5_KerasNueralNetwork.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.7.6"
    },
    "colab": {
      "name": "KerasNueralNetwork.ipynb",
      "provenance": [],
      "toc_visible": true,
      "include_colab_link": true
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/SiP-AI-ML/LessonMaterials/blob/master/KerasNueralNetwork.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2-cLl1LzuC1O",
        "colab_type": "text"
      },
      "source": [
        "# Basic Nueral Network Practice\n",
        "Today, we will built a nueral network in just 10 lines of code. We are going to use very simple data that we don't have to pre-process since it is already in a numerical format. We will go over how to build our model, how to initalize the model, how to train the model, and lastly how to validate its performance. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rr4uv9_ZuC1P",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "8f1a5222-f917-4434-ac24-5eb2ed86998b"
      },
      "source": [
        "from numpy import loadtxt\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dense\n",
        "import matplotlib.pyplot as plt"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Using TensorFlow backend.\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0AHKmodBuC1S",
        "colab_type": "text"
      },
      "source": [
        "We can now load our dataset.\n",
        "\n",
        "In this tutorial, we are going to use the Pima Indians onset of diabetes dataset. It describes patient medical record data for Pima Indians and whether they had an onset of diabetes within five years.\n",
        "\n",
        "As such, it is a binary classification problem (onset of diabetes as 1 or not as 0). All of the input variables that describe each patient are numerical. This makes it easy to use directly with neural networks that expect numerical input and output values, and ideal for our first neural network in Keras."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bsPt9P7WuC1T",
        "colab_type": "text"
      },
      "source": [
        "[Dataset Link](https://github.com/jbrownlee/Datasets/blob/master/pima-indians-diabetes.csv)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OrA8mBocuC1T",
        "colab_type": "text"
      },
      "source": [
        "We can now load the file as a matrix of numbers using the NumPy function loadtxt().\n",
        "\n",
        "There are eight input variables and one output variable (the last column). We will be learning a model to map rows of input variables (X) to an output variable (y), which we often summarize as y = f(X).\n",
        "\n",
        "The variables can be summarized as follows:\n",
        "\n",
        "Input Variables (X):\n",
        "\n",
        "Number of times pregnant\n",
        "Plasma glucose concentration a 2 hours in an oral glucose tolerance test\n",
        "Diastolic blood pressure (mm Hg)\n",
        "Triceps skin fold thickness (mm)\n",
        "2-Hour serum insulin (mu U/ml)\n",
        "Body mass index (weight in kg/(height in m)^2)\n",
        "Diabetes pedigree function\n",
        "Age (years)\n",
        "Output Variables (y):\n",
        "\n",
        "Class variable (0 or 1)\n",
        "Once the CSV file is loaded into memory, we can split the columns of data into input and output variables.\n",
        "\n",
        "The data will be stored in a 2D array where the first dimension is rows and the second dimension is columns, e.g. [rows, columns].\n",
        "\n",
        "We can split the array into two arrays by selecting subsets of columns using the standard NumPy slice operator or “:” We can select the first 8 columns from index 0 to index 7 via the slice 0:8. We can then select the output column (the 9th variable) via index 8."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ia2BsusvuC1U",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# load the dataset\n",
        "dataset = loadtxt('https://raw.githubusercontent.com/SiP-AI-ML/LessonMaterials/master/pima-indians-diabetes.csv', delimiter=',')\n",
        "# split into input (X) and output (y) variables\n",
        "X = dataset[:,0:8]\n",
        "y = dataset[:,8]"
      ],
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "x85MDccWuC1X",
        "colab_type": "text"
      },
      "source": [
        "Now we will create our basic model. We will use the Keras Sequential Model, just like last time, and we will have 3 layers, also just like last time. The input dimensions are 8, since we have 8 factors that can determine whether they will get diabetes. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SNN8fu8guC1Y",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 252
        },
        "outputId": "467a8ff3-414a-43a3-d7ff-fed7f6f18ff6"
      },
      "source": [
        "# define the keras model\n",
        "model = Sequential()\n",
        "model.add(Dense(12, input_dim=8, activation='relu'))\n",
        "model.add(Dense(8, activation='relu'))\n",
        "model.add(Dense(1, activation='sigmoid'))\n",
        "model.summary()"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_2\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "dense_4 (Dense)              (None, 12)                108       \n",
            "_________________________________________________________________\n",
            "dense_5 (Dense)              (None, 8)                 104       \n",
            "_________________________________________________________________\n",
            "dense_6 (Dense)              (None, 1)                 9         \n",
            "=================================================================\n",
            "Total params: 221\n",
            "Trainable params: 221\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NQANV56nuC1d",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# compile the keras model\n",
        "model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])"
      ],
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EAbBduGyuC1f",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "a01a3c61-392d-4798-c7fc-62cbc7cfe4c9"
      },
      "source": [
        "# fit the keras model on the dataset, train the model\n",
        "training = model.fit(X, y, epochs=300, batch_size=10, verbose = 2)"
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/300\n",
            " - 0s - loss: 0.4874 - accuracy: 0.7721\n",
            "Epoch 2/300\n",
            " - 0s - loss: 0.4870 - accuracy: 0.7721\n",
            "Epoch 3/300\n",
            " - 0s - loss: 0.4955 - accuracy: 0.7721\n",
            "Epoch 4/300\n",
            " - 0s - loss: 0.4775 - accuracy: 0.7695\n",
            "Epoch 5/300\n",
            " - 0s - loss: 0.4753 - accuracy: 0.7799\n",
            "Epoch 6/300\n",
            " - 0s - loss: 0.4892 - accuracy: 0.7552\n",
            "Epoch 7/300\n",
            " - 0s - loss: 0.4954 - accuracy: 0.7552\n",
            "Epoch 8/300\n",
            " - 0s - loss: 0.4772 - accuracy: 0.7630\n",
            "Epoch 9/300\n",
            " - 0s - loss: 0.4846 - accuracy: 0.7708\n",
            "Epoch 10/300\n",
            " - 0s - loss: 0.4847 - accuracy: 0.7617\n",
            "Epoch 11/300\n",
            " - 0s - loss: 0.4804 - accuracy: 0.7604\n",
            "Epoch 12/300\n",
            " - 0s - loss: 0.4821 - accuracy: 0.7682\n",
            "Epoch 13/300\n",
            " - 0s - loss: 0.4823 - accuracy: 0.7773\n",
            "Epoch 14/300\n",
            " - 0s - loss: 0.4749 - accuracy: 0.7799\n",
            "Epoch 15/300\n",
            " - 0s - loss: 0.4817 - accuracy: 0.7669\n",
            "Epoch 16/300\n",
            " - 0s - loss: 0.4894 - accuracy: 0.7526\n",
            "Epoch 17/300\n",
            " - 0s - loss: 0.4778 - accuracy: 0.7656\n",
            "Epoch 18/300\n",
            " - 0s - loss: 0.4824 - accuracy: 0.7708\n",
            "Epoch 19/300\n",
            " - 0s - loss: 0.4851 - accuracy: 0.7656\n",
            "Epoch 20/300\n",
            " - 0s - loss: 0.4896 - accuracy: 0.7656\n",
            "Epoch 21/300\n",
            " - 0s - loss: 0.4932 - accuracy: 0.7695\n",
            "Epoch 22/300\n",
            " - 0s - loss: 0.4939 - accuracy: 0.7552\n",
            "Epoch 23/300\n",
            " - 0s - loss: 0.4867 - accuracy: 0.7643\n",
            "Epoch 24/300\n",
            " - 0s - loss: 0.4811 - accuracy: 0.7682\n",
            "Epoch 25/300\n",
            " - 0s - loss: 0.4772 - accuracy: 0.7760\n",
            "Epoch 26/300\n",
            " - 0s - loss: 0.4733 - accuracy: 0.7656\n",
            "Epoch 27/300\n",
            " - 0s - loss: 0.4720 - accuracy: 0.7643\n",
            "Epoch 28/300\n",
            " - 0s - loss: 0.4735 - accuracy: 0.7617\n",
            "Epoch 29/300\n",
            " - 0s - loss: 0.4815 - accuracy: 0.7656\n",
            "Epoch 30/300\n",
            " - 0s - loss: 0.4791 - accuracy: 0.7773\n",
            "Epoch 31/300\n",
            " - 0s - loss: 0.4772 - accuracy: 0.7695\n",
            "Epoch 32/300\n",
            " - 0s - loss: 0.4783 - accuracy: 0.7747\n",
            "Epoch 33/300\n",
            " - 0s - loss: 0.4765 - accuracy: 0.7617\n",
            "Epoch 34/300\n",
            " - 0s - loss: 0.4693 - accuracy: 0.7826\n",
            "Epoch 35/300\n",
            " - 0s - loss: 0.4666 - accuracy: 0.7799\n",
            "Epoch 36/300\n",
            " - 0s - loss: 0.4717 - accuracy: 0.7760\n",
            "Epoch 37/300\n",
            " - 0s - loss: 0.4738 - accuracy: 0.7852\n",
            "Epoch 38/300\n",
            " - 0s - loss: 0.4793 - accuracy: 0.7760\n",
            "Epoch 39/300\n",
            " - 0s - loss: 0.4762 - accuracy: 0.7643\n",
            "Epoch 40/300\n",
            " - 0s - loss: 0.4755 - accuracy: 0.7734\n",
            "Epoch 41/300\n",
            " - 0s - loss: 0.4907 - accuracy: 0.7617\n",
            "Epoch 42/300\n",
            " - 0s - loss: 0.4675 - accuracy: 0.7812\n",
            "Epoch 43/300\n",
            " - 0s - loss: 0.4671 - accuracy: 0.7747\n",
            "Epoch 44/300\n",
            " - 0s - loss: 0.4735 - accuracy: 0.7682\n",
            "Epoch 45/300\n",
            " - 0s - loss: 0.4703 - accuracy: 0.7943\n",
            "Epoch 46/300\n",
            " - 0s - loss: 0.4710 - accuracy: 0.7760\n",
            "Epoch 47/300\n",
            " - 0s - loss: 0.4727 - accuracy: 0.7682\n",
            "Epoch 48/300\n",
            " - 0s - loss: 0.4643 - accuracy: 0.7734\n",
            "Epoch 49/300\n",
            " - 0s - loss: 0.4619 - accuracy: 0.7708\n",
            "Epoch 50/300\n",
            " - 0s - loss: 0.4721 - accuracy: 0.7812\n",
            "Epoch 51/300\n",
            " - 0s - loss: 0.4670 - accuracy: 0.7839\n",
            "Epoch 52/300\n",
            " - 0s - loss: 0.4671 - accuracy: 0.7799\n",
            "Epoch 53/300\n",
            " - 0s - loss: 0.4618 - accuracy: 0.7826\n",
            "Epoch 54/300\n",
            " - 0s - loss: 0.4718 - accuracy: 0.7839\n",
            "Epoch 55/300\n",
            " - 0s - loss: 0.4712 - accuracy: 0.7682\n",
            "Epoch 56/300\n",
            " - 0s - loss: 0.4727 - accuracy: 0.7799\n",
            "Epoch 57/300\n",
            " - 0s - loss: 0.4585 - accuracy: 0.7799\n",
            "Epoch 58/300\n",
            " - 0s - loss: 0.4654 - accuracy: 0.7708\n",
            "Epoch 59/300\n",
            " - 0s - loss: 0.4777 - accuracy: 0.7773\n",
            "Epoch 60/300\n",
            " - 0s - loss: 0.4663 - accuracy: 0.7826\n",
            "Epoch 61/300\n",
            " - 0s - loss: 0.4742 - accuracy: 0.7734\n",
            "Epoch 62/300\n",
            " - 0s - loss: 0.4627 - accuracy: 0.7760\n",
            "Epoch 63/300\n",
            " - 0s - loss: 0.4740 - accuracy: 0.7591\n",
            "Epoch 64/300\n",
            " - 0s - loss: 0.4718 - accuracy: 0.7630\n",
            "Epoch 65/300\n",
            " - 0s - loss: 0.4755 - accuracy: 0.7539\n",
            "Epoch 66/300\n",
            " - 0s - loss: 0.4659 - accuracy: 0.7773\n",
            "Epoch 67/300\n",
            " - 0s - loss: 0.4612 - accuracy: 0.7760\n",
            "Epoch 68/300\n",
            " - 0s - loss: 0.4600 - accuracy: 0.7773\n",
            "Epoch 69/300\n",
            " - 0s - loss: 0.4687 - accuracy: 0.7799\n",
            "Epoch 70/300\n",
            " - 0s - loss: 0.4669 - accuracy: 0.7747\n",
            "Epoch 71/300\n",
            " - 0s - loss: 0.4684 - accuracy: 0.7747\n",
            "Epoch 72/300\n",
            " - 0s - loss: 0.4638 - accuracy: 0.7734\n",
            "Epoch 73/300\n",
            " - 0s - loss: 0.4654 - accuracy: 0.7812\n",
            "Epoch 74/300\n",
            " - 0s - loss: 0.4581 - accuracy: 0.7799\n",
            "Epoch 75/300\n",
            " - 0s - loss: 0.4577 - accuracy: 0.7917\n",
            "Epoch 76/300\n",
            " - 0s - loss: 0.4641 - accuracy: 0.7747\n",
            "Epoch 77/300\n",
            " - 0s - loss: 0.4601 - accuracy: 0.7826\n",
            "Epoch 78/300\n",
            " - 0s - loss: 0.4693 - accuracy: 0.7695\n",
            "Epoch 79/300\n",
            " - 0s - loss: 0.4573 - accuracy: 0.7852\n",
            "Epoch 80/300\n",
            " - 0s - loss: 0.4624 - accuracy: 0.7786\n",
            "Epoch 81/300\n",
            " - 0s - loss: 0.4636 - accuracy: 0.7786\n",
            "Epoch 82/300\n",
            " - 0s - loss: 0.4621 - accuracy: 0.7865\n",
            "Epoch 83/300\n",
            " - 0s - loss: 0.4564 - accuracy: 0.7656\n",
            "Epoch 84/300\n",
            " - 0s - loss: 0.4612 - accuracy: 0.7708\n",
            "Epoch 85/300\n",
            " - 0s - loss: 0.4614 - accuracy: 0.7878\n",
            "Epoch 86/300\n",
            " - 0s - loss: 0.4642 - accuracy: 0.7799\n",
            "Epoch 87/300\n",
            " - 0s - loss: 0.4572 - accuracy: 0.7852\n",
            "Epoch 88/300\n",
            " - 0s - loss: 0.4663 - accuracy: 0.7786\n",
            "Epoch 89/300\n",
            " - 0s - loss: 0.4640 - accuracy: 0.7630\n",
            "Epoch 90/300\n",
            " - 0s - loss: 0.4653 - accuracy: 0.7773\n",
            "Epoch 91/300\n",
            " - 0s - loss: 0.4613 - accuracy: 0.7734\n",
            "Epoch 92/300\n",
            " - 0s - loss: 0.4613 - accuracy: 0.7786\n",
            "Epoch 93/300\n",
            " - 0s - loss: 0.4516 - accuracy: 0.7786\n",
            "Epoch 94/300\n",
            " - 0s - loss: 0.4607 - accuracy: 0.7786\n",
            "Epoch 95/300\n",
            " - 0s - loss: 0.4504 - accuracy: 0.7839\n",
            "Epoch 96/300\n",
            " - 0s - loss: 0.4604 - accuracy: 0.7852\n",
            "Epoch 97/300\n",
            " - 0s - loss: 0.4539 - accuracy: 0.7904\n",
            "Epoch 98/300\n",
            " - 0s - loss: 0.4530 - accuracy: 0.7852\n",
            "Epoch 99/300\n",
            " - 0s - loss: 0.4505 - accuracy: 0.7839\n",
            "Epoch 100/300\n",
            " - 0s - loss: 0.4579 - accuracy: 0.7799\n",
            "Epoch 101/300\n",
            " - 0s - loss: 0.4568 - accuracy: 0.7943\n",
            "Epoch 102/300\n",
            " - 0s - loss: 0.4715 - accuracy: 0.7852\n",
            "Epoch 103/300\n",
            " - 0s - loss: 0.4550 - accuracy: 0.7826\n",
            "Epoch 104/300\n",
            " - 0s - loss: 0.4706 - accuracy: 0.7773\n",
            "Epoch 105/300\n",
            " - 0s - loss: 0.4530 - accuracy: 0.7839\n",
            "Epoch 106/300\n",
            " - 0s - loss: 0.4591 - accuracy: 0.7773\n",
            "Epoch 107/300\n",
            " - 0s - loss: 0.4673 - accuracy: 0.7786\n",
            "Epoch 108/300\n",
            " - 0s - loss: 0.4580 - accuracy: 0.7773\n",
            "Epoch 109/300\n",
            " - 0s - loss: 0.4474 - accuracy: 0.7799\n",
            "Epoch 110/300\n",
            " - 0s - loss: 0.4529 - accuracy: 0.8047\n",
            "Epoch 111/300\n",
            " - 0s - loss: 0.4537 - accuracy: 0.7786\n",
            "Epoch 112/300\n",
            " - 0s - loss: 0.4584 - accuracy: 0.7812\n",
            "Epoch 113/300\n",
            " - 0s - loss: 0.4576 - accuracy: 0.7943\n",
            "Epoch 114/300\n",
            " - 0s - loss: 0.4537 - accuracy: 0.7839\n",
            "Epoch 115/300\n",
            " - 0s - loss: 0.4500 - accuracy: 0.7878\n",
            "Epoch 116/300\n",
            " - 0s - loss: 0.4502 - accuracy: 0.7812\n",
            "Epoch 117/300\n",
            " - 0s - loss: 0.4546 - accuracy: 0.7865\n",
            "Epoch 118/300\n",
            " - 0s - loss: 0.4469 - accuracy: 0.7760\n",
            "Epoch 119/300\n",
            " - 0s - loss: 0.4429 - accuracy: 0.7865\n",
            "Epoch 120/300\n",
            " - 0s - loss: 0.4448 - accuracy: 0.7865\n",
            "Epoch 121/300\n",
            " - 0s - loss: 0.4506 - accuracy: 0.7786\n",
            "Epoch 122/300\n",
            " - 0s - loss: 0.4517 - accuracy: 0.7917\n",
            "Epoch 123/300\n",
            " - 0s - loss: 0.4507 - accuracy: 0.7891\n",
            "Epoch 124/300\n",
            " - 0s - loss: 0.4684 - accuracy: 0.7760\n",
            "Epoch 125/300\n",
            " - 0s - loss: 0.4416 - accuracy: 0.7839\n",
            "Epoch 126/300\n",
            " - 0s - loss: 0.4581 - accuracy: 0.7917\n",
            "Epoch 127/300\n",
            " - 0s - loss: 0.4434 - accuracy: 0.7982\n",
            "Epoch 128/300\n",
            " - 0s - loss: 0.4483 - accuracy: 0.7786\n",
            "Epoch 129/300\n",
            " - 0s - loss: 0.4476 - accuracy: 0.7812\n",
            "Epoch 130/300\n",
            " - 0s - loss: 0.4451 - accuracy: 0.7839\n",
            "Epoch 131/300\n",
            " - 0s - loss: 0.4585 - accuracy: 0.7839\n",
            "Epoch 132/300\n",
            " - 0s - loss: 0.4484 - accuracy: 0.7839\n",
            "Epoch 133/300\n",
            " - 0s - loss: 0.4543 - accuracy: 0.7852\n",
            "Epoch 134/300\n",
            " - 0s - loss: 0.4453 - accuracy: 0.7760\n",
            "Epoch 135/300\n",
            " - 0s - loss: 0.4449 - accuracy: 0.7865\n",
            "Epoch 136/300\n",
            " - 0s - loss: 0.4433 - accuracy: 0.7839\n",
            "Epoch 137/300\n",
            " - 0s - loss: 0.4566 - accuracy: 0.7930\n",
            "Epoch 138/300\n",
            " - 0s - loss: 0.4375 - accuracy: 0.7891\n",
            "Epoch 139/300\n",
            " - 0s - loss: 0.4493 - accuracy: 0.7812\n",
            "Epoch 140/300\n",
            " - 0s - loss: 0.4462 - accuracy: 0.7865\n",
            "Epoch 141/300\n",
            " - 0s - loss: 0.4455 - accuracy: 0.7904\n",
            "Epoch 142/300\n",
            " - 0s - loss: 0.4474 - accuracy: 0.7865\n",
            "Epoch 143/300\n",
            " - 0s - loss: 0.4413 - accuracy: 0.7930\n",
            "Epoch 144/300\n",
            " - 0s - loss: 0.4587 - accuracy: 0.7760\n",
            "Epoch 145/300\n",
            " - 0s - loss: 0.4392 - accuracy: 0.7799\n",
            "Epoch 146/300\n",
            " - 0s - loss: 0.4371 - accuracy: 0.7969\n",
            "Epoch 147/300\n",
            " - 0s - loss: 0.4668 - accuracy: 0.7669\n",
            "Epoch 148/300\n",
            " - 0s - loss: 0.4455 - accuracy: 0.7826\n",
            "Epoch 149/300\n",
            " - 0s - loss: 0.4551 - accuracy: 0.7930\n",
            "Epoch 150/300\n",
            " - 0s - loss: 0.4477 - accuracy: 0.7852\n",
            "Epoch 151/300\n",
            " - 0s - loss: 0.4496 - accuracy: 0.7865\n",
            "Epoch 152/300\n",
            " - 0s - loss: 0.4474 - accuracy: 0.7852\n",
            "Epoch 153/300\n",
            " - 0s - loss: 0.4486 - accuracy: 0.7839\n",
            "Epoch 154/300\n",
            " - 0s - loss: 0.4476 - accuracy: 0.7760\n",
            "Epoch 155/300\n",
            " - 0s - loss: 0.4374 - accuracy: 0.7930\n",
            "Epoch 156/300\n",
            " - 0s - loss: 0.4431 - accuracy: 0.7878\n",
            "Epoch 157/300\n",
            " - 0s - loss: 0.4368 - accuracy: 0.7956\n",
            "Epoch 158/300\n",
            " - 0s - loss: 0.4381 - accuracy: 0.7878\n",
            "Epoch 159/300\n",
            " - 0s - loss: 0.4356 - accuracy: 0.7943\n",
            "Epoch 160/300\n",
            " - 0s - loss: 0.4445 - accuracy: 0.7865\n",
            "Epoch 161/300\n",
            " - 0s - loss: 0.4490 - accuracy: 0.7695\n",
            "Epoch 162/300\n",
            " - 0s - loss: 0.4438 - accuracy: 0.7839\n",
            "Epoch 163/300\n",
            " - 0s - loss: 0.4488 - accuracy: 0.7812\n",
            "Epoch 164/300\n",
            " - 0s - loss: 0.4668 - accuracy: 0.7956\n",
            "Epoch 165/300\n",
            " - 0s - loss: 0.4347 - accuracy: 0.8008\n",
            "Epoch 166/300\n",
            " - 0s - loss: 0.4476 - accuracy: 0.7773\n",
            "Epoch 167/300\n",
            " - 0s - loss: 0.4428 - accuracy: 0.7878\n",
            "Epoch 168/300\n",
            " - 0s - loss: 0.4458 - accuracy: 0.7721\n",
            "Epoch 169/300\n",
            " - 0s - loss: 0.4505 - accuracy: 0.7773\n",
            "Epoch 170/300\n",
            " - 0s - loss: 0.4417 - accuracy: 0.7865\n",
            "Epoch 171/300\n",
            " - 0s - loss: 0.4412 - accuracy: 0.7904\n",
            "Epoch 172/300\n",
            " - 0s - loss: 0.4406 - accuracy: 0.7878\n",
            "Epoch 173/300\n",
            " - 0s - loss: 0.4408 - accuracy: 0.7930\n",
            "Epoch 174/300\n",
            " - 0s - loss: 0.4553 - accuracy: 0.7656\n",
            "Epoch 175/300\n",
            " - 0s - loss: 0.4317 - accuracy: 0.7982\n",
            "Epoch 176/300\n",
            " - 0s - loss: 0.4335 - accuracy: 0.8008\n",
            "Epoch 177/300\n",
            " - 0s - loss: 0.4642 - accuracy: 0.7773\n",
            "Epoch 178/300\n",
            " - 0s - loss: 0.4536 - accuracy: 0.7852\n",
            "Epoch 179/300\n",
            " - 0s - loss: 0.4314 - accuracy: 0.7839\n",
            "Epoch 180/300\n",
            " - 0s - loss: 0.4367 - accuracy: 0.8086\n",
            "Epoch 181/300\n",
            " - 0s - loss: 0.4378 - accuracy: 0.7878\n",
            "Epoch 182/300\n",
            " - 0s - loss: 0.4314 - accuracy: 0.7956\n",
            "Epoch 183/300\n",
            " - 0s - loss: 0.4373 - accuracy: 0.7786\n",
            "Epoch 184/300\n",
            " - 0s - loss: 0.4253 - accuracy: 0.7943\n",
            "Epoch 185/300\n",
            " - 0s - loss: 0.4323 - accuracy: 0.7799\n",
            "Epoch 186/300\n",
            " - 0s - loss: 0.4332 - accuracy: 0.8021\n",
            "Epoch 187/300\n",
            " - 0s - loss: 0.4463 - accuracy: 0.7799\n",
            "Epoch 188/300\n",
            " - 0s - loss: 0.4439 - accuracy: 0.7760\n",
            "Epoch 189/300\n",
            " - 0s - loss: 0.4465 - accuracy: 0.7878\n",
            "Epoch 190/300\n",
            " - 0s - loss: 0.4523 - accuracy: 0.7904\n",
            "Epoch 191/300\n",
            " - 0s - loss: 0.4408 - accuracy: 0.7812\n",
            "Epoch 192/300\n",
            " - 0s - loss: 0.4323 - accuracy: 0.7865\n",
            "Epoch 193/300\n",
            " - 0s - loss: 0.4377 - accuracy: 0.7930\n",
            "Epoch 194/300\n",
            " - 0s - loss: 0.4342 - accuracy: 0.7943\n",
            "Epoch 195/300\n",
            " - 0s - loss: 0.4369 - accuracy: 0.7904\n",
            "Epoch 196/300\n",
            " - 0s - loss: 0.4361 - accuracy: 0.8021\n",
            "Epoch 197/300\n",
            " - 0s - loss: 0.4332 - accuracy: 0.7982\n",
            "Epoch 198/300\n",
            " - 0s - loss: 0.4398 - accuracy: 0.7839\n",
            "Epoch 199/300\n",
            " - 0s - loss: 0.4363 - accuracy: 0.7839\n",
            "Epoch 200/300\n",
            " - 0s - loss: 0.4384 - accuracy: 0.7917\n",
            "Epoch 201/300\n",
            " - 0s - loss: 0.4356 - accuracy: 0.7852\n",
            "Epoch 202/300\n",
            " - 0s - loss: 0.4408 - accuracy: 0.7878\n",
            "Epoch 203/300\n",
            " - 0s - loss: 0.4388 - accuracy: 0.7773\n",
            "Epoch 204/300\n",
            " - 0s - loss: 0.4419 - accuracy: 0.7839\n",
            "Epoch 205/300\n",
            " - 0s - loss: 0.4303 - accuracy: 0.7982\n",
            "Epoch 206/300\n",
            " - 0s - loss: 0.4450 - accuracy: 0.7943\n",
            "Epoch 207/300\n",
            " - 0s - loss: 0.4457 - accuracy: 0.7826\n",
            "Epoch 208/300\n",
            " - 0s - loss: 0.4320 - accuracy: 0.7891\n",
            "Epoch 209/300\n",
            " - 0s - loss: 0.4363 - accuracy: 0.7891\n",
            "Epoch 210/300\n",
            " - 0s - loss: 0.4470 - accuracy: 0.7812\n",
            "Epoch 211/300\n",
            " - 0s - loss: 0.4321 - accuracy: 0.7839\n",
            "Epoch 212/300\n",
            " - 0s - loss: 0.4286 - accuracy: 0.7904\n",
            "Epoch 213/300\n",
            " - 0s - loss: 0.4323 - accuracy: 0.7904\n",
            "Epoch 214/300\n",
            " - 0s - loss: 0.4264 - accuracy: 0.7943\n",
            "Epoch 215/300\n",
            " - 0s - loss: 0.4394 - accuracy: 0.7943\n",
            "Epoch 216/300\n",
            " - 0s - loss: 0.4288 - accuracy: 0.7930\n",
            "Epoch 217/300\n",
            " - 0s - loss: 0.4241 - accuracy: 0.8047\n",
            "Epoch 218/300\n",
            " - 0s - loss: 0.4316 - accuracy: 0.7943\n",
            "Epoch 219/300\n",
            " - 0s - loss: 0.4265 - accuracy: 0.7852\n",
            "Epoch 220/300\n",
            " - 0s - loss: 0.4230 - accuracy: 0.7930\n",
            "Epoch 221/300\n",
            " - 0s - loss: 0.4216 - accuracy: 0.8034\n",
            "Epoch 222/300\n",
            " - 0s - loss: 0.4370 - accuracy: 0.7982\n",
            "Epoch 223/300\n",
            " - 0s - loss: 0.4350 - accuracy: 0.7878\n",
            "Epoch 224/300\n",
            " - 0s - loss: 0.4437 - accuracy: 0.7826\n",
            "Epoch 225/300\n",
            " - 0s - loss: 0.4257 - accuracy: 0.7956\n",
            "Epoch 226/300\n",
            " - 0s - loss: 0.4342 - accuracy: 0.7995\n",
            "Epoch 227/300\n",
            " - 0s - loss: 0.4278 - accuracy: 0.7891\n",
            "Epoch 228/300\n",
            " - 0s - loss: 0.4260 - accuracy: 0.7930\n",
            "Epoch 229/300\n",
            " - 0s - loss: 0.4233 - accuracy: 0.7891\n",
            "Epoch 230/300\n",
            " - 0s - loss: 0.4225 - accuracy: 0.7943\n",
            "Epoch 231/300\n",
            " - 0s - loss: 0.4323 - accuracy: 0.7943\n",
            "Epoch 232/300\n",
            " - 0s - loss: 0.4322 - accuracy: 0.7904\n",
            "Epoch 233/300\n",
            " - 0s - loss: 0.4307 - accuracy: 0.7852\n",
            "Epoch 234/300\n",
            " - 0s - loss: 0.4232 - accuracy: 0.8021\n",
            "Epoch 235/300\n",
            " - 0s - loss: 0.4293 - accuracy: 0.7904\n",
            "Epoch 236/300\n",
            " - 0s - loss: 0.4237 - accuracy: 0.8008\n",
            "Epoch 237/300\n",
            " - 0s - loss: 0.4396 - accuracy: 0.7904\n",
            "Epoch 238/300\n",
            " - 0s - loss: 0.4378 - accuracy: 0.7904\n",
            "Epoch 239/300\n",
            " - 0s - loss: 0.4343 - accuracy: 0.7930\n",
            "Epoch 240/300\n",
            " - 0s - loss: 0.4226 - accuracy: 0.7982\n",
            "Epoch 241/300\n",
            " - 0s - loss: 0.4290 - accuracy: 0.7943\n",
            "Epoch 242/300\n",
            " - 0s - loss: 0.4181 - accuracy: 0.8034\n",
            "Epoch 243/300\n",
            " - 0s - loss: 0.4272 - accuracy: 0.7956\n",
            "Epoch 244/300\n",
            " - 0s - loss: 0.4362 - accuracy: 0.7943\n",
            "Epoch 245/300\n",
            " - 0s - loss: 0.4328 - accuracy: 0.7969\n",
            "Epoch 246/300\n",
            " - 0s - loss: 0.4334 - accuracy: 0.7878\n",
            "Epoch 247/300\n",
            " - 0s - loss: 0.4309 - accuracy: 0.7839\n",
            "Epoch 248/300\n",
            " - 0s - loss: 0.4377 - accuracy: 0.7839\n",
            "Epoch 249/300\n",
            " - 0s - loss: 0.4373 - accuracy: 0.7943\n",
            "Epoch 250/300\n",
            " - 0s - loss: 0.4242 - accuracy: 0.7943\n",
            "Epoch 251/300\n",
            " - 0s - loss: 0.4473 - accuracy: 0.8021\n",
            "Epoch 252/300\n",
            " - 0s - loss: 0.4379 - accuracy: 0.7904\n",
            "Epoch 253/300\n",
            " - 0s - loss: 0.4291 - accuracy: 0.7956\n",
            "Epoch 254/300\n",
            " - 0s - loss: 0.4261 - accuracy: 0.7891\n",
            "Epoch 255/300\n",
            " - 0s - loss: 0.4243 - accuracy: 0.7904\n",
            "Epoch 256/300\n",
            " - 0s - loss: 0.4419 - accuracy: 0.7982\n",
            "Epoch 257/300\n",
            " - 0s - loss: 0.4319 - accuracy: 0.7904\n",
            "Epoch 258/300\n",
            " - 0s - loss: 0.4272 - accuracy: 0.8086\n",
            "Epoch 259/300\n",
            " - 0s - loss: 0.4161 - accuracy: 0.8034\n",
            "Epoch 260/300\n",
            " - 0s - loss: 0.4429 - accuracy: 0.7773\n",
            "Epoch 261/300\n",
            " - 0s - loss: 0.4264 - accuracy: 0.7943\n",
            "Epoch 262/300\n",
            " - 0s - loss: 0.4287 - accuracy: 0.8073\n",
            "Epoch 263/300\n",
            " - 0s - loss: 0.4169 - accuracy: 0.7956\n",
            "Epoch 264/300\n",
            " - 0s - loss: 0.4338 - accuracy: 0.7982\n",
            "Epoch 265/300\n",
            " - 0s - loss: 0.4252 - accuracy: 0.7956\n",
            "Epoch 266/300\n",
            " - 0s - loss: 0.4356 - accuracy: 0.7904\n",
            "Epoch 267/300\n",
            " - 0s - loss: 0.4276 - accuracy: 0.7904\n",
            "Epoch 268/300\n",
            " - 0s - loss: 0.4214 - accuracy: 0.7956\n",
            "Epoch 269/300\n",
            " - 0s - loss: 0.4282 - accuracy: 0.7982\n",
            "Epoch 270/300\n",
            " - 0s - loss: 0.4185 - accuracy: 0.7943\n",
            "Epoch 271/300\n",
            " - 0s - loss: 0.4190 - accuracy: 0.7956\n",
            "Epoch 272/300\n",
            " - 0s - loss: 0.4260 - accuracy: 0.7943\n",
            "Epoch 273/300\n",
            " - 0s - loss: 0.4201 - accuracy: 0.7982\n",
            "Epoch 274/300\n",
            " - 0s - loss: 0.4187 - accuracy: 0.7982\n",
            "Epoch 275/300\n",
            " - 0s - loss: 0.4198 - accuracy: 0.8034\n",
            "Epoch 276/300\n",
            " - 0s - loss: 0.4241 - accuracy: 0.7982\n",
            "Epoch 277/300\n",
            " - 0s - loss: 0.4329 - accuracy: 0.7943\n",
            "Epoch 278/300\n",
            " - 0s - loss: 0.4174 - accuracy: 0.8008\n",
            "Epoch 279/300\n",
            " - 0s - loss: 0.4245 - accuracy: 0.8008\n",
            "Epoch 280/300\n",
            " - 0s - loss: 0.4350 - accuracy: 0.7891\n",
            "Epoch 281/300\n",
            " - 0s - loss: 0.4155 - accuracy: 0.7969\n",
            "Epoch 282/300\n",
            " - 0s - loss: 0.4171 - accuracy: 0.8021\n",
            "Epoch 283/300\n",
            " - 0s - loss: 0.4208 - accuracy: 0.7930\n",
            "Epoch 284/300\n",
            " - 0s - loss: 0.4228 - accuracy: 0.8034\n",
            "Epoch 285/300\n",
            " - 0s - loss: 0.4306 - accuracy: 0.7956\n",
            "Epoch 286/300\n",
            " - 0s - loss: 0.4297 - accuracy: 0.7865\n",
            "Epoch 287/300\n",
            " - 0s - loss: 0.4167 - accuracy: 0.8047\n",
            "Epoch 288/300\n",
            " - 0s - loss: 0.4158 - accuracy: 0.7943\n",
            "Epoch 289/300\n",
            " - 0s - loss: 0.4175 - accuracy: 0.8073\n",
            "Epoch 290/300\n",
            " - 0s - loss: 0.4271 - accuracy: 0.8177\n",
            "Epoch 291/300\n",
            " - 0s - loss: 0.4228 - accuracy: 0.7891\n",
            "Epoch 292/300\n",
            " - 0s - loss: 0.4279 - accuracy: 0.7917\n",
            "Epoch 293/300\n",
            " - 0s - loss: 0.4140 - accuracy: 0.8112\n",
            "Epoch 294/300\n",
            " - 0s - loss: 0.4261 - accuracy: 0.7969\n",
            "Epoch 295/300\n",
            " - 0s - loss: 0.4201 - accuracy: 0.7891\n",
            "Epoch 296/300\n",
            " - 0s - loss: 0.4210 - accuracy: 0.7982\n",
            "Epoch 297/300\n",
            " - 0s - loss: 0.4155 - accuracy: 0.8008\n",
            "Epoch 298/300\n",
            " - 0s - loss: 0.4141 - accuracy: 0.7891\n",
            "Epoch 299/300\n",
            " - 0s - loss: 0.4239 - accuracy: 0.7904\n",
            "Epoch 300/300\n",
            " - 0s - loss: 0.4288 - accuracy: 0.8047\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JvTdw53JuC1i",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "outputId": "b8bd5d78-1cd3-4a79-e9cd-e7b2938a06ee"
      },
      "source": [
        "# now lets see how the accuracy of our model increased over training.\n",
        "# summarize history for accuracy\n",
        "plt.plot(training.history['accuracy'])\n",
        "plt.title('Model Accuracy')\n",
        "plt.ylabel('Accuracy')\n",
        "plt.xlabel('Epochs')\n",
        "plt.show()"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEWCAYAAAB8LwAVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOy9d7wkVZk+/rxV1emmyaQJMCRhAImCgIqIiMAKplXYNbDripFdXfX3NS2iq7u6rmld47quoiumFUVFQcQIg8CQGQkDDJNh5s6dm/p2qKrz++PUe+pU6q6+c/umOc/nM5+53V1ddar73vc9z/u8gYQQMDAwMDAwiMOa6QUYGBgYGMxOGAdhYGBgYJAK4yAMDAwMDFJhHISBgYGBQSqMgzAwMDAwSIVxEAYGBgYGqTAOwmCfBxEdQkSCiJwcx15GRH+cjnUZGMw0jIMwmFMgoo1E1CCipbHn7w6M/CEzs7LIWvqIaIyIfjHTazEw2BsYB2EwF/EEgEv5AREdB6Bn5paTwCsA1AGcS0QHTOeF87AgA4O8MA7CYC7iWwBepz1+PYCr9QOIaAERXU1EO4noSSL6IBFZwWs2Ef07Ee0ioscBXJjy3v8mou1EtJWIPkpEdgfrez2ALwO4D8BrYud+DhHdSkR7iGgzEV0WPF8hok8Fax0moj8Gzz2fiLbEzrGRiF4Y/HwVEf2QiL5NRCMALiOiU4lobXCN7UT0n0RU1N5/DBH9ioh2E9FTRPR+IjqAiKpEtEQ77qTg8yt0cO8G8wjGQRjMRdwGYICIjg4M9yUAvh075vMAFgA4FMBZkA7lb4LX3gjgLwCcCOAUAK+MvfcbAFwAhwfHvAjA3+VZGBEdDOD5AP43+Pe62Gu/CNa2DMAJAO4JXv53ACcDOAPAYgD/HwA/zzUBXAzghwAWBtf0ALwTwFIApwM4B8BbgzX0A7gJwC8BHBTc46+FEDsA/BbAq7TzvhbAd4UQzZzrMJhnMA7CYK6CWcS5AP4MYCu/oDmN9wkhRoUQGwF8CtLgAdIIflYIsVkIsRvAv2rv3R/ABQDeIYQYF0I8DeAzwfny4LUA7hNCrAfwXQDHENGJwWt/BeAmIcQ1QoimEGJQCHFPwGz+FsA/CCG2CiE8IcStQoh6zmuuFUL8WAjhCyEmhBDrhBC3CSHc4N6/AukkAekYdwghPiWEqAWfz5+C176JgPEEn+GlkJ+zwT4KE680mKv4FoDfA1iNWHgJcudcAPCk9tyTAJYHPx8EYHPsNcbBwXu3ExE/Z8WOb4XXAfgvABBCbCWi30GGnO4GsBLAYynvWQqgnPFaHkTWRkRHAvg0JDvqgfw7Xxe8nLUGAPgJgC8T0WoAzwAwLIS4fZJrMpgHMAzCYE5CCPEkpFh9AYAfxV7eBaAJaewZqxCyjO2QhlJ/jbEZUmBeKoRYGPwbEEIc025NRHQGgCMAvI+IdhDRDgCnAfirQDzeDOCwlLfuAlDLeG0cmgAf7OyXxY6Jt2T+EoCHABwhhBgA8H4A7O02Q4bdEhBC1AB8H5JFvBaGPezzMA7CYC7jDQBeIIQY158UQniQhu5jRNQfxP7/EaFO8X0Af09EK4hoEYD3au/dDuBGAJ8iogEisojoMCI6C+3xegC/ArAGUl84AcCxACoAzofUB15IRK8iIoeIlhDRCUIIH8DXAXyaiA4KRPTTiagE4BEAZSK6MBCLPwig1GYd/QBGAIwR0VEA3qK99jMABxLRO4ioFHw+p2mvXw3gMgAXwTiIfR7GQRjMWQghHhNC3Jnx8hWQu+/HAfwRwHcgjTAgQ0A3ALgXwF1IMpDXASgCWA9gCFIAPrDVWoioDKltfF4IsUP79wSkoX29EGITJON5F4DdkAL18cEp3g3gfgB3BK99AoAlhBiGFJi/BsmAxgFEsppS8G5IvWM0uNfv8QtCiFFI3eYlAHYAeBTA2drrt0CK43cFLM1gHwaZgUEGBgY6iOhmAN8RQnxtptdiMLMwDsLAwECBiJ4FGSZbGbANg30YJsRkYGAAACCib0LWSLzDOAcDwDAIAwMDA4MMGAZhYGBgYJCKrhbKEdGLAXwOgA3ga0KIj8deXwVZvbkwOOa9QojriehcAB+HzCRpAHiPEOLmVtdaunSpOOSQQ6b+JgwMDAzmMdatW7dLCBGvrQHQRQcRFPR8ATKlbguAO4jouqAFAeODAL4vhPgSEa0BcD2AQyALh14ihNhGRMdCpiQuRwsccsghuPPOrIxHAwMDA4M0EFFmOnM3Q0ynAtgghHhcCNGA7EtzcewYAWAg+HkBgG0AIIS4WwixLXj+QQCVoGjIwMDAwGCa0E0HsRzRHjFbkGQBVwF4TdDO+HrI4qY4XgFZtJNoXEZElxPRnUR0586dO6dm1QYGBgYGAGZepL4UwDeEECsgK0y/xT37Adm3HrKi9E1pbxZCfFUIcYoQ4pRly1JDaAYGBgYGk0Q3HcRWRBuirYDWkjnAGyD74kAIsRayo+VSACCiFQCuBfA6IcRku1waGBgYGEwS3XQQdwA4gohWB9OsLgFwXeyYTZDDTEBER0M6iJ1EtBDAzyGzmm7p4hoNDAwMDDLQNQchhHABvB0yA+nPkNlKDxLRR4joouCwdwF4IxHdC+AaAJcJWbn3dshJV1cS0T3Bv/26tVYDAwMDgyTmTSX1KaecIkyaq4GBgUFnIKJ1QohT0l6baZHawMDAwCAFf3x0F57YNd7+wC7CjBw1MDAwmIV4zX/LUeEbP37hjK3BMAgDAwMDg1QYB2FgYGBgkArjIAwMDAxmGTx/diQPGQdhYGBgMMvQcP2ZXgIA4yAMDAwMZh3qrjfTSwBgHISBgYHBrEPdMAgDAwMDgzTUm8ZBGBgYGBikoGZCTAYGBgYGaTAMwsDAwMAgFSxSE83sOoyDMDAwMJhlYJG6YM2siTYOwsDAwGCWodaUDMKxZ5ZCGAdhYGBgMMvADMKxjIMwMDAwMNDAGkTRMSEmAwMDAwMNnMXkGA3CwMDAwECHCjEZDcLAwMDAQAeL1AXbMAgDAwMDAw1GpDYwMDAwSEW8UO6bt27ETeufmvZ1dNVBENGLiehhItpARO9NeX0VEf2GiO4movuI6ILg+SXB82NE9J/dXKOBgYHBbAOL1Dw46H9ueQI/unvLtK+jaw6CiGwAXwBwPoA1AC4lojWxwz4I4PtCiBMBXALgi8HzNQD/BODd3VqfgYGBwWwFN+vjwXKuL9Bwp3/KXDcZxKkANgghHhdCNAB8F8DFsWMEgIHg5wUAtgGAEGJcCPFHSEdhYGBgsE8hziBcT6DpTX8DP6eL514OYLP2eAuA02LHXAXgRiK6AkAvgBd2cT0GBgYGcwIsUisH4Qu4/vQ7iJkWqS8F8A0hxAoAFwD4FhHlXhMRXU5EdxLRnTt37uzaIg0MDAymE3UVYpIOwvN9NOdZiGkrgJXa4xXBczreAOD7ACCEWAugDGBp3gsIIb4qhDhFCHHKsmXL9nK5BgYGBrMDCQbhCTRmIMTUTQdxB4AjiGg1ERUhRejrYsdsAnAOABDR0ZAOwlABAwODfRpcKMcMwvVnRoPomoMQQrgA3g7gBgB/hsxWepCIPkJEFwWHvQvAG4noXgDXALhMCPmJENFGAJ8GcBkRbUnJgDIwMGiBT97wEK67d9tML2NOYu1jg/i7b94B35/+sA6QpkH4806khhDiegDXx567Uvt5PYAzM957SDfXZmAw33HtXVtx2qE1XHT8QTO9lDmHt/7vOgxVmxiqNrCkrzTt109kMfkCTW9+aRAGBgYzCE8IuDO0A57r6CnKvfN43ZuR69e1OgjfFxACaLjzKMRkYGAws/B8AW8GUiPnA3qKNgBgpNackevXNAbRDL7DeaVBGBgYzCw8X8CdgbDEfEBPSTKI4QnpIB7YOozXff12tbPvNpQGIYQKMxkHYWBgMGWQxVXGQUwGvcwgAgdx16Yh/P6Rndg5Wp+W6zc4xKRpD0aDMDAwmDL4xkFMGqxBMIPg+P906QBc86AziPlWB2FgYDCDcI0GMWnENQg2ztNlpJktCAG4XqhBBFUA0wbjIAwM5il8YTSIyaLoSNMYZxCcftpNyOQCgUIwbpT1CCHCtNfpgnEQBgbzFK4vpt2gzBdwgdzIhAsgFIing0HwtcqOZDF1Law13TqEcRAGBvMQnDtvNIjJockOojb9GgQ7hHIQ5tKvOd06hHEQBgbzEJ7qAmocxGTAcf+ZEKn5GuWCNM+6U5juVFfjIAzmLKoNFy/+7O9x7+Y9M72UWQe9RcNsws7ROl7w77/FE7vGZ3opLeGqEBOL1PJxt+sgXv/12/H9O+UYnUohCDE1w2umOYg3fOMO/OWXb+3KeoyDMJiz2Dlax0M7RvHwjtGZXsqsQ9gmenZlMT05OI7Hd43P+u8si0HUu8wg1j4+iDs27gYQOogIg0iZCVFzPXQruck4CIM5C3eW7pJnA2ZriKmuDO3M9DjKC8UgajGRussOwvV8jAXXLCkG0VqDaLoCBbs7ptw4CIM5C8408aY5N3wuwJ+lznO6duJ7C04Pnk4G4fsCvgBGAweRyiBSHETD81FwjIMwMIiAjd9M9eyfzeDPZvYyCPn/SK2JMz9+M9Y9OTSl1/ngj+/HVdc9OOn38/znhuuj1vTCQrkOHMTu8QZO/9df44Gtw5nHCCFwwef+gJ/eu01jLdIpKZHabe0gmp6PYlAzMdUwDsJgzsKbpUZwNiBkELNrp86hJRZenxquYeueCazflm1EJ4P7twzj3i2TT17Q6w2anj+pOoiNg+PYPlzD+m0jmcfUXR/rt49g/fYR9V2NxRiEHo7LchDdCjF1dWCQgUE3YRxENmYrg4iHavh/DuVMFSaaHiya/K5ad6yeL8L1dlBJzffU6t74fBMNTzmlsUbgINLqIFJE6qbXPQ3COAiDOQslxBoNIoHZmubKO3BmEDx7mcXgqUJtL1ti6C1KPG0edMPLL65zimyrmRLMDqoNV31n/OtcSq2kTtEg3O4xCBNiMpizMAwiG+qzSWnN8J4f3Iv3X3t/23O870f345D3/nxK18U75gSDqHbGIF71lbX48u8ey3x9oumh2mhtzL/028fwqi+vTX1Nd6yeEJMqlBtpwSAaro8zP34zrr17KwCg2vASKcnMINo5iKbno+h0R4MwDMJgzsI4iGwwq2qmaBAbdo6hmGPHec3tmwBIPcOypsYAMYNg5sA76E4ntz361ChWLe7JfL3W9NomLzzy1CgeeTq9HkM31p4vJpXFxKxoJMVB7Kk2sHXPBB4M9ImJhqfaezC4F1MekdowCAODGIyDyEarz4a7hebFVOoDCQ2iOTkNwvVEy918vemj2vRatseuNtzMc0QYhB5i6sBBDKsQUzJ8xg5xT3BMteEl2F6lKM2z7pQaKYywmxqEcRAGcxZs5Px5rEF0YshdbV5AKw3C9TobJDQ4PnVT1FQWU0q6aydo+j7qrmQJcabg+QINz4cQrXf81YaX+brrCTgBa/L9yfViioeYRmpNNZFOCdjVhlxL00uwPVUHoTMIHkWq3XdjrjIIInoxET1MRBuI6L0pr68iot8Q0d1EdB8RXaC99r7gfQ8T0XndXKfB3MR8ZxCbd1dx2Puvx7V3b8l1/OEf+IXSFnTBM25AfSFyOVU2ULvGGp0suyVCBhEVqSfLIC756m349xsfjrxW03oXTbTQISYaXiabcn0fpaD4zBMiFNc7SHNVjmGiieFqE8d/+EY862M34ZcP7FBtxPmYiYabmN1RapHmevJHf4UL/uMPEELMzToIIrIBfAHA+QDWALiUiNbEDvsggO8LIU4EcAmALwbvXRM8PgbAiwF8MTifgYHCfHcQ9wcFVjc88FTbY9nQXHO7bPSmfyZxtuD6+QYJ9ZelRDnYDQcRE6vZYOaBEJIBNTwfGwfH8eTuauT1Cc1BVJvZDoJF7DRW4PpCGWjP9yc1MIhZ0fBEE0+P1lR20vbhCfV96SGmuL6QyiCCY/ZUm3hoxyi8oK37XGQQpwLYIIR4XAjRAPBdABfHjhEABoKfFwDYFvx8MYDvCiHqQognAGwIzmdgoDBb+w1NFcbq0mj2ltrnkmwOjOTSvhKAZAxdR14Noo8dxJSGmAKR2k2K1Hkr4vne6k0fdddPGO0og8h2PBMxoTxyDU+EDMKf3MhRPbTE36W8nh9xHnKdXuI7Kae02mh4InJ/XDsxF1ttLAewWXu8JXhOx1UAXkNEWwBcD+CKDt4LIrqciO4kojt37tw5Ves2mCNQDGKeahDjgVHpK7Unz1uGpIM4aGEZQFSXiVdTu76fq8K6L3BMXQkxNaP/CxEWiAHy3rMchl7VXHe9hIHXDWirVNdqcL00BtH0tBCTL5QhbnTQZJBZUa3pY6gafoYN11dpvfw1VRte4jvhVhu6A2x6PrbumQjPFXwWc5FB5MGlAL4hhFgB4AIA3yKi3GsSQnxVCHGKEOKUZcuWdW2RBrMT8z3ExA6iJxeDkEbjoAUVAMlCLx2el49B8DkGx6aQQXjR0FJNM7hsNMfqLo750A349K8eST0HG+tsBhE+bu0gooK5Ds8XqlCt4fnq8+okzXV4oqmE7m17aur5hsYgGBNNL1IlXbAJdvDeaLtvX7FFIHSWc06DALAVwErt8YrgOR1vAPB9ABBCrAVQBrA053sN9nHMdwcxVpcGrC+PgwgYBIejdAYRn2Ps+vmymHhnPpUaRKg9eJHHQBizf3pEGtOf3rcNaeAahbG6G2QqZTOIdiK1XEvU6LPGUQp28Pr58mYx+b7AaK2J5Yukw94+HO76666XKsrrYaiibcEOWoXENYjNQxPqmOYcZhB3ADiCiFYTURFSdL4udswmAOcAABEdDekgdgbHXUJEJSJaDeAIALd3ca0GcxDzPc2VQyClHPFl3lV6QZhiKjQI3rlOpQYRFsolC8/iwm1vMd0xxruexg38RI4QU8P11XniRp+f5899YhIOYqzhwhfAykWymC/BIFJE+VGNVRQcSzEIdoAFm9DwBLYE3/WCnoIaIDTnHIQQwgXwdgA3APgzZLbSg0T0ESK6KDjsXQDeSET3ArgGwGVC4kFIZrEewC8BvE0IMbsnjBhMO6aKQYzWmi13mjMF3lHmuT/eVaa1QE9qEDkdhMsOIptBeL7oKATViInT+u6fjebugLGwSB5HU2MQ8hzh/Y3VXezW1lvNEKn17zs+b2FXcD8cYqppx9ZdH0+PhsaeEX+Ox+CuXCwZxLZANxgoO2h4fiqD0Cuui7alqtf5eyg5dsAgpINwPT/UIOagSA0hxPVCiCOFEIcJIT4WPHelEOK64Of1QogzhRDHCyFOEELcqL33Y8H7niGE+EU312kwN8Hi9N42pDvuqhvxwk//biqWNKVgDSLP/W0PDJCf8plMlkGw4d3dwkF87teP4uSP3qSMat5z6v9znJ4ZATOW/ozQGmsjTBz1ENBbvr0O7/r+verxREaaa7WpZRVpx3xr7ZM4699+CyCdQWzaXcWZH78ZO4ZDh7DuySGc+rFf45GnZNuODU+P4rX/LQMehy3rAwBs3TMB2yL0lwuoN5MahLz/cE0FLcTEn1HBJjQ9X7GRuuvPaQ3CwKCrmMqBQXpmyGzBeKBB5Lk/1hnYeLaqg/ByahB5qod/9/DTABARTvOck7WHWtPDgkoheE7eL2dNZaX3xhmRziA27a5G7i0rxFTNYBBb90yox6EGIR+zFtT0BDY8Pabes3677Ke0cdc4gDBh4D3nPQMvPVEmX27bM4G+koNSwUI9g0HoIaaSFmJquD4cm1AINAc9+2ouaxAGBl1FOHJ0hhfSJYx1wCDiNSG6LjNpDSIwvGkN4hgDgXHP265br6QWQs5ZUA6CQ1qBg8ga5xAX3XUGEDe8WQ4iEmLSHIx+PIeY2CDryQIc5gHCFGMOxTGbuuj4g7C4pwjbIvhCvr9oW4EGkRZi0kRqx1LzLCSDsFCwLTRcoRiN6wvlvIyDMDCIIRyKM/ne/62auU0GDdePhB/2BhyGyGPM4zUhulOJV03nqYNglkEkDXLW5zRQlsY9b6sMdgK+kOetN330xx1EEGLKqlqO3w+/TwiRMLwTDRfDE00MxcJk1ZiuoB/P4BATh7B0TURnTFsCxsBaDDuKJX1FWBZhcW8RANBbslFyLNSaHkbrbsIB6mGngq2L1JJBFB3JIPQ0Xg5DGgdhYBCDYhB7EWKa6iH037tjE8799O+mJPWWd5TtCgFFClvwMzQI3xfwRfvPjHfVelglDQMV+fqear5UWH23zkVu/SUHRCETYAaRVuEMJFuYy9YhPsbqMnOIsaBSQLXh4aR//hVO/OdfRd6ji9fZDCIaYuqNMIgJ7WfpLDg0NjhWR6VgoyfIwlqiHISDkmNj93gDQoTPM0ZrUQbBNr/henAsUuxD11yYZXZrHoRxEAZzFiGDmPw5xuv5QiN5sXOsgdG62zIskwf6bridMfdS2ILOIHSDmlfYTzqI9PvhVNS81dYNz1c757ord8PlgiV31sE1OUTTqtNqHLJ9RfS77C87qS0sgOwQky5Icy8mfq6sZQrpDIJ/ZuYwONbAkr7Q+HP7k76Sg6JjYU9QELikVz7PYnyUQVAixFQu2phoeqg1PQwEbMYwCAODDHCcfW/qIFgInipwEdfe1mbU3TCFsV1jPd3+sQPIYhCturxGry8/l942DoKvlzfVtd70lNOpu7JVRsmxUXLskEEEhraWkYEUn7zG54pPpesp2hjPSHONhpjS6ybiWUy6A2LdYazuYii4Ln8Gu8YbWBI4BQDKWYQOQt4fh54W9AQ6jp7m6tiRSmrbIvQUbIzVXTQ9od4zZhyEgUE60nbLnWJsihmE20HYSwiBh3aMpL6mx/TbOZs0QVr/TLYOTSijFB+lmQXevbODyGpSx44jb7V1w/OVblFvynkMpYBB1F0fvi9UWm0Wg4hPXpPHeonU0UrRwZOD6dlVepfXoWpTsYDUEFPwHH+Gaw4cwK6xBqoNV73PsQiDYw08uG0Yu0brWKqFj5gpyBCThfHgfOwg+P9IiMkORWohAMcm9BRtpaWwsG8chIFBBtJ2y50ia4c5WTS9UIRthzs2DuHFn/0DHt6RHHupG4s8gnJ4bLKB4Tu+dw/O/PjN8vkWPZp0sEPob6NBcCVvnmprP2h6x23Ea0EvpZJjo1ywUXd9jNbdtn2P0hhErZnMDBooO3hs51jiWCAqRn/2pkfwyi/fGpwnmcXEDOI5hy8FAJx3zAEAZHU0F8AduX8/Hn5qFBf+xx+xfvuIMvpAkkEw+JiBcgFE0RBT0Ql7MQHSAVWKtmJX7GQ5xJRnhOxkYByEwZwFZy/tjSA8ljM9My+Y1eRxWjxdLK3ITI+Lt7s/nUEo4T5mRHnXqjuOVsyrroRZaSSbWbv5DhgEO50BlbXkod70UHIsld2jh3syRepUDSLZ3+ikVYsyHdtEI7wfX8jvQggREa+5DoIdxOXPOxR3fvCFeObKBQAky+NrHr5fX+T8eohpaV+YxaQb8kWBgygXLFQKdmStRS2LCQAcy0JP0VbXSzAII1IbGETBNnCvHMSUh5gCp5VDg2BjlJarr8f82zoIzXaHDKL1+gAkZiDrYGPeToPg4/JUUjMjUCEm10eNQ0wFGWJix2hRtCtr1j2oczeTIvUZhy2Jvk+7h2rTRdGxUAgqkH0hHUGqBhE8V3JsLO0rKeM8ojmI1Ut7I9da2pcRYipoDCLQEZhB6ShoISaAQ0xhFlXoIDx1fDdgHMQchBACd20amulldBV57tGLGeMHtg5nCptZCLNApmYH1uyAQfDOtNpwse7JoUi6qm6Q22ksXkSD8CP/69i8u4rtWtO4rNDV4FgdDwXVwf05NYiRmptZce35AjetfwrX378dAFQGTq3poRGEmEqOjbrrqc9PtqQIv8t7N+/Bz+7bhj3VRmYW0/BEM1JbcMKqhWqmAgCVJfX0aA0/vnsryo6lwkiAZJMTKYVy/D4OD7GDG6k1VSqynrUEIGLc+bX+koOiHV5vccAySo6lpsfZFqHkWEGaazTEpDsRZmEmi8kggdse342Xf/FW1ftlPoLvMUvEBaIMYqTWxEu/cAt+fHdnXeGZQeiGYm/ABjMfg5DGaO1jg3jFl27FPUGDNyBqkDsJMYUNDJPH/eP378GV1z3Y9rzP+cRv8N4fydnWve00CO35PRPpYabbn9iNv7v6TrwvOGe8+rpcsFAuWKg3QwbRX3YiBXCXfPU2vP07d+NLv3ss1VnVmh5GJproKzko2ITlCysoOTaef+R+6hg2/p+64RE8NVLH8kU9EU1gz0Qz4ozjIjVvInj3ziGmvpKDNQfKwZhvPuswAMBRB/ar86xc3IOSY2HVkt6oBtEjHQczKEA6ghWLKli+qKJ6MQGAY8sQE0MxiFp3NYj2jeYNZh24Z8toSsOv+QKm7kPj2ffIu2RfCFTrHlxfqFbRecFprnlaaudBWi+kLLCD2BRkwuhN8Zo5xWQgPaU1jUEMTzQjYZs0ZuL7IlIL0C7EpD+fVfm8YyTa52qgEq2+ZgYxMhHWj/SXC9gyNBHoAp5a01PDNRy8OBrOAYI6iIkmFlQKuPODLwRBGtfPXXoCfnDnFnzwxw8odjlSa6JcsPD9Nz0b53769+ocrAkxVC8mbrfNDCIoDhyZaGKk1sRA2cEphyzGfVe9CAPlAt569mGKZQCyDmLdP52L3qKNdU+GrJhF6pJjKwZRsC38/O+fC8eiyPdgWxRxEAOJLKbuaBDGQcxBsBHI2tXNB7ChaBUyUkVfnpjUUHkAGKsHox8ns8gUqCymHMvgTJqnA8Okx7+bHYnU4c+tigddL2r808775xhj44yjLJE6WhmdFbKSjq/oyEpgDjENB2mjUZE6ZBB8Tl18HhxvpGsQQZrrQLkQYYMlx9Z6PYUDgg7frw/95UJEE0g4iOA8/DtVsCz1fLlgKQbBxpqdgu4cGFz7oW9EFvUGGkTBUuEjWwsl2dqXyFlMDFUo15AtO/Rw1FTChJjmIHTDOF+Ry0Fojek47JCV+ZIFFvnSUifj2DFcSw3rDY03cN8WGR5ytTVVGy7u2Lg7cbzr+bhlwy7lENgw6fFvvn+i1hrEPZv3ROYd+y0YRO8F0FoAACAASURBVMPzI9dIcxBrHxuMPOZK6XYaBJDd9XXXWAMFm/D8I5dFrssVxewgdJGatY/bHh/EFq2txa6xRnoWU1M6EnYGOtjgcuZS3fVQDoy/rhXEZzqwMeffKUfbpQ+UCxiZcDGiOYg84HMWHUuxM3Y4QJQJWDlDTAXbAmV1NtxLGAcxBxEWQ01tH6HZBDY8Wf38gejAIDYsWZkvWehk5sJzPnEzXvSZ3yeef8WXb8VF/3lLZN2eELj27q149VfWJip8v/L7x/HXX/sTblz/FIAw1BLpDxScp1KwMwVv1/Px0i/cgr/5nzsASOOSVgcRHh9lEGn3/MDW4cjjPBoE26YsJzI4VseS3hL+9jmrAQDHLZdpohwOLBW4DsJTnx8b3cv+5w58/uZHAQArFlUwOFaPOHOlE7gedo83sLAnaaw5fMOholrTTx0nmsUg+HfK0XbpCyoFxSDSnFIWWIPoKdroKdhY1l/CIUt61BodKzTJiTqIQloWk9s1/QEwDmJOQjmIecwgGsG9tTL4ukjdnCSD6MRBZB3z+M7x8Bgti2lkQjaPi1f48uyJuEHSq3vZIJcLduZ1uZhuRzDDuWBbqZXU4fr91LYb8XPqm1EVYmrBIPoCllHPcOaD47I30bMPXYJHP3Y+nn2oTD9lzaWnYCcZhNY5lVnb6qW92D3eiKyFHUmt6WPL0ARWBDOgdfDuPJxB7WnGv4WDKIQMwrEosksfqBQwUmtitOamhpSywA6tp2DDsS3c9r5z8LITl6u+T7pTiIjUMQ1Cb5HeLf0BMA5iTmKfYBBunhBTuFsPQ0ydahCBg+iguV7WGEvfF5E6CHZWcRbUU0jPmEoLMVUKdqYGES8MK9hWy+ry5Ozl5D2P1l0cvLhHPc5TB9GuHcfgWF0VjhWCUZpFxwodRNFGqWCj1gwZhD574akRabgPW9YH1xeREagci98yVEXd9bFSWzuDQ0z8u1Rv+olOrQCwcyzOIOQxvoiGl4C9ZxCsJ9iB4wlFai3EZMXrIJIOQr7HMAgDDWljJacLjz41ig1Pdz+9NleIKbh9PcTUzkGse3II24fDmDY7CL9N8zodWVXDTd9XO39PG+YSL4TryZiUVk1xEOWClekg4syk6Fiq+C2dQUSfSzvveN2NGNm+oJK64foYHKvjtsejGkXT89WchFYaxNJYa+tKwVYOolLUGISWxRQHF6M9NRJqBcwgeMLbykXZDoJ/l+quHzoNN5tB6CmpBStqKgfKDobGGxiruyqrKQ+4DkIvepNrDNJcY8aeGYVjWVGRerY4CCJ6CREZRzKL4M5giOlD1z2ID/90fdevo0JGeRiEHzKIdoVyr/jSrXjOJ36jHutGOU/tApBdNex6IYPw/TDcFWcc+k5QRzWlBXW5BYPgIi02LkXbioxhjcem4ywgzYmM110s7AmNua5BXL32Sbzuv2+PONKmK9Qxac5ZCIHB8XqikKxcsFTjuZ6ig3LBhhChw9ZDTIxDAgexY7gWieUXbQuPsYNYnB1i0secMjvQv/Kn4w5C+/ziDGKgUsC2YDDUZEJMldjvQKhBRK/DYaZywVJOpRCMH+Vji1OUop2GPGd+NYBHiejfiOiorq3EIDf4D3RvZw5MBmN1N3OM41SCNYhOReo8ISbd4OrGO6/DzWIQrifCOohgnCYQDR0B2TUXegM5ZiKVgp0ZSuQQk2449HkP8evEheY0xzNW9xRrkGsNejEFc5Qbnh/TSvwIy4ij2vBQa/qR3kSAdHyjgTPoCRgEEBZ+xR1Ef8nBfv3yHE+N1FEIKo5lDYWljPWKFAYRF6llg8DkTn1PLJnA0fohxXf2eohnsiK1DmY0cUfE3/3yhT3qPXws38OMahBCiNcAOBHAYwC+QURriehyIupv81YQ0YuJ6GEi2kBE7015/TNEdE/w7xEi2qO99gkieiD49+oO72teo5UQ2W00XL+jeP1kEaa5thKpw5TSPIwjDbqza6fpsGaY1blUhphCVlPPCDFlMYK0EFOlaGf2VeIQk15kpX8meo5/GtIdRDMS/2fj09RSZPUhSw3PV8enOQh2pvHpaRVNh6loDoKF9/iufKBSUOfYMVKDY8t2FCXHUgLvsv5SoqcRoKe5aiJ17LgDBsqJ9xVsUjv4Qmxnr6+vkzTXtg4iFsrir2jl4opiHXxsUTmIGdYghBAjAH4I4LsADgTwMgB3EdEVWe8hIhvAFwCcD2ANgEuJaE3svO8UQpwghDgBwOcB/Ch474UATgJwAoDTALybiAY6vLd5CzYYM+UgulGgV224uO7ebepxPpG6MwaR1utoouGpP9Z2BWnx6Wlbhqr446O71OsyxCTUtXjHGmcQmR1GYztzi9joJ+/pxgd3YOMumT3FhoMdhBACni8SxiYOvt+Hd4zi7k1DcIN5x70RByHPoTMHvcFhUxOp4xlkEw0Pn/jlQwDCqWoM3UBLBiEfc3eAvhiDGKgUVPdTzxco2JaqnygGTmxlSgYToIvUPkTA7MoxBhEPgQGSVUw1g1AhpkJcg0iK1DpWLu5RyQ2V2eQgiOgiIroWwG8BFACcKoQ4H8DxAN7V4q2nAtgghHhcCNGAdC4Xtzj+UgDXBD+vAfB7IYQrhBgHcB+AF7db674CNhjTsZOPo+62H3g/Gfzi/h34+2vuxhOB0csjUrsak2rkcCi6Yd6+pybZkC/UbjCv4+Nd8Vd+9zje8u112vn9SKuNkEFENYis7y2iQXi+GlwfD32N1pq4/Fvr8N9/fAIAErFoX8jPpF11LX9+533293jZF29VLcH7Sg5e8+xVWNZfUsan6QoVAtNbpDc9EZkQp+OPG3bh50GDvng77IrGbsqOrdhOJoMoOyjYljLGBZtwwsqFOOagBTgq6IN06upo91aGHcxznmh6aHg+hAgd1OXPOxS9RVvdw6KeAt53voykFyw9xBT9LI/YX95P0bFw8JJkWCsL2QzCUmtNw8pFPZI12ZY6lp1qWu3HVCGP/P4KAJ8RQkQqhIQQVSJ6Q4v3LQewWXu8BZINJEBEBwNYDeDm4Kl7AXyIiD4FoAfA2QASyigRXQ7gcgBYtWpVjluZH2D7MhMidVr8dirA1cC7xxtYvbQ3Vx2Er6V0skPJyqQBoprN5qGqMjb9ZQc7RtozCHY+HGJ6erSm4uiANLh6oZwSqWNOK20iGhBvtSFFZseixEQ5ZjBuLNTIoqrry8ls7RxEnJlw6Ki35OCjLz0OH33pcQCk0Wp6vlofH+f5kqn0ZaS58gCfm991ViL9tKzthi2LlLHjBn69pagB5e+K00sdm/C11z8LAPB3z10tGVOLnXSpEG3lwb/D77/gaLz/gqPxxqvvBACcunox3nTWYXhT0HSPP8J4FtOJqxbhsX+5AIRoOmo7lDIchB4mTAPrL5Vi2Bqc2dapqxfnvn6nyPOXfhWA2/kBEVWI6BAAEEL8eorWcQmAHwohvOC8NwK4HsCtkKxiLYDE1lAI8VUhxClCiFOWLVs2RUuZ/ZjJNNeG66HZBQbBxoTj6nlabaQJwq1CTLrz2Ly7imozKoi2Ev2bnq8+b2YQcbFaP0ZmMQX3EAsx6QxCNxRRkdpHwZE1A/HvOT7/mc/HQ2N8X/5uxDNi4nA9EXGKuoPQUbCjDoJDTPx5qRBTzJnztL60lFVudcH3zwxirN4MdslRA6r6HQUppbrBJqKWzgGQBrjW9NQa45scHid6xmFLI8+zk00byGNb1JFzAMI013gWUzkji4nB1+nRHATPwo6veSqRx0H8AID+zXvBc+2wFcBK7fGK4Lk0XIIwvAQAEEJ8LNAnzgVAAB7Jcc19AmoW80yFmLrAXDgjhx1FHgfBjtILRlnK9bUKMUUZBBs8Nj6tGIS+Dk5z1Qu2+PwRBqHmPcQchHYdPfYdF6kLNsGxKLGuXQnHlMIghGhrvDxfqCpsAIoN9ScchBXp48SGn++1FAzeeXq0jk//6hF89fePwfeFciR9KXUfbCD5f12kZvFZh84ggGTIpx3K7CCC34+4SH3vFtliJD5kyLa4DffUsGZ2hFkMwm5zHZ1BMI49qHvybJ4QkxNoCAAAIUSDiJKKThJ3ADiCiFZDOoZLAPxV/KAgdXYRJEvg52wAC4UQg0T0TADPBHBjjmvuE+BUxqxQRbcggorlbojUHFpIOIgWBj/sXJqvF5POLnaO1pXB4x1uK0amayGcDhmvh3C9MMTk+yGrSYSYNEe1pLeEzbsnZHvnDA0i7iDiWVQc2uHwhOfLdNu2DMIX2By0GgeyGUTRtiSDCBgXNzjk3wMpGNv4+X3b1Pf4gqP2x3jdhUVhfF1HOWYow7CJi0IwPa2/5OADFx6NH921FSetWgQg1CY6NdjlgtQgahkM4sMXHYP/vHlDQithYjJVqaQLewo4bvkCHLd8YeT5UkqzPgA4YeVCnHzwIvX49EOXqIyrS09dicGxRlv2tDfI4yB2EtFFQojrAICILgawq817IIRwiejtAG4AYAP4uhDiQSL6CIA7+XyQjuO7QkQCrQUAfwh6n4wAeI0QYmpnQ85hcB3EdDMI1xcQojv1F8wg+P+GG9RBtKi54M/BF0DDC1MYs6DHyJueCBlEEGJqJb5zaKKv5KDacFF3PSWoMlw/JlJn1EHonx+PplzaV8LOMTkXmYjQ9KQGYVOKg4gxiDDEFDoIX4hIN9A0eDEHMZYR/y/YViBSRzWIpuaY9NYZgBTmx+se+kpOaqdRpUEUo22wR2tNNU3t/g+fBwC45NRQX9RF6k4gQ0x+yCBiA6IuPXUVLtWuw+A01yljEI6Nn17xnMTzYR1E9Do/ftuZkccfe9lx6ud/ffkzp2RNrZDHQbwZwP8S0X9Chno2A3hdnpMLIa6H1BL0566MPb4q5X01yEwmgxToO+fpBBu8bjimUIOIGp9WjEDf8fNxzSCunibQxltTV2Mx8lahMw4xLeot4KnhesQYhucP01yjIaZ4FlN4HR4as99ACTtGamh4cgRn05UMwrHbaxDMnko6g/BF2zCM5wts1lppc8gsHhIqOOkiNV+3YFOiarvakA40LbwEhAaxRxV9hQziwIXJmgTGgAoxdWawSyrEFHxWbWpEGLadnsU01VAidZfmOkwWeQrlHhNCPBvSYB8thDhDCLGh+0szyALH3qd7YBAbhDyhrUefGsXP79ue+9yKQVTzaxC6g9RDQN+8dSO27ZlIHB8fbhOGmJhBtA8xLe4toeH5qoGcDn2tsg4ivVBO/964wnhZ8D+vSYrUpEJMV6/dqJzSrvEGerUYth7q4fvw/HwMYovGIDhklhSpLdQ9X30GD+8Yxffu2KS+o2IwQ1nHRMPDeN1NnIsRF6k55OQGNQ5ZYAYRz+xqh3JMpC7nHDGrCuW6GMYBQofZrcE/k0Wuuw4K194K4B+J6EoiurLdewy6h5nq5trogEGc+5nf423fuSv3uTl7qZMsJt1B6JlCH/nZelybMptadxD6jphDTGkFaer8gWFZHOSc66GZ8Bh9GE9Y1R0PMfH3dtzyBbjwuANx7PIBnLBSxqR5TUqDIMLu8Qau/MmDuOHBHQAkgzjqwAE89wiZvcJZZZxpw+mnjkU4+xnLsHxhegGZ64tIB1N2EPFdf9G2MFpzVd+iG9c/hf/3f/erEBsXremoNjyMN7IdRKUY7UmkV1a3mm/A31Wr0GMa+ksORmuu0rTyMggW+tvpOXuLSkaIaaaRp1Duy5D9mK6ADDH9JYCDu7wugxaYqVYbHL/tpPOpyLnT48Zz7CBUHYTrZ55Db64XL6iL6wPynLEQUzMmUrdgZDqDAGQWFCALq9LWoKfeJrKYPIHDlvXip1c8B8cuX4CfXfFcrAqKrao6g7CtSFYLO7jBsQaW9ZXwrTechuNXLFCGm1Mo2UFYFuF//uZU/L/z01uoeb4fEe4Hxxpwgh5HOgq2lWgtDujzkEMGwTvgasPFWD1HiClwEHq7ilbN5/i48YyW61lY0lfE4HgjM801C9PHILrfV2kyyHPXZwghXgdgSAjxYQCnAziyu8syaIVwYNDMMAgAuWshsmYExM/LxpUNEbfa0FNY48gKMQHRfkH6dQC5W5O9hYKq3Ur7EFNNOQhmEDKEdeT+YUsyHmnJ6+fzxdfGxl8HN9wLQ0xBoZxmMJhV8QAeIFqkxQzC1RgEkB3X1ivQAckgelNE5YJNSiPSETIIUkZdhcqaMsTU3kFwo8FwpGZLBhE4CP2zzoMlvSUMTzSVU4uL1FnIqqSeaszlEBMnSleJ6CAATch+TAYzhJmaSa3vNtOuffemoUg/JSBfKECfa8BMQheUs1JdIyGmFAdx9dqNeHznGL72h8fx7h/ci7s3yV6QvSVH9hbKmeZ62+ODuPYuGbLifkBbhqooOhZWaRXCuiPQU1uf2DWO//j1oxHmFzc4bByrWo1BwaaIjtD0BFzPx1C1obQLfeoYG1ZfCHhCZPYRYnB6MO9aB8caqQa9HYMoaiGmZUHFb7XhYbzuZWsQKospOQSnFYNYoBxE5wwCAHYEs0DSUm/ToM9j6CZKjgWi7jOVTpEni+mnRLQQwCcB3AVAAPivrq7KoCVUu+8ZymIC0h3Ey754KwDgouMPUs9VGx4WtmlVw8aH2ygAMQfR8FJ77kcZRHRHORjE7a94weH4/M0yp4LbFfSXHclaGp4UWFnczWA7l3z1NvUzdxTdumcCi3uKOOfo/bBxcBx3bByKOKlqYDxZZP70rx7BmYcvwckHL0bT8xMGhw0WC9sNzmKyogxirC61ADaUOoMI7yOayZW1+3V92Q5kQaWAXWMN7Byt46AUvaLoZDiIwLEXHAvFYEfOhlhmMTUjrcMj96uN3mQMlAvYPlxraST592C8Qw2C04m37pH73U4ZRLdDP0SEl56wHKd1sW3GZNDSXQWDgn4thNgjhPg/SO3hqHiqqsH0Ikxznf0hpjyzIzh8sXJxBaO1ZtBbSag+OFmprlkiNQCVxTRe1wrcguv0lmzFIHqKtjICeTSdRT28E61hQaWAFx97ID7z6hMARNkS3/dCLbZ+6wY5jc31RMLgsMNgJ8WtNvSQA3dbBdIHzOh1EJ4vtFbVLRiE56uwzWjdTe1MqrcR16GL1Oyc+ssFVAo2JhouxhvZDCJeSQ2Eob48DKJTMOPi34vcaa7TFGICgM+8+gScc/T+Xb9OJ2j5KQkhfMiW3fy4LoQY7vqqDFoiHBg0zWmuXmsGkYY8ISbena5c1ANfAGMNNzJnIKujqy9CQzvR9KCHb7cHA2RGtfAVO7jeohPUQXjoKdiRCuR20HfIqi9Q8H7dGTKb0HfDtz4WOAg/ySDYAPF3qgrltJtqeEJ9FmndP9V9iCiDiDsjjkpxiEk3ummzDbJ2z6FITcrg9pVs9BRtDFWb8HzRNsSkj95UIaaWGkT+8Z46uN6EHUSnaa7dDjHNVuS5618T0SsorRzSoCvYNFjF5256tG32znSL1PowHg4B/e6RnfjJPcmUUjZOd2zcrdpSZ4GL47jj53C1iabnK20gK9XV9cOxmhNNLxI/Z+OVFhrpKwUhpqaLisYg4hXiNzy4I6Gp6OM4VV8gK3RSDHYW3OtoxaIK1m2SYaiml9QgwhoGjUHYlMIg5HmZQegaBe+8v3HLE3hox2imBqHXS9RdPzb8Jl2DSMOorkEEx/QWHVSKtprvnClSx+oggDB81IpB5A0NxbG0N2QQMtafz5xxCK+bYz1nM/Lc9Zsgm/PViWiEiEaJaKTL69qn8csHt+MzNz2iujXGEW/zPF2IMIjg2q//+u34h+/ekziW/wCvXrsR/3L9n1umuz76lDRmzzpExl83PD2GpusrwTOtahmQO2D+w601vNSuoboAzugtOaoOoqfoKAMfZxBv+tY6/P01d0eMnP4z77Y5tKM7Mk7DfNe5R+K8Y/bHBy44Gg3Xx11PDsH1k1lMvAZXMYh0DaKmGEQy64UH5/z4nm2R15LhLG4C6CcYRFoIR9/Rf+ovj1fzD8ZqyTTX3pKDnhwO4tBlvXjh0ftF+gwN5GAQAPDqU1bii399Ustj4hioyO95vOEFgnA+BxEyiH1zf9yWrwkh2o4WNZha1LTh6mkIezHNTCW1vHZr9lKwLNTgY8dITfUlShsHCcjQyzNXLMBzj1iKomPh1sd2oekJHLqsF/ds3qNqDuLwfKF2oBNND8tTJooNTySzXfrKYYipUrTVbl53uNuHw0rsRb0FxUgqKTtejvHr4TT++aSDF+GKc47AaK0J2yLc+thgaiM9NZjHi4rUugNo+mGIiUM6OoOIOx2uNo6HR3hSmpsWYkpxsvp5zz5qP6w5aADnf+4PKnxX0Lqv9pUcVIoOtgYtPLJCTL0lR81zUNcO1mG32d1/4pWd9yAiIizpK+KpkXpHLMSx01nYvoK2DoKInpf2fHyAkMHUgY1AVux9piqp65FK5NbOqeBYQD10dmN1N9VBjNVd3Lt5D9501qEoF2yctGohbtkwiIbnY/nCCkqOlVq1DISjJwFEJsPpSMvf7ys5aHqy+dySvqImEIf3tDbQC+LP61k38dbTaSEmNpz95QKeuWJB4PxSGESaBhETqZuurwq9wvbQ4Tni57x/63DwfJJB+BaFE/UqSR1Ahz4LoUebH61rEBEGUbAzq7Jbga8dnysxVVjSW8JTI/XcKa5A6IBnW4+k6UKeT+o92r9/AvBTyCFCBl0CM4dMBpHRi6nW9PD+a+/HrrE66q6HK3/yALam9CSaLCIMIuac4iGkuFFKK1wDpEbh+gKnHyrbRpxx2FKs3y4jmEXbwopFFWzePYGJhocPXHu/6tUERENMAFAuJh0QOwhu0UAkwzMNz8d4w0WPxiA838d3/rQJr//67fjMTXL8yOLeYsQxOlrGzkBLDSJZkHXGYUtw75ZhNRFNBzup8bqL9197PyaanpoHwXB90TLEFHcQXMwX3/3awShNZjm9JUcJ/Okidfj+kmNFWnMD8nvi76Gv7ER0hXhn2Fbg76hVe5W9AScYdFJr0K6WZL4jT4jpJfpjIloJ4LNdW5FBWweR1c31rk1D+M6fNuHUQxZjounh6rVPggB8+OJjp2RdjRYMQs/g8VMarqW1vgCAjcEM6jXB0JNjl4fDTwqOhZWLe7B5qIp7t+zB//5pE846chledMwBAKRYr8erD1pQxqWnrgQR4Tt/2iSvGzimxb1FjNTcSFHXyISLSiHUIJqewDfXPoHBsTpWLenF5t0TSpjff6CEvzlzNQAZZmpMhKEZIkrMc+DPQ9+tHnXAADxfYNdYIxH2YYd64/oduGPjUPBcrNWG1jAvbcCM3j5icW8Rn3iFDMXEw1mORfCDGc3yfTYqBRvjDS+VQTz/Gfvhns17cNQB/SAKW3FwAkC5YCtH2FeyI2G4JYE4nAfsnLrlIC4+YTlGai7OOjL/9MnpqoOYrZhMztgWAEdP9UIMQoQaROv8/3jWzZZgx7hrrI4HgvDCqiW9U7YufdZCXIPQhWRPJB1EFoMYju3wl/aFBqVgW1i5qAd3PTmk3s/irx/MptAZRMmx8OGLj8WP796qHARjYU8RGKzK3a7NBq4Rq4PwsWWoitc++2B84MI1+OQND+HLv3scAHDZySvw5mBOcU/RxvBEU60ZkCEinUFMNELjy9B31nGDw5+XHrOXDiI8Rq+DUAyCkucAgA9eeDTOXbN/4nkgGJXpk3JixYAVjDe8yD0xzjpyWcSo8jS2wfEG7KB3kwoxFUMGYRFatu6Oo7fYOq15b/HKk1fglSev6Og9RqRuAyL6PGT1NCBDUidAVlQbdAlsXLJqCLJmUrOYOzje0HahU/eL3YpBPD0adgXlQi0dWc3VRiZkvx6m8Es0B1G0CSsXVzBSc7EtqGvgaWac6qsziDQjy+CmesVgPCbfQ08xrIPYMVxHreljxSKZpVN2bHUfuqHnHbK+2y5YViqD0Auy9J11IsTEQrn2uRZjDKKp1UGoNNdIiCmqFcTPLWdLyxRbzyctDBaGjRb0tC9EYwbRcH0MlGXvpqL22XNtw4ELKh2Fc7jDa7ccxGRgQkztcaf2swvgGiHELV1azz6LezbvwU/u2Yor/2KN6j2U1YNIn1qmg8XcJ3aOK+1BN+qA1Ao+/NP1eMVJK3DcigUdrbGusYYfrNuMJ3ePq8c7R8PZxk3PTzCMsXr6vcR34tzKApAGnxnFQ4EuwUyC711nEKFQmox7cwV0UWsLASBSB7FxUN7PysUV9RpDDxWldSB1bIq22tCMb/g+jXHEQ0yWFXmfvH/KTHNlx2On1EHItWtMJDh30bYghK+ymHQhne8vTeiPQx+FyiJ0MZLFJD8f/hzzolIImiZOc3ZeK5gQU3v8EEBNCOEBcl40EfUIIdJTSwwmhcuvvhNPj9bxtrMP1zSI9BBTKFJHX+fpYH96QsvAiTmRkZqLb9y6Uc7G7dRBaOv5yT3b8MQu3UFEGUS8T1RWiGmk1owYWj3TqWBb2C+Yv/vkYDVynjQHwbvVtMwZbrJXsK3ETpuNMN/PyoBB6IPtI6GiwJAtiDgIK3KP3FNJ30Hru/p44ZVlESyKajkFx4owhKbno970QBQ6niyROpVBOBZ8IcMlnh1qJkXbSmVFrVByLFQbHvoC537Wkcvw2mcfjBWLKirT68AFnTmIow7ox2VnHILXnj57pglMV7O+2YpcldQA9G+6AuCm7ixn38Vi1SV0QjWea5vm6qUzCL3ArhljEJzV02qUZxbirbt1dqI7CNcXCQbRSoNIy5wBpEFbEKRgcosEFrvZ8elGUd/FxqGHmOI7bSK5o+bPj0NMlYiDSIaKBiIhpqgG0UhxEPr50mLajiYcA4AQ0ePcIMRUdmxV6BVp1udkXMsOjRx3iLUtUmG/omOh7Egm1ZOSCZaGkhOGlABZBf/PLz0Wjh06NW6QlxeWRbjqomNw2LK+jt7XTXCa63T0YpqNyOMgykKIMX4Q/NymP6dBpzhwgdwpb95dL+MM6AAAIABJREFUVZkz9Q7qIGpNL6IDMOI7+WHlIDqP88bDVbrD0CeTeb5IOK+xLAYx0UzsWsO5AKRCHltV871QpAaSIjWQ7iC4RYbeWA4I6xq4cGxpX0k5AD2sVIqFmGyLImM/HdtCWrF4ti6Q/NMraGEfQDpdOxFi8iPrSmv3Hb9WGGKS9QqOHc26Kjk2ygUbCyqF3BXGYdZS8rPmhAVdT5qr4I90trXhni7kuetxIlJ17UR0MoCpS643AAAcGLRZ3jxUDQvlNGOxabCKt3x7HWpNL+zFpBl/NqC8a1vaVwpEyRiDqE3OQXz194/hh+u2RBritWIQ8W6vYxlpriMTzUTcW1Uo25bapXMtAu96+d5LuUVqXYNIGlIuhNLj5vouvBzTLVicZaTtMB2LIsf0RHSBDAahfeejNTfiAFik1tcVCTFl6B1WoDk4tqVSZy1KZjGlZTBlgZ1UbzH5Hv7d4vbqcxmcJGCymLLxDgA/IKJtkCNHD4AcQWowheDd7+bdE6EGoYnU6zbtxi8e2IErXjAOtvn6Ln1PEFY6fL8+7BrbjZWLK6jucBOhnpFJMoh/uf4hANLwMBvQHUQki8lLMojMLKZassX0QMXBjhFp8MsFG0XHUtdisZt1GN3Ys87QU7Txj+ceicd2juEnQV8iDjGV7HiIKVpwxvoDENVDdAbxqlNW4sRghjQjraV2fNep7/xTGYRNqn3FhccdiLe/4HDcu2WPep1F6nKWg9CcVCUWKnIsCl63lAahO4i/fvYq7EphoFlgBpHmjN957pEoOTYufObcnyumGIRp1pcOIcQdAI4C8BYAbwZwtBBiXZ6TE9GLiehhItpARO9Nef0zRHRP8O8RItqjvfZvRPQgEf2ZiP5jvneT5bDRlqGqluaaTCsdb7hau+/wdQ5HLQ+m86xc1APHokQ6Kk9sm2wqoR4q4o6jQJRBNH0/wm56i3ZqFpMbDMCJdxBlBsG7Nt2BcIgpTYNYGfRiIiL8/TlH4LjlUoQv2pYyZAWHYmKufJ4Nts4gyhki9bMPXYLXnn5IZM16Owped5xVEKUb8/B9UkQGgI9cfAwW9xZTKqn9bAdhpYeY5PUsJdDLLKZoaO7sZ+yHvzxlZWJNWWCH2Z/COpb2lXDlS9ZMuvPqbEK7mRrzHW3vmojeBqBXCPGAEOIBAH1E9NYc77MhZ0mcD2ANgEuJaI1+jBDinUKIE4QQJwD4PIAfBe89A8CZAJ4J4FgAzwJwVkd3NsfABm/z7qrKgNEZBDuQsbqrtAc9zZVDMMuDwqSViysoOlYixDS8FyJ1HA3XVw5IdxDxXjoHLaykitQsOMcZBBsd3uHqoQ8OVaVpECsXR6UxZmUlJzrvOC3E5KQyiPQq5TQ4KRXNaV1Jw6yYtBBTMlykh5gaLjOI8LwsotoWxWoikr2e2ElwN1fGZFpZhyL13HcCraBCTEakzsQbhRBqZy+EGALwxhzvOxXABiHE40KIBoDvAri4xfGXAriGLwOgDKAIoASgAOCpHNfcawghcOlXb8PND7W+3PuvvT9Rrbs34FDQ5qEJFU7Rw0DsQMbrrtplur5QPZC4ypk7mkoGIVMRX/XltTjnU7/FuieHlAbRjkF89Gfr8cN1W9Tj3pTslomU9cWfLzoWFvcWU0VqXktcgzggEOzZ6ekOZCzGIHTjptdQAGGaaqkQpnHGRepK3EEszggxtdkN64yAr5tmVPi49BCTlpFlJ9NYXV86CD18xG+xiVrGyR1LitPcADDaJrxzB8GfTVa31vmCUKQ2DiILth7eCZhBnvy15QA2a4+3BM8lQEQHA1gN4GYAEEKsBfAbANuDfzcIIf6c8r7LiehOIrpz586dOZbUHrWmj7WPD+It325dLP6dP23C+6+9f0quCaT3V4o4iMCBjNfdyDH8PjamJ61ahLeffTjOO+YAFBzC9uEJ3L5xNx7bOY67Nw0pDSIrQwqQTvKa2zfhV+t3qMfVpofDlvVG+vBnjaPgdb/8xOX40EvWoL/spDII1WYjxiDed8HReNvZh+FFx+yfeJ21DO4WukzLlIlHIUMGYatQUpZIrUJMGoOIiNRtOoAyg+jVCu/SMl/4ufQQU5IB6E6mqaW5MphhEEWdSfK6kkG8/ezD8cbnHop+zbDnHb+po1XG2HyCpRifCTFl4ZcAvkdE5xDROZC7/F9M8TouAfBDrRjvcMh+TysgncoLiOi58TcJIb4qhDhFCHHKsmX5G3C1AqdutkprG84Y5LM3SBt3qYeB+PXRWsxBBCEeNso9JQfvPu8ZWNRbRMGyIjOZa00vV4hpqNrEeMNTekXd9SEE8PKTVqj+Pq3Aazl+5UL89WkHo7eU7iD4/AmRulzAe847Sn0HaRoE1ywcsjQ745p3/aWCFmJy4mmugQYRFKrpvYM6YRBsyJf0lcK4dcrvEF87zeCwk7I0Y29FspiSIjUbsDgrSFtfwSa86JgD8Lwjl6nOpgBQsjsPEymROiWLaT7BNnUQbfH/IHf2bw7+3Y9o4VwWtgLQVa8VwXNpuARheAkAXgbgNiHEWFB38QsAp+e45l6DQzWtfiGyBtjsDdKmw+kpj2GIyYsxCGnomUHosfJCrLp3oump8Z6tQkxsfNmZsBagVx23Ajsf/gx7S05qiClkEK2NjB6CanoCdddTraxXBSGh/pSdLO+MS44dTBGTBrqQkcUU7x1UychiSgO/b0lfsWV7hjRmwOCJcPoanEgvJj8hUvP3YRNF9IrEdS0rct5Iz6u90CD6OkiNnYtwWrDBfQF5sph8AH8CsBFSV3gBgES4JwV3ADiCiFYTURHSCVwXP4iIjgKwCMBa7elNAM4iIoeICpACdZ5r7jVYYG1FKdmA5m1LkAeeJxD/+66ldE8db7iqDkI+H4SYmkkH4dgUMcy1pp+rUI4dIGsE3B+op2jnKqTic3PmR1/MQfzx0V04/V9/jff+330A2n+O/DrXeJzzqd/h5oefxn79JbWeow8cSLyPQzE8YrK36MjOo2oXHw66KdhWoneQ/lm2F6kDBtFbAv/qpP0OcbZTmsFhBqEznOhMapEUqa0wxGS1cN6ySWG6XjMpB1HILpSbT7BaJBXsC8j8donoSEjh+FIAuwB8DwCEEGfnObEQwiWitwO4AYAN4OtCiAeJ6CMA7hRCsLO4BMB3RXTizA8hHdH9kIL1L4UQP+3oziYJxSBa/EKwAd1/YOoKgVxfYP/+shpyT5TOIMZiGkRTMYjoIBkgg0FMtBepeXfOzoTXUckIJ7zy5BUo2HKH/p0/bVKOjXfJZa2OAQDufHI3tg/XcOmpK3HAQAUHDLRuCc0MY/+BMnaNNbBlaAJbhiZw8sGLcOjSXnzggqPx8pOS8lbIIOT/H3vZsTjqgAFlgHWx993nHYm+UtRRWUEr61bjUhlsfJf2FbHh6bD3UdZxqVlMVvJ9AlG9KVEop2UxtcJ7znsGFmqdWvW26u3em4Z4q435ilbhwn0Brb7dhwD8AcBfCCE2AAARvbOTkwshrgdwfey5K2OPr0p5nwfgTZ1ca6oQD4+kYUvQFK/Sxmh0As/3saSvqBzEgkohMsmMmQKL1Gy4PE2ktigudMoh7Qxdg2g11pEd4Fhd1lyoEFPG/T7n8KV46YnLse5JObAo/Awt9b8vpI5iW4ThiSb6Sg7+9eX5Zgszg9h/oIwHt42o51cuqoCI8MbnHZr6PjZibNwvPkE6kQktZMZ4wVHp2kq5YKPu+m0zfdjILukrql1nWrV00Q4ZSxxpAna84LDaSC+Us9owu3OOjt7fkg77JMXBzrdvvqe5qqyzfZNBtPqtfzlkBtFviOi/AoF63n9K9ZQma3FwiKkxhW2J3dj4zEU9xew018BBAFGRuuTYsWKs6D3Um74KGzU8P1UYB8L7E0KK4tUUg6ojbAZHai1AaCB5HVyTMTKRrJ5uBdYg4q0b4nUPcZS0EJMONsA9OQTWckGGpFqFb4AwDLekt7VI7bQwOE6KBpGmTenMhx1Du/XF0cmktzS0qqSeTwgHBu2bDCLzroUQPxZCXAJZRf0byJYb+xHRl4joRdO1wOkGp3+2CjExg2hkzGvIgx/dtQWXXx2O2nA9EbnmgkohEgbyglDSaE1qEBwDZqNbd/2EkKrn2lskNYVa01eFaHEd4q+/dhs+e9Mj2Ly7qvSQkVoTE01p/OLtG9R1rGjOvvoMYzvi792xGSf/86+wp9pIrcDNAg+x4fcsXxjWerSCSnONMR/HtmBRPgZYKdht9QcgDMct6Su27AAahphSnIc2tyE8PnkOfT22JlJ3gk47rcbBm4X5rkHw9zUZnWY+IM9M6nEA3wHwHSJaBOAvITObbuzy2mYEzCBaxWXZGMTbWHSCuzYN4bePhLUbni8iRmNRTwEP7QgNuN5qw9UZhB+K1OVYKmbRjjocrnber7+E0ZqLWtOL7ABv2TCIWzbIWRInrlqIuzftwfBEU2MQ6b8uRSdqECeaUQ2C/8g+dN2DAGQIi5vn5cGphyzGh16yBpeeugoHLqjgVc9aiR/euRnnH3dAy/dxKCbNwBe16up258hTJ8C/E0v7Smo3nxaWasVMCykM4rjlC/AvLzsOQ9UGPnnDwwCi7S0sFWKSj7/y2pNx+H7t22Uv6t07B/GyE5djv/4S+nMMGJrLeMnxB2GgUpjShJS5hI7cohBiKKg9OKdbC5ppsNjb6g+Zd97xNhadoOH6aGgaguv7kR2nDDH5qlLa09Jc00JMdddLMgjN4QxUCtg1Jtsws0DZSqg+/1hpfEciDqI1g+D/WYMoxNoU8C5s42A1cwZE6vltC39z5mqUCzb+9jmr0VdycNmZq9saJ/480orcCtqQnFYoF+xcPYW4pkOmufK6szWItN8dxbi0vk5EhL86bVUk6ygyqMiKhpjOO+aAXPMU9lZ03X+gjJef1Nl857mIZf2ljudYzyfsm7ypBZhBtBKluFfS3jgIvg7HrlnA5V0kzy/g4yJZTEIoo6XXQSRi7drjBZUCdo9LBsEOIqtYrr/s4IzDlgKQO+MwiyndUMYzc2oZDIJDOg3Xn5YdmV5JnfZaPgZhdcQg2mkQ/P26fvKzj2s2OvRwoV4XkjeLycBgMjAOIoZ4Dn8cni9UymbdzecgvvCbDfiH794deY7PwcbX9aUGsSbI52cDql4PnNFYzYUQ4e64GROpdRRimgbrnRx/fuGnf4dv3PJEYr2nrV6iQhAjtfYMgo0eGylmJmHIhEXh8P15Zh/vLYpBY7q0dZcLdiKtNQ19JSeXI+GBT4t6CmEWU4qh52l18XAgkC5SM3QtQmdfVs4sJgODyWB+K0yTQDsNou7qukA+B3H3pj14aMdI5LmGYhDyfMwgvn7Zs3D7E7uxuyrDQdz6g0NMbHy5xYGaPuf6iVBKIWPXqVfRfvKGh3HZmasByHtec+AA3nfBUaqDqmQQkuWkGTX9OmzgarFiwziDANpXT08FiAhfee3JOOag5Oztf3vlM7F/m/oLAHjXi54RmfKWhf9942lYv20Ejm1pDCL5O/SBC4/GCSsX4vTDliReSyuUY+jfpc6++OnJEIifXfGc1CwpAwOGcRAx1NvUQfCOvr/koJpzpkKt6SXYRj3mINxApF7SV8L5xx2oOqnyeuKjQ+Mtseuun8jKcTJ2nXqRFHdOFULA8wXOPmo/HLasD74vZObThBSzKwU7M5XSiTGIeLsSlZ3jpBu5biKe/8/gEFo7pFVop+HABRUcuEBmV7ViEOWCjVdkxLTDEFN29hMQdfbWXoSYjl2edJwGBjpMiCmGdpXUrD8MVArwfJFZS6Cj2nAT3VNViKkZ1SAYbEwbHjOMqIPh9MKqYhBeSr6/fGxRNPNFL5Jio8a3oYue/eUCtg/XUG16LcMs8QZ0HKYLi8LYcYT3MB0hppnCZKtvw0K57PoJIPpd5i2UMzCYDAyDiIHDI1k9h9j48R9p0/NhW61j1NWGpxwLox6Ejj780/W4b8swViyqRIe42NGMoHhFLTdJ4/BPrZmsgyho2UNlzXnoDILDLCya6k5qcW8R/3eXZDIrFmX3Z+TQiG3FQkwxg1fVxo7O57RBlq86nSEQZjGlpOXqNS3ad8TOyDgIg27AMIgYmEFkMYMJNeVMGrhGDh1ioumh4YYpq0CoHdy3ZRhAUCgXGTrDDCKaxcTgFM8wxOQlNAI2zCXHRlljAHqRFBsxvl/dQXz85cfhxFVy9rLOIG585/PwrTecmjiHExOp46M39Vh+J2mucw1hoVynDCK7fiKL0SoGYf6SDboA82sVA4dBfJHuINiBsMjazJHJpOsEjLhjqbtedE5wYCQaboaDKMU0iBQGoURPx1LOo1ywIsVxnAXFDkI3RKcdugSvDuYUb99TU88fuX8/nrl8oXqcFKmjtSRFxSBCBzGfGUSrgUGtEIr6KRpERiXvZCupDQzywDiIGFgUzmYQgQbRCYNIcxAxx1JteKlzgvk9nu9HdvHlog2L5LmFEEEdRHolddG2VFVxX8mJNHuLz7eOi50s5o7G5jnognO8fUS84SE7Kv0znY4spplCOOi+0xBTizTXDIpgaZqRgcFUwziIGLhVdVb6H++OOUTSdFuL1EIIFXvXheq0rCbdOLOx/81DT+O4q27A0HgTC2MVtD1FB+u3j+AZ//RLDE80MxmEnMksf+4tORE9grUNN8NB8IyE+PO6g2DDxofUY7Uk8R2xbREWVvau1cNshspi6rB/T1qrDfWakxFiMhqEQRdhHEQMzCD8LAfBIaZAJG7HIOqurzKEWjEIAKkM4s/bRzBac/H0aA0LtP5FNhEqRRsPbB1W50oUymkhHg4x9ZUcOLaFH7z5dBy0oKxqOfwMB0FEuOEdz8Nv3vX8yPN2MKKTz8/HOhYl5kHoBu+kVQtxzRufnavNxVxF2AG0QwbRIsTEr8VrXUyIyaCbMA4ihnobBqHqIIIQU7tiOX3oj15kl+YgbDvpIHhE6Hjdw4JKtElbT9HGrrG6ei6rrXWpEIrUrD8865DFWNxXVOt3UzQIxjMO6MeqJcnOqbxG/T22RUrXSHMQi3tLOHX14sS55hPYaHfaAbQVg2AnHG+YGLb77niZBgZtMX8DwTnh+QJbggE5tkUqfp4lUnO6Kqe5phl6HXoxnd77KI15OJEQU+AgtAlwvUUHtkVB51dCpWBD92NZdRClGIPQX2fHwPpAJ6GKgm0lQmOORWCXpZr1RZjR/N/p8kfY6QyBeFpwGuLFkKYOwqCb2OcdxJ5qA2d98reJ5+N1B4xaI6ZBtGUQobir2EnGsB49i4l3n9wEDpA78t6ijZGaC8uiRJgmPhbT0XaylRiDAKQBb8ZaeXQyOavkWKg3rUjNSGiwwlh8mqA9nxFmMXUaYspmHvzcaTH2pUJMRqQ26AL2eQfRW3Lw6VcdDwD4px8/oEZ0ZjKIZmd1EHpqJ+sbWe8ppDCIMS17yLEt9JcLGKm5sCnZhC7OIIqqm6mlYtf6iEjHpkQard3BrrdoWwkjqI8ZVc9FxqDuAw5iryupk8b+gAVlXPvWMxKtP9gxZBV2GhjsDeb/X2sblAs2Xn7SCrz8pBWREZaZWUyuTEflHXnbEFMjmbmU9Z40DUKHYxF6AwNvW0ClEPXvcSOh90DiEFOvFsN2bEv1eFJprp2EmBwrUQzGzkB3doV9jEG06sXUCq3SXAHgxFWLEixRtfs2/sGgC5j/f60dQHcQehbTWN1VoZiJho9ywdYGvwRdVhteYoQnP8/gEFNWm/C0Vhs6bItUiMi2kvMMBjXBGoi22mCH1lfWQ0yk2ohn1UG0gmQQ6Q5Cdxx6Dn9xH7BkKoup4xBTew0iDv5oTYjJoBvoqoMgohcT0cNEtIGI3pvy+mf+//bOPciSqjzgv+/2vXceO/ua3WWF3WV3Qcj6AnazoohGwbexXCytOMRUjGUVUQEfpQlYVhG0tCqxKtGQkChWSGliQIIvKkEFAQkJKA+zvAvcXUntUuC+2GUf87r3fvmjT/c93bd7Zu/M9My9M9+vamq6T3ffPmd67vn6e5zvE5Ht7udpETnk2i/w2reLyIiIXFRkXyFZ49jXIN59zT184+6dQKhB9FaCOC49EhyX/duvuOJ7j7R8pq9BRE7qXA3Cm0hFpEWLqJRKsZM5KLXWZ4gys8bnB00T00BPmWpQYtXiZh6mciCxryVrJfVkVDJMTEHGW7B/zkLSILKE/EREDvx2rjMTk1EkhfkgRCQArgXeCuwBHhCRW1T1iegcVf20d/7lwGbXfhdwjmsfBHYwCzWwo0Vh0JwwVZXdLwzzrEs1MTJWp7dSaikd+cyBYwxkZCg9nuGkPhENAsLoI1+YBIHEJqKSNM1cb3nZSXzqLWfyilOS9mm/4PqinjL/8YnXs94LVw1NTJEPYvJa3Gmq5SwNItxfkpFxFNpfPNaNTFRydCJiDaKNSK+mickEhDHzFPltPRfYoaq7VHUMuBHYNsH5FwM3ZLS/H/ixqh4voI8JfA0iEhBjLuIoWh08UgtrI0QTYzTZvzhS44gXcRQxnLF6OhIUi6rZIYsRrRqExCaismdiWtJX4ZVrlrb6IOJUG+F5Z65enFhMF5qYwnFGTvl2BUR6Eoyu95PxiYiX+nsBCIgpOqkn80FkUbIoJqNAivy2rgF2e/t7XFsLIrIe2AjcmXF4iGzBgYhcIiIPisiD+/btm2Z3Ya2vQWjTtwDNiX5k3PkgykkN4vDweBySemRknBdHwu3MKCYnKJb1J9NNtGgQKQEReCamUqm5aCqvtkJsYsqpqVwOSrEPIhIU7ZiYqkGpxRxSzhAQ4b0iYTX/JzKJBUR7Y52oHkQe5djE1NatDOOE6JTXuSHgZlVNeHlF5GTgVcBPsy5S1etUdauqbl21atW0O7E2Q4OIJvjIAT2cNjHVGoy4dN4vjoyjqpz9hds46+rbEtf7nxEJiDWpGguTaRDlwItiEokXTeVlRp0ofXR0PB3F1E7St5UDVZanhFw0hnSfpjL5dStTzeYa/S1XLDrxPFXTqShnGJNR5DqIZ4F13v5a15bFEHBpRvsfAD9Q1VbbTQEM9JS5588v4B9+vpNbH30O8AVEOKmP1OoM9JRjW/p4XePVzuN1ZXi8nljdPDxWo7dSYryunokp/H35hS9lSW+Fbdf+D9A6obQIiEQUkyRMTFn4PogsyqWmBhFpTO1oEF94zytj34XfR0j6IMK+TG3S7EaiybrdldQvPWmAOz/zRjauXNT2vcwHYRRBkd/WB4AzRGSjiFQJhcAt6ZNEZBOwHLgv4zPy/BKFsW6wn75KEL9Rx5pDQoMI4glvrN6IzUkQ1nD2OT4WpsjoKZdi30OkQSztqyScxum3wHTyvXJJ4joQgbcWYzINIm2qij/Pi2LKy+Y6EUv7K6zwqtP51+dqEAvASR291U8lrchpqwbaikiyKCajSAr7tqpqDbiM0Dz0JHCTqj4uIl8Ukfd4pw4BN6omly6LyAZCDeTuovqYR1DKNzGN1hqhk9q9HY7VGol0GIdTjurhsTp91cAJiORK6nQUUPrtPf3mH5RKKQ0i8kFkK4J+mGve8agv9Xr7AiKLaF3IQvZBxFFMs5BBr2liKvxWxgKk0FQbqnorcGuq7arU/tU51z5DjlO7aIJSyRMQUc3npA+iVJLQhl9vJLSG3QeTwVbHxmosqpap1dUTMuHvnnKQEAItPoi0AzgQ1izrQyS0V0eTg+878VnaV6EalDhpSW/m8UogzWR9U4hiyuKYSw2SdpwvKB9E5KSeBW3JkvUZRbLgczFlEZRao5giH8TxsVr85l5x6xR8reE3+48lPuvgsTEGF1UZrdVbUm1Uy6WE1pAOGW0pAFQSzt04yH1XvpmXLO1lHXDf5y7k5KVJZ3fEsv4q/33FBaxMmYGanxcKQlX1FspNb1I75gRqi4lpCquEu5U41cYsOI7jgkHmpDYKYP5/W6dA4E2csYmp1vRFRLb/ShBmQ/V9ELv2H0181oGjY6wYqNJTDlrCXKtBKbFiOp0or1WDCM/3V0znCYeIk5b05k4ekY9ivK6eD2LCj5uUY6PJmt3xvcoLyEk9xXUQU7pXYE5qozjm/7d1CkRftoY26zkMj9UZrzcYryv9Lry0Wi4xVlcOH28KiJ37khrE/qOjrBzooadSallJHWkIPXFN50nCXGf4LTHKl1RrNLyKcsVoEM3EgfN/IosEcrsrqadCs+Ro4bcyFiAmIDKIvtj1hsb1HEZrDY67t+NIg6h6GkQ0me/yBES4NqLGikVVesql2Ew16mkQ0LRVTxbFNNOx7pHA8TWI6QqhKNQg7YNYSCupVy/pZWlfpSXzahFE8txMTEYRmA8ig8jhV29oYqHbC8fHADwfhMQ+iMH+KkdGxhMlQA8eC89fMdBDTzmIHd7NGtLurTqnDkBLqo0ZfiONJuuwgFHYp5maaFo0iAW0DuK9m9fw9lesnh0Tk+ViMgrEBEQG0Vt0XTWRrvtgLCCSPoixWoOlfRVEiAsOAew9Eib4G1xUpbdS4oXjzTDXyP8ATft82ryTlWpjJokm7VpDiWoYzZQZa3Eq9PZEymnOF4KSxDXLZ+NeYBqEUQzz/9s6BaIvW72e1CAOHg0FRGxiKocC4vDwOEv6yi1mlecPhwJiZeSkjlZSjzcS2kE1xweRFhAz7YOIIovGPQ1ipsxY6UJC1UnSfhhTQ0QQsTBXoxjs25qBr0Ecn0SDGK01ODpaY6CnnFgVDfD8i6GAWDHQw5K+MvuPjtJohOk4fPt0tRxVics2MUWrp2fa6RlrEH4U0zQnmjf9TnZOrKmksjZOjDXL+lizLHuti2FMBzMxZRBpELVGg+Hx5iK4F44lBUTVMzH1lAP+9uJz+O3hUW574nm+9J9P8tzhSEBU+d31g9xw/26e3nuEg8dGEwnZorfrligm97a9pK/CkdFaoVFMcUW5aQqh6z/06sxyrZEjfiGYmGabuz77JvNBGIUAdJ/PAAAPKUlEQVRg39YMoom40UhmY42czlEt6EpZGK9r6FMol+gpB5y6oj92Yj9/eIRqUGJxT5nzTl8BwL07DsRrIyKqeVFMlaaACPs1s4+r4kUxTaWiXBalUmslPP9eZmKaeSpByXwQRiHYtzWD6G2s1mhwfKwex5gfnECD8CfFKNroucPDrBioIhKmyNiwop97dx7gwLGxRJK7SuyDyF4ot9QtOpvu232aWIPwTExF2bIXUqoNw5gv2Lc1g8DTIIbH6nGe/liD8HwQY7UGY/VGYuKLhMXeI6OJokCv3jDI9t0vsP9oysQUaRApAbBx1QDL+ivxaumZNzE1M9I2ZkiDmOxeMx2qaxhGcZiAyCAo+RpEjUE3mUdO6lhAlMNsqKEPwtcgwu0Xh2uxtgGwYeUi9h8d48hIjZW+iSl6u05Nzm88cxXbr3oby/qLMTFVvXUQU0n33Q4LKd23YcwX7NuaQaxBuHUQy52AiJ3ULgKpx5mYxuuNxJtxNBkeGRmn10u4t26wGeWUMDHl+CAiohXVMx7FFAvC0AdRkuLqCkxW3c4wjM7DopgyCLyJ8/h4PTYHHTg2RjUoJRZ9jdXyfRBR7YiIdV6JUd/E1JPjg4jIc2JPl2gc9/x6P3uPjBRav8B8EIbRfZiAyCCaiKNUG5EGcWSklkghUSkLo7XQPJPwQXjbPb6AyNMggmwfRPwZUYjoTEcxuft9/e6dAAlhNtOcuXoxm16y2GonG0YXYQIigyiKKdIO/Ld936dQCUpxAr+EBuFt+5PuikVV+ioBw+P1pA+inL2SOqKnKA0iJXCKnLwv2ryGizbPSf0nwzCmiOn7GUQT5VGvOlo0d/ZV/RXQzZKdvtbgaxO+D0JEWOvMTL4GMZkJqRovMpvpZH3Jz7O3e8MwfExAZBALiJFQQPRVg1gT8DWIakZoKyQn3t5Uyu51g/30lEssSmkikJ/mojANIl2QyASEYRgeZmLKIJqIjzgNor8aMDhQ5djBYforzT9Znt/B3/Y1DoA3nLESSEYLvXLNErauX567GnbTS5Zw2qpFubWlp0paINhqXMMwfExAZBALiJGmgDh3wwp2H9yTqBNdyTErJU1MSQHx4fM38uHzNyba3n3WKbz7rFNy+3P2umXc+Zk3tT+QSUhHFJkGYRiGT6EmJhF5h4g8JSI7ROTKjONfFZHt7udpETnkHTtVRG4TkSdF5AkR2VBkX31aTUxlXudyKe0+eDw+L5GyO8dJPRtVxaZKel2F+SAMw/ApTIMQkQC4FngrsAd4QERuUdUnonNU9dPe+ZcDm72P+DbwZVW9XUQGgEZRfU3T1CDCWtP91SBOtvfMAU9AZCyOC7c9H0Slc9086bBZExCGYfgUaWI6F9ihqrsARORGYBvwRM75FwN/4c59OVBW1dsBVPVogf1sIXIWR1FMfZWAU5b1cfa6ZbxvSzNU0xcKPRkFgKJrOxXTIAzDmIgiBcQaYLe3vwd4TdaJIrIe2Ajc6ZrOBA6JyPdd+8+AK1W1nrruEuASgFNPPXXGOp7lgwD40aXnJ87LNTFN4IPoJMwHYRjGRHSK/WMIuNkTAGXgDcBngVcDpwF/kr5IVa9T1a2qunXVquxKZlOhNYopW46emJO6U/7EraTXQVjZSsMwfIqcvZ4F1nn7a11bFkPADd7+HmC7qu5S1RrwQ2BLIb3MoJzyQaRDVSOyUnyH7b4PonM1iHRivplOBmgYRndTpIB4ADhDRDaKSJVQCNySPklENgHLgftS1y4TkUgtuJB838WMU0pFMfXnCIhq2XdSN7dFJN7vZAEBcO6GQU5ZGq6vsLKVhmH4FCYg3Jv/ZcBPgSeBm1T1cRH5ooi8xzt1CLhRVdW7tk5oXrpDRB4FBPhmUX1NU/Z8EJVAcjOQVoPm5N+TqnMQXdPJTmqAmz56Hn903nqAzFrShmEsXApdKKeqtwK3ptquSu1fnXPt7cBZhXVuAkpeFNNEE7yvNfjCIjxWAuodr0FAMx3IeH3WIokNw+gCOteDOodEtvijo7VcBzUkF8RVyknzTLdoENA0g43XTYMwDKOJCYgMfFt8nv8B8vMvhfuRD6Lz/8R91bCPpkEYhuHT+bPXHOAvGMuLYIL89Br+vpmYDMPoVkxAZOALiIk0iLxsrtExkVbndSfSWzUTk2EYrXT+7DUHJDWICXwQCSd1q4DoKZda1hp0IrEGUTMNwjCMJiYgMkhoEBOYiPxSoelaCtVAusJBDU0z2piZmAzD8DABkUGpTSd1NcOMVAlKXeF/gKYj3dZBGIbhYwIig/IJOqkjH0TWQrpKUOoaDSIyMdVNQBiG4WECIoN2ndSZGkS5RE+XCIiJhKBhGAsXKzmagYjwwdecyo69R7lw0+rc8yInddpBDfC+LWvidOGdTqRBGIZh+JiAyOHL733VpOeICNWglKlBbDtnTcYVnUlv1RRJwzBasZlhmoTJ/Do/lHUisjQgwzAMmxmmSbWcrUF0E92wVsMwjNnHTEzTpBKU5sUb+OfeuYmtG5bPdTcMw+ggTEBMk0pQyq0X0U386RtPn+suGIbRYXT/zDbHzAcTk2EYRhY2s02T6jwxMRmGYaQxE9M0+fgFp7O8vzrX3TAMw5hxTEBMk25a72AYhtEOZhsxDMMwMjEBYRiGYWRiAsIwDMPIpFABISLvEJGnRGSHiFyZcfyrIrLd/TwtIoe8Y3Xv2C1F9tMwDMNopTAntYgEwLXAW4E9wAMicouqPhGdo6qf9s6/HNjsfcSwqp5TVP8MwzCMiSlSgzgX2KGqu1R1DLgR2DbB+RcDNxTYH8MwDKMNihQQa4Dd3v4e19aCiKwHNgJ3es29IvKgiPxCRC7Kue4Sd86D+/btm6l+G4ZhGHSOk3oIuFlV617belXdCvwh8DURaUkWpKrXqepWVd26atWq2eqrYRjGgqDIhXLPAuu8/bWuLYsh4FK/QVWfdb93icjPCf0TO/Nu9tBDD+0Xkf+bRn9XAvuncX0nMV/GMl/GATaWTsXGAuvzDhQpIB4AzhCRjYSCYYhQG0ggIpuA5cB9Xtty4LiqjorISuB84CsT3UxVp6VCiMiDTmPpeubLWObLOMDG0qnYWCamMAGhqjURuQz4KRAA16vq4yLyReBBVY1CV4eAG1VVvctfBnxDRBqEZrC/9KOfDMMwjOIpNBeTqt4K3Jpquyq1f3XGdfcCkxeFNgzDMAqjU5zUncB1c92BGWS+jGW+jANsLJ2KjWUCJGnZMQzDMIwQ0yAMwzCMTExAGIZhGJkseAExWULBTkdEnhGRR11Swwdd26CI3C4iv3a/l891P7MQketFZK+IPOa1ZfZdQq5xz+kREdkydz1vJWcsV4vIs17SyXd5xz7nxvKUiLx9bnqdjYisE5G7ROQJEXlcRD7p2rvq2Uwwjq57LiLSKyL3i8jDbixfcO0bReSXrs/fFZGqa+9x+zvc8Q1TurGqLtgfwvDbncBpQBV4GHj5XPerzTE8A6xMtX0FuNJtXwn81Vz3M6fvvwdsAR6brO/Au4AfAwK8FvjlXPf/BMZyNfDZjHNf7v7XeghTzOwEgrkeg9e/k4Etbnsx8LTrc1c9mwnG0XXPxf1tB9x2Bfil+1vfBAy59q8DH3PbHwe+7raHgO9O5b4LXYNoN6Fgt7AN+Jbb/haQmctqrlHV/wIOpprz+r4N+LaG/AJYJiInz05PJydnLHlsI1z7M6qqvwF2EP4vdgSq+pyq/sptHwGeJMyj1lXPZoJx5NGxz8X9bY+63Yr7UeBC4GbXnn4m0bO6GXiziEi7913oAuKEEwp2MArcJiIPicglrm21qj7ntp8HVs9N16ZEXt+79Vld5swu13umvq4ZizNNbCZ8Y+3aZ5MaB3ThcxGRQES2A3uB2wk1nEOqWnOn+P2Nx+KOHwZWtHvPhS4g5gOvV9UtwDuBS0Xk9/yDGuqYXRnL3M19d/wjcDpwDvAc8Ndz2532EJEB4HvAp1T1Rf9YNz2bjHF05XNR1bqGNXLWEmo2m4q+50IXEO0kFOxItJnUcC/wA8J/nN9GKr77vXfuetg2eX3vumelqr91X+oG8E2a5oqOH4uIVAgn1e+o6vddc9c9m6xxdPNzAVDVQ8BdwHmE5rwoI4bf33gs7vhS4EC791roAiJOKOi8/0NA15Q3FZFFIrI42gbeBjxGOIYPudM+BPxobno4JfL6fgvwxy5i5rXAYc/c0ZGk7PDvJXw2EI5lyEWabATOAO6f7f7l4WzV/wQ8qap/4x3qqmeTN45ufC4iskpElrntPsJKnU8SCor3u9PSzyR6Vu8H7nRaX3vMtXd+rn8IIzCeJrTnfX6u+9Nm308jjLp4GHg86j+hrfEO4NfAz4DBue5rTv9vIFTxxwntpx/J6zthFMe17jk9Cmyd6/6fwFj+xfX1EfeFPdk7//NuLE8B75zr/qfG8npC89EjwHb3865uezYTjKPrngtwFvC/rs+PAVe59tMIhdgO4N+BHtfe6/Z3uOOnTeW+lmrDMAzDyGShm5gMwzCMHExAGIZhGJmYgDAMwzAyMQFhGIZhZGICwjAMw8jEBIRhTIKI1L3Mn9tlBrP+isgGPwOsYXQShdakNox5wrCGKQ4MY0FhGoRhTBEJa3F8RcJ6HPeLyEtd+wYRudMlg7tDRE517atF5Acup//DIvI691GBiHzT5fm/za2URUQ+4WoZPCIiN87RMI0FjAkIw5icvpSJ6QPescOq+irg74Gvuba/A76lqmcB3wGuce3XAHer6tmEtSMed+1nANeq6iuAQ8D7XPuVwGb3OR8tanCGkYetpDaMSRCRo6o6kNH+DHChqu5ySeGeV9UVIrKfMH3DuGt/TlVXisg+YK2qjnqfsQG4XVXPcPtXABVV/ZKI/AQ4CvwQ+KE26wEYxqxgGoRhTA/N2W6HUW+7TtM3+PuEOY62AA94WTsNY1YwAWEY0+MD3u/73Pa9hJmBAT4I3OO27wA+BnHxl6V5HyoiJWCdqt4FXEGYrrlFizGMIrE3EsOYnD5XySviJ6oahbouF5FHCLWAi13b5cA/i8ifAfuAD7v2TwLXichHCDWFjxFmgM0iAP7VCREBrtGwDoBhzBrmgzCMKeJ8EFtVdf9c98UwisBMTIZhGEYmpkEYhmEYmZgGYRiGYWRiAsIwDMPIxASEYRiGkYkJCMMwDCMTExCGYRhGJv8PglcFF7K1vz4AAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yM7BJOYtuC1l",
        "colab_type": "code",
        "colab": {},
        "outputId": "000665b8-29ba-4e8e-c4b6-f9e9a15dfd09"
      },
      "source": [
        "# now lets see how the loss of our model progressed over training.\n",
        "plt.plot(training.history['loss'])\n",
        "plt.title('Model Loss')\n",
        "plt.ylabel('Loss')\n",
        "plt.xlabel('Epochs')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ysAonDBZuC1n",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 101
        },
        "outputId": "581364d0-d03c-46b5-9c69-b509afc8010f"
      },
      "source": [
        "# evaluate the keras model\n",
        "_, accuracy = model.evaluate(X, y)\n",
        "print(model.evaluate(X,y))\n",
        "print(accuracy * 100)\n",
        "print('Accuracy: %.2f' % (accuracy*100))"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "768/768 [==============================] - 0s 35us/step\n",
            "768/768 [==============================] - 0s 26us/step\n",
            "[0.41956742852926254, 0.8033854365348816]\n",
            "80.33854365348816\n",
            "Accuracy: 80.34\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "k1XaMjn7uC1q",
        "colab_type": "text"
      },
      "source": [
        "What score did you get?\n",
        "Post your results in the zoom chat.\n",
        "\n",
        "Neural networks are a stochastic algorithm, meaning that the same algorithm on the same data can train a different model with different skill each time the code is run. This is a feature, not a bug.\n",
        "\n",
        "This model won't perform very well since the architecture is not very good. However, it is a very simple architecture, meaning it is a good place to start and learn from. "
      ]
    }
  ]
}