr5vEL;^UZTvn>rP>jDxd*< zggYOXM9;M$JiL6w=Q^<$WkHFB&44pw1(Xtak+Li~ z?xi+omqLT_{(XJp(bk*B6#PIY_dbFf&2`DN(B1oob);0hp;E(I!T zn#iiTwLAC6H}>0ZTa`1QTA*l2EwU}u@8O7(ypV_8WKHap;I=4qTD@5i<-%(OGNzJtEAW+WO0 zX@6bPDr`3+eP*i!8cDjilV24@hQnQ743GTe$r*w*4s-U}a}mDt2(uxQ=*-&76W_Y+xNCs$t8Cb*$=9_ZGV&cV{!p zHw8yF=V7rNK|8k8&$&@>a~+E$aG_p>=nu(P7WwzG4{oP-pGr;Dx2J_TI=%Nl-U-Et z{9NmAjhx=No#9@x8+sEcK=gS-ynT~}i-*$v(4kcqC}L*|onR4Nm%l`gPzptT{w$ov zfSooxALDT05oE6cYH%eg0SfVyny;e2&hKE)zK_wQepX*Xc|%@- zQB@AHWAsqIhRh`f-i+O#-%_-CR+`#yX;Cpnv?8OHCt3i1$=9}xipt)4N2vZqDVoj>|Gv|5L{ypkhfZ^s z!2E2Mp)08`-bM8-z=Dgsb3bKpeA~yVu4T0DZjNg3aBLLCYhsW}dfA5HvtI0N>0Ei~ zXO>#hm!x+PDsCA|a%1DXMrL;6sbDW)97}q5*p27WJ}ahHst2OWrHr1vsedC! z%DD7?MGA}auxN~rA&KK!ABWTEdQ_EV!QPc}paNAeEkKgY;i}v5GjU5*{7FBbRNhI2 z6p;B12}_`&bw-m4v2HhWZVRwRxS$%WW>s&-wD`tE^!(y05#Bu(7 zPo{vwJ(0Re^B&qEJO0B@#06zvhlQJrnQ@My5bJ~;vkmL! zg$iO`)HP60ic~WcF+u5s5Ust?g_a z9Fh6(!sx*Kuq8p{3sor*VKDp8{{_&%$A8qPH*_7}8*9zI-IP}uuM}-Q@={CMRuWy) zYF8H$RW+Z~G-~sX5&1&XUl2dD60|(m``DXG#_djOiMqn12TdyP35SexhQ2TG%F@mr z&f)T;m#>UkTwd;iawC|!^a|PZz+!IVkANzP$JGZGmVXPzfOZl(9(G%?>E|5PgEwLk zUnv=ny#3!KLvtq~(YD zPbR2unOz{8Hiq)Z%Sh2=ph9x36`ZGK0P8M=o>uFVBevJO_(^FF-=xM`=YMedKFO_{ z2IMKWr=iIre1N9B_c7%FDEi&<{SkaB;PfzuL{(Wi29Tazmo+nNI+}{psRaMw+Yd58 zO!O1=JSv$km=iq)ei^-)6Vq_Wz4c*{zhu$(mf3PfQ)%6e4r9;6PuQ)c1o%Dp_dSx? zQez&>KRP-&*C;ZgqT0_$UC4S2>c`Vt&Qgn=D%iR@9Vcy#^OWD(cSAxH&NacgcDm!) znk-bgcSGFL-S*Xze#!CW=36?S6XU#1biUlMRv&0bV}j?z^9vvwqBB0krxj$XnE z(H*pE?AIlaBvfhKjhH~JylyZgnQZ&3Ui?DytQIqy54Gjy5_wdu4XlJ&b@PGAKDz)q zropb#s))vDy)yrqtiK?%b-3f>?(&D|ngiM-Yowo6c0@b>Qh70=En?9+v=KSu9sKbZ zkMYv4_zVc5{7)uP6CgnP5(&lO5$4C_GgjT1eB!J#sW?C%-HjHpqlv3anN8|K3UjAV zDpnQABQ0feeaD&$;($i8Z2#+Jl1oLG053S^DN z?cN`XS!gD#AK(@`V?@7^oSqD3)xNS*;jQ!}uYYN^G;S&%c0?lSUj|0LJyq^VhNo6@ z^S{3dY}k4GWrH8kepQv_h!*~(x{_(@Z{%f0L*w%;yZo}|OY4Z1{$Gs9OQU)UFtBxp z2txfU-RT#`p}h0XKh_VBLCF&<^2cs zrB8hWta?B)-DG!aLG0^*xL6l*JAQpyy>dzaW%9C7{yTE>G;$~|Oqhqh4H_j~zW4i0z zSKr}w8$Nq~XJLas{>=K8CU@)$CS0@T1d$74fdqrN3kZi8i&T=xYgFI+BWORpeh^_7 z{}OTj^Poj?DDh~3Oxqu{Xe4N%{|7CMfh@7^(EYfEMZ~;ahk^GqpM3MnH<2Mp(Baix zbkM?_d=-UNY2KEeKX(Na-sOFMq!|fMAH!mLP$v(lHJ(vx&d>5V_T)=^{5|*pPf$+! zUOhK1zg&QwP!SaBYd^pf>Ta6}fAWNfnX}bDc;XN+^+_rocYfc>qk#;EWtL)_v&tbg zovf4}RQ67v#w|F)DpuruUP#lKqIs0c_j%YwFqGsG%4nY2o_xcly0Rfka_Ao$DZLxm zpC>m?tDjbVDr=i3hmMQ}f7}ZddA~NrYk5S@zysZ^pKJfeZrucgIlFDhLXXp@)4ckU zn5r`^fmh>NjO9^scx^oFQ};dQyP|3rQ*KpS$4*}XHgNpcrRUwv5YDyiR-Q=xwOSx@Fidz!SiMl>btU$0RIuu}gY_TDlq%J%CQRs1L>5_Bndr_n+InpMAW~d+g)=@NPdT*UUB7d9FU!`mJ?|cZ7}E z`MW4<-jv<1@`_x3iXs_(lGNZ9KDICK;L&-INoRY=@RQmNgTlH5T0tC&?`yWzfDiku zy_+m;<(b(qDEQ?rTgx~j2`zTkFQ`&_4i*4 zX?Ox3al0~`pXEde0nq4WE-=LYaWvUGH~n(^do`H&KWPH1#Zs0smE-t6Z8xHQMQER5*uvPtUkah*aQPzj?!>o&Y8^D0r89){5xJ?R8&j&Yg6XO^ODqjE6brYP< z`3O!Xqp;bz=eZ!$fb}-t*-mHM zy`=DFV(Cr!f9?TE|1A_$FMv?!TB6I}UX5oYP;a~oIIUeAeVMr5vAyce2B~@tjJ6ay z#?osTy};QX^N2n4>kkZV%0~a3pM#o6mM(HH0&dmrGDVx+!N27gs2s`w!DzGM4UguH z77-Gh9$}$v-%WfiLb3w8)E(y1^a$5_(A=eIIsDSFoxu*!nTx}$O@9Azwm<*(8K4yA zUM0sk`8EuS^s=;03aWz`QNK2>`#e&0Z|ZWtYFx(W=%q9&%a8D@+$lHKY13skQ`mFM zx#)?LRNur4CC9udoHAfV-w|Gu=i;LE4(je@0 z920|O=I2;`2c3ysNRIO`H&CDylCMJT+mpM({6z0Og$zY^oi^0>6P{T)Oe#AWEc2Zm)4_n~z+v5{y( zO*;Sm-&9r*sH{)PtJTQ&Lq}|WxvqWngJ{=s6`@XhJ_I9t^q7pJHrDxuuo(L_z48!R z9=KXfVv%QE>lW&m5x&4Kd+Z%~dg<=@z^xJ=9|qm}4Kl^g4qQoi(E(Y59V07n@9DRiS+q9$bB`_*^WNN|TWP9X$9ExE ze3xkSzJ#jP(%oojIXkB!Efq zPfA$?zkPQFC|y~E{lgPHO$DWQ7Ppab=HDRScXfenav>WyL5^!%;C4>0i>LlEU_UkOzagx3Kbd#E04SUK;W`}m#N`AqddWd1YWjB<1)kO}pD(I&;s>6` zg3Tgty{;9?MTJ;|!B+mK& zQ%J}YvIRsf^)Zk{nOot=6PN4ovBpbPfd1bMdpQckV{CxB@9*pqR?0{K6X98$Sr>l$ zt|8Ec;BI^U6Ou#?0Dml-qoK?v9)%V5yE?#ly`U*Oahb{lNSs000{r=j`#}{h1FXOx z2M*Ubak&k9d=O?YllZsdypsb`*J8z|W!VuzRc3F`@z*VP#i<-t&C{*wE%@-)b-I7k z7VH{+DHpF#FfxSddY^_W(NBf+zI%;|BOp&=rpmyml%({8+F+lX6$8v-<5;Q%yC<$+ z&fJW_FAaQN1C(zmzC~HImRXbk_aSh90d$8pwTVT1-b)7jZD&6e1LFNmuDPpU%I7l` z`}fYyhk$3fjOjS4dE(kVb$Id@HZB$T0G8som^J^fe%-#`!uk&e;3tV30&;5huG8kk zMFjkEAay(7a)B+iEZ%;B*>8|wK<>XvK1c`1(RpPG*0R5GpKZX_K2_r=Io#6WESX67 ztz(D^3(@;-=iCq3zuxe)L)7>eE#A3d8De2j%PgS=RXT~^!AV*KzS#x%$Btx?- z6xh!GvVw!yo~&9w7NVJ|)HkpG!O{F8i1%HPJPpnF0 zl1OZ+WB3<={*ix>an&cLa8^%@q!iWezpeU<7T#GPn-`ErKleufz5fW{9|8RTqXi_I zpOX&Vi_>Y0NT5$jQ)O%!Wd_;dC?5~mzN6ZyUH_I%v;FI|puLqhEfcxeZ%J!7DJ+nu z!eBn}fR%cbR*e~rEJU^rKleMctkLju$4_yf&a=1TRLxI5Rzaf&u4%YN82wp83d#?W$&Fe+QNqKUFuNl16* zj{O6j{L?fy(YaIk;`+o5Y_$WN;a^{X^4dq~&{Dkc@-4|$h@cYsO85;Uj~OEIFCkQV zo;#%v4u)I!f|DF72GjiMrT<>>0b$9ZI29TJ{4ESoWdwXY&FYe9ZO= zZu+hEk$Ov!yeq|&iMv6?#nAhh2s1Nrla(NXb?sdo zwYwYPLlzZGJA3&>Y;$Xg48yq;*LKwX@|mfhGxm7l#%^!JHrruck#{F+bp%~k`m=jJ ziM}EzH#By^by^;}6OBU1s1;}%W4+6Rl_yECTH)!-q})!`@QdED(%;)ilQ`9R10am! zu-#Ny_GnhEvN+>o?i`xTz}uX11@J`~%+u@e+uh!f|J2c}3XcsUuXJoWb+EZso!1~+ zm?WZk8N)<<%@L<)Z?Y`Qvf+Fc6=3VxwGzWu8tHb((+iHw&z+QVBC{AMUcg^zw1!Vi;D=-qqi-^YY{M=}(&ED3D9R|(=Q>_g^#~}9&U2oI`CM3|y#btO=;#Z?EZRoz?>+IlGm}YK z8~OS*u0y<3f?r>q9HKxF)CUP&zsH~7cCK3{`jSQkg$-ZA&qntx-74ga-g0tR*<-SW z(Qbl$R2j@js~F;@iczUyMv2MuNg(R=nB>Q5P;pIbID$IAkvg{)kgH!e?1@&b*YKbW z$UsMdnF{uqq>HE!myfhwU2EYr_Np&S7#K^}eF+r5Uw!&d@lRY?H0}6&>X%Q@HL_<< z7ywAwS=)l9^U%MY35?S`RN2#?b1MU3kR&>5h<&L}0a^XB^t$>kTmP z*ce~ynfGGG{%*VTOqgGYv_w)?}89G4}HLcY`Da)dk zEOt;Mwr0SAQFra#X8=jd?->Qhqo4Po7_Mj!Qv$6Ch7CsrCYCm2X#1OJzXPHM=ZJT4 zA$^phHl#0n-8W{{-|!=1oEy)Z+!-AiSR#&@o1l-txK zIIiLqdtp4ctbnLR-MVM-PB)VQC|eT|cO@#|IG(9iWlguaWjWc4GvIu?od2h;^}shZ z_-)$W>5ETGOp_dH;pph(u5jLbt$1WD_YF$uBi#h9{=174`MK5TxfhfD*ZC;3y&?9^ z*;CTe+|^qs^he!te*UZ)2DrJBV#()z_<{as!Q=CH<_a!Fhzkj> zw4u`6kNYP|n!HMOiSB9M;Kmujv)`|hhpzH8c1<$G6{oYS>FOQWXD}M*>Vdx))DHCI z}r zhNrvExG$0k+8O1Vtj4+bYJb$aJU;=t z3$d^JfTo)XJPxip(7a&MA)CAlt0KSe{K<2Dqq!BLP!E@lW>cQhz7^~P8AeUaF_tVbaP;GH zgvM4@44*=dT7Qzc7OZ*>^O?0-R(EuWbfvraO{a&gAzoM9PTcX}kb2*YY~J}H+EC=A zHScWqw&C}+8@=3Vg2K9bm0PB6+tsylU7?+OKfLZ)Puz`p3MICN7QWWp-kqb)HIdq{ z-oQ^6`_8LeBDP}e$V4LaP1~dy-B{G=+awywUc0A!MPyNf@+cx?BHGD%^k8d;VGOAO zkA~S(YM~Zr3TuxZq3PSf_|g8welE*_OfLPknyFo>dY0fXS)}C}bG5ej9T$Hzi`Tbj za@hEwV(b{DZ+;~;O=MNctewplnVm_Q7>>s$6J$Qv#nilyigJ}-tzEhzu>MlCHIA%C z!LXhMF`lISRnSS|Lx2{qbDF~&Z&&=p=~ZW9#>k#wNm?t*g9||(;=S^IS5IqUMWkaO zA97?vi0qc^pTMvzyggf>3gO80l3o>HkSa-~O|fx-MbM7z%q;FGF(x{Oa+Avs zP0SG8I`pT95|H8WUYyx z2&@7Fqu$T2hUXY}NjIiKBV29E1kc z@-^E`VZsVBM?GoL#(i(1U95S>OKRM=>!H7PdTGvd@Qz!z<%sA3WXo4a4RZX)#|~YuinkAUxRf%#XCy`w(Hv zVN*%j5{pC;Jum99Sr>fg+?q><3(H>;<6|2;ujNe+Y`gx>R0{#NW8v;_#zec#gPvGE zR;8_jo^VEo=`>7#2|Pt#ZH3hN?XW+i@`l)o5YsIh_K6MGO#Ok+T2FTZRP?oCs_O=Z z9Ab*)rm0$H_%d2r9FFI%mzi}xnrlC%Wt_JxRE%=yB%3WAj=|CDVK6j?o#%*|WXbE; zO?&YpE~socDF&qh>&`$o_}+F(YeyfD*&ZU+^X-+A*}8m&dSaGe2Hg>2Y-iKG>jqZG-P?P0wHNWkpq-O)!hLZdGO@e-L#Sf@XpWQm+ zqv%ln;z!Txi_f_y3{Oun3^|S~to%T-M}E(|Ec=zKtYVeaWKdT~?QH@Fzd1BQJ%=Sm zz%+)aA@NyM^(o?2t|739vh$h;3wJt+NSWmoWIqq_>=NCexWo*~5)iBo=0l#jZcY+@ zTZ@>aO%*+hQ(endXV?oa?Ciup=GnV^`alnH`u&szFxUFxXbOi#J@OiO&TJt+isKML z`~HsK_!pLeH2V$QZ0!Tp@LiL_9{%&2t59yEA#<-C{fiea7_E}PBB;TfGRT^4jZxc- zdU%75W(+{V+FQU3R@dVg2KaSEV?8ttg0$e z997e0@t|=p5ht+6x6KFQdnWzEJYKEKaKfEVp?0-BVCg>VD2A<@h#3aZ%qrhM{B$7S z*KWDYoKsKt1efJT%j_C2QaO9|`QG7Uoi6`*>b)7SeITW# z-A*XkESHP6Y{iRb7?b&UQ(k0!ctyAa9n;g>=uK2EJu>#v9eoXqEhM=!$}ps`#u>RB zODVH7P{fT(wuQiWFKQl zkv?jC1I<%?(WS)d;)l=B}hYUlKfOB-(0{ z?av`St}eq5dPT*omKUnybO=*KFQ$D>FVe^UM&#V5p=hx!z36wIBS38`9L5czVk zOPCd$?{?-(>Zoh@vruwp4 zwiip^R156)KbALcI;3oA{bGRng2h&LnRu|Q()pDN^ANYw7E1X$yVu{j>be{$i(!ro zz(i2t-CL+Za$sALOVWX&f5n;*yb8X;|uYYdpabt26M{KDL4J=QCXiLMW=H zNn3|j*Y;&7h<~fDRZVOCLA3y)2RCf=Fjqd>3vcgyT5D4bYnt-pTqbO5OYn)E^BoWk zMI;F*XVmPJNqQ}FO11{^X%Cv+8pzjT@Y?I8P*zgW=u1i*_#wg4UU}-( zK7^sm%=g6)uBnSgh7*R~pBTm$=99a^XhO1&i-LRY543D$%58MW#^<86Y@?y8&^oR1 zUIoFnhD-xfdA<>i`bUv?+|8wCqfolO)N+P|<-DS^>NL;&jcvN+mc%Uelydp(Xg)xw zq7kD!^Br%8_75B0vcy3zRoAw61YHi>I4HgoP)~6NU6`?xsjoJ@aBSh?%OJj~=^{(L zgoxa$EK?hs(yc}hxNsUtf>;0)ClI?b}7Nr~y#fZwH-gZIz8du24+Hgsxgz~i9h%T@VM zg~a;@<}T(vNc(QOXz8abqOkZwLn|=uMb;Y?f7atvAbiF%8Hp6i4{)C~)-0UiJ zHDc5zWW0)HMtfl`hx6m?`v{9JZ9SMt&?OPW*X;@W+0Li(HA~+w*eXVrS&ziHJ|rWJ z9Poi;eF`A77O96b8Scz-BvPt9ZAWy%qgM`#Xs*6V|1g*yK54_Qmuk^Sn7#Mi#7h?} zxi(e8(mGe+fm;hz zPUCV{1b#eR7^sPN?kZN-gX*M)WtG>;Mt5#*26Rau?yjJBqIu}DFQsRmIU9){kBmexe=qjI zrTTMjRi-g;)3|=&)G6I0PR&~FNBL!8Pbzd)a&mT*cX!~OA-M*o*9Qu|-b;rUX355x zUH>FTrH&X~Zb_Q>_-*Y*4ZQc7Kq32ONDmoJA$tv+o6#Per=l-kFxg=Nr>B*1B#Y|M zdXJ}Lyo8$9a)QU)*j78dJ$To5PBhevhsg#H9W1!EdC@Rv_9F;BHfL^xw3Uro^$uA? zky3#@o|cxngM0r4qdsZRdsNIFKQr3E{fopc5YuFaNe`3{D8KeqQeUm-jA>lz)<=T-tdn8fXIy|X0S)lW7@mbIa+iE&Pyq^6Ow6wz&Zx>nD< zU~%*#$K}ti&8#?Q1PI&17GSU7Y)N zE&6biBMk>1n((tYV9Y^SqVw?6gt$`F-ellHPm!VT*%b}_;8Say4))y|v`d49(VT|O z$fbzPmbaJ5rs7_#j!8?PR~`uL#(bm`LF~$*DkR!WR7u)%^0)ZL6+e$FI23ehKpB6; z@jM=Ir;ZGumogIDq&AZW6CcV`3e74xH}2?vd|2R7J{%(^-`wy&ABBxQ^R^LOUDxqo zMsGXW-);Gqhdv3=?eu7GqnpxJV9+f48m~|mTjt1n*$2ErMsu!x|4|g17QGPfq;}cz zxa|EG`q%VEEyG;^w#HkV=-p##F53zB{mtCMH1Po9o+@?rg?`J6_`YA%!iInYl$N?X zru&HX&?VX+P>m(RY~hig8cQL>x)H;>LzFLd{+#zqE^qC98a_%Yt)8d_!&uj$w znlxdKdQvaOwN2f0Qu9^W%8jql68N=z77s8wKEf+@8oxiBVym|>4D)SEn6db}=C$s9 zbhJe_)DjUjhxd3yF^2(%;a9+y`=&i+E?o1PT0bHjGc$33YPLqB5h3)EhsH1gkM ztH3uYVCnkAOp9!;`G@z>jjIH0mW*MSw+zsYJ;@2t8hw4ykT$#F9(B3>x?ylMi+fC^ zUOO6x8^F*ir9YeVl*@6EC+vyT6~P&OFafva$NtdG)4~Xs%SN&bQ(*Z;Y_-3<=^(~D*`#jJFh|n)d~=+7GJ2_SDr_)W1g-a z&5v(RG9*XkUD-=LmcNOpaqH{*M{Tzwd>}$!j+~&HYQ!jYS<(+&HKc1IW9IE+EuI9e(rqhK1i#28@D*DH`c{f@b z*Lau0vzA{{cy4o&dhIxR_7S$bU!Yee{I8WHr`7LE^wv1DC zOvS&<@dA6wNbfF4yhsjo+7VkbxvuQKe^QZ%5I5>SEJ}?U3>4c*PH<&&QC&BYkcGsg3YUTcJ3quR)9s~H1`*>?wn zYAi|;?0cSgCHZ~Ba|yvqer0rhW)UjqP@ktCAe+FA+}-NEm9Nbh7VexN>s09EZ-%C| zjYvFlBnXn)=bM)`@>p@!UXc+}i97Und+$eh8j5SbQ`M1lX;Cm-lEOo~NzWl4Q;vY* z61H#8cE5hNf5dg?BAn<7ff(@^{d$T2f9GaJ{{iB1R=YOfv(Zbx{DxFna@PEHIa?=3sn z4-*+S)F5IHD=gZHxRz%kj&|g5nb+GJOlK>}Y~J_h4mRLD0@Zk$ud&U^F3L@Tl&SSoNkdAB-8lbAUIgp!ai#Y@66-SJ=DxQc4T66yx6g zC(BsyQKQ(%PusQMC3bgU^fC@8&Axn%M|P)feo%s{J53rWiUom0Qd%$&CPxmggfxA% z%RiDt)%s9MrbjfBS_BYu%6l)7jMKJRpJc5*&=QheD$wRtVj>xz{E(P|2dFe=Ff_?d zwxU$Ux0T2{>$ouD2s^utIU207KO}c{Lza&fmsf9$ATN)6qN>8VNWvQ4tsWc+-dWlG zBbYJfUR(Ii8=Dl@Wb~ELHy24<4~2i_`G(HgRqzpzbF80&!0pq^S33o5BE_xY8~&`T3+wWT+K&$>u55!rHtYR#OX^RQ!p<50FCoxkkSc zj=SeQQaVLdNnktg-<0Q&uFt#0i9k;%e_gte+=niGZqy;?RlE9Fv=B*)KnoX%KebHX z-m0Gp^y#xswa^5$r{Vb|_CiI^FEFQb^W&liO%&icks_;=nX0O)F1eH*dy~b7HdXfq z-Wx-%?QJwC&xC#n3Ex>7JcL?K)+SmNx6)4)I8Xxhk9){a>%u}zCq%``hIEH6CR`#AWq=k5WhE|8oqvywj3Z*!G}Xg z_~d@pkamf-jEJBFt@6ZOOHW@><*!@RUOK~LFi`rd6)CiS1V}{TMfz<_%sjT)r^lWfZ}2sXE8ut_XZpE{L9N5nY0_F9zA_L$qlX$} zSS^hoKZggqm8g&L+RQec@n#fAdl^s@sOM;gac_^9>?b?k*m_s{lAh%CN{a z>HVYJ8uK8G*$T5o3B3t!$xleg?wGkhCh$6q@nXr4@lCA~vpa^(fl=t)QT5}!OrfJM zAN1!`IEuo6ELqbMc5o2@BTggCt3U;eI8F=?i3&&VTjpg4*i7l_T8B8FJG-$%mx5H> zpgl9=*qn`*N23o}TFwmXUzO@~WK=7K7sWRf#NHTpPx;#Ju{+w&Zn`i`k^=f!>%sGH z!uA-U(*4}YsQo#jEYkK#n=0$sF<4jlnX@bGcQQY7tu9&urDOMK>+*c#Fx<|N?4NkO zXs!0lj5ib$xh^0hBU8;SV0o*0@96PD724?@>?m3M^n@X#q4T*q{D`7ggTvv_&&I20 zPKnbtK{Y=IO$Y7RjKf?kzs>GyZD?8|i6p*{stH18eK4uXz}y26vV8((9A11Zct+lp z2dy!7+7xvT3SLvK&}r{^Z8Tk$gwc-es$zZnc6%&2WwM^hd~?)~y`y5f2ov&iz`w;` z#X`%(M<@7M@fD5w1LeLK54=ukuF^+j-gLZDHX<0qnK!zl7W$ zIER_}h)UNPEdvz7M=9H_(oK(nGVVc3E<5500)Goapi$~p+8(!)eQ5|C21CGD2>C>`e(?{k-A1`-@QXzQSPvHJ1f%^miI zGK%!0CT$cMe~=3<4R{E=?;u~$ex|ScSzTfb-ASDe01WS5Tg&UE6oD!NBd8+e{#33w zl^Y>Q)+o5wpz~s1_tg-rRy~dbVv>&EY{H1Yz8^_I62zwrv@7WKNQ((!drr!tE^nsA zK2jN=7o_n6yfAUF+Jo>1owF`u$W1d3v)j1)-@GW>zD6kSv}|o{&BhlBE4E><4ygxO zxIJPuQ4?P+WG&OPo7Y}@3Jn45-^%$ubT`JlC4883bKls>(<(T)i#;9|pJ+I} zm1~~Jh{eaDywqqBk$mc<%L3UNFKm?;rpWaOTA8;;RM|WW~w7 z)3u;^B%U8G{kA&^YsoEZB+&ALh(`!Cloo0ifQtJRP-tlOdPR@ zlo}_Kvv%h=Vm0o0^fcXA{w`NwI_#>mjR~>f#d1%r-gzMI@_cl{xK>2i4Qn~%%vRid zK7hqa&1@mZ#)c?Jll96>EA~S|x(YMyW&FQ_PF3>lCvqP~q$L#A2tzsrY~K%>7+$cF zP7aLhpa%E_*e8?iw7;QJYXWRl+Xe#D+ux;=*FJ~^huYSn2P|Qncx7pdO>Oif--HHc z(nv*~0e%AX!sb%)Gu}ffKM7u1ckR5> zo7(j~o1RQDr5#peZ>W2#skD1Yo##r}%WEyyvGD*hjF1c(bBUiU{?64kbmq3Z zOaNlDR?8>$HgQ+|YT1Vm59d5N1sUzfp)otF{jwQyo^=0`%6}kI!+%4h-kVvCo`6Wz zKci?mzfQqTPvl7j?5_%hW9j(^!tLt1F1`OYTYv$d0q8OtDr2BYQQ?ePZZ!Pm0eK<-RpL;Dr>OW%@Ux(XQ4vZk-b~@wl%QEL6LD zECn<5yqmAp+O^6~7E*T_r)cCFXCxl%=TWWO(*%hSmIR=Xt@$NI!tL91`{Y+Pi6xn3)0VexEb=LF z7Zw#}osE8Y{gWIPY{DX&Uisx&9pon%tc!hGIh-3XF_JsmrtHitur%gBEsc$U?&iqq z9B{eq(Z%pmKF31Eq(KWo#<7DdUUrSu8_J~?Z?O>yfGzqHF&jmXA1|vQ?l0gCyDbU- z%js6H2}41{(zpt6DyJ`U#85uRxiyT|nD^K++c})8JF^>IwI<^}XDu!5cVssxgf?~k zKAf4Zhp0V%aHIa>I>kldb2~IGzif>zGC%Ekd=;8UtsSA6qt@i9`i(u0TdAUdC%n9s zvYM{@@>W9aY|DL1M|q27=0yT}P2A<{SY?D3XKhIVWsK#ENCo^&vzk$30sK{~jh{Bd z^7kp0zu#>#Wh;;yMFR831nHeY!*m3KDyhP*o@l5>JL#f!xT!1J#A2@WiqCTKXi~9( zxZk|{VLo5$cVF_jukP!i5i`?@CT>fT=oM{&wK3R@GGlQL04h0$#iH62wWE6+DQn!U z@7OAM-g~x1y5>d2a5Z2&H}lh{hr;nYMHSJqv(J8&=ap_?J;;CXQqp=Yvd?CFFfy}K zn7%aF1m+*eRMdYUQ*J>)$^9ehJQPuAgPDzs|*{Z5o>kj;vOyMzo6DUX3akO*t7jove_E^fTH(5 zkf~k>C+60Cap2m)*2LqxiWj_`^SuNl{M4j}qrcE<9?QB5iaTf?`uo4bCpw+`^I)Gt z%k<#p*#si`j=E5Amgis;xJHnyClETra+?~&g}&gN5o|0PvPFlOu0IY32_g^|CU}`Q zJpkDEOE3Ef`x7kUxKD9`1APG6nn%Hm92ZjMl+P_CBBfRkc)7=RJTv6M(O2AszU+_5 zZ=>mJSV(m2V^yC%bwzeMo!wx!KfC1$@>BY?t^#&6m4``(Gi9TOD7K3N>UXLv+RX%J z(?(zxgLx1dFuLgux2fkM`NLc_R^x+Fm2~TPUK5HZS0o#u?d>!3AjM%>4?sNP{$RySD|*B(-t{1poXa|mKt}*21s%D%cVS*M$(^J z(iN{Vhdn3DR2XO`rE1+8e9D?Ulc3`iEmG+CXp5@lhN3W?dNApHuk%V}>}~t6-aYwH zzO^Ld!rI-+$23Z|9?@)xv$G`05xXm+hbyrhk}vDs9qg6(!i$T!NMC$SIfj_5R(hX7 zg^fFpKlMcKsJk}0hM?EKgdEfVgOM`*d>b4d=4MiDXn%==gTs8%v$-U5wZnJ*P}e52 zeS7@U5s63Nol0|a24sf^=kJRY_$Ov-RXrMfKUbT|W}YfucjYH!j0Wl!;&a;cCqhaX zA=f6c+VFW^eRZZ%AmlxWrh3!$)w;L}X9p%~+8@2S%)X*tZf>;G+&OI8Db$sKTsHGE zqrml(_&7nz;XLLVt#{Lsg%aB?#*plyS~hMH-=qd>KLs$_m!%&0T$^LppfXMLNODJw zqwqBwh#9J)^?NQuVu`RPp=Sz~4DdaQUTqEsCIjWK9}&@{TQiyj5hPj^Go!O zuPz$4yoVy{h@!iXyG(@5=RslF(=EV0hcR4M+S~^X15S-C^U*r{mp08n{^gbjNIdXY zA}+E0a7~#h9ft3Q3*K8v;O=CKe!v!qs2=)uY;sS&1a13;s%3Odj>MT`ubI-zqRwl* z{#+X!i7u58bl2=tFy7pGZ75n~M&zo{GTw`+(0cohy`ni>N!nshv>Uc@v|8BJ-)iX7 z9fdT9c?GUT+xga_=G(0M`=zGeo9-FVKL^|M&ZIIw1>dDm(c#GLG&;^Wl&%;x@9ZbF zWXIEqm}!YqOtbIC^|O*X7JC)9n}M+v)9Zd=I3-pm;jMjW1LO3K3pU?byjQ%%R^`7H zuer8?Ani>zY`zkA#arVtIQk+#7|lGq3+&j#aG1+GcTeMB`FfVvl&y)10pE&+HwZ=DoKqW36X#Di_G zPi%R{njiMfB)x`4eQ$KEkrPWuEm8zc=4U{?e_k(J%OJmIVv$Q`{ z`y*?A)ZmZt_+xVac!NJa^iSmRCn)+8O8#FJfcuS!J!`eyI)_iR6Z+&CiJw}2SW6{4 zZKX{CcMV&&8dLNr*LecJ$41{%#kstWaUfMTiOXNh3%8et#f%l~@BKiD>^<~6cJ;F= z?O#j702~w3tGH=hxJ9H~;(d8q!tgIw1T?@~gZmnia=kYxxJ9-s2lB$OJxz)ttnaUo z-){*P+>It~Y@9onWbnF-u>FzvI9;KmUM6Bl=aDNbUGX#BJcPzU1Py__QsdZM*)~xJ z5j_{~GI2}f8gCUWYx)OhDwa$tk!W3#wVZp=(6J;dnZzS5jbTPxDREdfupw4f306Hb zxc9pA!%CI1nPS<9M8;+UU#apG>rPUWofV}=aEja1&hp026CxLd?>jA1Rjrlf29gO% zjFehyITQ81#dqK8D7W8JYL*cfe{)HARRexpnR_`P^!egNP{1;9HSb}bm?G`Xl#XW8 zS>$^*K?Tm`XX!qMJ^39|g%J1jthGXnmvJ4HJj2cbJ3y8ynz6^c`D{%)w(_EN@MW9&Mxt=eM%|J>Q0Qg{*^7%a?*Na%UJ9PjBOAj?hrbwt2Z~- z^tQ#tV>nctJJ|bJea;o&i@G=v1&+h3+N_I?9X$N7kRu(?x<%csOxYb znZrNl+k{uEjnbDpduq%QcjTzmMpq~Ll#U$wUwDUq{qz#9{YF(sB`uR66=<7|;xPUW z`lbK*=hc^VOg~R6NC%t*O(Ff^h)cnFiPJ12s+%J#j496ry#iWtP=?~%UVfeU9j~am z5bNREFIv-u-BC8|g5o8fpjM!8eBVzZ?I?bWuDEtru)rxjiLt{&T&V4_Keodfs>*e(n~}xr zzOp5GPE<(SJTrjdkTNz-zjTworD{GBIqv1_xw&?3?PuEX;6}2C0$Y_qUvuf`b#lF#dZI!D5u05T7?B4IsXBfS?8~r2x z>xg&~By)|U^MqaQZTof6%OmERA4>JrKJ5#HPcA%9uKe#Ir?yaj|DniV^m!Kz z2s4KzjH%~g4d?+Tsx3IQ{N555d+a{$2Qu8!o`)U=0XYLQ3-hlUN%68rANQAlMrK%- zke+Vx(Oc7%E4M4pybBTnTVwVbb5zXE%34h-B6Mn&pv|1>K1?E9#F7pm^VAtjIT__T>>s)L$-Zc=yu!=jZ}0HhE!+V_{AK z!2`qTopcdjddky<>=Ce!e7x8}NrKoy9Nu_W_)({1<6P>j^1(1Cb=amv>6YN7aGxZS z{iU|KBy61zTWSU1y4^kGsrh3zulPIXzWGbwIzTK{}Rg|yD*ZLs*7mUi{ z|Gx zCf|PEj09V=T!Hrzih#i@|Ja_tEK!mMcH%+7&v{9^aLA@63N@EXRtu`MrVQ^pE#Z-b zHV);8T7&w_%1<|={;JXT7$c2BM6Rx9>$$IzE{r-u!ElX+ijBb*nh>GytQ7puDdCu; zk)wWFr2E!|zDskq<>O`ctHY+Pr(f+xhjJeX{tb*8=~{LAF;;(2-%QD1Du6f7rc){9 z=U~o%!Nfd*cj+y_j7649Cm&sy?at3wG{Cz9>{d*Tv*+CwcjeWl! z=ff2XlYVqzk^bE7jqziHj*kWt{MnjMp&8cyE5H=d1Iq!+13g1Mr3m&SYW;|x5TW;y zy&-Q9ieQG+>Yj_oLTBqqS9kCK%d0*mg^_<+2)J78)6!}}|5 zMy0y1lt%YwYj3}Fh$TmnG~cHLMf=*lDOrJ|EJKfVMPWBR5Q{8z4JJQlH=Nb&Z1b=J z)v)S=&5Jwn9LaRvPjl6gd649`mo+H=Q5rlL|CNs>V+OH@wb)`AB?%sP^5+Wt3B#sz z+Gh_<6CBIr5)Y!vkKaf5x|!P2TP^4P7{#E#0F=3V@WTQ@;49D4S=wyh2UCau_F6`C zZHUyH&1QHktdr>M9NmlHs!2$is*EXR*r)u3pO~SO5E_dqyA70|AEbzZURQ5-rl*1C z`;M`J+II6}FmK603OjFU)z~2?3c*4oG9Ig1xrZtgSjuH@oz@O+u3-roz|PXTdW=fB zB($#@@&G{4Yxfto(vG_1+|}4>j${fu=Ekd9M|(49t-IQmFFxTw{v2bYA9Z4k%|8&S z*1sWAdGI6Sx`&YJJ=4OD4nAFWNy?{C3@nZ8R)w;0ecmx07%ocH?H~TS@_UCu`18St z+(F6o7XzZ5FL<7TsXhQk6SpjFb$;k)otJn1>6ap~#kVXAL1BZ@6JzG@5(cB+vc!_p zfKK|3t9USruBqcL|C#d5xtu5=2BLsfc_y9eX<2I5H@fivbs{@EbuOZe&ov$Fl+jip z@;Xh|j)qpL1!Ho@0YBr0Y=mx<5{d3q>F|r*-OcWEYji(%$`O<7u_d=F<;!-XM!#%U zh&T{hTtng3(7oBO?odOouD$Cl;IWt88i(|JbtMYK*DlsgA>=d+m&wC_La>P2KPLu< zug{t8=l$<+QDUF|3tSY4vSl+~G=V7FbfV-jh_Yof+JhpnF61rNg^ZD!FlR_LUM)ps znFUt6G|wmcKrY%o?Ppi;wtlvWqh;9jtc?pmZ6YRPsA;4vF~P89()Er`)EI9(X1Qds1X?u+?>dvirMa{Xs{;(vR#2)wjvKETYJvsY=rr z9q5h*sr~0~KgW(0>(3=TQ@iH;7P5(>QGwJ4ix{CcTgdd~S2xPz5n#nP>>W~m; z+q1nKw}GG{rpCT3W5WLjdv6^U<=eFjD`EgD3IZY$(m22n(o!NFD%}XkAT3goBdOBe z1B!%ngXAbk43{Xn`x`5ZE#&s*4CBi^?w!n!A#sz*i}77m#>y>uB!%rcZswyMh)iKUB zTB}HU>_|x7%q(-@S>3rE?6EA@KQiY41YO<^_Zt1^KCqtO8P>eNdbqbwd|p^u*5i$L zbwgJx5zE1$&;J^C)_C+JL;kY$sIGmz8%u(YtIE#)lyh!bL|yLcFER}j1yn5?w!1OJ z2;h5I_s(%}xl?2-J3>D+w|r$1sQV`0(JkkxLZ<45gO%d)`!B|X!PP=t09c34GSM`k zdp^itN4=h@Zn`Y}~%V&tvN=5+2bigBK!`Z%s z#U2Z5^*YmoJ;`mOodWhGgNooJ;P2zMVU$z^d$Jh2C%J*kjOQdE!aF-+kkVVhS@qBe zIg$4$$>5lZ)Xq;ZW=cyR85i%I>rGR&RF;vW`D8+1%wtgz>vuEWqrN{yzqcC^H@g1> zE~L>Rw6o-6rFA_J$outkMwN!3{e0{Ba>~`hpG(Bcg8=nQsk76SJwUoXGo?Yy6ZID@ zMVS2`Xes*s{qLF_5!JLTqig5i+@C7cV8`UEvL8M2y5W=h>9~ieb?lpt!Pa87@|y8- zNG_4(vn3;7i5iRJT|OJ|n~rg!B#ZwZ@PNY@OT~c)oFGWz3HZap7(aSrnfvB}2fQ2y zV9T>yLK=IOEgiQ44jKwnm?3ufCj?7JE;BONk`XEl00}VgjlR(PXa;y=5r^ZCTxCGu zqWnmdcqOLhNZHj<`y)jm$0ERjnIi6gqy#L5Z=s<0o7_pkrO5w^Ffs)&{L4w9HHlQ0P49=f>v=)JRdc2U4t7qaFv0CcOYOSGRP7 zzOY1BNd6}k927>gtHcyV&sR8q?Q`+*@jh%^Z1ZEi;CdNQD)lFOX$55*WRFuO-Dt5G zXM2}(tlo;WwO*nMBGEd8~y$%{;#=!fPiO1 z2@z>_8A?}_Z9Cm}+Stk)J^Ik;?TqWALOrP_ZP8bT3TZR}Uc`Irk;5wBM`Nmep4UNctjJbYv&RtU)|yCtEaN zlH!r%$%qFismn3X2aVuZU-_aq@Vgc-(_9UYf*7BBi)9!k6ul>or4Y>+Pm_hVNyg|h zs;Q}s78N7THU=bmN>R<=am3o_D}_rwe8V0W9JdHKl;Qx^vEq?+m9XP#QK(%SGP^M_ zT16>K*v=hW`UqNea@oyfjGiqq?^8}k3abs3nHkTVwR)xsTy@|$E0DSo5H{6@?~E&!U!MBxqN z*)0=~;Cs@|i}H5csCNipR7`te3%?$J&gH(eZWDIcUn-y7FZT?sK|9W?gvbfu77L%v z&(;=c5vW>XhwCq5i{zM++~o3Zb9ZE{4mofOOuavsrn6%`D%_^;zU=T$DrIz^T~DyB z4JyL4JY2pr%nhK8-g}@Sj4|y8iE>UkSn0ShOkKo0g#`kM5Wh&dkZdx;P7S<}LH7Z@ z=>Y?Hlt$ueA+^VDVS0tGTNg9Az=Q(!he5H|;`09lLP}Eezd%SKA8pPkNlaDO-g_o% z6IfllIn$Ba4E!lQUf=>2qbw>)ouI{!4eR9(`mAngVG8ZGNaQ~Pjm>@_=tFP|L(ALv7&Ld7VI^U@dEuF-)nQ<(_Wj%sEiu>(WiuQ{3~24 zC?{7ha>)Zxnrp9M#vUs`!pSlx@&J@)CEXU%W|*Jr^gpu;>)nYE&OI=SX z2~gKRZ|}pK;Z!a^2AYsj`kS zdr?ry+Fl`OyqxylqWhZG#yeH^c$%;)EdlR&mKBf3{AyF?8^67I9fVV{jfNRt<)I}0 zPh8=>!tUeN`_ea1;pBoynhwb51ArS9*x`i4ytNZRJRf{uBRB{f>L0v!smOfGJdS&o zK&#R+Zu5tm-C&UubpE@ixSLhJ>;#to>7^PCs8_Cb-k5I7uMH<<)0P1Qo~au7Z1lIU z>PQ*^=32%(A6~fPl`c){`EzO)Dx{_O>R@b9N>~5%8=3^aLY|!R`OG||(U!ye-+8N^ zW`d!~6DPD9a$XAeC;i*5iADt=pW^UgPOkg@gIh40e`!F6MO%FZLnCNJf2!fSWTYy1 z>CX=>f*+cf<2)VOsn$-{+UYj<-}&%R{&0Wk%$c)mvJb^oPj~D8ExYv)e&)=X<6{>^ z8!-7|XRwnGnpGs;klf#0Gj)WWDSU@flv$NE!wr{tvyhp!)5ERO zUyP=Fs|juSnQgg!hun>Cw3Wa!1z9^O$rzWABR9p?aGcyP%BX;a^gi{}>jj zY+`k+Ru3MHy)I127*X~6`}-&O>7Nm2@VJwIs%1sXUW-Z**Ez9WI1FC<#OZUwaz1HP zKmUuGKKa0FF>@Sxl(HT70yCp{AU+e~54)!GSt3~L8|HM+KoVnq&69^%Ud`XV06A2_ z_kHkGOxIEHqW>tV_2;pS29UDSYm}-}vrEUu{2vhUU((=1qAPNJt@Lk|$KP|F=-?(+ z2l>~y)q5fOCI2MXO0qtG3G|G={L} zPb20@BWPNlKlv+cH+USK_rAP|TYa-XVdx(pI=oI}I1^F@0!Mf$s{GF{pR9sLf>4eg z$hSa}Eq}7yzg($*XF(CW)~uzlYwgK+>YoCO&_x8R2FjQq?=Shqj z{aqRY`v0ZbE>_Q}j9?ac_&x?35qa7!V`SluD z>6`eR#*#G?k{18c4;s6sm5ISD#3wwF=EO5n<)8nf=T@L+6_JOWSDxkEtpEDYR{EI>VK^RK} zbMcKWOwnIUDgB3@Gx5MIw6}ly7>mn!H_`r&o~MBVSg2ECFL0>`jz2yBcWa*X>iqH; z>>R%fym8i?c;SJhCVzyb(F2d?@m*Z+8%x^MWd5U2;tMr+d-NHi28p1N^Yv~#*;$|> z1CP7r0ZQ{p{;RHk`YOFZ&uKHT-X{uPH%XKP|LA!gh!{!uMByAcr13?8*Wa!A=lec^ z?~^!}pimx*D=QUH^hemxJg}kc&({bi6{eJr_Wtu&3juFmt$2pw>Z~$))>#uVGl$7+ z3F6u8s?$=;XthyHP^L{gsxXQJHX2s_qK5!cK)=Xm8H#Y(iC#fhnUP26n%(2m0UZB)6_2p%GSPdjXqhvy9tnkE@0&*2%dU z3`v=n?m9QrRcl88Aj$?IY?-k^48I9i-* z^e*1nj(qw8=J=8`vl`vcZil9PS>UmFHdHcC>-qY&^R8G+XXm3e1pj5vPiMvqGL!M> zQSk)?WrI2o^g7EJMJw+W^3nFCai^}wB;-1WwlZ!nmbduGhO(?GMPXx+q?#>7`}tQX zhuTJ6msuv8g0O7oy<-hUojBi6vuEhFFs)JWR=u4ONT;Hkcvf~c`lI*P_e-Daw?S0P zkZZ1)Ze*GD6B7sgD1>V+b~3G~NclqYi|UaV7@h?lt~WZ77Rj7wVwmS+3p@t_6|V%s zf};%H{H{2POl`WT0PDzKm4iXs^Pj)^6s82Pl#riApMSYruDJ7RzTcw*il$B1a1}sq z`p4kqQaY@qt?y6qA@Wd(oO*z+Lt&28{g?Rb)%?va*;pQq%Vp?C>#o6NWfyKOz0~lZ(K}|&QHfEUXk-9uHasrqVmC0zsH*lP6_1y3Xza&P{x9he~^TNA_$jG8{KRCDBJqI9BY56U)e` zTs7_jyKCt&7k{-!Ua6TKXnK*4`@jlKS7#yejT2C!!EM|< zz{E3BF0lLlsAw?{UP|=5Z5r(~7QrANQ+d#+CoH8TWb&!1!|NCaa7*u}|J3U>J?;GwF{qxmYo)lxkk;BQIKFJy_E(m$DUB-R@|+BTLUpvH^|! z)mYhSfL3ykwx@zDd{mIX&LK^s0Ut)adp{&ReyA)YQdu6nA-{8JVc-ah$Y9iUMF=^W z?SOmH!&db}gB#q09iJYxpv9YbCUYNym#Mv7qC@-md&_>n#4&$o=3DDPnLVd^24#Id zttHa_=i&php@52QegOV@<33eF)vA*FRG~CgBxr0`u^*xDqLp`3SbfRVWH3LwDLC3N zBWYsA^sCNUqnTVMMswGh!FODaBe=7<>+9=Nev5fw#KEHd?iL3k!;V}ti#9WY!-Co^ z3n3zjj2|91v(v?v zXnIW|s>c%MZ?fWi<1V=oI&aP1sm7(xwM)mjN@r8&yF<>Wk~z`DCb9<1Ee;0Ms{A0*v{&8Y=aaR%W5q~>LZusFR9nD&Xb9m+Di{N&R(RX4IqQs zj;>FEz~T_zBuz@lHGIO()1`NVEAG9v?S5NqH(=KqQN6CE)S~L z zqNT187sn0G@I}^;46pcB$_=jrlX}zYQq{MQj`2?3-5EOL;r%7f`j=iJk);Hs#9er0GwMZ*1 zQW}`ng^pjt$4}99Hb29pt5yZ4ms{3ja?2eD7NH|24M;BYC1i_IPA>a^qQwTWApHe>nN$h#GdfYR)Z_6 zqY^zKshA|;S(nG7ks|4&!RGuhOTGI0Gc8Kq1wr-RuhG1;PvgbdkIUjn`o`)MoG14ATlD*u>`v+Hd?k$c}s5&MH-xbKx#>QGjU4 zUcsJC`HqPJH4{%8~nggRNYVKmyJ;)`ZG9`0Be(B=g7o${}atgN61G zkyz4lBG(sAL-VGB2j}_`tuq~QY#`fa3ra^_aqMP27VkUdmHE}|45?TnmthsFM;9Av zMJqA1hU(dqtH_E&@*Yk@(o3ZBhGS&85sWTEjaep-Wcyavj``mg>_ixibqss>YQR}Z z2=Z;C9S%eCj3p^^&A4$Y0hg6c!b)M!2TY=Y;6vFl-j^Qp6ha7XkN&q|Tvjfdyn$@m zC84&Us~AXUFjpO0QI}KDqjd2}<5|A!XmEds*9!E}x}qwcuS1{zIlq{6aI-kFQ;3h< zP`}M8^=XwW8iT#(7P@!P4ro&QL|m;}Bj2Ts9S`;`3n-Gg#&vZrk5Bie=(x7?`n)Nk z9dkfULx?sYL;9I$bq1f|YzwyF9djsgkVn4jl+J(4|Lcp#`mb|44Y_$GCA}5l#bsYC z@)g2Ig;pL{wIZqTdl^xVKMY+3jz{dKLeE~5|I%s{jGPf9f>AYo5Kx<5H&tlzi4#^T zX$mOKZ#_Sm=Awu;@anhD$CN)^#0wu_bdhvTbt~K&8J?pi>0KF^H=u0t7u@Ss$5jB_ zH)FgJ6Q9ft_>rnXf#k~*hg@{QSvz$~v4ijz_jkV$Pu*T5+y^a@QhQyh&*O3T(PcOu z>>jmRy2-@$y&|hZgP*8yliRlFJ|& z`3(`|&(g81%AeW^)OWB*7MtFX(ssJ3d2L76-+5H)vqJ04ZIbew1tI&}kSuGtgKHdW zG%@gDg`iA1%*3HG%dMGG-;K8F*sLFPLRzp+dC~}N*~+FQ!P5)-?nnLcc16STc1A`iTIF_v=VyJTRiR< z!-uzLkhmGH>oKjS6vtNO7)F@>NKrEh5l2P?i{oqG7P=~666>}-IIF}%W?=szk?-Zh zy1U0+`AAm1{VJ82=mhIf8r%)#$X;VzI0%<1h$>dEnJ~%JT^D;WpMo$TYk|CIRb_*VXGM#I7l7|| zNW_iXAH7G;x$gF+R4qGteQOCyM1&OCyF%`;W}+0=&S%4v4Bxk{TVFy|gB?(9RyMZw z_G56=Dl&Oim3!YG-WFH?=#l2M__+(MfQUAci?}f;O2zN0EYQp-a5gvks{~3<*1FcG zNlv1n_1k{TTiy{ZC_7O!772Lgyor;{ud~ur-*Ny;Ju9JG1XlsT zf5v!KjI)%HV5$52Ia`e%azWE8H{bfiv4<&_BY^qA%L``ozHuIh4v*Shc&4kAq{@_} z5YmVNrOjWJbzDGX89>bRBi@Bjtqe>n*EBklYrPsS;Il2kHQTYrb!QgP8H8=H-<*y# zQ}SXPPblPO0IQiOA^#$!$5a>B>YF_!ioDN<$^wh1alB!>ALG(3htzQU)_Y7}qkpyk zh_eMw7CDpR!48n{*oUbj|C;C(-Ew4b2_xT}T%SQDE%*2}t zQFWCV*Mq#M`M9cX&0K-g`s$0$1}Nz5JKdlm(zHaF>^Suqc*6bSjCfU#sktCM-j@^uN3CYROjao#+wTdHdl*{(q z9hHpgxPtpa+)yzTmreY#)BW+(2$TMt{6us1nrD%Lx1TtrJydZRfcp(1%ExG>sSFjb zZYWCudzA}t`-ZnAbFvY`mIXAjv}CT>J*_I@OFn__Se0G-UML-tO}QaLNo8 zyyaaK3??c4n)cz+O~%kRq?Mnqa?8pK4;FG#<68}TCJM}@ieQqJ+shI~uG z%^v0ssPThLJ!zR(GcvB}Tc{lkn8bs+oW|nc7ZnU@%2s6dA+j^u^^>c_dbPTc%-1rf z$^;0RLg)ey_usp=)I4gba9~Vtp_~y3viZYpdN!XqJnBl4fT4!k0j~uBZhjc1bfa$>Z#Jf~kn@8{q5C!Hc7D#vYow=^CsU@33x0{+w% zM|4+!vl1q3kb-jIsPg%!A$O0Skk`$xVdO(Uo3lDEYkGSnuv1-^O7nBH)fV{2Y=_&i z-uXm4A=OKN&>frA^!O7T0VPX>TaOeiH79{{HOY#1bWkaJP_%M8TRJ3ns)og;As05o zl-!zGpVGO&!rUB_ZwkM&y{zn<2|}9+2+iIy?A;;A2pax9_ILE?!#DVXXV|wsn`%A} zx2PZdIrjR-w!1ja*Y?EaK6oud6tW*vsODYQnhy+8RjTL=pz@(y zrz5Z>$@z%273)ww%=^O^Ymr=wkLq&zdycZl5sdy*r7y~S_>P!KHgYdkNvFt?mTTHl zv&@mWlSbD7s@FKoWp0E)rIJ_m<~ro%;Pgkz{gjjkaYv&!=6ef`KRzBg{weWt(W(o% zV2yU+-rKF8?=CWFlbZe!>JIJ__eq6;f@9B%Z4(XhiRHrXKQg%-QBgrz1rgnj72Vx^ z9kl__%EuiP0x3@fcy~cy8)b0(s&sb2RAInJ>v=!~L$7lzMMtaefOfgbVY8kSDUP%9 zub9P=VNSVJ?N%fOW0=3FQJYyw`l99&I-_|1*G~XEYZ#__jb!EWY)su(&i7`C1{ZLj z)U$)-%<&wbHx4Z4^S&Ua(ROG}CRb1|<|O|LZJDXbFRo4 zw|La-bO?y&GzlZLS8I;ND=@)x#ceqKdd+!MLLh#TwvsjEu-M=7t5~WzN6SE<4-3hH zxp62Cl~^N`N^d8e2a~QP*8W9j=0p?k-#^56a_g;x7`L;T=8EQI>AtGjVjX^zVk$vH zzbDKd^K=+9QvA`SPc`#!n$$_J={)dH1+{x%xC+;r`^I>QjWQ&t`H}W-!)vX8f6C`5 zSf1z;*U_btv5bvkkYb}47amyU$c8pO^lEU>$sUk@Rlh^~p?3BZ{R$FNjwR$Tdm9ESS>$Ru|bN(V9Leh=_Q&**KBPI5#dQEVi*_@Q$Yjz(b(X| zwAL?u;HcXqDFsbWPjme&pc~0XE$u9f9iJI=?G&Ew6g`gQ%kBIDMIJ7w+s-q5d^YvA z80|pX3-Sf&N~{az^cufC^Q9ky8nJ^K@qIP4Mq^2w%7w}}023?r$qJpf5Ttxmzt>N# zA(`KsJN#iFyaPnWQ_a|LV3(#AY!b-3RFGte1`ACxv!q}*+!l5pR;p7N`;poE66^-D zD@3p>*nWSD_uE(>f8Xu**b49)@G#VKhY4?>D%*ospg?5Wu(veqLs5ADL7hSf|NM73Fmz&I2rKdr!O0WHFh(sFH5O<#C-XKfjYKszSE;O{ zSGipq?dAb<+xd7AG}LbV5=$*Vfs>zm{rHRLn^qGt1A26olSN6$l_R`s+{a5Qna-1q zLD(>pvuUg9V2POuO2>Yo_lXo{5l7(YvoXWXoGN_2^lOk>0#G~%pF)O+j+adSiMP1f z5MTL3++5lj+mBZqUvQpJROW$RN5M*mrW0gt=&j#?JByk;Rn)KaZ}8`jb@aF3iZ_8H zqW~P4Co}^d!wP{PTHS#oBOth69<(8Ud~>1K8?@PM$#Y#|>E}7yu?Xijm3eb@*5xH3J;e%EYcd(9n=){dXhNKr6V|cmo!sWZt^G{L+yU?e~ zbtriIQu*sbqwS!~v;Rnez`^o4YWlK6%7HiGdg=L}EQA09HsAQ-G~aldZ#>O6p5`0> zU-FII@R6vjDA{Xib>hZ)r#rdaMVmVkoz(V&Q%Hx_cQT4xKgY zBj3nUD=UYD1#ez%4HmM50%mkZ@8J7&=ngLzoQ>IIX76ADeQZy)UwBPzR|X6|>eJbM zmaNz&?-2$m>(c1QG29HF_s%TB%{Oi#D#EgOlq=?>@bVMT&y#~G6aW@@PATE~chLIJ zW7s|<#(H`7u$#@ur8H(kKL1tR_n{|b26sWPt)BT-6d;{WP5tI$-F1-~7i1o(5k6Yy z>Oq7G#NwY@SiJw5^rP=rWqzvB*H?u{k^GKt{46RTd5t&^2*2DB-eHKuhS&DsA)>Sh zK}QA6V8rO9Yuq-8+-d>tlHImUN%UDlS)Qs~Ydmah#Sd%;MG8LHj-4!sK zGh5&2BuJD3c>pM>VqaoFGG{I^%-L7DpuIhO!2BS+POH3qzy7X}nR?CsWTjO+`@D#H zZzKXn_r6q&{iw{0lB4}r=U$rDmwOiMvoWx<-0v$M-QOxv_&noM&h$Iw9uD9M5ZMC$ z2HDs?Z@e20F?Gwi-{jw~6loklz3UfewG((_s>=@KXnOmFX{%=FB}qv>ZcSR{&5${`029Vl>(!mtLE?W!RF; zDfwZYGUxFFJe!<@!@lKSQ`ydlDoeMh)hcbR@oq{d9`Q`m%q08_&~FKh-rAA&`W@fgC!@~U1mHRaeb zx-j9mUe;Xei6Y9>oJo@Y5PqkB`PL(+fozQ$vYM#!DoG?Z$ht4vV}y(P9c1NB$?$;k zBYQ)_I@myvRlvoi)P1nXgcB-_QzVb_UZJdd|Yp83`! z7tM%zbwvLf>P(%OEeyJO?mR=+yKpgkuFvKT^qJc0k3YT?wEFHq^i9bRbKuW;wG_jF znpYVM>p+x0e+ch%Cv2v48lAQrb6HGUUm3Q!cenn64er}GzK*!82Bmw{FTUR?dbLJ0 z%l3$hOEV9_k=HdmOh~jmaC}Xx#-+4)HBZPUqW&UP|EJp&)x?_B&)x-Am?yezuAvwq z42J#m0kag?{H{eEE4*!k{)cU-Yo6-i3-JhzkZGr?$6k~dq}sH~VXVnQs!Ox+^ZVTZ zFP6=>EG1@&O{Vn-SXY)HHsoVu<_DEJpu&^sI@xsnRs~53yVZxYGdd=i{VR&R$v#>U ztd@C^jIOJt8Y+^5(IW`JE6$t+hVg2v{S@vl5^3Kb+qt}-K~MK zsbAtv&5c16Xzk%r3rK=hS=^6JqnzU=YWiiUk3+*T}Q$%ztx;!HvQfy zfg8+A+I9@u_D+4_(y8VR1d7pph99xGA4@YjoOIBc%_hODP_M>zW3pR1VcLjjTjK}Lk8G*$A`}L|My01!24+m&Tl()BR zX~+E;$-S2^%6RO^kGM~7bc8jh`>s@46M*E+(ng3!KX;@${|KKxJU8)`7fCUHKJlYl zcMUq?mzOAQw?(JNV)Sok&?gkLh@pxy6ZUqJb`Jx%_Or4Pd}aBghjZR$wvfV$P8kS~ z^kUJc=uFF%#6e|M)tlw_g^HLsoL0&^>}Jg^Kr=Vt>21IUaf@Ue@#yQ&Ck*XM~i| zRpM9#>xVc1wBsSpifaJbX)%(`O38Kq13`;0MN#>^YP(6InbHv{{IaS6v;ny1GTGh7 z^aiZNLHd0fqTVb5U!7G(Zm$P)SM3C2!5VACc~MNs9{qzO$m6r?-!Tqr7MjspOCY7U z^cl9HC#_SynXn5|c5lxSE-?{`zAzYBkdv23Km;rZs<&I|=W8afmS!I=eh{;&vwTBY z7PHYQEPa*yMFDOO>&&i29-ZNg;q7=8HFkKTy+Z8ZliF7NOPG?Qf(G&4y55!R)NL@!jtt7wD|4v=A1ZQpgZpimF$9A!IK)EMA zxj>hE;xHGRydT{c?!yA`0I^9Sm- z+ia6d17$TXEjl*ULj{)>|9}wdhLasSD!8nM4myOA9Ty(YLX(Nqda70pK=yuw<>e~c zsrAE~%Q7H)51IlNJLaO2Fj1SsaAfQ)EM%3r5Q|#?OXn-NQ9U~6)XKNMGB>=o+A-#m~tO{V@C~xnL&FY1AK%-I&^>qNeBp}mYxG2J8 zyZRZZEgzWykVY0wnI|BTaASPpHWq1^eq?2rrBQnapcP2ouvS*g<=f~G@RUy#S~nSw zMN1FF1V_*$AFXneyDZFKb>_GkxiBp1nI4RD2Dwk? z9!un8sES{UX|nr@s@<9-S*tUhG=0uGHDfv>+d6&tX(WhYf?;D98G82mR*e%w>7Ec%GrqD;!b_e z>gEGz8d^h`5D+&o`)UoXVY)7vJdS!gjM5Y zDD+B<8D5o4ji$eW2AblY%T-^-u|xS0+V z!L?#k0_=b?gnvPE@94l;iRWYfZh4}>;iiuo)xD?hS#?XAS-1^HGzCoNLV;%Ha{>O# zrY1o}=VYz>3R2O1VC%G%&^g*FB+4B931D`Yq7s=uusf6+vL<%=Ru6bb5VM`l|(4<%UVpw z4S`j!CrWWM^goDd$%RyiXkaXbdoJI{rK#99JYrW%#{Ju%QNEEmxZ z2vUIRTbPp9xy~W`8|LQ!G0PpKj(E6WB>@|4f&2?U2mtIrN->vr1Hjx~oxt3LS}N;u zerQ=J;p9AzY3ZjV-$1OZwIxMmjZx=2GhvRpY>84m*3&+JjzYze)L7}>$LwpwIxKw` z7eVY9R;O5Yh6DTsktOL^*%9GNECd=eZhBc>QwIY<9|E%xM|mU-*G z%s3m9f7qzev1j7;Hfos$kK07zqhdW4mHFv6a%kIms1+_r{aLeSU2%~db{r!nNYiB+ zD;c25=B}C*Et>qkdZ-GZR`=#68rXf^Z4y_4zZN=Oh z06?>mKIqXztR$!fYu|6poavUAYH7ai0I(`*P8iczwDSSm%c<-AQyfqeQ1Qb;j9-2S z3%=ws>_2+kvg#2=fdnA$9Mjm`mje{+VAC;vVDppWdjkQdOxTrx=f!ydIYXg)%m20E zZ{&<7bJWt+*_d!c&h#e$`sJfs6%3#T?9W&Hn7&+wN;}hq2M$&>9-|{Y}RKtzXzG(PPf%KfJJ(Su97JMMcmJOQ(E_Mt~)e*;)X=;?ER| zY&R_Jyr-j!`ZyhW5=dMTC;3w@h;_ynEw!XuM{f>^t8&*Z_4 zt)+8#+`x`i&c;ID?qL_OUu7c8hj8ML7X2FxtvMHq<@sIG9!jh~tvA&~QJSN}O-<^J zhH`+NTGRyEZ)T~b1#eK6qgjD>PMf#)^G-%HnL)?P?U4O}y_aA2W*gbsf|mejacK}X ziCBIXhxurM1a08uMibW6N6cX!>zRpytou1hUS4i{F56^AW9(5ef}l*T_noAbJofG| zu2WdYPkTPAk^Ad>;x%+mq=9A*K7 zblT922Im4k)*$s9SIv`8G-&nx_UVGH2RVsSuAoGnLiP)mK}~=;pdbZ~OA@$`waW)9 zhWZrk`!rDomv!D1y@VL1Z^(lx7kH8|U>G{~lz<=FHZ?2q>5o0T$!qz8rV3+4jjdJK zKHlki^9%{^)kE}8iCd!l!8wF)(LcJGU^CO&)zo)*$5X8>aAzz6u>A08jpdf3?*L-* zf?}H(J;!KSfIy$KEaISW$Ci^4xYo&Gy%T`h^oRYiVVLsX<=wKd=fi<$!I;_sq#4+f zeF=_q?~8xilDGNvhC24rY`H)5(W4aJrXbycUtfF#*0At0^46$Xii|#i zsr;+S5M|CWmjYM%_-D2OBs*bV`dy}mgTqUvF@v2PJ$q@O?m!-jq>a%N8TtUddpB}^ zRWVG7QivD70YfN4Up`8T$zJKt<%Tk}& zJ#P?qjhl1_Kj90pSj9lt^p=;Zc2fwZs{vV6l(9DwTWPr!qT!kc@FR+`zGeJc4-juE z#sSbT*FmQ-WW>SrPAJoPctuAI<;RCvX@^yYM47vEWrsBdw&) zKN;@;NY2~0&&7k~EOw1&2?vk+ig)_c2(X-WAA{vwj9PX4fGc#_zT9V^JsKc6(K_5S zgDqSZIH|muFq57fywGtX7J~(i?FZO^%Nge{HZSyb$?jvT+L)-|s(^U?r{DNi zi%m1A1iRDM*Vj>Imn(ZHpVQ)|e+1$=@R6R(t>y5lIg z>A4~QhIxKissrk7oMG1{8w(IZ4nrLvieV&b@S?v@>WwQPb7Zs;VPUp@aeZf9n`S5se~{oGZdI-SyUQh&JsYFbeq`9~PS(3+hQ zm7~LrN6tO_ftdBSd~1>zeJ(kvVE$n}Q*lZF4q5n?;vxTX3+tt$f)Vb%aqekaXbA20 zGC3UygyFln5T4Ia|8mwS^R#H94aQb&Qx088>B2UDU%gj8zMuA~pIF{eS=@EuE!VGm zX&!<0)^}!OI=XtwmqC?r5JHY^_5fnnMPu0i1vL39>nKLyoDZ^D)<8agwM^FEXWGx* znKQgJchv1nL{IeYEkYc*kynNOrm3>V%x=v9SjgqdKr8fg$KrMgP|kM8(JxCIL}wH2 zR?e^YJQ8JcYkJs8r0|MHNEo}L@$BTHrY1}kW_{wKu16f^(SeY|_BGc4`kbzsmV>xASt&ih5@YcV+;r^C%`3N}t zx7G-6zxt%>xF*?M<616t@Z)DPh{tt`M8)%2k6=Sip-$-vQHlw?4b_qvN|YwONcPal z5FvH1ua(&^jPVO6s=hez9ZDY^>=G*pDV@Y#I(B%wUJHVDeSu~Oj%+-g7aQL6z0;Q$ zR{|Z%z4C&dze2|-mrgywvynS|bD?xv?RWqbr7koAr-m3?l-mC7X4VK3aB4n&gb(v0 zH$y%s32Rn4B9)m|E}(fjlPQBq(>2vYs1A8HAs zjg-CMXC)wIja;og*4Hk1en;4KGn7iWEc`KZ?`IA`%|K)3sk;{#2Mp6I)-=6JIBc#0 zejL6_)?Qhk5=fmANSzW$of1f$5=fmANSzW$of1fayxl2*)G2|~DS^}} zfz&C1)G2|~DS^}}fz&C1)G2|~DS^}}fs~KdDS;H0+WEgCkXk;rS*<JRnK(Oz!k$%W==^*-W!Z}iQF`L0)9 zHVyyX_o(-{XT#23h^;6@F4Q{hD~(Ore|nJx$6QnyS4G%)JW?B*xUZ+ylij$ASFcfX zf6umW3Q^eunTYq_)8=G1%Vl_c)l(lctR6f^f+P#hpJ-y9Ar-7O3f8Z-is1<0bCfi% zSj}INl&G`xGZ`a2=pQpt*SdGm<&i*&_`IRff;Mdt>I`pM`gZir`C6)WuAlY=493*s zeb~(e2$g({HF?Kg-JszyW;?;6)Px-Wt(M;py)5(g(g8nW{eAym8KQc{KdH!EVt#s2 zXve(+Z8sQER%+b=vzjQpVtP8#3!pXYCRE61{FO#iT{BOH&pU?*r=RS)pMO{w*Q&FY z<@8@?&KH9oSwU>>1q^UEX{;G3r;`fQw%%yx-={Ssl}mNNp@!=O z6SWPlKCR&-Ev_i3x`*rhh=pHUTl2oW0aki(7sOXSZ*+jN)>TlIHbUe$>@ZL;kOZmQ z9zLK`nKD*J`QJtiS|!BBUdq##7)n+7EWxC^zxi>YcOQQdu z{|&xwh&YtPqBB=jha8thjo-oV_sbjlprrKI9dq;d77dJ7xp6Q3{%-7#U!!@fd={@b z{6bljHE3ukDX;bWq1byB@Kou1F?zJrsBF}n?%&4bh6~1I3sud%AdXLRHR#UomuY07 z908DTc_fqmG5tjUDukFT4iRj*BFY@%O_PS7(QzVJ9f3?NiO566EB3iRz8Rj3^Y5_a z?$Q{lcYi7CX7i@Wea<3(BDFnu+MoHfY`97t)vsgqZ&HMY0F8*O_-6Id$b6{#!1?>- z`%Un}w!S|~jda(#wNC{5qub|?fJTJNm90#!rK*nilAoyJ6%e}a&8JMyJdy-{)Zfwk z)|Qwom#@J^&q0#5}=P>!wx{(~;rLH{-t>(@g$+L~$}m)^ok%U+Rp69pxH4;3E@RIz)Fhv$2#5%HUo zasHhQ=1d^8?73mN^2=T&DQT=H6Z#$dG{bAqdR3cK$ENi9zfFY*(8%6rvMhCP8ejZ3 zjVD5MgQ;jBF&glh;m;42v^n!$nX&yD&B@%AFo9)|q?1|3bxAqv#lq#^FPW=>(Bl%lyfK|>8R;xGW{=*VF9ua5}Y|PA#YZKP)HSWE~Tl zO&iRY$K*q7s-$d$eQOSe6DwoB3uq0-9;eQ7waB(Mu1j2;oY(V^<9-FO_R{gZs@1DI z!oR*09IRAHl1J#9mBcoX%>x>4U!DPdiabE5Q-IkXesh3c{aNsAZDLoj`y<=Y9@OLd z(f&X}-Kgua3jc2H&hisqBd02jj=~~()Ln@xpW3vGi~|8xK5}EEhDMl$mc@oE0}L0n zOMhI{A;^Wal^PM~sYjmiWXIoZ9V=rGA59_{J1 z@HfxhUYyiX^%D=>fWHubwc|QUrq0w-q$n!1RN($4+d5K}4h~N{0|4pmanCor6*o z1O${SRS}g=Xc9>jq=qWeOQeJJ4k00T;yE8N-}!#`-sic0-jBcbGwhw2S+mxfwchp4 z-lZ2qyX!d|Z{{b%3;CKZ>@x`oNQRySd9&fNVu>#@V&G>FJ8ZQN0U(07W@F~jM<89U zoyu$8Z^=AWf^jIU6DVDmMO%g$n$EfT)G+0X7q7pGfij}qh7do^K&DJny*!S^wh8&x zUjgrfnVe#6GI3rQR(eF!YH1#Jk92P7RswpiWF7nCDN3IwY@P{UDu3?w z!BPi%F_*elz8LFp9vAJea)+1KoO}d6;Y;%EJ}(rlvVTb0{;b_QY-48df~K|N7I@*#ln%y;rnt**G}pbUg7Ft2s%VS+$NF-c#_ZeH@Fe z>U-xV79VpGLP0GY>Z8YV!wa#-^dL6~ zgP$;--Ol#D`CPRB>w}3u>U*GyjJ1w<)7Mw6+9Mi7LUBg^IyGZewk$hDmvk99dBXvI zw)U$zkGbG*_;6IDbk2@K%IS?S8}j@<%Z!}I`hjdzNCWwPLZ5cRvx56jMYgF@srXvw z*_D)jLnm&NJ!04{6l;m^4$Wkso{otT?hX_R_1HFPJ>2S3+rf&;lea;|KKnRQmi-eO z%hT{PmAI+~v#?0JS$oxQFI0LXd)Q%Nv(fZkXhO zJE+I@Q=gGBOOZSH=%=#&r>^BmyFuVq?aK`T(T!ski#|EnPDq+{nT z4GG1cZFNC2`IvMxc9PBawgInPYjLZuftSO?}_ru&Cf z^I>@+t=Y`oG~ZQr3wG(`bcYst*+~V&AMo08GHG%N%bbhIH{@ork!6Rv)C!gA8^qLE z_%6K?2|}P}gqdRF_X6_ z0uFsYi-_nlNF^rd1cxY$opEI<5wzMPrg5qqi^5#+48#rV_}RkqC;b;bE$VpBb;jc! zj+iZYLN}-C6mCLS)BHUxHb*8EQPZ{^r;ZZFMY@CdIF^c?Tje(-YzG}t4%PZ54otQ) zwwyR2o6Y!)bKwK~qzJ84TM%nj|xTVsvKpXjsPkZH3ZREG(7iTm?vh(G4 zcu0_)(Mx=-df!GQfUnXToL|^}CF0BTX-Uf|Z=u|MW1`h>`L(COd$R#~(gBeV`T6H` zE||D&S;3Y@(mk;zQ7NB4+`(jH5RUL;^uG0vtG9=?^A^+T{UyGqM4E43k9YXquz@pI zsI%0Jzk`npPJfJvW|N4CgtMWhXGPb$xBS^eAO5gWnR6K`Y5h9K91?ZO=6e>LK-MrZ zW4nt1J-;+;wlFN67?IDCangyE9lBI2{5BS|xye`^Dg4TZc9e|3-p@*5pp##1ABOI< zR-YaOP^NBtkLDg6H0el5Le06Rw`q{MmHD=^)gO3MH{vDOW@B87Mo(GI5Ukyg(>{l@ za-ALKGZ)FaQmFOy^5F+7LS=y=M4X!BJIq!|e&9a(PdaeFiLcWFI0l97k0Rfh6(__Z zk8*A=J|tMjUP_JlGd^=Pal#xIFZXTXm5A2RiKQtuMR6a(S6eu7$Ay!yidd zMJYCq)P0gIcqx^D@FU25bUlin2$_6=b?oasy6Q#oS*c6?a#$!ieE_^ksDD>4XQx*rhuWmm!tRC4okBhw9YiXA*`3#|{y(aU%3~G2JErQk8#D z}m!ENEM3^#M)6aipP$6F6#!|7c3U{k@-1JzEPwo?m7$pE{ zEu(Nm1#-Y5@!6kQleJdsjxkd)Y+Ub!-yfDY`p+lX7%HwJ86W` z{IuWdYtHTs0%4L}k2Kz)+{xy#s2FmLHF!oez1Kl?*%Xq9??5CT-r&V?1?*St;NOE4W=LNnUISqHMbt1a6RL(bf*7Gup>`X53FzQ8)sdLon zc?;^~q_P~k;ujbeRhGp?La%?T;{vtVylh{NQ78@@EqAT1?RO z0+yfj1&gvuPw*p^L`V(-uBKt%r0NWAa^Gxmzl?4?)y?7sK2@#qB6RQ?4mubf8QA+` za^x}Qv@$7n?c<9J9X_Sqm|T-1G(4zSPb6B;m(&EbndimMKf{8x687uu1cUzBnF*Omxp+ z@yU|07%$X;N0&Fie5YUk?PW-bNJM_dwYO|@4}4ypjSyQ`t`e+$$CTZS##FF^7nGfE ztvQsMGyp3tRw$D`Qc$y@DB{UIYEen((Sja^`x95cr;Kd#)n0k76(d3}X$oT=E*HA2 zOgnjn^cJzZAI6lAWvwuk$nk*P_K0N7pd=Q#PC> z7M=n~0Ow=7yj^W)fblj48a)WVb_!!31a`|d)VY9TvQ3GD! z(W=ha9uMF+F*S~HkY^Y%MNMxuzQnA)lttg*9I0~Gt;#7#fBF_ip3epHe9|8)#EoHi zq2-tTW)RVccp~P&8v)odyLF6et)rX(P}0~coX2~NA zj?-{E;pvd_204eBUy+gzU`={`_lV7)@ArB8OdwMkg)3L3)@;pJe#TTit<%As7c&@4 zLRZfc7rYT3%Q4YkV5wE00Kmk1GCX({_;0H!`|Y}rFN4s|C%DBAuVWDLgI)2d(qaW2 zE7hMXBNI&J#K1PbGnSy5kq`-1?zHJK9$oQYD_5ioDUyxIR}RjU2nQ?2`=Z_LYL#Hc zbEdp6r>dgPk7Q&XVXWclQl zD=Ty{gE=HjIMD|OotrP}Gqa}grIXv-L;mR68n=xv;VjuDZz|X1Vhuqt2EN$*F7GdI z)UO~%JMVepY|&YRNQCF5_rx6r zSj4|?=J=bDhdD`8AvO@P)FQB#)4r;j*U@N6LT0rjfnBI@_%UZZsRg`(v^&_}bMZ{L zzRzejlw)$RqK(<5?9d0z#sX`De0UH>u0PpaDnRu$AUus@tVuqpmbz25c=Jg z?39AZ9{F>N$kBb+)JxKFpP-g1%a-e+VxC!FHJ5i1777gAG?%ZZnAkwVtQEShJU=8q z_T0JO!EUC7v^InYI-5dxdwSGhp)yWJ0^dMsf7QOD$5qHk zfBUt`c8+18+UD>UF1ajG!*3^S@e`a6*NME>Zc#o`1P$sNz2zGUia7eO4DIdJMN;8n zyiLkDT|*y?)4Ad7?<$wwSD34N!c+TLT27JltJ=Om>0a*S{>0Xib_PZ9anqb8&HSdD zy@vTrH1yd4`)}~KVg37Wu(VXUH7wmJfBxdbm5BUXK^H%?T#r{}%2Sr?+Au&=gUo8Y z7c+OJ0q#CD&}KH3SNIl6%lEWs7&z_~J}_T@?}n$tRh*ck*dPWS2IL#s%vQ zfR#f@Teb_Gu7X{Gy`L*n84tS4l%8m5zvQIOf!##b`uJviV}=krOjsk~w4F<$gl)Kr z+n7`HICo;$nz0zSr52Cwg+fGSje5%c0;kBuDsC`7(uKRru7)MIBthxaey=qrlT((=Dx7Rkv8(5YB!QNM zn>7gmMv}Bw&3zOI(QA~e>>06*pCuc(>iJZrpfZ&Y33;L}8=T54Ju(e3b1}5fS1)yi zs_L;%ET934YR*h8+q* z!J$%(_!?Fecgetp!A8azm}GU4Q<(3}f_S!IBJmzZD@=P}D|>7nIhJe#e1#a62uF?{)m&4n_P6L3-d6b5N6JwND!2R%L&|p12_+K|4#R8k>IB z5&39{rTwWJh7K`d^~K&=~Nz5m}8Fb{zm>bI{=!oRC8|tM?N<76(?C| zKoEQO{t#m5qp;kMEA(^?@_W`=Nv02O0{nCi{_{~O1!bR8yL#3G92o)5fsmCt`TQ#PJAfuEMUzu9XS4Jx7c>zbtm% zS~a3&wp2?~Gq5weI}k@2riV2|^T1k}1gb)ryaNLQ0(ygl&M2i>1(-|MH3DFhWuWo1 zdwNlg!kU_xzM9jK&STvnKc{k84dO;{;${M|+*~qOxIjR*)ekW;ywPD=XK-KHs6OH{ z`sgW0IJc6;#r;114hO%{VCFG|QVGrt~qe|z()$q!>OU-%_z zR?uHj1J?$GWq*VG{?mVz{Jm}fEz5KMFURL!Y?tHz<@o%U# zI8wNnQ^9~FCv|e}F%@GpR2ZaUtB3D%?y~Fx*Z_9);H$x<-4^o2U0%SuJ5>@&q11c@ zq(}&~sW_hM_lyAp(QdE_LQy=gYk*;Bqn!Ji(scOv)O`R!rqE$p7|^7xb2=9E3~x+>V}nDk|->5z{Ala)!3E0IEQNV^UmRSgX;r92AyQZfdu_Z#{| zc4rsB0Du>eLsaaN=YCy~bd{=RIT_AVTGJ3fFkVvYUfRDAHv`NW0M9h@-i!d{2O;1- z?i9jm5Co7qA9_ecQ2k!T zIlxKfxsIX`Ds``c-xI%fCzY~BVCoic{|xNaJhAed^d&%q4GKh5x21gra^a!N%ll1X zeo;fO^i!6IEI(RspV(D(Rtjl%42&d}@slm31$4_rW^-$Pyh9B(=va6#Whu!E#|*Tx;>mpjiWnNfQ1nZm{G_Mq zxQT}gCK(rMMY-?rckn3N;olmLbG6ZcPAZRE`BGM?4QQHClCvZ=bN?Mswv3~3wG@7S zy*`lW|3DMZ{_B7O*<_VYf*x)Ls-LCOt27lLX1YBN6dO!y0{2-(nAI)7`?6Bl-XzHN3wy3RW3q}Nl|oVMYFY!`4?Lp{Ek$rnpe zBQW+fdoU!|tHu6FA;AG==F#WOsfa-8bk+|e)*x0wx~F<%Kz@|b=AU1!9DP3Nj<4Hn zgs9h?EXWA4*11SPz!n2^{upCl5%-KZVrfg^Y_n``yY#Ok2^)RVys+50rObbF+wFtB zNu_|B*;Pe=u}nzvqR+ptKP~{*WRP66n;$A8gE0h@i%+IdKj~?jRPLg zZ_qArc8irgR&B+INj43W@|kT!gj=%2Ap80ca-Z{OtQM%0L6;T?TWdtu=#w`~A4x?$ z&%U6r*xlgUSF`MB)s3UR{TvJz_VsJ$qedrV)RDCej7BuH^GkhWEk!DrVCm+zEF}mpO4Q&(A8peDdeC)mw2;^J!|3{m`~yspj+1ewGazR5^De z8Sc?ycdm!9PnbKO;;ZOP6_|_xi-YF+E;c!=FAC1S!A9 z-2C@(??TxKqq!RC*q0ACe!4`swTg$yul*#;sHQ!W+K<$mfbH$V-V)jRU^+S3hRW6M zIL(1AtmuR0_s9`FqhcR%b-h#X@>x;i1ETKLb=EJDT)qKc%%Bi}6tus)H}=7x$#bLU zhhsi1QI<`?NZ3t&ogKR z^?yQy^d)LH?hZ6muM5y6xM}m=Rif;d^-y~EEeA7MZIX7_O_%U(n8DY)ap5>7Ny7TO z=XWg<8{^fX7E^!f1U8Ng{$$b`Aa0Mjgykn9zcscA>mf3<@;Fe#jIh~QkyfLws)s;8 zW;n-%!41nY)ec;$D;Y2B`oT8y)>?A=XQ!)Xm0;4Sh7gq!Jesjg#CPvv)cP%_a8t(571pG9U3utXIS!xSN zJyxP-+oix*24%a}^-7EkQlB`Ed}!J+kiAeN;fBIEO&V@!bqu|B6?7P$pU1?Jg*4Bh zu};~p$eT$G*t3&Du^GNY?D$v?)70!d3@zn7k*I^hph}0&%Ep`=7fVJ8P7hR6ZIlTI2;jI`Dzko;`S8mUtZlh(RfGBl+%S zF*^EtiNNd*Zuq#qt{4{3(VSm483Yxi_SKfCr^jH`<2ngOUws;Zv5Pw&MQv2Z@3jp@ zJ&eeFlJqSv8GioD+v0vezBt!(OXl4y^>MGGNl0WCvS>}1Uzl+FAF59zj!_SeWXWCP5CL305rE~pl_DXD7#`a#z zF5iygjVW(2@+#n@yUbbz*C*Jp${)SmKoyJ^)>S!LpF-_g%gKscXymhKpfZIkQjqe) zc2`Xu<1`Yr{SVb;Co9B(gR$~Lw6k8y@T_VBIr-al3yw{#uUHIsazqPOR^cpCrhZqK z#gW4aHBDLk;HyC+Z;S14VPpZAOd8%OI6vPd9n0-_6g59i*4m(bjl$IXku!NbXYS)) zZa6RpwkzsjIPLP@388V_F@e3&_zz1Mo+G6*%-RLbCG^-n=J+*%g}>{tgTtRA)X3Yz zH$z3Anke;0b9uF&p8gA4|KVNOpD{h(Djl2WblMvck(r)2qyP#AF`^m^-~reW ze+00c+ky~IA4gCY{e~ZyIQJQXBYhfs`t0%AGF-~{%roLzho1Mc4rDK}f!+GC79w8g z)gP>eez(E>F5zz-Hk%n^_{1YFO=!Il;Q}!R0(B4D0cn@w4QU`6(26?lUcNqbwaF=-fDetxb`I+nry25Rotiex@+bfjJp zUGjuRY>D z(!@QX2RV*>5un#oxlnUbS8pKGRZw-{h5H;ED5qx*u9NL2_S+S!GAb{t%MZ5d_s-;~ zL$rWk6yw@}%KeKE#yUaUm4h(Mk>T}Z`;qG!TO!I6fKD&i^ut|uH-B#+`gdI}F{|hX zaJty^Z)7$l|HqGFru7_o}qWM|18OVDZ?Hnme*Qm&* zawKT}bdsBzMYLcxPqR-E2nj$m^uB$=lFw_)eufPMN`uD9@VUnEGBZHm1f^AWS6k|k zU>4}9q;a?3B_Eut@|4LfsJKfjXLWsN{?Ue9IIG5Y(wvExu3>SjH!CQR31?SUH}QD~ z1b8d*(at`!e?9Vlu2eNyadTNGh$8YCfcYSs?&-7?K1)O033tsgiB6cy`n4A~(TOop zg{n#sKw*=9(9f?owqdG=!+98IglT}nI>{v#rU+_8i~LouQb)5ff2fv{EG0)PfDbDA zFcle{B~GWUpLkA6t^1RZXX-#J-=x4km~+jHGKl}2By4xfL{d??f5uzQ42Jjlh`6hW zp?ojb6(Q&KwOR@t)vb}K=JLSR7fGYP|1%eL7=fPZ;Z^_aBObVu*`ZOlaFh{2IujU> z8za}Np#I=(exW0wYb6?mRcHg+Td6chhwRI*%u?*s1z#|58|%{TeSiURuDe&~raWx( zLmc0c7S^|M35}*oP2zkD@04+2>`UWCg|33k>!kv25hYr2{lS$UU;SNM7i5%IEv_N?fsXPmE8XQZLF|U>8gp14~=!w_}!}1fIg&@2-4+Ft3 zm1_->}a$?Wt(-^I3JL$@a@u`tR4Q^pHo8V6~lO4_R33` zgYUOEk3Q*DJa<8V4=$`Gw9dM?R=V`mb=3=Yy3*5*Q>A}b0U!kN z{NqyU?l_bZfqvfq*(XJH-+#!K4}kR%u+Wuwy89^kq5}M2m)k8_N}{OC0esN%*IPPN zJG+G+Xx(zip#U0fE1+X}|HFU$E)^vmkS(uWjdzRQ?d9*&-o2r_E#!+4;CI8QPCWk; zlOPLQ8X!mG`}2Q0tStt5s5#E#Kp8PPpZM$Cb+P=e6ap9kCh_uec?EUQcnnPAV9=4@ zc4%vW12mr$)A?;__w2PI3k!aGjUp_iIN+Wt_TQJHcG&m8?R_%p#PsgHQ2*`^-a2(UcS&q_F=4yMAzjc@R~&4^T!z z&UN>LMdCS}AxzHM|5||SAW{_+GTNp_=(x!U-TBjf6c>|b0UkX-^JIeZC^_%H09xM{ z^qr>&`T-fCI~(D3g-U`nK>R|paLQ5bfE0j~%}g)O$n8D~F8%}l_)m}$OzC(k5R8vl zhb3j@62OqHuH|x5T7W3)Z-UvW`P`uhkiP;v*?#toFSVe_ir~oKcW8Go|2pnmLA?WN zf%#G&AnyUlo4ww#4yxbN)d#;zYl>+Jl{13B|p%`R&RE~ zZVUOM2)wre+hVLm=_LcSgbCk0tw|+k2QoYrV9iL8LM1maX@RIN9ZJ*w!>gr03Ql#< z1&Ue<@^isTxguXudH`eBWdQSt-hWGqGLJWBfITOzp6&in>t&B==TYOk42d`&}IC)ND^Wd(nkLnZkQ64_DRPQQzaO5gto z0P2&lj~dmZ3=!O$mId|!FByUiwFjZ<6?{yIW!C`zdGR0c_J7=TJzn}CRlihPu;0rV zC|g!_{_jXK^+fgnxkc=1`#&%Kvwc!`2}A*>?w9KS{}siqaQ<&Zu}IpN{;1llFH5!J Q5Adh1s(m$2+2Yav0-xBKjsO4v literal 0 HcmV?d00001 diff --git a/contrib/machine-learning/assets/cnn-pooling.png b/contrib/machine-learning/assets/cnn-pooling.png new file mode 100644 index 0000000000000000000000000000000000000000..c3ada5cff41c3b35ecc515bde1f3a6311f2d4787 GIT binary patch literal 79460 zcmagF1zc3!*9Hm*Qi_CtG=d5egMf62l!$b<#84w0!!QgWO1CITw{-WAk|NyzRH=p#^-8TgD~zbbShh8=B$VCb)lP7VTsgYa-@gJtwevga;gGxOInsN6 z_Iyvu`o?u_HUYYamG11%7PFF`h_w=WlvDl~yN|c$`8y42W-Q}ZcP#z(&miO;Zf>}K z!1QY<1bm12@U$SyXglrtw%Ug9GsGJ!+9r4u7jk73`~b^O_u%P>H&(C<()?|+Y@`{U zsKz7a4&qf|nUKO2VdW4s`-de95z<<$S3eFsNA6*b@yMj=w+{NzZZ(mLh2gRAV$r?} zhPk@35i+X;Oi@k0VUF9b9Pt_IIZLvdPJb2^&l~*BRsEq$I=kwZ_M}taAIrY+v-AkR z2OvG$gpnb=^JFum15lMH-h#G6VyWsEdvNvR8)kC$oyYR|2^B-QITCgqgUhYS*b>1h zc#W-2!4b?PRO_My9x>s}A#XL>t`cU+M2JY2);%>N+xmaVhEJL6yo!jnmm2ss{Y@J%7Gu*dCAS)gHsMGmfN%ivxxXbV(tiP z=og9mC}zQSGbd)11|r#IUABpOfFUgEyQz6yfkp64qf zK0i|7g6|*66hj$#8@oQf-e;429W5jbcs0z4ix+(>`irA9lG%9JF#S=?{nXdV#BI_f zjq%UrL-3oP`m;roUAK!p(2)!m!qDL8K^Vz z_rb!JSEiov9i%zsfu#(?u+-Ikaec_MqhwG@LrM$#Yhz&!jpt<9Zhe#Cwl zv!zR?cdpqxYHMm;td)7{)$MKN@TG>28`VDRYuDI*g9kS^H|_KFWy@GQCw?=xL*Rip zJShe{tYoo{ctzr`^ne!91;f{OP%rPi$KmqhPL3jW!bi3;;eLEUink=q_Z=&1@bk72 z(GNeG_W<_0Iqzxo?sK%>Eu;B>-_R;Kgca!*yKt|L0Ptyh0bnGNvVgaNx7TWoEB%Im z!HoHC@N@G3VP*gU@qwIwG{F-pMR|>wd$QrbqG^7KXFt^csKuNVO>4>HLLd^7tS}uj z^Ghb14ie%0n)hqqR(e;SfG9mMQk1Dwm)x$u;_v!}Rux%0-I^CD#k}@FzaOIzoi+ST997;RP)b z3NxjJ54KS+Q=WM=GCKY&U3zd}az*!1UnO-TOVvBkCc{_Z+`6~fAelQTTi`_UHu_8kQVqHNSqVm5;CHA z{%$;TS`C(WFvhiIWG7M`>6Pdebs?{;+;7yc$t|y(EvQgzWFIvxQXOQ+WIO#5N*g2? z`7Db&!#*n|eeOq7#^~mRo39sq6av$wE_g-+bxe zvXS?ocSG!B?4whq&Wc$Iv6;5TJ;N)u10zLa;@J?*O4YCof!CaGIdwth6C}k)2FBA* z>_NNC_8(8Zcl)PvtC;v-1mb;aU*KQkt$CrsFX|pFDky4{!j+rq_3Y-E{#>^fY9?+OKU ztc#I`GkR?LJY^1LsEXm*#p#FBW7D&hIWrdsm73z3CM$(13j}bcxU%6x)N}*FdRlkd z)a4b!EQ2h`HvM!ESs;Xj;z1#!$7syjFApxKB6shN?DdC>S}S7<$1ZiN`(GD(gVGpo z`Vsvi-H@=OIkIWiHu;AUHna!RoydSajZGK(Cos$XUkFw@g=QCbCrWq;xM{V^iEg*1 zwHUEm*bvk()-=#G(>}sFp7neAnNkpCzd@hqy5j!5{p4Lp$5B>%gKqPrWNAwyLsMlF z*b^L4LGzTp8aQUb$1#LkI41Nr>L?SzLIK@9Wzgf3$xY<)(LfRTYCKpQG0FUxp6p@ zWMlzIGi=|ze?*W(GeNFG97SA8;QP>w=`m(z5s@TxO)kexZwlWyTgp5PUXm{khcnqz zNQhiG-Q#^G`h#!Hb|-j>rrdh@ku$}zM@LiGgJ4ryU!k9Z(s4t{c`?F(@@EGyjrSv= z*^#{M-F@_~yoeY2UuVC+eyjKnkvJkGXtT3bp0YgKMH$t=i?q5LJ6LCXr1C4l$t|-> z7TIx~|9LEXj3M9DQ@zp7#FSR3n4>s;aIti;h4hSAhY%hh*+$uBy(ziLInu!A&8K7K zW_v=T80>YMy1by&=2h~TLR!fVF+hR5{di3&nW8&u$FxBc?cHtdH4 z=RLrcy{Kf0R1+i+5nj*FIAA$A`;s`}n_wZNrgP`KQf&QmA3`!-R0xXNlN*rZ3`Gto zB~HGa_q<6ykMbov&#BRC(lhXO{fPFzPka!N8DJkK4=`xhFbqjwd7bZStG^lFdDThn z!*BpPFTGY8iLc^zH--DqI45`_B36BRYfFa=)a#ZGeh+hQ>+GlGCzl-jzkoPqEv?oaO z*N+Rk?X8Obj()tPGzZ05ms6n$M*+k8wQ&yDaOK8g3+T>i9Bou$B7?kG+(c$|#WX5q z*~0`$Syxy-Q0_45&@`6=ZDmBHu0u7^$SLaPQiv4DcMiMojy66GzS36Ah3!ex{e=4r z5j_#RL2V=gLJ%Pa@f|nM8L~>&Zwl^F010^s!CSrCgSKTD+Y>e~6{}?LS@H67R6cKR zNp(k6RSJm|4E1h5LDL#x9hG(@0$ypHJeQ%?qoxJGO^Y1eK@w*TGkf?mvol*f5F@+B z9Y}pXOzX0o(??guq8ZCOdb30~o#_`E)oxAI7f25(mp>eINLknEfY5izU1WkZ` zivl5epzn{@GGIF~5vE}O4ZkJ(n=KSqN!R^VC&|SxxCU@A(EcG0E!5Gyr zxThrf$b0H0=Hy@}H(Pa3QuX}xg~rV|3N%`~?})x|y}W!X{#^Bd)9+{KkK8-^UUXQ; zmROJWz2zAS#c!VzWr|mFke;dPLFBa_g0Z$1ZlUiUTs71awx)9LpW&SdbYYV!<1-^E z(peQqCO%dsE*U-Nk@X~*dUPznEp-1XDO23_h7>)>9bE|bPn+?@wt3+FqSTt+9fK%B zfqIHo%F0+Am@)wtE;cO|9;SqiIV7>^{#};GX2-ho*L55$tRPz~+<(?l!JPkmqA|yx zHvc%^`4WhQkNJ-rb9koX{Hyleh4efBD&vG=?qNx5$|x#g&YBi3K%j%GjicM^JExJD z3xrN@^j)#AsGj~guoX3)>|^?$wAIpc(^GyYYT;-PFtcKhxQnHg zsD`ZkKbvFzlVGuNb8`~q;Q@oe05Ctm(Z!mFS42dF=LH`RA0IcS2DdB3!OhH*+rgFf zAD#TGA6cNQg^R6|o2{b*^Phgr%pKj`Bv@Gf4D|2MKfWi>)Am0jIk^6FT9^s){JF!! z3wXiv@4hij#s6FtRkQU3e$bb-wZ}X&%ovh_0x!h>s{jAH^Pds_tEJw5TJj1B{=b_3 z*RB6=Q*BqEi;SZ^W=uE9|E$+P$Nle{|7a%yr`)dlW__^{3+ok@qO7!*C-zPX{`-e=Gu>%V zz79djHb**;Wz=etP@XAH4^k~sGD(qob2>+)V z|LKAYhrsC_?*GvtruyeMSl9sB?^<~O*Wg&#xWa4y<5&C7z^|Y9VNtWZEAaon>*CP6 z{;%`>N2ia;U-=P}J=2!{zw2U0iu~t`{7=vGBPk(sYXCuP!T(g@KTpCBtDF9RUXFha z^(q*bbm8@9MeYA(>Hb;wm7mfG{n^&c^9qMKcwY7T+S*$alhl(^s|ebWbRP-ZiK6_m zLPIvY1h?k>)ya9qpBE=8)S%V%-jar$jgFHisp6?Pw)D>hVM+T#miq7y3FEywN!4HZ zRU{=NpvlQ4{ibDVH)o*4+(YgLk2tFSB*@6fbB%9nW{~=ui_5gb-*kJzotN~kpWSwb z20~`o=mbAR`>qefcPhpG{*t5glt)N4Q3zyyBT(I?H*aPZAQu+@+dm-eouQKDNn-U# zp4lDSiK>vQ55q=VlZA9seMx)X%u-iZh3>^(CtEJ+^ZeN!y9OU2mE6325J^}U*SYn8 z(Sx=Wk12qKZ>hM@mz6gVUryxZhNgaP&9NiFwZ!;3faY;(jZTRHH~QW8C!K4)v>tO1 z{LSh;laF(Y_4hhnv;2*%>;ExY#ZQte({$PC>C73^+FxgUW{OuAkP@*qLvd9i#=D$khT;BUi;>QyMeb=c-j~ARuzH(f)eVa*~gg2j)=vR4$jLw;T!naJe*>Y>fVIy#&^wU+8Yq>$;L#{U0QzpQ(K88^z zKMb1Yye4rIgnU27vQ6}qi0!enr8&73ajBP{Uy=)C)vFciaBuqFASxh*aVmn8M5RqW z$wjV`9=~FGFVfg{r>i^;vyT+b%woI`b58kh{II%e!Eza_(_Lo(b%`q11sM(2Uq z;w_Y*h}A25u=3Y{Fe(VGav)Z`jp3fpetF~R1XQI34<iPlFXu)T;7b*H`U*V*ML~ zBF$sR#;pq_57%gB!AaK1XPMpSn#^Q@Y|w1zqlmH33u3Cf&kLip8(5dIvfI zDK*hygJPp0)}w?tMph9XW@Z`Tv_eSdl!7KECo|LT>uk|@=vjI0O84!oXcrWmdn}^k zwNnLaL==15oeSO|@hwMxX!If~IBXD&_4gP(Gl0*n)?qiRLR@6Mk+q1%Z}(`OG8mL+ zmcQimt~BE_3B2w#&y6B*A_?Vi{<7g1KxDqPHoS+@H2cim`YpFuyHK}xpU!~_7B#il zSgG=@xMs&rA;sc{Wc5!vud9R}oTdxenl=#EEWC+`63JI6CZ- zvFRN7h%Joo{8ZJ<^-F96gSq)c*@46FxH8{sVozS3vvPto^P7FVz$V!d}G6{ zv3zcZ>FRG@?%T3o4Jti~%Xg+oV~}OicSf55`0;@_mGeW>?sz$bw<()A)eK7*Z`SzR z@@kHv7+O7g{5?L2O(=?yakx1}VeVD`OZxclm&WDhIjbcaG}U}M+*@sY!|8zGwZ0Ud z-Z-AxQ&u}ujk-j#bdYwPAx+RU6$wcU8OPZK<~K&QkANVTd|d}sUnrJ9n+ zeYnjnddGLwv2w_yQeN41)QcZ6J6r0D^5k%n(1l0Aw2g*pWxxUgC|I+*wczi+x048Bs+GfTabvc7sB zvX+8qGNc~7JjCp9-*l0l-I}Z8y7#_XDJKcr3PrRCz1|n?-BGM*LL~K({Fa{^I!}v_ zjg&I*m;ow?8TKvnv|K-@U{gL_Q3o|X?xG#Mt~qJgAH}&)VHEigom=GV$iWZ=3N+$S zrWHC|Ar{j}t9gQcW1jc~r3_IX;FV+H; z6v5HP+I7Y2i<4}N%bK;6iN4=D6XLbo6;X+%Qn=JQTBr53k zd~}|mX&#T5nkKBrc1y52&ISeYTr zAFEfTK^4egNV{fNZyV(4%Xy4G)nR3H23s`J+Ygp;f z_Jm$-QolQINq%`L8&IyPBHTqjUUeV50lq2XhPK5s^k7|3w*x@kuBIjcWJ??{dMRw29@QNod7F;sbRRO+jDi3cwp*m>(3p47P; zv8@49(H3{ZuMUJ9j0?>v5~fAb3r+=VpQO0d&N`>%ZHtI)h?#Gs-*}9P7Igkzda7i( zJFQW^9=NFcc1{!-MRg1rK}U(vgU1%t6B>>-&uiyz*-M*FQj3hLes>S!w8#7y9cbDEf=u(QN*&w ztBe^yl8S4;A|=oQ=KA>?MG$|V(m7zS=)8S=%1ZWfluJeDr6I4y+WdrMu*%OzY?PD> zZ6B1-Ay+@ERkLqMqG%x69rO74J^+h_W%9obuO8wkk}ddLU#lkUmrcT@(S+)SLM{6X z&D$oIC}7bN(Mnipi#wY_a+=$$dWB(Biuk4-w+s5SB3W2W>bsI)gJ5IiEsnzePymw7 zv0m_O$ zLd>R&9?-}bBjJEY(uj^oxN;X$lY+k zqu*wsXOZ(QNFYk%d8@?fm}?vSK69C?%nm(_ch{vty`oA)MuYzdNBq0p_29-Ja-$-P#t-}bs~1l zrJ5)|*0hGfhAD2<9*mv`^(x|hv7vo(*U<~j_)UN-|BE%`1BWjNFenr>&)r|@|HZ)f zLMF{3>Lt^=9j@}VdSiJhxEilqw}?R**W?|p*Vi~ka_fyXOjVA0RX+JoR|t-VduJiU zE2q$B^4?_>#YRnV_u})>3OQFcrJ=oQSKdg97A)#LN4aeFI15;Kc1fbtJC;5px$aC2q-1*QsxvwOaQEZnP6Z&_3 z({L$3ZH#dnvt$^U2-LTg)fx*;1OLKTQKo*C+dSJg`phIv?`Z3|eB7kZ0+UetT)*SB zqb}jT(#~Ub2VEoP<*{;eur2Z9AZF}=G&TbxxjIn9ws^I;bI^-!jy}j&4s3wU&&hta zX~;akN{oHY)>=J1h%xYK>$!0$M3j_NJ?t6f{)fqOGh!z@n&S?$0ASf|%k> zeQw(oC+9=Tf2az8#+*jlg!4Cy?kV{Z!>QeC#kZy($H@VUASFO;%W-fboF1UMPwYgQ zax@(ziz|Hnemz*`!B=%5q*aliQoy7qD#yLB-*Ue-EV)R!WpbcXSmg@GsIi68LH%6- zN`_>1zJfz*TlY4I*Nn)C(tew1x!~ApnXr{1lkg(n3or*;r2n55oY-mWv?vS_CPxkT zJuh}EAA2B+emY$NqVBEMir#)@>^B~ucYljk37zMRPiYBHkT7?P5}7MpF-E}rGW0M#W(zHUe`iw$BA%0*m-sB6{1MAMxf-ER-JaO)$nJ-S}f`(eLIcI9&u1X z-#g{F!jyPp1uEz#TH6%2$z$+Sq=nEjOZAuw=T141D`U=EzlOb^-7H4IsCvT(2sl)i ze_qb2|3;zt92G>soAIf!0lzKsv$JuYVmrS~d^O9!5N*X$0>pMG` z)8ISUd}i=@eCXUd&Ed2y%q?~U5WfMDThAOA$7h7JRh;UO^)YQ-*Yf;;r&n3f@9<>rcYw;4TFEj3cC zPF0|pX|BXHP7a;#)$a9s9~Htk`1)f9cU|_iEqYEu$hx&`D%;KQ8JOXFaST6_nsiVv zCO{~)Q3Bm%Z)kLWu)siFW;i3;X^2NP=uDWG#)Yw|Mq}wVdAWnk94q6vi`K@0b9FPQ5Yp0F z=$p;Pdb^Gd0+4jImH6aN%GWg0(E=rEa(Bi7LBPA~pG~jmsvJkjtQ<41(K9oqeJ*qB zVCnn>y98LXn{%UP_o+_8$@Zk7#VQki3L5UKdwol_)J|Vmwys`)TL)5@d-(WkB%|b* z#q+$&q&bgC++O3{0dTd8?dT8fj?@@nzYL{ZXm~;H-t%fZ?(z@)yu|0DD<2hA-*!pzM=Z()>r*l|JXg2Oz-RCe2v~ zwz-t>XRY=iAE+0dZYJKE@nO4ai$Pn$uAtC__#}FHXfcb`2vGPQhSbj$hLUgHBwXAa zMAe`6|HGUIFq_YS?sdPdGsZDW!;}g>!!Zc*W9+% z`oswD%6?1mau?h_i5y+9^+k8Sd-c9@cpVZ$>K1wq)yZEL2Tvf0OB)|E`JM)>yU3ya zjP!_rT^XT6R>nK3}&TiFLSf2#7;tBb5)XU{F*BjY?VVSdP5cr=gDR0JV7Tx@OBJ zO1z>fbsKuMSLi5iFW@ltt1hpxgKT^!>?dF;9LtKT$Ni^zF=b#>vq;;wA5GwB<0#lY zmpi#_^B9L?WgVhq6jsgrLHFU4Rt8K|-^Ed+wKLj?;S)nJd}4zK@AgK`BQ2#LAccq{ zqsaqCFNjpPdp_(hwl_`;qXZbhxEnlu{Zv>YnQ25$u#{}2-;v>ltw?_%m9CG{cMN5y z1b5@RoDUG!F&kjr^&MK;is|A`h6fSzZ`_2tuJ@z0^20Xm4o3U+9(o8C(P9{%&&4_N ztp&RXp<&HmvP%g;H+=9{zTL0~Ua~%7gM{7mW?Xt6! zQk6Ycp3~fJ(4Uh5+e~&<*cjK_MVvQes(0+cWEY8^;r6%Nv``Kwa+I=D6*uc`uG;tB zstgOhogG?(CK4~jgs+c`vXlKCRTupXSL0bLUOk%z^;Zfu36yjO>VdYFOHg_f!UOrU z?pAD57%aGMWNh7^)Y9srlxv=7;KOC41qa8+4kyp@d3i42tj+BCK5=ZkZWb$CH$q#F zN?6)`ms?C%Bf~Cf(tojoVXaer<}6PQ4-6m6QuI|xq3`P5pQdtexjM~Vlthf~2HvOD zgQ~5rO$+kp#DBt-=$CStVrJ7U>NFA46o>d@HeUz(ZJdUZbdmC9kC|UIxykQ=&U?ic z@wSx^PMVhsuf?`RBAZSCpn*DyIh@&`u9Q_JNpKu$8|Cs7$0%~vtzF>b(dsZ;+PA&% zFfDMITuRxW2r&Cv`5P8`C#;kf?QhyZCYM|&CPx`~l4gLNPpPLCY|RMo%~uj}cg;mn zUrTc{kHb4jE>JWtKUjVy?)6!NoJ#4BT3<69ME=QBbit64R`YGZ%r*tcCuBo2>(Sxc zM@G*5>`?)#(Vzhv`%UlW?VF%*_b<9yV^>C6Io(95U7<-qsrdSzYk8s3_u(rzz5+Vl9UF!$MFsZQ zfFhC7`*IWu{yukJwn)b6nu4=Z2_>SP|6l(6%of( z@A^`uhg|8%7S=Xq>Q!Bs=fu|*GwXKgnmXsd1Wiwk8eyE4YL@KZZq|vUZxXQ9kx5<6 zV0X^^*D+!nl0dj=g4h*?^;~|ouMYlFUAIHM;FhuOYwF9{a&-~3`lVNs`R&RCFKddN z9hk~kt8rAXvF}CLinMe9J*8V_z#)N?9b^c@-T=FIYa%vAhp@n*HF6lBKgPWdWrA(h zTl%`+^sJly{=<*guhrVPnvO{n-o4*|d;;Ot=!&FY<7(_S*s0zQWy8P&&9GlHj*SB) zZ6bftczlayrNMQd$QA~!RQ*O_S8h^L;3*8h+k2fY)-S31c&ddh!|*-v5+nh!Cy##g z2QR3g-D%zXNqu!VZ`Q<^ei`n;HtK&+S=ybDo$3XML{+utlP`=ltuUY^H!9zsFdxY# z3fOZ5F?jpsJ|uhZ<()q`@W! zQuY&zn4mk|@xOP8idu+jdHe z*L~=Ca|ZIoRX_O6xP zT2-V~>0rO|I7Hu$RF*U~j=Nbo{5-d$gO&yjIk}xV;>;TWl`vV)QkEPV(F%H|m%p-I zQmkZU^1e!Fyt)Tsm^`OmF1+5G0WJ{2neJ*vqRKG^G;9q1CXyoPm zxk0d)`HkPG#6?$l-EYkn2EwgE)O9XZm|2we_oFwLX&l%pGa8R?q$}e`f3zvnMsU*4 z&JYMSHMJ>=AE|LpN|5d)oCUv2;^oieu4di_BNG@owfi4h>OWK$miNEDAx@;=4aGAV zf#K2#<~bI9o_NHWQJyO9A(kp@`XugbqR5s)T`1oy#Ic-fo+4Jh>#+dXd#5S5Cd3~knwGD9M#0XG99y7M2pl6lR};<$7&ERo zNqijdHSM9`Hp?b{RadQ$B_=Te!O#l8UqEWQ_$l||4;X}5{`<1D9fkWc z+~Zs}a@BsD@+fBTvN_6oY^3&xib1?T_ptu>IkPL_>YdPgz}-8O6%f~D|HHOm zJkN$77&bvEqOj)Fer8IaT=Z@|R>QHmW}34u%|jFK6Xjt!9q`5+$W)}y`}}rXXug+X zuz92TiF90$N1(s|3E#@JI>toSR#txZcjaV;2Gg2zpBm6C^(|8`WD@F1FBQ+=t)iuF zxT3^w!sn{ocWh)Ss;3BvYUc!gy6A5;V@$yWdSnbix2w*^a|>mpc!xzKKE_zYb)iG^ zqJb(z>~Pv}FF|5gvIawA$+1#KwJ{dHSdZ~|`3)zzruM@*(Xd3@7+yY9mA{GVp<8N! z+74C0azY`;vZ~La$nbU8RfcWsGZ+2aBd*0+U*)-%k(flXZ-rJbCq#t>a7GSyM08$Y zJ$}P{0OaPkBX*J|UPux3h`ymt#_&)Apo!mSP$q(^D@`TiYK_Q~pRdM3?*MT1A#s+y zG0TynV>`EXpjY|7vi<#Wj#A-2T_4A3$Fo_8uu!IiNw@f7U>|UolM0kt*!kTcvRctq5mN8kwCrltMuwIQ~#_)aDL?@8z&GmV%u%bWt`uI0{H^FQh=xJCfFs0!yE6E%; zG*Ra=fR&H*&gcxyKl%eDy}z%_$-EdJ1QtP8Y5$J*d?WBN9Fv-n-o>PEPs*GP$MAOL zGAB$F9#Q*5L;w>Q#S!5JQu1H3i|)7;1cKd)_LwAlHHZ1LtXqxP6<5GAK> z&xsI{Y7seviD3y@;Pc6$J~yOKT5VG^;?ePO+=-W!z3ovFz+LH;W~Z%??%x102d0a^ zHjtKls)gC`wE;tB48QS$%C0he6~U!G%+8+d(#AVtvCfdJbkNs68Nq%Lsrqo7=ua*$ z-o)FdLsG$GRT^dBj*Q^fNfz<=AskuG)waa$0IVO-YZZ7wLkAtW#%S z-Y|kaEVvP`cMpe#2a}1M|AtBRdLN|5Ct2^(EPHV0hXqZV_{OrO z;XJ7DgI=8LhHtkxEe1< zHoKT*5;*uf>akT&#oA(gSLSICR+s_h$ORoW$bRO7gD$0K<8&}(VIR^KwUDah0T2Bk zm2}Oa-RFSFO!zQS5df6PT$=DYB5O4t?8foPKZ~rduh)TjptcjIHgRxOoaNH zcSPhVm}Y8_Mi(i*abrcBuz61+Hd6IP=2%NoWv0H;^88tC%A-8 zcB{n`Na@8~Q{VV{t6AzRw1ZxacfFA}DuEF_1ez>~Rc$$Yi}bv_4(4$4Rd>g4vN_OL zHm}!8kNl-4?@u&$0WEqQGaqU*9;f>-Ovkk7_^{^0&l&%!a&eEO_IuOC71KfXQoS4c zqwC*-Bo$mf6F);+D(6yO^Dk2O>dH!;#SjCEB-8EkJV;Ti&nv5|BIiEmyQ&GowiOuh z4BUP5gFYf}VV&e;dpwqtXODU@67T_A1+UC7oOJC$alfP+yfg{blf9_09H=y>R4{`{ zc6Ar7J&t!D?_J z^AcMnAVB89tm%#8Ev%S@qeS9tBbLu~Vm_5ai=M`oOZ`IzYhO7LlriVEAF;uZ&U-k% ze!L2U#7@`zkze2b!7*`8`C5hn6nQ&jLAF#SnCakF`(p1Q<)y#BWSG4{?H6kk&+)pR ztLfMr{m~ykTC7E(Ke&ymWV?_q)}QxjpZG&LIIN%idOLNC8ZB7o(POu8sP9ddL*|4^ z*gd7>&ij^eIJ}S1Fv;(vL8cYELs0EDLFq8+U65Tp6?Y!qc1#x>!E`)_9S0g|qKvfoE*V1>Z?Wz0}qU7y2wh{pnOilUGOTs~YEycHyLGT%0sc^5ylz zv0J?(qoaAxnK?c%(pz=dK>kV_TWP)gE+ruo9~FFiqWEZKG#JCbf(Gyjo#69Jt?VVA zt}%2f=dW=mQc%-mIqN6%>X_GWI`VT26(paTsV6>YsArEj2^ZK4fe=$osoTy@={+>j zH&@?yk{F9gbAwF={x;36%^H1xlT7k!h%8I4#RVLMQEV8XdbPAJFxC$28OcHQ4%E*m zClRq&I5zROvB*%z1)5B#N?tC3{!+$NoGM1ol$GHL+iwqCGFBLY5WDcG2wK4^D*J1P zwwF-`xLZE%c4ihqu&GyoLdE+cphiMip)ec@2gw0r^CQ@a}lHzg8`rgP>lfxxNH&(B`tYQF7^S zaeXv!$y6WPF(}8Q;&TB3F>$MQ3A1v9<_F{luA6u%OJ2|Grhcg*Hd59K%c66DOAHC5 z=c>JuomUm2iBG-R{}-TB)Df!Zu(M?^4`o=cM}c#;>mQL4Vmwm)$EqYP>g1IEn=c&G zc5ebPJHa%OZGavy0SIzOq|(EOy<9|k?cIF2=%nVefLSR zLCsf;l&ME|-iK*1Q_#kAI{ot9OYT3u$DNI%_1j6K97ti5!d|I&V{ZW?6oVk0rG2Sw zkQf8$m-$=ytoFB8-$)l2Axrw_H$iE?1?1OL+!H@k1E~vl{*hNg8$wt0sL0`u479Yk z;I-2j#wEgLdu2H)&KQgsR#e5zVGfuy`eC8xXA`tXx(O<8L~QUi>eq^rnvkIk{t0;2 z9uaR5LL0I|sOdNAVfrV8mVo@Gn8itsYnjv2tuG$WgdLd1VNnX4KDJH%cpo{r^7#9FWQ z)c)xj?p$fRXG;}`mCHZ~0u*2sI=BegPp{NVi7ip#$3AYg6@%X*(=>At%Fo=~h@A{x zgo~u*VdcglYiC7%*8Mnvr*Ln}yO9`s9n@!1MfVUK++gAgF?DsWahzmSVeaJ$yd6W@ zab4~<1&Q^D@@&_Kv%mUVt$E&S+=9H-&UqFb?R)KnwR<<9!dSNdH2XO8Vn%Fx-^uqJ7u{94;V#s z-Gl$i3p2;aTwUKwR)0{#UhM}w;bJc zA70GtA?AXGw=x$)DLH|^Ra-nuPWUfj`TKT3%){@`hI(_HvjRP+wjXh|ZeD>taQVi0 z9>Q6-f=xP49 z@IXDC@4#w0Wy}uWv&yL>+}Qx$wu@*O8L!128h<ZN!A7mzAt z!hah!zWt-h`&QQnoocroob&oc*!`7A-^)4BnM@eD&N9@(UUdE=;kp)eyxUU!PfC5P zR|gZ11Tx%V=)VxGhMM@m14ND{j8VgNSGM7*(Y@I5Px__KtI8HQxj=y{z zC)Q|hRNrJ~1V3-bD1L1SwGw?!PK17QRKD0C!5}6*K}YQa9u#puqddHOOm|*|@Zr+f zp+#8aZZY&UOa^r`weGq$P}^}IxDF1BkJV%`LRGG#Dkd2v=t1NSMX<(^I%}i#X|M{l zy}|<~JJXuTgBA5!ovpTBlC2VTw9HYE=X#{E3i?zd9t5Y3_Gu~w5Y}QS1UTI`hK54{ z9wfmcuUyv;Z>*%tl{fv()slRA73_I1NLM6}idt3t^2^oXHWTAdP3q6Jb6!d!iyN#@ zq*{K=t+rY`vH``MiaDGLic0HTtp7YmuL{XW(h4lMNuqcBQQ$6K02dJR_WxDKkjA58xon_r(| z4G%z&C+zQApv*^Z{av$T&r=S9Li^4!8$;N@{aVlpAf>kKH0P*xInwb(3ymp^zmpt` znuqY_m-L=>)SO2gvgx$ih^wJD2*!vJb5wbl3Q%QUIDZn`=%$&mADv5CNsd2h6rs7J zLT;~AxKG^HpD0+(ReL?&ePndeOB(cVF7Cfnr`jQ}G;r&F<7QvhVPY}8xP*Yli<51| zq<%xiVuSN4pC09OIUQ(% zoae|*zC8z*kTJ(8g;Diex-Q>UrkVoSncKhSy=<=au0Xk z%dL>@JFL-<+nmX@(+s@~H zaj!5@39tjg?e?H7<|}Iqt=PgDCV{6R^FSpppm*e(8-64`5JjEseTp2wsHT0^(}2{R zlf2>ec@r}nC4~aD&8}__YV5`fz7zTgbS%SUZI>6hHuyHGU{Pqpyn+9gv_ zcB<82Q&t}mic9cWnK^Yn^u^4(@uRg`lqM&{y|ajk@=T_2FsFsy`R`zpnl=G*tM<2; z6-gEXSy#5alGhmT|AFhtb{J|CeWQ>TGH;(?ph80HWG0)k#fO?!7lO=Mns_7iM{+b3 z*m!V^2#KoiaqwN<(|MVlx-)-x;s>3+M#X9`-k<=-y|1&y}TN9k8)U z8pH9|yvne$4u-`2Vo?-eFND+q)K|+(7aI0tU zIcLV%=j=c4eV+UKYaIIf>Z@9{YOS|ct@T!vL$2?%h~p^Bh^egR#U5kLYHhK+p&%@#@KU*_L?b)?TgDs~r2LVX>7BP8e2)tbw_Eo@N|9VV)?02RcK%mnnQ zj5_Qyxr1B@$*nt@jkp?^rC@^$?>>L4vy;7du8VYv^u|iK(hvqnet}`lK%M}ou__XR z@`{LJ4UXhXO1?7)JJ$~u>YCu(%7rUr$JTG;)7Wt1%~77(Ck-2bG}|B$xmQwYu;SmkD2+y>nXoXpG%()@^@ZOCk{YKk9#XzRH%Oo*!1vBry zjH|Tm=(9bZah??Buc$5D40kBHzU=pWt`YViRn6Pz|FDT<3&JyqQ;dPWRee0RO|^|p z)wS-Q<0$pT&>1-y6;r9ThLB~b%qtF5Ka^h?EO|9pY#?gSxIwC*(6+rC)A#P8ADu*{ z12^uhY0Z>|D-b)+U&g7oa%ziN9gf;V*F$xT67rV=cM*XnXbQ_ukKAZtB}5FlgC!s0 zq8^6_JN(J2xKcSGr-9M8wq2c^se!j;lYFvRML7)+VuuIX!Qa$>%E=;OxFPPMJ+1m4 z)dxJ^M(Q}8WUps=yKiIwF`PbF`x5b7t zI{$xE&!*V)mw$2r{JDbu$CCb!CH=oy=>OLp{pDIbgYvzt1q_&{$SguHIuRiasAAP0 zewjs`U}oyh&7Rd9QUcn0ELD5AY5TP#F8p(`a{~XiB1E^#VTpqn>p%Mc*dJZMd3+L@ zgnSa_*FpH-=f{Y%=ZgK7(LYJEZj{Ss_tKV z2S%BHT~_sX?<(ho>m0%jRV05Clu`?5=e+`B10udWJK*tnSziZNSx?0u_mB6eaI#P! zj-OEF1zj54&JxCrB~Af!w6p{1Yz>;tS3fEz`+XV`R6##=l_CPNsla>XbmS7}a2>V7 z;hB*5cgT;R_jDBGX1H$r)poxq2Oor0%FDM4yakWk45#)F29Ib5068w&Ds}eYey*`p z>iZoLrQzV09uw5K=sx&H-jB!b+_;%~iW4t>sG43pXf2M>*!j}$$TS*9(eTzaR|1Ng z;XXr1_T?Zsud6^@8(V$DbxJ|q!%?Wle|(M&e1+@&htkUj0H-kM;)7=8yuio0Bixqw z8o-bBQg8lA1Tx18q))-IEthr)49FU94hJh=aNP&ue0S+aRn%o5tk7%O2Qm;$B%XC{Kn9cLdGdJphCkI&bN?rA>re zoiFY!P~&FjuQSN92%6T3NIswcXE5L@uBXc2>rv{;xktEhoL+~7{;p}4aT@(ALC5kh z=>J6lZg&yant~Ax$Ae*X@$i0sOPd_DyBMTMNpb+xOFWew#aYjy_IKtWB4H#zduZP!v7iTHXO2?B+RWbhC zOVHy6=Yca~MeOoT(KuNpc;wRm{$3sz7#){LDFjsE=v|?)-<$EI)mU$x9J%8K zBqqt%7R;vL;GX%+vuNQFfBa;vLbOS~SOq*ym?YyP-i}OL^Jk$@JQ^bPchQHEoC<0F zytKW#Nj(V=(Z|u>3h?l2WPq{tSw;aIMoN06L zZg^4oYq793@})f`+0rBK_(U@iS7`zwnIhyOWia*H*<=1b`ZYFnGe!w&E zS-1O#EfatNnJh(^J?K-4@KL)ddjgM%z(b`e9|3lf0m|+y1)tKrI=`rv(N$yfSNC%p zuo9h}Y5}Rw@CiC^f-i{t{I&)-Ad-y?ISwQ_d7|HryuyQt399~8ss5#$Kyx}64i0u! z2KMK&rMzF^p}s~9k`S2cY^iP?F&LewINhGs=4#Y)sd%_eN29WJGb$XCr`DfmntgR> z!ZFr(qR4wLS&rrEg*%%_$RIHHwx$P?+3wX7g>21ZH^hEl0(gXTz*y4A=9l-@c#Ii~ z?;czzeZEDf#Y{OXB^FJ0FG!8c{Psa%h~r12`6k78nR|RhI)yc3BnWaFY?Y_%aQSO7 z$y|iF14r?_D8@yS7w_4@X!}^@gVau{e|c-JCSXk-1npd9CI9;Z0%N_E1SC54en6%@ z_*`n-zRVGKdMb%4rE+vU&@-WO{Wb@gVvgW$H9u8-7l(q0%D3gt+#*G{4Fa|8Ibcmg zN6{y<(G$Y

hKxcQ#`}D(_RVH|vpD)Lh*0)~xvwY)pdB_bB<%`QN{51l$&1EAi0^ zO;*RyNW+sUg^A&CtJd!t0)bz-xFhgjwO?}+^8Sl{4$okUW3lYqmX zj(B?CIY3aQ*xB0~#mOvV783DX!=}gdDTWcB&r|*l9?Wz&e_Ax&iHHMQKeS$HZtqOV z(_DDiW7t)At0XqRKj3o?6$3s&f@+XYY6^QO=&{?+&E&xnzI+bC>QWLn#^2W^%Vpe( zjaE_2xJFIgdBIkSP4+}!WOXwbwSF6hnb9JbI+7u*R=s`?c-#9H*uoM=6YZyC@w)lS zs~csS?s*B}Gp0|F(NacOYhKRGVJ{+`2Fy)P#05fuaIj>hl|b#D@O;aRcQ zt!Osp-dIHo4U|tF1?xCSvLl8> zyTDA)YJ!;vI0t6#_%4`<1wqYnBEsqiZ=-N%;O~~}agWM74U(%5`p4T_3nFU$2!{*C zSVH#`Zk-+1P1XuHy1Iz(#4>5}5sDEjsA>U=88lI!JKUGUw{~`i&9ppHm&bF@kkf~Z zaweUXO}Z6mKGL}vpi{{k$M6`_#>GNL{G$4WTg1O>{g^sV>lM$seinD`@Si=y_*OBY z7yGeV@yW?LYs}!7k5(JZzB|7$jNPFrk3+$sUWQ|6$ZAVb@QhR~6(6nAt4W_>GeXvh zzE&a6(njEF*=`s8&XU;WSuEEvUrTNH`uCtjVc8u?#e@Y(zKz}!T9u{$^(5KVF}0BM7XG`Cg#ey=@zOQ zoyU{9)I}W-37xVD-`tuu#XnT=9Z*51P*M=Um2aI9@3b`OPo$jXQjOgHUG$ORzzGUi z4T7RKj!-7ihw;URwwQzXex)wTP)EjoC}B0pjoT^!4=~v{P!urlz|U`28S@K|6O%Y7 zelgs!=Dlrd%+|69JVOomV6u zJdr4#wFI~6;zHapCUa=&hH32N8^Lzys8!+8Y7dsvjujjr)|k+Wp3Ez-*osR2$kim5 zi(RU8gae5{EdHUy_%y7p|+-eXQZ%>s`2{^FRtbsx? zv(xV1Z%cQXrMxTwAfs(bH!43K+o{K~!Rg>wa&Nn?ZF<59qx9WuF%a4|wElGOG-lG9dGJHP&?=Mq+_jYnixV=YY$G6>8~4re&APRX3|6O23fGdV7J3ZTd1gG9e@ohR?9?0J=0va;7FOpea6io*tUm5g zs1$IlZSuFbq`(RkD|-|VG>F4#{C9rF7kRktQr`Q8z2`P27#)D>7%R0zPQ?yLnk=gSsy z=omg7iq|bOTwQm5P%U9j>5Mb<`(};TTzEQRCQaI_fIwtXHNA6rQ~}oCx74xVhJg1b)4LIA3g+(p9xSZe%RvvZSJxt*nwi(#P?NUE~ zZ4=T%IZz@m2}stKr7X&yz3n;Ow=nwhIBHOI^EP}9F zmz^cB(~COAf+T+vG&u5l;0}4!#1^mvr&tJ04Xj8kG;gRi9o)Z^Yc}CmpiYAEmjp?; z4vv$C-s4Z|-k&tM&nT7lT|F!`EPtNMV_2^2KvNQ^8KLOUPPhP_P z|C#I2e-6j>m<~4m$rQ?2N=6^^k|`eqsmbiWoU!i;N#H6aI=gGX&yww1vOMiFWa3;J zBX*y~5;p}d%eW~}$3e)yOu=<`G?|XJcBD4!e^Wzd)=tomgDf_W3W>QV_h^cOa4Wu-*;7%O#h8z}zcNp~Y z^)(`7kG^QvGA>C4u6LpqaAZG6o?peY!``7^j}p!u7Han&{x7b{QiidHc1`z#uML@+ z@bQP+-(aTGmAhjM;!FBQAj5np5;2s2-i!?}Z@qg~ohJpO6ji*R18n&HHlxEhM&C)o zLx*v;e0J|CzXFEma#dZMkEnOS@)Uk$S}iqF`kLF6>6em=bIGF_$Cw5085VJ2wzEhL z4GKFVf}d9(PlxuEvk9{`>%BH_>)!U(Y@?0OK2KP^t^N5DVQHA052096q_L~%Gam8r z%IQf>Rjj2{Si{zW3YK1&k%x;bx?gmLHH;dzE?n%-B=TJD``fH1UqI63tEcvB2S;_x z3tQQ#-gHxLSSRV>BzjTjSk4rmhmXW=apwppH|BRaz;{N8ib;uEo- zIj!W;{`=!o$V{@V;t_LM1jm)74(pS?7%wMdz_#s>0g5|h4v$HF)}|A(TUrQnl=umZ~)QjNd9OGUOJ;Rg*%*d-!%qCldZGRX;S06NQ9ab zryUsqD^7Bk%9Dp!o8zuBv1{~Iag2W%%x3CFORbAvTt11!*eYJH(Tt7vcVik8B= z=P!+?0JGsFi5vP5HuL2HPhEsaQdRW5bA08W?Vz|!=E?op&!w*Y|BHlDXV@j!_;YyE zKK@vF?x1gd0lmwgOqRSx`sVwXJLAR`z7?06e%kc!zhT7jFpPnW3$HIe#`1*PfS-xR^uuXN0so%Uyzg!fJb;As$hOV1-Ol!+v|2 zgEQaGd&Il%y~GO#<|#)&*{tSZ?RRpJy%Fm+V&SADIG013ITGHWpioD6 zXUP&~8wp~a$VVRuOh4{OJu7g%w?bR{Mj23*?#ukj8IuJRA5-5 znnmfISrkL*TxO?3)!n@_J(#M7(ugYi-=Z4cg=tg(^ z(1#7)T)x{=KLLvZZ0ziK?S|9Tl;I*gp;|9R_ufU#Xi%*u-bXomCH1oPiGEJf_dnP! z;dzA&UfM3=X(KZ6g4|QtTP4{kG-or!ZxNe!CXUb_a^275Kllb+VsectpY{_3#ZCfS zHSeo0T!~f$Ce7|yuGk;27x@fQ45Do!AW+r4)ZaG^1@Ya}=g^nc*6d{knCHa2PskbmIYdY()jJ1s(b@uC&F_T`)US~;iC=Y(RC#R9Z}`w!O4 z(L&Noi%XropF@?(T~H52wFSxAIhF!EpO{YRaedbRo~K}Dc$03BV?4QL2gm(8-12-U zOGCq-HzPs(i1uJDLMh0k#=fF`NFVT5vh76}*SN5#deMnw2kdPz6OA;Vpd zPd}qS+}Wt|C>ni|%~%JEZb1;cK?rZ}`EpxvdfsuxmX#A6=G|oyuf&{!L$bMlHRCqpzg5T3pQi zM$^BqQ+{_41qBS2!)!+cmiLC!%%$zD_gn{1k5XTxeQA**mgwuI-+nJQ^aq2bfU;>F zU>#XHbe_`v%=v;H1ZPLYgwuRRgFRvG{(Ef{x#Z>-c(T`k+eXGmR9P?N zc{k&}&-QYe>rk3nve1d#&pb-YZ!Wxk7%Etz477c;Ni;5^U09wk?q)c>IEG_BLkjW` z6_2<(b+*xIYq9H*4+@AEL5_Rnfc;39=;p59nyFENa1+ey%M5%iJV(NFTzzLcg%fb! z#DL7dL~mn3ZWE)O$)au<`v5B6-IE+>_)0sQJ!|gf=>wwb?H->x=2Mi`KFuIj50)8t zEZln>a?N=7qlp*#+Sc(hyUxRSD+U7UAow3P1?z{P0KlTn-fah`+^wrW&CO8xI~KbOcWwd+1$ zSteZxLr5cyX9pU}h6`>}-6>b&62eT{Qar3urIR^WfcOBADMc!kSQHl3us0=`ml?d_ zw!Kux{E^7Yb0tG9=pL(5t>@T%Tz*_%%T7k8R(EV_z}s81-k6LqEkH$Tu-on7{R4VE zU}$lmndhawGVqtj0VA%7Sx>&{y<(%@8=>9WMDNQ+|8KI(s}MkTc|9r@4!?%^8{4H1 zTg}QG3V^Vu-?OXO*$%nGl!A8Z9_aQjD5jEjy^t@$ykYjLiwUP{-VD^lyfGdpMo&Bp zW#oZw_P5S(a2B;p%xg70?mAn{7~|x64+pic@i|#e)$&a{eennr57+2eEI?-4?taU4v+clEv6TRtk~ zUYmXN1JD87w%G_=A8YMtWX6sHqV9oK9ulRbtDu05@ASiU|1vYZ`|WAA7?j1?mpxJs z6Zlx%YKqMS{Yd6fa^CZ0a%TS{*}t+~@(hl6cjeAzw!j#78Si>%2TIyA4k;AWMkG}| ze$i7U^VE=olQFQoy+u&DLv4^ilU_Hk*+I4&2%sQK?guT5{*DFwd5RGqY^24XTs3U%8Pn|ue|T^g{T{6JNE$_#}!eue(D z%^Ghe$juBl)aR0LNJ*P<@E$8gcu!b!?j$oTT7{vi?k6+?*6>pIAue03iUzM4JpjD# zWHU)$@a0^Qp1E_Zs?}&1^<}~r$Lmcz@_c#o*T*OWg=7E7Rg?hrxuQ&F;FEXhulmNm z0EkAw^~drbCz0+Tioi`Q^v+5O{QREzzB7nV{hJJPzuS3=^FiVCam%YZZlM0%hHX?G zG!}ttnh|2ME7hN^s+c5w28N6mHZH`lQL{@Ym{L|;uB$&IikSTrW-2=ERb?HA>aw{Y zWnex$Q&O$H|7z8^i~0Dz*^iegGentn5q(X^iYY!nsl21b`=Uy09_GiFy#Dgam`FzG z@V1eC!B7{uIIhfGfePV4X|$}3=P3^Vq>tSxhJ;bFJ!NnY$sPj?$Zn26ccddFYP^wRG4 z?CJOv5u5@qioY>p8u%HM^!qYHmU?3<9MlR8rD)Y35G^6YsrvOzyf7A}uiL+TSLLPv zbvrsi(Dhst4=W3SW-K8b+s@?k?As6^bS<=1im2g8SC_ib8fxxbV>eiDuZQFN>_Axr znAtq6K=ivAtVv4(QuGxiiq<0fHA+tM<(dx+0j@4}oz)SwN|Ogr&Mtd)`?^-J38>-` zpR%J8HwY+hI%zz-)9heZWj(^@4)#_=`h`$IKBL;fnyg{OIewXux1l~dQ8hWtS0?w< zMuIK0XC|>U53UF<7GaqICCbA)TuKS zQr$tlp!CM{OX^*~O}J{lo zTAJSG$gUlU349^R7Q_D+W=z=scbG91!s@pW*z|af)s!F1A}$pP08M^EouK-zx|qsl z900`3>tN7i4I6gbv*YwBSeV3U`W?6(0>X{t8z=DV4h4($!#O`Qg57lbYp#QmLl6E{ zXP?0ed7iK?P_f&c2dXoWgNv>{FD-)dmy<_iZm>145KAgC_@=$D-4Zc!G_AjR#ds~5 z@p+W)4M0gzbG`koMNrs&6eP!FNO3Vj0J>Rk| z!ddcu!jnhRu#HC76k41~vP@pKPQ;y8@8G9HaS`bI;H01CigI4(EnJvY*wmx)3duLX zqJlP}DWeJF^Elf!S$Ue-#=UK9v}SHg6$3=T@Y2EQn}fx;(O`=^KP*OC-r z@mH9f+p+=ze^l0at}NWp60{hav0EiARd$>%NLZ}Z+#lOEseMLdZzM3A-k3tPUrG(F zF#S7e=Jnj27Ul>cNcu=1+#eDK?uqM&?wpV-nz@<)%J_Zwk~p>dryKPs227lhw;2o6+-4eh_!Xct?}+M=Wu zUp0SJm^j!w51a90cAIJlVgzN8e1=2Tg_6s4#XPX=M2o`cVupZMbvckC{gRX3_6yDi zFhrYsa{aF^`t2=S3&Pl*s8(_bIl!LHIcuyk0LnjZruuKip^k8rnx-aRdl`B{KpAr& zm={VZU#jy+mMAbiDYkTjM43)sX|3>4{MlPh@6*OdNb>2P_Cgw%7E~W0hu9#J^K^P< zLEn94Yl*7wEY(-S-{5(-HHY;Ed`emvmfRdk*Zd*wBiGkZcyjjWqb@8}marP;)im=e zpcsB}Ud)XY$HEk==UIN&eZQeVVy+=Y+2o#c{6u0?1)pgAm($2R)W8hi)-En8aBBI0 zLNEx2EiYX$!&R#M`vL|Xh3_CGqPv>)bL;J!mo4fdCK%Y96?7uW~9E~ao8lu|2Cd(N_U&{`W$hv6q4S9;a}A{id%^)8Pm2c8TN%V{XpLL*vLdzO?7c?Vwr6^824`m_UA~3 zkkdK_%X#(DN^ejKrkdFopFh#TybD8=gbWVuc#sI8-~KcmoHrEw5H6ja5OTTi1j}(% zO5{@AGv>1$5c+~EiuBYCkHQ*!&H37^iCiT%V**M>K0`&8WNMo^t!)FjM%t6^(ZEynla z>amUYG&V@sG+SnKygSo-GsO;~D6+?3e@n+fNBSgr;OG9XEF<||Qs+#ke4x*Ld;6GH zad zjJ+s1wH+qR)BaCkn0=+yflP?4_|NLk86zRmI^G=Sb<$zP8+8W5Klt3-uj<$PQcua~ zCbw!zFLXm9Uc@)~MB(_rSK$U^D);cWPG)2EvE z;j+52fi@qRv~t4)gLdW;Mz#fsT#xEkH(@DP$}s_!E3@J&6&-z8EjHuDB?^Ye@D zRG6|jO_XasT7C6FCW`Yc*W!*a&yn>$70roEu5HMk6c)8(Nlz}|fV;m(m^q>0VUxBg zIFEQ%T)4VgQQ8}yWP+d4V1eSJlU|a^-+$Yn(ewyxAaXKNOl!5|FT!;dQOezw?-@z} zkIw>YQK*n;&_v5_toTn@`;A7HGprzg2b4J=*`rXCQ8JO0t55T_{T^ZkF!6g&?+2X- zZ0Vxvk^Br5RV_oJ)A1s)S<>FxbFWkNCqps^W;aDSg?rj!(SYl=-Za9EbT&e2mj%gP zz`BmyHa}^qvR+p!9#=4tg-f5W;6&<;e9#;(M*FWE+TIM|tMESkHLqTCk|I0Ce^nNa zpKQY4t={_6`r@n(ba^Fs*AVR%H-2^LJyJ9F>$%=zZ>Fj^SYOoe47z(rzUPapgj6yt z4AJ*d;`mcQLXmVp$VaB&SMT+*BN)}cRCEv;7Dg{=s23{n^*vb?KO#Dr3Os`7bht1> z27eNFM@+alG?j-)!$kuF<|>LzUH4W7>J0iaoSgf9ZM8G7PB>R647i!sS8v4MUHWj5 z&NqB-d02z*?q;wlHCj63o?BL?cQIM7?bofPvW7;Ke%8GESOC&#za!aa$CP3yZ6~Ld zqI5|VBH4_Vu3bY&c03_Qqr9F^PvBJkWnvbg>v-7SI5(d<_PEN0X9AM4va>Y@3A821 z+wh_s?sI8MvR^XrzhRfu=z88&Vz8K|&o2v)50ul>Ty2b&l|6c3JmnD}if+qpmEYf- zm{{$vbv#tv*?bpa*R+Y$bY41Ng}Q^>e5Ve}OT7v%*seP5&U(_{SKrMiCV6+nsx{?! zWV=Cs)%M-+EVbZ5KKbW$(@Cl0=N1yhoxVbZB(v%Lnu7uh1ca`vE-ZPWcmyQRZ^+@H zC~veqPz}Q4)9F>PCS{qWZmuEI%8=KVIe;~B3~7hT4r+JSRQL8)_Qu;BG+AU&oMod{ zei)UZdMZLPk#{tp0X={*)X3`T$sT>N*9pN62$aU2H6HuURS#0ZQ#DYLor61{#qmvL zcY=Q!HHqKOHLprM)YpNsT3zfZFADTn5q`FxlXPc=e*X0N0I%n>nd!t=dF6u}g2*ou zyg!ihN@fA~=F%K3_Qr`9-QSzwWq&Q4s%z%!MkkEgAKW>tSmF5G{Rlb#5VZ#3YY<(( zP4lrPZ!$h$8{8nnywtxgZ(yg`zTHrXTGW10aZ;ezaV#h;SoN>t+z&S+k zaP3~u;56@Eda5*ysSKABNw;7xRvZHtwm#aoOT9~su7c6ZxX=i?{CwI!%RZy9X(r3w z-L*=_SG#?b!6f4>53#9@-y25yj0x2nmbk|WUn?g6ykaV{I->=`##X4T`e6&`4x^D} zNd4e+`c-wU>Re=2cHHOqS0byamHu6WEBeZ-lTqmX*E`i&1yBr+h$zV4Vvik!cI+kB zVMs<;FfYpjlZ@YEjuPozwAS>Th+WO4cYHso<0nsS;z^sA=D!-{N7cRvIQy_xWc z2`^-I&<%twzfDd0^u{x1$d4w!3ay0?)eSJHHq#1yJoL$MiWL3TzSAuy7)xA@su=E88B!x$)EzdgS_ zeSMG#9x6HMdr=fQckaC?NWV{)N235rU}de-=3}!GJgI;mrh888Bpuc8;d|%nw-zSy z?i}#QpKXky;i`TrXsHKHkLufZpC{Ni-taaN>gbo_L`G}jWlK|dosC754*lvyu6pEl zyT0@KAW|i<{3ZFwx)M@ywt9^AmW{MM#W17zxXWGloY%4*!ZlIPKM* z|C$TAB5_Vcv~Q>9%SVa!$E$PH%`5eqBNjextDlC~XeWl>QBkYRpaaXwf^++D!V@`W|2exuk7Jlh9Tu~~3;5lP2SogT+B9vYz1XhPjR{Q6fRBD7GnWqQ zC7Y$AHy50MZ0sE_?@5kl)pXbI>$GZ120;<5-PW%O)uJJD738p^AdO2OK+dlM%i;WeCp@JN`HZR)J)~2BSn`Wl#x*MSz z@xH3l#Yxb@_OnyKr<7GjAmQ=+EWkfGi2!PG(=3)I>M~1E*aN|bonJ^VHwQNa=P(nH zu*`bq&yB`2PsW3=&h+H@|f%Yh+sQ!X?QXcJg@bQbbzns{Yh3(VrKy!~lhd=Mk6q=}{-3I{S9p=FJ+hy}0w{AV zcq;Flt!O`G?EU3~4oKYTF6`khrgsYAHccsC+a?X-1uEbk523UoQqI>n^I(6C`G zvP+tGUc8v77f*6}+HzZPUn^^|g}X?NK0qt9Wp%hB%u^MoS^|(iRI4V?Ss0?1$>(fg zE_qI@il0VwY_#Xz^X+1tuu}80Hov#n1X}yP8Zq#jIX_J*RB_zZ$+r>MKgjNsN3S(v z?&_Me=TOaAd3*Z5te5EJM|jyI11zU~HhkF_;Wdo4W_JKIfxmEH^>jc2hSDY-30k3` zIMr)RlHf%fkH1>)L24?mX66ysV0cs|#ks7d*bh~t+VWC{ z69K0W03nK>?b1sQYLPkoDMSXrDCd6XXDHSN`^XSXNuJ|CgBXj;FTfa1{| zX?PBBU~8fPVdm*6a^_V+yklR@#7-PK;YSTcQ}V9fK;}9gBHz}W)SGm++;%OFCN;^! z2}SC(Pt0_fB&_rq0$i7?AAX!nH`#I zt1!P4xUl2f#e7(e(IngUUx}kwBxDDc=)Iu9ZFit1Y%Y(TDF9qMPzl0|P-2Y$cF4`q zbED=q0Z`-9p+h$%=9EY8CukxOm)m5jOYk=Q1!0rAw;`QlsMW}Lfa7EEPq2paBt>^D z6eVGGiVl56yO||Ev$fu6`;as#X|&qox*JveyX7Mxd($1ZZuA7}lY;6#Hr1+Ci#35w z&rXCJAgL!rhoQ`YTv}2m$q-mIMyi~`uq?}%uW{8En~BjyS=RPm%FSU-=NoHWuORQ( z`Z$Qkr|Y^%JuMNMHfSywcrqcou4W=j{PD*^@(es0D4!dufx=Kee!LiyEf_?Poy*@| zrFW6^HfeoL?m>I8k2eTe0~Hc-HTmASo0x#yD9kXUqqBRy@Tc{;G`s5?2IHZEmnny` zcp~hwu#?pOn!>ld&JWc{<Kc7C6`sDP^ zZ6ETlX*00i^XWm&05B1JUb^4LBZp_uQ`p_W3O3$kKyS#X3)++bUXpR@g)ec=dg5hU zsuhom#t7hj+tJ{g$VZ z?vCS6O1mT27vIS5Y0W&T1omen84tnBBC*@-Z(U{P<-pVJDsa>fgKJMOQE9Y)EB}P8>$}xcw zZW!v9?9FTb930n%2#l+oIZ!EEN##Bf>n&3u4nrvC@|Cug>AqnZ!NY8m<~Y9x^TkEN z(zD}9Os=)2j!Qu%x@K^(PNyObwDfURlPou62@#{97H^mmM@8k1rlL^|K3p-K4AJmB zPlvFa8s_vrDXibxgQoAS@rsNLOouqjuMZ6l8belxhKFxbS%!uo!i|S>lAOJrRq5#N z*oC(9Z~Nekk^tsDHtf*yWuH#SoD}Bue4%)+z?5nw%I|y6nERSppt6J>3RC~RXD#6< z*`D6iXNtGfU2hxPO+E`Qw673w#y8~`9NY<}UD{BueUWOu%>4dY@ldC3mQgC1plz=g znsz4FQb}-g#b?;anYGwtETJ;|-}|)TjjZFMA6IujsFY_rnVD@176nSV)*pV(bZlS{ z9O6WJ@fT;xORZ^VAvhd2)oiB>DN@14@?iWmoq{${x|iXKag=g+FQkBWZp+Qbb^yPq z(0HK+Kb7buSkB!|YDKkoSNNy9Rl5tsEtBQdR$-wAX5J11Ze6p(=QcL`>K})kV#SCv zuLxR96&Flf<`P300pbrG%kp$^cvjTawZPxb8GtFIjb1Amj=rZ7#q?im43qYrv2AV%{7x|KcXz^U9%K#+JRF!qc8Kuv;qHCbVvG+HqMFQ=e@$89`xiS(3H zJxPM*WU}t(s#{`3gCRn@(;#A|^CQxV{uZ8-ts!f7dDi&$`Dx}6X+4ugyuOIpPh4J3g zXWFOO?KL}v<>PdTj1_PLO3&qPGJMbR7vYXS(o9m-%o4Zj?w=jVr;ZJ#9NI+M9}bl$ zbq$+1PItsXX3&@lCV^J`K)YiQS+Xf=9F(2AFz3HgaBi)i5+E^6lZVL5I3!l>mP3a_ zVxIsK3;lSFiWa!o4k0Qk-blGf{7HcwqT#rk>Z)c4@sn$U>sB-0Y^uZ^Jy49IW-+_L zo3B-7I$j0(NV%mn2w44M$%loX4?tv^O;oyf9@=J+DWQN##UDlABvCW-P4HRvFb8)) zt)X7hyM+84_{rfp%844t9tp7-t)DTp5COP=IbgEKayNYEvLKClZbGXaf4%`c1V(2p zv-JUG)sj&h6M>}3ZQw8UA9F}e?!IeIz1x_8qX4HtuD^)stujn8bs*w1M_zPlOo?uYAxS0r^K zQ81@P4CLAaJk(xrP~+}s%vbilSP(FhmKF|oQ5*1k!EAqB;Ic* zE(t2M%2X-Lgf&1hv4K)p7eVzUC6!Q*mC3QeeZmzGtg7&zJ`Q#-yHKL@%Ao49RBoD- zMK$J%=E*uTeQL?eVWtQUNTf^BQ(1?yXr^LpPOAo|4ShEbZBIlfH5+$V>FsuGPq%55 z-XU6Zx(C2r$$9k76~pke?1py~;*VX{)cl;nnq;;p`tjOotP;B)%3_%U)>TlSY)0qV z)$5#|I3a5-Y<=&IBr=)p4U{Q)i|>eXj`P!>tSpu-i+Cr7l#J`@L!aHp_{LMkdGfP(_Z^b`QXdSr#fFxclu=+TRhZ| z;3n89`Gqz0-SKd>|5AW)r?(&NHQvN5J>1xFw--O;1CINr$bOE$8RM`3GQFXi;;!^D zI-JyxxJubJ!2e7;lCUm&#C=)u(OTMz8y+|cx7p%(Dm zGGBE-X(P*J0idms!X!f%9srAa*@o}RlNT)K#H?#~6X}tS8AH?8Bs^k5R8s@*KBVR| zUf^0~$kI3$1p*43=;D4%Sf*N^r0yxm^4cywt54RCnE3LX3A9}j`MYvqwn5ZWt zf>FZIVmyodYF4}b2RV4LCUkUE&3`kj=_72#?d9IC-AHBlU6W`Lp7@l_WCsn4wE+-` zu3?(D#uIfqq7Yc3pfJXCqt22oTpfPpIpsA3uD2B`C2e#wG!Lcb$~Sl}|3LU1N{`qJ z)`1}6SL`M`V3LeE892If+(>--)EtV`HyngYzyza>nJi zL>m?wh4po>IPZ1oEvfw^0|bcXrPBHP(g!q$UhclMdo^PL<2!OV_BfB~i9x0;x`I^( z3yEtP)-4f$;Fme4U9EOi?4j( z_Ne1?Zhg9Xr-sgcMQ4V4@211lY17~*4>UR7DC%)Ldyt>m81d%hRtZbq5(eyQ9W7{&H>JFEy}1ZQ7% z7i%27hhV*5FinhNbv`_Kn+;sE)C9b4&kzKGEw$f)T3*8YTwvQV5c0{m@#5v?!!EU- z3M`jB@b*VIQ{uq|Olyxhe_%OToqLYZ-?aXH(e?$EW8RgL5DKi zW|*ywrZV50aH`XLTE~oUy2P4s1?Fis<^paYooVn=n?UVl`io-V`|f0X zgQQ7Lf2?IaKmoPjv_qj^+-KV;89zBi_WN#;`<{OHF3njuk?YuM`n^>u8hDJzdl;}= zi#WTLZH~Jce#L0-zKFz#zw=JCHB?`C*@JY;lvHc%8Sp-bhb8<8;)rPtaW-tXd{1&j zo$W{DU2H=5re~5%&zNM7(S6;uq6RlUaGCZuKZvou<%=Mu&#mIqUL(xNmwHn$JKT0L z!Mqk)yP=8R%6oy*u=%JVd2N?12uTV8SN3@MtqVmKP1BfwrsG_yT-p^`Rim7`PPmif31O zy2Zc>?5U(`u3&ssKObq;iDx>X&bKd=ZwV!E}3WH4O8>A&vwr zVyYHo`I~GE*wNVmb~*BM{(Eh&GyH864;zJ^SI|`P-z{$3^}~^V`M<#+rGt>DV%_n2 zBg9qD>-{kShon21uJT$6&?cBurnN7IfrzTv4(?g~!`9-UxM-<-dFLo8CPIr;uuM7> zcu3=WWC0cJj`mt&Al1F0y=eGXhBLN)G)tat)ieRY%CF>DmWjLH*RIXMpV{YlbGz-e ziXB!*)MGj=hQTGoXMLtvYk+SA579H0-C&g*wpR@WS?UA6cO*aLTJT4~y+;v8Qrv>l zWs}rsjvA|#fp0~h@ENb=Grwh=RUa;+$Gr-GE)sqnV z8Ec0`cW0eAPSwT;F`0YLYqost%U|>*Uc(xeUBe}q zjaQ@B$}Cf}cT`p)9VBQXPce<;D%B6q0t5zk0~i{|w0*wyKw^*9#Zct~?vq6c<2w&O zK)|+=c901$$ez@ze?0Kixtwl6;?g*F(|zy>v6{7YFm;sMR(j!jOT+Vs$(L`~DWrc~ zR9jzreyAIYMzYM7ZC#DN1RweFNuOfWw(q^g7h5Pf^YaBa!PrGn#GJeNj87WiaG8#NOM0S{$={>h9O>3}GS`mBIPGsxH>ZE-q;U1on1 zW(uWQVP;sbKB+7{o32axDQl7u_v!2WSp%ji1D&1Xn(QbdB)$7XeEJmIwYEJ?Ja0N) z{>{(XTeOHcE7%XwGW9aaf)NWD^9}308k&2SIWif$dsy6bmGAgu^Gdsrbz7LNTU>&P z25Risea=7yHvB0DBRK!EOiL6J8{ai^x>ew3C9?_d7H^Uq>hX?e@RQiLg(BeXcC!+D znIi=${s`Mgq0fgNW0u*;Mf#fW`^-kQCuesW+1&c^VrgG}bs*W+9I6meyF#vO)%peY zirC4jR}JRNh+3;y^8OB@Og2P8q||&K8>$6sC+lBMMA92>O(d|!pO^wG`Ds75)pDsU z<21l&{bXseqP#Kpg)0mJFHD)p-H-0Z3Sbz_+y4x49Nhxek)2NgXB}Cu0PFbBQakR? z;0TG3ksQh`Ni^^UkBG%Wg~N!xugj=zULIGHspkK~-do0H(Y24FfT$oUC;|eKN|(|N z1}cq!bQpwmiFApIfP_-gC=CMA9ZE_{cSuT?aMSE-RFvo0=l!2^_SgM>f%(lnYpttu z%^F3SuOoM+hFi%sATLOx%C*EE99WQbXKQE|;aP|s)0dPjUry4>8Xe)MN?i)%u7MM!DUrFAqu%l6AzrX^$Dp|QRiew4yPD!sC9 z<7PV&I%f7og;S$l85}J-9_vb2AENJEW$9~|aM0i*%C{Sd*tXW(zkJm}iWT15Qp+{e zUHjUBQcOU-S>171bh$w*8)|66zx_i&YNV&TP{2U*$oRY0u}4y_-xt)>9G}%h*A zgxl9I=@8+xGdibMMhhj2jW|hX#a=;96v~jcNU5EIWar26v5xinX{jVl? z=m@m(?ILkv!j$yNN8LWFm8VuRvc>7AiETR9@@Y>6*wq-&)sMIxIa_pM+MputuZcVhW{(os;> zZv>+Li}{Jsq)ht``;|93Jx-EeN{n7ftuk_;=|;sn35tx-LBv_kJ>}~J-I7)GAZnyQ zga5M9`1mQieBmF;@t_ew*A|8qvnLlABO#P6#(R4S`V+i5=bk|{6R^UzCyx*sl2^D|H0}zI-TC>Ph7OF#ZhWJ44z*M5{uk40~hB_V6 zHg#>f4IikX_BMW&qx*p}hl}ovyuSTIEs-r&gZ82zC_B525|YG4RT^Q2UHsa!=rtKQ zFUXUG>WAcre&5mAc(kqGwDkD}mmg<5yVoTYhCU-d zuYpx2`h4@9L>ZK0Hf0)*B8PR!Vk)Jp>6TsI@pIhTB=QOFelLd83vUJL4hz{BJx))7 zw=jE7oS(af8nD*qd$eDg7J_MxX})fCrY3r}J0t&TQZ5>3C@C;~Ij|>#S^R9Z5N4 zwuZiboVE%is+FY=o+;a}jSk87%#58~m#;QByyIx_y8kW>8Dfv~%!t^$1* z>POssni+X{GUQGx2Cc5LA?Qxqb=lZU5(b#zN2clJ%~Flxf(Go zuO`n3Mfs_yW1e>`Vxqng)U6j&m*mQ1gudv856RiAs%P(^`Yv;DF2QHOqMc)UKX11M zJ@;V%G#MyMZctJHj*cC?$3cQ$Mge1g^t~eC(RM{k-+U1W9u^|ZKY#E<;Mt8_KASmr zD6`=vor9B`u%01R;AcM9f*hg0z8d)!=guiv^twc5SYMBkMb#U;gsHkWsocSSMdA&fkMOA zFo;0a-BJJvs*l0LCodbb8p8kgrdp<_|1n0if)4aP`W^$f7Qne|XtqFzjycK%Pz~q# ztw!QkR2Qxl+8ka544@;Rx)L-dh`@Qnu#uUu#UQM_mBUB-hdJ@ z3}SElc*Jc0E-dp$Z^NXN^$2lHg=V5M0Q54;2xRBXA*$o)1JJzr!l;E%V3U5}IYexq zsR5fhnyybcLddj=9sB^6Ka7(X{GOf^9ukg&Y`vj3NQX1Yu+AyL&5wH;rI%sF=N3~b zhnE59S^<3TC);wp{2|`$Rh`CsW0CgSd%Neoj zuoB7q4HkqFlRyA2W&hP!#Ew)e#Ww$Ispkl#;L|F*@4=@qdFhSQ!jx_|fhr_4qpsAz zd8W@dCxQ>HgWz`{lT~1;SWl-l3c)sN3~Ees!J`&1!uw&WRm;EDX(FI>HA<(Px_A!^ zg6&CU5nDj6IySPfK?e}h!PTOIIS!BV_XdPmB`&y&UIa4vpk*T7fFAMk)CRXrW5wa7 z^yuil!!$r(V+Ii-$HWvwd?eayYGfrWWJ6!f>!kNnFeWNMp7I~-xQdzts8+gAD zzweP+9Yv(!BwXKrbaVzU2f+9`s)2Ns5AinUJPh)BM^NbT6f84A33d_!J>E-rn6gph zk`9cfGQxu+5&y9k9yWk~n-f3iWHCJ0qyBOW!FmYbvm3k3&kb={U0l#xtSwk z5KERMRt0>a9&%Mg)Vl5t(Eb+vX(mRP^1Y#qDZK4)7(hx&xbGTMC}n}bXR~c-1H^S@ z5k=FNT;h-c@GJ~X7T_So!(%F#fOr{~80fm-_e@hb!44hs0qh+kimjPc!iRgpMbWzNeImiI1L!8C zYdrxt(KsL}b*F}6*F^xskCNfv7%3r88Je1QvBB@Zs0;Tc@-wD4)lt3r@4xCIW8uJBakft4bMvW-g z)0^Xj{~{m;vVSTf0*5db5V?}ah(gRt98e`b{A2tEFx;!$r||k=lt2m>;eH8EeARhS z@9K}Og}UHn;$b{yN!hx3d)_rEBf3BMB0Fs2gp}dc@53v;Oa~Eo+|Qae8Ea}UPQJfQ=Q)q)`LHt9h$CI_+G>MNCtVOVS8sc6 z);d?4;kY3K!)m#V20D5Z+IP@uA)sbWpvyPsq@NL7a#0F|=G3&5^I_2=t# zV&2MM&_!HC37J(w6>nh_Tsp0-jnwOzjA(NMbKBMS9>3-{$ekx)STv-4noY#9+ds2R zpoFJyW?0(J$Y7*xR6b6J@fv@m5LZ9oOvTY*)*S&l+tm)Hvla1POaGEbG`AX%9ul4t zyJr&j_YCajo6Op0#L+@-j`knJ;2nOH_Efn{orX(trZ&*Hwf2W8Y4A{STJuTi!qI&a z`!9D*)igyYBFgW^KE`x}1yMeEsz>ob!PqhaHf?0IjH9BU^~gZ?{9m@T3h8c*p4vs} ziv%#G?`BvNIoVqW=FWCCL=-lpbz_=;;$lUQDDP^}GK_q?oo2~_|6xsY%Kp<`3Z?q3 zqZ3_aQdc8qJ&kWrKEn`1Q5m08pus4(*Zy;mH={VQj5W{h50rRZKLI$8u5tM%02a+_j37i3 z78m|0TL-&?UaJ+)`k6J9qRJn3+I{yoU1=Pn6SuQtK-X|ndyUqty%8p+o3*klVfFg# z>@{67!y6v&nT@oW9_;0tJ}LYVsz_yRG>ZyR?2_moRPyYw30uv#4D@xC1tti8;?7n9 zbA&Nz%fH|J!MS=Oo``)vb6)iWt>`zpTneQXe_~ zIB($@x%qPgISk(Y!R{vdkWOu-w-sr^I@#$ehW4KfZWR>0Fo~Zxj5932p6UzL@jdGH z5@%_^BLf|KWXB92w7J|uAp^X$;uYV2#rC2VkLi**yhHGEQ}-ryEdqn6KUz+x@qPb( ze4>_E@`F%_c0ovl`M}GnJ77t?M8T3^jBrOQb}FdfUhAfP){qm}AwEl{B&mA*bH3CK zW52g(El%G_UZFJ)(|#M~+uNE{=l_oZ_{-cq8fCvA{XMRsl(qGXTVG;P6ihTUr%GKQ z<7tn9_^qfro%J;cxA=Vh_!S!Kq%TF*lbaa^wz!X^8IPq0dHHm2sWV92s?bM|a)@?Y zWsX8)l`Xg+J2uS6ckQSnu)ktN<6n763_TJLbi86KU(b!O`^2H-)1RB>pD7znD^!#( zS#$)7vofehN{K?ZWkb4S<~M)PD1FOI{%jXg$R2-l`*UQ>@IB{bNIiVuqJN6_p-? zGt;|gE9~m0weQ1spF)oKI?4qzYno_|uan}wc1H~}&p*X9H@M_x46e|xH|&U!qwXk1 zP`Hw0gEQ-=Fl=TeXE_=!@0sfb;57`gw^MH14xx7?z@(-G^Ez~s+I2P2WW&Z+BwZ^= z0&Obj(!{yL!O1T}i>C}~_%?5+mq?h2t!yjE8pWa(SYKT}TzPji>iE#>c*$;+?bCf` z3dCiqvPMqhPj+>eDp=E}K4xx$ZEoUeRQ#MG%c#2>VQFNe>Oo>q`p^k${)yv!iadNo zIH>oo$R*uQLZi6q7qf$>bxASa5rPReTLl`obMFY$QYdFg(73Kjys`9xI5_>hdXU54 zeCTB=_%tc?y{YmoZ+I1{v1D=23kUT^n}UaWah2~15_&(0o}9~l5v`#mvZK@B(srhms0dnC;4yAsBJvxNe73pYl= zQqEB>?1ODvhpe*yuLSaph#y7%>>GPIeq2z zci1t6kp(bkuV!MJlT#L*`&4eFA$Kju6OHwT&(*uv=s?MfXy08sr2OL$iH7iK=E&|$ zt&UC9hDwr@de%nms03#u;a-4^@sz~u1>x4s&Np z08`Sr@^V|?F(KuTDdA^6 zD+)oN*8RZo#LIiXa^+TtqrGnm+>tyK#aKcGqd|EwEZQ2A=s{>Jm?ycF4Zi1YnbwurxqX5 zxVOm=R;bs7c@mr|ZkUa8bSIe=8C1zGkm_n;HS5U`nvmL1>Z9e0GvamS!u-sJ#53IB zuZ(}jGaL971T~SVHxH%o2hm!c7?77J{yaq`k-~EW$DdYRD#bsyol~b_C_XJtEb{w+ z8B0+Hs$3p|vuFs;oQ^fe+F}$4eial1Z*Kdt_t|@_kucZlcbO0&dus+*RsIzzF=Y2^ zmCtN-8kza{k(GIy)wK5urSn+5`Ylvi=Ol@JAW`DN7;jw&qls6NEeNC0ou&!y689#b zjko-+#$cctQWnjEbwd#eIZr{z+4D2xT!xVIm4LF(3y>DhK2Np7(()&}M{==GPrEx` zcBIqi?lE0aMZ|)^ME6B_;}9QfkAx)`rXVOk>O5!5D3p4hjpf`%Lgb>&q>@AB{{uNh zN79FvP4RejcY$8Y48F)-jrX80MDy#h>zferPfXRQE+ay8OTa4RF9rVGI)hM0#r=*Z z2pFX-7&*ajyew@1L-N7{9*}ktmWHpg*u;z_f(SqT#OB7?QgD4&#%YoZ;8a zLJOsXo(0XHXhQVhmhM5ii6pdj3IMk(uZ%ZRcZdu3tKSCJOmSSUMpCce;Nr!=Z>~KW zC)_k3;Is@H41xkIi|B6eRRo)>YEFY^G7oPNK)DKJtY(WLs- zh~SdXH){Unbk)JFn@F|y!Ks_pNL@-0;QlWwfuH|n<^Qc($sk{xep9JQ`~IFf!$i}~K&lym;of2g z+7-?$SE`fii28g(^|FO&U!gAT%h~+yd%fl!&hqjSXG4xv9Yc)8H4Mk0TTuK+$pEWS zwj$4qabO#f#XG>Wz3+O^(TxL@HnwvNu_~aiUb(u3Yy>JlnjAwndgzoE^Np!Z9CAax zl@l}QvNU9o)d(Xl+0FmNk{t!IG43dMWhN{EA~DLJ5f{ZeVic;ORHDqx^2Lc0{8>7K z+67V`uqOq%wvG!n^I3+(Jd-A>`IaW53B{(Z3Vo>AOS7rn4J9r5_4U0C_}fKh{p#Gp zmaA1pzv__hsVb(N5YKW9iIR}!ASF_zk4WZ4n%XZ9Oz0Hs{0gDFkXU^Us7Bayp&>ki z{mGj~TYa_N((qOqkI0W{k4f6N?%z>XU>X+OecrfhtoNY5OMmf_3Ab^yZPSUL9N9R! zQG)A2dZ+Rp5aa3Eq#@S$C72f<|2Pu+y@ooTwD1@1tVrqUo+a8(R4_Hg>Qr$jAo|Vi+q0#4N{IkN0m4+6LdS^ny zsD#vTF+N$TVrGHk(rv$wycQ=X2pp?!4;~QzRphYH*#T~ic28Oqj@hI;iW_!!spJcM z|JW2aweF;8w(IR799SJdq_S=Q>N|_O-@Wq^%JdL%z%CDhkyYWiQ`d7(&;m~`+q1qB z{gv0xA!Rf#c!Mhaty+M?m*0-m8_5AKs0)#!c0NgGNUpu%>S|gaws|<6-7nfry8k0; z|H}%oWSSG{LXYe*@{;6P3uB~%iHDU4*dR63yS#wS=Uyva%tLJ%IYq!mtTSRj3VBzH z@0x`QGf?+I?_6fVG`0&F9S&Vn?n3VShe5^;3#{@=ccLCIN+vJ|F5^1-_v9_Ncu1m5 z?2$AI@r`)S=)#MY!#6uFtH4V4N()Ia_z;I*Rdp2+>T#HudidFZC^@WT7h@q|jKpv4 z9ht?g@$E!krd;f`O!Po)&jzRFNydsJW_-=X-e{kXZw!QIx;C85*>vnJiLXrnSPX0h%u@&e@pc!C!T?AW^ zO_gPgVoOe`MLSfdElcxx#|Kq5E+eBid8l5Km>Au?rU(}^|-wlZK8ve z>=1_6Aqs1EXp}90YtoO^_TVpdS&s zJwLtd8BS(fmyq@C^zPY=d4k>1Cp$+d>e>vvKS=8aO=RuT6nA?(Ucb8YOU(Z$ttg_j zqgDo(?1<97xi5eytp*UXx-)7(+7vDd`79KX2xzo-q1w3{pJMk0p;W}fLA`#hdEHbF zo3mF+M+SmBag@K@$}hXW=GDF%zViTQx~TevsYe!a7=gw8Wy|jB*v%ihOS{gK@0@evumy`+j5*og0P_FsZmGrY=$xhz!yA@Q1?&tngzdQ z9nZis2hr!_Qeh)R*N^Gi)<2?Cu@|mWn%lC!5yc!?*2Aw?8M*bP!kxRxUQpm?%iuF& zy?OiG)SgL#7;DU5HNpD`=-A>RZsdP!4~{^TYGwFQix~r-g5$cJOcKS@tjda0IvdIT z`lY$3bf>#rX&Odk92+n1w>`WO{h>A+|0k&~Af!UNX9b8rb&_`+Lx<&jkE0_V!UJFz z>cezGLt?EnosKBI5eRa;cwz_yk?isTW6p_fz$sZ^D`}-vMK7vRED zbkx5Go~q89_TH<}FXPcH+wQKaatgj1DDL6IFX@3!`tq%#NiWJ*pEpi6(A zIjxAsE{HYYWItj{N1->$l)gT0YKe$+lV|KG_;jL{ zCw}dkZ2K0TQqIm#nWHl~=kp}%CE{k6qOw}uuE&eb1{-YmB9csrM}`=@cBPIEZN%U? zF_5{H08v&Rj9o&wV+q`OZN$zv=LTK&<1f$M$$=YU{W$|FUZPEXB&u94&p&YQ2j$TR zbPe`N=q4)3%hx-aF({TGx=+mld`D6>2c*n_PAx_`k_*#xzYj#V5IjR{bj19%v>l`! z)*aFodytI+)(B|GpWni8@y{+c#DF@2P_7cc9legATCev`o(fDBj`(#sI5GOCfkpoO zX>>2aR)c1vJM~Dy`xzP0{N$a|9?)`gR&x4aOM%0okw|Y-=2X`iEKg8^+aym75l9YF zzdxHk4(~xY0GAx>sh#WXGk;p5NC5N=^b-d}k;w%FduJ{F$3Y4#^czQH`rqpHH-1GB ziG~4=#}YQ=0NCeQs!58sfiu_I@Dz_A3JmBVe{kWRplBQxW(h(SwZIj)>4un`=h>P` zLFItU-j=+54ovla(76K|*MCMrlQTlW-5L&i1ERcn5kmqbcl}*DaHNL;*Ls;1885*M zR0dcnZU-!ZhmONNa!Pqj3@xyqb}SNhh@2v<`rOOfQU6 z9D|B8>~3*m5DtWC5hWrx*s^$lBfwm71Rlc0W57VqfOSd;;|1X++Wudqo+?seACl98dW}+apIGMxh6LVjsh60((<)!z$S-LQjQ*r*1;_J? z#`q)GGi+MfH_NokG$uc&(afvo1rC|?OJs1HJWaz?#)lm>7?|v!1G;(laSv(p9i&vI z=Dn#n{Dn#ROQlHTn+&B7QpMMjzYIP3-g8xErd@$Q_tERT=u<<80-1%WjaD(FPBex) z-yV!uUUMtX30ZaN6dH4WLWy@2x^GLlXQBIMvQqu@DEnk~i%#2nF$1Ct#(O?iT}v!F z{icI$>zSW^2dufgU=X}w$qI4)YC0xt{q>y^B@e%%9u+E-amOpVCewd4Nogs4^fPo7K~!% zvDK`iS;N`!susF>uP~C-9<3k(%j%K6w$k_1)9e)CXbY2z3F1w%+FB;PWoa5hZ`l&x z_zOBi|Ncsnq(!0zhh|w=UJV?F&Q?^fb+NzjLg&Y1Yu#i;r9f`ud#Bv6P9TlNR&Wa@ z0hP7BL8?r%9^DJQMY)0>YsuRRQy48;gJ!$F@R^m;;1jau)z735!W#5Y+y09)11fYZ?+G8P1sYQmcqJol zs-x#Jf&XWCSo4wX=ZMDxODX`Cw1>;C@{|adk*^?Ew)5!^dub`-&CudJN4o8SA?_IXF=YRp@x?pO8} zA&sSm>-cna*YRo9h`x8F1no5ZFmA}1+iDciZQ9y8gPwl=%C*yT!y@hPj8VT0F1;E+ z^WH?w6h6T8+a6QC0GeDee!Ly5{gy*pSS?GDFio8;JlEu6`D}N?P|&*7USF=|%tpr9 zezRU3f8&9g(tHgFdRGAFhH%zId~G|k&{st+!Pav%2~+)*v0D?5$&XV(c8qlN$EqV0~`dK67GnnAbrB6|uwFmbYVp8q{rI)zt9QjpYGohIr*A zEWo)F=oecm|E41la~Wa`Z~bDP&Gcc_oExUmE2Y*^#BTCYoY&=)0@*2)0;sG6-C$c> zk|1vJVzS*zmAChYY~0>l_XsBq<;|V&&XwH{hN=ZRr9Y-y<&Ui;2cnlUekTq|yps4_ z7EyeY_gv76aV!>OH+Vm=KY^J_ZK9s%AA&3N>VBS+Dd8{oeb5UyvcteVLS4QYH;XRmXP?W9~(OpZlrZ) zzB;^F8YR56C9n3f^aaL{I|@|TYIhPa}HZ#JK(jvT8ys&2-@^8MSQ zLoAps*qTlaKA7lRP3+8HeqqHyEE4a!#I8CLw23o;+pd3!+jJ=1=T)EWMI+Xrs+&rJ z$oA$4&O0L|VzUz;t;!`!x`J5A8q%~(E05*9iKl(L{UbHVliYIM%|_#^dLC(4pYs=4 zOllgWQ9ViRij!xe7s~hTG=%{7{QETZz=l%-bM*bzi_PU0ZI4TaA&MwvpdmzFAz03# zvH4m0_^FewXy_MOL~pv~6S4*uC+t$@TCaI11_jj*>V5yh%A&3(=t1^tSLS!6dWSLP zy5}~W&+0FFB&(RRDNGtQyuGBj;jMs>?1Y6I}4x%6RIbu>*(Tu8&ki z(M`P;ZKl-J+3}=O0SP-*%MCs4-FNy#br<*5?v)4jH_y@l8sa5-N8OTsf*21Vnvq6s zr#aH!=DR-%AaUP&TSW)#oL?g*7hy>&QRBqRqS{aPFpA{g2i(; z;X|r=gXsySBQnnOcQ>C{IAJAPRS4OPxPcR?pEu95`yJt(O9PI#aG10m_ZE2rm14A<2?ut(-!UIvA@Cd@6Ei;k5fs?G}qU3LQC~bM*Y9d_tjM&*%nXD z%Btd4dI~|^YRlHVY)C@~ac7pTDtw?pg@zwTFX{2?XxW|*=0NmM#dGmED zJP9!Ka@^i`8K5WKH(5R1KH+^7M`zNDAWOEq%u=*`Pn9jNHoK+1-g+$>MQ2u&98uJ( zr|PG#zXyPB_FL#JLvTpoLTcslSLlDgl3P49&A1veMC8Jy>a3wDQWZs>qaPw;X+Nwb zuOvTef0}d`F&`}(CH>BUxaArKn)d?sf>yqA?Xqhb54~Tr+^`$=6v~cYcQ#`F*r0M= z0FOQfoICn5X(1=N#zoPl9;a{nL$7W0nzJN&vLuN%#}o#^MDl&heGY6p`q>X9~pFccyKJn+#pV%uC5*3 zPGZWU4IatI>N3|SOV`eMiaKltbDn5^dFFPo%sCFnPj8bvS$&#s+*B;lIs3`6FzFTR zS?Rj-%GBp0$;I45@4UZ4e08FWUfPPdm~=~G*z9pQ9Xvo720W@O$`zFX6^(GkLZx-ln_ZuA8%n-52D^974ZWfyzRo%Zti3q9OMH)VGQ za}#UZ38ly^w9U@esaDuUo7p5d!f|JW$}9C{j{Cb7gJd=*8{g+|eBS^3oTz4H%$L+c ziDK^)?r5+Fg{kTGBuz-wQ(}&=jI=lcm5o;1c(SjLFc~i`eJ|Hgde{5*1Ud#L79J%; zDsJ{T5m}>?acDvq1#wYiWOJW{9L7)!yQ)9Pe5Im!%w%@9 zxKuq4dOUb8yllu&BXx3eujO4EH&~u3FTW&X-xu&nO?z3ytHs4ya%VzBkXT0BZrXF1 zG$f*@@bfFF7>8{oFmumqHq&iK6*jbR#aWekp(4 zs=QO0NxG;`);(;eY!#=>qIH2nGxPmspM*U+x`w^NMWtI%dyUizgPs1rae@0jEpDj` z>iDRI0U;{4)DSlBY@){Kuli1pjM32tdT!3oS&XHyED!HFD_Fg(88pcd*xyiDe>I(H zF^j*vOR?zHiMwEA>N3gDK88b~O2VQ4gL^ppT8uG3Ehs?iI%}+L>U>YyanvK92bW#H zFQ?$gkM@VtJDZsnEl~v~yIU%y)<#y_?Ng1Oq-;g>lZ4x<{B)GAvk__V0zQbd3vLtm z!{6XmoB@!|c!urT#ccGCm=)}%zx}>?O=M|KcUiI)dj!7QgEhiWBfx7qcD0aVcVKH{ z5@qwsE@su)LPldXPrquQ_jCT0%sKyD*Hz(`b$E6-#K^2%$W$#`uki>A)82#DTg~cc zTF57^1U)=`&%*)-Yotb6x0d7t)`(%xV+C4Z{!dVnK0t1jm1ZK8xwgGt+rL_gKWBK2rEZTAkzSH20dUZ#h zFOACQZ>;sU#Nl0Bory+E7uYR$I9B5}?5X%+Rkz6ovlvgo=X1oJ!U~?!-&ks!2&}? z{TR2xG0~@XFU9zo-$oz6q9=1Snk{o${_0DXX|Oo-R>aAkS^XoWOsG(?`z_8Nw=pmZ zUMrT|x+e#YLkf}oU!0Ib(6hETpd~nqS8+EfGy!^-ZiMhKAusn`C6KY+t}qo0cn0If z^`9#9b6DLC=tzoR47_?T$ra5s>4i0LRZ$t99vlkm{SKpBO^m6aj(+;@kR)HghQE|E z-~=|-2xEto0+pz%C=N`Fb0M6OiPAOI*hre$4azF$@2P$%C$f?bHN^gLXX+_0D~h7= zC|1G?Wtb_*)23!k#*aaSzH{f{K|kR5-yw>>HADP4CZ;)oVL~#%5c3=sKk6$TL;y(U z<r}OjmxQhVE$E505P)R@XcSy=b zfJ26@Def8|e1fZ85n%Xw4J-skZ9+ZB8Dqa1z9c0S{9PYm{g!%|G!$lGH+aU5brNPj z)%j5d=sg2Zr}V3HkZJ;wMLsM5c=Ipz{Nz(7z;-t=L{FXy$n`wTfDvGL1mrbxd#Ip$ z6w^G&B+Ziss6ZpiiTP)kn-t6v%6W|#^99WQ!cc_{=xq&8ch0AU`mh3$gEhcs-l^Y$ zfLOyBfGwq6k|Wt4$aSrWlK{ZZfr)b~GLy3a$@;vms>6Yg0%80p1=g5S)~1U@(O z{Sj`_7=8qfA8k^Mmw;Dt9MaDLY=pcl=Ef<%G@mOI6Q03*`Qoom_DX4>GFw6-0GR*s8fwAHZJpV zww+#^Ec$}_dS;Cgt^Jl1q(4Cw0Z=#BZioed`wSO!xJ7$81+Z5nAq5K~#{g5t_ZhH{ zodeavuQ$zw|D9R$xo;n9YrAMHbwyXZx$hoX-SsvKt|6Ax?|QI63k@zee(a6142AqA zYe9f_nZDOUGz27B^Oy}Uz$^6z=6^Fn_Cy};1=KH&gaQHB&HPtrnU_Y2b}csV;+}SN zp6Q4xSj%pBsJ5Jl=n-ZL&B%6{+Ae4KI*ghpC2bc+==nbu7o(ZGq9SmUN1qF(NO4;k}ixc$W1%Xc%B%)-%G4A0icO5XmbSFD7ObTS;RhuXQAb$5_asnsi(lndND>mncvlUyjS{525tI~AU zelA5kMsI5f~E2XHZb9!#glUwwSa^GEvo=xBy(Wawmg|{ zCLoc06|1fFKy=38(LR}_;nK0ao;%M>$5idTqF*?yR?}6P?aq8&o->8IBFNrtkiDBt zauSFJF!#k^VN?QjD}_p^0rtCad3XMUT`oBE+_ucL1$d??8oe`Oza~w#Dt=Lm!`KqX zFPh&xq(8JJjI7`L-hyvuakxz#S?{vn@OC3Fs*&%O)>Z(*4|bv@r)WF(KpqB|MEt&u z97Y<`Q7*YBjId3bBjcf-BB2k&_FcM4n7_^3tT{w}VJw>#h94PeL6cOjOY3XEj}}*YlM~>77m743O8v>QG>465>@@6Ayd?4ZaMYUM~=TYSSYQ5r8yf zxw}4ChlY}?^|M|T1j-`kQ6gC`BPtor(>n!QY2HPa8pnS^x^QBnY%x7QRvH2+Q_BIe zAkp8(1#yE1A!IGY8v@6ei@Rh1(&FIoB`l1^t*VR5l@9ACjqN7vR>#8fu}5y+TOmXg ziVdLr!NTcgfk5e|LxeGiK&d{kj6lf;P=2wn>G*;`nO+_{#R53Vt*}|_AvSU; z6Um$BJ*O{hq{MkGSE(VR4COY88a;h_ZUqva^U7NG4^jtBE+Ryh!gL(ghu}(w2Bt_M z^P;*;55XZV5}1+0D)v8mnLR-EUpUx`{SGMCQm{& zLu6QYQ5ApJT?C!S9M!MlHkjnz`_uVog4_(>H9gJ*?mPXGkGx(fuy1W9zxOPqoHmH2 zlY{nDD9*^tfRh!4Brr7Ce_r?8M!%V{kF1lynCP^S&=tIj$jrx+FBJ;6+Gj_j_e`cL z@>DtOM;FtDgrf59&Ai;U|EU%$XF)=hMHB-(AW-lVPI_8#83KjNr`Ztud=S4Jgc+p( z&3d9J4LbRdIVEekp~;W6Q!L@kLi(_kq|Lw$JoQ=PZYt{z=mKP)L%t+;7(M zvM~F(*~16D(D%g%1^3hPfvP?f3EJcGMbR;RLDtAjM#&I&c?V{$P7iBWfn_tF8E0ez zvdx1TwFCQW>4WgKEk|Z@HJ5um&q&KppH{DiX@Ayl?n8S zSdc-;ZjILI-wQgGW%e3_4#>;nxp7+h&k)K?c=}4xH>@~#G+vJL7)-SS!Za~_ksty7 z$MbMKIEc{?5W<2-A9;&c4MzMK$@>0-$ScX$XUSktUHOXD9~NwgxEuXn$iU>_k`r{{|_Y-$V&dpqW`i8q~-ru ziyrlQ^mlg?k*?4Co!Um#c0@&~a9ZgJI@SDmEB`^g2&sTYFb|n;z#nuq1B3^jmZ72I)!2KwYBQ~mZuw~ z+9J9g^W1p!?oc!lgaA6Xe4ws1xrWqU;w?YMHMzoZoj zEc7*e)K0;FuB}Iee_+BSL5c@g2R2quS+u|Kc_`5s)<(Lhk;kiaq5NK>)BLY+x_sGj zYtqcj8@_4saIDw*UbM^p%F$)*p_LyNi)Obl_)b8vY}N}ceN|{Pe0#xLc6XeP1(!2- zqyHv_AZnn%eECiAcJJGpVESrW={u2tKRZJYw)FopeR8q@mC*K3#2Rj%iJyqZs@&?@ zk5LiP($Qy$)uXB$W1`Ev^$y+wqs|)}zLV=Wu~2j^+s}CmZJFcat-#ba_mUHD9%OjV z;}dUP4syI{9gmWu5s-x4LhJ60+G%)kYB{REjC6O`y*8A?#8AU>YdTYBXvg-0?a7Jm zu5=%12h?d}V?pEXT{dTnofU&b>pc2c2h4adFHrkHYFvLD16Cm9FUnaj;+awRv|V<} zMat-&knwX~o=fJ?uR0#g=B=RDc>c-TeFaQ&mx9SpypK0Tv?6owX*oVB+~?&w$*_!r zuJPdOnHLTX9u!E$mb7Q3xcrHHiDC85nRWY>G(HZ8hm)^{=!QSujCls7+nq3^bSulM zp3B&^NbXp&{+`cj)QU(0kFGhaQ}FwO`|&Z&;Z2!=Qr(6MhgyRD(?4>%%eWI{L)!*w zq1tToADuzx#)Sfhsq~EsC)|@*oO@j;3L#W)repKm8L>Un-G|H7)_$jrjRt5?@kEit zr9wWks{By2nte7C@Nk{r+v-uXUDXS`yl)xztGOh{Ud3}n4}2nVJAp`qaSB?xl~bWv ze_p|;`miDS2**@31snf#+~n*(+|qCZ}nspx(DTfYMJ1Z#C!&> zp5W>3_(N?&(7#m8hUkc;kS-AtZ8e?^Qk4e17O(w9yN2k2jlf)MDEkU56lH8KL<{W` z9V1uhb*z+-2cF3{#~yPqAKWY0NjZLeD4IW)&UANf<*m|=0hOFC`jk`DAVGz7IvdY` zmCb-TO6bUOUc2?V4MPk4{znZ=rdyVC4iA<*(I|MQA_{DKvs$05IsY!BD*i63olPWIt*Px^P#C;A#{zYt+s;N z&DPQ70Nisn1zdYx)sjRkB+K^%yj^MwmzcCYD`oq9XO+`mWfQVqfX+(-lr!QjI3>6M zxX!s$Rj^>c^sWQ4A?21biq#Kh{adz`j72Ek`f_@ zKQP~}-FIGlr>4tIwtRu0wpw8LvHZ?PqQls?we5!uW3pAFiL!qBY*ECCk^GgB7Wka& znLC(>;>F#04GG!y%BI7jXUmW^N|b6fLDt($%QCsv*rtF5qu}%r(y|V~UOr!T`L9)a z0rBb!3)?GiD&Fnw*>(?#_(T)=y)$N()N6Vr!a{!4U^)U~ORGGe*TZYK{Hkyl2RY3B zX}l9^n>9Ti%=&)1<#$SGg@I1WCv_-`zS%jwSjeMDg)I^3GL&JD&}OjpqZa+x_I@NT zE2l=6RF*G)|59_2o+T3EPV27j?-LNz1{J%c&>aLo2Zigcml)b=$~#H}Tc3u#jOLd5 zC1re+qaS~Ul!C(tq8gvEN3lD_HUks_USCZFC1ovYr^4BUshnq(g?x<#wbs8VLmN!r z3>?V3Au6THyE9&zIlNM`|s%XT|Ujq2(Sz}toL4NBHJS~1DiTlbz1HL2(o8* z;^Tt|6aAn{)$k2txG#rBc=c;bxY^$5PN78;TZHhEIwQQveF-8!8ak8H1<)6=m@e+- z)aFqlBp0zBVa?}=<`T#w`ZPe6YQwYWFOW4xQH~0=SO~T9E%`y`CXWruYZ3y9N~x;$ z4()($suG#?SHAXynp*9ZHeOVq!j3@m9scRz|0%+}NwCozY_gn)D}2^z*S#2VD=tcc z_Lkp%OvgIcyx;_ubNXKNS=o)A^Fr+|2fSPHB5L&X+iy<3^3dKQwcme2eR%kc%Muak z=m@_1{t+x%2krZu!IprE2Tr)3dzQQs%}y-h|+?71^sHdL1j(B z@)T*9y8gL+s+sEzvffk8R>$S=>gy1TK?fNh0=BpPjwa8c1O-4o0U&pWIGwjmuAZsjBEMzV63k*TKc8Q_ z>nAeAJbtPnq_(()KlN~1@Lqdox1v6~-rxKnW)dSg!^Hu5wei!MW?Q*|mBdCT-OAoO zMLXzd4rEzKTD8N&wrJ@R zm=D?HTOmoJ-{YJbO>^a;Ex+Y#6L%_&F`kh!%jIf|(GRK@=NC#bF_PIt&j#vXVe~@Y zm1Xe!+8^7DmBZq{wWW;K{X~-1%+;yd#kFX`&11~NZ}Ab~a?W)7sC$bRp``w{pPY0c z>j$inB3HrY>lzko?iTwOw~wzouSrhTcA9B&Hp}+;h}wa>9-NItOLyo>^X(|WU#M3% zGg(;4Bw!j1@+RWZRu^aT-@>?LVtw=CI;0Gkt$!~>If_Rl-7Xw0Jt)Ma z6A#Fe;Rj?<&F7aN!$xB=D`Hp&@BXphEZR;_V|idIZ|kLr307PgnnCP23L5`N(EZ|I zdN|PSw~AIg`vnT^fY+DpdlwyxXUeu8R+b%UqK1!>neI&Kv-tEXBybLAw+Vr6PkktN zsBKo+UfsK^Lrt~1>;o(D{c7d^QNS3%s?d7Jlz<2 z7-_%i<3a9)#DCVsPpZWvL#m8|S&i&I;+U54J)yX;Q&zrW*Wx|wxVNglU^Z3-P`OH# zoQ=NZ9}OUvMh{z+0g$w&N&{1_5L2FUv<}lq@YP*A*hZvk*hpwvE+IhwUyWC3mcxLW#Ir(v1mv3yLB5cA|d-c5cCj(rj z-$QkWXDFXPy?ixYk19j0gjcxef!Jzp&<{k}+ms*O!1&X;I53u!{%VwKYZ+*))*k%4 zLw8YWo(?vrg}oM4aIMx>bS>7Oc9ln3e|XrpFAkUo5JsvL|0!WT9$U!kj&yNkcBuLB z2sCl~f3^4J;ZX1I-?mDp(;`I@Dp?}5U_zFYq79WomUNC#_KA$0X+2q6wAnkdWjo2f zSB9b?Tee|D$TnjvGZ@2je`XNh@Aq8K?|QE5`~2}d{gta(KJWMazTfxqx?lG_>K&Vt zMF>rV*-!fM75r)$@>bQ~?E!MvXaS???s^Jh_F}?JY1m}hdPhg$%hj1z(f$1i6?Ic% z7Tr;*=(g#J*B*mkdZ`ID&v{W7UCw-xp}uj^-9EA6Sf7-_ek=*FgnUQRpNp!(r7XrW zR7tvG!ZZy^&Re1r-3&e1n`0r&-~=O8;u66|diJ$^3f6w`g?O$>H>zE2I&Xr4CzH20 z-}*miEr6X6pkarP#UMq>jq=Cu{a91^ZswX$x|1I}>>0%j3blX~?j#^$>|-2}b*bZ) zI{79GtUqhdXew~6fM2J7?RJHM{asVhN01~Sp%V~rF=Kn0Tl>A*LHFeMUWJl55ap!L zB57fWilW1|N!8q}wj7b`rR8Smr_pGw!n{dqoAeSp^rq9RIzUnDup8GNzf=dn&Lb@{ zjBsB7iXN_cSv=(!*BkJ&`@ZIi&2;TJhgX z-PKor7u|N24aIR{6|prRlH*JZZM)V2RIgsNa24O4KlQC@NS!=uRUc{$thxUn_^T?1 zB}7|~vz!MOSsdG)GBpMU&DjaH<)>tW-f?|9asV&-h+5s$X9zDTh z;ZCwc$(!jn?hF`9?1fJfxW0P9L|P6pBweI&cen~&ufI>-2D!B(wJlo z({;JHzrniIh#iS3bXN4zPJSV302DOCxA9@nmdK?AMmET=E431Q23yjhXRE)QfESZp zn`|=XGJe-A?8_!=`ObA__MHmf`ZU7FW6yJxy0Emc=2TXetwnuD5vQQ1{~UdmzvQ6L zRA#-6<>d_ULy0&1$aE=}+HGflVS1vtq2Sjx2x-iIP$z;78wPot%@?P7p$*`~Zamg_ z4EsDk5TEz!=3KkRS4JhCv{I9R-cR>x%$$CUIB09&{Ines!-C?m$Xw5=Cscg}25a>& z3(%Fu_{OZT@7`_MxXr|A1Lx{!rMVfm+T#)m<{}p^6nJ|UQl0=AI#7?F9;y1y2QQR@ zxGW6=II1@`cmLSWbIh9|4eYqrmEgJA(g+NFRgh>oY*a8kl<{g8xSp3M6nmLWrNqBV z*U|_=D1vuQX@hQ*s!Fjjp+&KE)IYp<8$Ifv{NGcYeG$C;awdwcX&nf7f&GNkY`<*i zxx9lJW;{J=viXcFs6mTLq7NKwpdQqe)FeF5gYR5%EX(Cn{36f5x7(U1EF-KwrGdU{ z(oJi>n17}APu_O-P7Tw^$LVEXqwfVQfti3S;-}#Qdm0w5k=){MJB3!RGE#JtLmEcK&D@Qk?a-1J!k=Ok|ZX7!Dw=MWU6|Q|+G0BDvgj*$m z(!MIh+g$AH>qSXW4=OP(%6AW3T9v&H!Z8F!-hLmJ52v)XJ^IhY0VH-fxZW1U5eJWb z{FjdZsqIdQCF&h>MPyujk>?-7RH4^%pF7FL3U_?kuO{F;Yt&TYC#d2FA2L~xP&4o% z>ctmZ^lfJygQj8QX{4UjZ*m3M;g+X>vzeZ9ef10pj{GUlbnyt5sLpO9&zQhDzB9|V zX*_cv3LtB*88NvaXFVm$PFh!R>svwi@b=SEKs1W8fu9azeR~sDTp+^-qNnKzLg4BR zi3uQ{DDQN7ioUsDLwo&9GSu&r4b<^gE?6XQ%<-YOp=SaGUJrJ@J=olYvg=#(bWia5 zC})RL*ZB#1`I+T#MRBr4PqF7>7SW88;LMy;D>C_ZAUBx==P+Y4~6um+NL<0 ziFNdB8u2!-E>2PoVwuRW;w^E#u$QoCws<|J5tZ75T#0w*;te!Gf@{MaDs`~`1YO|0 zS+{Kpp$V1Twc7iiPa28lPRJr(`)ymL#9l!5QGxFRw*Dedx>Ydx#U%W|T(kdXN+m8J zJUx|pdleAPi4C#Zm(2=Tn|rC-_r}8^p4zE`ya5RwH-j>sJ-gD9cMDRu-5G$g-Sye< zQ1G=L9Tcuu^X^J3Mw@k%fZ-upp2Ehx=RtN2noz1%owj?z!%PvYr8t(8Coe;)z_F|9 zxRie>7pMpCt+CfBD~c(fHPL;&_}30iX}MSWAG+b|EM*3ZI(x${Xm4+F(oLrd`*4nG zb5k<3oCnB0M>hcL0l0V%p0Wgdk#SzIIL6zcq z^QOR~|HBe)GNLNK@`8Z&+d0W`he93RaCD`V?-8sV(bv}$Qz69J=HcnlwTA!VmZ}gH zhca*VnVTYIGV9d|1S@2A3A5u+(kTy+_0PlkFb$t%2s~?a{6t(k>sf=k#BFh1 za2fnj&Ev>b^V0}Ij;i5VU4t~_XvvrKb&_^FA;_YR?H76Tkn9Y*w=gnhsrstr2q?h#j4&2rk;v(2^A5hjm#kfILLKAhG> zayW%u@;$#mf-Xfc)fW{1`#Q1M?Mg*(b0)^|Q?#1u6B(lo37?`=?L-jcqY*QM#5F|C z16Nqulvd<3DUSsbzCsM=E9K{gs<4lVp@IvGDrBN}dm%7eO`S%9T!K!Iw6Y(p#9t6Ra{b%6Su) zX9(WYuqY~whq(siIJ{YjoVkts+4 zkQBfimJOzHp#TdTc61~~u;hCMU~P_<6Sa0?r)5D*dDC&mKR%Yq5~A_dHmyq$a3j%& zTY^|0lZ-^3R>v;+>_63vStoW`!tdG`tg)ws|NQ(0h7qN@TC!7g!v}3}ntF2QmO%ez ztoX!LrE83f#Ykd3>!;fPpQ!zzEJ9H&ZwQrE9{)XPs+K=tVjMUA-FLv@BE|^p!|$C5 zzn0j8L2~P>c5|tTGTQ6yamcx&@Zfb4Az`RT3~K}RvE)X*Hv-XqV?JKGksZ1ASi%ZYzt z+n-||>lBGObqMXy+B+oB2V6QuLVfxSq2T)9gPbs5S@qf6$#(w2y%)VkKjLXANP0BJ zB&S+H5l1B1$eN}HU|x$&kZ#NZ?Qh{8vSD@St#O8l&{j`dbWU5vI()D_E;{uIib@JF z-27Wa+!$zawcs4yn}$p98GZC*Rak2q?-KgPv6wvoh-bK*h};d}@@|Ftz{e1c=$zNu zaesk7Id7N14*tSY*iB~rU0Z@YTho-}CBEQ7d(W7kBlJ^KqBLKpx^3zaV(42`-M>T{ z-Fd4(gS#nnpaIIA`WZL7XBUYYccaskty>#%zd?)I$HFO!5oJqdda9XVIF%&SoQGEb zPB)+Fn(V399ke^6CGq)pDCpnCu8ocZ2(&8u2;9R%)=5$vyny5Be8MLzo(lIjNmd+g z@zkVYykt6|!kmWR)*MU8{V474H{7EstdLg#^+pN0RmDxlP4@O1UF?seBQk7{nKu35 z6fO>QxbVI0xoU1Lug<^rp^>oO{P^d*(FL714x+!L^jc#?4s9V%S7W+JGA1iKR2<3! zCtGdJ1?v+I37kV=vJc1X0#QDcKk@uHi1N77qqR>!lox(qx${02sxmjlCFk85;!<|9 zM0q5rzkF{{xbJNJi@euUBSH>+{FrQp2T!T$^%j#`PC*P;p;`OfoC#8p^PO8RdhNtP z7M(v>@!eu-I%Zw1u71&_?l`_l+W@?{uw!A*6H&SDdYq4S$lvD9tE9R11JA&kur2QT zsZw|1zSitjx}a39zpdPUA0k27SSv{;A)e((!dX_@Uh!qP$ecUn&yUIUUd8oOt8kI_ zdU7Nu!K+jBVgbs;tN>^D6WSpjG8F0S96C-yY<7}M>D}H%%7w}QHj`!c2M$st(=)wiQQG^?JTvi4!D5Tm0j zr!!4P_^A8ksi9DzeT)xle8uPMQ;;{{pmahgH`Lh{Sfu);MvAL)L>znpvm2;HWjWC$SuS9o5D4 zZ@En+3~FQ-s&i}AngHz+#mR1&->y_u?|kLf_re{I7a_F^o%C2q{?(@Vs@QW7x!^l(7%KHbR`7kAn+}t&3uM0CK89At+5xS%1`k5~~{Z!m0F`XS}4&4}i>aTsq zSXmU%|G+^vt@fpvbe?B-?V}*=)5hHA&Vw8ptHTVX;r2;KRy&;hpV*;$6|?WV9y|mx zhfIr+>|nb{`fv+zl)D6T#lbG91MA&!mLNl`>q=8PX$_t}5-nukpvhoRX35@iio@-^ zwmIZK8F!-!3&yS|MDmeU=|j)HsJM|r6z|Yhhw8wBvV)UQx8Y!J|3SC-pLo_fT#fJd z&k=6B;(y&nQQw3qF*`?C(PT5l*y-mvt;Y119r?wRI=NK_ zsmeCZOSuv2h;;V`*^cd}jX@k&O%g^Fp_*miDK>Kj%+akl0sqpM@CD*UD}I67J#R7Z zpJK)ZvKMTX)jz|OZQ9s-c7mY4#&DwrmyL)d)g1FH)~wI^bh{|{m$tV5l(nb${Y%zf zDbrafhnYL4z7NW-WrVZ7QgCMmrUSK^1ja{6ZmSht06r z^WXnc+Lr7$Hfw~dBS(W}^S-UMYjLm4Kc)b@8y)`=Lo zlcp~xDaW6Ubcmghm_6(O?Jv^Jvweb?w`T5x>g6@rYwtJ(5&h_uIR1)zZ>+M^b9uNg zOLZ33WbUgRNCcl4TPV6t)=%&OF8s#^5rY*I2XUsD-bF$67L*E!6s>%^;r7u9EX@fFpemMugDq{S> z6iK-@MwD-K>Ie}FbuD0%O--o%;bM>{B5e^HC#K4apHuW0^{b?2Zx>uGYwZ{IG|OO_Yp`#dskq$9N8*$kQz@H{MTd4v95&uh)4^{@j^2 zy8s#RZo^k_hiZbN2dgFRPX$F!?K6s9-C4dd1L=8Pvh!_#v8lA~YQt>(Y9e~ec8`R% z`OMc)v)C^lj_3qJ7EZ`H$kJxKVnW}Q&(~NPprOG~yE2rpha`zmsPtvHbu+(=*~at> z`6)ebWpk%`o4rRmOVte3$nKQz8eg9(hW)PlvGDkcw8W%(5Q**@JufeILfNnvOu1RI zjsXBPrO%K>_V~;Vi{i+oLKo7YF3;<~-Qi_V6~sio)S4~LE*|K85V9Q@rR-pJ_k#7k)^O9lxkyFUEbCPsBmZg1p^Mp9&WvSk=CI^(>UbKbKX7Vy zXQ5uwbJsoWrp9`)i`fAZk7=laB@r7Oqp!mU;cHb+PbB*jv7X)BC7e0wZZnrHO3d#1 z&JAaDFet5246*R5d6ft?(A>OuZZ?+`tg>?5=eBxY^W2elMm_|zsPM>p^~v7|qa*L} z_JCJDG}>#i_fLo7tG6{)!yOR2@^IrL=pldfNL=tVv(IhGkpN^T3V!Qj7(?B^pw&T^ zM|^uWTOq+21{lq4mP9Q za{3-ljddO9OAq9=TBjDAh5iOGTe}-^Aa}O9AnbOb4W(=HyZ>fsGq6FyF<2r|#kxA|b^6f|ny?7TNh|B%%+ZC|~a=-I2UDI-GH zg6vDqa=GxYU;TY6`9dAovbykfLX_TBU&|QWe`u*KzHKOKRQlYn%(X2BN|MTI3iT%0Ws6L9KsiDHci0kDp0<%)rHp&%JWA zBGjrYY#r2m+WnWDtc(wB*b*0szFESDa!H=h1-wA{g~#|aj2+?Nh^G_KS=DVUy3=y2 zY7UBan9dx?YI^Rw=|1X&G+1H#!A%`^z%s2ZT<3L;**voB!5TeG>;>KZv|Vx2zBW;u zW3i5$Pv3B8RPn()@HIq1!_ni`M#ab&ftG)$tUK@oYsFN7NM&!TnDW8Dds;WZk5Np= z{IcW@a&M3XpgGNyan>H`w>u>myNC5^j6JAh+pU$$#j!ADK^Hn*PCD8gVZ9eRaM?|T zy-%-Lc2c3xie;GQlccB*uf_+RK^oJ*7Jkl#=h+|TTu6slfg_73o4-2&^bLAc;1_<8 z_x^htfgDXGSR`q~l5^on5DL|oOTsUKAn(2cut1aI3-;yq7HQMjs@M)wk`S%j%EUXN z_hzB{c*$9_FXAb9X!3n})gpABj-CnT(l-1v4A5S(=FlVv-<7N+=Z55aQp~r}T8W7a(eSG+}`L3CHgs5_5cwYAEwM*|FJBI?`A&r0c$4AU}NX@?&aiKre7*U#^YOF8% zQ;qe1TH~MAc+ofgw8qS60157Y1{Dkb^k-197=`>-gNhR!im3B0_;>dv}@TYo^uNUS=51ojt7 z1ZkOO&R29(0(Rpo|F2dD+_)^G>=!DZ^H=9GYlMfM4B2-=FQkMNt~BD>jecafiM_4r z?jf6N7-&S*qvC^C(N}TiufTCisViAF-OnW@iF!mbd;QVl`&b>C9jbr7U)rHLQ~EC* zn)qit=AOg07So{;HwSvFAwfR^=kic%_kkox(05yByd|(B<8c;M8To?=8{7+P&zI1Ec>HO>j{ij8MWHmcBeI?%UG$D5!v_5u>*1#xofXM?*#b2+85_0TK zOev|33Mk*+nX&Gp5|IsF-j^+hsl$|V69=Wy{c2`I7%Dei0rU6a?G|uw<@htk6@X<3 z#oBvb`iK1RM8i0A<~=LScXW8ur7lUb^YlR9Z97&)q~ZTervapsS=e{a&6gbYz;*26 zPRI>+e&D&2_cf}Z`%VE_p!wS9Pz$B8t2(>(i6LRUrTlBs$DHMF&AuH>tCMP!tWr)Yd7D=_vH$x6{G>x7{{ z`?W7l9smM}tkPBI((=gY5WbLTubABTsT8_{oLOdKp8faAYR*+%@6x^E-1ZK3UIUH` z8n`Z+nAf^7@{m*4WuADk8rHLVAMbiG)ES z8)RbHSTSp2Sp+S%&BZ1FsgafcU;!Wkkb)HOL_@o}49SO%M)R9mxLZ=_j*eTu(uj9- z`%;gxVMa#CrA4s->B?;m`76^!3v@!Di4vIaJvTeqrRTfVH_PL?Gi`w!m-q@E`YKW( z-2irNpd8r_y~io(@Uu=!GToRk$jmlQOst0*#%!ir=4Fe#6pU|SwzY57fG_zop?PZ& zC%rH(*Mt;@Gjo1YjQ}Ry%B_iar|I1cBK;!t21q~+iif^l)g9Bf*!Gx;86k++|4QbS z42Lz?3F$lj0;>26j#>ffv*Q)6<>@mtF{K?*o%{SkiE`|eK;qv?H5Z9!o*b#7P81?U){Rte_mQ zDocYV?Mvpwzd2-0GY+b`Rp;$%G;?}y%52#v2I4293~n2*^GOE$EV#{??&aX65CUEf zjDCh8O=LfM2u@Ve{3}nW&|o%Onnh;pI>Mw(%!;FIO>T!;R86wW5HO6UKI`T@huUK< zOja9ALHJ8B(5Nu|FhBmQ5omFR)Ds_H<`mN9Xw(FPMh#FZ=)GlOvv0Le&z?Z9BH%0t z?5J^6r%y6>XEkr>^^`V_+~C?^&(42h-gOJV59X>A2lUINq9Ob+w7njtG?&AynDD1X z%k;=k1K=6Mzs5QrdB5o<%_ebfHaEKESYmL?DfJ0aSzd14wppIj%Ysj7MnL)ryVG zSw0JaIuoaAK+Zfvtpu?X+ln~Pm%N%3%0&f_s5sI6EA$kpZPzDh#QCbAdB*c6$)-r> z`$P|&95P7CwdiCkQdOngULzOSQz3-vV*fXG;L4ld!MbvXYJg9@A-;QG9{5zHs~PMo zoIXmfJ+chZQb2fBDJ^rcxyi%5U8HQ$svfYHzwwj_U7+&4XcOTNi89qE$TS&QfxECH z>_#O_*3ZjC5R+BJ9Q^Bj69bH|#ooHkZX-c$hZ(-lLA2E1L2_a8g%8Pma={^KMzu(d z>^bO)<-ikTec+MkACDK&W=;fJ>Pfbr_6DJU^dL-HCMVy4I;7;>ZPiuXN;nH~%s8!V zu0NVja6NHHQ!`cis0M~|-!F|4CB_8g`aknHE=>0&h8?gsHuZoR7cR3i&AHZkglr%0 zl#cwYJVw8BZ&?t(*(T=z#vWu`U!k0BUqgE^SX`l^YHtI!CM6TmMFt&W^_eqA#fFAs zkCsyV?poodw-i>{XYMXc;qFM=_Lb*x>@b;e54@1L1FA#sq9=VAT_2`kK`LIxKm5c# zq_JHTv;!G>*91~9ixs9}ffF8|!E7`5;a;u)5*fCs$GFA(kj-fNv{gB9MPgs8xhTm_ zchz-ah!+hA0;ow_v}TSm`4;#9-dEZo?iJ-hoF%JrT<=-UV%(K+O?J2&4D!&k_Usx|pmf3Zw^eD6aYx!CH5ZDX{=Y3}qkMx(+yYLNc$B*hBNjrS$ G#{U8{`h7M4 literal 0 HcmV?d00001 diff --git a/contrib/machine-learning/assets/cnn-strides.png b/contrib/machine-learning/assets/cnn-strides.png new file mode 100644 index 0000000000000000000000000000000000000000..26339a9fa6fda4afbc7c2c6d3271326753c86957 GIT binary patch literal 75501 zcmeFZcT`i|_BIL%qM(8zpi%|tO;G7on)FWSLFpj9ga83S5s)rjn$mmk)k3cVp*Mj* zBtYl{63VxEkMB9}Ip6r5d+#6L9pjGsW|(BN_g-_YJ=dJieCAv$hH9uO+#tSBjE9GJ zL-Dz+79QS}Qarp%M^~=^XNY&MIO5@5m$a3U(NL6;VbE}Ova)?;iHG+*G#*5xuiZnL zZltAXgHMqCcAnrCJAU@tAD5o<21Y~^TxIT*(x}J^BQ79VITa7^`}umTsIEO?q+@wUx%I6AEY)Cpk_hlAo0-(QEVehpPuB5m#6Su<_k~ z=@F8)TeJdW)f@d(GfC`A@`c;2 zC;dknmVV`w66&`kQ%O7x*sNHu|GK)@DSN$Nj^X-stwGxL+waMjbVTSM9^+T^RtV{S zATO!x8VC}8;5tw9a-umkDljwjzl_7PgEqRGnC3|=UZm{&d)w6f}gLYz7X*1 zkA8U=O(^RienWg)ZHoN5mlQGoYciq2q<~wB@9FrOI|H6=v&ue;78K%<8s;P*j6N6X z<|vC~Fc~)bbT{Tk`m!xo$sE=1 zP$>GpcYX5fHTG+varS8v$?Sc*#zp*kQk~%5dOF9eG%lcD!4Wv=A;;*I{G|lzpySxX z1*`4vqZ(71sqAc)S35FZ;`W|<_;}PwQ$S1o`=j{BtH1iXUGecd96El0yEOUk{8cD- z`=PmYMeHPW3sP*_^5KAzeRW3X`T@XdRnLS;<_!b2=jt*o!BhA%Y*{HpOouAbt* z9@;rOJNv%SP`-?}aqxQP{0C$(fl!jx4lh--JyDS)j)td=?5EMQ%jhSU%`S1h=1z?w zaVDDoLQfFzh>UPanm-9IYv}E|F>&teduBXrS8~km>EGb^a;5ymLxf$uAw7}$0}o)(;Be+9XJ=t*iN`6()*)#6jokUb-;K) zn_FHrTR@@2*dc0KxF*<+-gf$l&;4M5NR}+_&kk8>pXPE~K98=AyZd_+i)XaHHI7> zk3~E}L_8ou0?!Od! zD=sT!F&y3a0Q9bTg1&I8j2F&3yinqAavqC6%FxX)W7Z)3<||BQ_Q(9APUrWF-rq;x ziOYDr^Cn#(<-XVG^oy1W-aXGYZqqnym8F*@VNSkGW*)j@xUAhN&f?8-#$v!yq!O9z zp2DvJQxVXBs{|Ta`0| zol>hUsco@Rs5YOnoGGbldKER@G-W*vng+W{(azG!-dv}d4!-pcdXt>Ah|Y5~X0_Xs z%emOY_g6OZ&Y}(i{&U~8eq}puu`f7-_H1BkK)4GUwl{Zcn(3STod}!zJJTKWL0ejD zt_}}j7TeuZOwXti&cz>x4vETQhA-RSU!m5>a`NnWn}PZcgt;pFYHraTA>DRJbL- za8vTJWIDKJfytl5Kj2*N-1?QMC#4T7&hoX}>nA_WzW78^S+gKz87S}G*1tVTnN1r< zrgLkbVNy(psa~L7u$|k&yE|0lnT^-4-5)8&<~$ij-#2gUT}`<+ep`(silpqS|D6}~ z4}d3&h`dSN;=14Ryy&@$h0L9hCHa!@MtTQwabc|Ubv_o6T>e$tjgZNE71qmlUC5U` z+grj&L%{d_1-}YNC;Ujwj}hXjVA+9bnT@>9j^z8^)lcKbH}%LMZZ_$elA_Ag#}PpR zn+;S&+VX5CMN|{tywwSO2g!O@tvlJ-J+o7GzWub|Ej$}eTLAXbY_>B6-xn<5C`lYz zEL&_NJ0j7$))*x5jpCd2n#3CCNE5#=zn+!5HP$l0@-RdwlX$qO_($;#gRQEp$~sF7 z^5M$HATu&|mF~8ESXBpitO|>28l1-^`i`;)i;t?QcmwD@&)mmgM8uGN|3rY~ZVA89 z6mMlPVJl%gA${&cwJAwm`$8|f^(Q#nuv~G@>pUk8A`)qmEzoyU;SG=J1}%nWpO7TS z3lu?XJ2sB1M3Ha%rz8?Z1bxt3a)WZ5@8<`VKTbSZ@H$I9j`F{DoKvgcqHpNy7Jv!7 zL9!E+8RQTq&tus1%P8~{;#q;4t-)Gi$4Lj3AMK9)aoMTzNMbd&2e{Gio=dXVRK$v3 zUtQS`L(Tf7o!(*2br2sYzSz6TdkM8?S!zi=t})R)?jENQCswdTMl| zF2V7%QMI|m+-Ku3;eOP|kF@fl3FDbHmDA|7Wlz(2iu$68!3xJw$CkMqpD%P%=|~@K z%={!(YZ+A1-hU3i=&~-+JtAdP4A%Cb-~I7S!n<(1a7S68Rql!T#Gwv)B3r>vtg_Jkpb z=80n&Dt)T^JdNODCl7n^qo$cHqM6wl6fe}+u6YC6Pyo}xRdD)&)J$(Qd#~QdOn0ou zOw2evoUQk8n;nQW3N$jQb8~EK)OJSh%?V?!VwBMK^SnN0`=79+O_H9FdI$~d?os)H z=4@6|)NvG8tjA|J$L&lTJ6dU-J7m{r+!*Oj1OfX=972&KJz+!Q9sYKw1_vXDD*xRQIi?OTkaK8TfJ~!|3wl_82z6IXhZC`oXBC&IJ;!LqB4zeS4eW<+7 z9S9!k=edsx=}A+=wJ+)1+ed^4Pdo9+REZeo$v-hE+#C<6`nY7w&MWJ6bMo%~Q*OZ< zCn=d?Zf9heA@1m+#=wjje|#HK-$!NEG#)@k@!nEj(MnYnj{`WqibsHdACC|?!Ux_G z_|*SAmd9tqyZp!bOL%y}ws-`;?@Q4#poHg~nObabi2KamGLH5CM)aPV+J{ z{BeuBy*Q)3ss@9MldC0z0M8?yM~o803=9lnt`=4zTC(!Ln*)D|GupVjJB#r0LLd+x z$YUNSS8HBAVPRq3NBq3}{M^7D+-^`u_ZME=j&4l<5b~!SSxYx_S6gRyTPH_`3%M^| zI(fK@GcsOu^v~Bn#%bwg`>&oH-F`n7@PND*S9tk&9`XKDHqcb;;;e{0d7WXH#7_OIH~u2cS=PiGR)4?|%P! z@pnTp-it^7iz@!1=ReK@ik2W2_R$9jke`AiQ9G4nRYGG&X5pKiGEe?o7 z_*LCf6Stq?;S-Qa{f{3EF9KLtN^eKvB`W-n_gq|XCQPbly!BrM0XJzq_|mS+@lA^L zzuyK3BKq%l{4MBj9sPgNj$B1uo?2}P9lOJ+TCeNBXh!DhMke8ij?|h zervVA7)KmpN=I^4*Pm6rJ5A+V78HE9;D>2B;oBoM$X@^E2o`~$9&#H#2|8RZc){C^ zqxVIi?Z*n^X*_-fjMO@${HT2;`)>~Z;m%J9Aqm(V?&pI4xrzMwBYD3w%jOe)zvk*! z!&;Q+k)GqSDjY5(f!nBaq9^+C`Tm~QJyA4ko@84$+;%i)IgmD68L5Uo-M+L}M=Jd6 zE$*iVr2yk=yh~RZUQ4~bl$>1uk>BH^*crTve@JYvo33d?L-4EHcW^>i>Rx=S=d_>w z)T`wiN`kh$3(&1MyMrMyt%nL{r+AvNFQDE7$&VFynm6ui3d*8#-lGFf)o!T#6vus2 z-aQFBj{#fvU)_H4v?iKcU{0fU&Q%+8B=sglFrWIJ|NI{BvEx1_6T5Jd!%}`YV=uVY z>Q%SeggB_1kH^uSNIK!X?aSDGHfyu$1}|NV^^XBxd}tKmVZPS%K0Opl1C(-pBx4Fa zxPg)In%-agOlxfOIkM7zYtte=K7n+OentL(C`J9IkEz7*#vXs3!v~KELlbYOrkrZ~ ziSdbdI5T=8&fl`h-bkTXORWYza`U!f8^umHoRfEEoOTm+^*+cPE;Kn3vy`G(u8V&) z+G^#6ESyKzIn}h4auA`Lg6qzLwTl<_zA)|1cat_dtWd)9%CI6CW*iuh=}P3|DG5|q z8KSlc9lJK_U*e@7M#*h~q^*sK5Y(TTP@D5jh0S*`lRK-rj#LEqeVC)d)R+_gTP*lN z@l!${#6i1RI7d{ZZeZd$5V5Ju?$u^R_rE}9a(?We#$Z%%` zk+80{>I(|59JL1N`u#Sp-xD@{X_m-dJ16r+4bJR9C29uUtS*0c90HwF!?>hlE5=d& zV-BTrYOoPh?`Ofj{^Z9sQ3e)^K`U@_Y@_-YcA9pWvvgp_pOX|URmfT@fjd!dKFRAz0agTA6}GHW$-a3xmp&3fzVk{LpUAc`9l5>vf`X z65^JlMBsI>jrxo^mw3nUqM@7#vwMG?r44exPmTCWJn2sc>K1xd2J=cjwymZ?Op&7L zO(GfJ?z6TDt0&r?*@-Xsx}FJJ%bp#bd=^|-^j&~#@4QBd)*1@=7Ykd>&UT#%ZmDl= z)_&j?nEEl(|8H~15Iqyhd7ycEjhq$tT-0-eVXxM1WhS@jw!pOyd7+#uP;^kD@|uPE z-qv-J6fT_%H4(&Xn8oKvlHP2^*tsrNmDDFaU9EHDDShK9PD?X-TkpO1JCl*ATM~Yi z`^G-##;iS>t*5i9D3=fFVtdAGju_s4o_laGJnOPDIMEdL#uqnd<22gKDUxwVlr3HT zJ(TJN{$I6FOdjI4^d9%qW>gL-aTIEup7>&@?N^gCQY{Vf>pQ9I)|;{AeX!3SE%azX z9Xp*LuN@_+LP4Fkj&$&E@cz-!0*MR8RC(RnDu5DftmZsy+;SkcJU|nALXv z6_Q-;V~l3MsNd~WNKmcKB39IECDfOXtp8YS^{ z1sF=c2DBjQzSTyzv$@rvLb_!zsfFuvw#>i0d8{QcHe#U3}NxJis{d7JjC!szu&%H$}~ONPlb^aAW2HvXWIb!0?wVTX&>6mv+EYSeF}`6iFHqwQ$aU{r8Je*MdnwZSHwyLSNXU(!SCi+& z?yWR7|CQ7+fC2Z!T)*K(xAlg%!PUm>_AfJ}Pt!Td*B`Q&VI$zQ&}-s&(* zgcH)2{zpU$xJvjkrCw`+dVgG$36YEh@4}Fpd#$_`M=vWA_ByhtOY0KqI5&9vo70#{ z{Ea*;g>Bke5P2E?F0Dpu?>~q$qRcl?Qv(ATOmnR!qQc8@<|_kr3PMrDzMYVMo|d|J zRRQCnI@J=Dbg_=puq;im@4d;1QWK*o|E;-|G>m+&D3d()FcH^2Mnn%oQ3m8>sBzHb zw<@fa6&17jplh>48$gqeyL;7DM1A4+9;@8ksWGu07_-a}_dIkcs?B5xfLmxR=A;Ld zq$-0^ygl$`0zVp2y-uMhnIuGg;j?DTzNNgxZ!BMyj@si<6S7EhQ&d?GoC;Bx!O7E> zdeTwB+iW$I&XkuX{u~Y-rZ6o@RY#b(JGOmUz7OA_GWPcPoQ!iz;TDK@sGIJui=g;x zLQoERjWfH2eshZXiqck;Jl_pd$v7)YJbFWv%7cQC=ae;V`RmVnLuA)jreMhV)`j>Z z&++n1Kc~-UKV5g@*fU#Glv@P!Ar4dKiL}1f+WvQa2ajcZXLS5^+mykxM2U9g;lB}E zlek^L#RCZ=ZHyIeOq;Vg^{}W*tA(Wa?@l3IPq`sF;?>(Q(2F5G9!NQs^RwHam;s!i z$6Q@}-G2OJ*wYS41%@sA#h?KC;p?Cbg6G*QNp9*+PBfc=f<9v}je-`uihg8o3`LUg zun`Gm@M~o#J%&B?Ub366UK`n(C4pTrdes=SKkuoWRa&(5rP1}5D!)hJ0yf(`Dw?^8 z*U`Pyo4QE9KtUv5j2%`|Zp@36e|0{hI$!Cnv&6gH<+HBL+weBkdH=<{2+3cGgIOIw z9KMEfqTe7_eHROOQ4kUosaHx{>3#1o-{S+p)`6}Aah-&MMaZSr*f@h^0a)W+{TX}a z+1#*xG`;1sb>ZudNPtLAalI*QkGc6vL$eeU{RBzb#lpAC#qX zZT-h(TsH?NH?NDDu|w_7l}wg=IEG3W@PVL}Phn;5bVfqM?re^@RdUdPq&#>t8x`~Z z@}Q8xbghGSNr97p7HjD|?7YpjOZXFSX7qAwj;hwpant+l7St)iNXLT)^Qu^Nw1&JqzlpZt@g-MZS&@8SWNib&fL{mS}0cD8jt*@^Hu+OrnWPY@5hcD=ZWtMV+Xf<{o}n06bK8|ZIgYM zQ1q-c++o5C%sG!;C=Mq@JWLL3*$$oyAFu)czO4>@)^mUtRhwlX#1b5TFoUpkqIuK# zTSrg$oOripFzv&>17ns)H0U>%AHMpX?+08Z zc$r);;`4(!peJ>YD0R_~pyh}40ql7q*e&bv@#C4atA?}37OEKZoW=TpAp4V;@~uV= zF}ff22(Gx4`i~RG*?n|P`w2M-`m?7514$Ki=I_)eJ1Mm9b_$nE5kq& zWXPsWrM3*Eh!Pwpuw%96F9Gt5UbIu|@_T!y4`t_= zk~)Kf<0a`+7#NM(d{z4iD}pHeJWxAmqU%3UbJl3z(>#b#dW!N`c7XIKWcX6n8z&B? zt$dR3>0JurzCQ_?Bumf+D1c>0PwBUclfR}!0k24$>BR*7W{K9371;SAd#6V6Z5+Wt zdrWMUxK<w|XD-ZdNburg2 zEOmHg;tCyZyf@ibzXR$%yE_Q}wQpP-%msImw$8L_e*9omZ@Su+R^*d6U_q@_h0dSL zmkO&-vwO4=8`4x%AJ31g6xBp`*e(*l*3GQeBc;-o9f1Xb8ZE2_< z1yZ$ESjwyV^oWKUxnUQ+x#hRG(2;UZL z=gjVx?@_{u`|5@T3oEJ6UKjM1^Ed2wm~#^G%x_VjOfSF=@CoxlrqTJrNx_4=4HSu6 zD`$BUj-{=n`Nu``p+!J!$k|{GE>ty!9Uq(nCLoAw%x2M46GlH@sQuO}j0E5J+JEuc z3>0-oEN4MV6Rc!}63+&r zu;*#kIl}^LGjWtJyrC5vmyA^_7)%rC1p?1X{KULFeTu&K2EQ^dbJxMC&sy$eLWa){ zMe^cgjqPaXackiaqR0vyS;SMkH13w1a~XEDTjYc&TzA@W053cF$DCrO=KeDXtG~wm zY~WjwsP}eMPv_SyZCKPpuJN5oI_fI$%;topy7wLo&Uz?V zGdc{Eo1g{IZ~8{cFaIXYD6$aCB_;7I+l8O&2v`_;546SKIRrb@MYoA89Oz&Rt(WZ5 zp{p?ukfc{tu2T$x9hOif5c%#oQZZv2^qEe^UD5q7+fiNMweY&*hYFv~&Mi*}gfZXx ze3ASHQ_d|O7MQfU35o2O1(%%Gl=V12BoMP;Nay9eU? zQUf~Vk|}gOK+b?<))#36&|B9`H>Tke3QFXPHxH)*t*6;uZZXrJP!9yy z*OmZ>9J;}V`5I4dq-tmaJ1q%dt<2AY13_cUcdM=^3&#mA!Ol~Gp{jJyje;NVW*lph zUsZMFe#Q|~H<0o?-zNX9lgv98I`P;BsBogu88oUt^{IAK+l=Gd9XTV)uuN~3QaEcP zHpBDflszb=o|to9&jc!WjKjyhP2XFc*@_{RIBTyf(7PW0p}_}u5!EuUo1V;98#p?+m%^>FvPDnU~fJ+*gym|xx1#o zAG_#o`Xm5O69mJI^m20X2Gh>6R^t-EK3TqdZCk!LVsa2_`5Z~w&|9Zo(my&PxpT{A z8~Kg`^h?+C8`nmXyqa+kMM2-8zG0;JZghDmbjI)e6mmRp%)8N4^zA*et)#J%rh372 zen*ohpVdp<`mzw~RkVaD^tglI2iNgJSsiy8m7vZ0?=Rxoyib!W6t0bMj^2O6F?;XT+@WUG0vx8n_#XmF*; zEQ0d|IX6q5A*AG~#nv2TL;~>yDOzfEa+oqe(|I+9i3MsZGZPs=@yuiQ1^$A7M`-|+F>0oE@MBS*gwMvtOegVQc^X0n03qK6 zBC1N`IZ`lWPtp>O1XBt?CC9RUMLE;eEQ}wgugJH^!*=40A)}IMIQMf{Db{@H^l*Bt z<$QdT1oO2~t6JuFm61H4o3&JDcz2jJLwLykZ0b~9m4my*0wP>Kly2Gv-v^2rk)CZb zg~wp(0winEHG%`J4tR~={3;LCne_@qnz}eG;vB9$**>N<_*7j?6ESRQ&MqM4LjEzz>Bd4gY5}e9rd;K{O>nE0KlT%s0h!CkU09K zjnuYIvy3S$$;a{19}8_hIrDL0?1A7%#NQlztAg{~d}280;5sh{I+Dzjj6yc4nRsyOwRO7h|}F_NA~MJ_Q<@BXhMf>FXpYmm^fGN8#S2 z|9O$(%}EHv;Iym=D+n@JmZ&EtYx?a`pI9!YV*VbV7P=YpRmvJ(_0flW=6l^~@490` zS?h6zlhK;2{N&?`vA2WwSn1bQ`cHVYbJY@g(x_Y|?C4M0$Nu2cLSQ>Xw64RF5-kLv zHNB*gkvnOb5qlsF1yORfIjXHyM^LPN-8?ko1iN0=>S>o#ZRY=Tu`9KUrFF%AG{?;v zd2gHuI#f=F8Rd2T{Mh!f{UH~)X@16${UR%KGWw_oL_Gn_sCy2qaU;;pmt4vYJP84L z=$qjS?|*{HQEapcXG}Lv9QzgBVv=MNk8kqbtx)K_JxAr1sN1y1)B4OxG~*pdGdznz z=W7}6;}hG;RPmz^QhUIpfPiu{CnhE!BNF@U3_5g~AJ!;Yi>jdFs2u<^Nghr11o^;H za2qu##Wh$kZ066x!awGuiJ+d8r!yA!bBX>L2gLzAhJ>f3(||)FAm>w!Z<*yBYtxRECZ`G=D7jQS)wBHIlqZh{-rj0fws7&()7Dp=&@P{YNRQ2{fpy&5Ol_jdabAq z>PKGBe^x+IxK!_beJyY9Q5nOUk6p()<7`NB-LFwS88EEki0$V{^p9_J6QH|Q*sDH6 zmIq*1xw_JFn^&i$@*&*ZSR_~ zQbJdgFX1wFGB(5s9xywZ-t^SrU2B3Rjf#Op|>6rGeYLhqnl6B zxFFrE47*I5st7^nsL&+|eepEQZDvq?)UF^ZWK@5S6};p-C%s37`8pt{|9j;!fQ{fn znQ7h!5p^p8!RHRMZL16B*(@%DG+v+kQ-vSz8wd5guLnRvZpfi2Wlm%N2h88oQhPrjeuh7vh#HkU@5kBChpm*bV$)I)*tztB9qOL{=&wc&5!_h1s* zLVC8;SGIzpcI19^_q%#wOYEn;{FNLMB58?LVZS-TICdPNQ@ zbVF(SqsZl%8KaB)&GMtWbJlZRn;lbLcP2JKMe z1@^9kc*R)OzAq#^Znc#3Qxio=g^*HfkfHz-;kOFj^#8Ot3d3eGJ&3v}!Z zzc=rAa3u4zpmM@d^gY(EklLwfbS957(u0n@?p0Kd?+m;bskT!np?2P~cWRJ>pdN9`?T(Tnobe@b;*?Xkq2?&M`5V!F~jXZ?t)lbu?R zElxeg<2Gxj`=P}aMPK|ZJv)V|j#$nqw`%iKy}xM&)QV8UA^PIN8&K$G1mxLnwIXmS zB!AXMB0$d@UEVW#`Ec(0QJ91BP!IoKa(93d zAm)NbJVFPHNS%H~A738artO23N4@!zqK0TNjE5FIlN{S??NU#ANY0{m=ipjyt%G6T zGg>P5vvoG0m;F?!)!F_L1OD<(XQCoS96MoUB-C)dA{v#s0@*g05o+b@e`Qj2Z2+FR7k*8Nkn) zPBl)G7sYZb*T1FBiD~;A=xye(*o>Kzk?@kIcc zgl6R+_HcmROx&>V`fIAnKpgN+#aM1wb1=pc|LL0B;mlsTM-tW+dHQhl#x`7#~KT>WM&($u%3N#)NqB%>%hX&58wu!%mJK2^Q20A!CG z4zy19ZgzkEE7mTSRXqkDNxCE6xZQ4kGchOO7u{I2FzH{-0$8;rCafn--;@i&nMDN> z{w@;dYrhUH62TX4X8XkAJlJ$Ze0+*oXo&!p2byu+;!ZLIButVg`i?EjrWh3f?gt^^ zn-AAZYrF}pOE>QfiJfPP9ry3ee+8;^%`~1G^iFAi3jUMsi(Yg}4MM-!?01q6`p1K6 zT~aY-W_DWh2o^v8Ahy%*HvbDKF>?bY=J|Q=J0xza{V~#1VD`;4-e6s@-VAKVQ8TqR z*X?4(5l%I$2JkfTR?D!2i*iTrHxrRDcZr+-u@9y@0hrubB5O|bHST|CVU5aciQ|;7 zSBKPkFm+9hLp4cgcoR~jSAv-J%TogEjFGmhtZguQ+`&*;JKOrHQx3)O&i3J^Qf(`?cgO(!vhU;9lR z_kk@sBXya~|30a|^PfNK)l%VrYCA&u3jV|1{#wTY%F*FapJG5$rnNWZ=3j+O~j}hpH_EP@_8m|UsSe-iQS>dFdOymDKck*j~{cF zFNmf->mzG5qxH`A*JL3|KnX7Y;RP{!VraGqIiLBZ>?#f@m_Yy^ZSXS7;V3k@-cCrE zN)|8R$+br=@wqmWqf{)V8==fh(s%)xc!B|PBnf0TJita=w_D@4q!>7XaBs=1wGR}+ z-3jWghc$cUQzltTk1U-9ZZmMUaLr!*05S;G$Kn*#awO#q&@Wyry$@vX)9jaD=&5y%d4Y_q6p?mgnzWY`zI08PnRg=%OM~#z@N;S5x_Bxp`U*hmH zFC`1WBh)1bN&UturFJZ&AFBU4ko~2G8c6T8F{kqP$g70iRag0h<7+O#6UN^O3EsFjD;J!4rw*ez+-`x$u9m1Q&Lm-xG!K$lb;QndGazjfXv=&re|K`y~e z4!TD4x^MlOhQ^xkuoC!nDnOWW8t@2X8}2E?Pr5uun>bu14 zb~zjueKRsndLbv&PrXsy@``S>t^=(!W=p-TNru9V#2lR8KmPjET_Ag0(k4V0d+v5p*T!M^ zPZzcMUE-bb>YLN?2;W|8M2}t-QZ2?{ByL8}t4S7=H(h|G!7u z!moU3;y$(EXJ2HuC%#T z+m3TiQaW%z=z3(Wa+~@u-#Fgs;;Y=K?CUNpDjx9_TxY(709||cl~LI0^~#f+$R~Ro z)vNRM44=O8C`HzVItO31K2KMD1kPd1pyY`ArA`)8f5yrOa+A6yi>&Q4I z8yBHvPj%<=4M}#3Ka$~Yk$tY^${=heO{^=Z4Q?)EK`cnx!I72AJWYM{{wEK7HY=J^ zC6A2#FOU0PLL^4X*ay*!zTAD98M*IbQ|9bcth(~_;e^=xFVFPs`B+LXeGa$&P60Au z0)gF3I$|-XDeW^TZlKaRk%wYt#f6nH-W0x7zwUU?3vE`$M-HNsGYd#1j@Fr#a@0j! z+sHNpj_0Hzb(x7Bq)LM-y{HS?#NIbjc2_kX2??OGV}YLcuh48 z;qR^_f*XkVTy*CmPYNA%rUPj>$9{lwBjI%%)d)bXFAigkNie;x+@0;ZUMfs;SpO&q z7S;^!kC4HRMvqvB+IcMz2%5{O4Oe=d{wx#tnFIuaL5KA%M@evP&eUP1tG<9Ap%7ov zAahKYho})NA4J}T7VWy;SzB@Vxb>j!W0(#&&+N~h9~rSs^!jM}$Jg=A$)qR0AC`y$ z2L3gFPlDzN-`q3Pwa^IM8msV8^moMQkUW&=k;{r^oC!l)?+Afsv^oCm%^UETX&pxy zbi_cjl2NRE&V9GUD$_7N7m9G(Y$L0|RQ8ltt?{^UOOGX8kb!|?%kh=9$%-sUS}>iT zteRw1TfVhw;l$MGP~2tA!qCjnC2~-_t66^%pX>P4Wy_;5Uc!%nEQTEcjclJgmQ>v( zl}~gb8jp0DU8$qZz189kV}R8>ICzn##*ul*E`GUQ?3#uVHfsP^zK}Kkbhof^_6?~x z*BjpY=Pbasef$3Q_$Vg#pc7)tYp))5U3w@_ozHYoUoSZBRtSG@yth870StX5$`)a7 zr-h==!?+F26zX!`JDYu6`71Mwpww-^W1?>bWW3_IHEeDRJzQhG|*(fZqdEZFbq3gdFsl3oY+j*nFcwcU~SsN~mqg?pjoXr_8 zK^T;{(j>Ct+bR23L+KBF;EnR8@rl%@rlj7iJ5zHMyr8mkdc6tE zSa^TdqpPG6EwcW$c3@eaxX=uaBfZOpne}ON>cXyAbU;!peA@@yjYQL5O zGMwlxa&91%JbCUp?Lq2A!#qujVsZ zu9wut<>s;oBWI;mjDuo>9GTnx5UCHP9u>u|Wq_RjI#Q?R_WxDls@cDa!ENDg!62DmkaYw?B;wd%~IOm+|B!# zckCn0W5KM*CxXiu-c>6U(DIj+@&#`5<;L)eV@^n1Wf#n6_iF7Y^v*<=(%w&aZS1fj zfbi!r!&R-J0L4XsIBSPS)rdJ-Z1$p3(qrx9)Ygm&5)cW=RD`j~RKiwhU7H_n#k|zFf}Y4EY)ZJ;-`VZObaDm5qTN+SQO}e0%zo{GmDmbxMmckbEq`VTt?lr zN-jznV;^Su_cG1+_ZYkSTZk2+?FmDd)KO)bayrTg1Dy}4T~-(7=KiO-C12w&Fk6-L zONJJV1G>EDk=<~w&7V(D8)zvfXk-V(-1mGDz6DPdr|XA_V6csZ3l8nxJF4dI<`2-k+ecv!&V^YeDJ$v-_z!oUTd>N;MDb`kUF30ITG4 zEBkf}VJMw>6pJmgxixkhN*VtUSz!QBy}Ks!!C{mKJTV}vKoyQ8XbjlOwD760m7_eZ ztz+jFxPJf+v>YBtxFB>dRhYo{CHy zkZ>BGeof1nRRHEq4)*O4-$FNS0<-i;X#f6bu4?SqgrWM`(K0FI+@WrkjoLx~q+0Kn zwD9zEz`I`kH}AT3u8!huQ)sdOCzvby17Q9SQjpxk_dj}822F2bkPxP7I3^WmKc1o731>eZeu{867TNGaxOC4 zFh(1zSIFF8xhnV7VpT3O*@q&MVS@PfSq2>J7v(+CUFA(EEI|q|#51-%i49D?M?^D! zthv_L{;(7NJ*|=+KDWoH=WdVv&R;s_0xeG*{}t5X1=^O&1GOJQ$ypnb=&`E2Sr;7l zNRG1gmSL3m+1pfK)EdA2I2U9_6BZR6t1dPVkippNjly3f9UHX(4}{NN;No2X7ZdGN zBa{E|9%n&RnRB9H-3RNMUKitZf+NCm$?0ZNoVc9{Q&~SE>3?%OA9J_I9s~`mkgCs= z=1aZwqx$3MmUDdJqI2Q-1}YQxTT^~B4%ss==VrI6y-{n{7K0nY=(AE2ssz@v(-8NK z$zl8d>rw4G!oe#JDwt^N+?UluS$J)DH|5hXi=RT*C;_w%|KHHMUQpla1PKNd&u`$KVcUwk}OfIyN~V z;IV3P%)nlh$*wCb_uSpgb`k8A(;w}_9)p12WcY4Utj#vWq8A(N6k;4H9e+FBKjP8_ zxOf2bg_0jLjJ6(oI~j$C;f!~EjagI@iO0AMX-(xF&r2i0Y;|*f!A$_$IQ+18<)+R2 z+k@NqyF~;tZZR~YZ#it`X$=BDH*~jDu;^4-Aq*c!e;6ws2ecMXP@x$X%c7oeLEAbb zsafZ!z{cKOEphIscEmGgf8s(Vk&Pm`*Qtea5$5iMY)S4pHlQ-RhFsp~r#Vhz&CJko zr@05JAl13TOK{^KISR!Vl*G+JRm;) zb-5uSN{CflX_ss5s`Dw=RShGn2TD`xOwjn2B+G&>1Slahey&wZ!Aj86U|qN&p(aFk ziw(67vPt10J?-(gs3Kx-fBK}PG(W%oLHn>% z;09BcnUc(%-C-hEZh99r?>L0P$gQ7P1F}!G{mg z1zeG(G`!v>H4d`VfgXF%$?0q|A(K zk@x_b3NuLXR`IiBH=Nq!^sw+WPpnCgwKoXUKRDwDY91k!>|}uHNRx4`UlO)7kkS^Q zAY+pi0+fxz=%E>SdY2CTr_`wh$mhg8hqqIzA{IsobNcxEqETs4sgGV;Si~e?o z?i!W1J^8UJ#`j!?h1db>&PHsu@CaKvy&8SQPn`Qi_5nSNVPf|dGFw40otYNDG=mTj z$R*3}1AACjNma#lkaW_kRb4p)LUnI+^4N>BPE=A>piV893L9^6YcT+(6Ud-HPUzZO z`8xNN&>a&|?!8d`8eHCI&L)KbK^U`y;c9!q90ERA~3 z`7J3_I@&(@YOE$}ay|c2ri&I}-HJUkJ|?cc^OLzREbge-hCGwNN|mlK4Jt-`5K|Vp zWGqPn2s8-MI>bIVwdA!jK=B>N=h#@*mBC}CB231;)k-b7UykAW+I2mc46~&G(;<{D zFD606S>mfsIkP2cQIdr<$&?z<%-*>XA*QVP=$DcmVLrDu^+y#*cyq7&k=6oqm-FC3 zW#$`rjNV&5@;3|<{I^)L)c9R2+vS$C0qg$op$^lCh-;}({oc)%G(r~SyB}cOr;r*; z#|%vXR-^qW0iIMvma}mX)7)#?UkNwgUjMlF7`AKGs;XJ_&8lzKj2jSyK7aBR?X$`$FJFD!8t?RO)n zn6O32W>{|p-L+^P2Mi@k^*$0XKadsBmfErJxsymI^8min>YT#RmqI1|?T#6ZZz8)N zIWT0UIAU|F3)0GItYlhlpKm#Wou#>lx&XTqx8`g?8UI3D=_hkzQUH%-U$te4W9_u3 z8ehKv$Uqk~l&MIP!w4e^uCrn}akGx%sDRY>>`-6V$p4u2`&t zcH*`QIP-_yqGasp>=I^zf@FFW;>j$lXee+`q<)04eh1K=!V*M_BFB&FC@P;Aoy1iX z@uq8qy?WN(Ucf5K?!>9;x5lgd#mx7WI={3Nx4@@@DlnX8Xx7|;6f_lNaoQ|FS1CID zdGsr=A$WCH#31vWs~I~1jVTG^$YW`x!*_rJPkW9B#wL9?ZRr7*FBT}x1jl^GCR6bZ+ZTZLE)2d1A5ev z-4A9t)IN8f41ShXHnV$bzA2l(5~m7X5s9{zNsNv%5yO@NMV%Cx(JP216V9RXNKoti zQ0{>a6c=iG0qqEVsXV>*r&ejBmwwzoaNB? zFr(EpTxx}WYn-bo&3YvN_b<^D8I?La3DI}$Ds`GhpS=YKSv^C+NC^(E{DK-%e@ zQv_Yj#?3=efIgwGvaW%y$dgY*TFZnuDS)Pd*7;}q1fj^XKRO=qc#1UHm`~+Ay#Dvb zp*g7Sw@X%exA7Zrj4y(QBcJ$j4RKKQvK z9%tF>snxNfYLM*FFz@-4z_u%u6#3DMNIus5MpNpQ;WO{O+agDU12xV(!+G&G#rrY+ zm7rUUnsPTk+@dt^PN=M|b~nvR(=ixr#EY8 zM9A%SC+zFnl&d$dCqYjGM~pkT1!IS_Q1Sec0$Pc_2cAc#K@2s%=esxAF0PNX=vkQL zgF^;7tXrq%-!;cS=yMHi@v1|QBvy5pdsh4AM{7X+{Q`3r%AxnG3xQ`Eo8V?f(!$Jr zq9_$9nnk(>phVRlWEbX|-?>$EmaNUSg>eu|Yo|O7(@m%0>5lgm_|13vx%n-+-zf=B zxBE61iEabR-A&MCeU520Rd?SfmSW0C8Z#O!W*u5vKecENKp~bkECOEV$cx;1(s@l} zZ9RbR?h$hX6C@*~3bHT3%udiRnrzDaoVf_xv_J^-8T+nv%R_DGXd98(i}%H@zxPf^#M}B)o2N_ei22 zBr7EfJ$i(jD%EoI9`1{jVx%+_e~_;Hk-}!i#$=B3-M6XVlyPkb7mm{PY`9b9poq;$ z_oUi|nt=1zmopUO<`i4)xv#P*yC4l*zoiwS-6k%KFM6KGRbO1*J=Gj*G2;dPC4(m4 zR9n;ft1*`Lucl#pM=0Q1uDB1MJL<`S1qwdKys|vfjdq`M%*n0p7Bx|Ft_%fy zF}MA!Yx^tj`$zs_UA_K`b+y31)JHwB4;W5~R;6H0e=XZP_KD z6^q)eEOv3V!T3pv(NJ^>^Wf(uD(-kXJb zW7g}H?hKyDTGGu^W+Ew#Q^)p)zjxp@Z_NeK^+61;Ng}uXZO2O@vCybC3OyTE>)g`B zk@x@l+S7~l@A!K}C#F70QXMD78sZ5aJBNH=l$c4bueMfFVqyrRy^fhuQ05DF1Ju7T z;s43cZ6cO(s*f!%aL0u34xU==d>XIknTQ}&@V)q=i_o`5vw}6hGr&ZPifEC|zbny? zn+6WnCg||@mln=9w|UXn%zfOQ4i{(V$}36vI|agLS81{N)5z}>h-~BKAt^hn9piuJ zZ{aNU*@B*KUSnB=`rT;$k8-w?XASj}bn&)juP0s5u3u!Kc5$~%o}oDU?mdT|P2Vr- z59T7%az2ZCmgR)M^{3Acrkvt`7_1Hy&w)hD?}DzC&Ktfz2uV2(UmCv-;e+8|Ewrwm zZEA;N`WH?FrbL#Afb%BYbm2N8H=9L+RxpBMwYB}(Bn{*T`;ENR#uJkqt_`%tN**tp zjyc9o)$Z1z4%Ujm%wD+HF7Ey)_x>UG{wnCzi|>C;?TTM;`v^SOOTaR1=d(g=H(VJr zHawePK_gzso{%JHBZas|AIGgY6qzS=&6pq)S5#Dex3co#5MKaSP^df_owzAm# z&GN3f5Z9>Kmr{^I-$XaHM=LV%uel`HhJq9DbvKHq19pb>?5o9;MYlafst}}KWLfkC zC(WFXW`h`}xMTHQ>ej0iwo!%h9~PY=pS6jv*Fb+KA|%uNoP-BhQI3F6;fGuMr`-Me z5q}U7v|Nf8YyY6zcx@h0F1)I1saUcd$!HPHqj)w~^*aB}_nh(*%d z+9Siy#(d#Y2%{ftHST3uH2Iw@cLGl>7-P#4=ra~wGygY^kUk(g8QtZD|FF=fusf6s z@X;R}81kuqxz0-oRKlLbu`sd*v&4zz>h!j$RiyY8!z2%I%0s&!K)yxkltlrCN%Efz zle;+^-iVt?euC{&j>nc6ttL>Z8>5+K9c2@U8>y0KlT#`DeVR_0`yrricKtiL82SH0UBtv<(M3EQgDy`hgdsz_$(V!>e{Ht|bY{`-ApE;gb5(DDo$* z5tdAeFuEJ%RsKls1A&vh%?}<=uw7YA>>q?sW;g>!HA3g4J2nzg%K{kZyxt#a?5kvW zsutAT&oKBUTzpS@tju%bF2I$Jf6$$>l85d>r4$W4%{SfGvQ+f4R17q|@=8{%_)MD# z%iFkx40z*hYE;v{SB}elF{8Kp6IJ$xVn{3qKw>ex8`kq2z~bBO|HiAU0Mx|#XNEPu z>z~e&2M|O4WIpa-Ks|u)EM*jc^sRFI#DL z8n}G_cM5B?;qsrnv1eN66OhT(r#{5vPu92>$8%tWATxDFJyLhp4f5Z`^WNVOk_!+ zf8B140l1zRfcwM7jbMfSKyZmeoGp%?SCY&2Pg^>TUEz>8^r_u>lR92l-WI0X_r z|3@?cj5rUFNC26tb%aQw21W?oWwu zD8bz>p8Jcl;DVM3>FOH0WWXdA|AQo|Q8CmUZ_DumlZr1lpwpoB{MeU=9#mTinKV9d zpMJ{Pb=|!FDsFzHojL5%cQs2g zOTJ3kMZ27!S~=?FOg6?p3ALKX_Mq)@(pe*mQOKzC2W#r1-0JCyk@tzj?!mngzV3R} zZnlfjv4#;u1UF|&O#+@}#n8^6b0`Fy@)D@`AtG_(WzJW%L9b&`0b>UH|1~q1>Qp07 zFD~SsE_o~&y7bAIp=`?|?Lu6g>jze|GLj-iikB)-)oupbiHAt${z6kZm;`~3sI1V% z!xkk-Z~`aWK4zb%HL*G4yDDEg0T7z`ROv{Jk>8O_q6-{$imbP{fuw)y1{#de_@YOZ@ zx=7}5QLVbexsILC1c~Una@bse$ml~53eElug`Qx8y(5_J-O8_idwdBT-r~FF3J-eD znyv>p>45wR*uStH;Nia?NU-DtCqgPmCjrMeS^v+?W*Ikc?MZOvy;Yzh0)5l!jj~8} z6;ag~?G$fg2AcO~y+i$w6YO$du7i3li?UiQGN%NkW3VSn88@(RETiBPd72(qV#Fx$7jj z9tWTQNjLsTSKW1S)Ozvx3$T*@5{JVeqP2;NagH}MWGOUp37YfALTr$adD4JK=?*xFF4)mlf#;&HA$UFUL*`q;4wt3=D5Zt+s%yYGb% ziCd%d;Xy+XJ?`Af=Dl0y!Ciq`GZY$mE(tWR{S#LeZMWyX%?4bgC#k_{yHf zlK@#kI)e08?6wNp+8Dir8t1*Y_SGUFATU(FqKuqiam+m^hIB%5B-ROJ1DC z(#-K;xawvBX8K`<9m12Z($jmG@b7c)M&qNrZ1qcp_?WQ4!|?94=Hf_1>m)}i zwLIe}lIG))gPGIXJy0L8APxNV?kR3AcWNl>O5>hp|F(=#&1tY?%1!)|uui7t8?2v3 z-oDCu6e*aG??|N|ppc)7DFm zQ1cR4&DrBq$a;*9R?}S>wxWx(bIa1qA5l}t^{o}mAE%)r4FC_rxo)QQEV2D&Z)>{g z7cw12s@T*KJ|dx%td-(djgRC@u!m#%WTRY6jyYEjQ+N%bG7PE|xVmyrNKLOq^;$pt zs*(8{4J5FFkVm26S@z8%A4E>ZPM6@zIGtPHBJK%T^;!?LR<7%;78hN0I3vT!PJs#nmTX3fBf4gYJufas|EdooBKLrmnxw6;4?Ws zzcUOy=B({h%hrabz0;~H)s;(+H?WFcD;>OIZ3GL&ERhF{#&^l@;>ATLXYdi+DO%CK zB}4IB^Wt*po2JmGR>wC}ZoRvb=U-}G(s#q)eVx2$-Ezo?QJ>s9nSkL)ov-h|-q_Q< za+{p9zghD-&-0Ee1<9NWlGCWwET7^~lP464w?97$rQXvOaN_ZfInG_jnUul>pGl@B z50mwcHhjXNnOm4i zzsD!HoX;kS{bGJe`%yGY62f<1qi(GE7C%AyHMLCH%U*fZBC!$*zFKbBn@P{WXX10D zK~n4Fd5hn14jXZ#Mcd^KN3J}H)m`-BF8w+_Vm+T-8%auvr~OEVrDa;>#imtuuHCxBTWZ>*I%4#8W0g|lDb>{mGZ70!Nzv%jIU1DpBm zN8}f-GD;`k2(XkwD?M~~n@GTo;d$?rm|987_(&pvMIiBF z@qgZu7C1&LYT&c{><4%Y+{W{q2%oO7|NH>($cQ&U2(uJ5_W$?>%sUzFfpDHn)^7hn zg5PBdFvGFYXSo`E@6u0Y^&}XrXcsdo2KxGmC^wjasBxKO@<00lNvmE%vnKrL1KrO> zr2?Y1JxF`>pS1NuR%tl4VE#_`$olV+|EDLl0ei`_at2=csor-efZpBfp2+=F)Zb)* zsC=p%?|$;JK=0*1?=`yuVm}v^4~Y8qwfv`_>%9ioMt|`mY>f1$qW*G`f9RL-|3ViD zb*u}a8mC>_?bxix#-hYIpY!$ItX{2)C=3IGH7e5}qA~pOFmf zxliGDj*@kdY{4rCU6F5;8k&c$vGSpgX&N)S#l^&X%Poh_tXBs)Oqwna6SH+CkQx@w z1*26w6_X1ZX60zf>zupX*I?BfF_k8u`pe^UnV0caf)*3AB z%Bh0?xquxUAOB==RQA+Z`n{nm3=>Ci_i40A5S{xa-kPm@qrM{sjVl=+#mjIj^vrQ; zgz^;!O$GY`2`FEJmKFQSrhE7cz>Rq!RAW+m%dv8?dHs~RGBPwLGBb2WU?~ALwiVao zm%qq0VrZSWWn^@EPup_5BxXM0Mz7+`oI)M-laD^S^Ch7<7XmM)lJ)AP;_U5|VSD9r zM`AIP6r56HeTW6|Fu4O;Qo7gtQLHtmvHB;D(5nU{9&@kR8tJ_^X#7wU7gr(0jzhQ_ zvDdU3yys1HXlXw+oS~gMhD z^rIEj&+pW^?z=Ze$}tR73~)gV1VE3H!UPN@nPqdIB>mo9%W+d@r+;M#V#KZQydQ1S z|M_9)Ha*GmgBh+^AMQ@k#X(u#Aw$#o?5iTCiAmn%xkn?OsDEl-LHccmpCQg(B556r4~Jj~>&eb$&F78JdGI~#CuJ04v(Jjt z#Smq39_xiI@m722v^a;xZ3Y4qqFQC(w)w{b$05@rsk5UVlc=t;Tp{@lt+hPV@m7*i z1tYjrw9;{jE5C-5)|*(UGx^-PD~=Bz+|`(T^9Z$c)yjyiTbO6Pyayh-OHcBYWvfN% z^mc;7T<$GodVCOJv7^9?uL0QDUODGtJ~3bEHD!Lhn?i(>TQ4fj_X0#>35qSOXY}W* zeeO4;-vi>(O!>gF4(+Uw~bloiVu_-2k(hT3${`9UePVn@43f^=CKBX z34W_>CL;!JRyC(>8Hw3vtSddx+b};H-RXKlWD`T@rTLa{`!^h@fY2OzHiyf-|6rNy)mkAFqy6=E7$Y@ecZJr zD-=a3Iz+q9Mt@)Ubtt^*;rw*u$)RP&iz&CM$`v~I2~(Iu%SZR29)`VBWMDYG>w|$p zWfI}7Mm3&3^5|s_@Z8?IIG~7OZV|qwf0)-lR0o8WyYTi91mh}8SHv5dA{*lz) zcOkU(;qgda|Cvh`@S6RIv%(ezVj47xHB`(8m2Lmhb=cEj{%GbRTbC${&tgvOFtL5v z*hEv{Qj(AdsESf!monIuPUnyNmK!u(JQfs%7-{HLuux~6b2*4f5JVMe9>5b=zRJ<4qN?Ht2<0CKb6CX*BT6J(Y@`041 zEPA9t1CiLu-(jg00{1S~#-q$*q5QP&jj#7<22aiUi0u;x_MViUmDeCn5qU`nlpy8z zDPM@TU1hzksNYQ$u&h2EfT=Hh>jm>Xp<zpcblou=i7+g%F;Sp+vR*6l`ei|Av%r!TOFjhJ8T@_~iq@(?Vf zDx4KQC0rZMJ86z73fL;#+9aD*YZfL^q0@ph!Gif096b7Qu@Irv{>xYHE>jinHJlvnu0iXbuI$=F*MbYW*9@!oEVIr))#d)_#2Ko^1m=4H zn>LxOaGb64hW72)l}d4Xb+#yHhU;uZ=e&$xg7-SrMB!WjV43>noM$=Co>ut{gN!(T zdTZ8s4^`k%DUX{m8#t$RZm7pU&>Fiq-4;mxa17U8M7QiYDU9 zE$}%!GLW+MEJx@&+X|n&N?kKO-9)ekB~AKScWf(Pxps3VXu_DZrSC1-NWhkoHc8H|Zv>GUbsBD+5} z)*?@{B~kNX=%*eCEL8csQJ)UmMlkbCgMr`r`ynbhWBLmoeO-Srra2!98)H`4Oa-^LQAm9w=Z*48Lt@!v=1MTR;f5 zaMHv$$54g?tZw>q)n@8g@%T!X3CnN&>DR7&?kFJxc#0)FiVn}F8eQE8^YbLx4=Kq|RGuejaPKCiTb~n4d@Ei%YW*Q+ntp-+ocp+j?X5JXj`#dt13E5D!r4`NA5&sGnHF$gVEeJG!2GQ*=5Af~Akh*K#hl4u%yBH+k z04A#m(XHVG%;Er@VUQo*a-JP~l`I#~nS4tn6VRDV$s3H$keK)bWEhR}R|@P-ctc(3 zPft_%5#q)(Q|q{-B}>Fxuttv{a)%!kAGf83iHpDiE=Na@v4VZhuMV zhe2lXLWjs^x;po+A(M^L5i4 zt^R`8sCs0BLkD;Hu-uT}MD~FmB(H8*2N613xs;egd^9JML>zB(Ql{0^vYr6)uzJ1c z^YRyLiSLc-J2F7p7&Sp{e1JP5C6Gi?we#J_M@>>LPleZe?|%q$eb(TB?N@O*V@_KL z1QH=KgVbmcNYdu@|1FSsV*-hU0)&$lfYbM}ecyoL9r8)I@!d zL=Vk2k4@9?Jy{8u@N@AVwWvS8s&e+-p7ll2^!@;5Zq4LJQ7FtMs zhf@dI_PzJ~dO?4Q<^4^sQ(VfIfH|4rce2O})v4IO=-jl^!7F8!a2`kZ(%~$|5&M8& z_APtbvH@wYvY3jJ7o4tYvExtA|dqTjI&2YQ~^}Am&w-bB%M)mBrxkMQTHD@IG ziQ3ex#LNOpTJE$ZCgo8=k3vkwOR_@6pi3XDo4`(h#?WW`+i~H2B1v}Iqex<*Zg1&i zb>OO^(R}%t{%}stCqAlJ5?UIahg!f>L01(M5fPY`glyitr2#VeqpNaNT*)m}Yb#bw z1RnO{H6f|ra0KBlAzQ7?;loBHXMU&`R4q`OR$QwyY<=6wlOx^@J=(r+8Zv;mPP+*Oa~($TYw8E*X&|N%4R2W#?&axpiy-+!FK5r%iCOmJ&lV~0UV&_jIAb ze0jbV?}SZpdC=t%BdL=W6)67n21*88b)6JM0})$oZEE^*v9^9&{N+W#{N`JM@lXea zOmu|z;q3I~$~D0W8@e#bIp$eSn67WD$VEIC&I02vja$t^Foxp;`8A_6-wAH88RYm_ zHz|y*guADuPR&*Y01yW3+y@(@ml49CAQy0(6r}6hCT>xFAYC7uc#K{RP**v7aM&Nq zn{R~rqmy6EhIH(+0^Xp%aZ)XU!t9`ZW+Ivz?frHqVq&Mj4fRb_3{rWE@A;HLqx&v8 z3zwj=#uNDhOrsUgtsA#8=@Li};!wBYc}Ehc_eXJ!oV4IH`?ucohyKzoV0n`m+!&)U zQ@=j(ROt<-awf}O@U$-lvtx0XTBm6&?vWIm0ViMXaMPZCYHH*WLy$Fx>bW&kUP~OU z2)>Rh@b~g#z#HJwjPnjkp_pgkk-XA1{%^biIlc1GoTsU(*Zxyc@J9_USr=5~vZKq# zeqI^`Gyq3(+dKF_uXAcBfYM;;j-Vb!r2kmAlP&?TNePuT`PmQ71z24gt;p7bKLIt6 zpGf;k3%sVJkNt-p|Kp1uaf4#%Q2T_<&#Sh5_&~3M)PX;*FJ`kC&u8m;#&iEf3{XDS z1+VFoO~ovXpL}mAC?9Kb*%$x3jtv^6^PJ;P$z*@}y_qqfo@%vdTlTZJoC|mgs--eC zp^ZNg188#SN`a$hUQJib&$P-&k_TS%%SrxSKmVsW$tlCRb7yBvgmeRc4x(J2@*Jg= zDLvf2`j$Q5q~ZOG-*5;g*(z}ElA1YRdg1G6A{Qe~@>u&$QR|g^c*%z59pz8nn})y ziHVYR18A+ildGrKF}L9`{`3!Hon(eUt`1UTkAL1c^ta3F0;ojYw6}^f%R+u&Ej;Z8A9E0Ov0lt&c{dYS( zk&T*u{P2K(=vn%{Klc6S&qA-@w&q;9Eu!(jKfZ>P$WY^kbfP9shlwe;dFIX6kPG+_ z!xxb<|1i+#mE;Cv1vtX7jjMj%+Szq-p)`x3dvrlZxUj5N&8f=`Cz~p%QBxqFmi+G@ z%t(TTPZ)qRzQxzZ)ii^3xkU-?|IF4gYaveNq~Rs&$O)lvAeYp9d-mU~i-kjY>5=g* z9JpYR_#0%t9+oxI7}(2RtX zG0z|lCl~gG%6RhLh&yvo`}G9s@AB7?<)6Ia%K4`s+>j3VLgHfF6afjqot$y1Y_ld2 zoB7mQ?El=}V`brkiNpOS)e1hGu3P_b13&yn##{XPWe^#|u`vJodFQB+>CvhZzMamr zr-u@hPF+2RY85;D`MLOi*GxmiIou2Pdjl@tH^ilIc;2zd3cI`RAH>vyO;k6V7G?eN z%jL^*v+X2zn*L#nLk=KmPsO+)xu>}p_Gws;k|ES|t-rmMKQ?6+soyFfHLoz+-cwP; z`HvNJAE<`#?nOaDPjc;s%RI61H5Urc(+7(a2p_(TRnDinAasW|kn4Ug;q$+H&P)c7 zRBGPJTXP|L1Mg5wQpgh>ny0|U+yk6(+WGp($s68ZM*L;OUn}wdYb)^)TfyNF1asn< z34Po~U%kP3{vwe^`32VX^oE_qj>WO1#3)<>iuV2yquB5;D(1Dk^8ENHb1gi^WHo>6 z=j{a|QkFaIMft*Lq_^r$sR>c&+f8s4M{Q>3gI4W1b}V|kH$AM9oTgAw{_oZojLnxIsN zp=9G!f&eN?!<&=_C$(jAqYisdkLnrF(!?I1QqGM4o%|3komIRdDWYuPP?Ha9W!5^)1b!yK*8C;1By$Pna;&t z(A45MUL>DkwKu$d3@T7P!I|5MOd$D_3LvL}jM8s_jORoOu7iV+tVz?(44%g;^>?)D znhxg829%`W>aAX?n-%lQo|{KYG^q5Ck}^ah%MH!ef+W9PWi0PEdkVWZ6G&+oF`%5p z!6cq%U+KNUB2dB_UP6aJ%csG$UJ*sHTa<5hmBrS`M_HEb4t!S%5!ahs=*s_g#8s_1 zow%T56uA!bJau-<(N%Hcj*l(tQN)Lg&FNN93Mu0cY&Q#HOVvI}SNteg3pb6{)QYtK z=$pP?j7$u5AC#OSig$Q%GF#ZPX4~X_I=zU5^UGZzywQhR?IjEz3AQIm-vlnU2MCd)oSJPqzW(-6Stey~w=FpdbRMa?06jd?KJ9CQKXWX6zlxG!B;{P?7%(g)~wI1OGU396L06$QoD=)8L%3^jL_ zUz15pYjd$%I_mits_PwLokzWGo0k+vmOoc`vow%~EzQZPfIeXSY{Ftt+Wz&|88-N>;kEnKN_AHvnul{=6=P!-%Dy}ck9EwbGxV23GFm%GE zHhR+i%${0_ET5XphKxjb^*yX@qj)p3zW87af+AErE_Yc`8?yyj^v%N8>cP7k2FHt`cPzaoWUtsPFCVDKewZ3FE}@NV6NdQlqJ)w_wVSLES6yGc<7 zwUoY-m=xv+jC@{#cXHID(w6V-^Qn(-tt@`AUT`#dyUKOGdMS|? z^rD`74SbLHz2|o@kfZBf94h_5e=^u%E9G7QGpq7j5ogG$_)PNM=Vs%B97X7ZiMA(6 zlU27l19L9m(%|Bho$ie`^E1|?yNb=gaA;J8=*6a~*Up(kRA0!9m~3$PELq zECD#0;{#tZdBdK-lbmusqidd3Mz^AE!Z36f;+>T1-Bz5d*5xODR%sb&5u&d!+XA3AFemiC7?%4oPCK4;#(UU!UmbW1W%ul`ocU4Ojj%J-d_Vbiw_O8N>f zLJzZzV1LA=Yb)$kLdYPQdmI0U0SxfPx4$whIvY(^JH88II1a6`=n|lTq3i}v`6>@6 z+e@LPH8uM&G$^-O0RSF)Rlx@3-K(U|a4#}n8AF-HL(APvIU!g@TYP@Y5!Vzef_^8r z_*xHi)8<{;#Yzgu&Mbdaqoz2pfWr?93^F-Vt$g}&y=EY+Q1TArBM_}x&&Okv<=d>H1P;#E`#vHPKkPkDIYuIQRV0qICvZ>a zs@4`it1IUqb(YY~i(q=f#DN7HkTe2p7OrPEs(hD|2(Dv-r8u3Gcn+qg`_0h5&%0xYE7*BsJq;(hgUEuoE)K~& zE&(uLluox=5#u}?>GAfFAiZiMcHu~($RPR(m@V{2f_eyiJB1VY_F(EV*A6FVi%Ptl zM3v_^CAf=3x6z8ACm1{b@ten~vqkDniPqAJx5Pg}jU#s+Ce)}pB~t`UN!4eFA1x%U za%Uu-To&>?qm zcm2l2SNpoS>~>h!98cxlIL74@U1R1k5wU2Gab!+j(&vd2lg$ucoZ0ux6xxnE~X}inGL${1VKv%EHKY2PkkLWr7WlD2b{sa<=^G6;1e{+~Q4t?^3&`uK(1i}N z$4RV>*rJNTO}`|+flYa@#_WI97CJWZU91nd3>)GaqmhWgYWsH;1 z^*45x&g$S!joVS?{a=|eE~uB}C@_EGY!*J{9^#&Y8$gfc6mqA4qv2bnl5c&p-vTnc z9gY%&QXu<|h}!|cP^a`fB8RlQcZ1yVVs9^QVjW#n```v3yjI1L^1Q1_e<67j(qgRB8bq#x;Pz>=IppBO(GuG$OWrminl+?YPJH`hslmPH1&m!_zHEu_lwcr zkFs9vny^LbRL&JC{Cnt$%G{5Y^B6WBNNc6=B-gY?2jck&N?AoA$%Qy6w zdV<@wr-bWiDk+3^|A+8%D{tw-#1LZAc)~M-ZsBdnX@I1KnNueD%I2Xn|XVvZK}w|Kn$|V6Pi67JIPdnW5N&x2t(^?CDGzs ztS#>ICw2?ElbwG$1HQ$Xlv{|cPL|1%Odn1)O6 z7>!y_zx-BHz49ZVC*NAXw<{$j=n=T{n$ybsfn+;rKi(X~W#hC2hAazhy*& zt)DRQ2Cj3YKD%}Du=sO=W5rb)Iq<#DqV;%nqs{F zlF)L`;yegzt<0F@`t83a*BFS6Ikb*u$(?y*?#Z8tU=NcBl&Md!V!_TuY_rIF9+q## zINxl`rDt0hSaY(|v9p#K?+Cb!Qq;<9v{E{(O5#P^A1>qcgM@I0YCMNxX0FC+b}Ov@ z@i+V|@_7q=57kuZ=+C!2wpRIeKjbXlXE-tLw;RXkR}qs^75{rm_57xFu=LO9YY(FD z%4gaFIndAy+=)&Ni>^OegxN}r0}jas%@|egxk_>yK&3eW=g(U~M!zmm`trDb8ZM>s z@-mH;t1{tcP{S0mGm5`kT<>%{+~P68xa-Kr#jo>qJYe^3XTxwAgUjgwE?4CSJB(G; zJ#tU-*uT(he17vENa`kRwEdJUe+T$%+mmq&C}(LYeXC`cpwURZclQj9fOZU84lEw% z6fXi!tIF5Md8IEq7GU&T-iXqSJw&~{83QlLGO)PomB&)pcu%i-?3a!#Y_}vuP#SqX z1UziIF=MnUp9}jpRa-~1I=l)_U{-3X|8<%AF*3E&EPRROHR6GyCa{m@qg2{4m>;d0 zYBrcj-0F{Y{@I>!-{PeRy+s|T8-J8Nw;&)!ikAT!{Z~le=SP;GLa$>8)FA(d8ltuS zcm!tW{G%|<2g=c}__w2u9HVNu7c`5mN!cc@zs-n_uh}RShL$4Rs$#`EILSBUW;s_Q z7s)GQ@;@J8u&I~+*guM?{?c;5HY*Q+SXD|H5(}DDB;Z+gQRc)#p2h619ZuW7zZ(t_ zSBSEl?%qwBK;7728F{Pm!B!VzApz4?3J^sBBKg-Jl>qX)u+FpXiqY(E+%?=Me_j6_ zvv<(}dzW~btg)b7$N*Chg=dUhNZk}(g!R~7i7x<7?GzBNf3{*G*nx>61Dkn3hkysm z(>wc|kf*xksA|eNPc0_aAS*8E{L)3$>Zr!(@!s2x|IiU7sDf0G;GTU41End|AzELR z3?WRjNby$w`uVl~)mDMn6jiuByMDAqTjW2A{n%v54J5|Bmv6?^@cGj&lT0pLzO(f* zRwYJ>dE?eTuKL4NhJe)Y2cf7p;zz&`Vw(Nojf_w1| zrWM-}V7l~WfvQwju#h01RWLAz^51~nkyUC*sBFt&Rvx!_m}nWEoOlE-B!3)jk)B4G z@5P(Dqxh^2Btn1gfQ#v8L&Lpq*0R73TIKxOR?&}sJaLBXYL#=JO3s+}2V~ z*A2SrqDFF|=Nq^~Es>SR$lKX&BM+3v_K}`0_z+3Ck*-qJ3MYP^y6)h{Xjr7V!4$rO z?N}ZoLgfl0?5)OAuKVDj=Il;yl(83d*sdSMvPkDgv#bm|0!jio@slW~23&7^!mdEv z)~rUCn!@MO(~U#P+`Ow>L~hKF6#ezNzBgCzT-su}l|)(FL_RalH#rKablS<5pDGBS zLhee~T-FLdhh2AR3tG0zoB~?nuTqgYGtZ__Bs4~hBIXI=@3A;oajsge5wMKp1~I<6 zv{zgBEw+KNv73niZ_iMV!h_2ma|_5D0>GH?oj2uU1enBF-XC!@{c{doP@`sP9c}H6 z066cn6CCr0$+WVq9B-33v4SsQ6jcY>4MwDfkn7U3+4!1aE2+|m(zI}EQzbmcnIvGs zv&ZehVBf_7{@U`BXGml8_hsBWU_$LPvV3fOuWt%08Jri04!6jz{?#i#PZ!pxnH*p- zR*73t*7@%pqGZI9_owGutN^YwsAAzm!Y;5bm9XnC;-ps7VeyP^RG`E;MHZc1Hu6Is zKQlOskYyXW7hG0bdVu*}Y41W~vvTlA7cog)3pqo}&;-=(_uVP>4@`&*6(*;MCgPHV#;mwW}l9)!dzn0>YIP8WnlmF}wl8%T(HN=SE`M zvD3|}7NXxub0n}PA@HB4JG3vvH70X>r8ovU|udw_pYy2Obb03XWV~Sus zlE?&bving?9jLaubY9c3uXXL9Z?2wq`~sK!tPpcE5|@CooxcarPlAY|l?2DMFXUBO zp@eYjViQ&q&LzM^X;}7?-$#6Yn`gV718YwZah5sWkCJek(7ZrI`+g%d;DBkQ=wjlo zj$2Nh#L<;eKt;j3LEvB;w|>vs9WY4T{c6Mv{CkitGzah4q!TeEopC_D+3+pd1dZ!k zPQ{z=xt9{W>rV1Ao(>(`j~09?8j8_tr5Z?Y_a4?ScBPy^0HBnZfPZ#8(&04wEw;7i z$*Dfv549niaf(KD$Vr}Dk8>FHIGUud78)OWiXWk_6!Mw78}g8IY@D6d%vpX_;Hcv2 zLV7&ZVA}PprV>p2e)ZX|q?TDa-&_H=+hQ!C-)o!YmeNZ%oD^(*Jy{74&9HNxnqkax zuL-t}=!HI|`2ei$qaODDrgT0_^;3HwHv2Pvo*HeQszYyExrx8geCkQuoxaeAAA>sR zv-MKM>CRjT?>G0lf&N`N0;lnJ>I=R1iIg|4t`a$MO6(-S`fZEWKfRBT;eao-K+Ybd zCmnS)6X8QH2M=C$;*L%)e7>vmJZ0;G5jFwfr8MAFlo8&Xc^RW%NcD`!;hN<1dBUnd zmM9B~-t%;C!4`iD&|JUn#|pu3v?O_`nSTBclph|Q!pDo7t|JC=c6HlcRN@{EsZumd z`UP^u#$!e1J4X`8$c&^@1=ALyiSQ^TA+(3e^);Ew`7p7?%B3Xx^?0{qGLePk^E9wA z#*BU^|6R=wiU)4Nm&WqEybb^09|=hkSZ|Q!d%5*oj+80hlQ8xuvh&L-`Gb!jX4luG9Vv%ZW&9bsw~s^TzX@1ZYihYVI;uKV?%}> zv$!G~dnGV@3uAM;64ew-uyrk8@3$7dGDY>@k9x-PtR&h>Z4AdTEg$m`>siCnltCfo zL^UXMSuGWMTxI|>#Joom zrC+S3~_MTf5!X&wD;9fS#C|=f+$FcL5Oq<5+YK9 zfG8p@A|S1TqI7qP2-4jt0(#Tkpdy_jodP!@-AI3PTSv}wo_L@4Ti;sWTF1Xw$aP(N z&z?Oqd*(O4?V`T4G~o=YCWX)4dRD6NOP+g~1o5a%>lbe2^~KFtQSoOhO57=V8c5|x zsY%P&{ej-s`IyCBh?B!zTi>7yjki^OFHh*1Zm022*QOET=~lbGa*_;V3u!LsxiMtF z(!4dF>a=sM7h2o{Nr|45nY2KR|86} zW+b&^P3EGYEMHHBD|zoFz_~es%qEsAd^RvYq)Z|dW>wE-qXOiWY1NylxpFBKaDYxg z?gHwiH%a|G!gMvqFZ2$j7~7FJ>yM^nf1-Y#|3v*X$2b%a zv8vqJm#i3?(A-_V4EQ7bjSAeYO5~kOJJFdhDF(}4!fItI79j|`Hu|k7Hziv4<+MI{ zoXmMWsUyv#W>KIwcQqql_fjEW92aXt@*5>GN(473c^M7arNS2@3znRTf}WUY3- zD4f#D6?H&cl}Ha=K`L8v#O_BAGQVLg4t+(lrdjf4cT{?A_<#Y3(QCcU|Azqp(8heZ zjxX++3-KRoCyu(MVz0$nkNPbav(FEn85wbRq+L)3k_iA%<;R()?&hh^4L?s5YfLX3 z=qZ%U$}GNT^nA}FlZ@VMVnD%&7y0?{j1#nCk`AB!2_UjkGyRC$3$^{R7<7H=l$Tdw zb=RgQ2&RCcs>x}xPq;;IS6Ct{J%8zrV2A(sc3VlU29aicxNPF7{)HmGfWfY z$DchkS-wC{gMY87rCMaz>&~LnFvY(TBsrH?Mu@#_Dd*DJ30AN(_W)A4K-5;Uj7aC>0uXn?!)$v znjmyAk+qZ#MRVCkr4FHF4zozxI2N~z0?ZYCM&j9K=lOzKVx+8-BVx#SS)8%%q`U0n zd8wYhhlQ<;^0Gqe!$C=HA;wtjQFU(~-{sR31?ZJ1M?yG-TPg~dkwY?>s~z?--ToL0 zbM3as>Cb&RBqp(3JwLGqWs@ij+*fY=+EG;Kj(#ktHFT)y+)BJrKdW46*e~=dMlFjcvAlIpGjMaoNz;gTAdqT(8~m#L6>@nEiNM z3IX1=uVI=dN%oV`E7RLNds8<|t1P;c*YON6VevgEe_)g*<2wHpM(Ikp1IV}IH~D(4 z{CCU$7l@v&W#~7rn^1K`n3Tai6>{7I#^z%d!#N$Vl52%YH}2jH2203iYQB z&ET6?KiMuLr>&N?Cu1!ACE2#(CRzklfjm7B+epz%uZix}uvF0ok%j(^vNc9kn=Z3o@Tp$=*bTD$`fb>2 z+8jS&a{nmtH>^=c#Xe=r4KKGi*m)O9my54h!lfIIm}GR7cGiv#`v5wit)(!#qx-`N zTcco08lAsjODHcjzu$b=bn)x4wJxi2y?B(B5uejJzsV>fQ=JO|?ZB&>IMvT+bZOM0 zQJ9hTX5P4@e&k?P*xRxPtN04edX0~eszy7@X_Lv40f6vBblINvoxCDuuv`IgpR)JA zK>0L=`9;$K(7n{LseEJWrTt|8u#~_aTxiG^_Z8W-JP~EowXwet64t0AIJl#i{(Gkj^DM-iK8`;9lTBGEQG_`#888-$wl@wJC^6ysOA}zY- zcVjQNe~#*V*xI&)+1SC9a-Dy*3_*BJZx=!awPC9tuIJ1QvF>3$LO+3@p}1(rIWAhHuE0A>uigQF>|8*Hv83`uOBu zV4M<+{Np09^g2h_u&R;9nX^!XW4=9rjW4iuwf&R?Ud8Nss1U z^R6-J*4TE)>1X3_tlRKQ6Z$mW)o&O^*r?R8;ddrGbYMA^RJ+cf+GkIHQQ&0-UzUFF zMbt%Hq)rwJ{X&KV1v#BWs&$R*l`NI>^=~TocPV`eV~lsSBQ7n~L7PV1n>@15_8|;G zp%%@cUQ=k@4u z@s;Vh=CHRK$M(F0?ND)(;ICjc41!hEnbW=^1x_9kh4rQG0*&Vja)f>aF%j!1#B;q# zQYQx#)yQ!?Fh5bJ2w=TTl#CDm{P>*Yak1r1Z-dFTz6)U6pV+LQ zuP;;@+vyx)oY7cjNxhRt=?0D`R0*P_!2i_;LRe~6VuRF|^Wd@t4X;1?IsO2DeUXU# z-k1_{<@hnjf+!}s~xSzz;qq`5?sdxrRAQSucxN&3WxjVLS8A^{kaZa za=&@sHH{=3IGdSAvDkgJZ+%Jke)D{kybAj=i*&LBf2)(>LQr0qn&j;#1y6PX@ zt$G{>K4NSfEaMY>`z-H~Ps!o&R5xR#0@POM{C(M7*9#&&G9%yeUKjFeTYpCLB>V-j~|l{5zlh3?UF@BLt`^rn&*S3IC5B8$B`;$ zObW^+Dlqzd4~^-a{M>h2@|AgWP+2=6yL9gOin@H>YDvt$eNhF z{11lu_wfFf-ru_Sm%I4yt?~EP_u7sN?FUy`%3KLYUzx z?QVs_&_6mt5x%FCd-!7H2DFow;b-UKyfPlwe_R%M#_fFO`}zkG0C2yBLeK+k0UkED z1QDPQ>*LAThjKeyX? zyLIpAYtA_9WOvP*>4aq-K6ojQhQ=0-9lsPR0+2px?xA3LXMx6JH#Y78W)7EoHH0m? z3=zyT<+`zJkAK#+n%3Fb3VkPTZU>7A7-L4KcA1OrZ3up-A-*X?!ifLqZG|R!@jYyY z)f!@n4ETkq@G?tXR_+u!dUE*4Q=EhJ8$B3euO#f)H3lq>ko^J705Srm2o8J;2OqXaH#;bfcol zUu}gx&@j_gz>Ht8>G%OMqeVIA_aT0AYsS(9Bm%gIaryxH(8HgX7J+<;x5$^(N6n0d zDptl_VC7O43(kQjjqk#fRqyWWO2bE`aS9qvQl&JQAkUeg87|jIKYnw=SuGVR$~TVG z<-W$u{8Qfuf;iq-p`3#@m0kwkuzpg-S0he=M(~U3TWScJs@LYYfzZ? zUd0WB^4hmZxKkLVW%$S`I)la-GN%{cL<$rqGpp$`O2)pyBR?(nvxDf{$2)-D>o#dg z7K2Oxb((e5T~@w_B7998MfIeQ1`N_7)p8RkhQg7`)*L$32fpZR0Kiq$I3h64UDSqM zcO1=>8Iwre?3i&74>55}=?SXtlDA{&5g?|fJLcC9eDuUQap(66;{IgGDj?&|t>@O} zrpzB+M=5ji*fjHUdSVsd_*tw62D0CELO2z+gCbr*3}JlTXZ^<=GTza*NiR<_wn%GC zUPCrx)N>h$d|4H#<4mclLLQVZJE*V?iQ+zyXu!M5@IH-`e7|v3^)-MEr5T;HujBkj z>aXvlYvdDMC*tecfPIAy%@jd*^Z4o0U^Bg+J@>?@%lEv;r+?j9z{_~?PxNYUodG<2 zxDNMY`61Eg^LTtBcbua=Z5I&&piQB_Kq@_|fCW(mEC>p}{M)eW*n~%zO9`*AV&Brv z@>uxNh73+|vACxU`cH-BBITlnL)R7>CLXJN{7}`w4T}crihW_~Y?xHQRjrPJSWdoi zOZ!W$zENe7i79&X$jA)qCo!xx$y}yQqz66;4XPs&Mzc>N`9z<`zcM6 z3(d^Ju@6IDQvPXOl|g|U0+GBSFw@t;Jaw)?@>ib^*1ZJMvu6=kp&6a4{q3_uHjhLi zIDV54%Fe04ccDLZF71%$QJrgvmJg7@eQkU~CWgzGDQS!tZf4k=mPjEMiG>Zp;QqT! z!?Ff6cSB>=R^d5M+`a0Kbd!U$P}Y4*OZJ8`%i+hgnieWKJ`S-nZ{Mr zUe=uW4C#=~SRx7QkUcvYa{a#VpX?24gH#^6#(kVR)?~!^Q+8eYN_K-p;u&JhK;}XVh@OG-YG^U?0sZ*w>QH4?7v>SIFhx5Dhe6p<$fffN{J*ASB-`i ziJW|Bz>bo?o%xeeYAfRhr4*e-;v6}eDXS{8v0&qmcFF|u1JV+PUIc6ZD!aMT%CfJcl1niQd8<1gpGI?! z6Y#D2yUIgt#x+_3Roy@(|4=QL{|5GDqZsbXgQAW4OXI2jC^7Clg|b?sCu2ps0stcGkkc)!v<9QHD*lo&zvi6INCSs*fcK$m^gl=VjYHc#XFm+%2Q!CziHY#F7%qmprV%%sUuX5u&`?DN8S+t z@XzRXdZhmds#TW@byl%ME{bSSoAPDm1oozAbEXjs^PkQ)g~9me{}}(A$BQ8C zmVpoHY7jQsTX=Y@747?XpNgb)RYV2Pv zl`Pbdd}<9$bu^VOw;3oh7+;oT|I$z2(+yKe5L(N7a&e_EuA8)=-!M&WX?KAO=Ax4K zSZBU!t#L-XNlb8H`{LAop!`P^W)6jCi7Ja(({)Q#x$jWvun~~CxpEL;4I z#_8BPm+13HQ+{6?=DX%yyaKxKbBnpDvKZsIpY+ynq-BRszYP?pnxTkjf`K4{Px%birUE5!;) zY_%E3{I!rS;%Kq^`nUO`v+Zq~Z~^vZw(c@4cn*`o{;|U|6rNaKR1(cuxeS)BClR4V^)% z3e%>uk%#zhP=phF-fU&auOev-`?gUp=;-_)auR$$otnwuWR0{k5G(yMp{l)X!q~N4 zwo!f!nSeR3hXTida(o=xU9oEjlHGeVWISCz>|e!5)(6P0{9WJDhNo^gGj_?cQOua+ z*1{s0i;;jfYWj{g0-~n}wm+#JqFV#iJqrSLM8~Zu(eL9xZ7FL^b9}l!o!NrVQrek} z7GDZ`aYrYpZ^V1NQR@)^cf)g0cYtBo1RKU?tfjfuP@j%);}hh&aXg#|sO?h29YUx6{!A-GiGX=ZX`osegNah^Y_XC^a_P1@Tj@8PBil6q=2e6nf zAfdchS_7_F4znJ_VvRBw{TY*2p9{x-d>7**P(Llo&a@pzDt5jnizB1OR6FsxG#f{* z?Fy#m5Bp>PgRm)%adm13^Woy>cW-Dl9IHc?W6WAo7lywklGfvOt4IAbq(ZBUwAw>u zA+18GyZH)pnHhb%q`6DU%Qd7%9i?hE)f*3Ib%Z`;*j?BL^?LoAdVL4h{CN`)!X_B0 zshAFdO@L!&Z|Y5EeARiOhVA;*4Kfv{bjg+P3g(_!w(O~2F`jakcEEPUCuih&B8nz1 znq!xxc$%@hBdV}}J=e7QVc7cIsaZWxv^Lq9`k>U|`XI0)SJ{fEoFe(R^;uQ&ZgTfO zh`KT3)4i|8-S=2(%%YAIP*+>uQCHt2$<~5m>8XEoU>l5J=j~~qZ?!$ngO9k&SFGU1 zT&qt=Rjb0)Q$YXLN=tQl&%0lxsD0;~J)A^Ns2b8{mgT$_Z1_dhL_2Tw_Udw}`(*cq z?A|no=`4p>5(*&(=J11_XsXH*V%-QuRT7hHXfQvUYq4)c;qO{2`?|2=7t4~6YQ4}& zQ|8zsKV6gRb?Il{L1HC<5kw)Lf4b2w(n^7u2>WyL&I8QyL0WJi(h_IRiy(rK z{U^!NjjDf5vgB3KC756c!0o#4z->6T{u!F|(_aec`IAAP^Z-I4bZi}bW%;R{?jLk# z-gDp0OtfG3@S-H&(kuCy#I)bpnG!W@d&f=ZZ8)dr^wO60krj3m4Z5l6L?~d9=<~-& zq0+(Q%d@$youv5WPFYehrV@GAT2oax^8SHeDS;c{JO7sk3)uHLCW_f}Uc{L!6#77X zxbsy^sbO)X>MZTB`3fs4if`;Us#G`PN2C-P+sg*HIDN};dG&SChiNcOYGy6>Dneuu zEok9M2Hbf*GN{I#V`5Ul-@uQm4r`s-LxX18&0JlNSC2WkMtlD3W-wePl1nsJxbCAS zsWfIE`i*Tc6!L;tRBv;S*1I5p%?P2*XP9}2uM5~d4Inp>^Jby8rvepQg}B(4J(|cj z?d#w?OPgN4z70{SWoW-)<;;d{HuEX`oY55zXA z`RmBhG6mXx?&7YY7c0s?`7~Fr{nME6)uCY@2+a9H+5I9Cj2}`7DW%<&+sBCjF5ZrF zx9IUy?TqJ|^fE&69Tr{1OI;1eGPOoWG_#Bo?8jeg!PQ2adiFY!lBGJS`Q|u=LD3S& z`6g4f!vxZSG&Zqwwz1mg6U_`!ahC>!7ytQa_(So(jpl;TLWNFjw8=#8vY(T{X;SFD zu?juyw_TD}I8>7DaA~od;|bs5g44cgvP=|PVQVBSTnp+%GuWk~S0uGJot*=BtV!(e zrCwiOdKx_)d>C4?jwLNj0!`pdTUqzM*rQUeMs}0d}1CORj!_l+s+N+aw`C^vTHA*dm2L#1ZNFNe|rpvgMy{Q}-~hU5j|;gsF_=*fE@b(rgR$xPsYOy{0lr6&K-&-u>F=;KxK22P?85yHwAilnz`55Pg2H>bQ_`6B0U6Ql_QV>qs&Aq z0aYuaW4z~24+C{*2Cz|>$v^_kj9N~J+El^;D@Gk;8-wR#b(*%j4ChS{4ozz`E>6(F z8x_ik;*G~Gtap{kM2|wV!`PvkJ?g4p_EO{OJEiUQE@EAe#8fg|_T1fXYbzmY}=c z@pZR6e0cJq9?3c$+EgO$_(hQ__C%{?gBBlR4wVZ!p$F#W25ARRomYLNzn^4 z!~5=10U<)C>~^yo0HzM}&&*kZN(pbmVO_cXQhlO57EL;E>w_+ zJ#zBVuGg!&c%kPR>CVxntTcN~KzhCRp?xjHUoS(f&Tlh7$MHje!f#WBNk_tyPdpXh znkChpFkU>ba+y!zB#S$t;c_5GGNh{I+wc2bi*+M+LJ^w$CZ(Pez(9K~hQkc!mNtqr z{xYP<$(?)D@4`&ozF2F?K!y;-8Gp=Q2PouE|Nbasi-d*P2iUCzI--1Nry0?C1Asjy(SSDz{G(=c6} z@=MS=PnDwPJq1<{%@XUKk-@lgw~O zir|1HTf6}#kSAf)j+%gE$2ZfbGeDlZqv}jhRN4Ixy>1GDdTiIl3YANN+@wQ zLdEY^2BX+%>Xf>t2viL{gr#;4`2W@jM=S_IMgTuRQ4s(3aC{8N$nx^6*$r?NoC%AM zXpq4?3Fd~>4o1tDd9E^`L~2f*)tB_sSu$5xmlQBVl_cTAl6-oI* z#~ytS0YisEPyic{AG{x&>Yt07>R`-(Q1b*D#wD0~I{_=XS(0%QtxVG?Gz?Gp4KolS z{XR~ZZVpbcMp1*{r7Pd9+oCK8+lk|$C{U0?BEB>4-oYBRaQs=&xy~ZfI|nBGo7}!h z!X+BEgXzCgUt*U!cHTV6la>2IC*TPJ?Aq~9#Y|6C`q#P9u+(k|W$=lp|S z;-^b>z6zG@Xtv%pRyg~`I?VGaXWdI+aoDxkoC@d2ftHxFu}7*l8H7#V6?*h(wpQ+~ z<)~RN(J$vl2!$06JC<@J`~_Wya6GO{b!G*#*+diJmu}hh9s?y3UgLWt?&D{$jCVzk zQ!VcImo6BeZ{S4xoXMJmLHsV)DE7 zyUu!)dtuQO+|8;j(N!ZA^c!U+qQJ3h&59kcn~WVu0JN=A@Pp z8cVZIXMNJSm!gzet>x3H^3( zY5J_7OoWVQH2h3Xe|(_9ofEI`-20l~UJI#|0CT&?4X>s?rN|~-dn`*ue{Vm8Mo!rI zkV`Ul5k5~yX%fR}Ro480Vzsn0jc$Xl4A5hJ+^&{mOiWnSo>E%#e9Gs|IT5`w3LSxU z`uXW`STVYx(b>ePhx>r+AkRoCPzbON0@pSiPAaJ?(9 zXfoDrY|d(ksmtfGRzn;z zXRk1GnU8FY9zRngXTSPUOni_}tHrySdyuYE2F-=@ki0Z~#%(%m z<`Y@_h(nU{aCPh@;kOP&Xs@43VSES`ew^ogMlUaGsr1yM{y>OljtWLeRjJmCR27n@@m1@<~JNZxB3>_h^?$v4$ZwOO0>@Awi=$AJjIB^c^CYy)r zjbml2?Ty)Ah*SyPDV@3vmRcKGl?05Oj8*zK*3|9mABulOr0&q{nKy^brIMenBM9$Y z=NJ?6At&B<>?7gOs!35V2q{7-LInD{=0lMua7(Sr(PN;ySY27v1sGQoK|d1Re1oFPYc#%kLv6Q|)m8wl6esSg;&i;1FZmWu3#!0uLt#z@{Y1Ly+inO58K9Pbbm5u|I z-evLWkQ3IoST$C3LgP?$lBW$NTGUtTyCJ=XJ~)oD0c#w$6kEBvDNewaj?3HKWn!#8 z-fyIuGklzX?v$*Q!^Jn0C9N7J$PxEY{nVEhU0 zm&7B3VZX};&JA-HjVI4&xy9|w6sGLa`7>p0wB6@0X`j`^FKQFPseT7Z&; zAA*h8c>=?1`tscFhE}sC-Z}2l@tTdMzquL$v=l~(f8q(*u770oO# zJki~nZALyR@Xu#C%=WKSa&4`CIsK`>f>U}FaT6JT6t>^lMRT~|SeP7RzJEkc4CBBjl6}GYg6Bd%+z||#PWc2`-+*0^r1C198S*yUN^@Imps9=XH zwCeGyd82Zelfn-9u4$cFRUYgXS9dvyQh6_tYPA=8d9;+ps+@c6kllO$-yB1f0SSBM zvP64w^mV0)dmy>D6J_Jq;tdb2Yg!~;IstcIBTgAp0?a+-A}Aus&BzWqsv6zVPsK9e zt7b5I-&dK3p_~;S?nl*j#T%HXD3Uzaywo6?vSl+II7Zd39KJdu}qdEV?o$e<#O zrk~AVp14b};0Br5o!RW|&rOx`+p=mg_M@+RfL4XGKb6j({+?NEQews_&@G9)Jd|$j z!lKO>>3--%e;`SBNk!P3OQq-QIF>axo0xbMi_7}L#>B*aINNvW#pMxqJiq(bHtM63 zcCFNW_-0mn9;`H9{I7;HzdV0;lvyuCFU{`rJNE~R^LF%!ELm^Smqu4)b$omF_%lN1 zy4e}LD`|7f8@dSm)gCun7}Y4?_nIw)UwpWpd5+b-58;2Nc}w|zQplJWd&SL2py?6i zxEG_T!=)rfCI5CpAV}=t5m0Kvdru@#Mi~o?@}W-u7tTGWuzZM4X!#IvJMzIAHOjR! zNXD0z@rM>QsEnVTiM5~p(qrfzc;=P^*o<(OTbDv z{x(|FibQoDI|Mok5;yojdm?E+3l5M?Dn^Uy=Z_+IHX^I5b-tOQK;=pPO0gnGs9&ivKejdZ{M1OAsmr$|L@eiLzOF^2W)f05@?QC3 zwjPmEZB)=xroB&YHN-{X^g#Ev&`$i}(YJ@!&-W^d&S2kU%bV}x*=mS$STF9;?I-vA zBvrgHG}UiaKWek>3Kw?pvz2{qe>J5ev|U)_I4~=#n{PB0E=-TkJsI-&2|pt!Jm0Z5 zjG*vZpzx$K{SAp;>53Y=6KbVr^n7!bC2&CF4HA=anJP9|jfJ&gS#xQPtv(OO90u$Yz???o9wc?MbUY9jT#o*Fb4T@3w}RJl(n!C+dM zNXAiv^@YKjMXQ<+Wp{d}4h`rLX*I8LtEzy}@4GHOjy^bo3IfJwH01n9I3MNFlW=?c z2}bVLKSOz<^Ly4dtMh}Yye$yZQ@E3xnA?Nn7VmX5h*RTDn>&r`h!uRwU+o-yVwvK% zaQu0czQ#_6!u>Lr!`(hyaEp-O*Uzp+V;QAmMoxCQyv;vM4O1*X)i2eAJ<}QDc@Y&5 z%0fVB)0NI0)T5hEXtcNHnLQB|m7NHC#+RZ6aeC@FD#k*nZhP9Q^INKs@hm@k5Jw-j zb5*~fzUHIVn~HiXmBHS18tX2e+`uhChft4xqhb!*^;u>81<10Y)w@BwJj1lOMQZe+ zVw?O-X2sTUMAf6+txw=PkC{{7GaEYM`7#n?wh|0|1nd}zaTX^2(=PkkTI?M9XTr^>mHM5!D3CZf%}jv+E0efg*raK>&~LSZEsJA~uh`FXz0JtmX)FBog1jP_&^3&~N2WMX8tH+L4@ z8(&RBkW5cid`senO*uxz=0ueBTz>ahCVSZl9M-&BEd809t1}yk;ZA#-GhA*aSaj%= zPCMdhWz)JJ*PX5e##IVjqPJaBi9YTUc?1N-a0%pup^NgBFJvci(XvBa5)o_=TCjN^ zeRB)6%x;?Z!gs@4Vbmo=4}l7Jq*>j9I9HN$x=}iX6EzJ&_|A* zSVb#FogUtab%!0SX?aX6DwcHzExtnF^DzXDXA+{yS0Mbp!sR0hC0BH_G1NIhfTlVc z{*Vd|nW$&~`~wr(i&TOa=zsA)7+Bs|CMt*Ens4#zSMPj#&P4!&E@Sx?5943G>HE8d zC4@(rMPL5)$G-gu6+V1Yprcv*pZ|?N1VR!4=l H_w@e&&uRUj literal 0 HcmV?d00001 From 814aeadcbb77f6b19eca81599c12d81ee113499c Mon Sep 17 00:00:00 2001 From: iABn0rma1 Date: Fri, 31 May 2024 02:32:02 +0530 Subject: [PATCH 11/37] added formulas --- contrib/machine-learning/IntroToCNNs.md | 55 +++++++++---------------- 1 file changed, 20 insertions(+), 35 deletions(-) diff --git a/contrib/machine-learning/IntroToCNNs.md b/contrib/machine-learning/IntroToCNNs.md index f832980..fb9027a 100644 --- a/contrib/machine-learning/IntroToCNNs.md +++ b/contrib/machine-learning/IntroToCNNs.md @@ -26,25 +26,15 @@ ## Introduction - Convolutional Neural Networks (CNNs) are a specialized type of artificial neural network designed primarily for processing structured grid data like images. CNNs are particularly powerful for tasks involving image recognition, classification, and computer vision. They have revolutionized these fields, outperforming traditional neural networks by leveraging their unique architecture to capture spatial hierarchies in images. ### Why CNNs are Superior to Traditional Neural Networks - -1. **Localized Receptive Fields**: - - CNNs use convolutional layers that apply filters to local regions of the input image. This localized connectivity ensures that the network learns spatial hierarchies and patterns, such as edges and textures, which are essential for image recognition tasks. - -2. **Parameter Sharing**: - - In CNNs, the same filter (set of weights) is used across different parts of the input, significantly reducing the number of parameters compared to fully connected layers in traditional neural networks. This not only lowers the computational cost but also mitigates the risk of overfitting. - -3. **Translation Invariance**: - - Due to the shared weights and pooling operations, CNNs are inherently invariant to translations of the input image. This means that they can recognize objects even when they appear in different locations within the image. - -4. **Hierarchical Feature Learning**: - - CNNs automatically learn a hierarchy of features from low-level features like edges to high-level features like shapes and objects. Traditional neural networks, on the other hand, require manual feature extraction which is less effective and more time-consuming. +1. **Localized Receptive Fields**: CNNs use convolutional layers that apply filters to local regions of the input image. This localized connectivity ensures that the network learns spatial hierarchies and patterns, such as edges and textures, which are essential for image recognition tasks. +2. **Parameter Sharing**: In CNNs, the same filter (set of weights) is used across different parts of the input, significantly reducing the number of parameters compared to fully connected layers in traditional neural networks. This not only lowers the computational cost but also mitigates the risk of overfitting. +3. **Translation Invariance**: Due to the shared weights and pooling operations, CNNs are inherently invariant to translations of the input image. This means that they can recognize objects even when they appear in different locations within the image. +4. **Hierarchical Feature Learning**: CNNs automatically learn a hierarchy of features from low-level features like edges to high-level features like shapes and objects. Traditional neural networks, on the other hand, require manual feature extraction which is less effective and more time-consuming. ### Use Cases of CNNs - - **Image Classification**: Identifying objects within an image (e.g., classifying a picture as containing a cat or a dog). - **Object Detection**: Detecting and locating objects within an image (e.g., finding faces in a photo). - **Image Segmentation**: Partitioning an image into segments or regions (e.g., dividing an image into different objects and background). @@ -55,7 +45,6 @@ Convolutional Neural Networks (CNNs) are a specialized type of artificial neural ### Let's start by understanding the basic architecture of CNNs. ## CNN Architecture - Convolution layers, pooling layers, and fully connected layers are just a few of the many building blocks that CNNs use to automatically and adaptively learn spatial hierarchies of information through backpropagation. ### Convolutional Layer @@ -67,31 +56,22 @@ The dimensions of the input image, including the number of channels (e.g., 3 for

-- The input matrix represents a simplified binary image of handwritten digits, -where '1' denotes the presence of ink and '0' represents the absence of ink. - +- The input matrix is a binary image of handwritten digits, +where '1' marks the pixels containing the digit (ink/grayscale area) and '0' marks the background pixels (empty space). - The first matrix shows the represnetation of 1 and 0, which can be depicted as a vertical line and a closed loop. - - The second matrix represents 9, combining the loop and line. -
- #### Strides The step size with which the filter moves across the input image.
-- This visualization will help you understand how the filter (kernel) moves acroos the input matrix with stride values of 3 and 2. - +- This visualization will help you understand how the filter (kernel) moves acroos the input matrix with stride values of (3,3) and (2,2). - A stride of 1 means the filter moves one step at a time, ensuring it covers the entire input matrix. - - However, with larger strides (like 3 or 2 in this example), the filter may not cover all elements, potentially missing some information. - - While this might seem like a drawback, higher strides are often used to reduce computational cost and decrease the output size, which can be beneficial in speeding up the training process and preventing overfitting. -
- #### Padding Determines whether the output size is the same as the input size ('same') or reduced ('valid').
@@ -99,10 +79,8 @@ Determines whether the output size is the same as the input size ('same') or red
- `Same` padding is preferred in earlier layers to preserve spatial and edge information, as it can help the network learn more detailed features. - - Choose `valid` padding when focusing on the central input region or requiring specific output dimensions. - -
+- Padding value can be determined by $ ( f - 1 ) \over 2 $, where f isfilter size #### Filters Small matrices that slide over the input data to extract features. @@ -111,9 +89,7 @@ Small matrices that slide over the input data to extract features.
- The first filter aims to detect closed loops within the input image, being highly relevant for recognizing digits with circular or oval shapes, such as '0', '6', '8', or '9'. - - The next filter helps in detecting vertical lines, crucial for identifying digits like '1', '4', '7', and parts of other digits that contain vertical strokes. - - The last filter shows how to detect diagonal lines in the input image, useful for identifying the slashes present in digits like '1', '7', or parts of '4' and '9'.
@@ -125,12 +101,21 @@ A set of feature maps that represent the presence of different features in the i
- With no padding and a stride of 1, the 3x3 filter moves one step at a time across the 7x5 input matrix. The filter can only move within the original boundaries of the input, resulting in a smaller 5x3 output matrix. This configuration is useful when you want to reduce the spatial dimensions of the feature map while preserving the exact spatial relationships between features. - - By adding zero padding to the input matrix, it is expanded to 9x7, allowing the 3x3 filter to "fit" fully on the edges and corners. With a stride of 1, the filter still moves one step at a time, but now the output matrix is the same size (7x5) as the original input. Same padding is often preferred in early layers of a CNN to preserve spatial information and avoid rapid feature map shrinkage. - - Without padding, the 3x3 filter operates within the original input matrix boundaries, but now it moves two steps at a time (stride 2). This significantly reduces the output matrix size to 3x2. Larger strides are employed to decrease computational cost and the output size, which can be beneficial in speeding up the training process and preventing overfitting. However, they might miss some finer details due to the larger jumps. +- The output dimension of a CNN model is given by, $$ n_{out} = { n_{in} + (2 \cdot p) - k \over s } $$ +where,
+ nin = number of input features
+ p = padding
+ k = kernel size
+ s = stride -
+- Also, the number of trainable parameters for each layer is given by, $ (n_c \cdot [k \cdot k] \cdot f) + f $ +
where,
+ nc = number of input channels
+ k x k = kernel size
+ f = number of filters
+ an additional f is added for bias ### Pooling Layer Pooling layers reduce the dimensionality of each feature map while retaining the most critical information. The most common form of pooling is max pooling. From 21612e4b81468a127feea94ee419d22efcb1133c Mon Sep 17 00:00:00 2001 From: Vrisha Shah <74671946+Vrisha213@users.noreply.github.com> Date: Fri, 31 May 2024 07:58:39 +0530 Subject: [PATCH 12/37] Update index.md --- contrib/machine-learning/index.md | 1 + 1 file changed, 1 insertion(+) diff --git a/contrib/machine-learning/index.md b/contrib/machine-learning/index.md index 46100df..e3a8f0b 100644 --- a/contrib/machine-learning/index.md +++ b/contrib/machine-learning/index.md @@ -10,3 +10,4 @@ - [PyTorch.md](pytorch.md) - [Types of optimizers](Types_of_optimizers.md) - [Logistic Regression](logistic-regression.md) +- [Grid Search](grid-search.md) From a953122b2681a430e0805de2031f5c20af2ee8cc Mon Sep 17 00:00:00 2001 From: Vrisha Shah <74671946+Vrisha213@users.noreply.github.com> Date: Fri, 31 May 2024 07:59:40 +0530 Subject: [PATCH 13/37] Create grid-search.md --- contrib/machine-learning/grid-search.md | 68 +++++++++++++++++++++++++ 1 file changed, 68 insertions(+) create mode 100644 contrib/machine-learning/grid-search.md diff --git a/contrib/machine-learning/grid-search.md b/contrib/machine-learning/grid-search.md new file mode 100644 index 0000000..3bf53ff --- /dev/null +++ b/contrib/machine-learning/grid-search.md @@ -0,0 +1,68 @@ +# Grid Search + +Grid Search is a hyperparameter tuning technique in Machine Learning that helps to find the best combination of hyperparameters for a given model. It works by defining a grid of hyperparameters and then training the model with all the possible combinations of hyperparameters to find the best performing set. +The Grid Search Method considers some hyperparameter combinations and selects the one returning a lower error score. This method is specifically useful when there are only some hyperparameters in order to optimize. However, it is outperformed by other weighted-random search methods when the Machine Learning model grows in complexity. + +## Implementation + +Before applying Grid Searching on any algorithm, Data is used to divided into training and validation set, a validation set is used to validate the models. A model with all possible combinations of hyperparameters is tested on the validation set to choose the best combination. +Grid Searching can be applied to any hyperparameters algorithm whose performance can be improved by tuning hyperparameter. For example, we can apply grid searching on K-Nearest Neighbors by validating its performance on a set of values of K in it. Same thing we can do with Logistic Regression by using a set of values of learning rate to find the best learning rate at which Logistic Regression achieves the best accurac +Let us consider that the model accepts the below three parameters in the form of input: +1. Number of hidden layers [2, 4] +2. Number of neurons in every layer [5, 10] +3. Number of epochs [10, 50] + +If we want to try out two options for every parameter input (as specified in square brackets above), it estimates different combinations. For instance, one possible combination can be [2, 5, 10]. Finding such combinations manually would be a headache. +Now, suppose that we had ten different parameters as input, and we would like to try out five possible values for each and every parameter. It would need manual input from the programmer's end every time we like to alter the value of a parameter, re-execute the code, and keep a record of the outputs for every combination of the parameters. +Grid Search automates that process, as it accepts the possible value for every parameter and executes the code in order to try out each and every possible combination outputs the result for the combinations and outputs the combination having the best accuracy. +Higher values of C tell the model, the training data resembles real world information, place a greater weight on the training data. While lower values of C do the opposite. + +## Explaination of the Code + +The code provided performs hyperparameter tuning for a Logistic Regression model using a manual grid search approach. It evaluates the model's performance for different values of the regularization strength hyperparameter C on the Iris dataset. +1. datasets from sklearn is imported to load the Iris dataset. +2. LogisticRegression from sklearn.linear_model is imported to create and fit the logistic regression model. +3. The Iris dataset is loaded, with X containing the features and y containing the target labels. +4. A LogisticRegression model is instantiated with max_iter=10000 to ensure convergence during the fitting process, as the default maximum iterations (100) might not be sufficient. +5. A list of different values for the regularization strength C is defined. The hyperparameter C controls the regularization strength, with smaller values specifying stronger regularization. +6. An empty list scores is initialized to store the model's performance scores for different values of C. +7. A for loop iterates over each value in the C list: +8. logit.set_params(C=choice) sets the C parameter of the logistic regression model to the current value in the loop. +9. logit.fit(X, y) fits the logistic regression model to the entire Iris dataset (this is typically done on training data in a real scenario, not the entire dataset). +10. logit.score(X, y) calculates the accuracy of the fitted model on the dataset and appends this score to the scores list. +11. After the loop, the scores list is printed, showing the accuracy for each value of C. + +## Python Code + +from sklearn import datasets + +from sklearn.linear_model import LogisticRegression + +iris = datasets.load_iris() + +X = iris['data'] + +y = iris['target'] + +logit = LogisticRegression(max_iter = 10000) + +C = [0.25, 0.5, 0.75, 1, 1.25, 1.5, 1.75, 2] + +scores = [] + +for choice in C: + + logit.set_params(C=choice) + + logit.fit(X, y) + + scores.append(logit.score(X, y)) + +print(scores) + +## Results + +[0.9666666666666667, 0.9666666666666667, 0.9733333333333334, 0.9733333333333334, 0.98, 0.98, 0.9866666666666667, 0.9866666666666667] + +We can see that the lower values of C performed worse than the base parameter of 1. However, as we increased the value of C to 1.75 the model experienced increased accuracy. +It seems that increasing C beyond this amount does not help increase model accuracy. From 50d2b654e03d777ef20c20a210732a9b1a44c54e Mon Sep 17 00:00:00 2001 From: Vrisha Shah <74671946+Vrisha213@users.noreply.github.com> Date: Fri, 31 May 2024 08:05:09 +0530 Subject: [PATCH 14/37] Update index.md --- contrib/plotting-visualization/index.md | 1 + 1 file changed, 1 insertion(+) diff --git a/contrib/plotting-visualization/index.md b/contrib/plotting-visualization/index.md index 32261d6..a8416a8 100644 --- a/contrib/plotting-visualization/index.md +++ b/contrib/plotting-visualization/index.md @@ -3,3 +3,4 @@ - [Installing Matplotlib](matplotlib-installation.md) - [Bar Plots in Matplotlib](matplotlib-bar-plots.md) - [Pie Charts in Matplotlib](matplotlib-pie-charts.md) +- [Box Plots in Matplotlib](matplotlib-box-plots.md) From c54f6128e9cdee0834d265c7b6ca7a42a105c540 Mon Sep 17 00:00:00 2001 From: Vrisha Shah <74671946+Vrisha213@users.noreply.github.com> Date: Fri, 31 May 2024 08:06:46 +0530 Subject: [PATCH 15/37] Create matplotlib-box-plots.md --- .../matplotlib-box-plots.md | 104 ++++++++++++++++++ 1 file changed, 104 insertions(+) create mode 100644 contrib/plotting-visualization/matplotlib-box-plots.md diff --git a/contrib/plotting-visualization/matplotlib-box-plots.md b/contrib/plotting-visualization/matplotlib-box-plots.md new file mode 100644 index 0000000..323b5a4 --- /dev/null +++ b/contrib/plotting-visualization/matplotlib-box-plots.md @@ -0,0 +1,104 @@ +# Box Plot + +A box plot represents the distribution of a dataset in a graph. It displays the summary statistics of a dataset, including the minimum, first quartile (Q1), median (Q2), third quartile (Q3), and maximum. The box represents the interquartile range (IQR) between the first and third quartiles, while whiskers extend from the box to the minimum and maximum values. Outliers, if present, may be displayed as individual points beyond the whiskers. + +For example - Imagine you have the exam scores of students from three classes. A box plot is a way to show how these scores are spread out. + +## Key Ranges in Data Distribution + +The data can be distributed between five key ranges, which are as follows - +1. Minimum: Q1-1.5*IQR +2. 1st quartile (Q1): 25th percentile +3. Median: 50th percentile +4. 3rd quartile(Q3): 75th percentile +5. Maximum: Q3+1.5*IQR + +## Purpose of Box Plots + +We can create the box plot of the data to determine the following- +1. The number of outliers in a dataset +2. Is the data skewed or not (skewness is a measure of asymmetry of the distribution) +3. The range of the data + +## Creating Box Plots using Matplotlib + +By using inbuilt funtion boxplot() of pyplot module of matplotlib - + +Syntax - matplotlib.pyplot.boxplot(data,notch=none,vert=none,patch_artist,widths=none) + +1. data: The data should be an array or sequence of arrays which will be plotted. +2. notch: This parameter accepts only Boolean values, either true or false. +3. vert: This attribute accepts a Boolean value. If it is set to true, then the graph will be vertical. Otherwise, it will be horizontal. +4. position: It accepts the array of integers which defines the position of the box. +5. widths: It accepts the array of integers which defines the width of the box. +6. patch_artist: this parameter accepts Boolean values, either true or false, and this is an optional parameter. +7. labels: This accepts the strings which define the labels for each data point +8. meanline: It accepts a boolean value, and it is optional. +9. order: It sets the order of the boxplot. +10. bootstrap: It accepts the integer value, which specifies the range of the notched boxplot. + +## Implementation of Box Plot in Python + +### Import libraries +import matplotlib.pyplot as plt +import numpy as np + +### Creating dataset +np.random.seed(10) +data = np.random.normal(100, 20, 200) +fig = plt.figure(figsize =(10, 7)) + +### Creating plot +plt.boxplot(data) + +### show plot +plt.show() + +### Implementation of Multiple Box Plot in Python +import matplotlib.pyplot as plt +import numpy as np +np.random.seed(10) +dataSet1 = np.random.normal(100, 10, 220) +dataSet2 = np.random.normal(80, 20, 200) +dataSet3 = np.random.normal(60, 35, 220) +dataSet4 = np.random.normal(50, 40, 200) +dataSet = [dataSet1, dataSet2, dataSet3, dataSet4] +figure = plt.figure(figsize =(10, 7)) +ax = figure.add_axes([0, 0, 1, 1]) +bp = ax.boxplot(dataSet) +plt.show() + +### Implementation of Box Plot with Outliers (visual representation of the sales distribution for each product, and the outliers highlight months with exceptionally high or low sales) +import matplotlib.pyplot as plt +import numpy as np + +### Data for monthly sales +product_A_sales = [100, 110, 95, 105, 115, 90, 120, 130, 80, 125, 150, 200] +product_B_sales = [90, 105, 100, 98, 102, 105, 110, 95, 112, 88, 115, 250] +product_C_sales = [80, 85, 90, 78, 82, 85, 88, 92, 75, 85, 200, 95] + +### Introducing outliers +product_A_sales.extend([300, 80]) +product_B_sales.extend([50, 300]) +product_C_sales.extend([70, 250]) + +### Creating a box plot with outliers +plt.boxplot([product_A_sales, product_B_sales, product_C_sales], sym='o') +plt.title('Monthly Sales Performance by Product with Outliers') +plt.xlabel('Products') +plt.ylabel('Sales') +plt.show() + +### Implementation of Grouped Box Plot (to compare the exam scores of students from three different classes (A, B, and C)) +import matplotlib.pyplot as plt +import numpy as np +class_A_scores = [75, 80, 85, 90, 95] +class_B_scores = [70, 75, 80, 85, 90] +class_C_scores = [65, 70, 75, 80, 85] + +### Creating a grouped box plot +plt.boxplot([class_A_scores, class_B_scores, class_C_scores], labels=['Class A', 'Class B', 'Class C']) +plt.title('Exam Scores by Class') +plt.xlabel('Classes') +plt.ylabel('Scores') +plt.show() From c3f715e98321ff5afe0fb50d5dd80e7ec7366a59 Mon Sep 17 00:00:00 2001 From: Vrisha Shah <74671946+Vrisha213@users.noreply.github.com> Date: Fri, 31 May 2024 08:09:21 +0530 Subject: [PATCH 16/37] Update matplotlib-box-plots.md --- .../matplotlib-box-plots.md | 39 ++++++++++++++++++- 1 file changed, 38 insertions(+), 1 deletion(-) diff --git a/contrib/plotting-visualization/matplotlib-box-plots.md b/contrib/plotting-visualization/matplotlib-box-plots.md index 323b5a4..4d76243 100644 --- a/contrib/plotting-visualization/matplotlib-box-plots.md +++ b/contrib/plotting-visualization/matplotlib-box-plots.md @@ -40,60 +40,97 @@ Syntax - matplotlib.pyplot.boxplot(data,notch=none,vert=none,patch_artist,widths ## Implementation of Box Plot in Python ### Import libraries + import matplotlib.pyplot as plt + import numpy as np ### Creating dataset + np.random.seed(10) + data = np.random.normal(100, 20, 200) + fig = plt.figure(figsize =(10, 7)) ### Creating plot + plt.boxplot(data) ### show plot + plt.show() ### Implementation of Multiple Box Plot in Python + import matplotlib.pyplot as plt + import numpy as np + np.random.seed(10) + dataSet1 = np.random.normal(100, 10, 220) -dataSet2 = np.random.normal(80, 20, 200) + +dataSet2 = np.random.normal(80, 20, 200) + dataSet3 = np.random.normal(60, 35, 220) + dataSet4 = np.random.normal(50, 40, 200) + dataSet = [dataSet1, dataSet2, dataSet3, dataSet4] + figure = plt.figure(figsize =(10, 7)) + ax = figure.add_axes([0, 0, 1, 1]) + bp = ax.boxplot(dataSet) + plt.show() ### Implementation of Box Plot with Outliers (visual representation of the sales distribution for each product, and the outliers highlight months with exceptionally high or low sales) + import matplotlib.pyplot as plt + import numpy as np ### Data for monthly sales + product_A_sales = [100, 110, 95, 105, 115, 90, 120, 130, 80, 125, 150, 200] + product_B_sales = [90, 105, 100, 98, 102, 105, 110, 95, 112, 88, 115, 250] + product_C_sales = [80, 85, 90, 78, 82, 85, 88, 92, 75, 85, 200, 95] ### Introducing outliers + product_A_sales.extend([300, 80]) + product_B_sales.extend([50, 300]) + product_C_sales.extend([70, 250]) ### Creating a box plot with outliers + plt.boxplot([product_A_sales, product_B_sales, product_C_sales], sym='o') + plt.title('Monthly Sales Performance by Product with Outliers') + plt.xlabel('Products') + plt.ylabel('Sales') + plt.show() ### Implementation of Grouped Box Plot (to compare the exam scores of students from three different classes (A, B, and C)) + import matplotlib.pyplot as plt + import numpy as np + class_A_scores = [75, 80, 85, 90, 95] + class_B_scores = [70, 75, 80, 85, 90] + class_C_scores = [65, 70, 75, 80, 85] ### Creating a grouped box plot From d552e04ded85206b731474e0e272f1701641c13a Mon Sep 17 00:00:00 2001 From: Vrisha Shah <74671946+Vrisha213@users.noreply.github.com> Date: Fri, 31 May 2024 08:09:57 +0530 Subject: [PATCH 17/37] Update matplotlib-box-plots.md --- contrib/plotting-visualization/matplotlib-box-plots.md | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/contrib/plotting-visualization/matplotlib-box-plots.md b/contrib/plotting-visualization/matplotlib-box-plots.md index 4d76243..1780cab 100644 --- a/contrib/plotting-visualization/matplotlib-box-plots.md +++ b/contrib/plotting-visualization/matplotlib-box-plots.md @@ -134,8 +134,13 @@ class_B_scores = [70, 75, 80, 85, 90] class_C_scores = [65, 70, 75, 80, 85] ### Creating a grouped box plot + plt.boxplot([class_A_scores, class_B_scores, class_C_scores], labels=['Class A', 'Class B', 'Class C']) + plt.title('Exam Scores by Class') + plt.xlabel('Classes') + plt.ylabel('Scores') + plt.show() From 208ff1ea06192dc51ebc45f1e32b29a185d4b561 Mon Sep 17 00:00:00 2001 From: Vrisha Shah <74671946+Vrisha213@users.noreply.github.com> Date: Fri, 31 May 2024 08:13:42 +0530 Subject: [PATCH 18/37] Delete contrib/plotting-visualization/matplotlib-box-plots.md --- .../matplotlib-box-plots.md | 146 ------------------ 1 file changed, 146 deletions(-) delete mode 100644 contrib/plotting-visualization/matplotlib-box-plots.md diff --git a/contrib/plotting-visualization/matplotlib-box-plots.md b/contrib/plotting-visualization/matplotlib-box-plots.md deleted file mode 100644 index 1780cab..0000000 --- a/contrib/plotting-visualization/matplotlib-box-plots.md +++ /dev/null @@ -1,146 +0,0 @@ -# Box Plot - -A box plot represents the distribution of a dataset in a graph. It displays the summary statistics of a dataset, including the minimum, first quartile (Q1), median (Q2), third quartile (Q3), and maximum. The box represents the interquartile range (IQR) between the first and third quartiles, while whiskers extend from the box to the minimum and maximum values. Outliers, if present, may be displayed as individual points beyond the whiskers. - -For example - Imagine you have the exam scores of students from three classes. A box plot is a way to show how these scores are spread out. - -## Key Ranges in Data Distribution - -The data can be distributed between five key ranges, which are as follows - -1. Minimum: Q1-1.5*IQR -2. 1st quartile (Q1): 25th percentile -3. Median: 50th percentile -4. 3rd quartile(Q3): 75th percentile -5. Maximum: Q3+1.5*IQR - -## Purpose of Box Plots - -We can create the box plot of the data to determine the following- -1. The number of outliers in a dataset -2. Is the data skewed or not (skewness is a measure of asymmetry of the distribution) -3. The range of the data - -## Creating Box Plots using Matplotlib - -By using inbuilt funtion boxplot() of pyplot module of matplotlib - - -Syntax - matplotlib.pyplot.boxplot(data,notch=none,vert=none,patch_artist,widths=none) - -1. data: The data should be an array or sequence of arrays which will be plotted. -2. notch: This parameter accepts only Boolean values, either true or false. -3. vert: This attribute accepts a Boolean value. If it is set to true, then the graph will be vertical. Otherwise, it will be horizontal. -4. position: It accepts the array of integers which defines the position of the box. -5. widths: It accepts the array of integers which defines the width of the box. -6. patch_artist: this parameter accepts Boolean values, either true or false, and this is an optional parameter. -7. labels: This accepts the strings which define the labels for each data point -8. meanline: It accepts a boolean value, and it is optional. -9. order: It sets the order of the boxplot. -10. bootstrap: It accepts the integer value, which specifies the range of the notched boxplot. - -## Implementation of Box Plot in Python - -### Import libraries - -import matplotlib.pyplot as plt - -import numpy as np - -### Creating dataset - -np.random.seed(10) - -data = np.random.normal(100, 20, 200) - -fig = plt.figure(figsize =(10, 7)) - -### Creating plot - -plt.boxplot(data) - -### show plot - -plt.show() - -### Implementation of Multiple Box Plot in Python - -import matplotlib.pyplot as plt - -import numpy as np - -np.random.seed(10) - -dataSet1 = np.random.normal(100, 10, 220) - -dataSet2 = np.random.normal(80, 20, 200) - -dataSet3 = np.random.normal(60, 35, 220) - -dataSet4 = np.random.normal(50, 40, 200) - -dataSet = [dataSet1, dataSet2, dataSet3, dataSet4] - -figure = plt.figure(figsize =(10, 7)) - -ax = figure.add_axes([0, 0, 1, 1]) - -bp = ax.boxplot(dataSet) - -plt.show() - -### Implementation of Box Plot with Outliers (visual representation of the sales distribution for each product, and the outliers highlight months with exceptionally high or low sales) - -import matplotlib.pyplot as plt - -import numpy as np - -### Data for monthly sales - -product_A_sales = [100, 110, 95, 105, 115, 90, 120, 130, 80, 125, 150, 200] - -product_B_sales = [90, 105, 100, 98, 102, 105, 110, 95, 112, 88, 115, 250] - -product_C_sales = [80, 85, 90, 78, 82, 85, 88, 92, 75, 85, 200, 95] - -### Introducing outliers - -product_A_sales.extend([300, 80]) - -product_B_sales.extend([50, 300]) - -product_C_sales.extend([70, 250]) - -### Creating a box plot with outliers - -plt.boxplot([product_A_sales, product_B_sales, product_C_sales], sym='o') - -plt.title('Monthly Sales Performance by Product with Outliers') - -plt.xlabel('Products') - -plt.ylabel('Sales') - -plt.show() - -### Implementation of Grouped Box Plot (to compare the exam scores of students from three different classes (A, B, and C)) - -import matplotlib.pyplot as plt - -import numpy as np - -class_A_scores = [75, 80, 85, 90, 95] - -class_B_scores = [70, 75, 80, 85, 90] - -class_C_scores = [65, 70, 75, 80, 85] - -### Creating a grouped box plot - -plt.boxplot([class_A_scores, class_B_scores, class_C_scores], labels=['Class A', 'Class B', 'Class C']) - -plt.title('Exam Scores by Class') - -plt.xlabel('Classes') - -plt.ylabel('Scores') - -plt.show() From dd24cd3a2e82ef40f55ead219868947afdfc12c6 Mon Sep 17 00:00:00 2001 From: Vrisha Shah <74671946+Vrisha213@users.noreply.github.com> Date: Fri, 31 May 2024 08:14:00 +0530 Subject: [PATCH 19/37] Update index.md --- contrib/plotting-visualization/index.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/contrib/plotting-visualization/index.md b/contrib/plotting-visualization/index.md index a8416a8..92ec106 100644 --- a/contrib/plotting-visualization/index.md +++ b/contrib/plotting-visualization/index.md @@ -3,4 +3,4 @@ - [Installing Matplotlib](matplotlib-installation.md) - [Bar Plots in Matplotlib](matplotlib-bar-plots.md) - [Pie Charts in Matplotlib](matplotlib-pie-charts.md) -- [Box Plots in Matplotlib](matplotlib-box-plots.md) + From bdbe355c485de10dfea943108434b756e624a889 Mon Sep 17 00:00:00 2001 From: Soubeer Koley Date: Fri, 31 May 2024 08:23:22 +0530 Subject: [PATCH 20/37] updated index --- contrib/machine-learning/index.md | 1 + 1 file changed, 1 insertion(+) diff --git a/contrib/machine-learning/index.md b/contrib/machine-learning/index.md index 94ca1e2..073bca9 100644 --- a/contrib/machine-learning/index.md +++ b/contrib/machine-learning/index.md @@ -9,3 +9,4 @@ - [TensorFlow.md](tensorFlow.md) - [PyTorch.md](pytorch.md) - [Types of optimizers](Types_of_optimizers.md) +- [Random Forest](Random_Forest.md) \ No newline at end of file From cd74a77397850249e8910638f41149a86476f37b Mon Sep 17 00:00:00 2001 From: Ritesh Date: Fri, 31 May 2024 10:17:29 +0530 Subject: [PATCH 21/37] Create clustering.md --- contrib/machine-learning/clustering.md | 115 +++++++++++++++++++++++++ 1 file changed, 115 insertions(+) create mode 100644 contrib/machine-learning/clustering.md diff --git a/contrib/machine-learning/clustering.md b/contrib/machine-learning/clustering.md new file mode 100644 index 0000000..7d5577c --- /dev/null +++ b/contrib/machine-learning/clustering.md @@ -0,0 +1,115 @@ +# Clustering + +Clustering is an unsupervised machine learning technique that groups a set of objects in such a way that objects in the same group (called a cluster) are more similar to each other than to those in other groups (clusters). This README provides an overview of clustering, including its fundamental concepts, types, algorithms, and how to implement it using Python. + +## Table of Contents + +1. [Introduction](#introduction) +2. [Concepts](#concepts) +3. [Types of Clustering](#types-of-clustering) +4. [Clustering Algorithms](#clustering-algorithms) +5. [Implementation](#implementation) + - [Using Scikit-learn](#using-scikit-learn) + - [Code Example](#code-example) +6. [Evaluation Metrics](#evaluation-metrics) +7. [Conclusion](#conclusion) +8. [References](#references) + +## Introduction + +Clustering is a technique used to find inherent groupings within data without pre-labeled targets. It is widely used in exploratory data analysis, pattern recognition, image analysis, information retrieval, and bioinformatics. + +## Concepts + +### Centroid + +A centroid is the center of a cluster. In the k-means clustering algorithm, for example, each cluster is represented by its centroid, which is the mean of all the data points in the cluster. + +### Distance Measure + +Distance measures are used to quantify the similarity or dissimilarity between data points. Common distance measures include Euclidean distance, Manhattan distance, and cosine similarity. + +### Inertia + +Inertia is a metric used to assess the quality of the clusters formed. It is the sum of squared distances of samples to their nearest cluster center. + +## Types of Clustering + +1. **Hard Clustering**: Each data point either belongs to a cluster completely or not at all. +2. **Soft Clustering (Fuzzy Clustering)**: Each data point can belong to multiple clusters with varying degrees of membership. + +## Clustering Algorithms + +### K-Means Clustering + +K-Means is a popular clustering algorithm that partitions the data into k clusters, where each data point belongs to the cluster with the nearest mean. The algorithm follows these steps: +1. Initialize k centroids randomly. +2. Assign each data point to the nearest centroid. +3. Recalculate the centroids as the mean of all data points assigned to each cluster. +4. Repeat steps 2 and 3 until convergence. + +### Hierarchical Clustering + +Hierarchical clustering builds a tree of clusters. There are two types: +- **Agglomerative (bottom-up)**: Starts with each data point as a separate cluster and merges the closest pairs of clusters iteratively. +- **Divisive (top-down)**: Starts with all data points in one cluster and splits the cluster iteratively into smaller clusters. + +### DBSCAN (Density-Based Spatial Clustering of Applications with Noise) + +DBSCAN groups together points that are close to each other based on a distance measurement and a minimum number of points. It can find arbitrarily shaped clusters and is robust to noise. + +## Implementation + +### Using Scikit-learn + +Scikit-learn is a popular machine learning library in Python that provides tools for clustering. + +### Code Example + +```python +import numpy as np +import pandas as pd +from sklearn.cluster import KMeans +from sklearn.preprocessing import StandardScaler +from sklearn.metrics import silhouette_score + +# Load dataset +data = pd.read_csv('path/to/your/dataset.csv') + +# Preprocess the data +scaler = StandardScaler() +data_scaled = scaler.fit_transform(data) + +# Initialize and fit KMeans model +kmeans = KMeans(n_clusters=3, random_state=42) +kmeans.fit(data_scaled) + +# Get cluster labels +labels = kmeans.labels_ + +# Calculate silhouette score +silhouette_avg = silhouette_score(data_scaled, labels) +print("Silhouette Score:", silhouette_avg) + +# Add cluster labels to the original data +data['Cluster'] = labels + +print(data.head()) +``` + +## Evaluation Metrics + +- **Silhouette Score**: Measures how similar a data point is to its own cluster compared to other clusters. +- **Inertia (Within-cluster Sum of Squares)**: Measures the compactness of the clusters. +- **Davies-Bouldin Index**: Measures the average similarity ratio of each cluster with the cluster that is most similar to it. +- **Dunn Index**: Ratio of the minimum inter-cluster distance to the maximum intra-cluster distance. + +## Conclusion + +Clustering is a powerful technique for discovering structure in data. Understanding different clustering algorithms and their evaluation metrics is crucial for selecting the appropriate method for a given problem. + +## References + +- [Scikit-learn Documentation](https://scikit-learn.org/stable/modules/clustering.html) +- [Wikipedia: Cluster Analysis](https://en.wikipedia.org/wiki/Cluster_analysis) +- [Towards Data Science: A Comprehensive Guide to Clustering](https://towardsdatascience.com/a-comprehensive-guide-to-clustering-9789897f8b88) From 90e238ebf593caf7cf99ed1d213823f5219a062f Mon Sep 17 00:00:00 2001 From: Ritesh Date: Fri, 31 May 2024 10:18:24 +0530 Subject: [PATCH 22/37] Update index.md --- contrib/machine-learning/index.md | 1 + 1 file changed, 1 insertion(+) diff --git a/contrib/machine-learning/index.md b/contrib/machine-learning/index.md index 46100df..1e0004d 100644 --- a/contrib/machine-learning/index.md +++ b/contrib/machine-learning/index.md @@ -10,3 +10,4 @@ - [PyTorch.md](pytorch.md) - [Types of optimizers](Types_of_optimizers.md) - [Logistic Regression](logistic-regression.md) +- [Clustering](clustering.md) From d00f146401afba7514f183291d40a62de92866ba Mon Sep 17 00:00:00 2001 From: Yogesh Vishwakarma Date: Fri, 31 May 2024 11:34:39 +0530 Subject: [PATCH 23/37] created Sliding_Window.md and updated index.md --- contrib/ds-algorithms/Sliding_Window.md | 249 ++++++++++++++++++++++++ contrib/ds-algorithms/index.md | 1 + 2 files changed, 250 insertions(+) create mode 100644 contrib/ds-algorithms/Sliding_Window.md diff --git a/contrib/ds-algorithms/Sliding_Window.md b/contrib/ds-algorithms/Sliding_Window.md new file mode 100644 index 0000000..72aa191 --- /dev/null +++ b/contrib/ds-algorithms/Sliding_Window.md @@ -0,0 +1,249 @@ +# Sliding Window Technique + +The sliding window technique is a fundamental approach used to solve problems involving arrays, lists, or sequences. It's particularly useful when you need to calculate something over a subarray or sublist of fixed size that slides over the entire array. + +In easy words, It is the transformation of the nested loops into the single loop +## Concept + +The sliding window technique involves creating a window (a subarray or sublist) that moves or "slides" across the entire array. This window can either be fixed in size or dynamically resized. By maintaining and updating this window as it moves, you can optimize certain computations, reducing time complexity. + +## Types of Sliding Windows + +1. **Fixed Size Window**: The window size remains constant as it slides from the start to the end of the array. +2. **Variable Size Window**: The window size can change based on certain conditions, such as the sum of elements within the window meeting a specified target. + +## Steps to Implement a Sliding Window + +1. **Initialize the Window**: Set the initial position of the window and any required variables (like sum, count, etc.). +2. **Expand the Window**: Add the next element to the window and update the relevant variables. +3. **Shrink the Window**: If needed, remove elements from the start of the window and update the variables. +4. **Slide the Window**: Move the window one position to the right by including the next element and possibly excluding the first element. +5. **Repeat**: Continue expanding, shrinking, and sliding the window until you reach the end of the array. + +## Example Problems + +### 1. Maximum Sum Subarray of Fixed Size K + +Given an array of integers and an integer k, find the maximum sum of a subarray of size k. + +**Steps:** + +1. Initialize the sum of the first k elements. +2. Slide the window from the start of the array to the end, updating the sum by subtracting the element that is left behind and adding the new element. +3. Track the maximum sum encountered. + +**Python Code:** + +```python +def max_sum_subarray(arr, k): + n = len(arr) + if n < k: + return None + + # Compute the sum of the first window + window_sum = sum(arr[:k]) + max_sum = window_sum + + # Slide the window from start to end + for i in range(n - k): + window_sum = window_sum - arr[i] + arr[i + k] + max_sum = max(max_sum, window_sum) + + return max_sum + +# Example usage: +arr = [1, 3, 2, 5, 1, 1, 6, 2, 8, 5] +k = 3 +print(max_sum_subarray(arr, k)) # Output: 16 +``` + +### 2. Longest Substring Without Repeating Characters + +Given a string, find the length of the longest substring without repeating characters. + +**Steps:** + +1. Use two pointers to represent the current window. +2. Use a set to track characters in the current window. +3. Expand the window by moving the right pointer. +4. If a duplicate character is found, shrink the window by moving the left pointer until the duplicate is removed. + +**Python Code:** + +```python +def longest_unique_substring(s): + n = len(s) + char_set = set() + left = 0 + max_length = 0 + + for right in range(n): + while s[right] in char_set: + char_set.remove(s[left]) + left += 1 + char_set.add(s[right]) + max_length = max(max_length, right - left + 1) + + return max_length + +# Example usage: +s = "abcabcbb" +print(longest_unique_substring(s)) # Output: 3 +``` +## 3. Minimum Size Subarray Sum + +Given an array of positive integers and a positive integer `s`, find the minimal length of a contiguous subarray of which the sum is at least `s`. If there isn't one, return 0 instead. + +### Steps: +1. Use two pointers, `left` and `right`, to define the current window. +2. Expand the window by moving `right` and adding `arr[right]` to `current_sum`. +3. If `current_sum` is greater than or equal to `s`, update `min_length` and shrink the window from the left by moving `left` and subtracting `arr[left]` from `current_sum`. +4. Repeat until `right` has traversed the array. + +### Python Code: +```python +def min_subarray_len(s, arr): + n = len(arr) + left = 0 + current_sum = 0 + min_length = float('inf') + + for right in range(n): + current_sum += arr[right] + + while current_sum >= s: + min_length = min(min_length, right - left + 1) + current_sum -= arr[left] + left += 1 + + return min_length if min_length != float('inf') else 0 + +# Example usage: +arr = [2, 3, 1, 2, 4, 3] +s = 7 +print(min_subarray_len(s, arr)) # Output: 2 (subarray [4, 3]) +``` + +## 4. Longest Substring with At Most K Distinct Characters + +Given a string `s` and an integer `k`, find the length of the longest substring that contains at most `k` distinct characters. + +### Steps: +1. Use two pointers, `left` and `right`, to define the current window. +2. Use a dictionary `char_count` to count characters in the window. +3. Expand the window by moving `right` and updating `char_count`. +4. If `char_count` has more than `k` distinct characters, shrink the window from the left by moving `left` and updating `char_count`. +5. Keep track of the maximum length of the window with at most `k` distinct characters. + +### Python Code: +```python +def longest_substring_k_distinct(s, k): + n = len(s) + char_count = {} + left = 0 + max_length = 0 + + for right in range(n): + char_count[s[right]] = char_count.get(s[right], 0) + 1 + + while len(char_count) > k: + char_count[s[left]] -= 1 + if char_count[s[left]] == 0: + del char_count[s[left]] + left += 1 + + max_length = max(max_length, right - left + 1) + + return max_length + +# Example usage: +s = "eceba" +k = 2 +print(longest_substring_k_distinct(s, k)) # Output: 3 (substring "ece") +``` + +## 5. Maximum Number of Vowels in a Substring of Given Length + +Given a string `s` and an integer `k`, return the maximum number of vowel letters in any substring of `s` with length `k`. + +### Steps: +1. Use a sliding window of size `k`. +2. Keep track of the number of vowels in the current window. +3. Expand the window by adding the next character and update the count if it's a vowel. +4. If the window size exceeds `k`, remove the leftmost character and update the count if it's a vowel. +5. Track the maximum number of vowels found in any window of size `k`. + +### Python Code: +```python +def max_vowels(s, k): + vowels = set('aeiou') + max_vowel_count = 0 + current_vowel_count = 0 + + for i in range(len(s)): + if s[i] in vowels: + current_vowel_count += 1 + if i >= k: + if s[i - k] in vowels: + current_vowel_count -= 1 + max_vowel_count = max(max_vowel_count, current_vowel_count) + + return max_vowel_count + +# Example usage: +s = "abciiidef" +k = 3 +print(max_vowels(s, k)) # Output: 3 (substring "iii") +``` + +## 6. Subarray Product Less Than K + +Given an array of positive integers `nums` and an integer `k`, return the number of contiguous subarrays where the product of all the elements in the subarray is less than `k`. + +### Steps: +1. Use two pointers, `left` and `right`, to define the current window. +2. Expand the window by moving `right` and multiplying `product` by `nums[right]`. +3. If `product` is greater than or equal to `k`, shrink the window from the left by moving `left` and dividing `product` by `nums[left]`. +4. For each position of `right`, the number of valid subarray ending at `right` is `right - left + 1`. +5. Sum these counts to get the total number of subarray with product less than `k`. + +### Python Code: +```python +def num_subarray_product_less_than_k(nums, k): + if k <= 1: + return 0 + + product = 1 + left = 0 + count = 0 + + for right in range(len(nums)): + product *= nums[right] + + while product >= k: + product /= nums[left] + left += 1 + + count += right - left + 1 + + return count + +# Example usage: +nums = [10, 5, 2, 6] +k = 100 +print(num_subarray_product_less_than_k(nums, k)) # Output: 8 +``` + +## Advantages + +- **Efficiency**: Reduces the time complexity from O(n^2) to O(n) for many problems. +- **Simplicity**: Provides a straightforward way to manage subarrays/substrings with overlapping elements. + +## Applications + +- Finding the maximum or minimum sum of subarrays of fixed size. +- Detecting unique elements in a sequence. +- Solving problems related to dynamic programming with fixed constraints. +- Efficiently managing and processing streaming data or real-time analytics. + +By using the sliding window technique, you can tackle a wide range of problems in a more efficient manner. diff --git a/contrib/ds-algorithms/index.md b/contrib/ds-algorithms/index.md index 31cff39..d3c3c27 100644 --- a/contrib/ds-algorithms/index.md +++ b/contrib/ds-algorithms/index.md @@ -10,3 +10,4 @@ - [Greedy Algorithms](greedy-algorithms.md) - [Dynamic Programming](dynamic-programming.md) - [Linked list](linked-list.md) +- [Sliding Window Technique](Sliding_Window.md) From ef42e881851eb20f9a9c8c28a9afdcff89c7fb67 Mon Sep 17 00:00:00 2001 From: Ankit Mahato Date: Fri, 31 May 2024 20:49:03 +0530 Subject: [PATCH 24/37] Update index.md --- contrib/plotting-visualization/index.md | 1 - 1 file changed, 1 deletion(-) diff --git a/contrib/plotting-visualization/index.md b/contrib/plotting-visualization/index.md index 92ec106..32261d6 100644 --- a/contrib/plotting-visualization/index.md +++ b/contrib/plotting-visualization/index.md @@ -3,4 +3,3 @@ - [Installing Matplotlib](matplotlib-installation.md) - [Bar Plots in Matplotlib](matplotlib-bar-plots.md) - [Pie Charts in Matplotlib](matplotlib-pie-charts.md) - From f26f8da60ec1814d6165a6cb908aa2a0ea7a1b2c Mon Sep 17 00:00:00 2001 From: Ankit Mahato Date: Fri, 31 May 2024 20:53:21 +0530 Subject: [PATCH 25/37] Update grid-search.md --- contrib/machine-learning/grid-search.md | 41 +++++++++++++------------ 1 file changed, 22 insertions(+), 19 deletions(-) diff --git a/contrib/machine-learning/grid-search.md b/contrib/machine-learning/grid-search.md index 3bf53ff..ae44412 100644 --- a/contrib/machine-learning/grid-search.md +++ b/contrib/machine-learning/grid-search.md @@ -1,20 +1,26 @@ # Grid Search Grid Search is a hyperparameter tuning technique in Machine Learning that helps to find the best combination of hyperparameters for a given model. It works by defining a grid of hyperparameters and then training the model with all the possible combinations of hyperparameters to find the best performing set. + The Grid Search Method considers some hyperparameter combinations and selects the one returning a lower error score. This method is specifically useful when there are only some hyperparameters in order to optimize. However, it is outperformed by other weighted-random search methods when the Machine Learning model grows in complexity. ## Implementation -Before applying Grid Searching on any algorithm, Data is used to divided into training and validation set, a validation set is used to validate the models. A model with all possible combinations of hyperparameters is tested on the validation set to choose the best combination. -Grid Searching can be applied to any hyperparameters algorithm whose performance can be improved by tuning hyperparameter. For example, we can apply grid searching on K-Nearest Neighbors by validating its performance on a set of values of K in it. Same thing we can do with Logistic Regression by using a set of values of learning rate to find the best learning rate at which Logistic Regression achieves the best accurac -Let us consider that the model accepts the below three parameters in the form of input: -1. Number of hidden layers [2, 4] -2. Number of neurons in every layer [5, 10] -3. Number of epochs [10, 50] +Before applying Grid Searching on any algorithm, data is divided into training and validation set, a validation set is used to validate the models. A model with all possible combinations of hyperparameters is tested on the validation set to choose the best combination. + +Grid Searching can be applied to any hyperparameters algorithm whose performance can be improved by tuning hyperparameter. For example, we can apply grid searching on K-Nearest Neighbors by validating its performance on a set of values of K in it. Same thing we can do with Logistic Regression by using a set of values of learning rate to find the best learning rate at which Logistic Regression achieves the best accuracy. + +Let us consider that the model accepts the below three parameters in the form of input: +1. Number of hidden layers `[2, 4]` +2. Number of neurons in every layer `[5, 10]` +3. Number of epochs `[10, 50]` + +If we want to try out two options for every parameter input (as specified in square brackets above), it estimates different combinations. For instance, one possible combination can be `[2, 5, 10]`. Finding such combinations manually would be a headache. -If we want to try out two options for every parameter input (as specified in square brackets above), it estimates different combinations. For instance, one possible combination can be [2, 5, 10]. Finding such combinations manually would be a headache. Now, suppose that we had ten different parameters as input, and we would like to try out five possible values for each and every parameter. It would need manual input from the programmer's end every time we like to alter the value of a parameter, re-execute the code, and keep a record of the outputs for every combination of the parameters. + Grid Search automates that process, as it accepts the possible value for every parameter and executes the code in order to try out each and every possible combination outputs the result for the combinations and outputs the combination having the best accuracy. + Higher values of C tell the model, the training data resembles real world information, place a greater weight on the training data. While lower values of C do the opposite. ## Explaination of the Code @@ -32,16 +38,14 @@ The code provided performs hyperparameter tuning for a Logistic Regression model 10. logit.score(X, y) calculates the accuracy of the fitted model on the dataset and appends this score to the scores list. 11. After the loop, the scores list is printed, showing the accuracy for each value of C. -## Python Code +### Python Code +```python from sklearn import datasets - from sklearn.linear_model import LogisticRegression iris = datasets.load_iris() - X = iris['data'] - y = iris['target'] logit = LogisticRegression(max_iter = 10000) @@ -49,20 +53,19 @@ logit = LogisticRegression(max_iter = 10000) C = [0.25, 0.5, 0.75, 1, 1.25, 1.5, 1.75, 2] scores = [] - for choice in C: - logit.set_params(C=choice) - logit.fit(X, y) - scores.append(logit.score(X, y)) - print(scores) +``` -## Results +#### Results +``` [0.9666666666666667, 0.9666666666666667, 0.9733333333333334, 0.9733333333333334, 0.98, 0.98, 0.9866666666666667, 0.9866666666666667] +``` -We can see that the lower values of C performed worse than the base parameter of 1. However, as we increased the value of C to 1.75 the model experienced increased accuracy. -It seems that increasing C beyond this amount does not help increase model accuracy. +We can see that the lower values of `C` performed worse than the base parameter of `1`. However, as we increased the value of `C` to `1.75` the model experienced increased accuracy. + +It seems that increasing `C` beyond this amount does not help increase model accuracy. From d83b6e1deaa84b2d0d2411c7a5eb8b4284ab5351 Mon Sep 17 00:00:00 2001 From: Ankit Mahato Date: Sat, 1 Jun 2024 20:56:34 +0530 Subject: [PATCH 26/37] Update CONTRIBUTING.md --- CONTRIBUTING.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/CONTRIBUTING.md b/CONTRIBUTING.md index 4a366da..8688009 100644 --- a/CONTRIBUTING.md +++ b/CONTRIBUTING.md @@ -24,8 +24,8 @@ The list of topics for which we are looking for content are provided below along - Web Scrapping - [Link](https://github.com/animator/learn-python/tree/main/contrib/web-scrapping) - API Development - [Link](https://github.com/animator/learn-python/tree/main/contrib/api-development) - Data Structures & Algorithms - [Link](https://github.com/animator/learn-python/tree/main/contrib/ds-algorithms) -- Python Mini Projects - [Link](https://github.com/animator/learn-python/tree/main/contrib/mini-projects) -- Python Question Bank - [Link](https://github.com/animator/learn-python/tree/main/contrib/question-bank) +- Python Mini Projects - [Link](https://github.com/animator/learn-python/tree/main/contrib/mini-projects) **(Not accepting)** +- Python Question Bank - [Link](https://github.com/animator/learn-python/tree/main/contrib/question-bank) **(Not accepting)** You can check out some content ideas below. From d372975f6c7accb7dd3e7a7852ba61d3e24ea724 Mon Sep 17 00:00:00 2001 From: Mohammed Ahmed Majid <109688855+PilotAxis@users.noreply.github.com> Date: Sat, 1 Jun 2024 21:45:43 +0530 Subject: [PATCH 27/37] Added file exception-handling.md --- contrib/advanced-python/exception-handling.md | 192 ++++++++++++++++++ 1 file changed, 192 insertions(+) create mode 100644 contrib/advanced-python/exception-handling.md diff --git a/contrib/advanced-python/exception-handling.md b/contrib/advanced-python/exception-handling.md new file mode 100644 index 0000000..bddcac4 --- /dev/null +++ b/contrib/advanced-python/exception-handling.md @@ -0,0 +1,192 @@ +# Exception Handling in Python + +Exception handling is a way of managing errors that may occur during program execution, through which you can handle exceptions gracefully. Python's exception handling mechanism has been designed to avoid unexpected termination of the program and offer a means to either regain control after an error or display meaningful messages to the user. + +- **Error** - An error is a mistake or an incorrect result produced by a program. It can be a syntax error, a logical error, or a runtime error. Errors are typically fatal, meaning they prevent the program from continuing to execute. +- **Exception** - An exception is an event that occurs during the execution of a program that disrupts the normal flow of instructions. Exceptions are typically unexpected and can be handled by the program to prevent it from crashing or terminating abnormally. It can be runtime, input/output or system exceptions. Exceptions are designed to be handled by the program, allowing it to recover from the error and continue executing. + +## Python Built-in Exceptions + +There are plenty of built-in exceptions in Python that are raised when a corresponding error occur. +We can view all the built-in exceptions using the built-in `local()` function as follows: + +```python +print(dir(locals()['__builtins__'])) +``` + +|**S.No**|**Exception**|**Description**| +|---|---|---| +|1|SyntaxError|A syntax error occurs when the code we write violates the grammatical rules such as misspelled keywords, missing colon, mismatched parentheses etc.| +|2|TypeError|A type error occurs when we try to perform an operation or use a function with objects that are of incompatible data types.| +|3|NameError|A name error occurs when we try to use a variable, function, module or string without quotes that hasn't been defined or isn't used in a valid way.| +|4|IndexError|A index error occurs when we try to access an element in a sequence (like a list, tuple or string) using an index that's outside the valid range of indices for that sequence.| +|5|KeyError|A key error occurs when we try to access a key that doesn't exist in a dictionary. Attempting to retrieve a value using a non-existent key results this error.| +|6|ValueError|A value error occurs when we provide an argument or value that's inappropriate for a specific operation or function such as doing mathematical operations with incompatible types (e.g., dividing a string by an integer.)| +|7|AttributeError|An attribute error occurs when we try to access an attribute (like a variable or method) on an object that doesn't possess that attribute.| +|8|IOError|An IO (Input/Output) error occurs when an operation involving file or device interaction fails. It signifies that there's an issue during communication between your program and the external system.| +|9|ZeroDivisionError|A ZeroDivisionError occurs when we attempt to divide a number by zero. This operation is mathematically undefined, and Python raises this error to prevent nonsensical results.| +|10|ImportError|An import error occurs when we try to use a module or library that Python can't find or import succesfully.| + +## Try and Except Statement - Catching Exception + +The `try-except` statement allows us to anticipate potential errors during program execution and define what actions to take when those errors occur. This prevents the program from crashing unexpectedly and makes it more robust. + +Here's an example to explain this: + +```python +try: + # code that might raise an exception + result = 10 / 0 +except: + print("An error occured!") +``` + +Output + +```markdown +An error occured! +``` + +In this example, the `try` block contains the code that you suspect might raise an exception. Python attempts to execute the code within this block. If an exception occurs, Python jumps to the `except` block and executes the code within it. + +## Specific Exception Handling + +You can specify the type of expection you want to catch using the `except` keyword followed by the exception class name. You can also have multiple `except` blocks to handle different exception types. + +Here's an example: + +```python +try: + # Code that might raise ZeroDivisionError or NameError + result = 10 / 0 + name = undefined_variable +except ZeroDivisionError: + print("Oops! You tried to divide by zero.") +except NameError: + print("There's a variable named 'undefined_variable' that hasn't been defined yet.") +``` + +Output + +```markdown +Oops! You tried to divide by zero. +``` + +If you comment on the line `result = 10 / 0`, then the output will be + +```markdown +There's a variable named 'undefined_variable' that hasn't been defined yet. +``` + +## Important Note + +In this code, the `except` block are specific to each type of expection. If you want to catch both exceptions with a single `except` block, you can use of tuple of exceptions, like this: + +```python +try: + # Code that might raise ZeroDivisionError or NameError + result = 10 / 0 + name = undefined_variable +except (ZeroDivisionError, NameError): + print("An error occured!") +``` + +Output + +```markdown +An error occured! +``` + +## Try with Else Clause + +The `else` clause in a Python `try-except` block provides a way to execute code only when the `try` block succeeds without raising any exceptions. It's like having a section of code that runs exclusively under the condition that no errors occur during the main operation in the `try` block. + +Here's an example to understand this: + +```python +def calculate_average(numbers): + if len(numbers) == 0: # Handle empty list case seperately (optional) + return None + try: + total = sum(numbers) + average = total / len(numbers) + except ZeroDivisionError: + print("Cannot calculate average for a list containing zero.") + else: + print("The average is:", average) + return average #Optionally return the average here + +# Example usage +numbers = [10, 20, 30] +result = calculate_average(numbers) + +if result is not None: # Check if result is available (handles empty list case) + print("Calculation succesfull!") +``` + +Output + +```markdown +The average is: 20.0 +``` + +## Finally Keyword in Python + +The `finally` keyword in Python is used within `try-except` statements to execute a block of code **always**, regardless of whether an exception occurs in the `try` block or not. + +To understand this, let us take an example: + +```python +try: + a = 10 // 0 + print(a) +except ZeroDivisionError: + print("Cannot be divided by zero.") +finally: + print("Program executed!") +``` + +Output + +```markdown +Cannot be divided by zero. +Program executed! +``` + +## Raise Keyword in Python + +In Python, raising an exception allows you to signal that an error condition has occured during your program's execution. The `raise` keyword is used to explicity raise an exception. + +Let us take an example: + +```python +def divide(x, y): + if y == 0: + raise ZeroDivisionError("Can't divide by zero!") # Raise an exception with a message + result = x / y + return result + +try: + division_result = divide(10, 0) + print("Result:", division_result) +except ZeroDivisionError as e: + print("An error occured:", e) # Handle the exception and print the message +``` + +Output + +```markdown +An error occured: Can't divide by zero! +``` + +## Advantages of Exception Handling + +- **Improved Error Handling** - It allows you to gracefully handle unexpected situations that arise during program execution. Instead of crashing abruptly, you can define specific actions to take when exceptions occur, providing a smoother experience. +- **Code Robustness** - Exception Handling helps you to write more resilient programs by anticipating potential issues and providing approriate responses. +- **Enhanced Code Readability** - By seperating error handling logic from the core program flow, your code becomes more readable and easier to understand. The `try-except` blocks clearly indicate where potential errors might occur and how they'll be addressed. + +## Disadvantages of Exception Handling + +- **Hiding Logic Errors** - Relying solely on exception handling might mask underlying logic error in your code. It's essential to write clear and well-tested logic to minimize the need for excessive exception handling. +- **Performance Overhead** - In some cases, using `try-except` blocks can introduce a slight performance overhead compared to code without exception handling. Howerer, this is usually negligible for most applications. +- **Overuse of Exceptions** - Overusing exceptions for common errors or control flow can make code less readable and harder to maintain. It's important to use exceptions judiciously for unexpected situations. From a9e3a4673c62076ba2a764cf6d77d5acda177fc3 Mon Sep 17 00:00:00 2001 From: Mohammed Ahmed Majid Date: Sun, 2 Jun 2024 00:39:56 +0530 Subject: [PATCH 28/37] Updated Files --- contrib/advanced-python/exception-handling.md | 6 +++--- contrib/advanced-python/index.md | 1 + 2 files changed, 4 insertions(+), 3 deletions(-) diff --git a/contrib/advanced-python/exception-handling.md b/contrib/advanced-python/exception-handling.md index bddcac4..3e0c672 100644 --- a/contrib/advanced-python/exception-handling.md +++ b/contrib/advanced-python/exception-handling.md @@ -1,6 +1,6 @@ # Exception Handling in Python -Exception handling is a way of managing errors that may occur during program execution, through which you can handle exceptions gracefully. Python's exception handling mechanism has been designed to avoid unexpected termination of the program and offer a means to either regain control after an error or display meaningful messages to the user. +Exception Handling is a way of managing the errors that may occur during a program execution. Python's exception handling mechanism has been designed to avoid the unexpected termination of the program, and offer to either regain control after an error or display a meaningful message to the user. - **Error** - An error is a mistake or an incorrect result produced by a program. It can be a syntax error, a logical error, or a runtime error. Errors are typically fatal, meaning they prevent the program from continuing to execute. - **Exception** - An exception is an event that occurs during the execution of a program that disrupts the normal flow of instructions. Exceptions are typically unexpected and can be handled by the program to prevent it from crashing or terminating abnormally. It can be runtime, input/output or system exceptions. Exceptions are designed to be handled by the program, allowing it to recover from the error and continue executing. @@ -35,7 +35,7 @@ Here's an example to explain this: ```python try: - # code that might raise an exception + # Code that might raise an exception result = 10 / 0 except: print("An error occured!") @@ -72,7 +72,7 @@ Output Oops! You tried to divide by zero. ``` -If you comment on the line `result = 10 / 0`, then the output will be +If you comment on the line `result = 10 / 0`, then the output will be: ```markdown There's a variable named 'undefined_variable' that hasn't been defined yet. diff --git a/contrib/advanced-python/index.md b/contrib/advanced-python/index.md index b95e4b9..febcbbe 100644 --- a/contrib/advanced-python/index.md +++ b/contrib/advanced-python/index.md @@ -7,3 +7,4 @@ - [Regular Expressions in Python](regular_expressions.md) - [JSON module](json-module.md) - [Map Function](map-function.md) +- [Exception Handling in Python](exception-handling.md) From 3a8ac54d7cd94a7ceda1552d8be823129230ff78 Mon Sep 17 00:00:00 2001 From: Ankit Mahato Date: Sun, 2 Jun 2024 03:05:00 +0530 Subject: [PATCH 29/37] Update clustering.md --- contrib/machine-learning/clustering.md | 19 ------------------- 1 file changed, 19 deletions(-) diff --git a/contrib/machine-learning/clustering.md b/contrib/machine-learning/clustering.md index 7d5577c..bc02d37 100644 --- a/contrib/machine-learning/clustering.md +++ b/contrib/machine-learning/clustering.md @@ -2,19 +2,6 @@ Clustering is an unsupervised machine learning technique that groups a set of objects in such a way that objects in the same group (called a cluster) are more similar to each other than to those in other groups (clusters). This README provides an overview of clustering, including its fundamental concepts, types, algorithms, and how to implement it using Python. -## Table of Contents - -1. [Introduction](#introduction) -2. [Concepts](#concepts) -3. [Types of Clustering](#types-of-clustering) -4. [Clustering Algorithms](#clustering-algorithms) -5. [Implementation](#implementation) - - [Using Scikit-learn](#using-scikit-learn) - - [Code Example](#code-example) -6. [Evaluation Metrics](#evaluation-metrics) -7. [Conclusion](#conclusion) -8. [References](#references) - ## Introduction Clustering is a technique used to find inherent groupings within data without pre-labeled targets. It is widely used in exploratory data analysis, pattern recognition, image analysis, information retrieval, and bioinformatics. @@ -107,9 +94,3 @@ print(data.head()) ## Conclusion Clustering is a powerful technique for discovering structure in data. Understanding different clustering algorithms and their evaluation metrics is crucial for selecting the appropriate method for a given problem. - -## References - -- [Scikit-learn Documentation](https://scikit-learn.org/stable/modules/clustering.html) -- [Wikipedia: Cluster Analysis](https://en.wikipedia.org/wiki/Cluster_analysis) -- [Towards Data Science: A Comprehensive Guide to Clustering](https://towardsdatascience.com/a-comprehensive-guide-to-clustering-9789897f8b88) From 03a168d202babc9839e00b9b9563dc49f39bed39 Mon Sep 17 00:00:00 2001 From: Ankit Mahato Date: Sun, 2 Jun 2024 03:11:03 +0530 Subject: [PATCH 30/37] Update and rename IntroToCNNs.md to intro-to-cnn.md --- .../{IntroToCNNs.md => intro-to-cnn.md} | 75 +++++-------------- 1 file changed, 17 insertions(+), 58 deletions(-) rename contrib/machine-learning/{IntroToCNNs.md => intro-to-cnn.md} (89%) diff --git a/contrib/machine-learning/IntroToCNNs.md b/contrib/machine-learning/intro-to-cnn.md similarity index 89% rename from contrib/machine-learning/IntroToCNNs.md rename to contrib/machine-learning/intro-to-cnn.md index fb9027a..0221ca1 100644 --- a/contrib/machine-learning/IntroToCNNs.md +++ b/contrib/machine-learning/intro-to-cnn.md @@ -1,30 +1,5 @@ # Understanding Convolutional Neural Networks (CNN) -## Table of Contents -
-Click to expand - -- [Introduction](#introduction) -- [CNN Architecture](#cnn-architecture) - -
- Convolutional Layer - - - [Input Shape](#input-shape) - - [Stride](#strides) - - [Padding](#padding) - - [Filter](#filters) - - [Output](#output) - -
- -- [Pooling Layer](#pooling-layer) -- [Flatten Layer](#flatten-layer) -- [Dropout Layer](#dropout-layer) - -- [Implementation](#implementation) - -
- ## Introduction Convolutional Neural Networks (CNNs) are a specialized type of artificial neural network designed primarily for processing structured grid data like images. CNNs are particularly powerful for tasks involving image recognition, classification, and computer vision. They have revolutionized these fields, outperforming traditional neural networks by leveraging their unique architecture to capture spatial hierarchies in images. @@ -52,9 +27,7 @@ The convolutional layer is the core building block of a CNN. The layer's paramet #### Input Shape The dimensions of the input image, including the number of channels (e.g., 3 for RGB images & 1 for Grayscale images). -
- -
+![image](assets/cnn-input_shape.png) - The input matrix is a binary image of handwritten digits, where '1' marks the pixels containing the digit (ink/grayscale area) and '0' marks the background pixels (empty space). @@ -63,9 +36,7 @@ where '1' marks the pixels containing the digit (ink/grayscale area) and '0' mar #### Strides The step size with which the filter moves across the input image. -
- -
+![image](assets/cnn-strides.png) - This visualization will help you understand how the filter (kernel) moves acroos the input matrix with stride values of (3,3) and (2,2). - A stride of 1 means the filter moves one step at a time, ensuring it covers the entire input matrix. @@ -74,9 +45,7 @@ The step size with which the filter moves across the input image. #### Padding Determines whether the output size is the same as the input size ('same') or reduced ('valid'). -
- -
+![image](assets/cnn-padding.png) - `Same` padding is preferred in earlier layers to preserve spatial and edge information, as it can help the network learn more detailed features. - Choose `valid` padding when focusing on the central input region or requiring specific output dimensions. @@ -84,37 +53,31 @@ Determines whether the output size is the same as the input size ('same') or red #### Filters Small matrices that slide over the input data to extract features. -
- -
+![image](assets/cnn-filters.png) - The first filter aims to detect closed loops within the input image, being highly relevant for recognizing digits with circular or oval shapes, such as '0', '6', '8', or '9'. - The next filter helps in detecting vertical lines, crucial for identifying digits like '1', '4', '7', and parts of other digits that contain vertical strokes. - The last filter shows how to detect diagonal lines in the input image, useful for identifying the slashes present in digits like '1', '7', or parts of '4' and '9'. -
- #### Output A set of feature maps that represent the presence of different features in the input. -
- -
+![image](assets/cnn-ouputs.png) - With no padding and a stride of 1, the 3x3 filter moves one step at a time across the 7x5 input matrix. The filter can only move within the original boundaries of the input, resulting in a smaller 5x3 output matrix. This configuration is useful when you want to reduce the spatial dimensions of the feature map while preserving the exact spatial relationships between features. - By adding zero padding to the input matrix, it is expanded to 9x7, allowing the 3x3 filter to "fit" fully on the edges and corners. With a stride of 1, the filter still moves one step at a time, but now the output matrix is the same size (7x5) as the original input. Same padding is often preferred in early layers of a CNN to preserve spatial information and avoid rapid feature map shrinkage. - Without padding, the 3x3 filter operates within the original input matrix boundaries, but now it moves two steps at a time (stride 2). This significantly reduces the output matrix size to 3x2. Larger strides are employed to decrease computational cost and the output size, which can be beneficial in speeding up the training process and preventing overfitting. However, they might miss some finer details due to the larger jumps. - The output dimension of a CNN model is given by, $$ n_{out} = { n_{in} + (2 \cdot p) - k \over s } $$ -where,
- nin = number of input features
- p = padding
- k = kernel size
+where, + nin = number of input features + p = padding + k = kernel size s = stride -- Also, the number of trainable parameters for each layer is given by, $ (n_c \cdot [k \cdot k] \cdot f) + f $ -
where,
- nc = number of input channels
- k x k = kernel size
- f = number of filters
+- Also, the number of trainable parameters for each layer is given by, $ (n_c \cdot [k \cdot k] \cdot f) + f $ +where, + nc = number of input channels + k x k = kernel size + f = number of filters an additional f is added for bias ### Pooling Layer @@ -135,18 +98,14 @@ Pooling layers reduce the dimensionality of each feature map while retaining the The flatten layer converts the 2D matrix data to a 1D vector, which can be fed into a fully connected (dense) layer. - **Input Shape:** The 2D feature maps from the previous layer. - **Output:** A 1D vector that represents the same data in a flattened format. -
- -
+![image](assets/cnn-flattened.png) ### Dropout Layer Dropout is a regularization technique to prevent overfitting in neural networks by randomly setting a fraction of input units to zero at each update during training time. - **Input Shape:** The data from the previous layer. - **Dropout Rate:** The fraction of units to drop (e.g., 0.5 for 50% dropout). - **Output:** The same shape as the input, with some units set to zero. -
- -
+![image](assets/cnn-dropout.png) - The updated 0 values represents the dropped units. @@ -263,4 +222,4 @@ dropout_output = cnn_model.dropout(flattened_output, dropout_rate=0.3) print("\nDropout Output:\n", dropout_output) ``` -Feel free to play around with the parameters! \ No newline at end of file +Feel free to play around with the parameters! From e4dd80f1ce9e18679d00ba97bbe99a337a15b73e Mon Sep 17 00:00:00 2001 From: Ankit Mahato Date: Sun, 2 Jun 2024 03:11:26 +0530 Subject: [PATCH 31/37] Update index.md --- contrib/machine-learning/index.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/contrib/machine-learning/index.md b/contrib/machine-learning/index.md index ea8b5cb..988a27f 100644 --- a/contrib/machine-learning/index.md +++ b/contrib/machine-learning/index.md @@ -9,4 +9,4 @@ - [TensorFlow.md](tensorFlow.md) - [PyTorch.md](pytorch.md) - [Types of optimizers](Types_of_optimizers.md) -- [Introduction To Convolutional Neural Networks (CNNs)](IntroToCNNs.md) +- [Introduction To Convolutional Neural Networks (CNNs)](intro-to-cnn.md) From 33543fd13c21329a4656f360a96d29c6483e8c71 Mon Sep 17 00:00:00 2001 From: Ankit Mahato Date: Sun, 2 Jun 2024 03:15:14 +0530 Subject: [PATCH 32/37] Delete .vscode/settings.json --- .vscode/settings.json | 2 -- 1 file changed, 2 deletions(-) delete mode 100644 .vscode/settings.json diff --git a/.vscode/settings.json b/.vscode/settings.json deleted file mode 100644 index 7a73a41..0000000 --- a/.vscode/settings.json +++ /dev/null @@ -1,2 +0,0 @@ -{ -} \ No newline at end of file From 3c95166575d6e25c2544a0be7d8d8cae4ab9b45e Mon Sep 17 00:00:00 2001 From: Ankit Mahato Date: Sun, 2 Jun 2024 03:15:39 +0530 Subject: [PATCH 33/37] Rename Random_Forest.md to random-forest.md --- contrib/machine-learning/{Random_Forest.md => random-forest.md} | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) rename contrib/machine-learning/{Random_Forest.md => random-forest.md} (99%) diff --git a/contrib/machine-learning/Random_Forest.md b/contrib/machine-learning/random-forest.md similarity index 99% rename from contrib/machine-learning/Random_Forest.md rename to contrib/machine-learning/random-forest.md index 59c44ef..0abd1ab 100644 --- a/contrib/machine-learning/Random_Forest.md +++ b/contrib/machine-learning/random-forest.md @@ -193,4 +193,4 @@ Scikit-learn Random Forest Documentation Wikipedia: Random Forest Machine Learning Mastery: Introduction to Random Forest Kaggle: Random Forest Guide -Towards Data Science: Understanding Random Forests \ No newline at end of file +Towards Data Science: Understanding Random Forests From 86e7c0d80613c236d05b4cd0895f0ee6df7601cb Mon Sep 17 00:00:00 2001 From: Ankit Mahato Date: Sun, 2 Jun 2024 03:16:18 +0530 Subject: [PATCH 34/37] Update index.md --- contrib/machine-learning/index.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/contrib/machine-learning/index.md b/contrib/machine-learning/index.md index 073bca9..3b8c95b 100644 --- a/contrib/machine-learning/index.md +++ b/contrib/machine-learning/index.md @@ -9,4 +9,4 @@ - [TensorFlow.md](tensorFlow.md) - [PyTorch.md](pytorch.md) - [Types of optimizers](Types_of_optimizers.md) -- [Random Forest](Random_Forest.md) \ No newline at end of file +- [Random Forest](random-forest.md) From 0189c285cc983d2a19e306967bfe4d1f0dda43da Mon Sep 17 00:00:00 2001 From: Ankit Mahato Date: Sun, 2 Jun 2024 03:18:23 +0530 Subject: [PATCH 35/37] Update random-forest.md --- contrib/machine-learning/random-forest.md | 37 ++++------------------- 1 file changed, 6 insertions(+), 31 deletions(-) diff --git a/contrib/machine-learning/random-forest.md b/contrib/machine-learning/random-forest.md index 0abd1ab..feaaa7a 100644 --- a/contrib/machine-learning/random-forest.md +++ b/contrib/machine-learning/random-forest.md @@ -2,31 +2,6 @@ Random Forest is a versatile machine learning algorithm capable of performing both regression and classification tasks. It is an ensemble method that operates by constructing a multitude of decision trees during training and outputting the average prediction of the individual trees (for regression) or the mode of the classes (for classification). - -- [Random Forest](#random-forest) - - [Introduction](#introduction) - - [How Random Forest Works](#how-random-forest-works) - - [1. Bootstrap Sampling:](#1-bootstrap-sampling) - - [2. Decision Trees:](#2-decision-trees) - - [3. Feature Selection:](#3-feature-selection) - - [4. Voting/Averaging:](#4-votingaveraging) - - [Detailed Working Mechanism](#detailed-working-mechanism) - - [Step 3: Aggregation:](#step-3-aggregation) - - [Advantages and Disadvantages](#advantages-and-disadvantages) - - [Advantages](#advantages) - - [Disadvantages](#disadvantages) - - [Hyperparameters](#hyperparameters) - - [Key Hyperparameters](#key-hyperparameters) - - [Tuning Hyperparameters](#tuning-hyperparameters) - - [Code Examples](#code-examples) - - [Classification Example](#classification-example) - - [Feature Importance](#feature-importance) - - [Hyperparameter Tuning](#hyperparameter-tuning) - - [Regression Example](#regression-example) - - [Conclusion](#conclusion) - - [References](#references) - - ## Introduction Random Forest is an ensemble learning method used for classification and regression tasks. It is built from multiple decision trees and combines their outputs to improve the model's accuracy and control over-fitting. @@ -41,9 +16,9 @@ Random Forest is an ensemble learning method used for classification and regress For classification, the mode of the classes predicted by individual trees is taken (majority vote). For regression, the average of the outputs of the individual trees is taken. ### Detailed Working Mechanism -* #### Step 1: Bootstrap Sampling: +#### Step 1: Bootstrap Sampling: Each tree is trained on a random sample of the original data, drawn with replacement (bootstrap sample). This means some data points may appear multiple times in a sample while others may not appear at all. -* #### Step 2: Tree Construction: +#### Step 2: Tree Construction: Each node in the tree is split using the best split among a random subset of the features. This process adds an additional layer of randomness, contributing to the robustness of the model. #### Step 3: Aggregation: For classification tasks, the final prediction is based on the majority vote from all the trees. For regression tasks, the final prediction is the average of all the tree predictions. @@ -73,7 +48,7 @@ Hyperparameter tuning can significantly improve the performance of a Random Fore #### Classification Example Below is a simple example of using Random Forest for a classification task with the Iris dataset. -``` +```python import numpy as np import pandas as pd from sklearn.datasets import load_iris @@ -109,7 +84,7 @@ print("Classification Report:\n", classification_report(y_test, y_pred)) Random Forest provides a way to measure the importance of each feature in making predictions. -``` +```python import matplotlib.pyplot as plt # Get feature importances @@ -132,7 +107,7 @@ plt.show() #### Hyperparameter Tuning Using Grid Search for hyperparameter tuning. -``` +```python from sklearn.model_selection import GridSearchCV # Define the parameter grid @@ -155,7 +130,7 @@ print("Best parameters found: ", grid_search.best_params_) #### Regression Example Below is a simple example of using Random Forest for a regression task with the Boston housing dataset. -``` +```python import numpy as np import pandas as pd from sklearn.datasets import load_boston From 56e7774648e3dbccd68dbdc1d8e61521524b83aa Mon Sep 17 00:00:00 2001 From: Ankit Mahato Date: Sun, 2 Jun 2024 03:20:23 +0530 Subject: [PATCH 36/37] Update index.md --- contrib/ds-algorithms/index.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/contrib/ds-algorithms/index.md b/contrib/ds-algorithms/index.md index d3c3c27..e3fc378 100644 --- a/contrib/ds-algorithms/index.md +++ b/contrib/ds-algorithms/index.md @@ -10,4 +10,4 @@ - [Greedy Algorithms](greedy-algorithms.md) - [Dynamic Programming](dynamic-programming.md) - [Linked list](linked-list.md) -- [Sliding Window Technique](Sliding_Window.md) +- [Sliding Window Technique](sliding-window.md) From 234df0ee0b4b4d29a799eac61b942cea57a785a9 Mon Sep 17 00:00:00 2001 From: Ankit Mahato Date: Sun, 2 Jun 2024 03:20:46 +0530 Subject: [PATCH 37/37] Rename Sliding_Window.md to sliding-window.md --- contrib/ds-algorithms/{Sliding_Window.md => sliding-window.md} | 0 1 file changed, 0 insertions(+), 0 deletions(-) rename contrib/ds-algorithms/{Sliding_Window.md => sliding-window.md} (100%) diff --git a/contrib/ds-algorithms/Sliding_Window.md b/contrib/ds-algorithms/sliding-window.md similarity index 100% rename from contrib/ds-algorithms/Sliding_Window.md rename to contrib/ds-algorithms/sliding-window.md