From 782cdb9493ee7bc905b74bd9fe16df4e30ed4c51 Mon Sep 17 00:00:00 2001 From: Petra Bosilj Date: Sun, 1 Dec 2024 16:38:02 +0000 Subject: [PATCH] Week 11 workshops and Week 8 solutions --- .../workshops/Week 11/Week11_workshop.ipynb | 894 +++++++++++++ .../Week 8/Week8_workshop_solutions.ipynb | 1181 +++++++++++++++++ 2 files changed, 2075 insertions(+) create mode 100644 CMP3751 Machine Learning/workshops/Week 11/Week11_workshop.ipynb create mode 100644 CMP3751 Machine Learning/workshops/Week 8/Week8_workshop_solutions.ipynb diff --git a/CMP3751 Machine Learning/workshops/Week 11/Week11_workshop.ipynb b/CMP3751 Machine Learning/workshops/Week 11/Week11_workshop.ipynb new file mode 100644 index 0000000..ac1837c --- /dev/null +++ b/CMP3751 Machine Learning/workshops/Week 11/Week11_workshop.ipynb @@ -0,0 +1,894 @@ +{ + "cells": [ + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "# Week 11 - Artificial Neural Networks (Part 2)\n", + "\n", + "As the last workshop focused on the _perceptron_, this one is going to focus on the next two \"steps\" in constructing a neural network: **activation functions** and **multi-layer perceptron**. An artificial neural network is simply a multi-layer perceptron, using Adaline perceptrons (with a choice of differentiable activation function) as units.\n", + "\n", + "This workshop will also be a mix of going through some implementations together, with **five (5) exercises**.\n", + "\n", + "- [Activation functions](#Activation-functions)\n", + " - [Exercise 1](#Exercise-1)\n", + "- [Implementation of multi-layer perceptron](#Implementation-of-multi-layer-perceptron)\n", + " - [Excercise 2: Non-linearity](#Excercise-2:-Non-linearity)\n", + " - [Exercise 3: Dense layer](#Exercise-3:-Dense-layer)\n", + " - [Loss](#Loss)\n", + "- [MLP Network](#MLP-Network)\n", + " - [Data](#Data)\n", + " - [Exercise 4](#Exercise-4)\n", + "- [Exercise 5: Initialisation](#Exercise-5:-Initialisation)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "### Activation functions\n", + "\n", + "Single-layer perceptrons were not able to succesfully classify when classes were not _linearly separable_. To design a classifier which can cope with such classes, last workshop we looked at the _kernel trick_. Another way is to chain single-layer perceptrons into a network consisting of **inputs**, one or more **hidden layers** and the **output layer**.\n", + "\n", + "In order for the **multi-layer perceptron** to learn, we rely on stochastic gradient descent, which requires a **differentiable activation function**.\n", + "\n", + "First let's revise how to use `sklearn` implementation of multi-layer perceptron, on a smiple linearly separable dataset." + ] + }, + { + "cell_type": "code", + "execution_count": 46, + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "# modified from https://scikit-learn.org/stable/auto_examples/neural_networks/plot_mlp_alpha.html#sphx-glr-auto-examples-neural-networks-plot-mlp-alpha-py\n", + "import matplotlib.pyplot as plt\n", + "import numpy as np\n", + "from sklearn.datasets import make_moons, make_circles, make_classification\n", + "from sklearn.model_selection import train_test_split\n", + "from matplotlib.colors import ListedColormap\n", + "from sklearn.neural_network import MLPClassifier\n", + "\n", + "# make linearly separable classes, slightly shuffled\n", + "X, y = make_classification(n_features=2, n_redundant=0, n_informative=2, class_sep = 1.5, \n", + " random_state=0, n_clusters_per_class=1)\n", + "rng = np.random.RandomState(2)\n", + "X += 2 * rng.uniform(size=X.shape)\n", + "\n", + "# just use 'identity' in this example\n", + "activation_function = 'identity'\n", + "\n", + "# split the data\n", + "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.4)\n", + "\n", + "\n", + "# initialise the classifier\n", + "clf = MLPClassifier(solver='lbfgs', alpha=0, random_state=1, max_iter=2000,\n", + " early_stopping=True, hidden_layer_sizes=[100, 100],\n", + " activation = activation_function)\n", + "\n", + "# fit the data, examine error\n", + "clf.fit(X_train, y_train)\n", + "score = clf.score(X_test, y_test)\n", + "\n", + "# below is used for plotting the train/test data and the results\n", + "\n", + "fig, ax = plt.subplots(1, 2, figsize = (10, 5))\n", + "\n", + "x_min, x_max = X[:, 0].min() - .5, X[:, 0].max() + .5\n", + "y_min, y_max = X[:, 1].min() - .5, X[:, 1].max() + .5\n", + "xx, yy = np.meshgrid(np.arange(x_min, x_max, 0.02),\n", + " np.arange(y_min, y_max, 0.02))\n", + "\n", + "# just plot the dataset first\n", + "cm = plt.cm.RdBu\n", + "cm_bright = ListedColormap(['#FF0000', '#0000FF'])\n", + "# Plot the training points\n", + "ax[0].scatter(X_train[:, 0], X_train[:, 1], c=y_train, cmap=cm_bright)\n", + "# and testing points\n", + "ax[0].scatter(X_test[:, 0], X_test[:, 1], c=y_test, cmap=cm_bright, alpha=0.6)\n", + "ax[0].set_xlim(xx.min(), xx.max())\n", + "ax[0].set_ylim(yy.min(), yy.max())\n", + "ax[0].set_xticks(())\n", + "ax[0].set_yticks(())\n", + "\n", + "if hasattr(clf, \"decision_function\"):\n", + " Z = clf.decision_function(np.c_[xx.ravel(), yy.ravel()])\n", + "else:\n", + " Z = clf.predict_proba(np.c_[xx.ravel(), yy.ravel()])[:, 1]\n", + "\n", + "# Put the result into a color plot\n", + "Z = Z.reshape(xx.shape)\n", + "ax[1].contourf(xx, yy, Z, cmap=cm, alpha=.8)\n", + "\n", + "# Plot also the training points\n", + "ax[1].scatter(X_train[:, 0], X_train[:, 1], c=y_train, cmap=cm_bright,\n", + " edgecolors='black', s=25)\n", + "# and testing points\n", + "ax[1].scatter(X_test[:, 0], X_test[:, 1], c=y_test, cmap=cm_bright,\n", + " alpha=0.6, edgecolors='black', s=25)\n", + "\n", + "ax[1].set_xlim(xx.min(), xx.max())\n", + "ax[1].set_ylim(yy.min(), yy.max())\n", + "ax[1].set_xticks(())\n", + "ax[1].set_yticks(())\n", + "#ax[1].set_title(name)\n", + "ax[1].text(xx.max() - .3, yy.min() + .3, ('%.2f' % score).lstrip('0'),\n", + " size=15, horizontalalignment='right')\n", + " \n", + "ax[1].set_title('Classification results'.format(activation_function))\n", + " \n", + "plt.show()" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "### Excercise 1\n", + "\n", + "Let's examine the influence of the activation function when the datset used is not linearly separable. Below sinippet of code runs a multi-layer perceptron with two hidden layers of size (100, 100) on the 'moon' and 'circle' data shown. The learning is done with an identity **activation function**.\n", + "\n", + "Try using different choices of activation functions. Options include:\n", + "- `identity`\n", + "\\begin{equation*}\n", + "f_{id}(x) = x\n", + "\\end{equation*}\n", + "- `tanh`\n", + "\\begin{equation*}\n", + "f_{tanh}(x) = \\tanh{x}\n", + "\\end{equation*}\n", + "- `logistic` (sigmoid -- used in lecture)\n", + "\\begin{equation*}\n", + "f_{\\sigma}(x) = \\frac{1}{1+ e^{-x}}\n", + "\\end{equation*}\n", + "- `relu` (rectified linear unit -- deep learning)\n", + "\\begin{equation*}\n", + "f_{relu}(x) = \\max(0, x)\n", + "\\end{equation*}\n", + "\n", + "Look at the results of the classification. Think about the following:\n", + "- Are there difference in the quality of the results?\n", + "- Are all the activations able to learn the \"correct\" classification?\n", + "- Which activation functions is the best/worst?" + ] + }, + { + "cell_type": "code", + "execution_count": 49, + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "# modified from https://scikit-learn.org/stable/auto_examples/neural_networks/plot_mlp_alpha.html#sphx-glr-auto-examples-neural-networks-plot-mlp-alpha-py\n", + "\n", + "from sklearn.datasets import make_moons, make_circles\n", + "from sklearn.model_selection import train_test_split\n", + "from matplotlib.colors import ListedColormap\n", + "from sklearn.neural_network import MLPClassifier\n", + "\n", + "datasets = [make_moons(noise=0.3, random_state = 0),\n", + " make_circles(noise=0.2, factor = 0.5, random_state = 1)]\n", + "\n", + "\n", + "##################################################\n", + "# options 'identity', logistic', 'tahn', 'relu'\n", + "##################################################\n", + "# change below to try different activations\n", + "##################################################\n", + "activation_function = 'identity'\n", + "\n", + "fig, axs = plt.subplots(2, 2, figsize = (10, 10))\n", + "\n", + "for (X, y), ax in zip(datasets, axs):\n", + " X_train, X_test, y_train, y_test = train_test_split(X, y, test_size = 0.4)\n", + " \n", + " x_min, x_max = X[:, 0].min() - .5, X[:, 0].max() + .5\n", + " y_min, y_max = X[:, 1].min() - .5, X[:, 1].max() + .5\n", + " xx, yy = np.meshgrid(np.arange(x_min, x_max, 0.02),\n", + " np.arange(y_min, y_max, 0.02))\n", + "\n", + " # just plot the dataset first\n", + " cm = plt.cm.RdBu\n", + " cm_bright = ListedColormap(['#FF0000', '#0000FF'])\n", + " #ax[0] = plt.subplot(len(datasets), len(classifiers) + 1, i)\n", + " # Plot the training points\n", + " ax[0].scatter(X_train[:, 0], X_train[:, 1], c=y_train, cmap=cm_bright)\n", + " # and testing points\n", + " ax[0].scatter(X_test[:, 0], X_test[:, 1], c=y_test, cmap=cm_bright, alpha=0.6)\n", + " ax[0].set_xlim(xx.min(), xx.max())\n", + " ax[0].set_ylim(yy.min(), yy.max())\n", + " ax[0].set_xticks(())\n", + " ax[0].set_yticks(())\n", + " \n", + " \n", + " clf = MLPClassifier(solver='lbfgs', alpha=0, random_state=1, max_iter=2000,\n", + " early_stopping=True, hidden_layer_sizes=[100, 100],\n", + " activation = activation_function)\n", + " \n", + " clf.fit(X_train, y_train)\n", + " score = clf.score(X_test, y_test)\n", + " \n", + " if hasattr(clf, \"decision_function\"):\n", + " Z = clf.decision_function(np.c_[xx.ravel(), yy.ravel()])\n", + " else:\n", + " Z = clf.predict_proba(np.c_[xx.ravel(), yy.ravel()])[:, 1]\n", + "\n", + " # Put the result into a color plot\n", + " Z = Z.reshape(xx.shape)\n", + " ax[1].contourf(xx, yy, Z, cmap=cm, alpha=.8)\n", + "\n", + " # Plot also the training points\n", + " ax[1].scatter(X_train[:, 0], X_train[:, 1], c=y_train, cmap=cm_bright,\n", + " edgecolors='black', s=25)\n", + " # and testing points\n", + " ax[1].scatter(X_test[:, 0], X_test[:, 1], c=y_test, cmap=cm_bright,\n", + " alpha=0.6, edgecolors='black', s=25)\n", + "\n", + " ax[1].set_xlim(xx.min(), xx.max())\n", + " ax[1].set_ylim(yy.min(), yy.max())\n", + " ax[1].set_xticks(())\n", + " ax[1].set_yticks(())\n", + " #ax[1].set_title(name)\n", + " ax[1].text(xx.max() - .3, yy.min() + .3, ('%.2f' % score).lstrip('0'),\n", + " size=15, horizontalalignment='right')\n", + " \n", + "axs[0][1].set_title('Activation {} - classification results'.format(activation_function))\n", + " \n", + "plt.show()" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "### Implementation of multi-layer perceptron\n", + "\n", + "Below we are going to work through an implementation of a multi-layer perceptron. We will use it on the digits dataset later in this notebook.\n", + "\n", + "Firstly, we're going to define an abstract concept of a \"layer\". Layer is just a collection of processing units, which takes some features (input features, or output from a different layer) as input, performs some operation, and outputs new features. Each Layer must implement two operations: a `_forward()` pass, which takes the inputs, applies the operation, and produces the outputs, as well as a `_backward()` pass, that takes the gradients of the next layer, updates the weights based on the input values and the output gradient, and passes the output gradient of the current layer to any previous ones.\n", + "\n", + "This implementation is adapted from [this tutorial](https://github.com/aayushmnit/Deep_learning_explorations/blob/master/1_MLP_from_scratch/Building_neural_network_from_scratch.ipynb)." + ] + }, + { + "cell_type": "code", + "execution_count": 50, + "metadata": {}, + "outputs": [], + "source": [ + "import numpy as np ## For numerical python\n", + "np.random.seed(42)\n", + "\n", + "from abc import ABC, abstractmethod\n", + "\n", + "class Layer(ABC):\n", + " \"\"\"\n", + " A building block. Each layer is capable of performing two things:\n", + " - Process input to get output: output = layer._forward(input)\n", + " - Propagate gradients through itself: input_grad = layer._backward(input, output_grad)\n", + " Some layers also have learnable parameters which they update during layer._backward.\n", + " \"\"\"\n", + " def __init__(self):\n", + " \"\"\"Here we can initialize layer parameters (if any) and auxiliary stuff.\"\"\"\n", + " # A dummy layer does nothing\n", + " pass\n", + " \n", + " @abstractmethod\n", + " def _forward(self, input):\n", + " \"\"\"\n", + " Takes input data of shape [batch, input_units], returns output data [batch, output_units]\n", + " \"\"\"\n", + " # A dummy layer just returns whatever it gets as input.\n", + " return input\n", + "\n", + " @abstractmethod\n", + " def _backward(self, input, output_grad):\n", + " \"\"\"\n", + " Performs a backpropagation step through the layer, with respect to the given input. \n", + " To compute loss gradients w.r.t input, we need to apply chain rule (backprop): \n", + " d loss / d x = (d loss / d layer) * (d layer / d x)\n", + " Luckily, we already receive d loss / d layer as input, so you only need to multiply it by d layer / d x. \n", + " If our layer has parameters (e.g. dense layer), we also need to update them here using d loss / d layer\n", + " \"\"\"\n", + " # The gradient of a dummy layer is precisely grad_output, but we'll write it more explicitly\n", + " num_units = input.shape[1]\n", + " \n", + " d_layer_d_input = np.eye(num_units)\n", + " \n", + " return np.dot(output_grad, d_layer_d_input) # chain rule\n" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "#### Excercise 2: Non-linearity\n", + "\n", + "The simplest layers we will be implementing are our non-linear layers, or **activation functions**. We will be using ReLU, rectified linear unit:\n", + "\\begin{equation*}\n", + "f_{relu}(x) = \\max(0, x)\n", + "\\end{equation*}\n", + "With the following derivative:\n", + "\\begin{equation*}\n", + " f'_{relu}(x)=\\begin{cases}\n", + " 1, & \\text{if $x>0$}.\\\\\n", + " 0, & \\text{otherwise}.\n", + " \\end{cases}\n", + "\\end{equation*}\n", + "\n", + "We will also look at Leaky ReLU, a variant of ReLU which allows a small, non constant gradient $\\alpha$ to pass through when the inputs are < 0, according to the following:\n", + "\\begin{equation*}\n", + "f_{leaky_relu}(x) = \\max(\\alpha x, x)\n", + "\\end{equation*}\n", + "With the following derivative:\n", + "With the following derivative:\n", + "\\begin{equation*}\n", + " f'_{relu}(x)=\\begin{cases}\n", + " \\alpha, & \\text{if $x>0$}.\\\\\n", + " 0, & \\text{otherwise}.\n", + " \\end{cases}\n", + "\\end{equation*}\n", + "\n", + "_Note:_ Don't forget, the derivatives need to be multiplied by the output gradient in the `_backward()` pass.\n", + "\n", + "The below class should implement the ReLU and Leaky ReLU functionality. **Insert the code in the `_forward()` method of both classes to achieve the functionality of the function above, as well as the `_backward()` method of the `LeakyReLU` calss.** The cell below will plot the functionality of the layer you implemented to check for correctness.\n", + "\n", + "The `_backward()` pass function is fully implemented for `ReLU`." + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "class ReLU(Layer):\n", + " def __init__(self):\n", + " \"\"\"ReLU layer simply applies elementwise rectified linear unit to all inputs\"\"\"\n", + " pass\n", + " \n", + " def _forward(self, input):\n", + " \"\"\"Apply elementwise ReLU to [batch, input_units] matrix\"\"\"\n", + " ############################################\n", + " ########### YOUR IMPLEMENTATION HERE #######\n", + " #### replace the following line: ###########\n", + " ############################################\n", + " return 0\n", + " \n", + " def _backward(self, input, output_grad):\n", + " \"\"\"Compute gradient of loss w.r.t. ReLU input\"\"\"\n", + " \n", + " relu_grad = input > 0\n", + " return output_grad*relu_grad \n", + "\n", + "\n", + "class LeakyReLU(Layer):\n", + " def __init__(self, alpha = 0.01):\n", + " \"\"\"ReLU layer simply applies elementwise rectified linear unit to all inputs\"\"\"\n", + " self._alpha = alpha\n", + " pass\n", + " \n", + " def _forward(self, input):\n", + " \"\"\"Apply elementwise ReLU to [batch, input_units] matrix\"\"\"\n", + " ############################################\n", + " ########### YOUR IMPLEMENTATION HERE #######\n", + " #### replace the following line: ###########\n", + " ############################################\n", + " return 0\n", + " \n", + " def _backward(self, input, output_grad):\n", + " \"\"\"Compute gradient of loss w.r.t. ReLU input\"\"\"\n", + " ############################################\n", + " ########### YOUR IMPLEMENTATION HERE #######\n", + " #### replace the following line: ###########\n", + " ############################################\n", + " lrelu_grad = 0\n", + " return output_grad*lrelu_grad \n" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "To check if your implementation of ReLU and Leaky ReLU is correct, you can plot the output of this layer for a range of (1-d) inputs:" + ] + }, + { + "cell_type": "code", + "execution_count": 42, + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + }, + { + "name": "stdout", + "output_type": "stream", + "text": [ + "[0. 1.5 0. 1.7]\n" + ] + } + ], + "source": [ + "import matplotlib.pyplot as plt\n", + "\n", + "fig, ax = plt.subplots(1, 2, figsize = (15, 5))\n", + "\n", + "X = np.arange(-2, 2, 0.1)\n", + "relu = ReLU()\n", + "lrelu = LeakyReLU(0.1)\n", + "\n", + "#y = [relu._forward(x) for x in X]\n", + "y_relu = relu._forward(X)\n", + "y_lrelu = lrelu._forward(X)\n", + "\n", + "ax[0].plot(X, y_relu)\n", + "ax[0].set_title('ReLU')\n", + "ax[0].set_xlabel('input')\n", + "ax[0].set_ylabel('output')\n", + "\n", + "ax[1].plot(X, y_lrelu)\n", + "ax[1].set_title('Leaky ReLU')\n", + "ax[1].set_xlabel('input')\n", + "ax[1].set_ylabel('output')\n", + "\n", + "\n", + "plt.show()" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "#### Exercise 3: Dense layer\n", + "\n", + "For computational efficiency, it is better to model a whole layer of neurons, instead of each neuron separately. This is referred to as the **dense** layer as every neuron in a dense layer takes all the inputs from the previous layer (and it's all outputs are used as inputs to the next layer).\n", + "\n", + "We can model this as:\n", + "\n", + "\\begin{equation*}\n", + "f(X) = W \\cdot X + \\vec{b}\n", + "\\end{equation*}\n", + "\n", + "Here, the matrix $W$ is are the weights for each of the neurons. Each neuron has as many weights as there are `input_units` (and one bias). So there are total `input_units`*`output_units` weights, and `output_units` biases.\n", + "\n", + "Just like before, **insert the code in the `_forward()` method to achieve the functionality of the function above.** The `_backward()` pass function is fully implemented.\n", + "\n", + "Instead of a random initialisation (or zero initialisation) of weights and biases, we are using [Xavier initialisation](https://andyljones.tumblr.com/post/110998971763/an-explanation-of-xavier-initialization) which helps the model converge faster. Instead of initializing our weights with small numbers which are distributed randomly we initialize our weights with mean zero and variance of 2/(number of inputs + number of outputs)." + ] + }, + { + "cell_type": "code", + "execution_count": 20, + "metadata": {}, + "outputs": [], + "source": [ + "class Dense(Layer):\n", + " def __init__(self, input_units, output_units, learning_rate=0.1):\n", + " \"\"\"\n", + " A dense layer is a layer which performs a learned affine transformation:\n", + " f(x) = + b\n", + " \"\"\"\n", + " self.learning_rate = learning_rate\n", + " \n", + " self.weights = np.random.normal(loc=0.0, \n", + " scale = np.sqrt(2/(input_units+output_units)), \n", + " size = (input_units,output_units))\n", + " self.biases = np.zeros(output_units)\n", + " \n", + " def _forward(self,input):\n", + " \"\"\"\n", + " Perform an affine transformation:\n", + " f(x) = + b\n", + " \n", + " input shape: [batch, input_units]\n", + " output shape: [batch, output units]\n", + " \"\"\"\n", + " ############################################\n", + " ########### YOUR IMPLEMENTATION HERE #######\n", + " #### replace the following line: ###########\n", + " ############################################\n", + " return 0\n", + " \n", + " def _backward(self,input,output_grad):\n", + " # compute d f / d x = d f / d dense * d dense / d x\n", + " # where d dense/ d x = weights transposed\n", + " input_grad = np.dot(output_grad, self.weights.T)\n", + " \n", + " # compute gradient w.r.t. weights and biases\n", + " weights_grad = np.dot(input.T, output_grad)\n", + " biases_grad = output_grad.mean(axis=0)*input.shape[0]\n", + " \n", + " assert weights_grad.shape == self.weights.shape and biases_grad.shape == self.biases.shape\n", + " \n", + " # Here we perform a stochastic gradient descent step. \n", + " self.weights = self.weights - self.learning_rate * weights_grad\n", + " self.biases = self.biases - self.learning_rate * biases_grad\n", + " \n", + " return input_grad" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "The below code is a simple test of your implementation of the `_forward()` function. It creates a `Dense()` layer which accepts two inputs (samples with two features), and returns a single output. Then, it creates an array of three samples called `input` and passes it through the network to obtain the `output`. Finally, there is an `assert` command manually testing the output for the second sample.\n", + "\n", + "**Modify the code such that**:\n", + "- you create a Dense layer that accepts samples with two features, but returns _two_ outputs instead of one\n", + "- the `assert` statement checks the correctness of the _second output_ for the _third sample_\n", + "\n", + "_Reminder:_ The `assert` command will not do anything if the argument passed to it is `True` (in our case, the condition checking for the correctness of the output). Alternatively, it will cause an error if passed a `False` value." + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [ + "test_layer = Dense(2, 1)\n", + "print(test_layer.weights, test_layer.biases)\n", + "input = np.array([[1,2],\n", + " [2,4],\n", + " [3,6]])\n", + "output = test_layer._forward(input)\n", + "print(output)\n", + "assert(output[1] == input[1,0]*test_layer.weights[0]+input[1,1]*test_layer.weights[1]+test_layer.biases[0])" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "#### Loss \n", + "\n", + "You can see that the above layer expects the output of the gradient during back propagation, but does not compute it itself.\n", + "\n", + "Therefore, we will use a `Loss` class which will be able to return the loss and it's graditent. First we define a template method and then the actual loss we will use.\n", + "\n", + "This is a modification of **cross-entropy** loss. This is a component-wise loss -- the expected output format is a single output class our classifier can predict. \n", + "\n", + "The network will return unconstrained outputs (which can be greater or lower than zero). The 'softmax' function is a way to map those outputs into class probabilities. However, the crossentropy loss can be more efficiently calculated without first calculating softmax, which is implemented below.\n", + "\n", + "However, we could implement any other `Loss` function to use with this code." + ] + }, + { + "cell_type": "code", + "execution_count": 7, + "metadata": {}, + "outputs": [], + "source": [ + "class Loss(ABC):\n", + " def __init__(self):\n", + " self.calculated = False\n", + " pass\n", + " \n", + " @abstractmethod\n", + " def _calculate_loss(self, output, target):\n", + " self.calculated = True\n", + " pass\n", + " \n", + " @abstractmethod\n", + " def _grad(self):\n", + " assert self.calculated == True\n", + " pass\n", + "\n", + "class softmax_crossentropy_with_logits(Loss):\n", + " def __init__(self):\n", + " self.one_hot_target = np.array([])\n", + " self.softmax = np.array([])\n", + " super().__init__()\n", + " \n", + " \n", + " def _calculate_loss(self, output, target):\n", + " super()._calculate_loss(output, target)\n", + " logits_for_target = output[np.arange(len(output)),target]\n", + " xentropy = - logits_for_target + np.log(np.sum(np.exp(output),axis=-1))\n", + " \n", + " self.one_hot_target = np.zeros_like(output)\n", + " self.one_hot_target[np.arange(len(output)), target] = 1\n", + " \n", + " self.softmax = np.exp(output) / np.exp(output).sum(axis = -1, keepdims = True)\n", + " \n", + " return xentropy\n", + " \n", + " def _grad(self):\n", + " super()._grad()\n", + " \n", + " rvalue = (- self.one_hot_target + self.softmax) / self.softmax.shape[0]\n", + " #print(rvalue.shape)\n", + " \n", + " return (- self.one_hot_target + self.softmax) / self.softmax.shape[0]\n" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "### MLP Network.\n", + "\n", + "Finally, we define MLP. Despite being a collection of layers, we can think of it as just a big 'layer': we can pass inputs _forward_ through the network and propagate the error _backwards_.\n", + "\n", + "However, will add a `train` method that will do the above, together with calculating the `loss` (pass inputs forward, calculate the error, and propagate it backwards through the networks).\n", + "\n", + "We will also add a predict method. We assume our netork outputs log-probabilities (logits), and the final class prediction is the one with highest (log)-probability." + ] + }, + { + "cell_type": "code", + "execution_count": 8, + "metadata": {}, + "outputs": [], + "source": [ + "class Network(Layer):\n", + " def __init__(self, layers, loss = softmax_crossentropy_with_logits()):\n", + " \"\"\"ReLU layer simply applies elementwise rectified linear unit to all inputs\"\"\"\n", + " self.layers = layers\n", + " self.activations = []\n", + " self.loss_ = []\n", + " self.loss = loss\n", + " pass\n", + " \n", + " def _forward(self, input):\n", + " self.activations = []\n", + "\n", + " # Looping through each layer\n", + " for l in self.layers:\n", + " self.activations.append(l._forward(input))\n", + " # Updating input to last layer output\n", + " input = self.activations[-1]\n", + "\n", + " assert len(self.activations) == len(self.layers)\n", + " \n", + " return self.activations[-1]\n", + " \n", + " def _backward(self, input, output_grad):\n", + " loss_grad = output_grad\n", + " \n", + " layer_inputs = [input]+self.activations #layer_input[i] is an input for network[i]\n", + " # Propagate gradients through the network\n", + " # Reverse propogation as this is backprop\n", + " for layer_index in range(len(self.layers))[::-1]:\n", + " layer = self.layers[layer_index]\n", + "\n", + " loss_grad = layer._backward(layer_inputs[layer_index],loss_grad) #grad w.r.t. input, also weight updates \n", + " \n", + " return loss_grad \n", + " \n", + " def train(self,X,y):\n", + " \"\"\"\n", + " Train our network on a given batch of X and y.\n", + " We first need to run forward to get all layer activations.\n", + " Then we can run layer.backward going from last to first layer.\n", + " After we have called backward for all layers, all Dense layers have already made one gradient step.\n", + " \"\"\"\n", + "\n", + " # Get the layer activations\n", + " logits = self._forward(X)\n", + " \n", + " # Compute the loss and the initial gradient\n", + " loss = self.loss._calculate_loss(logits, y)\n", + " \n", + " # propagate gradients through the network\n", + " self._backward(X, self.loss._grad())\n", + " \n", + " return self\n", + " \n", + " def predict(self,X):\n", + " \"\"\"\n", + " Compute network predictions. Returning indices of largest Logit probability\n", + " \"\"\"\n", + " logits = self._forward(X)\n", + " return logits.argmax(axis=-1)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "#### Data \n", + "\n", + "Let's work with some data with more than a handful of dimensions. We will use a toy dataset of handwritten digits accessible thorugh `sklearn`. These are quantised 8x8pixel 4-bit images.\n", + "\n", + "We will load them, display some of them for illustration, \"flatten\" them from 8x8 and split 50-50 to test and train (not best practice, just for illustration!).\n", + "\n", + "We are loading the data before setting up the network, to know the number of our inputs. Here, it will be 64-dimensional samples (from 8x8 pixels)." + ] + }, + { + "cell_type": "code", + "execution_count": 9, + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "iVBORw0KGgoAAAANSUhEUgAAAxsAAADSCAYAAAAi0d0oAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy81sbWrAAAACXBIWXMAAA9hAAAPYQGoP6dpAAATiUlEQVR4nO3dfWyV5d0H8B8gpRVMKVEBHbR24gtObRQ0/qVACwmosEiNW+ZaNQgZsjmTWaPZwLcF5mI1kjDYJsvCtmwovmzqHCQtMUMXwYFuyzIDON2yKUhpVJhVej1/PLFbB1uLz309h9bPJ2lC73Of77lO2197vtzn3GdISikFAABAwYaWegEAAMDgpGwAAABZKBsAAEAWygYAAJCFsgEAAGShbAAAAFkoGwAAQBbKBgAAkIWyAQAAZKFsHIOGDBkSy5YtK/Uy4JhmTqB/zAr0zZzkM2jLxiuvvBLz58+P6urqKC8vj1NPPTUaGhrioYceKvXSjinf//734+yzz47y8vKYNGmSr88njDnp26pVq6KxsTEmTpwYQ4YMiebm5lIviRIwK//dG2+8EXfeeWdcdNFFUVVVFSeeeGJcdtllsWnTplIvjf9H5uS/O3jwYNxwww3xmc98JiorK2PUqFFx/vnnx4MPPhgffPBBqZeXzXGlXkAOW7ZsiWnTpsXEiRNjwYIFMW7cuHjjjTfihRdeiAcffDCWLFlS6iUeE1avXh2LFi2Kq666Km655ZZ47rnn4stf/nIcOHAgWlpaSr08MjMn/bNixYp455134qKLLoq//e1vpV4OJWBW+vbEE0/EihUrYt68edHU1BQffvhh/PCHP4yGhoZ4+OGH47rrriv1EsnMnPTt4MGD8fvf/z5mz54dNTU1MXTo0NiyZUt89atfjd/85jfx4x//uNRLzGJQlo177703Kisr48UXX4zRo0f3uuytt94qzaKOMQcPHow77rgj5syZE4888khERCxYsCC6u7vj7rvvjhtvvDGqqqpKvEpyMif9s3nz5p6jGqNGjSr1cigBs9K3adOmxeuvvx4nnnhiz7ZFixZFXV1dfOMb31A2PgHMSd/GjBkTL7zwQq9tixYtisrKyli5cmXcf//9MW7cuBKtLp9B+TSqnTt3xjnnnHPYD3tExMknn9zr87Vr18b06dPj5JNPjhEjRsTkyZNj1apVh12vpqYmLr/88mhvb48pU6ZERUVFnHvuudHe3h4RERs2bIhzzz03ysvL48ILL4zf/va3va7f3Nwco0aNil27dsWsWbNi5MiRccopp8Rdd90VKaU+79Nf//rXuP7662Ps2LExYsSIOOecc+Lhhx8+bL/XX389/vjHP/aZ19bWFm+//XZ86Utf6rV98eLF8d5778VTTz3VZwYDmznpe04iIqqrq2PIkCH92pfByaz0PSvnnHNOr6IRETFixIiYPXt2/OUvf4l33nmnzwwGNnPSv78pR1JTUxMREfv37//YGce0NAjNnDkznXDCCemVV17pc9+pU6em5ubm1Nramh566KE0c+bMFBFp5cqVvfarrq5OZ555Zho/fnxatmxZam1tTaeeemoaNWpUWrduXZo4cWJavnx5Wr58eaqsrEynn356OnToUM/1m5qaUnl5eZo0aVK69tpr08qVK9Pll1+eIiJ9/etf73VbEZGWLl3a8/nf//739KlPfSpNmDAh3XXXXWnVqlXpyiuvTBGRWltbe1330ksvTf35tt5zzz0pItKbb77Za/v777+fhg4dmm655ZY+MxjYzMnR//obOXJkampqOurrMbCZlY//UOHzn/98Ov7449OHH374sTMYGMxJ/+fk/fffT3v27Emvv/562rBhQxo3blyqrq5OH3zwQb8zBpJBWTZ+9atfpWHDhqVhw4alSy65JN16663p2WefTV1dXYfte+DAgcO2zZo1K9XW1vbaVl1dnSIibdmypWfbs88+myIiVVRUpD//+c8921evXp0iIrW1tfVsa2pqShGRlixZ0rOtu7s7zZkzJ5WVlaU9e/b0bP/3H/gbbrghjR8/Pu3du7fXmq655ppUWVnZ6z709wd+8eLFadiwYUe87KSTTkrXXHNNnxkMbOZE2aB/zMrHe6jw6quvpvLy8nTttdd+rOszsJiT/s/JT37ykxQRPR9TpkxJL7/8cr+vP9AMyqdRNTQ0xPPPPx9XXnll7NixI771rW/FrFmz4tRTT40nn3yy174VFRU9/+7s7Iy9e/fGpZdeGrt27YrOzs5e+06ePDkuueSSns8vvvjiiIiYPn16TJw48bDtu3btOmxtN910U8+/hwwZEjfddFN0dXX9xzN2pJTi0UcfjSuuuCJSSrF3796ej1mzZkVnZ2e89NJLPfu3t7f369DgwYMHo6ys7IiXlZeXx8GDB/vMYGAzJ33PCUSYlY8zKwcOHIjGxsaoqKiI5cuXH/X1GXjMSf/nZNq0abFx48ZYv359LFq0KIYPHx7vvfdev68/0AzKF4hHREydOjU2bNgQXV1dsWPHjnjssceitbU15s+fH9u3b4/JkydHRMSvf/3rWLp0aTz//PNx4MCBXhmdnZ1RWVnZ8/m//lBHRM9lEyZMOOL2jo6OXtuHDh0atbW1vbadccYZERHx2muvHfF+7NmzJ/bv3x9r1qyJNWvWHHGfj/PCq4qKiujq6jriZf/4xz96/SJg8DIn0D9mpf8OHToU11xzTfzhD3+IZ555Jk455ZT/Ux4Dhznpn7Fjx8bYsWMjImL+/PnxzW9+MxoaGuLVV18dlC8QH7Rl4yNlZWUxderUmDp1apxxxhlx3XXXxfr162Pp0qWxc+fOmDFjRpx11llx//33x4QJE6KsrCyefvrpaG1tje7u7l5Zw4YNO+Jt/KftRfzP6Udr+MIXvhBNTU1H3Oe888476tzx48fHoUOH4q233ur1wq2urq54++23/XH4hDEn0D9mpW8LFiyIX/ziF/GjH/0opk+f/n/KYmAyJ0dn/vz5cccdd8QTTzwRCxcuLCz3WDHoy8a/mjJlSkREz7nyf/7zn8f7778fTz75ZK/m3NbWluX2u7u7Y9euXT2NOiLiT3/6U0T880wE/+6kk06KE044IQ4dOhT19fWFraWuri4iIrZu3RqzZ8/u2b5169bo7u7uuZxPHnMC/WNWDve1r30t1q5dGw888EB87nOfKzyfgcec9O2jp67/+1PIBotB+ZqNtra2Izbbp59+OiIizjzzzIj4Zyv+1307Oztj7dq12da2cuXKnn+nlGLlypUxfPjwmDFjxhH3HzZsWFx11VXx6KOPxu9+97vDLt+zZ0+vz/t7+rXp06fHmDFjDjvV3KpVq+L444+POXPm9OfuMICZk49/mkI+WcxK/2blvvvui29/+9tx++23x1e+8pWjuBcMBuak7znZu3fvEb9G3/ve9yLin8VssBmURzaWLFkSBw4ciM9+9rNx1llnRVdXV2zZsiV++tOfRk1NTc+bC82cOTPKysriiiuuiIULF8a7774b3/3ud+Pkk0/O8k7B5eXl8ctf/jKampri4osvjmeeeSaeeuqpuP322+Okk076j9dbvnx5tLW1xcUXXxwLFiyIyZMnx759++Kll16KTZs2xb59+3r2/eIXvxibN2/u8zBiRUVF3H333bF48eJobGyMWbNmxXPPPRfr1q2Le++9N8aMGVPY/ebYZE76npOI//1fuB07dkRExAcffBAvv/xy3HPPPRERceWVV3p61ieAWel7Vh577LG49dZbY9KkSXH22WfHunXrel3e0NDQ8xx1Bidz0vecrFu3Lr7zne/EvHnzora2Nt5555149tlnY+PGjXHFFVcM3qcdZj/fVQk888wz6frrr09nnXVWGjVqVCorK0unn356WrJkyWHvK/Hkk0+m8847L5WXl6eampq0YsWK9PDDD6eISLt37+7Zr7q6Os2ZM+ew24qItHjx4l7bdu/enSIi3XfffT3bmpqa0siRI9POnTvTzJkz0/HHH5/Gjh2bli5d2uuc0B9l/uvp11JK6c0330yLFy9OEyZMSMOHD0/jxo1LM2bMSGvWrOm139Gefm3NmjXpzDPPTGVlZenTn/50am1tTd3d3f2+PgOXOenfnHx06sQjfaxdu7ZfGQxsZqXvWVm6dOl/nJP4t9ORMjiZk77n5MUXX0yNjY1p4sSJacSIEWnkyJHpggsuSPfff/+gfY+NlFIakpLzP/5/aG5ujkceeSTefffdUi8FjlnmBPrHrEDfzMmxYVC+ZgMAACg9ZQMAAMhC2QAAALLwmg0AACALRzYAAIAslA0AACALZQMAAMhi0L2D+Pr16wvPbGlpKTyzoaGh8MyI/33Hy6JVVVUVnsngc9lllxWeuX///sIzIyLuvPPOwjPnzp1beCaDT3t7e+GZ8+bNKzwzIqKurq7wzBz3n9JbsWJF4Zm33XZb4ZmnnXZa4ZkREdu2bSs8czA99nJkAwAAyELZAAAAslA2AACALJQNAAAgC2UDAADIQtkAAACyUDYAAIAslA0AACALZQMAAMhC2QAAALJQNgAAgCyUDQAAIAtlAwAAyELZAAAAslA2AACALJQNAAAgC2UDAADIQtkAAACyUDYAAIAsjiv1AorW0tJSeObu3bsLz+zo6Cg8MyJizJgxhWf+7Gc/KzyzsbGx8ExKa/To0YVnbt68ufDMiIi2trbCM+fOnVt4JqW1ffv2wjOnTZtWeGZlZWXhmRERr732WpZcSuu2224rPDPH44TVq1cXnrlw4cLCMyMitm3bVnhmfX194Zml4sgGAACQhbIBAABkoWwAAABZKBsAAEAWygYAAJCFsgEAAGShbAAAAFkoGwAAQBbKBgAAkIWyAQAAZKFsAAAAWSgbAABAFsoGAACQhbIBAABkoWwAAABZKBsAAEAWygYAAJCFsgEAAGShbAAAAFkoGwAAQBbHlfLGt23bVnjm7t27C8/cuXNn4Zm1tbWFZ0ZENDQ0FJ6Z4/vU2NhYeCb9t3379sIz29vbC8/Mpa6urtRLYAB4/PHHC888//zzC8+cN29e4ZkREXfeeWeWXErrxhtvLDyzpaWl8MwLL7yw8MzTTjut8MyIiPr6+iy5g4UjGwAAQBbKBgAAkIWyAQAAZKFsAAAAWSgbAABAFsoGAACQhbIBAABkoWwAAABZKBsAAEAWygYAAJCFsgEAAGShbAAAAFkoGwAAQBbKBgAAkIWyAQAAZKFsAAAAWSgbAABAFsoGAACQhbIBAABkoWwAAABZHFfKG+/o6Cg884ILLig8s7a2tvDMXC688MJSL4GCPfDAA4VnLlu2rPDMzs7OwjNzueyyy0q9BAaAm2++ufDMmpqawjNzrDMiYu7cuVlyKa0cj2l27dpVeObu3bsLz6yvry88MyLP49mqqqrCM0vFkQ0AACALZQMAAMhC2QAAALJQNgAAgCyUDQAAIAtlAwAAyELZAAAAslA2AACALJQNAAAgC2UDAADIQtkAAACyUDYAAIAslA0AACALZQMAAMhC2QAAALJQNgAAgCyUDQAAIAtlAwAAyELZAAAAslA2AACALI4r5Y13dHQUntnQ0FB45kCS42taVVVVeCb9d/PNNxee2dzcXHjmQPo52b9/f6mXQMFyfE8feOCBwjMff/zxwjNz+cEPflDqJTBA1NbWFp65b9++wjPr6+sLz8yVu2nTpsIzS/V32pENAAAgC2UDAADIQtkAAACyUDYAAIAslA0AACALZQMAAMhC2QAAALJQNgAAgCyUDQAAIAtlAwAAyELZAAAAslA2AACALJQNAAAgC2UDAADIQtkAAACyUDYAAIAslA0AACALZQMAAMhC2QAAALJQNgAAgCyUDQAAIIvjSnnjVVVVhWdu27at8MwcOjo6suRu3bq18Myrr7668Ewope3btxeeWVdXV3gm/bds2bLCMx988MHCM3N4/PHHs+SOHj06Sy70R47HiJs2bSo8MyJi4cKFhWeuWLGi8Mzly5cXntkfjmwAAABZKBsAAEAWygYAAJCFsgEAAGShbAAAAFkoGwAAQBbKBgAAkIWyAQAAZKFsAAAAWSgbAABAFsoGAACQhbIBAABkoWwAAABZKBsAAEAWygYAAJCFsgEAAGShbAAAAFkoGwAAQBbKBgAAkIWyAQAAZHFcKW+8tra28MytW7cWnrl+/foBkZlLS0tLqZcA8F81NzcXntne3l545o4dOwrPnDdvXuGZERFz584tPPO6664rPDPHOjk6t912W+GZ9fX1hWd2dHQUnhkRsXHjxsIzr7766sIzS8WRDQAAIAtlAwAAyELZAAAAslA2AACALJQNAAAgC2UDAADIQtkAAACyUDYAAIAslA0AACALZQMAAMhC2QAAALJQNgAAgCyUDQAAIAtlAwAAyELZAAAAslA2AACALJQNAAAgC2UDAADIQtkAAACyUDYAAIAsjivljdfW1haeuWLFisIzW1paCs+cMmVK4ZkREdu2bcuSy+AyevTowjPnzp1beOYTTzxReGZERHt7e+GZzc3NhWfSf3V1dYVnbt++fUBkLlu2rPDMiDzzV1NTU3hmjt89HJ2qqqrCM2+88cbCM3O5+uqrC89cvXp14Zml4sgGAACQhbIBAABkoWwAAABZKBsAAEAWygYAAJCFsgEAAGShbAAAAFkoGwAAQBbKBgAAkIWyAQAAZKFsAAAAWSgbAABAFsoGAACQhbIBAABkoWwAAABZKBsAAEAWygYAAJCFsgEAAGShbAAAAFkoGwAAQBZDUkqp1IsAAAAGH0c2AACALJQNAAAgC2UDAADIQtkAAACyUDYAAIAslA0AACALZQMAAMhC2QAAALJQNgAAgCz+B0C1OQxUB+zyAAAAAElFTkSuQmCC", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "from sklearn import datasets\n", + "from sklearn.model_selection import train_test_split\n", + "import matplotlib.pyplot as plt\n", + "\n", + "digits = datasets.load_digits()\n", + "\n", + "_, axes = plt.subplots(nrows=1, ncols=4, figsize=(10, 3))\n", + "for ax, image, label in zip(axes, digits.images, digits.target):\n", + " ax.set_axis_off()\n", + " ax.imshow(image, cmap=plt.cm.gray_r, interpolation=\"nearest\")\n", + " ax.set_title(\"Sample: %i\" % label)\n", + " \n", + "n_samples = len(digits.images)\n", + "data = digits.images.reshape((n_samples, -1))/16\n", + "\n", + "X_train, X_test, y_train, y_test = train_test_split(\n", + " data, digits.target, test_size=0.5, shuffle=False\n", + ")\n" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "### Excercise 4\n", + "\n", + "Finally, we get to put together our network and train it.\n", + "\n", + "After running the below code snippet, experiment with different values for **learning rate** and **number of epochs**, different **number of neurons** (in the middle - hidden layer), using different **activation functions** (ReLU v LeakyReLU) as well as different **batch size**.\n", + "\n", + "For **learning rate** and **number of epochs**:\n", + "- try LR of 0.01, 0.001, 0.1\n", + "- try a mixed LR (e.g. 0.1 on the first two layers and 0.01 on the last layer)\n", + "- how does this influence the convergence speed (= **number of epochs** required to train the network)?\n", + "- how does this influence _the smootness of the loss graphs_?\n", + " \n", + "For different **number of neurons**:\n", + "- current configuration is `num_inputs x 100 x 200 x 10`. `num_inputs` and `10` can not change (they depend on the input, and the number of classes -- 10 digits)\n", + "- change the number of neurons in the hidden layer: (100, 200) , (100, 100), (200, 200), (200, 100)\n", + "- careful to consistently change the dimensions\n", + "- remove one of the hidden layers (and the ReLU)\n", + "\n", + "For **activation function**:\n", + "- Can you observe any difference in the network behaviour between using ReLU and Leaky ReLU?\n", + "- How does the behaviour of the network using Leaky ReLU change for different values of $\\alpha$?\n", + " \n", + "For **batch_size**:\n", + "- try values 16, 32, 64, 128, 898 (dataset size)\n", + "- How does this influence the convergence speed?\n", + "- how does this influence _the smootness of the loss graphs_?\n" + ] + }, + { + "cell_type": "code", + "execution_count": 32, + "metadata": {}, + "outputs": [], + "source": [ + "layers = []\n", + "layers.append(Dense(X_train.shape[1],100, learning_rate = 0.1))\n", + "layers.append(ReLU())\n", + "layers.append(Dense(100,200, learning_rate = 0.1))\n", + "layers.append(ReLU())\n", + "layers.append(Dense(200,10, learning_rate = 0.1))\n", + "\n", + "model = Network(layers)\n" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": { + "scrolled": true + }, + "outputs": [], + "source": [ + "def iterate_minibatches(inputs, targets, batchsize, shuffle=False):\n", + " assert len(inputs) == len(targets)\n", + " if shuffle:\n", + " indices = np.random.permutation(len(inputs))\n", + " for start_idx in range(0, len(inputs) - batchsize + 1, batchsize):\n", + " if shuffle:\n", + " excerpt = indices[start_idx:start_idx + batchsize]\n", + " else:\n", + " excerpt = slice(start_idx, start_idx + batchsize)\n", + " yield inputs[excerpt], targets[excerpt]\n", + "\n", + "train_log = []\n", + "val_log = []\n", + "\n", + "for epoch in range(100):\n", + "\n", + " for x_batch,y_batch in iterate_minibatches(X_train,y_train,batchsize=32,shuffle=True):\n", + " model.train(x_batch,y_batch)\n", + " \n", + " train_log.append(np.mean(model.predict(X_train)==y_train))\n", + " val_log.append(np.mean(model.predict(X_test)==y_test))\n", + " \n", + "# print(\"Epoch\",epoch)\n", + "# print(\"Train accuracy:\",train_log[-1])\n", + "# print(\"Val accuracy:\",val_log[-1])\n", + "plt.plot(train_log,label='train accuracy')\n", + "plt.plot(val_log,label='test accuracy')\n", + "plt.legend(loc='best')\n", + "plt.grid()\n", + "plt.show()\n", + " \n", + "\n" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "## Exercise 5\n", + "\n", + "The last thing to examine is the effect of weights and biases initialisation on the speed of convergence of this network. In order to do this, go back to [Exercise 3](#Exercise-3:-Dense-layer), and the implementation of the Dense Layer, and introduce the following options for initialising the weights and biases:\n", + "- introduce an additional (string) parameter in the __init__ constructor, called `winit`, to chose the way of initialising weights\n", + "- if `winit` is equal to `\"normal\"`, use the same initialisation which is currently implemented\n", + "- if `winit` is equal to `\"zeros\"`, initialise both weights and biases to 0\n", + "- if `winit` is equal to `\"ones\"`, initialise both weights and biases to 1\n", + "- if `winit` is equal to `\"zones\"`, initialise all the weights to 1 and all the biases to 0\n", + "\n", + "Then, come back to [Excercise 4](#Excercise-4), and repeat the training experiment. Which kind of initialisation works best? Comment on how well other initialisation options work." + ] + } + ], + "metadata": { + "kernelspec": { + "display_name": "Python 3 (ipykernel)", + "language": "python", + "name": "python3" + }, + "language_info": { + "codemirror_mode": { + "name": "ipython", + "version": 3 + }, + "file_extension": ".py", + "mimetype": "text/x-python", + "name": "python", + "nbconvert_exporter": "python", + "pygments_lexer": "ipython3", + "version": "3.12.6" + } + }, + "nbformat": 4, + "nbformat_minor": 4 +} diff --git a/CMP3751 Machine Learning/workshops/Week 8/Week8_workshop_solutions.ipynb b/CMP3751 Machine Learning/workshops/Week 8/Week8_workshop_solutions.ipynb new file mode 100644 index 0000000..ec858cb --- /dev/null +++ b/CMP3751 Machine Learning/workshops/Week 8/Week8_workshop_solutions.ipynb @@ -0,0 +1,1181 @@ +{ + "cells": [ + { + "cell_type": "markdown", + "metadata": { + "collapsed": true, + "jupyter": { + "outputs_hidden": true + } + }, + "source": [ + "## Regression Trees" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "collapsed": true, + "jupyter": { + "outputs_hidden": true + } + }, + "source": [ + "# Machine Learning: Week 8\n", + "## Decision trees and Random forest\n", + "\n", + "This week we will learn something about trees and forests for regression and classification. We will start with regression trees. Please download all files from blackboard before starting the notebook (the notebook itself + `regression_train.csv`, `regression_test.csv`). Also, execute each code cell in the correct order. \n", + "\n", + "Please read over the whole notebook. It **contains several solved exercises and four (4) that you need to add some code to**. If something does not run it might mean you need to alter or add a statement :) Also, **make sure you run all the code cells** (even the ones with examples) to ensure the rest of the notebook runs as intended.\n", + "\n", + "We'll cover the following topics:\n", + "- [Regression Trees](#Regression-Trees)\n", + " - [Training a Regression Tree](#Training-a-Regression-Tree)\n", + " - [Exercise 1a: Training the Tree](#Exercise-1a:-Training-the-Tree)\n", + " - [Evaluating the Tree](#Evaluating-the-Tree)\n", + " - [Exercise 1b: Finding the best fit](#Exercise-1b:-Finding-the-best-fit)\n", + "- [Decision Trees](#Decision-Trees)\n", + " - [Training a Decision Tree](#Training-a-Decision-Tree)\n", + " - [Exercise 2a: Decision boundary analysis](#Exercise-2a:-Decision-boundary-analysis)\n", + " - [Evaluating Decision Trees](#Evaluating-Decision-Trees)\n", + " - [Exercise 2b: Qualitative analysis](#Exercise-2b:-Qualitative-analysis)\n", + "- [Decision Forests](#Decision-Forests)\n", + " - [Exercise 3a: Qualitative analysis of `n_estimators`](#Exercise-3a:-Qualitative-analysis-of-n_estimators)\n", + " - [Exercise 3b: Quantitative analysis of `min_samples_leaf`](#Exercise-3b:-Quantitative-analysis-of-min_samples_leaf)\n", + " - [Exercise 3c: Quantitative analysis of `n_estimators`](#Exercise-3c:-Quantitative-analysis-of-n_estimators)" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "collapsed": true, + "jupyter": { + "outputs_hidden": true + } + }, + "source": [ + "We start with a simple regression task. We have to learn a continous function with a 1-dimensional input.\n" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "collapsed": true, + "jupyter": { + "outputs_hidden": true + } + }, + "source": [ + "We first load the training data with the pandas data frames and plot the training points and the \n", + "ground-truth function. The ground-truth values are stored in the test data set in order to evaluate the quality \n", + "of our fit. " + ] + }, + { + "cell_type": "code", + "execution_count": 1, + "metadata": {}, + "outputs": [], + "source": [ + "import pandas as pd\n", + "import matplotlib.pyplot\n", + "import numpy as np\n", + "\n", + "data_train = pd.read_csv('regression_train.csv')\n", + "data_test = pd.read_csv('regression_test.csv')" + ] + }, + { + "cell_type": "code", + "execution_count": 2, + "metadata": { + "scrolled": true + }, + "outputs": [ + { + "data": { + "text/plain": [ + "((150, 1), (150,))" + ] + }, + "execution_count": 2, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "x_train = data_train['x'].values\n", + "y_train = data_train['y'].values\n", + "\n", + "x_test = data_test['x'].values\n", + "y_test = data_test['y'].values\n", + "\n", + "x_train = x_train.reshape(-1, 1)\n", + "x_test = x_test.reshape(-1, 1)\n", + "\n", + "x_train.shape, y_train.shape" + ] + }, + { + "cell_type": "code", + "execution_count": 3, + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "## plot the data\n", + "import matplotlib.pyplot as plt\n", + "\n", + "plt.figure()\n", + "plt.clf()\n", + "plt.plot(x_train,y_train, 'bo')\n", + "plt.plot(x_test,y_test, 'g')\n", + "plt.legend(('training points', 'ground truth'))\n", + "plt.show()" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "collapsed": true, + "jupyter": { + "outputs_hidden": true + } + }, + "source": [ + "### Training a Regression Tree" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "collapsed": true, + "jupyter": { + "outputs_hidden": true + } + }, + "source": [ + "We will use the `sklearn` package to train our regression trees. `sklearn` is a generic machine learning library \n", + "that offers a lot of learning algorithms. A regression tree can be generated by:" + ] + }, + { + "cell_type": "code", + "execution_count": 4, + "metadata": {}, + "outputs": [], + "source": [ + "from sklearn import tree\n", + "\n", + "regTree = tree.DecisionTreeRegressor(min_samples_leaf=1, max_depth=None)" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "collapsed": true, + "jupyter": { + "outputs_hidden": true + } + }, + "source": [ + "We can set the minimum number of samples per leaf in the tree and the maximum depth of the tree as can be seen above. \n", + "\n", + "The tree can be trained by using the `.fit` method, similarly to most other models from `sklearn`:" + ] + }, + { + "cell_type": "code", + "execution_count": 5, + "metadata": {}, + "outputs": [], + "source": [ + "regTree = regTree.fit(x_train, y_train)" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "collapsed": true, + "jupyter": { + "outputs_hidden": true + } + }, + "source": [ + "We can use the trained tree for prediction calling the `predict` method (again, similar for most `sklearn` models):" + ] + }, + { + "cell_type": "code", + "execution_count": 6, + "metadata": {}, + "outputs": [], + "source": [ + "y_predict = regTree.predict(x_test)" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "collapsed": true, + "jupyter": { + "outputs_hidden": true + } + }, + "source": [ + "#### Exercise 1a: Training the Tree\n", + "\n", + "In this excercise you are supposed to train the tree for our regression task.\n", + "- Train a tree with `min_samples_leaf` set to 1, 5 and 10 and predict the output for `x_test` and plot the predicted\n", + "function values.\n", + "- Do you see a difference in the functions?\n", + "- Based on this qualitative analysis (visual inspection), which value of `min_samples_leaf` would you use? " + ] + }, + { + "cell_type": "code", + "execution_count": 7, + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "from sklearn import tree\n", + "\n", + "######################################\n", + "##### YOUR CODE GOES HERE: ###########\n", + "######################################\n", + "# your code here: tree with min_samples_leaf = 1:\n", + "# initialise model:\n", + "# regTree1 = \n", + "# fit the model:\n", + "# \n", + "# predict the values:\n", + "# y_predict1 = \n", + "\n", + "regTree1 = tree.DecisionTreeRegressor(min_samples_leaf=1)\n", + "regTree1 = regTree1.fit(x_train, y_train)\n", + "y_predict1 = regTree1.predict(x_test)\n", + "\n", + "regTree2 = tree.DecisionTreeRegressor(min_samples_leaf=5)\n", + "regTree2 = regTree2.fit(x_train, y_train)\n", + "y_predict2 = regTree2.predict(x_test)\n", + "\n", + "regTree3 = tree.DecisionTreeRegressor(min_samples_leaf=10)\n", + "regTree3 = regTree3.fit(x_train, y_train)\n", + "y_predict3 = regTree3.predict(x_test)\n", + "\n", + "fig, axs = plt.subplots(3,1,figsize=(7, 18))\n", + "\n", + "# plot for min_samples = 1:\n", + "axs[0].plot(x_train,y_train, 'bo')\n", + "axs[0].plot(x_test,y_test, 'g')\n", + "# plot the predicted values:\n", + "axs[0].plot(x_test, y_predict1, 'r')\n", + "axs[0].set_title('min_samples = 1')\n", + "\n", + "# plot for min_samples = 5:\n", + "axs[1].plot(x_train,y_train, 'bo')\n", + "axs[1].plot(x_test,y_test, 'g')\n", + "# plot the predicted values:\n", + "axs[1].plot(x_test, y_predict2, 'r')\n", + "axs[1].set_title('min_samples = 2')\n", + "\n", + "\n", + "# plot for min_samples = 10:\n", + "axs[2].plot(x_train,y_train, 'bo')\n", + "axs[2].plot(x_test,y_test, 'g')\n", + "# plot the predicted values:\n", + "axs[2].plot(x_test, y_predict3, 'r')\n", + "axs[2].set_title('min_samples = 3')\n", + "\n", + "\n", + "plt.savefig('regressiontrees.png')\n", + "plt.show()" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "collapsed": true, + "jupyter": { + "outputs_hidden": true + } + }, + "source": [ + "### Evaluating the Trees\n", + "\n", + "We can also use `sklearn` to evaluate the tree. There are different metrics that we can use. We will use the \n", + "mean squared error criterion to evaluate the trees. We can compute the MSE on the train data" + ] + }, + { + "cell_type": "code", + "execution_count": 8, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "(0.0, 0.07406137873038969, 0.1486237118341035)" + ] + }, + "execution_count": 8, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "import sklearn.metrics as metrics\n", + "\n", + "y_predict_train1 = regTree1.predict(x_train)\n", + "mseTrainTree1 = metrics.mean_squared_error(y_train, y_predict_train1)\n", + "mseTrainTree2 = metrics.mean_squared_error(y_train, regTree2.predict(x_train))\n", + "mseTrainTree3 = metrics.mean_squared_error(y_train, regTree3.predict(x_train))\n", + "\n", + "mseTrainTree1, mseTrainTree2, mseTrainTree3" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "collapsed": true, + "jupyter": { + "outputs_hidden": true + } + }, + "source": [ + "And we can compute the MSE on the test data" + ] + }, + { + "cell_type": "code", + "execution_count": 9, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "(0.03965319971792027, 0.08128469781895029, 0.16326245803469855)" + ] + }, + "execution_count": 9, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "import sklearn.metrics as metrics\n", + "\n", + "mseTestTree1 = metrics.mean_squared_error(y_test, y_predict1)\n", + "mseTestTree2 = metrics.mean_squared_error(y_test, regTree2.predict(x_test))\n", + "mseTestTree3 = metrics.mean_squared_error(y_test, regTree3.predict(x_test))\n", + "\n", + "mseTestTree1, mseTestTree2, mseTestTree3" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "collapsed": true, + "jupyter": { + "outputs_hidden": true + } + }, + "source": [ + "#### Exercise 1b: Finding the best fit\n", + "We can see that trees with `min_samples_leaf = 1` have MSE 0 on the training set as the training set is learned by heart. However, the error on the test set is the one that really counts. Here, also `min_samples_leaf = 1` performs the best, but **can you find better settings of `min_samples_leaf`**?\n", + "- Go back to [Exercise 1a](#Exercise-1a:-Training-the-Tree) and try different values of `min_samples_leaf`. Then use the code in [Evaluating the trees](#Evaluating-the-Trees) to evaluate these on the training and the test set.\n", + "- Can you find a value `min_samples_leaf` which results in a smaller **MSE on the test set** than `min_samples_leaf=1`?" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "collapsed": true, + "jupyter": { + "outputs_hidden": true + } + }, + "source": [ + "## Decision Trees\n", + "In contrast to regression trees, the output for a decision tree is a discrete class label, not \n", + "a continuous value. A decision tree can be used for any multi-label classification problem. We will use the \n", + "decision trees on the iris data set. We do **not** use the same features as in lecture examples.\n", + "Instead, we will only use the first two features `sepal_length` and `sepal_width` in the training data as input\n", + "for illustration purposes. " + ] + }, + { + "cell_type": "code", + "execution_count": 10, + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjcAAAGwCAYAAABVdURTAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy81sbWrAAAACXBIWXMAAA9hAAAPYQGoP6dpAACJWUlEQVR4nO3dd3hUZfbA8e+dmfRKGiGht9A70kRAUJoI6toWO7q64s+2uq69IrqrrmUVu9gQK4IFqQrSe++9BUJIbzOZuff3x02GDMm0kMkkk/N5nnkgM2/eeyaTZE7ufd9zFE3TNIQQQgghAoTB3wEIIYQQQtQkSW6EEEIIEVAkuRFCCCFEQJHkRgghhBABRZIbIYQQQgQUSW6EEEIIEVAkuRFCCCFEQDH5O4DapqoqJ06cICoqCkVR/B2OEEIIITygaRr5+fmkpKRgMLg+N9PgkpsTJ07QrFkzf4chhBBCiGo4evQoTZs2dTmmwSU3UVFRgP7FiY6O9nM0QgghhPBEXl4ezZo1s7+Pu9LgkpvyS1HR0dGS3AghhBD1jCdLSmRBsRBCCCECiiQ3QgghhAgoktwIIYQQIqBIciOEEEKIgCLJjRBCCCECiiQ3QgghhAgoktwIIYQQIqBIciOEEEKIgCLJjRBCCCECiiQ3QgghhAgoktwIIYQQIqBIciOEEEKIgCLJjRBCCCECiiQ3QgghhAgoktwIIYQQIqBIciOEEEKIgCLJjRBCCCECiiQ3QgghhAgoktwIIYQQIqBIciOEEEKIgCLJjRBCCCECiiQ3QgghhAgoktwIIYQQIqBIciOEEEKIgCLJjRBCCCECiiQ3QgghhAgoktwIIYQQIqBIciOEEEKIgCLJjRBCCCECiiQ3QgghhAgoktwIIYQQIqBIciOEEEKIgCLJjRBCCCECiiQ3QgghhAgoktwIIYQQIqBIciOEEEKIgCLJjRBCCCECSp1Jbl566SUUReH+++93Omb69OkoiuJwCw0Nrb0ghRBCCFHnmfwdAMDatWt577336Natm9ux0dHR7N692/6xoii+DE0IIYQQ9Yzfz9wUFBQwceJEPvjgAxo1auR2vKIoJCcn22+NGzeuhSiFEEIIUV/4PbmZPHkyY8eOZcSIER6NLygooEWLFjRr1ozx48ezfft2l+PNZjN5eXkONyGEEEIELr8mNzNnzmTDhg1MnTrVo/FpaWl8/PHHzJ49my+++AJVVRk4cCDHjh1z+jlTp04lJibGfmvWrFlNhS+EEEKIOkjRNE3zx4GPHj1Knz59WLBggX2tzdChQ+nRowevv/66R3OUlpbSsWNHrr/+ep5//vkqx5jNZsxms/3jvLw8mjVrRm5uLtHR0ef9PIQQQgjhe3l5ecTExHj0/u23BcXr168nIyODXr162e+z2WwsXbqU//3vf5jNZoxGo8s5goKC6NmzJ/v27XM6JiQkhJCQkBqLWwghhBB1m9+Sm+HDh7N161aH+2699VY6dOjAI4884jaxAT0Z2rp1K2PGjPFVmEIIIYSoZ/yW3ERFRdGlSxeH+yIiIoiPj7fff9NNN5Gammpfk/Pcc8/Rv39/2rZtS05ODv/5z384fPgwt99+e63HL4QQQoi6qU7UuXHmyJEjGAxn1zxnZ2dzxx13cPLkSRo1akTv3r1ZsWIFnTp18mOUQgQuW1YW5j+XQWkpQd26EtS+vb9DEkIIt/y2oNhfvFmQJERDpZnN5Dz9DEVfzQSr1X5/8AV9afTG65iaN/djdEKIhsib92+/17kRQtQtmqaRddffKfpyhkNiA2BZv4HT4ydgy8jwU3RCCOGeJDdCCAeW1aspmb8AVLXygzYb6pksCj78qPYDE0IID0lyI4RwUPTd9+Bqt6LNpl+uEkKIOkqSGyGEA9vp02CzuRyjZmfTwJbrCSHqEUluhBAOjMlNwOS6zpQhIR5FUWopIiGE8I4kN0IIB+FX/wWsLs7cGI1E/PWvtReQEEJ4SZIbIYSD4N69CBs/Hqo6M2M0YkxOJuL2SbUfmBBCeEiSGyGEA0VRaPTGf4n8+10ooaEVHyBkyEUkzp6FMS7OfwEKIYQbUsRPCOGUWlCAZfUaNIuZoC5dMDVr5u+QhBANVL3oCi6EqPsMkZGEDr/Y32EIIYRX5LKUEEIIIQKKJDdCCCGECCiS3AghhBAioEhyI4QQQoiAIsmNEEIIIQKKJDdCCCGECCiS3AghhBAioEhyI4QQQoiAIsmNEEIIIQKKJDdCCCGECCjSfkGIAGA9dIiib77Feuw4hrhGhF91JcFdu/o7LCGE8AtJboSoxzRNI++llyl4+x0wlJ2IVaDwgw8JHTuWuLfeQAkJ8W+QQghRy+SylBD1WOGHH1Hwv7dB08Bm029WGwAlc+eS88STfo5QCCFqnyQ3QtRTWmkp+W/9z/kAVaXo66+xnTpVe0EJIUQdIMmNEPWUZfMW1DNnXA+yqZQs/r12AhJCiDpCkhsh6imtuNj9IEVBKynxfTBCCFGHSHIjRD0V1K7t2UXEzmgaQR3SaicgIYSoIyS5EaKeMiYnE3rpJWA0OhlgxNiqFcH9+9duYEII4WeS3AhRj8W+8DyGpKTKCY7RiBIaStzbb6Eoin+CE0IIP5HkRoh6zNikCUlzfyHitltRIiL0O4OCCLvyChJ//YXg7t39G6AQQviBomma5u8galNeXh4xMTHk5uYSHR3t73CEqDGazYZWUIASHo4SFOTvcIQQokZ58/4tFYqFOE9qYSHFP8yi+Ndf0YqKCOrUiYgbbySoU8dajUMxGlFiYmr1mEIIURdJciPEebAeOMjpq69BPXkSFAU0DcvGTRR+9jlRD/2D6Afu93eIQgjR4MiaGyGqSbNayZw4EfX06bI7yq7w2vT2B/mvvErxTz/7KTohhGi4JLkRoppKFizAduSoPZmpxGAg/51ptRuUEEIISW6EqC7zkqVgcnFlV1Up3bIFNT+/9oISQgghyY0Q1aWpqmcDrVbfBiKEEMKBJDdCVFNwz56uExdFwdisGUpsbK3FJIQQQpIbIaotbMJ4lOhol/2dIu+4XSoECyFELZPkRohqMoSFEf/RhyjBwY7tD8qSndCxY4m45WY/RSeEEA2X1LkR4jyEDBxA0sL5FHz8CcU//YxWUkxQ+zQibr2ZsPHjUdx17RZCCFHjpP2CEEIIIeo8ab8ghKhV5rVrKfjoYyyr14DRSOjFw4i87VaCOnTwd2hCiAaozpwzf+mll1AUhfvvv9/luG+//ZYOHToQGhpK165d+fXXX2snQCFElfLfmUbmhCsp+XUuakYGano6RTO/JuPSURTNnuPv8IQQDVCdSG7Wrl3Le++9R7du3VyOW7FiBddffz2TJk1i48aNTJgwgQkTJrBt27ZailQIUZF51Srypryof1CxUrPNBjYb2f93L9YjR/wTnBCiwfJ7clNQUMDEiRP54IMPaNSokcuxb7zxBqNGjeLhhx+mY8eOPP/88/Tq1Yv//e9/tRStEKKigg8/ctwpVoXCL76spWiEEELn9+Rm8uTJjB07lhEjRrgdu3LlykrjRo4cycqVK51+jtlsJi8vz+EmhKgZllWrnffWArDZMK9cVXsBCSEEfl5QPHPmTDZs2MDatWs9Gn/y5EkaN27scF/jxo05efKk08+ZOnUqzz777HnFKYRwwoOt7rIdXghR2/z2W+fo0aPcd999fPnll4SGhvrsOI8++ii5ubn229GjR312LCEampAhF7m+LGUw6GOEEKIW+S25Wb9+PRkZGfTq1QuTyYTJZGLJkiW8+eabmEwmbFWc6k5OTubUqVMO9506dYrk5GSnxwkJCSE6OtrhJoSoGZG3TwJnDUQVBYKDifjr9bUblBCiwfNbcjN8+HC2bt3Kpk2b7Lc+ffowceJENm3ahLGKvwYHDBjAokWLHO5bsGABAwYMqK2whRAVBHfvTuwr/9YTmXNbUAQHE//xhxhd/PEhhBC+4Lc1N1FRUXTp0sXhvoiICOLj4+3333TTTaSmpjJ16lQA7rvvPoYMGcKrr77K2LFjmTlzJuvWreP999+v9fiFELqI664juE8fCj/7HPPKVSgmI6FDhxJx440YU5r4OzwhRANUpysUHzlyBEOFxYgDBw5kxowZPPHEEzz22GO0a9eOH3/8sVKSJEQgUUtKIDjY4Wehrglq25bY52ThvhCibpDeUkLUQWpREdkPPEjJvHlQagXAkJRE9EMPETFR1rAIIRoe6S0lRD2mFhVxst8AtKwsx/szMsj55z8p3bOb2Gef8U9wQghRD9Td89xCNFDZ99xbKbGpqPDDjyg9cLAWIxJCiPpFkhsh6hBVVSlZuNDtuNxnn6uFaIQQon6S5EaIOkTLyXHdzqCMdfdu3wcjhBD1lCQ3QtQhiqfVuoOCfBuIEELUY5LcCFGHGMLDUWJj3Y4LGz3K98EIIUQ9JcmNEHVM1L3/53pAUBBRDz5QO8EIIUQ9JMmNEHVM1J1/I+yaq6t+0GQi4duvMfiw2awQQtR3ktwIUQfF/fc1En/7leBBAzEkJ2Ns1ozIv91Bk53bCenb19/hCSFEnSZF/IRwwbxxE6W7dhHcqyfBaWm1euzgrl1J/ObrWj3muTRNw3biBFgsGFNTUYKD/RqPEIHEarVSWFiI0WgkIiICRVFqZN7i4mLMZjPh4eEEN9CfWUluhKhC7htvUvDf16G09OydoaHEvPA8kddf57e4alPRd9+T/+ZbWPfvB0CJiSbixhuJuv8+DGFhfo5OiPrLYrGwfv16du3aRWnZ75i4uDh69epF69atqz1veno669atIz09HQCDwUCbNm3o27cvkZGRNRJ7fSG9pYQ4R85zz1P4nvNO89HPPUvUpNtqMaLal/ff18l/5VVQFKj4K8JgILhXTxK+nun5tnUhhF1paSlz5swhKyuLqt5++/fvT7du3bye9/Dhw8yfPx/AYV5FUQgNDeWKK66o9wmON+/fsuZGiApsNpvLxAYgL8CrA5fu268nNuCY2ACoKpb1Gyj87PPaD0yIALBlyxaniQ3A6tWrKSgo8GpOm83GH3/8gaZplebVNI2SkhJWrVpV7ZjrI0luhKgg/5ln3Q+y2Sj4aqbvg/GTohkzwGh0OaZg+qe1FI0QgUPTNHbs2OE0sSm328sK5IcPH8ZsNrs87sGDBykpKfFq3vpMkhshKrBs3uzROPPSpT6OxH9KDxxw3QJC07AdOeL2F7QQwpHVaqW4uNjtuNzcXK/mzcnJcbsYWdM08vLyvJq3PpPkRogKDFGercMyJiT4OBL/MURGuj1zo4SH19jODiEaCqPR6NHPTZCX7VWCg4M9+mOjIe2ckuRGiAqin3rCo3FRj/zTx5H4T9hlY12fuTEaCbt8XO0FJESAMBgMtGzZ0mWCo2ma1zum3M0JEBsbS0xMjFfz1meS3AhRQXBaGobUFJdjTF27Yqznuw5cCR0xAlOnTlWfvTEYIMhE5F131n5gQgSAnj17Ok1EFEWhcePGpKS4/h10rsjISNLc1OHq06dPgzrbKsmNEOdIWvYnBieXnYytWpLwy0+1HFHtUkwmEr76kuAePfQ7TCb9BhgaNSJhxpcEtW3rvwCFqMcSEhIYOXKk/RKRwWCwJx1NmjRh1KhR1UpCBg0aRPv27QE9STIY9Ld3o9HI4MGDz6t+Tn0kdW6EcKJo3nzypkxBy83DkJBA7EsvNqjWB5qmUbphIyWLFqFZLAR17UrY6FFSpViIGmC1Wjl48CBnzpzBZDLRokULEhMTz3ve3NxcDhw4gNlsJjo6mjZt2hASElIDEfufN+/fktwI4YSmaZRu346amYmxSROCaqj9gpqfT+nmLWiaRnC3rhhcXAe3padTumcPSmgowT17SmIhhGiwvHn/lvYLQlShZOEicp97Duv+A/b7grp0Ieb5Zwm54IJqzakVF5P74lQKv5wB5TUpgoMJv/YaYp56EkN4uH2s7UQ6OU88Qcn8BfZCekpsLFH3TCbyrjsb1LVzIYTwlpy5EeIcxXPnknVH2YLZc1oPYDCQ8PVXhPTv79WcmtVK5l9vwLJyJaiq44MGA8G9e5HwzdcowcHYMjPJGDUGNSOjyl1LEXfcTuwzT3v7tIQQol6T9gtCVJNmtZLz6ONlH1RuPYCqkvPEU14XsCuZ+xuW5csrJzZl81rWrqN4jr5QueDtd5wmNgCFH3xI6b79Xh1fCCEaEkluhKjA/OefqKdPV05syqkq1p07se7Y6dW8hTNm6Gd+nDEYKPxyBpqqUjjjK7d1Zoq++car4wshREPi1ZqbnTt3MnPmTP78808OHz5MUVERiYmJ9OzZk5EjR3LVVVcFzKps0TDZTqR7NM56/DhBnTt5PK/16NGqz9qUU1Wsx46iFRejuWuap2nYjh/3+NhCCNHQeHTmZsOGDYwYMYKePXuybNky+vXrx/3338/zzz/PDTfcgKZpPP7446SkpPDyyy+7bOAlRF1miI/zaJy37ReMiYmuz9woCsaEBJTQUHD3B4LBgCE+3qvjCyFEQ+LRmZurrrqKhx9+mO+++47Y2Fin41auXMkbb7zBq6++ymOPPVZTMQpRa0KHDkWJikLLz3c6xti8OUE9uns1b/jVV2NZs9b1mGuvQTEaCb/yCoq++db5pSmrlfCrrvTq+EII0ZB4tFuqtLTUq0Ze3o6vTbJbSrhTMP1Tch933mMq7v33CBs7xqs5teJiMsZchnX//spJi9GIsXlzkubNxRARgfXQITJGjkYrLq481mAgdPQo4t9/z6vjCyFEfVfju6W8TVTqamIjhCcib7mZmBeeRynvH1VWU8bQqBGN3n7L68QGQAkLI+G7bwgZfGGlx4L79yfx+28xREQAYGrZkoTvv8PUprXD8TEaCb/+OuLeetP7JyWEEA1ItercrF27lt9//52MjAzUcxZJvvbaazUWnC/ImRvhKbW4GPPCRdjOZGJMSdEvWdVAheDSffuxrFoFmkZwvwsIKusHcy5N07CsW0fpjp0ooSGEDhuGMSnpvI8vhBD1kU/bL7z44os88cQTpKWl0bhxY4dKqYqisHjx4upFXUskual5msVCyR9LUE+exJCYQOiwYfrC2Fpk2bad0s2bwWQk5MLBmJx09tZUFfPyFdgOHkSJjiL04osxyPeBEHVaZmYmp0+fxmAwkJqaSmT5WVXRoPi0/cIbb7zBxx9/zC233FLd+EQAKZo9m9wnnkLNytIvn2gaSnQ0MY89SsSNN/j8+NbDh8m65/8o3bDx7J2KQtj4y4n998v2Sz0A5mXLyf7HP7Adq7CNOiSEqMl3E/XA/SiudjMJIWpdXl4eixcvJiMjw+H+Nm3acNFFF8kSCOGU18mNwWBg0KBBvohF1DPFP/9C9t33nL2j7CSglpdHzr8eBUUh4oaJPju+7fRpTk+4AvVMluMDmkbxnJ+wnTqltzQwGDCvW0/mxBtAPWeBrtlM/mv/RbNYiHn0Xz6LVQjhnaKiImbPnk1JSUmlxw4cOEBRURGXXXaZ9FkTVfL6T9UHHniAt99+2xexiHpEU1Vyn3/e5ZjcF6ei+bDmUcFHH6Nmnql6y7SqYlm5CvMfSwDIe/nlsvYJVV+FLZj2LrZz/joUQvjPtm3bKCkpqbLViaZppKenc/ToUT9EJuoDr8/cPPTQQ4wdO5Y2bdrQqVOnSqcFf/jhhxoLTtRdlvUbHC/vVEHLzaVkyVLCLr3EJzEUffOt66q/RiNF331PUKeOWFasdD2ZplH8089ETrqtZoMUQlTLnj17XPZwUxSFvXv30rx581qMStQXXic39957L7///jvDhg0jPj5eTgk2UGrWGc/GnfFsXLViyM52PcBmw3Y6AzXLzTgAgwE1M7NmAhNCnLeqLkdVpGkaRUVFtRSNqG+8Tm4+/fRTvv/+e8aOHeuLeEQ9YUypejdS5XFNfBdDcjK2I0dcDDBiatoUQ+Mk+2Jnp2w2j5+TEML3IiIiyHdRKVxRFKKiomoxIlGfeL3mJi4ujjZt2vgiFlGPBHXpgql9u7MF5s6lKBgaNybkwspF62pKxMS/uu7XZLMRfv11GOPjCb3kEjAanY8NCiLs8nE1H6QQolo6duzo8nFN0+jQoUMtRSPqG6+Tm2eeeYann35aTgc2cIqiEDtlip5cnJtglCU8sS++gOIqoThPETffhKl166qTFkUh7PLLCe7bF4Dox/6FEhbqNMGJefwxDDExPotVCOGdTp06ERsb63TpQ5s2bWjcuHEtRyXqC6+L+PXs2ZP9+/ejaRotW7astKB4w4YNNRpgTZMifjXLvHIlOU8+jXXnTvt9pjatiXnqKUJHDPf58W1ZWeQ++RTFP/0ENn1xsRIRTsRttxH90D9QTGevvJbu3EnO409gWb3Gfp8hOZnofz5ExLXX+jxWIYR3SkpKWL58OQcOHLAvLjaZTHTp0oU+ffpgkNpUDYpPKxQ/++yzLh9/+umnvZmu1klyU/M0TcO6cxe29HQMiQkEde1a6wvNbadPU7p9O4opiKBePTGEhzsdaz1wEOuhQyhRUQT36unTs0tCiPNXVFTEmTNnMBgMJCUlSfG+BsqnyU19J8lNw2Y9eYq8qVOx7tuHITqayLvuInTIRVWOtWzeTN5r/0XNzMSYmkr0I/8kyMl6s9IdOymeNw+tqIigtDTCxo5BCQvz5VNxoGZnUzR7NrZjxzHExRE2/nJMqam1dnwhhPA1nyY3a9euRVVV+vXr53D/6tWrMRqN9OnTx/uIa5EkNw1X9sOPUDRjRqX7Dc2a0nj+PHuPKdVqJfPy8ZRu3lJpbMiwoSR88bn9YzU/n6y778G8eLG+nsdggNJSlOho4t56s1YuzRV8+BG5L0wBqxVMJr2ooaYRcestxDzztJyZEkIEBG/ev72+YDl58uQqq0IeP36cyZMnezXXtGnT6NatG9HR0URHRzNgwADmzp3rdPz06dNRFMXhFlrLDRpF/ZT3yqtVJjYA6tFjZIwabf/4zLXXV5nYAJh//4OsyXrLCU3TyLrjTsxL9CrI2GxQWqo/lp/PmdsmYanY88oHCr/5ltynn9GPq2n6v6oKmkbhJ5+QN/Ulnx5fCCHqIq+Tmx07dtCrV69K9/fs2ZMdO3Z4NVfTpk156aWXWL9+PevWrePiiy9m/PjxbN++3ennREdHk56ebr8dPnzY26cgGqD8ae+6fNx2+AglK1ZgTU/HsmqVy7HFs+eglpRgWb8B859/Vt3+oeyEaP6bb1Y7Znc0m438/7ziYoB+VsfmSRFDIYQIIF4X8QsJCeHUqVO0bt3a4f709HRMJu+mGzfOsa7IlClTmDZtGqtWraJz585Vfo6iKCQnJ3t8DLPZjLlCf6O8vDyvYhT1n3ntWnBT7RSg4N33MLVq5X5CTaPoyxlYjx7VLwNZrVWPs9koWbgIrbjYJ+tvSrduxXbihJtBpZQsXEjENVfX+PGFEKKu8vrMzaWXXsqjjz5Kbm6u/b6cnBwee+wxLrmk+j2EbDYbM2fOpLCwkAEDBjgdV1BQQIsWLWjWrJnbszwAU6dOJSYmxn5r1qxZtWMU9ZN6yrOGmFp+vvuWDmVsWVloLqqnnp1UQy0u9mhOb6l5HhxfUdAKCnxyfCGEqKu8Tm5eeeUVjh49SosWLRg2bBjDhg2jVatWnDx5kldffdXrALZu3UpkZCQhISHcddddzJo1i06dOlU5Ni0tjY8//pjZs2fzxRdfoKoqAwcO5NixY07nL0/Eym/SRbbhCerezaNxpvbtCe7m2djgvn31AoKuGncCSkyMz4oDmlp7dpbJo3FCCBFAqrUVvLCwkC+//JLNmzcTFhZGt27duP7666tVe8BisXDkyBFyc3P57rvv+PDDD1myZInTBKei0tJSOnbsyPXXX8/zzz/v0fFkt1TDlD5gEKqrPlRA8tbNKLGxpLdq4/xSE6BERpKyeye206c52btv1WtuAIxGIu/8GzGPP3Y+obuUed1fMa9YUXUMBgOGxo1JXr1SdkwJIeq9el3nZsSIEbRp04b33nvPo/FXX301JpOJr776yqPxktw0TJbNmzl92eVOz7RE3nE7Mc/oBSgLpn9K7uNPOJ0r7sP3CRs9umzsdHIffxIMCqgVfpSMRkytWpE450eftnUo3X+A0+Mu1y89VUxwDAYwGon//DNCB/uuv5cQQtSWGt8KvsrN7pGKioqK3K6DcUVVVYcFwK7YbDa2bt1Kkya+6zwtAkNw9+4k/vozxpYtHe5XIiOJeuxRe2IDEHnLzcS+/l+UuDiHsYbkZOI+/8ye2OhjbyHuw/cxdTjb5E8JCyPiphtJnD3L5/2qgtq0JumXnwi7bKxD36yQQYNI/OF7SWyEEA2SR2du2rVrR+vWrbn99tsZM2YMERERlcbs2LGDL774gk8++YSXX36Zm266ye3BH330UUaPHk3z5s3Jz89nxowZvPzyy8ybN49LLrmEm266idTUVKZOnQrAc889R//+/Wnbti05OTn85z//4ccff2T9+vUeXcYCOXMj9CrFpdu3Y2zcmOAuVe/KK1e6fz/WQ4cJatsGU4sWruc9fgKtuBhjagqGWqxOXE7NzcWWkYGhUSOMCQm1fnwhhPAlb96/Pdq7vWPHDqZNm8YTTzzBX//6V9q3b09KSgqhoaFkZ2eza9cuCgoKuOKKK5g/fz5du3b1KNCMjAxuuukm0tPTiYmJoVu3bvbEBuDIkSMOjdGys7O54447OHnyJI0aNaJ3796sWLHC48RG+Ebp9h0Ufv0NtpMnMSYmEH7VVQT17HFe/aXU3FxyX5yKeclSsNkwde5M7DNPYTrnzIu3SpYtJ+/f/8Z2/ARKeDjh115D5F13YjinjIGmaZRu2EjR999jyzyDsUkyEdddS1DHjpXm1Gw2ShYupGTub2jFJZg6pBFx3bUYz/OMomXbdvJeeAHrvv0QHEzYmNFEPfQPDE4KVxp8uHjZb06sh02fQkE6RDaBHrdASuU6W96wWq3s37+f48ePo6oqycnJtGvXjpCQkEpji4uL2b17N5mZmRgMBlq0aEHLli0xyhomIeo0r9fcrFu3jmXLlnH48GGKi4tJSEigZ8+eDBs2jLhzTuPXRXLmpuZoNhs5jz5G0ZczwGTU17MYjGC1Ejp2LHH/exMlONjreYt/m0fWHX+rcn1M5H33EvPPh6sV7+lrrsWyfEWl+5XISJIWzsdUViZAM5vJmnwPJXN/0+vYqDZQDGCzEX7DRGKnvohSlnTbTp4k868Tse7eo18WKv9xUhRip7xAxI03VCvWrH88RPHMrys/EBREwvffEtK7d7XmrTdspfDjrbD1SzCUvQYGI6hW6HYjjP8YjF6X6SIzM5Nff/2VkpISe/KtaRomk4lLL72Upk2b2sfu37+f33//nYq/IjVNIzIykssuu0x+fwhRy+r1gmJfk+Sm5uS9+hr5r/236gcVhfCbbqTRi1O8mtN68hSn+vQ9myRUIe799wgbO8arebMf+idFLhadK40akbJNb7mQ/a9H9YStqsXHikLUPx4k+oH70VSVjEtHYd2zx+mOqfgvPiN02DCvYs3/4EPynnnW+YCgIJrs2uH0DE5AmPcQrHwNqOr7QIFBD8MlL3s1ZUlJCV9//TUWi4Wqfu0ZjUb+8pe/EBMTQ0ZGBrNnz65ynKIoREREcO2118oZHCFqkU97SwkBoBYVUfDe+84HaBpFM2Zgy8ryat685593mdgA5HrZL0lVVYq+/dblGC07m+Kff8GWmUnRjK+c16/RNAreex+tuBjz0qVYd+50vhXcYCD/f297FStA/htuWjaUlpL/+utez1tvFOfAmrepOrFBv3/1W1DiXbXxPXv2YDabq0xYQP8+Kd8MsXnzZqfzaJpGQUEBBw8e9Or4QojaI8mNqBbL2rVohYWuB5VaMS9d6tW8JUv/dDvGduiQV3Na/lzmsm5NucIvv7Sv8XFFy8/HvG49JQsW6petnFFVLKtWo7r7OlX8lKIiNA+qJBf/4rzBbL136HewuWmXYS2GQ394N62b7xtN0zh48CCapnH48GGnSRDoZ2+kr50QdZckN6JaNLPFw3Gebeu38yAJcXdmp9LwIs+SC81sRrN49rzwYqynXysAzYMeWIC9+3hAsnr4PWP18GtVPtyD7y1bWWKruqk8rWmafawQou6R5EZUS1DHDuDBbqggJw1QnfFkN5QSGenVnEF9+3o0LrhXL4I6e7DzTlEwdeyoj3XzBmdISsIQ6/kOJiU21qFejTOmtPYez1nvJHev2XFlEhMTXe7gUxSFhIQEFEWhUaNGLudSFIX4+Hivji+EqD2S3IhqMTVrRsjQIc7fiI1Ggrp3I7hLF6/mjXrkn27HhF/zF6/mNCUkYOrYwfUgRSHqwQcI7taNoC5dXD6v0OHDMaWmEH7llSihIc6TPIOByFtvse+s8oTBYCB0+HC342KeetLjOeudxI7QfDAoTl4DxQQth0JCmlfTdurUyeWlJk3T6FyWjHfx4Pu2Qwc331NCCL+p1m6pRYsWsWjRIjIyMiqdvv34449rLDhfkN1SNcd6/ASnLx+Pevq04xkMoxElOorEH2cR1Lat1/Oe+dudlPzya5WPGVo0p/HSJZXq0riP9TinBg8BJ5fJYp57lshJtwFQumcPp6+4Ei2/oNLzMiQlkTRnNsYUvYZN8dy5ZN35d/3ximMVheABA0j44jOUKuqnuKIWFXHygv5O195E3HYLsR72Uqu3svbDRwOh6AxoFb+uRghPgEkrIK6119Nu2rSJNWvWoChKpUSnY8eOXHjhhSiKgqqqLFy4sNI6nfLPGzJkCGlp3iVXQojz49PdUs8++yyXXnopixYtIjMzk+zsbIebaDhMqSkkzZtL5N/uQCkrHqdERBBx800kzZtXrcQGIP7994j61yP6JZpyISGEXXN1tRIbPdZUGq9YTvCggXrfpTKGFi2I++hDe2IDENS+PUnzfiPiphtRyqpxKzExRN75N5J++9We2ACEjR5N4pwfCR010n62x9isGdFPPlGtxAbAEB5O8ppVhI4Z7bBg2ZCUSMzLLwV+YgMQ1wbu3AD9/g9Cyn6JhURDv3vhro3VSmwAevTowciRI0lOTrbfl5CQwLBhw+yJDehn0EaMGMGFF15IbNn3oaIoNGvWjMsuu0wSGyHqOK/P3DRp0oR///vf3Hjjjb6KyafkzI1vaJoGFgsEB59XZeJzqRYLWK0YwsNrbE4ANS8PIiMdKmBXxZvnpakqWK3VKlzoilpQAKGh1UrqAoKmgc0CxmCP1nl5qvyss7vvgfKxiqLU6Pe2EMI7Nd5+oSKLxcLAgQOrHZwITIqiQDXOUrhjCA6GGkoWNE3DsnwFhV9+SenefRhiYwifMIGwK6+olDxpFgvFv/5K0TffYcvIwNS0KeF/vY7Q4cNRnKzHUQyGGou1IoOXC6gDjqKAyQffWx4kNWd2r2DH2t/JKASjotEyMZK0i/5CWHxTt59b206cOMGqVavsZ9Dj4+MZMGAAjRs39nNkQtQ+r8/cPPLII0RGRvLkk/VzQaOcuWmYNFUl+x8PUfzNt3qrCKtNf9PUNIzNm5Pw3beYUlMAva9V5nV/pXTLFv0Slqrql5xsNkKGDCH+4w9RArk6sABg8+w3WX0qFEWzoSl6aw0FjSDNzJgLu5HUxbvK0760fPlyewHCc/Xs2ZO+Hu4YFKIuq/EzNw8++KD9/6qq8v7777Nw4UK6detGUFCQw9jXXnutGiEL4VsF73+gJzagJzZgr5djO36crEmTSJz7K4qikP2Phyktf6MoXzBftljY/Oef5E55kdjnn6vN8EUtO7r6R1af0hNYrXzXlqKgoVBKMHOXbeavLXsQFOl6y3htOHjwoNPEBmDjxo2kpqaSkpJSi1EJ4V8eJTcbN250+LhHjx4AbNu2rcYDEqKmaVar61YRNhulW7dhWbsWY5MmlPz2m/NCgapK4ZcziH74IQxy5i9gbdmyFUVLOJvYVKApRsyEsffPr+k0+i4/ROdo7dq1bsesXr2aK664ohaiEaJu8Ci5+f33330dhxA+Yz10CDUjw/UgoxHz8hUYU1PdV0A2m7Fs2EDo0KE1FqOoOzRV5YRadWJTYRTHT57Gg5KPPpebm+t2TJaXPd6EqO+83gp+2223kZ+fX+n+wsJCbrvttio+Qwg/c1NKH9DX36gqaB6MBVC9Lg8l6hEN97uivK8Q5j/VKGcmRL3mdXLz6aefUlxcXOn+4uJiPvvssxoJSoiaZGrRwl6HxymrleA+vQnu3duDCU0Ede9WM8GJOkcxGEhSMlE01601khM8b6vhSxFltZhciXH3/S9EgPE4ucnLyyM3NxdN08jPzycvL89+y87O5tdffyUpKcmXsQpRLUpICJG33uK8RorRiLFlS0IGDyaobVuCBw1y2X4hbMJ4jNJXKKB1bd/S+WUpTcWElfaDr6ndoJzo7UFCfsEFF9RCJELUHR4nN7GxscTFxaEoCu3bt6dRo0b2W0JCArfddhuTJ0/2ZaxCVFvUffcSMuQi/YOK9U0MBgwxMcR/8pG9B1Tcm6/ra28qJkNl/w/q2JHY556trbCFn7S+6K90iTwD4HAGR9FsGLBxSY+mhDZKdvbptSotLY1WrVo5fbx9+/a0aNGiFiMSwv88rnOzZMkSNE3j4osv5vvvvycuLs7+WHBwMC1atKgXWw2lzk3DpVmtFM+eQ+Fnn2M9eBAlOorwK68k4sYbMCYmOoxV8/Io/GomRTO/Rs3MxJjShIiJEwm/+i8oYWF+egaiNmmqyrE1s9m+fSsZpZEYsdEq2krngWOIaeldR/LasHPnTjZu3EhBQQEAUVFR9OnTh3bt2vk5MiFqhjfv314X8Tt8+DDNmzevt2XIJbkRQggh6p8aL+K3ZcsWh4+3bt3qdGy3brLQsq4yr1pFwYcfYV65EhSF0MGDibz9doJ796r2nGphIUVfzqDwyxnYTp3EkJBAxLXXEnHjDRgqNr4ELJs3k/PEk5Ru2qzvTAoLJWzcOGKnvFDjvaOqopWWUvTd9xR+9hnWgwcxREYRdtWVRN56C8bkunGJIeDZrLB1Bqx9GzJ3Q3AkdL0eLvg/iG3u7+iqzWYuYvfiz9l5JJM8LYoQSmgfb6DT0KsJT3S8JFRcXMyOHTvYs2cPJSUlREVF0bFjR9LS0jDVQv8wTdPYv38/27dvJzs7G5PJROvWrenSpUulN4zS0lJ27tzJrl27KCwsJCwsjLS0NDp16kTIue1W8o7D6rdg65dgzoO4dtD3buh+Ixgdi71648iRI2zbto2MjAwMBgPNmzena9euxMu6N+GCR2duDAYDiqKgaZrbMzY2m+sdBv7WUM/c5L/7LnnPT7G3EQD0/6sqsS9OIeIm7xuh2rKyyfzL1Vj37NHvKP9WMhgwpqaQOOsHjE30DtpFc34i++93VzmPEhtL8srlPi2Kp5WUkHnTLViWLz/bUgHAaESJiiTxu28J6tjRZ8cXgK0UZl4Be38BxXB2271ihKBwuHkRpNa/NgGlhbn8+uVbnNISAU1/boCiqYRoRVw+chixrXoC+u+fOXPmUFxcXGl7dkJCApdddhnBPuhPVk5VVRYvXsyBAwfsv9NB7w1nNBoZM2aMvWN6SUkJP/30k71XVTlFUYiIiGD8+PFnd2qd3AzTh+lJTfkapfLXuPUI+OvPXvcH0zSNNWvWsHnz5kqxAgwfPpzWravXHV7UT968f3u0oPjgwYMcOHCAgwcP8v3339OqVSveeecdNm7cyMaNG3nnnXdo06YN33//fY08AVGzLOs36IkNnE1syv+vaeQ89jilu3Z5PW/u449j3bdPT2oq/qJWVWwn0sm67wH9Q4uF7Hv+z+k8Wk4OZ2651evjeyPv9TewrFxpj8/OZkPLL+DMbZP0rt7Cd5b/B/bN1f9fsZ6QZoPSIvhqvJ4A1TNrZ79DhpagLzpXzv5K1RQDZiWMBfN/s39vLVy4sMrEBuDMmTOsWrXKp7Fu376dAwcO6PFViEHTNGw2G/PmzcNqtQKwbNkycnJyKs2haRqFhYVni7uqNv21q5jYwNnX+OBiWDrF61gPHz7M5s2bq4xV0zQWL15MYWGh1/OKhsGj5KZFixb224svvsibb77JnXfeSbdu3ejWrRt33nknr7/+Os8//7yv4xXVUPDxx863NgMYDBR+6l2NItupUxT//ItjsuQwwIZl+XJK9+0j/403nY8rY1mzFrVsIWRN08xm/fk5S15sNmxHjmJessQnxxfol6NWv+m8SKJmg4J02D2nduM6T6UF2ezKj0JTqv5VqilGspV4Tm5ZQEZGBpmZmU4L6mmaxp49ezCbzT6JVdM0l0sKNE3DbDZz4MABioqKOHjwoMtYT5w4oSc/e+dC7mHHxMZhsApr/gdWi1fxbtu2zeWVAk3T2Llzp1dziobD6yJ+W7durXLbYatWrdixY0eNBCVqlnnNGtfJhc2GedVqr+a0bNnqUeVfy/oN+qUgdzQN8+o1XsXgKeuhQ2h5ea4HmUxY1m/wyfEFkHcMCk+5HmMIgqMrayeeGpJ9aBNWxfVlJEWzcerQLjLctQBBv2x05syZmgrPgdlstu+kcsZgMHDq1ClOnz7tUVXjU6dOwbFVYHCzVqgkG7IPeBMup06dchmDpmn68YWogtfJTceOHZk6dSoWy9ks3GKxMHXqVDrKmoU6STG46pFTxsuFjIqrM0EVx5lM4MnxASW4+osOXfIkVk3zbJyoHo++BzT3b5J1jEc/Wyj2dYueMBi8/rXsEU+PryiKd7F6+PPt7WvrSQy++lqJ+s/r74x3332XefPm0bRpU0aMGMGIESNo2rQp8+bN49133/VFjOI8hQ6/2O1lqdCLh3k1Z3Cf3nDubokq5g0ZMICwy8e5n9BgIHjAAK9i8JSpVSsM7nZD2WyEDB7sk+MLILqpvnvGVc8m1aovPq1H4tr0JUR1ve5DUwykpvUhNTXV7XxBQUEkJCTUVHgOQkJC3O4wUlWVpk2bkpycjNFNsq8oCk2aNIHWl+ivnSvRzaCRd4t/mzVr5jbBadq0qVdziobD6+Tmggsu4MCBA7zwwgv2NTdTpkzhwIEDUuK7joq47dayxY5V/KJQFAgK8nq3lCE6mogbJjpW+3UYYCDs8nEYU5oQftONEBbqcr7QMaMx+GgbrGI0EvX3u5wPMBoJ6tnjvLbECzcUBQb9E3BymUExQWJnaD28VsM6X8aQMLol4bSLpqLZSFFOEZ82gNjYWLdv2F26dPHpdvDu3Z0XH1QUhaioKJo3b05wcLDLM/GKotC6dWsiIyOh+SBo0lt/DZ0Z9E/nvyuc6Nq1q9PLUoqiEBwcTPv27b2aUzQcXhfxq+8a6lbw4l/nknX3ZH3tTflaGYMBgoKI/+gDQod5d+YG9IW6Z+74G+ZFi89uMS/7N/iCvsR//hmGyEhAr3Fz+vIJYK38F54xLY2khfN9eopZU1VyHnucos+/OBtr2ZZwU+vWJHz7tdS68TVNg/kPwcrX9EsUqhX9TI4GsS3hlj8gtv61CVCtpSz58iX2mhujaDa9J5WmgmIgTstk7F9uIixeP8NQUlLCr7/+SmZmpkN5DU3TaN26NRdffLHPL7WsW7eODRs2OGyvBggPD2fcuHH2Jps2m40FCxZw5MiRSrEmJyczatSos9vW847rW8Gz9p7dAl7+Gve9G8b8z3lvNxd27drF0qVLK8UaFBTEmDFjaNy48fl9MUS9UuMViufMmcPo0aMJCgpizhzXuxkuv/xy76KtZQ01uQGwHj9O0RdfYl6xEgwKIYMHE/HX68/rTV1TVcxLllA482tsx45jbNyY8Gv+QuiIEfp6mwrU7Gxyp0yl+Lff0MxmjEmJRN51F5E33nC+T82zWDUNy7p1FH45A+u+/RhiYwgfP56wcZehhLo+syRq0LE1sP49yNgOoTHQ+Rrocj0E+76Qo69oqkr6xt/YtXU9uRaFUINKuzYtaTXwaowhju06VFXl4MGD7N27l+LiYqKjo+nQoQMpKSm1Vvk9MzOTHTt2kJWVRVBQEK1ataJdu3YEBTmue9M0jWPHjrFr1y4KCgoIDw8nLS2N5s2bV07CSoth+zew9SsoyYHEjtDrDmg+8Lxizc3NZceOHZw6dQqj0Ujz5s1JS0sjVH5mG5waT24MBgMnT54kKSnJ5V8ViqJIEb8GSrPZ0IqKUMLDPV5s7NG8JSVoqooSFubyF7+qqqhZWRiCgjCU/eUpRG3TNI3S0lJMJpPfFrtac0+CYsQUneh+sKdzWsyU5GcTGtUIU7B3xfiEqCk13n5BrbDlV5VCZ6IC28mT5L/9DkUzv0YrKoKQEML/chVR90zG1Lz65fSLFyyk4J13sKxZC4CpdSsibr+diBsmOiRPqsVC9r33UzJ3rv2SlxIVSeTttxP90D/O78kJ4aGSkhI2b97Mzp07sVgsGAwGWrduTY8ePRyaDPuKZrOx97vH2ZIZQlaQvnA5ofQI3ZMVWk94ptp/cJw+tJMlv88nyxKqX1bSNGJMJQwePJiU9rJGTdRdXq+5KSkpqdenA+XMTc2xHj3K6cvHo57JcqyjYzSiRESQ+OMPBKWleT1v/rvvkff8C45tEsrO2oSNG0ejt99CMRhQLRZODbwQNT29ynlCR44k/uMPvT6+EN4oLi5m9uzZ5OfnO6wLURR9C/jYsWPtLQ18QbPZWP7+fewwdrOv9dEf0P/fgy1c8Lf/eT1v+p6N/PT7auCczQhlz/GSfh1p1eOiGngGQnimxtsvVBQbG8tFF13Ek08+yaJFiyguLq52oKJ+y/nnI5UTG9BbGhQWkn3v/V7PWbpvn57YgGORwLIWD8Vz5lD842wAch9/0mliA1Aybx7Ff0jVYeFbq1atqpTYgH6JSlVVFi1a5NMz3sfmvaonNuDQ/qH8/5voxsnf3/N63vl/LKNSYgP2jxevdl7tWAh/8zq5WbhwIaNGjWL16tWMHz+eRo0aceGFF/L444+zYMECX8Qo6iDr4cOYl/7psv1C6bZtWMp6w3iqsHw3kzMGAwWffAJA0axZbufLf+llr44vhDdKSkrYv3+/yzYFhYWFHDt2zGcxbD94EsVZ6wP07eg7dm73as4Tu9ZhJsT5DidFwaYEsWfVPK/mFaK2eJ3cXHjhhTz22GPMnz+fnJwcfv/9d9q2bcu///1vRo0a5YsYRR1Uunu3Z+O8bMhZun2761YRqop1Z9mcHpw1tB454tXxhfBGbm6u27MyiqKQlZXlsxgyDY317edOaIqRTFwX7zvXsYOe/XynH5efL1E3Vata1J49e/jjjz/sN7PZzGWXXcbQoUNrODxRV3m6ddrbLdZKWLh94aJT7iojV5wvyEctHYQAjwruaZrm08J8Js1NJ3VNcz/mHEHBwYD7jtumoPrVLkM0HF6fuUlNTaV///789ttv9O/fn7lz55KZmcmsWbO47777fBGjqINC+vZFKSvQ51RQEKEXebfgMGzUSNeJjdFI2Nix+n+raOB6rhAv20oI4Y1GjRoRERHhdlzz89g56E5rk+vLUqDROsxN49hzdLxghOufQwBNo0u/+lVRWjQcXic3iYmJFBUVcfLkSU6ePMmpU6dkUXEDpISFEemqpYGiEHHzTRgaNfJq3rArJmBo3LjqdTeKAgYDkbffBkDM44+6nsxgIObJJ7w6vhDeMBgM9OrlfEu0oii0adPGpzszOw27FqNWqu+OOvf4mo0QrZi0Syd5NWdoVCNSIizOExxNIy6ohJjGvkvahDgfXic3mzZt4uTJk/zrX//CbDbz2GOPkZCQwMCBA3n88cd9EaOoo6Lu/T8ibr1F/8Bo1G8mPSkJmzCBmCe8/34whIeT8E2FVgjl8yoKSlgY8dM/Jqisn0zY6NFEOatlYzQS/8XnGGuhxoho2Dp06GBPcMo7apcXnGzatClDhgzx6fEj2w1mTFoIwVoxaBqKZrOfyQlVCxnbPZ6wps57Sjkz5rq/ExdUon9QtluxPNmJNJQwYeLfauw5CFHTzqu31JkzZ/jjjz+YPXs2X331FaqqSoXiBqh0716Kvv0OW3o6xoQEwq66iuAunc9rTq20lJLf5lHyxx9oVhvBPbsTftVVGKKiKo21Hj9O7nPPU7p1m37ZasRwoh5+CEN4/S3nL+qfvLw89uzZQ15eHiEhIbRt25akpKRaa6lQmnOC/b++zskc/Ux6Snwkrcc+hCnSu8XE5zq8ZRkbNm6guFQj1AjdunambZ/61b1dBIYab79Q0Q8//GBfSLxjxw7i4uK48MILGTp0KEOGDHHZdbYuCMTkRtM0bAcPoZlLMDZvjsHFGgDNYsF68CAoCqZWrer0glvVYsH85zKwmAnp39/lJS6tuBjr4cNgMunPy8V2cjUnB9vxEyjRUZiaNfNF6DXHXADZByAoDOLaVqv5YJWOrIS8o3o35/g2zsepqt4M0WaBRm1qrP+TJe80+en7CIqIISqlA4qLVgVFRUUUFxcTFhZGuIuEVVNV8o/toLQ4n6gmbQmuwfYDvnD48GGys7Np0qSJ6waQmgZZ+8BaArGtIMT5WjebuZjco9tQUIhp0RVDkPPF9xaLhfz8fIKCgoiKiqq1JKycqqrk5eWhqirR0dEuF11brVby8vIwGAxER0f7rbWFJ1RVJTc3F03TiImJwejid1FpaSl5eXmYTCaio6Nr/TWob3ya3CQlJXHRRRfZk5muXbueV7C1LdCSm8JvviX/jTexHToE6LuTwq+9huhH/unQY0krLSX/zbco+PgTtJwcAAwJ8UTecQeRf7+rRvtBnS9VVcm+6++UzP3NoZBfUI/uJHzxuUOSoxYXk//KqxR+/gVaob67w5iSQuTkvxNx880Ovyysx0+Q9+KLFP/0s327eVDXrkT/82FC69rC4+JsWPQ4bPpEf1MDiGsHQ56E7jdWf95l/4Ylz0FphZ0wUU3hqi+hZYXF35oG6z+AP1+E3MP6fUER0Ot2uPgFl2+wrhRlHGT13C/YXxKPquhvZnFaJn06taHl4GsdxmZmZrJ69WqOHz9uvy81NZV+/fqRkJDgMPbgnzNZt+MA2Yp+v0Gz0jY0iwtG30B4Ustqxeory5cvZ/t2x7ozBoOBCy+8kA4dOjgO3vQpLH1BT24ATGHQ81a4eAqExdqHqaVmNsz+H9szjZgNegIYqhbQrbGBbuMmYzCd/SOmuLiYNWvWsHfvXvs29tjYWHr37k2bNi4S3RqiaRrbt29n8+bNFJb9zAYFBdGxY0f69OnjkORYrVbWrVvHzp07KS3Vd3xFRETQo0cPOnXqVKeSAVVV2bx5M1u3bqWkRP+ZDQkJoUuXLvTs2dMhITObzaxdu5bdu3fbr3ZER0fTq1cv2pdddheV+TS5qUnTpk1j2rRpHCp7Y+7cuTNPPfUUo0ePdvo53377LU8++SSHDh2iXbt2vPzyy4wZM8bjYwZScpP/xpvk/fs/lbdOG42Y2rYlcfYsDFFRaDYbWbdNomTR4ioXCIZdMYFGb71ZZ35RnBpxydlaNudQIiNJXrsaQ3Q0WkkJmddeh2XDRsdqxmUibp9E7LPPAPqlq9Njx6FmnVNRuew5N3rrDcKvuKLGn0u1lOTCRwMhczc47IJRAA2GvwiD3Symrsr8h2HFK04eVODmxdBqqP7hwsdg2dSzx7QPM0JyT7h1iddncYoyDjHrh28pUiId67KUtQm4qIVCh5F3AHDq1Cl+/vlnVFWtsqXBuHHjSEpKAmDnb+/z55Gz89jHajYitAImXHUN4YktvIrVVxYvXsy+ffucPj5o0CA6dy67pLt0Cix+gipfg4QOMGkFhEajWkuZ/+lUjliTHCsUA2gabUIyuPimx1EMBkpKSvjxxx+rrKgMMHDgQLp06XL+T9SFqpI70F/bxo0bM3bsWIxGI1arlV9++YWMjIwqY+3SpQsDB55fx/GaomkaixYt4sCBA1U+3qJFCy655BIMBgMWi4XZs2eTk5NT5fPq27cvPXv29HXI9ZJP2y/UpKZNm/LSSy+xfv161q1bx8UXX8z48eOr/MYHWLFiBddffz2TJk1i48aNTJgwgQkTJrBt27Zajtz/rIcPk/efsjeqc39AbDase/dS8P4HABT/8islCxc53flQPOtHzL//4cNoPVf4zTdOExsAraCA7H/+Sx874yss6zdUmdgAFH74EZateon4vKkvVU5swL5IMueRR1GLimrmSZyvFa9UkdiA/Q1u8ROQc9i7OYtzYMWrLgZo8P1f9f9mbC9LbCoc0z7MBic3wNq3vTs+sG7+l5UTG9DfkDWNZYcsmHP1N7KlS5dWSmzgbEuDpUuXomkaJdknWX7Yqr+O57yxa4qRQiWS9fO+8DpWX7BYLC4TG9B/xwGQdQAWP1l2bxWvQeYuWPkaAAeXfc0RW3LlxAZAUdhvaczRNT8CsHHjRqeJDcDKlSsp8uHPQUZGhtPf75qmcfLkSXaVFf7ctWsXp06dchrrtm3byMzM9Fms3jh69KjTxAb0S5Dlf8Rv3rzZaWIDsHbtWvLz830RZoPi1+Rm3LhxjBkzhnbt2tG+fXumTJlCZGQkq1atqnL8G2+8wahRo3j44Yfp2LEjzz//PL169eJ///O+KVx9VzTza72xpDOqSuFnn+vl3z//3PVYo5HCL7+s+SCrIf/td9yOKZmnl3wv/PQz1wONRoq+nIGak+NwKaoqWmEhxT//4lWsPqFpsO7dKhKbihTY+LF38/7xNJXeJM9VkA6ntsOGD8HgojibpsLaaV4d3lqUx97CWOeVdBUFFSP7ln/P6dOnyc7OdtnSICsri8zMTPYt/x4Vg9O1SJpiZE9hDNZi/79ZLF261O0YTdP0N8mNH1edrNgH2mDdNNA0duw96Lb9wq4d21BVlV27djn9upbbs2eP2zira9euXW7PEO/YscPhX2cURWHnzp01Ftv52LFjh8vnVR6rpmn2f12N3e1hBXjhXJ1ZlWWz2Zg5cyaFhYUMGDCgyjErV65kxAjHVfojR45k5cqVTuc1m83k5eU53AKB9dAht0W21MxMKCnBuv+A07MbgH6mZ9/+mg2wmtRTGe4HWSxAWWsFV18Dm43SAwewnUgHq9X1nCaTfd2SX1kKociDv0azXJ8BqCTDw95C6ev0uVU3X6+cQ+6LvFVQlHkYm+J68boBlbzcXI9/RvPy8sjNy8OA6/YHNiWY4qzjLsfUhuzsbI/GHT9+vOz1dfP1LcwAawm5tjC37RdyLEGUlJTY16244svfka7OWJx7fHdxaJpGbm5ujcV2Ptw9L03TyMnJwWq12tfjuBIo71P+5PfkZuvWrURGRhISEsJdd93FrFmz6NSpU5VjT548WWlXQePGjTl58qTT+adOnUpMTIz91qyu747xkBId4/psDEBQEISEOCwsrnoyBUOsmzG1RAkL82CQ/heSwV2FZIMBQ0wMSowHa6tUFSW68jbzWmcKBYObHWyKAiFevl5hHhZTjEiG0Fh9XYcrwZFe7dwKjnBfb0hDITg4iODgYM/mDA4mJCgIDfdxBIXHejSnL4V42DYkIiKi7DVw8/NtCAJjCCFKqetEU1MJMdgI8nBnpKdf/+oICQlxe+am/Pju4lAUxeOvqa95EkdISAhGo9GjnV6+fA0aCr8nN2lpaWzatInVq1fz97//nZtvvtnt6UhvPProo+Tm5tpvR48erbG5/Sn88nGuz0YYjYRdPg7FYCD8yivcJkJhdWQxbdhll7kdE9RN36EXduUVrjuIqyrh48djSk0lqHt3118DTfPo2D5nNEGnv7i+LKRaoct13s17oQcLkI0h0OYS6Hyt68tiBhN0m+jV4UPjU0lRTrm8fKIpRtr0HEZKSorbX+4hISGkpKTQusdgl2ctFM1GqnKS0EbJXsXrC/369fNoXPfu3aHLta7PnhlM+hiDgbZJoSguz/IotG0aT1BQEM2bN3eZXGiaRtu2bT2Kszratm3r9pJMu3bt7GPdxVobu7s8UR6zuzEGg4FWrVrVm+dVn3mU3Fx55ZUe37wVHBxM27Zt6d27N1OnTqV79+688cYbVY5NTk7m1KlTDvedOnWK5GTnv7hCQkKIjo52uAWC4IEDCO7fr+o3d4MBTCai7v47AOETJ2JIiK96rNGIMTWV8L9c5eOIPRP9yMPg5o0t5vnnAIi843aU8HCnz8vUsSOhIy/V5/3nQ2WLTqv4pWJQCL/uOkxNm553/DVi8KP6m1eVC0SN0HIYtBjs3ZwpvSDJTdmGgf/Qv3fajYaUvlWfvVGMehI0wEllaBd69+4LKE7aBKi0CT5FbOtemEwmevfu7Wau3hiNRuLa9aN10CmUKuYsP07vvv29jtUXGjdu7LJOD+ibLIxGI7QcCi0ucvIaGPTvj0GPANBx6LWEaoVVJo6KZiNSy6XdRfo2e3etIpo3b15pm31NatmyJY0aNaryzV1RFIKCguy7tbp27YrJZHI6Ni4ujpYtW/osVm+0b9+eiIgIp7GGh4eTlpYGUGlb+LljU1JSXL6nCc94lNxUvKzj7na+VFXFbDZX+diAAQNYtGiRw30LFixwukYnkCmKQvwnHxNSXtrdaISy+hCG+HgSZnxBUFnNDGNcIxK++w5T67JGkyaTPSEwdehAwvffur/EU0sMkZEkzv2l6qacRiOxb79FSNkbn6lpUxK+/Rpjkyb64xWeV3CfPiTMnGEvUhg6dCiN3n4LJSL87FiDvhA1/NrriH3xBZ8/N4817goT50J42ZuMIejsm1y70XDdj9Ur5nfHGkh0ss231x0wfErZ8Yxww1z9DRb0Y5dfKotMhpsXQbz3f9036TWKSzo1IljTf74NmlVPSjSNNiGnGXLtg/axXbp0oU+fPvY3gYr/9u3b9+x2aWDodffTJuS0HqqmYtD0Mx7BWgmXdkkgucelXsfqK9dccw1hTi69JiQknC1roShw/RxoO7Ls4wqvQXgi3DAPGuuvZWijFMaNHE6MlqMPrdB+IY4sxo27nOBI/bJgUlISo0aNsl9GMRgM9jfkli1bMny4bxthGo1Gxo4da9/GX7FVRUREBJdddhmRZT/7UVFRjBs3zt6YtOLY8i3jdaWYX3BwMJdffjmNympwVYw1NjaWcePGERoaCkBcXByjR4+2fx9UfA2aNWvGpZdeWmfKctRnfq1z8+ijjzJ69GiaN29Ofn4+M2bM4OWXX2bevHlccskl3HTTTaSmpjJ1qr4tdcWKFQwZMoSXXnqJsWPHMnPmTF588UU2bNjgcW2GQKpzU6501y5KFixEM5sJ6tSJ0EtGVFl5WNM0LMtXYF6zBhSFkIEDCL7ggjr7g1T4zTcUff8DWG2EXDiIyP+7B0MVVUw1mw3z739g2bwZxWQiZNhQgrt1q3JOtbiYkp9/wXroEEpUFGGXja07Z2zOZSuF3XPg1BZ9LU7aeEiqej2aV46thj+nQnEWxLfX6+ZEJlU9Nn0T7P0VbGa9mnH7sXrycx6sxYUcWvkdOWdOExQcRKueFxPdvOqzSiUlJezfv5+ioiIiIiJo3bq1/U3iXLmHt3Bw4+9YS0uJTUiiZf+rMIW579jtD0ePHmX16tWYzWYiIiIYMmSI/Y2xklPbYM9PejHHxt0hbRwYq/j5VlVObPiVk4f1nTYpbbqQ3O2SKqs/22w2Dh06RHZ2NiaTiZYtWxIbG1uTT9GtjIwMjh07hqqqJCUl0axZsyp/F6mqyrFjx8jIyMBgMNCsWTMSE+tm9eny7ewnTpwAoEmTJjRp0sTp8zp06BBZWVkYjUb7WS3hXL0p4jdp0iQWLVpEeno6MTExdOvWjUceeYRLLrkEgKFDh9KyZUumT59u/5xvv/2WJ554wl7E79///neDLeJXHbbMTCwbN+mngHv1whgnP0wNjs0KR5frVZAbtYbkqhNBQG//cGSZ3n4huQfE1nIXaE3Tk7GCkxDVBFIvcH7WymrWY7UU6EXuEtKcz1ucDUdX6muLUvpClPPLAPn5+Zw5cwaj0UhycrLrhbknN0P2QQiLg2YD9TVU58tSpD8va4l+Vq9Rq/OfM4BlZWWRl5dHcHAwycnJdebsjjh/Pk9uvvvuO7755huOHDmCpWxbbrkNGzZ4O12taqjJjZqXR84TT1L84+yz9V6Cggi/5mpinnlamkw2FBs+hsWP68lCueSecNm70PSCs/fZrPD7k7D6TSgtL+qm6GduLnsPolN8H+vun+G3+yG7QpmCuLYw6g1oX+EPGk2DVa/DkuehpMJ26+YXwrj3IbHj2ftKi2HeP/Q6Mrayy9+KETpfDWPfcdhVVlhYyJ9//smRI0fs95lMpkqXzAA9Afv5Lji56ex9kU30s2I9b6ne81dt+nNa+RpYKtTpaXMpjPug9hPNOi4zM5M///yT06dP2+8LDQ2lT58+TnfgivrFpxWK33zzTW699VYaN27Mxo0bueCCC4iPj+fAgQMu2yYI/9FKSsi85jrHxAagtJSir2Zy5sab0DyofyHqudX/gzmTHBMbgFObYfoQOFH2h4mmwY+3wLKXKyQ2ABrs/U1vDVF0xrex7v4JvrpcbxpaUdZ+mHGZnviU+/1pmPegY2ID+pmZjwbCmbKaQKpNn3P9e2cTG9DP3mz/Fj4ZotcZQr8cNnv27Eq7K61WK5s2bWLJkiVn7zy+Tv/cU1scj1+QDrNv9brgod0vd+t9wComNgAHF8OH/SHfeQmMhiYrK4s5c+ZUqlhcUlLCsmXL2Lx5s58iE/7idXLzzjvv8P777/PWW28RHBzMP//5TxYsWMC9995bZwoqCUdFP8yidOvWqiv0qiqWVaspnvtb7Qcmao85Hxb8s+rHNFVf37PgYf3jY6th65dUWUROs0LeMVhV9Y7GGqGq8Os95Qc8NwD9n7n36uPyjsOfU6qeR7Ppl6j+eFr/eM8vcGBhlbu10GyQsQ02fgLAli1bKCwsdLptee/evWRklBWcnP+Qvm27qnnLHzcXVP2YM6e2wvr3qfI1UK16Ab8V//FuzgC2Zs0abDaby5YGnhTPE4HD6+TmyJEj9mZlYWFh9h4YN954I1999VXNRidqROGMGa5rvBgMFM2cWXsBidq34/uz3cWrotn0MwK5x2DTdDftF2yw4YMaD9HuyJ+QewTnFXo1yDkIR1fAFjd9o1QrbP9GTy42fuy+OGHZ83LXpkBRFL1NQc5hOLzEdV2g0iLYNcv1cc/l0WvwoVdVogNVcXExR44ccfl6qarqsveTCDxeJzfJyclkZWUB0Lx5c3sfqIMHD7otqy38w3Yi3XX7BVXFdtz/5emFD+Ud82yXU/4JyD/uvv1CwSnfvbHmefi9mHdMH+suYVGtekuL3MNuenZpkHdMb8jp5q98TdMoLCzUv17uGEx6rN7IO+78TFA5c56+hqiB86TRp8Fg0F8v0WB4ndxcfPHFzJkzB4Bbb72VBx54gEsuuYRrr72WK+pIlVvhyJiU6LouisGAIamx88dF/ReZrK85cTuuMUQ0dn3WAPTdQL4qIRDp4fdiZGP95i4JUIx6vFGp7hOhyGSPyvoriqLXKYnwIFbV5tk4hzgau2+/EBQOQR60KwlwzuoGVaSqqkfjRODwOrl5//33efzxxwGYPHkyH3/8MR07duS5555j2rRqLpwTPhV+3bWuB6gqEddeUzvBCP/odBUYXVR+VgzQbBDEtoDuN7o+c6MYodekmo+xXIshejLmSlSqXsG360TXyY1ihA4TIDQaetzs+syNYoCetwF6xVl3JfLbt28Pca2haX/XiYgpRP/6e6Obu9fABD1u9V2CWY+Eh4eTmprqtiu3tDRoWLxObgwGA6YKhdSuu+463nzzTf7v//5Pmn3VUeHXXIOpTRunbQqCunYlbFwd6KskfCesEQx5qurHFIN+u+Rl/eMWF0H7cVW/YRtMeuXk/vf7LFSMJhj5musxI1/VL7M1agn9/q/qMYpRTyyGPqN/3GECNB1Q9dkbgwliW0Gv2wHo1q0boaGhTt8wW7RocbaJ7yX/Pvs1rMrQZyDUy+rtqX2g83VQVVNQxQgh0TDIyQLxBqhv374uk5uePXvKmZsGplrVjbKzs3nllVeYNGkSkyZN4tVXX7WvwxF1jyE8nITvvyVk6BDHBxSF0Esv1dsU1JHuusKHBj8Kl74Cwed0P49pATf8Bs0H6R8rClz9jf5Gf+7lqZQLYNJyvaCeL3W9Hq78Qm81UFFEElw1Q28aWW7kf+GiJ8F0zptXQhrc8oe9TQHGIP15dr66ciLSchjc9qc9CYmIiODyyy+vVAlXURQ6duzI8OHDz76Zthist8uIOafuTEi0nqRVNwm54lO44J7KXeKb9NRfA6lzY5eUlMTYsWMr1T4xmUz07dvXba8yEXi8LuK3dOlSLr/8cqKjo+nTpw8A69evJycnh59++omLLrrIJ4HWlIZaxK+c9dAhLGvXgaIQ3L9f3W09IHzHUgT75+t1YRq1huaDne+mK8iAg4v06r8pvfUKubXJVqpv385Ph6gUaD28ytYDAJTkwf55ZRWKO0LTfs4v2+Qeg0N/6Jepmg6AhPZOQzhz5gyZmZkYjUZSU1OdnwFQVTi8VK/NExan94WqiTUxRWdg/wJ9t1tydz25EVUqb39QXqG4adOmritKi3rFm/dvr2uDT548mWuuuYZp06bp3WvR+5TcfffdTJ48ma1bt1YvalErTC1bYqojnXSFf5g1I4cMHTEbzUQTTXNcnMLN3KX3t7Ka9S3NiZ2dJkIZGRlkZGSgKAqpqak106vIGKQ3C/VE7hG9uJ85D1KOQ5NeYKr6Unm2GsEJU080TaOx1ghnnYo0TcNisWCxWDAajVitztfBWKxWNp0OJztbT4B6pJYS7Sy5KcnVv65FmfqZs/Zj9UtoVQmPh67XuXji1aCqetKasR2CI6D9Zb4/G1cLFEWx93Nyx2KxcPjwYYqLi4mMjKR58+YOSy58TdM0Tpw4Ye8t1bx5c3vTUHH+vD5zExYWxqZNm+zt28vt3r2bHj16UFxct7cmNvQzN6Lh0jSNdevWsXnzZlRVRVEUNE0jNDSUiy66iJYVk968E/DBBfq28IqMwXDVTOh0dmdkXl4eCxcurFQdtlmzZgwbNsxpo8saYynSK/ZmnPOHlWLUL8MNuN9+V0lJCYsXL+bYMcet2YmJiQwfPtzhd8KZM2dYtGgROTk59q8VQOvWrRkyZIjDGYHVq1dXWQU3OTmZyy677GyrBk2DP1+EpS/oZ2IUo372KCxOb//Qxc3i/5pwZDl8P1HfGq8Y9JgUg75IfPSbzpOsAKFpGlu3bmXdunVYrVb7axscHMzAgQP1heI+dvr0aRYtWkReXp7D91a7du0YPHhwrSZZ9YlP2y/06tWLnTt3Vrp/586ddO/e3dvphBC1ZO3atWzcuBG1rOZR+S/UkpISFixYcPYN32aFtztXTmxAb6D5zVVwZKX9c+fMmcOZM5XbMRw7doxffvkFW1WVsWvSuz0qJzagJw3zHoDNeqE/m83Gzz//zPEqajplZmYyZ84ce32b/Px8fvrpJ3vV9Yp/Ax48eJB58+bZ79u4caPT8v4nT57kp59+OnvHny/C4ifOFlQs371VnA3fXQ+75njzzL13cgt8NgLyytpKaCqgnS0K+OMtvj1+HbBt2zZWrVplPwtX/jpaLBb++OMP9u3b59Pjly/hKC+AW/F7a9++fSxcuFBqxtUAr5Obe++9l/vuu49XXnmFZcuWsWzZMl555RUeeOABHnjgAbZs2WK/CSHqhuLiYpf9dTRNY+3atfoHK14Bc46L2TT49W4AduzYQXFxcZW/jDVN48yZMxw8ePA8Indj7zzI2ut6zPx/AHDgwAGysrKcxlpcXGz/w23Lli2UlpY6HXvixAlOnNAL+LlrFnzq1ClycnL0S1FLX3Ayquw4C//l26rDS1/Q1zFV2YJChW0z4dQ23x3fz6xWK+vWrXM5Zs2aNT5NLjZt2uS0VYSmaRw5csSh+aeoHq/PfV1//fUA/POflXcAXH/99fZTbIqi+P4vNiGERzypIH769Gny8vKI3vCh+wnLul/v3r3bbZuCvXv30rZtW2/C9dzyl9yPKcyAnCPs3bvX4RLAuTRNY8+ePfTs2ZM9e/a4fV779u1DVVWPfs9t2LCBi+OPu26BgQaZO/W+Usnd3M7ptdJi2PmD61o/BpPeV6zx1Jo/fh1w9OhRSt00CS4oKCAjI+PsVv8apKoq+/bt8+hnJikpqcaP35B4ndz49K8wIYRPlJSUuHxjrzgu+twu1M6oKmaz2eUQT1oZnJdiD0tQ5B11eoapopKSEjRNc/sGWP68Cgo8a4hZUlKiLx4uX2PjSrGPOq6X5Lo/Nvi+47sfefq96KvvWZvNZr8s7Iq7nyvhntfJTYsWLXwRhxDCh6Kiojw61R4ZGam3CijMcD1QMYLBQFRUVJXrbezDFIWoqCinj5+3mJZwyoNL4PFpREcXOb0sVS4qKgpFUQgPD3fZs6j8eSUkJHgUZnR0NMS29Cy5iPHR79iwOL0WkNXFpg9N06tUByhPvxd99T1rMpkIDg7GYrG4HCe7ps5ftYr4ff755wwaNIiUlBQOHz4MwOuvv87s2bNrNDghRM1o1aqVy3ofiqLQrFkzwsPD9WJ/7rQcCkDHjh1dDtM0jQ4dOngTqneGv+h+TKO2EJFAhw4d3CZ45c+nY8eObtsvpKWlkZiY6LYPFaDXBGs3Vk8wnFGMes2huNZu56sWUzD0vNVN3zANut/sm+PXASkpKfr3uBOKohAfH09cnIvX6TwoikKHDh08+t4S58fr5GbatGk8+OCDjBkzhpycHPv15tjYWF5//fWajk8IUQNMJhMDBw6s8jFFUTCZTPTv31+/o+v1kNjJ+WTGYBj/MaD3YEpMTHT6y7pVq1akpqaeV+wuNe4M7ca4GKDAVfpuqaZNmzpud684SlFISkqiXbt2AHTp0oXo6Ginz6tTp07Ex8cDuC1c2qFDB307vCkYxk7TYzq3rYJi1L+uo99wOdd5u+gJvcqz4iTBGfoMxARuYU+DwcDgwYOrfF0VRUFRFC688EKfxtCjRw8iIiKcfm91796dmBgv23WISryuc9OpUydefPFFJkyYQFRUFJs3b6Z169Zs27aNoUOHVqp1UddInRvRkB04cIC1a9fatzgDpKamMmDAAMe/Vm1W+Goc7JuHfScP6BWNb/gN4tvZ77JYLKxevZo9e/bY/9gJCgqiS5cu9O7d+2yNF1/68VbY/LnjZZ+IxnobiZZnkw9VVVm3bh3bt2+3r6sxGo2kpaXRr18/h7NbJSUlrFixgv3799vP+ISGhtK9e3e6devm8OZ08OBBlixZ4nC5wWAw0LVrV/r16+cY6+6fYeEjcHrH2ftaXASjXq+d6sO5R2HufbB79tldU1GpMORJ6P23BtGM89ixY6xatcqhbVDjxo0ZMGBArSzkLSoqYsWKFQ4L/cPCwujZsyedO3d2eWanIfPm/btaRfx27dpFixYtHJKbvXv30q1bNyniJ0Qdp2kaWVlZmM1moqKiXK8vMBfo24NLC/UzJBWSmnNZLBbOnDmDoigkJCTUfiEym1WPtShTTxZSejkdarVayczMRNM04uPjXTb9LSkpsVeRTUxMdJmspaenc/r0aSIiImjVqpXzsZoGGdv0xbsxzX13KcqV/HQ4sweCwvVqzoYqGooGME3TyMnJobi4mIiICL+cLSkuLiY7OxuTyURCQkLt/CFQj/m0/UKrVq3YtGlTpYXFv/32m9vr70IEJE2D42th1yywFEJSF/3STogPF9KeByVzF/HbZuo7jRq1gW43QISThbEhkdD7drdzWvKzWPfTe6TnqyhAy8Qweoy7B0OQ86TBIzlHYMsXUJAOkU30WJ01jDSaoPsNHk1rMplITk52O85qtXLkyBEyMzMxGAyoqkpycnKVf1kXFxdz+vRp8vPzsVgsxMfHO29BoSi136frXFFNAqLlQnUpikKjRo1o1KiR32IICwuTbuU+4nVy8+CDDzJ58mT7lsk1a9bw1VdfMXXqVD780IP6GEIEkuJs+PoqOPR72UJNBVQrzHsQrvgMOl3p7wjPspr1yzfbvtLXXCiKfhlnwT9h5KvQ7/+qNe3ueR+y5JAVlHjKL2FlZips+PhDxvRpQUrvsd5Pqqr6pZsVr+qtARSDfgll8RMw6GEY8ZLPL58cO3aMhQsXYrFYMBgMaJrGli1bSExMZOTIkQ4LU7du3cqqVavQNM0+dsOGDbRt25YhQ4bY+/AJIWqH15elAL788kueeeYZ9u/fD+gr0J999lkmTZpU4wHWNLksJWqMpsH0oXqvnkpbfBX9DfnWJdB8kD+iq+zHW2HzZ1VXpwW46iuvGzSe2ryQ2av03wOVkg1NQ0Fl4hWjCU9q6V2sS16A3590/vjwFz3b1VVNWVlZ/PDDD1XWJFEUhdjYWK666ioMBgN79+7l999/dzpXWloaQ4YM8VmsQjQUPu0tBTBx4kT27t1LQUEBJ0+e5NixY/UisRGiRh1ZBoeXOqldUvY3w9IptRqSUzmHYdOnzhMbFPjjaa9L/y9fvars06s4i6IoaBhY9evn3sVqKYTlL7ses+wlvWGmj2zevNllJePs7GyOHDmCpmmsX7/e5Vy7d+/2uNifEKJmeJ3cFBcX24tbhYeHU1xczOuvv878+fNrPDgh6rQd37uuGaLZYN9vPn0T9tiuH91cxtH0xaWZu7ya9oyW4Pby0JESL9ceHVwMFjfJgDlPvxToA5qmceDAAbcl8g8cOEB2djZ5eXlu5zx06FANRiiEcMfr5Gb8+PF89tlngN7d9IILLuDVV19l/PjxTJs2rcYDFKLOcvcGDIAGpXUgubEU6JfJPBnnBe3cei3nUhRUvFxv4mkMXsbqDXf9ospbNLhr0wB6IlTegVoIUTu8Tm42bNjA4MGDAfjuu+9ITk7m8OHDfPbZZ7z55ps1HqAQdVZCBxeXecqExUGY/3Zj2CV01Bc6u2Iw6XVsvBCkmV1fytI0win0ak4SPKxonOCb3ZmKorivoVG208ZVob9ymqY53zUlhPAJr5OboqIie12M+fPnc+WVV2IwGOjfv7+9FYMQDUKPm/XKss4oRujz97pRPyRtHIQn4vRHXjFBp6shPN6radtF5bu+LKUo9Gjl3Zw06QnJPZ1/bRUjpPT1TefsMp06uajQzNm2EmFhYbRq1cplghMWFkbz5k62rwshfMLr5KZt27b8+OOPHD16lHnz5nHppZcCkJGRIbuPRMMSkQiXlV2KPfeSj2KEpM5w4SO1H1dVjEH61nSDoXLSYDBBZGO49D9eT9v/qgcIV/P0szcVz+CU/T+BDDqMvMP7eCd8AkFhlWNVjHrRuct9W3aic+fOTuvZAFxwwQX233f9+/cnLCys0tjycv7Dhg2T4mxC1DKvf+KeeuopHnroIVq2bEm/fv0YMGAAoJ/F6dmzFkqHC1GX9JoEE3+F1P5n7wuJgQEPwq1/1q1Cfu1G6TG1uRR7byNTKPScBH9bC9He94AyhYZz3c130cx4ioptGgzYSAvNYMJt/6perMnd4Y610OkvZxdtG0zQ+Wo9Vh+etQG9JcOYMWPo2bOn3heqTHx8PCNGjKBHjx72+yIjI7niiitIS0tzqGfTtGlTxo8fT9OmgdurSYi6qlp1bk6ePEl6ejrdu3e3/0WyZs0aoqOjfdsBuAZInRvhM4WZYC3WexqZzrMyr68V5+g7jiIS9TMkNcBaUkTW3jUYgoKIa9sPQ021XzAXQPEZCIvXKybXMlVVKSoqwmg0uq0ma7VaKS4uJjg42KNu4UIIz/m0t1R9J8mNCEgn1sOGjyD3CIQn6G0KWg8/vyq+liLY/jXs+QVsFkjpDb1ur/oMT+Fp/fjHVuqXjlqPgO431tqZK1VVOXToEAcPHqS0tJTY2Fg6duwo3ZVFnZWfn8+uXbs4c+YMJpOJFi1a0Lp1a6lm7YIkNy5IciMCiqrCz3fBhg/0yzaqVV8crFmh1XC47sfqne3I2AGfjdB7OpW3PihvgzD+Yz1xKbf7J/j2Gj0B0lTsl7zCGukdxFP71sQzdaqoqIhffvmF7OxsFEVB0zT7v3379pXL5aLO2bFjB8uXLwdw+H6NjIzksssuk/cmJ3xeoVgIUUcsmwobyhbXlm/11sr+PfQH/FSNxbyWIvj8EijMKJtPPfuvaoVZN8PRlfp9GTvg6yv1vlX2bfGafivJhc9H6p2vfUTTNH777TdycnLsH1f8d+3atfY2MULUBceOHWPZsmVomlbp+7WwsJBff/21yrYfwjuS3AhRX1nNsOIVKi7kdaDZYNvXemdtb2z/GvJPOGkrgb61fcWr+v9Xl9e2qiIGzQbmXNj4iXfH90J6ejqZmZkuqwlv3LjR5eNC1KZNmzY53YWnaRp5eXlSVqUGSHIjRH11fC2U5LgZpMH+ed7Nu/dX19WMVSvs+Vn//64fXRcH1FTY85N3x/fCkSNH3BbRy8rKori42GcxCOEpm83GiRMn3Lb2OHLEyz9IRCWS3AhRX9ksHgxS9DM83s7rrvKyWqrXsvEkBmuJd8f3gqqqbpOb8nFC+JunZxDdtf8Q7klyI0R9ldTFdeNOADR9l5M3mvRyfeZGMUDj7vpOrNS+rqs0G0yQ2s+743shISHBbeISGhpKeHi4z2IQwlNGo9HtQlhN00hISKiliAKXJDdC1FeRSXrLBFdtCpK6QtP+VT/uTK/bXSc3mgr979P/3+9e52tzAFQb9LnLu+N7oXXr1m7ryXTq1EkqBIs6QVEUunTp4nKM0WgkLS2tliIKXPITL0R9NvoNiGtTdZuC0Bj4y0zva91Ep8L4TwBF31Zun7Ps10XXidCtbCt4uzHQ7z7Hx8uPXx5fkus+TefDZDIxYsQIjEZjlZenmjRp4lBNWAh/69SpEy1atKh0f3m7josvvlgKQNYAqXMjRH1Xkgur34L17+m7nEJioMcteguImPMo/X90pb4ram9ZEb/kHvqZmm436j2qymka7PgeVr0Ox1bpSU7rETDwIWh98Xk+Oc9kZ2ezefNmDhw4gNVqJSYmhs6dO9OxY0cpiibqHFVV2b17N9u3byc7OxuDwUDLli3p3r27XJJyQYr4uSDJjQhomnZ+VYnPd15fHd8L5UXRhKgP5PvVc968f9dQ8xch/CczM5Nt27Zx6tQpDAYDLVq0oFOnTkRGnkcfotIS2DoDNk2HgpMQ20Jfi9LxSr3Ddl2i2mDnLFj/PuQcgogk6HGzfvko+JyFtOZ82PwZbPlCL64X3w563wntL3M8G4NeUGznzp0cOnQIm81GUlISnTt3JikpyXksdeCXdE2+UWRsW8z29cvJKA7CqKi0jA+m4+AriWjcusaOUVOys7PZvn07J06cQFEUUlNT6dy5s7SgqOMksfENv565mTp1Kj/88AO7du0iLCyMgQMH8vLLL7tcTDV9+nRuvfVWh/tCQkIoKfFsu6mcuQksW7ZsYdWqVfby5aD/sjAYDIwaNYrUVO87XVN0BqYPg4ytjq0HNBVaXAQT51ZOGvyltAS+uhwOLNDXuWg29KV0KiR0hFv+0BceA+QehU+G6AkQANrZz+kwAa7+xp64paenM3fuXGw2m8PXtSG1NNjw/WusOxOJotnQytYQKZqKkVLG9Esjucelfo7wrD179rBkyRIAh9cLYPjw4bRuXfeSMSG8VW/aLyxZsoTJkyezatUqFixYQGlpKZdeeimFhYUuPy86Opr09HT7Tao5NkzHjx9n1apVgGP9CE3TsNlszJs3z+Ok18GPt8DpHWWTqY7/HlkO8x48j6hr2KJH4eAi/f/2XUtlsZ7ZAz9MLHtMg6+v0hOc8vYIFT9n12xYOgUAs9nMb7/95pDY6FOcbWkQ6EXGDq/4jnVn9DN/WoXF2ppiwEoQv63aiSXPd20lvHHmzBn++OMPh3L+gP3jRYsWkZeX58cIhah9fk1ufvvtN2655RY6d+5M9+7dmT59OkeOHGH9+vUuP09RFJKTk+23xo0b11LEoi7ZunWry1O6VquVXbt2eTdp1n69C7az7c2aTb9UVZzt3by+YM6Hde85L7in2eDAQr3/0/G1cGLt2b5TlQfrrRSsZvbu3UtpaanTgmOKorBly5aaeQ511JbtO1GcfQ8oBixKKHv+/Lp2g3Ji27Ztbi9tbN++vZaiEaJuqFNbwXNzcwGIi4tzOa6goIAWLVrQrFkzxo8f7/IH12w2k5eX53ATgcFdGfPyMV45/CdOezWVs5nh2Grv5vWF9A1gdddWQIHDS+DQ766L7QGUZMPpHW6/ZpqmkZ6eHrD9mjRV5aSa4HDGpopRpJ+qG2dujh8/7vK10DSN48eP12JEQvhfnUluVFXl/vvvZ9CgQS6LHKWlpfHxxx8ze/ZsvvjiC1RVZeDAgRw7dqzK8VOnTiUmJsZ+a9asma+egggInr5h14E3dk+TC03zbqwA3C/yrE9fqUBNRIVwps4kN5MnT2bbtm3MnDnT5bgBAwZw00030aNHD4YMGcIPP/xAYmIi7733XpXjH330UXJzc+23o0eP+iJ84QfJycluT8c3adLEu0mbX+h+jCEIUi/wbl5faNITTKFuBmnQYrC+ENpVJWHQ6+MkdiI5OdnlsPLLwoG6y0MxGEhSTju/LKWPoklSo1qLyZWUlBSXr0X5zikhGpI6kdzcc889/Pzzz/z+++80bepd0bGgoCB69uzJvn37qnw8JCSE6Ohoh5sIDN26dXP5F6nRaKRDhw7eTRrfDtqMdN6zSTFC9xshPN67eX0hNAZ6TnLeKkExQYsh0LgrNBugF+Fz+rwMcMFkCAqlffv2mEzOq0RomkbXrl3PP/46rFvHds4vS2kqQZqZ9oOvqd2gnOjSpYvbMzOdOvmuSrQQdZFfkxtN07jnnnuYNWsWixcvplWrVl7PYbPZ2Lp1q/d/oYt6r2nTpvTt2xdwrBVRvhX8kksuISwszPuJr/gM4tqiX5oom7c8gUjtC6NeP6+4a9Ql/z57tsme5JTF3aglXDWj7C4Frv0BIpNxuORS/gbeZiQMeRrQG02OHDmyUkuD8v/36tWLli1b+uoZ1QmtBl9Hj1h9DWDFMziKZsOIlZF92xAS46LeTy1KSEhg8ODBQOWfA0VRGDZsGLGxsX6KTgj/8Gudm7vvvpsZM2Ywe/Zsh9o2MTEx9jelm266idTUVKZOnQrAc889R//+/Wnbti05OTn85z//4ccff2T9+vUe/XUidW4Cz8mTJ9m+fbtDEb/OnTuf3+trKdSL3W38GApOQUxz6H0HdLkOTHWs74utFLZ/oxfxyz4IEYl6+4Uet0BIlOPY4hz9OW35vEIRv79Bp7+AwfFMRX5+Ptu3b+fw4cMORfwa0h8SJzbMZfvmdZw2h2LERstGBjpdeDlRqV6eEawFmZmZ9iJ+oCf/nTt3drtBQ4j6ot60X3B2nfiTTz7hlltuAWDo0KG0bNmS6dOnA/DAAw/www8/cPLkSRo1akTv3r154YUXPC4qJsmNEEIIUf/Um+TGHyS5CTCaprceWP2GXsvFYNI7VQ94EJrWgUW/tSH/JMyZBPvng2rVL0+lXgCXvQfJ3fwdnRBC1AhJblyQ5CaAaBrMvRfW/K9C6wH0BEdTYcKn0P0G/8boa9kH4X8d9do7lShw4zxoc0mthyWEEDWt3rRfEOK87JylJzbguM1ZterJzexbISew2wTw2QgniQ2ABl+NB9VJBWMhhAhQktyI+mv1G66r7mqavsg2UJ3eDdkHXI+xFsOmT2snHiGEqCMkuRH11/G1rgvTaTY4XgfaJPjK3l88G7dvrm/jEEKIOkaSG1F/OStIZ6eAMbhWQvELT59bIH8NhBCiCpLciPqr3Rj3CU7b0bUTiz90/atn43r/zbdxCCFEHSPJjai/BjyoLxyuimKEsEbQ/abajak2hcfp7RVciUqFlhfVTjxCCFFHSHIj6q+mF+jbvQ2mCguLy1oPhMbAjfMhNMC3+984H2KdtC0JjobbV9VuPEIIUQe4W7QgRN3W/Qa94/X69/XFw8Zg/VJU9xv1BCfQmYLh3n2w/gNY+SoUntZbLvS4BS56Qn9cCCEaGCniJ5zSbDbU7GyU0FAMkZH+Dqfm2EqhOAuCIyE4wt/R1ByrBUqyISQagqrRMFRUTVWh+IyeODeEhFmIOsqb9285cyMqUQsLKfjf2xR+/gVqdjYAIYMGEXXfvYQMGujn6M5DURb8OQU2fAjmPL1NQbsxcNGT9btVQ94JWPoCbJqu17VRjNDpKhjyFCR19nd09ZfVDCtfg9VvQUG6fl/TATD4UUgb59/YhBAuyZkb4UAtLCTzqqsp3b7dsbKt0QiqSqM33yD8yiv8F2B1FWbCRwP0dgUVa+MoRlAUuP4naDfKf/FVV84R+LCf/vw069n7FaN+puHmxdCsv//iq6+sFvhyDBz63XHRumLQPx71BvS/13/xCdEASfsFUW0Fb79TObEBsNlA08h+6GHUnBy/xHZeFj1WObEB/WNVhR8m6n+p1ze/3gNF5yQ2oD8vm0V/XtJ+wXvr34ODiyvvxiv/eN4DkH2o1sMSQnhGkhthp9lsFH72ues3Q4uFou9/qL2gaoI5H7Z87qKasaqvwdk5q1bDOm95x2HPz3ovrapoNr09w6E/ajWsgFDes8wpRb+8KYSokyS5EXZqVpZ9jY1TRiOlu/fUTkA1JecQWEtcjzEEwekdtRJOjcncBbi5qqwY6t/z8jdVhTN7cPm11WxwenuthSSE8I4kN8JOCfNsh40SFurjSGpYkAc7ojQVgsJ9H0tNCtTn5W+KAiY33+OKUb6uQtRhktwIO0NkJMGDBumLh52xWgkbO6b2gqoJjVpBYif0An9OaDboWM8WSqf0gYjGrscYTNB+bO3EEygUBTpe6bq1h2bTxwgh6iRJboSD6Pvudb7mxmgkuE8fgvv2rd2gzpeiwNBncHqZQTHqb1QJabUZ1fkzmmDIky4GKNDrDoh0kwCJygY+jL3a9bkUEyR0gLTLazsqIYSHJLkRDkIGDaTRm29AcHDZ6XmTfgOCe/Yk7pOPURQXZ0Dqqs5Xw+g3y1o1GPR/y/8ybzsKrvjMv/FVV9+7YcjTgKInaRWfV7eJMOp1f0ZXfzXpAdf9WFbkUXH8uiZ2hBsXgDHIjwEKIVyROjeiSmpODkXffU/pnr0oYaGEjRlN8AUX1M/EpqKCU7DpU8jap1eb7XwtpPbxd1TnL/eo/rxyD0N4op7YSAG/82fOh60zIH0DGEOg/WXQegQY5O9CIWqbN+/fktwIIeq0/GM7Kco6QXhcClFNO9bMpFYLZGzT184kdgqsNhxCBChpvyCEqPdObVnIqlUrOUX5mqH9JPMd/QcOIqnLxdWb1GaFZVNh1et6bSPQd531uRMufkF6cgkRIOTcqhCizknf8Bs/rdzDKS3R4f5TWiJzlu8ifdM87yfVNJh1I/z+9NnEBqC0UE92vhiln9ERQtR7ktwIIeoUTVVZumYTKgZ98XfFxxQDKgb+XLUBzdu2EgcWwbaZVLlrTlPh8FK9krUQot6T5EYIUadkbFtMriGuUmJjpxjIMcRzescS7ybe8IHr2jWKAda9592cQog6SZIbIUSdkp953KNxeaePeTfxmT3O+3CBfvYme793cwoh6iRJboQQdUpweJRH40LDvdztGJ7o/GxQubA47+YUQtRJktwIIeqU1J6jCFGLXI4JUQtp0uNS7ybuNlE/O+OMYoDuN3k3pxCiTpLkRghRpxhDwunTPMTlmL4twjGGeLltu/O1kNCx6nU3ilE/s9PnLu/mFELUSZLcCCHqnM5j/k7/ZDNGzQKahqLZQNMwahYGNLHQafSd3k8aFAq3/A7NBukfK0b9Bnohv9v+hIhE558vhKg3pEKxEKLOsuSd4eCqHykqyCM8KppW/SYQHB1//hOnb4IDC/UKxc0GQfNBei81IUSdJe0XXJDkRgghhKh/vHn/lstSQgghhAgoktwIIYQQIqBIciOEEEKIgCLJjRBCCCECiiQ3QgghhAgoktwIIYQQIqBIciOEEEKIgCLJjRBCCCECiiQ3QgghhAgoktwIIYQQIqBU0R5XiABmNcPuOXBmL4TGQMcrIaqJv6MSQghRgyS5EQ3HzlkwexKUZIPBBJoKc++DPnfBqNfBKD8OQggRCPx6WWrq1Kn07duXqKgokpKSmDBhArt373b7ed9++y0dOnQgNDSUrl278uuvv9ZCtKJeO7AIvvkLlOToH6tWPbnRbLD2HfjtPr+GJ4QQoub4NblZsmQJkydPZtWqVSxYsIDS0lIuvfRSCgsLnX7OihUruP7665k0aRIbN25kwoQJTJgwgW3bttVi5KLeWfxE2X+0Kh7UYO00yD1amxEJIYTwEUXTtKp+2/vF6dOnSUpKYsmSJVx00UVVjrn22mspLCzk559/tt/Xv39/evTowbvvvltpvNlsxmw22z/Oy8ujWbNmHrVMFwEi9yj8t7nrMYoBLvkPDHywdmISQgjhlby8PGJiYjx6/65Tu6Vyc3MBiIuLczpm5cqVjBgxwuG+kSNHsnLlyirHT506lZiYGPutWbNmNRewqB+Ks92PUYxQnOX7WIQQQvhcnUluVFXl/vvvZ9CgQXTp0sXpuJMnT9K4cWOH+xo3bszJkyerHP/oo4+Sm5trvx09KpceGpzopnry4opqhUatayceIYQQPlVntodMnjyZbdu2sWzZshqdNyQkhJCQkBqdU9Qz4XHQ6SrY+YOexFQlKAw6X1O7cQkhhPCJOnHm5p577uHnn3/m999/p2nTpi7HJicnc+rUKYf7Tp06RXJysi9DFPXdiJcgJKaKMziK/s+YtyEkstbDEkIIUfP8mtxomsY999zDrFmzWLx4Ma1atXL7OQMGDGDRokUO9y1YsIABAwb4KkwRCBq1gjvWQNo4ffFwucSOcO0P0PMWv4UmhBCiZvn1stTkyZOZMWMGs2fPJioqyr5uJiYmhrCwMABuuukmUlNTmTp1KgD33XcfQ4YM4dVXX2Xs2LHMnDmTdevW8f777/vteYh6Iq41XDcLCk5BziH9TE5CGiiKvyMTQghRg/x65mbatGnk5uYydOhQmjRpYr99/fXX9jFHjhwhPT3d/vHAgQOZMWMG77//Pt27d+e7777jxx9/dLkIWQgHkY2haT9I7CCJjRBCBKA6VeemNnizT14IIYQQdUO9rXMjhBBCCHG+JLkRQgghRECR5EYIIYQQAUWSGyGEEEIEFEluhBBCCBFQJLkRQgghRECR5EYIIYQQAUWSGyGEEEIEFEluhBBCCBFQJLkRQgghRECR5EYIIYQQAUWSGyGEEEIEFEluhBBCCBFQJLkRQgghRECR5EYIIYQQAUWSGyGEEEIEFEluhBBCCBFQJLkRQgghRECR5EYIIYQQAUWSGyGEEEIEFEluhBBCCBFQJLkRQgghRECR5EYIIYQQAUWSGyGEEEIEFEluhBBCCBFQJLkRQgghRECR5EYIIYQQAUWSGyGEEEIEFEluhBBCCBFQJLkRQgghRECR5EYIIYQQAUWSGyGEEEIEFEluhBBCCBFQJLkRQgghRECR5EYIIYQQAUWSGyGEEEIEFEluhBBCCBFQJLkRQgghRECR5EYIIYQQAUWSGyGEEEIEFEluhBBCCBFQJLkRQgghRECR5EYIIYQQAcWvyc3SpUsZN24cKSkpKIrCjz/+6HL8H3/8gaIolW4nT56snYCFEEIIUef5NbkpLCyke/fuvP3221593u7du0lPT7ffkpKSfBShEEIIIeobkz8PPnr0aEaPHu315yUlJREbG1vzAQkhhBCi3vNrclNdPXr0wGw206VLF5555hkGDRrkdKzZbMZsNts/zs3NBSAvL8/ncQohhBCiZpS/b2ua5nZsvUpumjRpwrvvvkufPn0wm818+OGHDB06lNWrV9OrV68qP2fq1Kk8++yzle5v1qyZr8MVQgghRA3Lz88nJibG5RhF8yQFqgWKojBr1iwmTJjg1ecNGTKE5s2b8/nnn1f5+LlnblRVJSsri/j4eBRFOZ+QHeTl5dGsWTOOHj1KdHR0jc0rfENer/pDXqv6RV6v+qU+vV6appGfn09KSgoGg+slw/XqzE1VLrjgApYtW+b08ZCQEEJCQhzu8+V6nejo6Dr/DSLOkter/pDXqn6R16t+qS+vl7szNuXqfZ2bTZs20aRJE3+HIYQQQog6wq9nbgoKCti3b5/944MHD7Jp0ybi4uJo3rw5jz76KMePH+ezzz4D4PXXX6dVq1Z07tyZkpISPvzwQxYvXsz8+fP99RSEEEIIUcf4NblZt24dw4YNs3/84IMPAnDzzTczffp00tPTOXLkiP1xi8XCP/7xD44fP054eDjdunVj4cKFDnP4S0hICE8//XSlS2CibpLXq/6Q16p+kderfgnU16vOLCgWQgghhKgJ9X7NjRBCCCFERZLcCCGEECKgSHIjhBBCiIAiyY0QQgghAookNzXopZdeQlEU7r//fn+HIqrwzDPPoCiKw61Dhw7+Dku4cPz4cW644Qbi4+MJCwuja9eurFu3zt9hiSq0bNmy0s+XoihMnjzZ36GJKthsNp588klatWpFWFgYbdq04fnnn/eob1N9UO8rFNcVa9eu5b333qNbt27+DkW40LlzZxYuXGj/2GSSH4G6Kjs7m0GDBjFs2DDmzp1LYmIie/fupVGjRv4OTVRh7dq12Gw2+8fbtm3jkksu4eqrr/ZjVMKZl19+mWnTpvHpp5/SuXNn1q1bx6233kpMTAz33nuvv8M7b/KbvQYUFBQwceJEPvjgA1544QV/hyNcMJlMJCcn+zsM4YGXX36ZZs2a8cknn9jva9WqlR8jEq4kJiY6fPzSSy/Rpk0bhgwZ4qeIhCsrVqxg/PjxjB07FtDPvH311VesWbPGz5HVDLksVQMmT57M2LFjGTFihL9DEW7s3buXlJQUWrduzcSJEx2KRIq6Zc6cOfTp04err76apKQkevbsyQcffODvsIQHLBYLX3zxBbfddluNNigWNWfgwIEsWrSIPXv2ALB582aWLVvG6NGj/RxZzZAzN+dp5syZbNiwgbVr1/o7FOFGv379mD59OmlpaaSnp/Pss88yePBgtm3bRlRUlL/DE+c4cOAA06ZN48EHH+Sxxx5j7dq13HvvvQQHB3PzzTf7Ozzhwo8//khOTg633HKLv0MRTvzrX/8iLy+PDh06YDQasdlsTJkyhYkTJ/o7tBohyc15OHr0KPfddx8LFiwgNDTU3+EINyr+RdKtWzf69etHixYt+Oabb5g0aZIfIxNVUVWVPn368OKLLwLQs2dPtm3bxrvvvivJTR330UcfMXr0aFJSUvwdinDim2++4csvv2TGjBl07tyZTZs2cf/995OSkhIQP1+S3JyH9evXk5GRQa9evez32Ww2li5dyv/+9z/MZjNGo9GPEQpXYmNjad++vUPzVlF3NGnShE6dOjnc17FjR77//ns/RSQ8cfjwYRYuXMgPP/zg71CECw8//DD/+te/uO666wDo2rUrhw8fZurUqZLcNHTDhw9n69atDvfdeuutdOjQgUceeUQSmzquoKCA/fv3c+ONN/o7FFGFQYMGsXv3bof79uzZQ4sWLfwUkfDEJ598QlJSkn2hqqibioqKMBgcl90ajUZUVfVTRDVLkpvzEBUVRZcuXRzui4iIID4+vtL9wv8eeughxo0bR4sWLThx4gRPP/00RqOR66+/3t+hiSo88MADDBw4kBdffJFrrrmGNWvW8P777/P+++/7OzThhKqqfPLJJ9x8881SZqGOGzduHFOmTKF58+Z07tyZjRs38tprr3Hbbbf5O7QaId99osE4duwY119/PWfOnCExMZELL7yQVatWVdrCKuqGvn37MmvWLB599FGee+45WrVqxeuvvx4wCx4D0cKFCzly5EjAvEEGsrfeeosnn3ySu+++m4yMDFJSUrjzzjt56qmn/B1ajVC0QClHKIQQQgiB1LkRQgghRICR5EYIIYQQAUWSGyGEEEIEFEluhBBCCBFQJLkRQgghRECR5EYIIYQQAUWSGyGEEEIEFEluhBBCCBFQJLkRQvjcLbfcwoQJE5w+Pn36dGJjY2stHndatmzJ66+/7vXnnTlzhqSkJA4dOlTjMZXLzMwkKSmJY8eO+ewYQtR3ktwIIRqsmk6qpkyZwvjx42nZsmWNzXmuhIQEbrrpJp5++mmfHUOI+k6SGyGEqAFFRUV89NFHTJo0yefHuvXWW/nyyy/Jysry+bGEqI8kuREiwH333Xd07dqVsLAw4uPjGTFiBIWFhfbHP/zwQzp27EhoaCgdOnTgnXfesT926NAhFEVh5syZDBw4kNDQULp06cKSJUvsY2w2G5MmTaJVq1aEhYWRlpbGG2+8cd5xz549m169ehEaGkrr1q159tlnsVqt9scVReHDDz/kiiuuIDw8nHbt2jFnzhyHOebMmUO7du0IDQ1l2LBhfPrppyiKQk5ODn/88Qe33norubm5KIqCoig888wz9s8tKiritttuIyoqiubNm7vtRv7rr78SEhJC//79He7fvn07l112GdHR0URFRTF48GD2798PnL1c9+KLL9K4cWNiY2N57rnnsFqtPPzww8TFxdG0aVM++eQThzk7d+5MSkoKs2bNqs6XVojApwkhAtaJEyc0k8mkvfbaa9rBgwe1LVu2aG+//baWn5+vaZqmffHFF1qTJk2077//Xjtw4ID2/fffa3Fxcdr06dM1TdO0gwcPaoDWtGlT7bvvvtN27Nih3X777VpUVJSWmZmpaZqmWSwW7amnntLWrl2rHThwQPviiy+08PBw7euvv7bHcfPNN2vjx493Gucnn3yixcTE2D9eunSpFh0drU2fPl3bv3+/Nn/+fK1ly5baM888Yx9THteMGTO0vXv3avfee68WGRmpnTlzRtM0TTtw4IAWFBSkPfTQQ9quXbu0r776SktNTdUALTs7WzObzdrrr7+uRUdHa+np6Vp6err969KiRQstLi5Oe/vtt7W9e/dqU6dO1QwGg7Zr1y6nz+Hee+/VRo0a5XDfsWPHtLi4OO3KK6/U1q5dq+3evVv7+OOP7fPcfPPNWlRUlDZ58mRt165d2kcffaQB2siRI7UpU6Zoe/bs0Z5//nktKChIO3r0qMPc1157rXbzzTc7jUeIhkySGyEC2Pr16zVAO3ToUJWPt2nTRpsxY4bDfc8//7w2YMAATdPOJjcvvfSS/fHS0lKtadOm2ssvv+z0uJMnT9auuuoq+8feJjfDhw/XXnzxRYcxn3/+udakSRP7x4D2xBNP2D8uKCjQAG3u3LmapmnaI488onXp0sVhjscff9ye3FR13HItWrTQbrjhBvvHqqpqSUlJ2rRp05w+h/Hjx2u33Xabw32PPvqo1qpVK81isVT5OTfffLPWokULzWaz2e9LS0vTBg8ebP/YarVqERER2ldffeXwuQ888IA2dOhQp/EI0ZCZ/HbKSAjhc927d2f48OF07dqVkSNHcumll/KXv/yFRo0aUVhYyP79+5k0aRJ33HGH/XOsVisxMTEO8wwYMMD+f5PJRJ8+fdi5c6f9vrfffpuPP/6YI0eOUFxcjMVioUePHtWOe/PmzSxfvpwpU6bY77PZbJSUlFBUVER4eDgA3bp1sz8eERFBdHQ0GRkZAOzevZu+ffs6zHvBBRd4HEPFuRVFITk52T53VYqLiwkNDXW4b9OmTQwePJigoCCnn9e5c2cMhrMrBBo3bkyXLl3sHxuNRuLj4ysdOywsjKKiIo+fjxANiSQ3QgQwo9HIggULWLFiBfPnz+ett97i8ccfZ/Xq1fYE4YMPPqBfv36VPs9TM2fO5KGHHuLVV19lwIABREVF8Z///IfVq1dXO+6CggKeffZZrrzyykqPVUwgzk0aFEVBVdVqH7cib+dOSEggOzvb4b6wsLBqHceTY2dlZZGYmOh2fiEaIllQLESAUxSFQYMG8eyzz7Jx40aCg4OZNWsWjRs3JiUlhQMHDtC2bVuHW6tWrRzmWLVqlf3/VquV9evX07FjRwCWL1/OwIEDufvuu+nZsydt27a1L5itrl69erF79+5KcbVt29bhLIcraWlprFu3zuG+tWvXOnwcHByMzWY7r1jL9ezZkx07djjc161bN/78809KS0tr5BgVbdu2jZ49e9b4vEIEAkluhAhgq1ev5sUXX2TdunUcOXKEH374gdOnT9sTk2effZapU6fy5ptvsmfPHrZu3conn3zCa6+95jDP22+/zaxZs9i1axeTJ08mOzub2267DYB27dqxbt065s2bx549e3jyyScrJRHeeuqpp/jss8949tln2b59Ozt37mTmzJk88cQTHs9x5513smvXLh555BH27NnDN998w/Tp0wE94QO9WF9BQQGLFi0iMzPzvC7zjBw5ku3btzucvbnnnnvIy8vjuuuuY926dezdu5fPP/+c3bt3V/s4oO/kWr9+PZdeeul5zSNEoJLkRogAFh0dzdKlSxkzZgzt27fniSee4NVXX2X06NEA3H777Xz44Yd88skndO3alSFDhjB9+vRKZ25eeuklXnrpJbp3786yZcuYM2cOCQkJgJ5EXHnllVx77bX069ePM2fOcPfdd59X3CNHjuTnn39m/vz59O3bl/79+/Pf//6XFi1aeDxHq1at+O677/jhhx/o1q0b06ZN4/HHHwcgJCQEgIEDB3LXXXdx7bXXkpiYyL///e9qx9y1a1d69erFN998Y78vPj6exYsXU1BQwJAhQ+jduzcffPCByzU4npg9ezbNmzdn8ODB5zWPEIFK0TRN83cQQoi66dChQ7Rq1YqNGzee1wLhumLKlCm8++67HD161Cfz//LLLzz88MNs27bN48tn1dG/f3/uvfde/vrXv/rsGELUZ7KgWAgRsN555x369u1LfHw8y5cv5z//+Q/33HOPz443duxY9u7dy/Hjx2nWrJlPjpGZmcmVV17J9ddf75P5hQgEcuZGCOFUfT9z88ADD/D111+TlZVF8+bNufHGG3n00UcxmeTvOiECmSQ3QgghhAgosqBYCCGEEAFFkhshhBBCBBRJboQQQggRUCS5EUIIIURAkeRGCCGEEAFFkhshhBBCBBRJboQQQggRUCS5EUIIIURA+X+fNEFp1YJzUQAAAABJRU5ErkJggg==", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "from sklearn import datasets\n", + "\n", + "iris = datasets.load_iris()\n", + "\n", + "X = iris.data[:,:2]\n", + "Y = iris.target\n", + "\n", + "x_min, x_max = X[:, 0].min() - .5, X[:, 0].max() + .5\n", + "y_min, y_max = X[:, 1].min() - .5, X[:, 1].max() + .5\n", + "\n", + "plt.figure()\n", + "plt.clf()\n", + "\n", + "# Plot the training points\n", + "plt.scatter(X[:, 0], X[:, 1], c=Y, cmap=plt.cm.Set1)\n", + "plt.xlabel(iris.feature_names[0])\n", + "plt.ylabel(iris.feature_names[1])\n", + "\n", + "plt.xlim(x_min, x_max)\n", + "plt.ylim(y_min, y_max)\n", + "plt.show()" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "collapsed": true, + "jupyter": { + "outputs_hidden": true + } + }, + "source": [ + "The iris data set contains 3 classes of iris flowers that should be classified according to their sepal width \n", + "and sepal length. " + ] + }, + { + "cell_type": "markdown", + "metadata": { + "collapsed": true, + "jupyter": { + "outputs_hidden": true + } + }, + "source": [ + "We split the data set into 67% training data and 33% test data. I.e., we have 100 training and 50 test points. " + ] + }, + { + "cell_type": "code", + "execution_count": 11, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "((100, 2), (50, 2))" + ] + }, + "execution_count": 11, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "from sklearn.model_selection import train_test_split\n", + "\n", + "X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.33, random_state=42)\n", + "X_train.shape, X_test.shape" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "collapsed": true, + "jupyter": { + "outputs_hidden": true + } + }, + "source": [ + "### Training a Decision Tree" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "collapsed": true, + "jupyter": { + "outputs_hidden": true + } + }, + "source": [ + "We will again use the `sklearn` package. A decision tree can be generated by:" + ] + }, + { + "cell_type": "code", + "execution_count": 12, + "metadata": {}, + "outputs": [], + "source": [ + "from sklearn import tree\n", + "\n", + "decTree = tree.DecisionTreeClassifier(min_samples_leaf=2, max_depth=None)" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "collapsed": true, + "jupyter": { + "outputs_hidden": true + } + }, + "source": [ + "We can set the same properties as for a regression tree. Similarly, we can train the tree:" + ] + }, + { + "cell_type": "code", + "execution_count": 13, + "metadata": {}, + "outputs": [], + "source": [ + "decTree = decTree.fit(X_train, Y_train)" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "collapsed": true, + "jupyter": { + "outputs_hidden": true + } + }, + "source": [ + "We can use the trained tree for prediction by:" + ] + }, + { + "cell_type": "code", + "execution_count": 14, + "metadata": {}, + "outputs": [], + "source": [ + "y_predict = decTree.predict(X_test)" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "collapsed": true, + "jupyter": { + "outputs_hidden": true + } + }, + "source": [ + "#### Exercise 2: Decision boundary analysis\n", + "In this excercise we want to train a decision tree for different values of `min_samples_leaf`. After training the tree,\n", + "we will plot the decision boundary of the learned tree with the existing python code.\n", + "- Plot the decision boundary for different `number of min_samples_leaf`.\n", + "- Can you observe a qualitatitve difference between the learned classifiers?\n", + "- Which value of `min_samples_leaf` would you choose?" + ] + }, + { + "cell_type": "code", + "execution_count": 15, + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "import matplotlib.pyplot as plt\n", + "from sklearn.tree import DecisionTreeClassifier\n", + "import sklearn.metrics as metrics\n", + "\n", + "n_classes = 3\n", + "plot_step = 0.02\n", + "\n", + "######################################\n", + "##### CHANGE THE FOLLOWING LINE: #####\n", + "######################################\n", + "clf = DecisionTreeClassifier(min_samples_leaf=5)\n", + "\n", + "clf = clf.fit(X_train, Y_train) \n", + "\n", + "# create a grid for the two input dimensions\n", + "xx, yy = np.meshgrid(np.arange(x_min, x_max, plot_step),\n", + " np.arange(y_min, y_max, plot_step))\n", + "\n", + "plt.figure()\n", + "Z = clf.predict(np.c_[xx.ravel(), yy.ravel()])\n", + "Z = Z.reshape(xx.shape)\n", + "cs = plt.contourf(xx, yy, Z, cmap=plt.cm.Pastel1)\n", + "\n", + "plt.scatter(X_train[:, 0], X_train[:, 1], c=Y_train, cmap=plt.cm.Set1)\n", + "\n", + "plt.xlabel(iris.feature_names[0])\n", + "plt.ylabel(iris.feature_names[1])\n", + "plt.axis(\"tight\")\n", + "plt.show()\n" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "collapsed": true, + "jupyter": { + "outputs_hidden": true + } + }, + "source": [ + "### Evaluating Decision Trees\n", + "For evaluating the decision tree, we can compute the ratio of correctly \n", + "classified samples on the test set. This metric is called `accuracy_score` in sklearn. " + ] + }, + { + "cell_type": "code", + "execution_count": 16, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "(0.8, 0.7)" + ] + }, + "execution_count": 16, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "from sklearn import metrics\n", + "\n", + "train_accuracy = metrics.accuracy_score(Y_train, clf.predict(X_train))\n", + "test_accuracy = metrics.accuracy_score(Y_test, clf.predict(X_test))\n", + "\n", + "train_accuracy, test_accuracy" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "collapsed": true, + "jupyter": { + "outputs_hidden": true + } + }, + "source": [ + "#### Exercise 2b: Qualitative analysis\n", + "\n", + "For the values of `min_samples_leaf = [1,2,3,5,7,10, 15, 20, 50]`, compute the `accuracy_score` on the train and on \n", + "the test set. Plot both accuracy scores as a function of `min_samples_leaf`. " + ] + }, + { + "cell_type": "code", + "execution_count": 17, + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "from sklearn.tree import DecisionTreeClassifier\n", + "from sklearn.model_selection import cross_val_score\n", + "\n", + "minSamples = [1,2,3,5,7,10, 15, 20, 50]\n", + "train_accuracy = np.zeros((len(minSamples),1))\n", + "test_accuracy = np.zeros((len(minSamples),1))\n", + "\n", + "\n", + "for i in range(0,len(minSamples)):\n", + " \n", + " ######################################\n", + " ##### YOUR CODE GOES HERE: ###########\n", + " ######################################\n", + " # declare a decisionTreeClassifier taking as input the minSamples:\n", + " clf = DecisionTreeClassifier(min_samples_leaf=minSamples[i]) \n", + " \n", + " # fit it to your data:\n", + " clf = clf.fit(X_train, Y_train) \n", + " \n", + " # fit it to your data:\n", + "\n", + " # store the train and test accuracy:\n", + " train_accuracy[i] = metrics.accuracy_score(Y_train, clf.predict(X_train))\n", + " test_accuracy[i] = metrics.accuracy_score(Y_test, clf.predict(X_test))\n", + " \n", + "\n", + "plt.figure()\n", + "plt.plot(minSamples, train_accuracy, 'b')\n", + "plt.plot(minSamples, test_accuracy, 'g')\n", + "plt.xlabel('min_samples_per_leaf')\n", + "plt.ylabel('accuracy')\n", + "plt.legend(('training set', 'test set'))\n", + "plt.savefig('classification_minSamples.png')\n", + "plt.show()\n" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "collapsed": true, + "jupyter": { + "outputs_hidden": true + } + }, + "source": [ + "We can again see effects of overfitting and underfitting. For `min_samples_leaf < 10`, we have a high training accuracy but the test accuracy degrades. Thats **overfitting**. \n", + "For `min_samples_leaf > 20`, test and training performance degrades which is a sign of **underfitting**. " + ] + }, + { + "cell_type": "markdown", + "metadata": { + "collapsed": true, + "jupyter": { + "outputs_hidden": true + } + }, + "source": [ + "### Decision Forests" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "collapsed": true, + "jupyter": { + "outputs_hidden": true + } + }, + "source": [ + "We will again use the sklearn package. A decision forest can be generated by:" + ] + }, + { + "cell_type": "code", + "execution_count": 18, + "metadata": {}, + "outputs": [], + "source": [ + "from sklearn import ensemble\n", + "\n", + "decForest = ensemble.RandomForestClassifier(n_estimators=10, min_samples_leaf=2, max_depth=None)" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "collapsed": true, + "jupyter": { + "outputs_hidden": true + } + }, + "source": [ + "We can set the same properties as for a decision forest (including number of trees by n_estimator). \n", + "Similarly, we can train the forest" + ] + }, + { + "cell_type": "code", + "execution_count": 19, + "metadata": {}, + "outputs": [], + "source": [ + "decForest = decForest.fit(X_train, Y_train)" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "collapsed": true, + "jupyter": { + "outputs_hidden": true + } + }, + "source": [ + "We can use the trained tree for prediction by:" + ] + }, + { + "cell_type": "code", + "execution_count": 20, + "metadata": {}, + "outputs": [], + "source": [ + "y_predict = decForest.predict(X_test)" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "collapsed": true, + "jupyter": { + "outputs_hidden": true + } + }, + "source": [ + "#### Exercise 3a: Qualitative analysis of `n_estimators`\n", + "In this excercise we want to train a decision forests for different values of `n_estimators`. After training the forest, we will plot the decision boundary of the learned forest with the existing python code. \n", + "- Plot the decision boundary for different number of `n_estimators` and use `min_samples_leaf = 10`.\n", + "- Can you observe a qualitatitve difference between the learned classifiers?\n", + "- Execute your code several times. Can you observe a difference between the executions? If yes, why?" + ] + }, + { + "cell_type": "code", + "execution_count": 21, + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "import matplotlib.pyplot as plt\n", + "from sklearn.ensemble import RandomForestClassifier\n", + "import sklearn.metrics as metrics\n", + "\n", + "n_classes = 3\n", + "plot_colors = \"bry\"\n", + "plot_step = 0.02\n", + "\n", + "\n", + "######################################\n", + "##### CHANGE THE FOLLOWING LINE: #####\n", + "######################################\n", + "clf = ensemble.RandomForestClassifier(n_estimators=30, min_samples_leaf=10, max_depth=None)\n", + "\n", + "######################################\n", + "##### YOUR CODE GOES HERE: ###########\n", + "######################################\n", + "# fit the model to the data:\n", + "clf = clf.fit(X_train, Y_train)\n", + "\n", + "\n", + "\n", + "xx, yy = np.meshgrid(np.arange(x_min, x_max, plot_step),\n", + " np.arange(y_min, y_max, plot_step))\n", + "\n", + "plt.figure()\n", + "\n", + "Z = clf.predict(np.c_[xx.ravel(), yy.ravel()])\n", + "Z = Z.reshape(xx.shape)\n", + "\n", + "cs = plt.contourf(xx, yy, Z, cmap=plt.cm.Pastel1)\n", + "\n", + "plt.scatter(X_train[:, 0], X_train[:, 1], c=Y_train, cmap=plt.cm.Set1)\n", + "\n", + "plt.xlabel(iris.feature_names[0])\n", + "plt.ylabel(iris.feature_names[1])\n", + "plt.axis(\"tight\")\n", + "plt.savefig('classification_forest.png')\n", + "plt.show()" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "#### Exercise 3b: Quantitative analysis of `min_samples_leaf`\n", + "\n", + "The below cell similarly examines the effect of varying `min_samples_leaf` for RandomForests (similar to our [previous analysis on trees](#Exercise-2b:-Qualitative-analysis)):\n", + "- Taking into account what you know about overfitting and underfitting, which value of `min_sample_leaf` would you chose based on the below graph?\n", + "- If you run the below cells of code multiple times, does the graph change? Why? Does that change your choice?" + ] + }, + { + "cell_type": "code", + "execution_count": 28, + "metadata": {}, + "outputs": [], + "source": [ + "from sklearn.ensemble import RandomForestClassifier\n", + "\n", + "minSamples = [1,3,5,7,10, 15, 20, 30]\n", + "\n", + "numTrials = 10\n", + "train_accuracy_single = np.empty((len(minSamples),numTrials))\n", + "test_accuracy_single = np.empty((len(minSamples),numTrials))\n", + "\n", + "train_accuracy_mean = np.empty((len(minSamples),1))\n", + "test_accuracy_mean = np.empty((len(minSamples),1))\n", + "\n", + "train_accuracy_std = np.empty((len(minSamples),1))\n", + "test_accuracy_std = np.empty((len(minSamples),1))\n", + "\n", + "\n", + "for i in range(0,len(minSamples)):\n", + " clf = RandomForestClassifier(n_estimators=100, min_samples_leaf=minSamples[i])\n", + "\n", + " for j in range(0, numTrials):\n", + " clf.fit(X_train, Y_train)\n", + "\n", + " train_accuracy_single[i,j] = metrics.accuracy_score(Y_train, clf.predict(X_train))\n", + " test_accuracy_single[i, j] = metrics.accuracy_score(Y_test, clf.predict(X_test))\n", + "\n", + " train_accuracy_mean[i] = np.mean(train_accuracy_single[i,:])\n", + " train_accuracy_std[i] = np.std(train_accuracy_single[i,:])\n", + " \n", + " test_accuracy_mean[i] = np.mean(test_accuracy_single[i,:])\n", + " test_accuracy_std[i] = np.std(test_accuracy_single[i,:])\n", + "\n", + " " + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "_Note: the errorbar function of matplotlib has undergone recent changes. If the below code does not work for you, you can use the commented lines which do not plot the error bars instead, they should definitely work._" + ] + }, + { + "cell_type": "code", + "execution_count": 29, + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + " \n", + "plt.figure()\n", + "\n", + "plt.errorbar(minSamples, train_accuracy_mean.squeeze(), yerr = train_accuracy_std.ravel(), label='train')\n", + "plt.errorbar(minSamples, test_accuracy_mean.squeeze(), yerr = test_accuracy_std.ravel(), label='test')\n", + "#plt.plot(minSamples, train_accuracy_mean, label='train')\n", + "#plt.plot(minSamples, test_accuracy_mean, label='test')\n", + "plt.legend()\n", + "plt.xlabel('min_samples_per_leaf')\n", + "plt.ylabel('accuracy')\n", + "plt.show()" + ] + }, + { + "cell_type": "markdown", + "metadata": { + "collapsed": true, + "jupyter": { + "outputs_hidden": true + } + }, + "source": [ + "#### Exercise 3c: Quantitative analysis of `n_estimators`\n", + "\n", + "Test the algorithm for a different number of trees, i.e., [1,5,10,20,40,60,100]. Repeat each experiment 10 times and average over the performance values due to the randomness. Evaluate the average accuracy on the training and on the test set for the given number of trees (n_estimators). \n", + "Use `min_samples_leaf = 15`.\n", + "- Which value of `min_samples_leaf` would you chose? Has your choice changed from [Exercise 3a](#Exercise-3a:-Qualitative-analysis-of-n_estimators)?\n", + "- Can you comment on the behaviour of the classifier for `n_estimator < 50`?\n", + "- Can you comment on the behaviour of the classifier for `n_estimator > 200`?\n", + "\n", + "_Note: the below cells might take a few moments to execute._" + ] + }, + { + "cell_type": "code", + "execution_count": 24, + "metadata": {}, + "outputs": [], + "source": [ + "from sklearn.ensemble import RandomForestClassifier\n", + "from sklearn.model_selection import cross_val_score\n", + "\n", + "numTrees = [1,5,10,20,40,60,100, 150, 200, 250, 300, 400]\n", + "\n", + "numTrials = 10\n", + "# store test/train accuracy for every trial\n", + "train_accuracy_single = np.empty((len(numTrees),numTrials))\n", + "test_accuracy_single = np.empty((len(numTrees),numTrials))\n", + "\n", + "# store the mean test/train accuracy for all the 10 trials for a certain value of min_samples\n", + "train_accuracy_mean = np.empty((len(numTrees),1))\n", + "test_accuracy_mean = np.empty((len(numTrees),1))\n", + "\n", + "# store the standard deviation of test/train accuracy for all the 10 trials for a certain value of min_samples\n", + "train_accuracy_std = np.empty((len(numTrees),1))\n", + "test_accuracy_std = np.empty((len(numTrees),1))\n", + "\n", + "for i in range(0,len(numTrees)):\n", + " clf = RandomForestClassifier(n_estimators=numTrees[i], min_samples_leaf=15)\n", + "\n", + " for j in range(0, numTrials):\n", + " clf.fit(X_train, Y_train)\n", + "\n", + " train_accuracy_single[i,j] = metrics.accuracy_score(Y_train, clf.predict(X_train))\n", + " test_accuracy_single[i, j] = metrics.accuracy_score(Y_test, clf.predict(X_test))\n", + "\n", + " train_accuracy_mean[i] = np.mean(train_accuracy_single[i,:])\n", + " train_accuracy_std[i] = np.std(train_accuracy_single[i,:])\n", + " \n", + " test_accuracy_mean[i] = np.mean(test_accuracy_single[i,:])\n", + " test_accuracy_std[i] = np.std(test_accuracy_single[i,:])" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "_Note: the errorbar function of matplotlib has undergone recent changes. If the below code does not work for you, you can use the commented lines which do not plot the error bars instead, they should definitely work._" + ] + }, + { + "cell_type": "code", + "execution_count": 25, + "metadata": {}, + "outputs": [ + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "plt.figure()\n", + "\n", + "#plt.plot(numTrees, train_accuracy_mean, 'b', label='train')\n", + "#plt.plot(numTrees, test_accuracy_mean, 'g', label='test')\n", + "plt.errorbar(numTrees, train_accuracy_mean.squeeze(), yerr = train_accuracy_std.ravel(), label='train')\n", + "plt.errorbar(numTrees, test_accuracy_mean.squeeze(), yerr = test_accuracy_std.ravel(), label ='test')\n", + "plt.legend()\n", + "# Always label your axes\n", + "plt.xlabel(\"numTrees\")\n", + "plt.ylabel(\"accuracy\")\n", + "\n", + "\n", + "plt.savefig('classification_forest_numtrees.png')\n", + "plt.show()" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [] + } + ], + "metadata": { + "kernelspec": { + "display_name": "Python 3 (ipykernel)", + "language": "python", + "name": "python3" + }, + "language_info": { + "codemirror_mode": { + "name": "ipython", + "version": 3 + }, + "file_extension": ".py", + "mimetype": "text/x-python", + "name": "python", + "nbconvert_exporter": "python", + "pygments_lexer": "ipython3", + "version": "3.12.6" + } + }, + "nbformat": 4, + "nbformat_minor": 4 +}