dataset_creation_and_model_development.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Programming Midterm.ipynb",
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "68uZpmgoLz_7",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sn\n",
        "import requests\n",
        "import os\n",
        "import re\n",
        "from string import digits\n",
        "from google.colab import drive\n",
        "from tensorflow.keras.models import Sequential, Model, load_model\n",
        "from tensorflow.keras.layers import ZeroPadding2D, Convolution2D, MaxPooling2D, Dense, Dropout, Softmax, Flatten, Activation\n",
        "from tensorflow.keras.callbacks import ModelCheckpoint\n",
        "from tensorflow.keras.preprocessing.image import load_img, img_to_array, ImageDataGenerator\n",
        "from tensorflow.keras.backend import clear_session\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.metrics import confusion_matrix"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bW1e-JhcGlY9",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "drive.mount('/content/drive')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "K4CaIDYaOOAh",
        "colab_type": "text"
      },
      "source": [
        "# Dataset Creation\n",
        "\n",
        "The images for this dataset are scraped from Google Image Search. This scraping method is adapted from here: https://tinyurl.com/vn83kgp.\n",
        "\n",
        "For each person included in the dataset, a Google Image Search query was carried and the following JavaScript commands used to save the urls of all the images appearing in the search results into a file:\n",
        "\n",
        "```javascript\n",
        "urls=Array.from(document.querySelectorAll('.rg_i')).map(el=> el.hasAttribute('data-src')?el.getAttribute('data-src'):el.getAttribute('data-iurl'));\n",
        "\n",
        "window.open('data:text/csv;charset=utf-8,' + escape(urls.join('\\n')));\n",
        "```\n",
        "\n",
        "A file containing image urls was created for each of the considered individuals. The following block of codes reads, stores and displays the name of each of these files."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Nb7glnVjMKBj",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 204
        },
        "outputId": "13fa2160-842a-428a-bdb7-c3890ab5513c"
      },
      "source": [
        "url_files = []\n",
        "link_folder = '/content/drive/My Drive/EECE 693 Programming Midterm/Dataset/Link Files/'\n",
        "# r=root, d=directories\n",
        "for r, d, files in os.walk(link_folder):\n",
        "    for file in files:\n",
        "        url_files.append(os.path.join(file))\n",
        "\n",
        "for url_file in url_files:\n",
        "    print(url_file)"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Tedros Adhanom\n",
            "Fadlo Khuri\n",
            "Hamad Hassan\n",
            "Anthony Fauci\n",
            "Marcel Ghanem\n",
            "Bill Gates\n",
            "Cate Blanchett\n",
            "Donald Trump\n",
            "Samuel L. Jackson\n",
            "Keanu Reeves\n",
            "Andrew M. Cuomo\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Obn5P1zjRFQx",
        "colab_type": "text"
      },
      "source": [
        "Next, the following block of code will open each of these files one by one and extract the urls as a list. It will then iterate over these urls to download the corresponding images. Downloaded images are named after the person followed by an ordinality number.\n",
        "\n",
        "At first, 80 images per person are downloaded. From these, 10 images per person will be manually selected for training and 2 for testing.\n",
        "In case we were building a large dataset, manual selection would not be feasible. However, in this case the dataset is very small and therefore we cannot tolerate any outliers. For example, some of the images scraped had the subject's face partially blocked by other people or had a young image of the subject."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7376tFg-9G81",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "for url_file in url_files:\n",
        "    image_path = '/content/drive/My Drive/EECE 693 Programming Midterm/Dataset/image_dump/'\n",
        "    print(f'Downloading {url_file} images...\\n')\n",
        "    url_path = open(link_folder + url_file).read().strip().split('\\n')\n",
        "    total = 0\n",
        "\n",
        "    if not os.path.exists(image_path):\n",
        "        os.mkdir(image_path)\n",
        "    image_path = image_path + url_file\n",
        "    if not os.path.exists(image_path):\n",
        "        os.mkdir(image_path)\n",
        "\n",
        "    for url in url_path[:80]:\n",
        "        try:\n",
        "            req = requests.get(url, timeout=60)\n",
        "\n",
        "            file_path = os.path.sep.join([image_path, '{}.jpg'.format(url_file + str(total+1))])\n",
        "            file = open(file_path, 'wb')\n",
        "            file.write(req.content)\n",
        "            file.close()\n",
        "\n",
        "            print('Downloaded {}'.format(file_path))\n",
        "            total += 1\n",
        "\n",
        "        except:\n",
        "            print('Could not download {}. Downloading next file'.format(file_path))\n",
        "    print('\\n\\n')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "RxE6KT8DW97u",
        "colab_type": "text"
      },
      "source": [
        "---\n",
        "\n",
        "# Loading the Data\n",
        "\n",
        "After having selected and partitioned the training and testing images into separate folders, the data is imported and processed.\n",
        "\n",
        "Images are loaded, resized and normalized. They are stored in numpy arrays.\n",
        "\n",
        "The label for each image is extracted from its name, the trailing digits and \".jpg\" are removed. Labels are stored in a one-hot encoded format.\n",
        "\n",
        "**Training and Validation Data**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yj2gr1G9XYWE",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "image_list = []\n",
        "label_list = []\n",
        "train_folder = '/content/drive/My Drive/EECE 693 Programming Midterm/Dataset/Training/'\n",
        "remove_digits = str.maketrans('', '', digits) \n",
        "for r, d, image_names in os.walk(train_folder):\n",
        "    for image_name in image_names:\n",
        "        img = load_img(train_folder + image_name, color_mode='rgb', target_size=(224,224))\n",
        "        img = img_to_array(img)\n",
        "        img = img / 255\n",
        "        image_list.append(img)\n",
        "        label_list.append(re.sub(r'.jpg', '', image_name).translate(remove_digits))\n",
        "X_train = np.array(image_list)\n",
        "y_train = pd.get_dummies(pd.DataFrame(label_list))\n",
        "y_train.columns = [x[2:] for x in y_train.columns]\n",
        "\n",
        "X_train, X_val, y_train, y_val = train_test_split(X_train, y_train, test_size=11, random_state=42, shuffle=True, stratify=y_train)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AtDieRevo34b",
        "colab_type": "text"
      },
      "source": [
        "**Testing Data**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7rE4mzgNo6LX",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "image_list = []\n",
        "label_list = []\n",
        "test_folder = '/content/drive/My Drive/EECE 693 Programming Midterm/Dataset/Testing/'\n",
        "remove_digits = str.maketrans('', '', digits) \n",
        "for r, d, image_names in os.walk(test_folder):\n",
        "    for image_name in image_names:\n",
        "        img = load_img(test_folder + image_name, color_mode='rgb', target_size=(224,224))\n",
        "        img = img_to_array(img)\n",
        "        img = img / 255\n",
        "        image_list.append(img)\n",
        "        label_list.append(re.sub(r'.jpg', '', image_name).translate(remove_digits))\n",
        "X_test = np.array(image_list)\n",
        "y_test = pd.get_dummies(pd.DataFrame(label_list))\n",
        "y_test.columns = [x[2:] for x in y_test.columns]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OyFrNsXkbnvU",
        "colab_type": "text"
      },
      "source": [
        "**Prepare Image Augmentation**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5pMp3n67ZIbS",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "datagen = ImageDataGenerator(rotation_range=10, width_shift_range=0.1, height_shift_range=0.1, zoom_range=0.1, horizontal_flip=True)\n",
        "datagen.fit(X_train)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DdqpMdv7pduf",
        "colab_type": "text"
      },
      "source": [
        "---\n",
        "\n",
        "# VGG-Face Model Transfer Learning\n",
        "\n",
        "For accomplishing this task, the VGG-Face model will be fine-tuned on the presented dataset. VGG-Face is a VGG16-based network trained by Parkhi et al. (https://tinyurl.com/q32eqa8) on a large facial recognition dataset.\n",
        "\n",
        "First, the VGG-Face model is created according to the architecture used by Parkhi et al. (https://tinyurl.com/s3un98p). Then, the pre-trained weights are loaded into it (the weights shared by Parkhi et al. are made compatible with Keras by Sefik Serengil https://tinyurl.com/trjp3rs).\n",
        "\n",
        "The model for this task will incorporate the VGG-Face but with the last softmax layer removed. Two fully connected layers and a softmax layer with 11 outputs will be added. Only these new layers will be trained, all previous layers are frozen."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yCrrocI9puWm",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "clear_session()\n",
        "#Build the original VGG-Face\n",
        "vgg_face = Sequential(name='vgg_face_original')\n",
        "vgg_face.add(ZeroPadding2D((1,1),input_shape=(224,224, 3)))\n",
        "vgg_face.add(Convolution2D(64, (3, 3), activation='relu'))\n",
        "vgg_face.add(ZeroPadding2D((1,1)))\n",
        "vgg_face.add(Convolution2D(64, (3, 3), activation='relu'))\n",
        "vgg_face.add(MaxPooling2D((2,2), strides=(2,2)))\n",
        " \n",
        "vgg_face.add(ZeroPadding2D((1,1)))\n",
        "vgg_face.add(Convolution2D(128, (3, 3), activation='relu'))\n",
        "vgg_face.add(ZeroPadding2D((1,1)))\n",
        "vgg_face.add(Convolution2D(128, (3, 3), activation='relu'))\n",
        "vgg_face.add(MaxPooling2D((2,2), strides=(2,2)))\n",
        " \n",
        "vgg_face.add(ZeroPadding2D((1,1)))\n",
        "vgg_face.add(Convolution2D(256, (3, 3), activation='relu'))\n",
        "vgg_face.add(ZeroPadding2D((1,1)))\n",
        "vgg_face.add(Convolution2D(256, (3, 3), activation='relu'))\n",
        "vgg_face.add(ZeroPadding2D((1,1)))\n",
        "vgg_face.add(Convolution2D(256, (3, 3), activation='relu'))\n",
        "vgg_face.add(MaxPooling2D((2,2), strides=(2,2)))\n",
        " \n",
        "vgg_face.add(ZeroPadding2D((1,1)))\n",
        "vgg_face.add(Convolution2D(512, (3, 3), activation='relu'))\n",
        "vgg_face.add(ZeroPadding2D((1,1)))\n",
        "vgg_face.add(Convolution2D(512, (3, 3), activation='relu'))\n",
        "vgg_face.add(ZeroPadding2D((1,1)))\n",
        "vgg_face.add(Convolution2D(512, (3, 3), activation='relu'))\n",
        "vgg_face.add(MaxPooling2D((2,2), strides=(2,2)))\n",
        " \n",
        "vgg_face.add(ZeroPadding2D((1,1)))\n",
        "vgg_face.add(Convolution2D(512, (3, 3), activation='relu'))\n",
        "vgg_face.add(ZeroPadding2D((1,1)))\n",
        "vgg_face.add(Convolution2D(512, (3, 3), activation='relu'))\n",
        "vgg_face.add(ZeroPadding2D((1,1)))\n",
        "vgg_face.add(Convolution2D(512, (3, 3), activation='relu'))\n",
        "vgg_face.add(MaxPooling2D((2,2), strides=(2,2)))\n",
        " \n",
        "vgg_face.add(Convolution2D(4096, (7, 7), activation='relu'))\n",
        "vgg_face.add(Dropout(0.5))\n",
        "vgg_face.add(Convolution2D(4096, (1, 1), activation='relu'))\n",
        "vgg_face.add(Dropout(0.5))\n",
        "vgg_face.add(Convolution2D(2622, (1, 1)))\n",
        "vgg_face.add(Flatten())\n",
        "vgg_face.add(Activation('softmax'))\n",
        "\n",
        "#Load Pre-trained Weights\n",
        "vgg_face.load_weights('/content/drive/My Drive/EECE 693 Programming Midterm/vgg_face_weights.h5')\n",
        "\n",
        "#Remove the Last Softmax Layer\n",
        "vgg_base = Model(inputs=vgg_face.layers[0].input, outputs=vgg_face.layers[-2].output, name='vgg_base')\n",
        "\n",
        "#Freeze Base Layers\n",
        "for layer in vgg_base.layers:\n",
        "    layer.trainable = False\n",
        "\n",
        "#Add Fully Connected Layers\n",
        "fc = Dense(1024, activation='relu')(vgg_base.layers[-1].output)\n",
        "drp = Dropout(0.5)(fc)\n",
        "fc_out = Dense(y_train.shape[1], activation='softmax')(drp)\n",
        "\n",
        "model = Model(inputs=vgg_base.input, outputs=fc_out, name='final_model')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6QY23CVl7bdL",
        "colab_type": "code",
        "outputId": "9aabb2f9-f240-4be2-b11a-12f7514491d7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "model.summary()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"final_model\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "zero_padding2d_input (InputL [(None, 224, 224, 3)]     0         \n",
            "_________________________________________________________________\n",
            "zero_padding2d (ZeroPadding2 (None, 226, 226, 3)       0         \n",
            "_________________________________________________________________\n",
            "conv2d (Conv2D)              (None, 224, 224, 64)      1792      \n",
            "_________________________________________________________________\n",
            "zero_padding2d_1 (ZeroPaddin (None, 226, 226, 64)      0         \n",
            "_________________________________________________________________\n",
            "conv2d_1 (Conv2D)            (None, 224, 224, 64)      36928     \n",
            "_________________________________________________________________\n",
            "max_pooling2d (MaxPooling2D) (None, 112, 112, 64)      0         \n",
            "_________________________________________________________________\n",
            "zero_padding2d_2 (ZeroPaddin (None, 114, 114, 64)      0         \n",
            "_________________________________________________________________\n",
            "conv2d_2 (Conv2D)            (None, 112, 112, 128)     73856     \n",
            "_________________________________________________________________\n",
            "zero_padding2d_3 (ZeroPaddin (None, 114, 114, 128)     0         \n",
            "_________________________________________________________________\n",
            "conv2d_3 (Conv2D)            (None, 112, 112, 128)     147584    \n",
            "_________________________________________________________________\n",
            "max_pooling2d_1 (MaxPooling2 (None, 56, 56, 128)       0         \n",
            "_________________________________________________________________\n",
            "zero_padding2d_4 (ZeroPaddin (None, 58, 58, 128)       0         \n",
            "_________________________________________________________________\n",
            "conv2d_4 (Conv2D)            (None, 56, 56, 256)       295168    \n",
            "_________________________________________________________________\n",
            "zero_padding2d_5 (ZeroPaddin (None, 58, 58, 256)       0         \n",
            "_________________________________________________________________\n",
            "conv2d_5 (Conv2D)            (None, 56, 56, 256)       590080    \n",
            "_________________________________________________________________\n",
            "zero_padding2d_6 (ZeroPaddin (None, 58, 58, 256)       0         \n",
            "_________________________________________________________________\n",
            "conv2d_6 (Conv2D)            (None, 56, 56, 256)       590080    \n",
            "_________________________________________________________________\n",
            "max_pooling2d_2 (MaxPooling2 (None, 28, 28, 256)       0         \n",
            "_________________________________________________________________\n",
            "zero_padding2d_7 (ZeroPaddin (None, 30, 30, 256)       0         \n",
            "_________________________________________________________________\n",
            "conv2d_7 (Conv2D)            (None, 28, 28, 512)       1180160   \n",
            "_________________________________________________________________\n",
            "zero_padding2d_8 (ZeroPaddin (None, 30, 30, 512)       0         \n",
            "_________________________________________________________________\n",
            "conv2d_8 (Conv2D)            (None, 28, 28, 512)       2359808   \n",
            "_________________________________________________________________\n",
            "zero_padding2d_9 (ZeroPaddin (None, 30, 30, 512)       0         \n",
            "_________________________________________________________________\n",
            "conv2d_9 (Conv2D)            (None, 28, 28, 512)       2359808   \n",
            "_________________________________________________________________\n",
            "max_pooling2d_3 (MaxPooling2 (None, 14, 14, 512)       0         \n",
            "_________________________________________________________________\n",
            "zero_padding2d_10 (ZeroPaddi (None, 16, 16, 512)       0         \n",
            "_________________________________________________________________\n",
            "conv2d_10 (Conv2D)           (None, 14, 14, 512)       2359808   \n",
            "_________________________________________________________________\n",
            "zero_padding2d_11 (ZeroPaddi (None, 16, 16, 512)       0         \n",
            "_________________________________________________________________\n",
            "conv2d_11 (Conv2D)           (None, 14, 14, 512)       2359808   \n",
            "_________________________________________________________________\n",
            "zero_padding2d_12 (ZeroPaddi (None, 16, 16, 512)       0         \n",
            "_________________________________________________________________\n",
            "conv2d_12 (Conv2D)           (None, 14, 14, 512)       2359808   \n",
            "_________________________________________________________________\n",
            "max_pooling2d_4 (MaxPooling2 (None, 7, 7, 512)         0         \n",
            "_________________________________________________________________\n",
            "conv2d_13 (Conv2D)           (None, 1, 1, 4096)        102764544 \n",
            "_________________________________________________________________\n",
            "dropout (Dropout)            (None, 1, 1, 4096)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_14 (Conv2D)           (None, 1, 1, 4096)        16781312  \n",
            "_________________________________________________________________\n",
            "dropout_1 (Dropout)          (None, 1, 1, 4096)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_15 (Conv2D)           (None, 1, 1, 2622)        10742334  \n",
            "_________________________________________________________________\n",
            "flatten (Flatten)            (None, 2622)              0         \n",
            "_________________________________________________________________\n",
            "dense (Dense)                (None, 1024)              2685952   \n",
            "_________________________________________________________________\n",
            "dropout_2 (Dropout)          (None, 1024)              0         \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 11)                11275     \n",
            "=================================================================\n",
            "Total params: 147,700,105\n",
            "Trainable params: 2,697,227\n",
            "Non-trainable params: 145,002,878\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mh7jaAac5x4t",
        "colab_type": "code",
        "outputId": "a5d4e65f-17d2-48fc-a32e-fff5415398ce",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        }
      },
      "source": [
        "#Check Layers Freezing Status\n",
        "for layer in model.layers:\n",
        "    if layer.trainable:\n",
        "        print(layer.name + ' is trainable')"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "dense is trainable\n",
            "dropout_2 is trainable\n",
            "dense_1 is trainable\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "diVbR4SB9Djr",
        "colab_type": "code",
        "outputId": "fe37bf3d-092c-408f-8571-c55ca6e68427",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 867
        }
      },
      "source": [
        "model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])\n",
        "\n",
        "checkpoint = ModelCheckpoint('/content/drive/My Drive/EECE 693 Programming Midterm/final_model.hdf5', monitor='loss', save_best_only=True,)\n",
        "\n",
        "bt = 10 #batch size\n",
        "ep = 25 #number of epochs\n",
        "\n",
        "history = model.fit(datagen.flow(X_train, y_train.values, batch_size=bt), steps_per_epoch=len(X_train)//bt,\n",
        "                    validation_data=(X_val, y_val.values),\n",
        "                    epochs=ep, callbacks=[checkpoint])"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/25\n",
            "9/9 [==============================] - 5s 560ms/step - loss: 2.3966 - accuracy: 0.1556 - val_loss: 2.2923 - val_accuracy: 0.4545\n",
            "Epoch 2/25\n",
            "9/9 [==============================] - 13s 1s/step - loss: 2.2727 - accuracy: 0.3034 - val_loss: 2.1675 - val_accuracy: 0.3636\n",
            "Epoch 3/25\n",
            "9/9 [==============================] - 16s 2s/step - loss: 2.0824 - accuracy: 0.4719 - val_loss: 2.0431 - val_accuracy: 0.4545\n",
            "Epoch 4/25\n",
            "9/9 [==============================] - 19s 2s/step - loss: 1.9679 - accuracy: 0.4607 - val_loss: 1.9293 - val_accuracy: 0.4545\n",
            "Epoch 5/25\n",
            "9/9 [==============================] - 10s 1s/step - loss: 1.8335 - accuracy: 0.5056 - val_loss: 1.8352 - val_accuracy: 0.6364\n",
            "Epoch 6/25\n",
            "9/9 [==============================] - 16s 2s/step - loss: 1.6652 - accuracy: 0.6404 - val_loss: 1.7428 - val_accuracy: 0.5455\n",
            "Epoch 7/25\n",
            "9/9 [==============================] - 16s 2s/step - loss: 1.6186 - accuracy: 0.6067 - val_loss: 1.6683 - val_accuracy: 0.5455\n",
            "Epoch 8/25\n",
            "9/9 [==============================] - 17s 2s/step - loss: 1.5811 - accuracy: 0.5955 - val_loss: 1.6058 - val_accuracy: 0.5455\n",
            "Epoch 9/25\n",
            "9/9 [==============================] - 15s 2s/step - loss: 1.3099 - accuracy: 0.6629 - val_loss: 1.5452 - val_accuracy: 0.6364\n",
            "Epoch 10/25\n",
            "9/9 [==============================] - 2s 190ms/step - loss: 1.4083 - accuracy: 0.6854 - val_loss: 1.4883 - val_accuracy: 0.6364\n",
            "Epoch 11/25\n",
            "9/9 [==============================] - 15s 2s/step - loss: 1.2942 - accuracy: 0.7000 - val_loss: 1.4383 - val_accuracy: 0.7273\n",
            "Epoch 12/25\n",
            "9/9 [==============================] - 15s 2s/step - loss: 1.2726 - accuracy: 0.6292 - val_loss: 1.4064 - val_accuracy: 0.6364\n",
            "Epoch 13/25\n",
            "9/9 [==============================] - 11s 1s/step - loss: 1.1874 - accuracy: 0.6854 - val_loss: 1.3572 - val_accuracy: 0.7273\n",
            "Epoch 14/25\n",
            "9/9 [==============================] - 16s 2s/step - loss: 1.1807 - accuracy: 0.6517 - val_loss: 1.3032 - val_accuracy: 0.8182\n",
            "Epoch 15/25\n",
            "9/9 [==============================] - 16s 2s/step - loss: 1.1555 - accuracy: 0.6629 - val_loss: 1.2659 - val_accuracy: 0.9091\n",
            "Epoch 16/25\n",
            "9/9 [==============================] - 15s 2s/step - loss: 1.1092 - accuracy: 0.6966 - val_loss: 1.2381 - val_accuracy: 0.7273\n",
            "Epoch 17/25\n",
            "9/9 [==============================] - 13s 1s/step - loss: 1.0320 - accuracy: 0.7416 - val_loss: 1.2028 - val_accuracy: 0.7273\n",
            "Epoch 18/25\n",
            "9/9 [==============================] - 2s 190ms/step - loss: 1.1098 - accuracy: 0.6854 - val_loss: 1.1960 - val_accuracy: 0.6364\n",
            "Epoch 19/25\n",
            "9/9 [==============================] - 13s 1s/step - loss: 1.0075 - accuracy: 0.6966 - val_loss: 1.1635 - val_accuracy: 0.6364\n",
            "Epoch 20/25\n",
            "9/9 [==============================] - 15s 2s/step - loss: 0.9697 - accuracy: 0.6966 - val_loss: 1.1391 - val_accuracy: 0.9091\n",
            "Epoch 21/25\n",
            "9/9 [==============================] - 14s 2s/step - loss: 0.9382 - accuracy: 0.7667 - val_loss: 1.1213 - val_accuracy: 0.9091\n",
            "Epoch 22/25\n",
            "9/9 [==============================] - 2s 191ms/step - loss: 0.9491 - accuracy: 0.7416 - val_loss: 1.1130 - val_accuracy: 0.9091\n",
            "Epoch 23/25\n",
            "9/9 [==============================] - 17s 2s/step - loss: 0.9132 - accuracy: 0.7191 - val_loss: 1.1236 - val_accuracy: 0.6364\n",
            "Epoch 24/25\n",
            "9/9 [==============================] - 12s 1s/step - loss: 0.8489 - accuracy: 0.7416 - val_loss: 1.0981 - val_accuracy: 0.8182\n",
            "Epoch 25/25\n",
            "9/9 [==============================] - 2s 190ms/step - loss: 0.9939 - accuracy: 0.7079 - val_loss: 1.0779 - val_accuracy: 0.8182\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dBkTAZJ3_Hvc",
        "colab_type": "text"
      },
      "source": [
        "**Learning Curves**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OLlBdTL6_TaS",
        "colab_type": "code",
        "outputId": "01386719-40d4-4df0-8602-6a6a8e9563d7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        }
      },
      "source": [
        "plt.figure()\n",
        "plt.plot(history.history['loss'], label='training')\n",
        "plt.plot(history.history['val_loss'], label='validation')\n",
        "plt.xlabel('Epochs')\n",
        "plt.ylabel('Loss')\n",
        "plt.legend()\n",
        "plt.grid()\n",
        "plt.show()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYYAAAEGCAYAAABhMDI9AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nIYQQRUhiEEIIUcT/A+s75175KPYVAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6Ti28Pvfun8L",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model.save_weights('/content/drive/My Drive/EECE 693 Programming Midterm/model_weights.h5', save_format='h5')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3IH146NC6_xk",
        "colab_type": "text"
      },
      "source": [
        "---\n",
        "\n",
        "# Model Evaluation"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-1wR35IyxP5d",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model = load_model('/content/drive/My Drive/EECE 693 Programming Midterm/final_model.hdf5')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "m3-69BGYnM7z",
        "colab_type": "text"
      },
      "source": [
        "**On Training Data**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0rk6VcxPooEC",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "column_names = y_train.columns\n",
        "X_train = np.vstack((X_train, X_val))\n",
        "y_train = np.vstack((y_train, y_val))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "outputId": "c1b1437f-63ff-4931-8da4-22297518d8e8",
        "id": "zNHXozE2nP76",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "model.evaluate(X_train, y_train, batch_size=1)"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "110/110 [==============================] - 7s 65ms/step - loss: 0.8040 - accuracy: 0.8364\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.8040437698364258, 0.8363636136054993]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "id": "DOJNZ2dznP8G",
        "colab": {}
      },
      "source": [
        "y_pred = model.predict(X_train, batch_size=1)\n",
        "y_pred = list(np.argmax(y_pred, axis=1))\n",
        "y_true = list(np.argmax(y_train, axis=1))\n",
        "\n",
        "for i in range(len(y_pred)):\n",
        "    y_pred[i] = column_names[y_pred[i]]\n",
        "    y_true[i] = column_names[y_true[i]]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "outputId": "7f9314c4-39af-4a48-8695-1fba977e9dcb",
        "id": "52uatTbznP8V",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 255
        }
      },
      "source": [
        "print(pd.DataFrame({'True Class': y_true, 'Predicted Class': y_pred}))"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "            True Class    Predicted Class\n",
            "0           Bill Gates         Bill Gates\n",
            "1         Donald Trump       Donald Trump\n",
            "2        Marcel Ghanem      Marcel Ghanem\n",
            "3    Samuel L. Jackson  Samuel L. Jackson\n",
            "4       Cate Blanchett     Cate Blanchett\n",
            "..                 ...                ...\n",
            "105        Fadlo Khuri        Fadlo Khuri\n",
            "106         Bill Gates        Fadlo Khuri\n",
            "107       Keanu Reeves       Keanu Reeves\n",
            "108      Anthony Fauci      Anthony Fauci\n",
            "109     Tedros Adhanom     Tedros Adhanom\n",
            "\n",
            "[110 rows x 2 columns]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab_type": "code",
        "outputId": "c962677a-fc90-4dfa-a4d5-9901af8600c4",
        "id": "iuVh18TynP8a",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 539
        }
      },
      "source": [
        "df_cm = pd.DataFrame(confusion_matrix(y_true, y_pred, labels=column_names), column_names, column_names)\n",
        "plt.figure(figsize=(7,7))\n",
        "sn.heatmap(df_cm, annot=True, cmap='PuBu', cbar=False, square=True)\n",
        "plt.title('Train Confusion Matrix')\n",
        "plt.xlabel('Predicted Class')\n",
        "plt.ylabel('True Class')\n",
        "plt.show()"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfoAAAIKCAYAAADLb43NAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0\ndHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAAgAElEQVR4nOzdfbzX8/3H8cdTpxRdSZyMQ6aIYlEY\n5iIXkwqZa2azofGbq202S0aY7IKRNSyZ612YYa2SbakV21BEoREqUadEEV2e8/r98f2cfB3nqtP5\nnu/nfM/zfrudW9/P+/P5vD/P7/t7eH2uzueriMDMzMwK02b5DmBmZma540JvZmZWwFzozczMCpgL\nvZmZWQFzoTczMytgLvRmZmYFzIXezD5D0uOSvpnvHJVJaiPpb5JWSPrzJvRzpqS/N2S2fEjr52Tp\n40JvVgAkrcz6KZe0Kmv6zI3pKyKOiYh7NyHLGZKmJ9telBSkr9S3vywnAcXA1hFxcn07iYgHI+Kr\nDZDnMyQdJikkPVqp/UtJ+5Q69jNc0gO1Lbepn5M1Hy70ZgUgItpW/AALgGOz2h6sWE5SUS5zSPo+\ncAswgkxR3hG4DTi+AbrfCXgtItY3QF+5shQ4QNLWWW3fBF5rqA0ow//vtjrzL4tZAUuOMhdKulzS\nYuBuSVtJGidpqaQPktc7ZK0zRdK5yeuzJT0l6cZk2bckHVPNtjoA1wLfjYhHIuLjiFgXEX+LiB8m\ny2wu6RZJ7yY/t0javFLWH0hakpwN+FYy7xrgKuDU5EzBOZWPfCV1TY6ci7KyvynpoyT3mdnvKWu9\nAyU9l1wSeE7SgZXG4jpJTyf9/F1S5xqGfC3wGHBasn4L4FTgweyFJI2U9LakDyXNkHRw0t4fuCLr\nfb6YleN6SU8DnwBfrPQ53S7pL1n9/1zSJEmqIas1Ey70ZoWvC9CJzBHxEDL/3d+dTO8IrAJG1bD+\n/sD/gM7AL4C7qikgBwCtgUermFdhGPBloDfwJWA/4MpKWTsA2wPnAL+RtFVEXE3mLMGfkrMUd9X0\nhiVtCdwKHBMR7YADgZlVLNcJGJ8suzXwK2B8pSPyM4BvAdsCrYDLato2cB/wjeT10cBs4N1KyzxH\nZgw6Ab8H/iypdURMrPQ+v5S1zllkPr92wPxK/f0A2DPZiTmYzNh9M/yMc8OF3qw5KAeujog1EbEq\nIpZFxF8i4pOI+Ai4Hji0hvXnR8SdEVEG3AtsR+a0fGVbA+/Vcmr9TODaiFgSEUuBa8gUsArrkvnr\nImICsBLYrc7v9LPKgV6S2kTEooh4uYplBgKvR8T9EbE+Iv4AzAGOzVrm7oh4LSJWAQ+RKdDVioh/\nA50k7Uam4N9XxTIPJJ/D+oi4Cdic2t/nPRHxcrLOukr9fUJmHH8FPABcFBELa+nPmgkXerPCtzQi\nVldMSNpC0m8lzZf0ITAV6JicZq7K4ooXSUEBaFvFcsuAzrXcB/AFPns0Oj9p29BHpR2FT6rZVo0i\n4mMyp8zPBxZJGi+pRx3yVGTaPmt6cdbruua5H7gQ6EcVZzgkXSbp1eRywXIyZzFquiQA8HZNMyPi\nGeBNQGR2SMwAF3qz5qDy6dsfkDl63D8i2gOHJO2bej33P8AaYHANy7xL5pJBhR35/GntuvoY2CJr\nukv2zIh4IiKOInMGYg5wZx3yVGR6p56ZKtwP/B8wIWvnCIDk1PqPgFOArSKiI7CCT8e/utPtNZ6G\nl/RdMmcG3k36NwNc6M2ao3ZkrssvT65RX90QnUbECjI3zP1G0uDkzEFLScdI+kWy2B+AKyVtk9zU\ndhWZU831MRM4RNKOyY2AQytmSCqWdHxyrX4NmUsA5VX0MQHYVZk/CSySdCqwBzCunpkAiIi3yFwO\nGVbF7HbAejJ36BdJugponzW/FOi6MXfWS9oV+CnwdTKn8H8kqcZLDNZ8uNCbNT+3AG2A94D/AhMb\nquPkevP3ydxgt5TM6eYLydyJDpliNB14CZgFPJ+01Wdb/wD+lPQ1g88W582SHO8C75MpuhdU0ccy\nYBCZsxzLyBwJD4qI9+qTqVLfT0VEVWcrniAz5q+RuUywms+elq94GNAySc/Xtp3kUskDwM8j4sWI\neJ3Mnfv3V/xFgzVv8k2ZZmZmhctH9GZmZgXMhd7MzKyAudCbmZkVMBd6MzOzAuZCb2ZmVsBy+k1W\nll8rln7sP6lIqQ7t0vNXTys+WpPvCABcN/n1fEfY4Cf9uuc7wgZp+l2xFGtdVO0Dr3xEb2ZmVsBc\n6M3MzAqYC72ZmVkBc6E3MzMrYC70ZmZmBcyF3szMrIC50JuZmRUwF3ozM7MC5kJvZmZWwFzozczM\nCpgLvZmZWQFzobcqXTdiOEcPOoLTzjo531FSkyUtOSpMfXoaRx83kKMG9Wf0XXfmLUeaxuWKgXvw\ng6/uxveO2o1Ljtw1bznSNCZp+T1JS47mmCWnhV7SYEkhqUc91j1M0rhc5Er675pk+2lWW2dJ6ySN\nqsP6u0qaIOl1Sc9LekhSca7yNraBA45l5E21DkOjSEuWtOQAKCsr49oR1zPmtjsY/+hYxk2cwNw3\n5uYlS5rGBeD2KXO5+R//Y+Q/X8tbhrSMSVp+T9KSo7lmyfUR/enAU8m/DUJSQ37j3lvAwKzpk4GX\n65ChNTAeuD0iukfEPsBtwDYNmC2v9undh/btO+Q7BpCeLGnJAfDS7FnsVFJCyQ4ltGrZioH9BzBp\nyuS8ZEnTuKRFWsYkLb8nacnRXLPkrNBLagt8BTgHOC2r/TBJUyQ9LGmOpAclKZnXP2l7Hvha1jrD\nJd0v6WngfknbSPqLpOeSn4OS5WZJ6qiMZZK+kbTfJ+moKmJ+ArwqqW8yfSrwUB3e3hnAfyLibxUN\nETElImZLOjv7jICkcZIOS16fnmScLennWcuslPRLSS9L+qek/ZIxelPScckyrSXdnaz/gqR+dchp\nBap0SSldumy3Ybp422JKS0vzmCglIhhy6C5ceuSu7P/FrfOdJu/S8nuSlhzNNUsuj+iPByZGxGvA\nMkl9subtDVwK7AF8ETgoOUq+EzgW6AN0qdTfHsCREXE6MBK4OSL2BU4ExiTLPA0cBPQE3gQOTtoP\nAP5dTc4/AqdJKgHKgHfr8N56ATPqsNwGkr4A/Bw4HOgN7CtpcDJ7S+DJiOgJfAT8FDgKOAG4Nlnm\nu0BExJ5kzpDcm4yZmSVGTZ7LLf94jTHT3uSgbp35Yuct8x3JLO9yWehPJ1NESf7NPn3/bEQsjIhy\nYCbQFegBvBURr0dEAA9U6m9sRKxKXh8JjJI0ExgLtE/OIEwDDkl+bgf2lLQ98EFEfFxNzolkiupp\nwJ/q/W5rty8wJSKWRsR64MEkJ8DaJAfALOBfEbEued01af8KyZhExBxgPvC5u40kDZE0XdL0e+77\nXa7ei+VZ8bbFLF68aMN06ZJSiosL5haRevtw1ToAVq5Zz+x3VlCy9RZ5TpRfafk9SUuO5polJ4Ve\nUicyR65jJM0DfgicUnGKHliTtXgZUJfr7tmFejPgyxHRO/nZPiJWAlPJHMUfDEwBlgInkdkBqFJE\nrCVzdP4D4OE65IDMdfw+1cxbz2fHtS5H3euSnRuAcpLxSXaENuqehIgYHRF9I6Lv2d/49sasak3I\nnj17MW/BAt5euJC169YyfuIEDj+0eV/NadViMzYv2mzD612L27F4xeo8p8qvtPyepCVHc83SkDe2\nZTsJuD8ivlPRIOlffHoqvSpzgK6SdomIN6j5Br6/AxcBv0z67h0RMyPibUmdgVYR8aakp4DLgAtr\nyXsTmaPo9z/dF6nR74GhkgZGxPgkwyHA+8A84P8kbQZsD+yXrPMscGuS74Pk/f26LhtLTAPOBJ6U\ntCuwI/C/jVh/o1x59VBmzJzB8uXLGXRCf84753yOHzS49hULOEtacgAUFRVx1dBhnHvBEMrKyzlx\n8Al079YtL1nSMi5tWxdx9kE7A7CZ4IUFy/nf4o8aPQekZ0zS8nuSlhzNNYs+PZBswE6lycDPI2Ji\nVtvFwO5kTo9fFhGDkvZRwPSIuEdSf+AWMjfJTQN2iYhBkoYDKyPixmSdzsBvkv6KgKkRcX4y736g\nRUScIelAMnf9bxMRyypl7AqMi4heldrPBvpGxIXJjXB9I+KqKt5jjyTrLsA64CXgEmAJmVPsfYBX\nga2A4RExRdLpwBWAgPERcXnS18qIaJu8rvxeV0ZE2+R6/O1AXzJnDb4fETXenrli6ccN/+Fag+jQ\nbvN8R9hgxUdral+oEVw3+fV8R9jgJ/265zvCBmn6XbEUa11U7VFqTgq9pYMLfXql6X/eLvSf50Jv\nTU4Nhd5PxjMzMytgLvRmZmYFzIXezMysgLnQm5mZFTAXejMzswLmQm9mZlbAXOjNzMwKmAu9mZlZ\nAXOhNzMzK2Au9GZmZgXMhd7MzKyAudCbmZkVMH+pTSFbvT4VH26rS/+Q7wgbrL2lpm8/tnxKy5fr\ngL9Ipir+fFLOX2pjZmbWPLnQm5mZFTAXejMzswLmQm9mZlbAXOjNzMwKmAu9mZlZAXOhNzMzK2Au\n9GZmZgXMhd7MzKyAudCbmZkVMBd6MzOzAuZCb2ZmVsBc6K1aU5+extHHDeSoQf0Zfdedectx8RG7\nM/OqY3nhJ8dy/zlfYfOi/P3apmVM0pQlLTmuGzGcowcdwWlnnZy3DBXSMiZpyuLPJ39ZmlShlzRY\nUkjqUcflL5W0Rdb0ytyl+9y2u0paJWlm1k+rBuq7r6RbG6Kv6pSVlXHtiOsZc9sdjH90LOMmTmDu\nG3NzuckqfaFjG77brwdfvmECe1/3N1psJk7Zt2uj54D0jEmasqQlB8DAAccy8qZRedl2tjSNSZqy\n+PPJX5YmVeiB04Gnkn/r4lJgi1qXyp03IqJ31s/ahug0IqZHxMUN0Vd1Xpo9i51KSijZoYRWLVsx\nsP8AJk2ZnMtNVqtoM9GmZQtabCbatCxi0fJVecmRpjFJS5a05ADYp3cf2rfvkJdtZ0vTmKQpiz+f\n/GVpMoVeUlvgK8A5wGlZ7YdJmiLpYUlzJD2ojIuBLwCTJU3OWv56SS9K+q+k4qStq6QnJb0kaZKk\nHZP2eyTdKunfkt6UdFLSfp+kwVl9Pijp+Dq8h66Spkl6Pvk5MOs9jMtabpSks5PX+ybbf1HSs5La\nVV4+F0qXlNKly3Ybpou3Laa0tDSXm6zSu8tXcfM/X+GNEV9jwc9P4sPV6/jnq4saPQekZ0zSlCUt\nOdIkTWOSpixpkaYxaawsTabQA8cDEyPiNWCZpD5Z8/Ymc/S+B/BF4KCIuBV4F+gXEf2S5bYE/hsR\nXwKmAucl7b8G7o2IvYAHgezT4tuR2cEYBPwsabsLOBtAUgfgQGB8FZl3yTpt/xtgCXBUROwDnFpp\nO5+TnOr/E3BJkvlIoMbDWUlDJE2XND3f154aQsctWnHsXiXseuWj7HT5w2zZqogz9ts537HMzJqM\nonwH2AinAyOT139Mpmck089GxEIASTOBrmRO8Ve2Fqg4Ep4BHJW8PgD4WvL6fuAXWes8FhHlwCsV\nZwAi4l+SbpO0DXAi8JeIWF/F9t6IiN4VE8lOwShJvYEyYNda3vNuwKKIeC7Z7odJP9WuEBGjgdEA\nrF4ftfRfreJti1m8+NMj59IlpRQXF9e3u3o7okcX5i1byXsr1wDw2AsL+PIu2/D7Z99q9CxpGZM0\nZUlLjjRJ05ikKUtapGlMGitLkziil9QJOBwYI2ke8EPgFH1a8dZkLV5G9Tsw6yIi6rBctuy+syvs\nfcDXgW8Bv6tDPwDfA0qBLwF9gYqb89bz2c+idR37y5k9e/Zi3oIFvL1wIWvXrWX8xAkcfmi/2lds\nYAve/4T9d+5Mm5YtAOjXowtzFq1o9ByQnjFJU5a05EiTNI1JmrKkRZrGpLGyNJUj+pOA+yPiOxUN\nkv4FHFzLeh8B7YD3alnu32Su+98PnAlMq0Ome4BngcUR8UodlgfoACyMiHJJ3wRaJO3zgT0kbQ60\nAY4gc0bif8B2kvaNiOcktaOWU/cNpaioiKuGDuPcC4ZQVl7OiYNPoHu3bo2x6c94bt57PPL8fJ4d\nNpD1ZcHMt99nzFOvN3oOSM+YpClLWnIAXHn1UGbMnMHy5csZdEJ/zjvnfI4fNLj2FRtYmsYkTVn8\n+eQviz49wE2v5Ga6n0fExKy2i4HdyVzDviwiBiXto4DpEXGPpIuAC4F3I6KfpJUR0TZZ7iRgUESc\nLWkn4G6gM7AU+FZELJB0DzAuIh5O1tmwfjI9kcyp/TuqyNw1WbdXVlt34C9AABOB72bl+QVwAvAW\nsBIYm7yHfcncQ9CGTJE/kszZgA3vuVqbcOq+IbW69A/5jrDB2lvq+gcb1thWfLSm9oUaSYd2m+c7\nQur480m51kXVXtNtEoU+jZK/z58F7BMR+TmXXBsX+s9xoU8vF5J08+eTcjUU+iZxjT5tJB0JvAr8\nOrVF3szMjKZzjT5VIuKfwE75zmFmZlYbH9GbmZkVMBd6MzOzAuZCb2ZmVsBc6M3MzAqYC72ZmVkB\nc6E3MzMrYC70ZmZmBcyF3szMrIC50JuZmRUwP+u+kKXkWfdp0vLrN+Q7AgDrHhia7wip42epm20C\nP+vezMyseXKhNzMzK2Au9GZmZgXMhd7MzKyAudCbmZkVMBd6MzOzAuZCb2ZmVsBc6M3MzAqYC72Z\nmVkBc6E3MzMrYC70ZmZmBcyF3szMrIC50JuZmRUwF3qr1tSnp3H0cQM5alB/Rt91Z7PPAXDRgP14\n4VffYebN53PxwP3ymiUt45KWHNeNGM7Rg47gtLNOzluGCmkZkzRlSUuO5pjFhR6QVCZppqQXJT0v\n6cCk/QuSHk5eHyZpXPL6bEmjqumrv6RnJc1J+vyTpB1r2f5gSXs09PvaFGVlZVw74nrG3HYH4x8d\ny7iJE5j7xtxmmwOgZ8k2fPvIvTnwx3fR5we/ZUCf7uzSZau8ZEnLuKQlB8DAAccy8qYq/7NsVGka\nk7RkSUuO5prFhT5jVUT0jogvAUOBGwAi4t2IOKmunUjqBfwa+GZE9IiI3sCDQNdaVh0MpKrQvzR7\nFjuVlFCyQwmtWrZiYP8BTJoyudnmAOixQ2eee/0dVq1dT1l5MPWVBQzev0desqRlXNKSA2Cf3n1o\n375DXradLU1jkpYsacnRXLO40H9ee+ADAEldJc3eiHUvB0ZExKsVDRExNiKmJv2dJ+m55MzBXyRt\nkZw9OA74ZXIGYJfkZ6KkGZKmSeqRrH+ypNnJ+lMb7B1XoXRJKV26bLdhunjbYkpLS3O5yVTnAHh5\nwVIO2n1HOrVtQ5tWRRyzdzdKtm6flyxpGZe05EiTNI1JWrKkJUdzzVLU4D02TW0kzQRaA9sBh9ez\nn57AjTXMfyQi7gSQ9FPgnIj4taSxwLiIqLhMMAk4PyJel7Q/cFuS6Srg6Ih4R1LHqjYgaQgwBOC3\no25jyDnn1fOtWGVz3nmPGx/7N4//5Ew+XrOWF+ctpqw88h3LzKxGLvQZq5LT7Eg6ALgvOQ1fb5K2\nBiYBWwCjI+JGoFdS4DsCbYEnqlivLXAg8GdJFc2bJ/8+Ddwj6SHgkaq2GxGjgdEArF5f7ypUvG0x\nixcv2jBduqSU4uLi+nZXb2nJUeHuJ2dy95MzAbjujH68s+zDvORIy7ikJUeapGlM0pIlLTmaaxaf\nuq8kIv4DdAa2qcfqLwP7JP0sS3YeRpMp6gD3ABdGxJ7ANWTOIFS2GbA8uWeg4mf3pM/zgSuBEmBG\nsjORE3v27MW8BQt4e+FC1q5by/iJEzj80H652lzqc1TYpv0WAJR0bs/g/Xvwh2kbc2Wn4aRlXNKS\nI03SNCZpyZKWHM01i4/oK0muh7cAlpE5Gt8YvwAelfTfrOv02X20AxZJagmcCbyTtH+UzCMiPpT0\nlqSTI+LPyhzW7xURL0raJSKeAZ6RdAyZgr+sPu+zNkVFRVw1dBjnXjCEsvJyThx8At27dcvFpppE\njgoP/fBkOrVtw/qyci4e8zgrPlmTlxxpGZe05AC48uqhzJg5g+XLlzPohP6cd875HD9ocKPnSNOY\npCVLWnI01yyK8DVGSWXArIpJ4IqIGC+pK5lr570kHQZcFhGDJJ0N9I2IC6voayAwnMxNfe8BC4Cr\nI+I1SRcAPwKWAs8A7SLibEkHAXcCa4CTgHLgdjL3C7QE/hgR10p6BOieZJwEXBo1fYCbcOq+ULX8\n+g35jgDAugeG5jtC6qz4KD87TVXp0G7z2hcyS5PWRapulgt9IXOh/xwX+vRyoTfbBDUUel+jNzMz\nK2Au9GZmZgXMhd7MzKyAudCbmZkVMBd6MzOzAuZCb2ZmVsBc6M3MzAqYC72ZmVkBc6E3MzMrYC70\nZmZmBcyF3szMrID5WfeFzM+6T620PHMf/Nz9qvi5+9bk+Fn3ZmZmzZMLvZmZWQFzoTczMytgLvRm\nZmYFzIXezMysgLnQm5mZFTAXejMzswLmQm9mZlbAXOjNzMwKmAu9mZlZAXOhNzMzK2Au9GZmZgXM\nhd6qNfXpaRx93ECOGtSf0Xfd2exzpCnLRQP244VffYeZN5/PxQP3y1sOSM+YpCnLdSOGc/SgIzjt\nrJPzlqFCWsYkLTmaY5ZUF3pJXST9UdIbkmZImiBp1xqW7yjp/+qxnXmSZkmamfx7fNa8lfXNX822\nDpM0biPXOVvSF7KmL5W0RUPmqqysrIxrR1zPmNvuYPyjYxk3cQJz35iby02mOkeasvQs2YZvH7k3\nB/74Lvr84LcM6NOdXbps1eg5ID1jkrYsAwccy8ibRuVl29nSMiZpydFcs6S20EsS8CgwJSJ2iYg+\nwFCguIbVOgIbXegT/SKiN3AScGs9+8iVs4EvZE1fCuS00L80exY7lZRQskMJrVq2YmD/AUyaMjmX\nm0x1jjRl6bFDZ557/R1WrV1PWXkw9ZUFDN6/R6PngPSMSdqy7NO7D+3bd8jLtrOlZUzSkqO5Zklt\noQf6Aesi4o6Khoh4MSKmSWoraZKk5ysdgf8M2CU5Mv8lgKQfSnpO0kuSrqnDdtsDH1RurG6bkrpK\nelXSnZJelvR3SW2Sed0k/VPSi8l6uyTdtZX0sKQ5kh5MdmqQ1EfSv5KzF09I2k7SSUBf4MHkfV1C\npuhPlpSz387SJaV06bLdhunibYspLS3N1eZSnyNNWV5esJSDdt+RTm3b0KZVEcfs3Y2Srds3eg5I\nz5ikLUtapGVM0pKjuWYpavAeG04vYEY181YDJ0TEh5I6A/+VNBb4MdArOTJH0leB7sB+gICxkg6J\niKlV9Dk5KbhfBE7ZiG2SbOP0iDhP0kPAicADwIPAzyLiUUmtyexYlQB7Az2Bd4GngYMkPQP8Gjg+\nIpZKOhW4PiK+LelC4LKImJ68r++ROQPxXl0G0grLnHfe48bH/s3jPzmTj9es5cV5iykrj3zHMrOU\nSnOhr4mAEZIOAcqB7an6lP5Xk58Xkum2ZIpyVYW+X0S8lxx1T5I0JSKyr8/XtM23ImJm8noG0FVS\nO2D7iHgUICJWAyQH789GxMJkeibQFVhOZufmH8kyLYBFdR6RipDSEGAIwG9H3caQc87b2C6AzJ7l\n4sWfbr50SSnFxTVdNcmNtORIW5a7n5zJ3U9mfuWuO6Mf7yz7MC850jQmacqSFmkZk7TkaK5Z0nzq\n/mWgTzXzzgS2AfokR++lQOsqlhNwQ0T0Tn66RcRdNW00It5I+ttjI7a5Jmu5MmrfgapqeQEvZ2Xd\nMyK+Wks/VeUfHRF9I6JvfYs8wJ49ezFvwQLeXriQtevWMn7iBA4/tF+9+2vqOdKWZZv2mVs0Sjq3\nZ/D+PfjDtNl5yZGmMUlTlrRIy5ikJUdzzZLmI/onyRxBD4mI0QCS9gI6JD9LImKdpH7ATsk6HwHt\nsvp4ArhO0oMRsVLS9mSu+y+pbqOStgV2BuZXmlXdNqsUER9JWihpcEQ8JmlzMkfp1fkfsI2kAyLi\nP5JaArtGxMtVvK+K6Zydui8qKuKqocM494IhlJWXc+LgE+jerVuuNpf6HGnL8tAPT6ZT2zasLyvn\n4jGPs+KTNbWvlANpGpM0Zbny6qHMmDmD5cuXM+iE/px3zvkcP2hwo+dIy5ikJUdzzaKI9F7bS/6k\n7BYyR/argXlk7jj/APgbmVPx04EvA8dExDxJvwf2Ah6PiB8mN6+dm3S5Evh6ctSevZ15ZIpnGdAS\nuCkifpfMWxkRbZPr8p/bZtLFuIjolSx/GdA2IoZL6g78FugMrANOBnYkc719ULL8KGB6RNwjqTeZ\nO/47kNkJuyUi7pR0IjACWAUckLyfC4F3I6L63b/V69P74TZzLb9+Q74jbLDugaH5jpA6Kz7Kz45T\nVTq02zzfEawpaF2k6malutDbJnKhTy0X+nRzobcmp4ZCn+Zr9GZmZraJXOjNzMwKmAu9mZlZAXOh\nNzMzK2Au9GZmZgXMhd7MzKyAudCbmZkVMBd6MzOzAuZCb2ZmVsBc6M3MzAqYC72ZmVkB87PuC5mf\ndW910OrSP+Q7AgCv/OCY2hdqJN1KOuY7gtnG8bPuzczMmicXejMzswLmQm9mZlbAXOjNzMwKmAu9\nmZlZAXOhNzMzK2Au9GZmZgXMhd7MzKyAudCbmZkVMBd6MzOzAuZCb2ZmVsBc6M3MzAqYC72ZmVkB\nc6G3ak19ehpHHzeQowb1Z/Rddzb7HM5StYuP2J2ZVx3LCz85lvvP+QqbF+XnfytLl5by46EXcv4F\nZ3LB/53JX//6UF5yQHo+mzRlSUuO5pilSRV6SWWSZkp6WdKLkn4gqcHfg6QpkvpW0X62pFGV2r6V\nZJopaa2kWcnrnzV0rsZUVlbGtSOuZ8xtdzD+0bGMmziBuW/MbbY5nKVqX+jYhu/268GXb5jA3tf9\njRabiVP27droOQBatGjBuedcxB23P8hNN45m3PhHWLDgrUbPkZbPJk1Z0pKjuWZpUoUeWBURvSOi\nJ3AUcAxwdT4DRcTdSabewLtAv2T6xxXLSGqRv4T189LsWexUUkLJDiW0atmKgf0HMGnK5Gabw1mq\nV7SZaNOyBS02E21aFrFo+aq85OjUqTPduu0GwBZbbElJyU4sW7a00XOk6bNJS5a05GiuWZpaod8g\nIpYAQ4ALldFa0t3JEfULkoSkf7QAACAASURBVPrBhqPwRyRNlPS6pF9U9CHpdknTkzME11S1neSI\n/TVJzwIH1TWfpJWSbpL0InCApHmSOifz+kqakrweLuleSdMkzZf0NUm/SN7HREktk+XmZbU/K6lb\nPYeuTkqXlNKly3Ybpou3Laa0tDSXm0x1Dmep2rvLV3HzP1/hjRFfY8HPT+LD1ev456uLGj1HZaWl\ni3jzzdfZbbeejb/tlHw2acqSlhzNNUuTLfQAEfEm0ALYFvhupin2BE4H7pXUOlm0N3AqsCdwqqSS\npH1YRPQF9gIOlbRXdv+StgOuIVPgvwLssRHxtgSeiYgvRcRTtSy7C3A4cBzwADA5eR+rgIFZy61I\n2kcBt1TVkaQhyc7L9Hxfe7LC13GLVhy7Vwm7XvkoO13+MFu2KuKM/XbOa6ZVqz7h+hHDOO+8i9li\niy3zmsUsDZp0oa/kK2SKJBExB5gP7JrMmxQRKyJiNfAKsFPSfoqk54EXgJ58vpDvD0yJiKURsRb4\n00bkKQP+UsdlH4+IdcAsMjsuE5P2WUDXrOX+kPXvAVV1FBGjI6JvRPQdcs55GxH3s4q3LWbx4k+P\nzEqXlFJcXFzv/pp6Dmep2hE9ujBv2UreW7mG9eXBYy8s4Mu7bNPoOSqsX7+eESOG0e+wr3LQgYfl\nJUNaPps0ZUlLjuaapUkXeklfJFNQl9Sy6Jqs12VAkaSdgcuAIyJiL2A80LqqletpdUSUZU2v59Px\nrrydNQARUQ6si4hI2suBoqzloprXDW7Pnr2Yt2ABby9cyNp1axk/cQKHH9ovl5tMdQ5nqdqC9z9h\n/50706Zl5jaUfj26MGfRikbPARARjBx5AyUlO3HCCaflJQOk57NJU5a05GiuWYpqXySdJG0D3AGM\nioiQNA04E3hS0q7AjsD/gH2q6aI98DGwQlIxmRv7plRa5hlgpKStgQ+Bk4EX6xl5HtAHeBw4sZ59\nnAr8LPn3P/Xso06Kioq4augwzr1gCGXl5Zw4+AS6d8vpbQGpzuEsVXtu3ns88vx8nh02kPVlwcy3\n32fMU683eg6AV155iScnT6Rr11248KJvAvDNb3yHffc9sFFzpOWzSVOWtORorln06cFj+kkqI3M6\nuyWZI+T7gV9FRHlyPf52oG8y7/sRMVnS2UDfiLgw6WMccGNETJF0D3Ag8DawAhgbEfckN8pdFhHT\nJX0LGAosB2YCayv6qiLfvGRb70laGRFts+YdDNxFZodhSrLcYZKGAysj4sZkuQ3rZc9L+v4TmR2S\nNcDpEVHz32GsXt90PlzLm1aX/qH2hRrBKz84Jt8RNuhW0jHfEcw2TusiVTerSRX65ix7J6LOK7nQ\nWx240H+eC701OTUU+iZ9jd7MzMxq1mSv0Tc3EdE13xnMzKzp8RG9mZlZAXOhNzMzK2Au9GZmZgVs\nowq9pM0ktc9VGDMzM2tYtRZ6Sb+X1F7SlsBs4BVJP8x9NDMzM9tUdTmi3yMiPgQGk3mq287AWTlN\nZWZmZg2iLoW+ZfJVqYPJPDluHTl+zrqZmZk1jLoU+t+SeU77lsBUSTuReYyrmZmZpVytD8yJiFuB\nW7Oa5kvKz1f9mJmZ2Uap9Vn3ki4B7gY+AsYAewM/joi/5z6ebRI/696akBb90nPrT9nk+/MdwWzj\nbOKz7r+d3Iz3VWArMjfi/ayBopmZmVkO1aXQV+wlDADuj4iXs9rMzMwsxepS6GdI+juZQv+EpHZA\neW5jmZmZWUOoy7fXnQP0Bt6MiE8kbQ18K7exzMzMrCHU5a77cklvAbtKat0ImczMzKyB1FroJZ0L\nXALsAMwEvgz8Bzg8t9HMzMxsU9XlGv0lwL7A/IjoR+bP65bnNJWZmZk1iLoU+tURsRpA0uYRMQfY\nLbexzMzMrCHU5Wa8hZI6Ao8B/5D0ATA/t7HMzMysIdTlZrwTkpfDJU0GOgATc5rKzMzMGkS1hV5S\npyqaZyX/tgXez0kiMzMzazA1HdHPIPN1tNlPwauYDuCLOcxlZmZmDaDam/EiYueI+GLy786Vpl3k\nm4GpT0/j6OMGctSg/oy+685mn8NZ0pdjzLDvsGj8Hbz4wC82tG3VfkueGHkFcx76FU+MvIKO7bZs\n1EyQns8mTVnSkqM5Zqm20Es6WtJJVbSfKOmonKTJEUllkmZm/XSt43rDJV2WvL6nqvGoYd0Ny0vq\nJOkFSd+SdJikcfV5H1l995V0a+1L1l9ZWRnXjrieMbfdwfhHxzJu4gTmvjE3l5tMdQ5nSWeOe8f/\niwHf++x3bF1+1vFMmj6bHqd8n0nTZ3P5Wcc1Wh7I/5ikMUtacjTXLDX9ed1VwL+qaP8XcG2DJ8mt\nVRHRO+tnXmNtWFIH4AlgdETc3QD9FUXE9Ii4eNPTVe+l2bPYqaSEkh1KaNWyFQP7D2DSlMm53GSq\nczhLOnNMmzmH9z9c+Zm24w7uw30TpgJw34SpHH9I30bLA/kfkzRmSUuO5pqlpkK/eUQsrdwYEe8B\njX8urAFJaitpkqTnJc2SdHzWvGGSXpP0FNU8L0DSEckR+ixJv5O0eTWbags8Dvw+Im7Pbpf0sKQ5\nkh6UpKTfeZI6J6/7SpqSvB4u6X5JTwP3N8RZgdqULimlS5ftNkwXb1tMaWlpLjeZ6hzOku4c2Yo7\ndWDxsswzvRYvW05xpw6Nuv00jUlasqQlR3PNUlOhby/pczfrSWoJtGnwJLnVJuu0/aPAauCEiNgH\n6AfcpIw+wGlkvsRnAJknAn5G8rz/e4BTI2JPMjc0XlDNdn8FPBURN1dq3xu4FNiDzE2NB9XhPewB\nHBkRp9dhWTNLRES+I5jlVU2F/hHgTkkbjt4ltQXuSOY1Jdmn7k8g85cDIyS9BPwT2B4oBg4GHo2I\nTyLiQ2BsFX3tBrwVEa8l0/cCh1Sz3SeB4yVtW6n92YhYGBHlZL4/oGsd3sPYiFhV20KShkiaLmn6\nptzYUbxtMYsXL9owXbqklOLi4nr319RzOEu6c2QrfX8FXbbuCECXrTuy5IMPG3X7aRqTtGRJS47m\nmqWmQn8lUArMlzRD0gzgLWBpMq8pOxPYBugTEb3JvM9cfDPfH8nsGE2Q1C6rfU3W6zI+/TPH9Xz6\nmVTO83FdNhgRoyOib0T0HXLOefWInLFnz17MW7CAtxcuZO26tYyfOIHDD+1X7/6aeg5nSXeObH97\nagbfGJDZ9/7GgEMYO21Go24/TWOSlixpydFcs1T7d/QRsR74saRrgG5J89y6HFU2AR2AJRGxTlI/\nYKekfSpwj6QbyIzNscBvK637P6CrpG4RMRc4i6pvWgQgIm6W1AV4RNLAWnLNA/qQua5/4ka+pwZV\nVFTEVUOHce4FQygrL+fEwSfQvVu32lcs0BzOks4cD15zEYfuszudO7Zj/l9Hcc2Yh/n5fWP54/WX\n8O1jD2P+4vc47cqRjZYH8j8macySlhzNNYuaw/UrSSsjom3WdGfgb2RulptO5qt3j4mIeZKGAd8E\nlgALgOcj4kZJ9wDjIuJhSUcAN5LZGXgOuCAi1lTa5oblk+m7gS3I7Dh8PyIGJe2jgOkRcY+kg4G7\ngA+BKUDfiDhM0nBgZUTcmKxzGHBZRR/VWr2+8D9cKxgt+p2V7wgblE2+P98RzDZO6yJVN6tZFPpm\ny4XemhAXerNNUEOhr8vX1JqZmVkTVWuhT/7s7OuSrkqmd5S0X+6jmZmZ2aaqyxH9bcABQMXfb38E\n/CZniczMzKzB1Pp99MD+EbGPpBcAIuIDSa1ynMvMzMwaQF2O6NdJakHmq2mRtA1QntNUZmZm1iDq\nUuhvBR4FtpV0PfAUMCKnqczMzKxB1HrqPiIeTJ6KdwSZR8cOjohXc57MzMzMNlmthV7SjsAnZB4w\ns6EtIhbkMpiZmZlturrcjDeezPV5kXn++s5kHgPbM4e5zMzMrAHU5dT9ntnTkvYB/i9niczMzKzB\nbPST8SLieWD/HGQxMzOzBlaXa/Tfz5rcDNgHeDdniczMzKzB1PqlNpKuzppcT+arVP8SEatzmMsa\ngr/Uxqxe/AU71uTU8KU2NR7RJw/KaRcRlzV4KDMzM8u5aq/RSyqKiDLgoEbMY2ZmZg2opiP6Z8lc\nj58paSzwZ+DjipkR8UiOs5mZmdkmqsvf0bcGlgGH8+nf0wfgQm9mZpZyNRX6bZM77mfzaYGv4Ju8\nzMzMmoCaCn0LoC2fLfAVXOjNzMyagJoK/aKIuLbRkpiZmVmDq+nJeNX+TZ6ZmZk1DTUV+iMaLYWZ\nmZnlRLWFPiLeb8wgZmZm1vA2+kttzMzMrOlwoTczMytgLvRWralPT+Po4wZy1KD+jL7rzmafw1nS\nnSPfWcYM+w6Lxt/Biw/8YkPbVu235ImRVzDnoV/xxMgr6Nhuy0bNBOn5fNKSozlmSWWhl7Sy0vTZ\nkkY14va7Sppdl3ZJwyUV3Jf+lJWVce2I6xlz2x2Mf3Qs4yZOYO4bc5ttDmdJd440ZLl3/L8Y8L2f\nfabt8rOOZ9L02fQ45ftMmj6by886rtHyQP7HJG05mmuWVBZ6y7+XZs9ip5ISSnYooVXLVgzsP4BJ\nUyY32xzOku4cacgybeYc3v/wM8coHHdwH+6bMBWA+yZM5fhD+jZaHsj/mKQtR3PN0uQKvaRjJT0j\n6QVJ/5RUnLQPl3SvpGmS5kv6mqRfSJolaaKklslyV0l6TtJsSaMlKWnvI+lFSS8C361ntvOSvl+U\n9BdJWyTtJyfbe1HS1KStp6RnJc2U9JKk7kn7Y5JmSHpZ0pCsvldKuj7p478V7ztXSpeU0qXLdhum\ni7ctprS0NJebTHUOZ0l3jrRl2ZChUwcWL1sOwOJlyynu1KFRt5+WMUlLjuaaJa2Fvk1SAGdKmglk\nP6HvKeDLEbE38EfgR1nzdiHz5TvHAQ8AkyNiT2AVMDBZZlRE7BsRvYA2wKCk/W7gooj4Ui3ZdqmU\n7fyseY8kfX8JeBU4J2m/Cjg6aa84d3c+MDIiegN9gYVJ+7cjok/SdrGkrZP2LYH/Jn1MBc6rKpyk\nIZKmS5qe72tPZvZZEX56uDW+unx7XT6sSgogkLlGT6bwAewA/EnSdkAr4K2s9R6PiHWSZpF5Vv/E\npH0W0DV53U/Sj4AtgE7Ay5KmAR0jYmqyzP3AMdVke6NStuFZ83pJ+inQkcz3BDyRtD8N3CPpIT79\n1r//AMMk7UBmB+H1pP1iSSckr0uA7mS+PXAtMC5pnwEcVVW4iBgNjAZg9fp6/1+leNtiFi9etGG6\ndEkpxcU5PYmQ6hzOku4cacuyIcP7K+iydUcWL1tOl607suSDDxt1+2kZk7TkaK5Z0npEX5Nfkzkq\n3xP4Dpmv0a2wBiAiyoF18enuczlQJKk1cBtwUrL+nZXW31T3ABcmfV9T0XdEnA9cSaZwz5C0dUT8\nnszR/SpggqTDJR0GHAkckBy5v5CVL/v9lJHjnbQ9e/Zi3oIFvL1wIWvXrWX8xAkcfmi/XG4y1Tmc\nJd050palwt+emsE3BhwCwDcGHMLYaTMadftpGZO05GiuWdJ6RF+TDsA7yetvbuS6FUXzPUltgZOA\nhyNiuaTlkr4SEU8BZ9YzWztgUXI/wJkVOSXtEhHPAM9IOgYokdQBeDMibpW0I7AXmbMTH0TEJ5J6\nAF+uZ45NVlRUxFVDh3HuBUMoKy/nxMEn0L1bt2abw1nSnSMNWR685iIO3Wd3Ondsx/y/juKaMQ/z\n8/vG8sfrL+Hbxx7G/MXvcdqVIxstD+R/TNKWo7lmURqvGUlaGRFts6bPBvpGxIWSjgduBj4AngT2\njYjDklPoKyPixsp9ZM9LTq2fDiwGXgPmR8RwSX2A35H5Ct6/AwOS6/jZuboC47LbK/V9AZl7BpYC\nzwDtIuJsSY+QOQUvYBJwKXA5cBawLslyBvAx8BiZywz/I3MJYHhETKn0fk4CBkXE2TUO5Cacujdr\nzlr0OyvfETYom3x/viNYU9C6qNovoktlobcG4kJvVi8u9Nbk1FDom+I1ejMzM6sjF3ozM7MC5kJv\nZmZWwFzozczMCpgLvZmZWQFzoTczMytgLvRmZmYFzIXezMysgLnQm5mZFTAXejMzswLmQm9mZlbA\n/Kz7QuZn3Zs1eWl57r6fuZ9yfta9mZlZ8+RCb2ZmVsBc6M3MzAqYC72ZmVkBc6E3MzMrYC70ZmZm\nBcyF3szMrIC50JuZmRUwF3ozM7MC5kJvZmZWwFzozczMCpgLvZmZWQFzoTczMytgLvRWralPT+Po\n4wZy1KD+jL7rzmafw1nSnSNNWfKZY8yw77Bo/B28+MAvNrRt1X5Lnhh5BXMe+hVPjLyCju22bNRM\nkJ7PpjlmaVKFXtLKrNcDJL0maac85JgnaZaklyT9Kx8Zcq2srIxrR1zPmNvuYPyjYxk3cQJz35jb\nbHM4S7pzpClLvnPcO/5fDPjezz7TdvlZxzNp+mx6nPJ9Jk2fzeVnHddoeSD/Y9LcszSpQl9B0hHA\nrcAxETE/TzH6RcRewBTgyjxlyJmXZs9ip5ISSnYooVXLVgzsP4BJUyY32xzOku4cacqS7xzTZs7h\n/Q9XfqbtuIP7cN+EqQDcN2Eqxx/St9HyQP7HpLlnaXKFXtIhwJ3AoIh4I2n7uqRnJc2U9FtJLZL2\n2yVNl/SypGuy+pgn6RpJzydH5j2S9uGSLstabrakrrVE+g+wfbL8NpL+Ium55OegpH1LSb9LMr4g\n6fik/b+SemZtb4qkvjUs3zPrfb4kqfumjmd1SpeU0qXLdhumi7ctprS0NFebS30OZ0l3jjRlSUuO\nbMWdOrB42XIAFi9bTnGnDo26/TSNSXPM0tQK/ebAY8DgiJgDIGl34FTgoIjoDZQBZybLD4uIvsBe\nwKGS9srq672I2Ae4HbiM+uufZAIYCdwcEfsCJwJjKnIAT0bEfkA/4JeStgT+BJySvI/tgO0iYnoN\ny58PjEzeZ19gYeUwkoYkOzfT833tyczSKSLyHcEaUVG+A2ykdcC/gXOAS5K2I4A+wHOSANoAS5J5\np0gaQuZ9bgfsAbyUzHsk+XcG8LV6ZJksqROwEvhJ0nYksEeSA6C9pLbAV4Hjss4WtAZ2BB4C/g5c\nTabgP5zMr275/wDDJO0APBIRr1cOFRGjgdEArF5f7/+ai7ctZvHiRRumS5eUUlxcXN/u6i0tOZwl\n3TnSlCUtObKVvr+CLlt3ZPGy5XTZuiNLPviwUbefpjFpjlma2hF9OZmCuJ+kK5I2AfdGRO/kZ7eI\nGC5pZzJH6kck19LHkymYFdYk/5bx6Q7Pej47JtnLV9YP2AmYCVRcFtgM+HJWlu0jYmWS8cSs9h0j\n4tWIeAdYlpxpOJXMET41LP974DhgFTBB0uF1HbiNtWfPXsxbsIC3Fy5k7bq1jJ84gcMP7ZerzaU+\nh7OkO0easqQlR7a/PTWDbww4BIBvDDiEsdNmNOr20zQmzTFLUzuiJyI+kTQQmCapFJgE/FXSzRGx\nJDnKbge0Bz4GVkgqBo4hc+NcTeYBgwAk7QPsXEuW9ZIuBWZJ+imZo/OLgF8mffSOiJnAE8BFki6K\niJC0d0S8kHTzJ+BHQIeIqDjbUOXykr4IvBkRt0rakcwliSfrMm4bq6ioiKuGDuPcC4ZQVl7OiYNP\noHu3brnYVJPI4SzpzpGmLPnO8eA1F3HoPrvTuWM75v91FNeMeZif3zeWP15/Cd8+9jDmL36P064c\n2Wh5IP9j0tyzqCldq5G0MiLaJq9LgKlkTuG3AYaSOaJeB3w3Iv4r6R7gQOBtYAUwNiLukTQP6BsR\n70nqC9wYEYdJagP8lczNdc8AB5C5s39epRwb1k+mf03mcsHtwG+A3cnsRE2NiPOTfm9JsmwGvBUR\nFTsUxcA7wHURcU3SVuXykn4MnJW8x8XAGRHxfrUDtgmn7s0sHVr0OyvfEQAom3x/viNYTVoXqbpZ\nTarQ20ZyoTdr8lzorU5qKPRN7Rq9mZmZbQQXejMzswLmQm9mZlbAXOjNzMwKmAu9mZlZAXOhNzMz\nK2Au9GZmZgXMhd7MzKyAudCbmZkVMBd6MzOzAuZCb2ZmVsD8rPtC5mfdWxOy4qM1tS/USDq02zzf\nEVKn5ddvyHeEDdY9MDTfEdLHz7o3MzNrnlzozczMCpgLvZmZWQFzoTczMytgLvRmZmYFzIXezMys\ngLnQm5mZFTAXejMzswLmQm9mZlbAXOjNzMwKmAu9mZlZAXOhNzMzK2Au9FatqU9P4+jjBnLUoP6M\nvuvOZp/DWdKd47oRwzl60BGcdtbJectQIS1jkqYsFw3Yjxd+9R1m3nw+Fw/cL285ID1j0lhZUlPo\nJYWkB7KmiyQtlTQux9udIqlvFe1FkkZIel3SzORnWDKvq6TZucyVb2VlZVw74nrG3HYH4x8dy7iJ\nE5j7xtxmm8NZ0p0DYOCAYxl506i8bDtbmsYkLVl6lmzDt4/cmwN/fBd9fvBbBvTpzi5dtmr0HJCe\nMWnMLKkp9MDHQC9JbZLpo4B3NqYDSUUNmOenwBeAPSOiN3Aw0LIB+0+1l2bPYqeSEkp2KKFVy1YM\n7D+ASVMmN9sczpLuHAD79O5D+/Yd8rLtbGkak7Rk6bFDZ557/R1WrV1PWXkw9ZUFDN6/R6PngPSM\nSWNmSVOhB5gADExenw78oWKGpP0k/UfSC5L+LWm3pP1sSWMlPQlMktRW0t2SZkl6SdKJyXJfTdZ/\nXtKfJbWtLoSkLYDzgIsiYjVARHwUEcOzFmsh6U5JL0v6e8UOiqTzJD0n6UVJf0n6QtI9km5Nsr8p\n6aSs7f0wWeclSdckbV0lzUnWe03Sg5KOlPR0cpYhp+e+SpeU0qXLdhumi7ctprS0NJebTHUOZ0l3\njjRJ05ikJcvLC5Zy0O470qltG9q0KuKYvbtRsnX7Rs8B6RmTxsyStkL/R+A0Sa2BvYBnsubNAQ6O\niL2Bq4ARWfP2AU6KiEOBnwArImLPiNgLeFJSZ+BK4MiI2AeYDny/hhzdgAUR8VENy3QHfhMRPYHl\nwIlJ+yMRsW9EfAl4FTgna53tgK8Ag4CfQWYHJOlrP6A30EfSIVk5bgJ6JD9nJOtfBlxRQzYzs9SY\n88573PjYv3n8J2cy/sozeHHeYsrKI9+xmo2GPNW9ySLiJUldyRzNT6g0uwNwr6TuQPDZ0+j/iIj3\nk9dHAqdl9fmBpEHAHsDTkgBaAf+pay5J3wIuAbYGDkya34qImcnrGUDX5HUvST8FOgJtgSeyunos\nIsqBVyQVJ21fTX5eSKbbkin8C5JtzEoyvAxMioiQNCtre5WzDgGGAPx21G0MOee8ur7NzyjetpjF\nixdtmC5dUkpxcXENa+RGWnI4S7pzpEmaxiRNWe5+ciZ3P5n5X+Z1Z/TjnWUf5iVHmsaksbKk7Yge\nYCxwI1mn7RPXAZMjohdwLNA6a97HtfQpMjsDvZOfPSLinBqWnwvsKKkdQETcnVynXwG0SJZZk7V8\nGZ/uNN0DXBgRewLXVMqZvY6y/r0hK1u3iLiriuXLs6bLqWYnLSJGR0TfiOhb3yIPsGfPXsxbsIC3\nFy5k7bq1jJ84gcMP7Vfv/pp6DmdJd440SdOYpCnLNu23AKCkc3sG79+DP0zLz/3MaRqTxsqSqiP6\nxO+A5RExS9JhWe0d+PTmvLNrWP8fwHeBSwEkbQX8F/iNpG4RMVfSlsD2EfFaVR1ExCeS7gJGSfpO\nRKyW1ILMmYDatAMWSWoJnEntNxQ+AVwn6cGIWClpe2BdHbaTU0VFRVw1dBjnXjCEsvJyThx8At27\ndWu2OZwl3TkArrx6KDNmzmD58uUMOqE/551zPscPGtzoOdI0JmnK8tAPT6ZT2zasLyvn4jGPs+KT\nNbWvlANpGpPGyqKIdFwn0f+3d97hclVl+76fJIRQg7QA0mvoJYkUkSpKLwoIUkRBkA+pPxQRPykK\nioiKoCCCUsWCdJBqgADyQQKBhKJUAYHQDBBaSPL8/lhrkjmTk5MAmbU3k/e+rnOd2Xv2nvXMPnPm\n3etdb5HG2Z67Zd8mwJG2t5W0PnA+afZ+LbCn7aUl7QMMtv3NfM7cwK+AQaSZ9vG2L5O0GXAyMHt+\n+e/ZvkrSrXmM4S1jz0byIuwMvAm8k8c9hRSNf032LiDpSGBu28dJOhD4NvAyKcZgHtv7SDovn3Np\n6/uVdCiwXx56HLBn1t48xuTz8/LG5OemybsT6vHHDYIZ4PU3q/ni747+88w+/YNmMWbb80dVS5jM\n+xcdXbWE+tGvj6b1VG0MfdAGwtAHHyPC0NebMPQ1pwdDX8c1+iAIgiAIZhJh6IMgCIKggwlDHwRB\nEAQdTBj6IAiCIOhgwtAHQRAEQQcThj4IgiAIOpgw9EEQBEHQwYShD4IgCIIOJgx9EARBEHQwYeiD\nIAiCoIMJQx8EQRAEHUzUuu9kotZ9EAQdSO9N96paAgATh15YtYQpRK37IAiCIJg1CUMfBEEQBB1M\nGPogCIIg6GDC0AdBEARBBxOGPgiCIAg6mDD0QRAEQdDBhKEPgiAIgg4mDH0QBEEQdDBh6IMgCIKg\ngwlDHwRBEAQdTBj6IAiCIOhgwtAHQRAEQQcThj4IgiAIOpgw9ME0uf3OYXx++23YYtstOfvc387y\nOkJLvXXUSUtddNRJS5U6zjnmAF649iweuOgnk/d9Yt65uOG07/Lon3/GDad9l/nmmauopgYlrktb\nDb2kYyQ9JOlBSSMlrdvO8XrQcZykI2d0fw+vc56knT/E+JtIuuaDnlclEydO5ISTTuScX5/FtZdf\nxTXXX8fjTzw+y+oILfXWUSctddFRJy1V6zj/2tvY+vAfd9l31F47cMvw0Qzc9QhuGT6ao/bavpie\nBqWuS9sMvaT1gW2BdWyvAXwWeLZd4wUzlwdHj2KpJZZgicWXoO9sfdlmy6255dahs6yO0FJvHXXS\nUhcdddJStY5hIx/ltTfGddm3/WcGccF1twNwwXW3s8NGg4vpaVDqurRzRr8o8Irt9wBsv2L7eQBJ\n35d0r6TRks6WpLz/Tcr1KwAAIABJREFUVkk/lzRc0iOShki6TNJjkn6Yj1la0ujGIJKOlHRcfryc\npOsljZA0TNLAdr25Ht7D8pJulvSApPskLddy3hBJ92etG2dPx8i8bx4lTsmvO0rSl/J5m+Trc6mk\nRyVd3BizHYx5aQyLLLLo5O0BCw9gzJgx7Rqu9jpCS7111ElLXXTUSUtddDQzYP7+vPjqWABefHUs\nA+bvX1xDqevSTkN/I7CEpH9J+rWkjZueO8P2ENurAXOQZv4NxtseDJwFXAkcBKwG7CNpgemMeTZw\nsO1BwJHAr2fWm+mGab2Hi4Ff2V4T2AB4oXGCpA1I72sH209kjQfZXgv4DPAO8AVgLWBNkhfkFEmN\nT8LawGHAKsCywKdbRUnaP98oDa96bTAIguDjgu2qJbSNPu16YdvjJA0iGbBNgT9J+o7t84BNJX0b\nmBOYH3gIuDqfelX+PQp4yPYLAJKeBJYAxnY3nqS5SYb1L00T3dln9vtqYqr3IOlW4JO2Lwew/W7W\nBrAy6Ubkcw3PBnAn8DNJFwOX2X5O0obAJbYnAmMk3QYMAd4A7rH9XH7NkcDSwB3NomyfnceBdyd8\n6E/ugIUH8OKLk+9RGPPSGAYMGPBhX+5DUxcdoaXeOuqkpS466qSlLjqaGfPa6yyywHy8+OpYFllg\nPl767xvFNZS6Lm0NxrM90fatto8Fvgl8UVI/0kx7Z9urA78F+jWd9l7+PanpcWO7DzChRXfj3F7A\nWNtrNf2sPPPfFczAe+iOF4B3SbNyAGz/GNiP5BG4cwaWGpqvx0TaeKO2+qqr8fQzz/Dsc88x/v3x\nXHv9dWy28abtGq72OkJLvXXUSUtddNRJS110NHP1HSPYe+uNANh76424atiI4hpKXZe2GQpJKwGT\nbD+Wd60F/JspBvGVPAvfGbj0A7z0GGDh7MYfR3KZX2/7DUlPSdrF9l/y+vUath+YKW+oK92+B9tv\nSnpO0o62r5A0O9A7HzsW2Be4SdJbtm+VtJztUcAoSUOAgcAw4ABJ55M8BRsB38rPFaNPnz58/+hj\n2O/A/Zk4aRJf3HEnVlh++ZISaqUjtNRbR5201EVHnbRUrePi4w9m43VWZsH55uHfV57B8edcyskX\nXMUfTzyUr223Cf9+8RV2+95pxfQ0KHVd1K51iey2Px2YjzQLfxzY3/YrObBud+BF4F/Av20fl13f\nR9oeLmmT/Hjb/HrNzx0CHAr8B3gSeDqfvwxwJikQcDbgj7ZPyMF642z/tEXjcaQ178nhmLYXl3Qd\nsF+Ti71x/EXAn2xf3cN7WAH4DbAg8D6wC7Bk471IWhL4G/A1YE/SssYk0vLFPsB44CfAVoCBH9r+\nUzfX4wxgeF4K6Z6P4LoPgiCoK7033atqCQBMHHph1RKm0K/PNIOz22boOw1JvYB7gb1sP1y1nhki\nDH0QBB1IGPpu6MHQR2W8GUDSYsBo4O6PjZEPgiAIAtq4Rt9JZBf+KlXrCIIgCIIPSszogyAIgqCD\nCUMfBEEQBB1MGPogCIIg6GDC0AdBEARBBxOGPgiCIAg6mDD0QRAEQdDBhKEPgiAIgg4mDH0QBEEQ\ndDBh6IMgCIKgg4la951M1LoPgg/F62++N/2DCtF/ntmrlgDENemOvoddUrWEyYw/a6+odR8EQRAE\nsyJh6IMgCIKggwlDHwRBEAQdTBj6IAiCIOhgwtAHQRAEQQcThj4IgiAIOpgw9EEQBEHQwYShD4Ig\nCIIOJgx9EARBEHQwYeiDIAiCoIMJQx8EQRAEHUwY+iAIgiDoYMLQB9Pk9juH8fntt2GLbbfk7HN/\nO8vrCC311lEnLT846Tg+v+3m7LbXLpVpaBDXZGrqck0ADtl8ZUZ+fzvu/9/tuHDfDZm9z8w3y5UY\nekkLSBqZf16U9J+m7b49nLe0pNFt0rSjJEsa2MMxt0oanB+Pa4eOujBx4kROOOlEzvn1WVx7+VVc\nc/11PP7E47OsjtBSbx1107LN1ttx2qlnVDJ2M3FNpqZO12Sx+ebgoE0Hst6PrmPtH1xN715i1yFL\nz/RxKjH0tl+1vZbttYCzgJ83tm2P/6CvJ6nPTJC1O3BH/j3L8+DoUSy1xBIssfgS9J2tL9tsuTW3\n3Dp0ltURWuqto25a1llrEPPO27+SsZuJazI1dbomAH16iTlm603vXmKO2frwwth3ZvoYtXHdSxok\n6TZJIyTdIGnRpv0PSHoAOKjp+H0kXSXp78AtkuaXdIWkByXdLWmNfNzGTd6C+yXN083YcwMbAvsC\nuzXtn0PSHyU9IulyYI6W807M2u6WNCDv207S/+Wxbm7af5yk32WvwJOSDml6nSMkjc4/h+V9S0t6\nVNJ5kv4l6WJJn5V0p6THJH1qZl377hjz0hgWWWTRydsDFh7AmDFj2jlkrXWElnrrqJuWuhDXZGrq\ndE2eH/sOP7/5YZ446Qs8c/LOvPHu+9z8yAszfZy6GHoBpwM72x4E/A44MT/3e+Bg22t2c946+ZyN\ngeOB+22vAXwXuCAfcyRwUPYefAbo7nZpB+B62/8CXpU0KO8/EHjb9srAscCgpnPmAu7Oum4Hvp73\n3wGsZ3tt4I/At5vOGQh8HvgUcKyk2fJYXwXWBdYDvi5p7Xz88sCp+byBwJdJNyRH5vcYBEEQfEyZ\nb86+bLfGEqz4vctZ6qhLmatvH778qWVm+jh1MfSzA6sBN0kaCXwPWFzSfMB8tm/Px13Yct5Ntl/L\njzdsPG/778ACkuYF7gR+lmfQ89me0M34u5OMMvl3w32/EXBRfs0HgQebzhkPXJMfjwCWzo8XB26Q\nNAr4FrBq0znX2n7P9ivAS8CArPty22/ZHgdcRrohAXjK9ijbk4CHgFtsGxjVNF4XJO0vabik4R8l\nyGTAwgN48cUpd5ZjXhrDgAEDPvTrfdx1hJZ666iblroQ12Rq6nRNNh+4CE+/Oo5Xxr3HhEnmivuf\nYb3lFprp49TF0At4qGmdfnXbn5uB896a3gG2fwzsR3K739kabCdpfmAz4BxJT5OM866SNJ2Xfj8b\nXYCJQCNO4HTgDNurAwcA/ZrOea/pcfM506L5+ElN25Omda7ts20Ptj14/32/3t0hM8Tqq67G0888\nw7PPPcf498dz7fXXsdnGm37o1/u46wgt9dZRNy11Ia7J1NTpmjzz2tusu8yCzDFbbwA2HbgIj77w\n+kwfZ2YEsc0M3gMWkrS+7X9Img1Y0fZDksZK2tD2HcAePbzGsPz8DyRtArxi+w1Jy9keBYySNITk\nAn+06bydgQttH9DYIek20qz6dpK7/O+SVgPWmIH30h/4T378lRk4fhhwnqQfk254dgL2moHz2kqf\nPn34/tHHsN+B+zNx0iS+uONOrLD88rOsjtBSbx110/K9Y49mxMgRjB07lm132pKv7/sNdth2x+I6\n4ppMTZ2uyb1Pv8Jl9/2be47ZhgkTzchnX+OcOx6b6eNoyqS0GiQdB4wDbgZ+STKUfYBf2P5tXsP+\nHWDgRmBr26tJ2gcYbPub+XXmz8ctC7wN7G/7QUmnA5uSZsEPAfvYfq9p/KHAybavb9p3CLAycAQp\nRmBN4BHgk6T1/uGSxtmeOx+/M7Ct7X0k7QD8HPgv8HdgiO1NGu/T9k/zOaPzOU9LOgL4Wh7+HNu/\nkLQ0cI3t1fLx5+XtS1ufmybvTqj2jxsEH1Nef/O96R9UiP7zzF61BCCuSXf0PeySqiVMZvxZe03T\nC125oQ/aSBj6IPhQhFGbmrgmU/NxMfR1WaMPgiAIgqANhKEPgiAIgg4mDH0QBEEQdDBh6IMgCIKg\ngwlDHwRBEAQdTBj6IAiCIOhgwtAHQRAEQQcThj4IgiAIOpgw9EEQBEHQwYShD4IgCIIOJgx9EARB\nEHQwYeiDIAiCoIOJpjZBj0ja3/bZVeuA+mipiw4ILXXWAfXRUhcdEFqq0BEz+mB67F+1gCbqoqUu\nOiC0dEdddEB9tNRFB4SW7mirjjD0QRAEQdDBhKEPgiAIgg4mDH0wPSpfv2qiLlrqogNCS3fURQfU\nR0tddEBo6Y626ohgvCAIgiDoYGJGHwRBEAQdTBj6IAiCIOhgwtAHUyGpr6TV8s9sVeupE5I+IWmN\nqnUEQRDMKLFGH3RB0ibA+cDTgIAlgK/Yvr2ghl/YPkzS1cBUH1Db25fSkvXcCmwP9AFGAC8Bd9o+\noqSOrOUW25tPb18hLcsCpwHrA5OAfwCH236y0Pi1+pxkTQsBXweWJn1eGlq+VljHcsBztt/L/9Nr\nABfYHltSR9YyGDgGWIp0TQTYdiU3zJI2YOq/zwWFNRS9Jn2mf0gwi3Eq8Dnb/wSQtCJwCTCooIYL\n8++fFhyzJ/rbfkPSfqQvy2MlPVhSgKR+wJzAgpI+QfpiAJgX+GRJLU38AfgVsFPe3o30WVm30Ph1\n+5wAXAkMA24GJlao46/AYEnLkyK6ryT9vbauQMvFwLeAUaQbwsqQdCGwHDCSKX8fA0UNPYWvSRj6\noJXZGkYewPa/SrvvbY/ID4cD79ieBCCpNzB7SS2ZPpIWBXYl3YVXwQHAYcBiJK9Cw9C/AZxRkaY5\nbV/YtH2RpG+VGtz2iPyZ2N/2HqXGnQ5z2j6qahHAJNsTJO0EnG77dEn3V6TlZdtXVTR2K4OBVVy9\nK7voNQlDH7QyXNI5wEV5ew+Swa2CW4DPAuPy9hzAjcAGhXWcANxActffm13Wj5UUYPs04DRJh9j+\nZfNzkqq4+QH4m6TvAH8kzYq+BFwnaX4A26+1W4DtiZKWktTX9vh2jzcDXCNpa9vXVazjfUm7A18B\ntsv7qoq3OTZ/p9wCvNfYafuyCrSMBhYBXqhg7GaKXpNYow+6kI3GQcCGedcw4Ne235v2WW3TMtL2\nWtPbNysh6T7b60xvXyEtT/XwtG0vW0jHBcDKwFXAW00CflZi/BYtbwJzkb6832fK2uu8hXWsAnwD\n+IftSyQtA+xq++SSOrKWi4CBwENMcVO7dNxC1jIUWAu4h64GtnTcT9FrEjP6oAvZoP8s/1TNW5LW\nsX0fgKRBwDulReQ4hTOBAbZXy1H329v+YUENi5DW4ueQtDZd1+jnLKWjGdvLVDFuNzyRf3oB81Qp\nxHal4zew/bCko4Al8/ZTQHEjnxlie6WKxm7luKoFZIpek5jRB12QtC3wA6aOBi06I8lahpDcws9n\nHYsAX2pawy+l4zZS4MxvbK+d9422vVpBDV8B9iGtMTYvpbwBnF+FGzSvj2/D1BHMdbhJrIwcLLkC\n0K+xr2TWStawHSlIsa/tZSStBZxQUSbC74FTbD9ceuzukDQAGJI377H9UgUail6TMPRBFyQ9DnwB\nGFWDgBVyIGDjzveftt+vQMO9todIur/J0FeyhCDpi7b/Wnrc7pB0HfAuLZHDto8vrGMo3afXbVZS\nR9ayH3AosDgpsns9kvu8qBZJI4DNgFurujlt0vIIKdL9KZK7vLL0Okm7AqcAt2YdnwG+ZfvSwjqK\nXpNw3QetPAuMromR37tl1zqSiue8Aq/kvGRnXTtTXTDPnZLOBRazvVVei13f9rkVaFm8qlzoFo5s\netwP+CIwoSIth5Jmi3fb3lTSQOCkCnS8b/t1Sc37qkpt27KicbvjGJLb/CWYXPfgZqCooafwNQlD\nH7TybVLk9G10DVapwh07pOlxP2Bz4D7K57weRMpFHijpP6S78KrSuX6ffxppfv8C/gRUYej/Julz\ntm+sYOzJdLOUc6ekeyoRA+/aflcSkma3/aikKtanH5L0ZaC3pBWAQ4C7KtCB7X9L2hBYwfbvs3Gd\nuwotQK8WV/2rVFAhNl+TNUkeBYBhth9o13hh6INWTiSls/UD+lYpxPbBzduS5iOt2VcgxZ+VNBfp\ni+LNHMVcBQva/rOko7OwCZKqKsxyN3C5pF5UG2E+f9NmL1Jxp/4lNTTxXP6cXgHcJOm/wL8r0HEw\n6WbwPVKhnBuAYsGjzUg6lhRbshLpJnU2UvrupyuQc72kG0iFnSCnhJYWIelQUgXFRmzNRZLOtn16\nW8argYc2qBFVrePNCHm9fnTpCN5ppLSNsF2yWmBj3FtJrumbbK8jaT3gZNsbV6DlKWAHKo7nyDpM\nutGYQPK4nGD7jqo0ZV0bk244ri+d49+crVI1kkYCawP3NcULPFjVso+kL9CUPmz78go0PEhacnsr\nb89FiuWINfqgCNfVwR0LoK41zHsBqwB/Ljj+QGBVoH/+cmgwL00R1YU5gpQvvpykO4GFgJ0r0lKL\neI4apfl1qTFPuvFYmpT+WLqYz6k5JfNS4E+2Rxcev5nxti2pEeMyV1VC8thX2r4sL6msJGm2CoJ8\nRdcSyROZkjI78weLGX3QTFPBj/EkdyxUl17XPEudAPzb9nMFx98B2JHU0Ka5XOWbwB9tV7LmKakP\nyQ0qKspEyDrOA5YF/kbF8RyqQaOSrGMkyU29NMklfCWwqu3iNeazod+V5J6el2Twi7vvJR1JSjfc\nAvgR8DXgD+1yU09HywjSuvgngDtIqarjXbiEsqQjSFULG96EHYHzbP+iLeOFoQ+CnpG0vu1/VK2j\nQY2M2rHd7a8gva7bRiW2DympI2u5Ly+pfIsUmHd6c1pmFUhanRRk+yXblcTdSNoC+Bzp5vQG2zdV\npKPx9zkYmMP2TypMlR3ElDiFYbbb1osgXPfBVEjaHtgob95q+5qKdKwHnE4qb9oX6A28VYF34X5J\nB5Hc+M1FUKoo4VmX7lvFDXoP1KVRCdSkxryklUkz+Z2BV0iZGf+vtI4G2bBXYtxbkKT1SVkz++Z9\nvSvSMpKUptsnC1vS9jPtGCgMfdAFST8mpbVdnHcdKunTto+uQM4ZpNanfyF9me8NrFiBjguBR4HP\nkxrc7AE8UoEOqJFRq1Ghmro0KgH4KqnG/Im2n8rZGRdO55x28DtShsrnbD9fwfiTyfEtJwMLk2b0\nlVXbJNU5OBq43PZDSg2qhpYWkT0KxwJjmLI+b6AtwXjhug+6kKNB13LX1rD3V1TFarjtwc0RulW4\nQRtjNnTk6P9httcrqSNr+QtwiO3KjVp2PTaYXKjG9rcLjd8I1pyHGjQqaUWpFO4Sth+saPw5gCXd\n1Ha6Ih2PA9vZrurmuFnLMk51/5v3DbF9b2EdjwPr2n61xHgxow+6Yz6g0WK0qnxkgLcl9QVGSvoJ\nacZWvLgFU4ISx0paDXiRNDspRotRezgXhKnUqNWgUM1Pe3iubRHMPZHTH7cnfbeOAF6SdKftIwrr\nmFzrHlhGFda6B8bUwchnLpW0ve3/wOSA3zOA1QvreBZ4vdRgYeiDVn5EWpMeSvqy3Aj4TkkBkvrY\nngDsRTLs3wQOB5YgzRpLc3aenf0vKfp+7vy4JD0ZtUqoQaGapW2f37oze1wuINUzL01/228o1by/\nwPax2UtWmuOAT5Gvge2RFRZ5Gi7pT6QiQlX3o/8GcEW+EVqH9H1XPCMCeBK4VdK1FMhYCUMfdMGp\nd/WtTCk/e5TtFwvLuAdYJ5eJPD1XyKss8Mv2OfnhbaR0sio03AbJ9Qi8YPvdvD0HMKAKTaQZa2uh\nmn17PGPmcqhSmdmzGztynvTlpBlTFfSRtCgpre2Y6R3cRrqrdV/VOu28wNukqPtmLcUNve17JR0C\n3EhqyPRZ2y+X1gE8k3/6UqACaRj6oAuSGtH2Y/PvFSWt6LJtNpu/naook5lESIuTZo135O0jmFKj\n+w+2H69A1l+ADZq2J+Z9Q7o/vD0olb3d0/adJcdt4bOkkqb9bP9SqYb6dcAttot6oZo4gVRu9o5s\nVJYFHqtAR51q3X+1inGbaSm+BamI0evAuUqNsoouaRRPQY1gvKCZ/A/RoB/J/TeiZCS1mkrOqpvy\nswV1XAJc3EgvlPRPUnObOYGBpYtsZA1T5fxKesD2mhVoqTQ/PGuYl1SwZxipHO9Ztk+rUlMdkDQn\nyaPQmEXfAPyw4QkqrKUfydNTWXpqS/GtqWh4zEqRb0q/zdTXpC3fszGjD7pge7vmbUlLAG2p1tQD\nA/O6pkilXhtrnKX7WK/UUkPgbdunAkgaVkhDKy/nYKKrso4dSHnSVXCLpC8Cl1WR7qcpZYnPBn4G\n3AI829hfxRpwHYxaHu9t4BhJJ+bHVVJ5emoNl74uJtU22JYUN/AVoG1LCDGjD3pEaZHvIdurFBxz\nqZ6et12kG5ikh5vft6T5bb+WHz9ie+USOlo0LUf6kliMdOPzLLB3yWUESSfZ/q6mlEueQFrvLJof\nLen3PTztigoa/YVk1L5Mk1GzfWhhHRsA5wBz215SqSXqAbb/p6SOrKVO6anDgQ2cmwzlrJ47bZde\n+hphe1BL6vC97dIRM/qgC5JOp2sjmbVIPeCLUcqQzwBv5viEfwE0GfmBpHr3xbH9BLCepLnz9rgK\nZGwJfNf2PBWMPZk6rP12w/K2d5G0g+3zJf2BtKxQmp+TZtBXAdh+oCn+pjSVp6c20cdNnQRtj8/G\nvjSNa/KCpG2A54H5ezj+IxGGPmhleNPjCcAlFQdcVcmxwDWSTmTKzc4g4LukClvFkTQ7KcVwaVKE\nNwC2Tygoo3dON+w2V71xQzSLUhujZvvZlqj7idM6ts10l576/Yq01GXp64eS+pPKEp9Oykw4vF2D\nhes+ACYHhyxk++GW/asAL1eUglI5+cu6ETQDqdzqKa6o7aek60nRwiNo+uJuxA4U0vAe8B+6N/S2\nXUkKYh3I+fN/JZUy/T3ZqNk+q7COS0lxC2cA65JuTAfb3q2kjrpRh6WvKghDHwAg6Y/Ar1vT6CR9\nBjjQ9pcr0LQdcK1zOd4AJI22vVrFGiqPtg96RtKCwGmkFESR8sYPqcLb0uqFauwv7IVq1VTl0ldj\nYvV1pr4mbYkrCdd90GD57nLlbQ+TdGYVgkjdt34h6a/A72w/WpGOOnGXpNVtj6paSF3IwV0HMqXj\n4m2kNLv3p33WTNfQY4nbdlU862G8V0iBgMDkuvv/A5xYUkfmSqZ4od6bzrFtJ6+Jrwr0q2jpC9I1\nGQbcTIEllTD0QYOeAquKt9kEsL1nzpPeHThPkknu0EtsVxIMVwM2BPaR9BTpS7N0yiGkmWKdOJP0\nGf113t4r79uvoIbm/58DgN8UHHsyOR32f0mu6ctJHeyOJ3V+vKQKTcDitresaOwuSDqLVAdjU1JW\nws6kSpylmdP2UaUGC9d9AECuufwr29e17N+K5PLbqhplIGkB0pf3YaT82+WBX9o+vSpNVTGt1MMa\nZSoUp7uCQVUVEcpjV7a0odSj4jbgH6TsiC1Jfc8Pd/lS1g1NZwOn18EL1ZTi1/g9N/A3258prOOH\nwF2t37ftImb0QYPDgGsl7UpysUHqfb4+qahDcSRtT+rvvTypScmnbL+Uq349TIpWbef4zamGU2H7\nkHaOP40x/w0gaWGaCrLM4kyUtFxOPSSXna0qwhyqqykPML/t4/LjGyTtAuxRRZyLpFGka9EH+Kqk\nJ6nOC9Xgnfz7bUmLAa8Ci5YaPNeeaPSH+G4ObH2fNtegCEMfAGD7MUmrkwp9NIK9biMV2SheNjPz\nReDnrbEDtt+WVKJ5yvDpH1KWfPNzKsk1+xKwFMnLsWpP53U43wKGZkMi0jWpY459EVpSH18F+ufC\nV6VTHyuZIEyHayTNB5xCSpk1yYVfhKpqT4TrPgg+Rkh6ANgMuDlXG9uU1FymWNe4Ono6cmT3Snnz\nn7aLBn01zV4heaAa6VpFZ6+SngYmUcPUR0mfBHrnzeedWlFXRv7M9LNdrC+8pB77dthuS3GymNEH\ntSXXLD+ZVHBElC+x2trxqgsu3PEq877tVyX1ktTL9lBJpXsRNDwdnwZWIdXsBtiFtKRShKZa960s\nr9SRrGSt+1rMXm0vXbWGBpKOBmZrimj/B6krZl/gfFIv+FJapvVZadSFeKJQVk+j3kU/0tLoA6Tv\ntTVI/1frt2PQMPRBnfkJsJ3tog0wmvhpReP2xNgcQHQ7cLGkl4C3SgqwfT6ApAOBDRszsxzRXLLc\n63Y9PFe03/msHAzZA7sAzUFur2YvVG/SsmAxQ0/Pn5U+wMqS7mq3N8r2pgCSLgPWaQQo5sJcx7Vr\n3DD0QZ0ZU6GRL966cgbZgdRA5nBSnnR/UvOUKvgEqXRnY9137ryvCDWtdR80Ybv5JvS0vG+iUte4\nkjp6/KxI6gWUzApYqTkLwfZoSW1rkhWGPpguko5riuQtyXBJfwKuoKnQRimXrKQ/2961Zf11MlVE\nDbd8cZ5fevwWfgzcn1O6RCpYc1ypwetWpCaYirklzdYoXGT7PJi8Nl5k+W1GsT1J0mcLDvmgpHOA\ni/L2HsCDPRz/kYhgvGC6SNrO9tUVjNtdG1K3q0xkN+MvavuFOuSuN6XlTPUUBeMWphpcWoRUSx3g\n/0rmaks6Nj9cCRhC7tRGctPeY3vPUlrqhqQlu9tv+5mCGk4CFgG+afvtvG8uUv39F20fXUpL3ZDU\nj6mrOZ7ZriDSMPRB8AHINcRf9Sz8j1NV5PC0kHQ7sE2jWqKkeUg9Eoq1ZZ2W16dBae9Pkx6RAr+W\nIWUjFEvDzGvxJ5IqFDZuipcEzgW+V3XUfZ3IPUV2s31QO14/XPcBAJJ+2dPzFaVMLU4qivPpvGsY\ncKjt5wqNvx7JPf0a8APgQmBBoJekvW1fX0LHNLTNSYp4fzrXNS9JT53yTEr/K8kAYHzT9vi8rySN\nqPvGF/WF+fce3Rzbdmyv3rydb87+p7CGicB3JB1PSjkEeNz2Oz2cVhxJi1RRNVDS2qTy3rsCT9HG\n4NGY0QcASBpPasH6Z+B5WvJwG5HWhTXdBPyBKV+ae5KqfG1RaPzhpN7z/YGzga1s3y1pIKnefrEy\np7lQzi9JNx3fA34FjCF1vzqqir9PXZB0DOnL8vK8a0fgz7ZPqkDLVOVvJd1nu0cvSAkkjWq9AQhS\n+W/b2xQaa0WScd8deIWUmnqk7W6XB2fauGHoA5hcT34XUse4CaQP4KW2x1aoaaTttaa3r8T4kh6x\nvXLTc0Xrmedxb6+dAAATPUlEQVRCObuQbjqGAmvYfjKXwr2lqi/wnBa0Ck3leG1fUIGOQaSGPwC3\n276/tIasYyRwkO078/YGpPbPRT6zTTqaAxV7AesAC9j+fEkdQVckTSJ5Jve1/Xje92S7CxmF6z4A\nwParwFnAWdllvhvwsKSjbF/Y89lt41VJezKl69bupJKepWiuD97qbix9hzzJ9r8AJD1l+0mAXPu/\nkrXOHAy3CcnQXwdsBdxB6ktQFNsjJD1LvuGQtGTJwLMm9gV+J6k/ySv2X6BI8GgLzaVWJwDXAn+t\nQEftkfSM7W6DF9vAF0jfrUMlXU/qLthdFcOZSszogy7ktbzdgS1IzW1OtV2s2lmLlqVIa/Trkwzr\nXaROekW+wCVNJBWjETAH8HbjKVLpzGLte/OMfhPS7Ozv+XHjC2KoK+jUlgO+1gTut72mpAHARaWW\nVpp0tNb/XxJ4tGTgWTea+gOULK9aN+oWtDktJD1re4nCY85FqomxOymm5QLgcts3tmW8MPQBgKQT\ngG1IDVL+CFwfUbH1oY41zCXdY/tTkkaQ+nu/CTxie2BhHZXX/2/Rsw2pyVDzckbRokZ5LfhIUgzH\nZM+t7WKBkrm+wrRwSS09UXhG3934nyAvm9revC1jhKEPYPLa0VNMmbU2PhiVtZSUtBDwdab+sqrC\nFRq0IOnXpGDF3YD/B4wDRpauWCdpuO3B2eCvnYufVNKPPpcBnpN043MOsDMpp7/oTUe+FmeRvHKT\nW/baHjHNkzqYHoorCTjG9vwl9ZQm1uiDBstULaAbriQFrtxMtf3Fg26w3UjXOiuvN85ru23VvXqg\n8vr/TWxgew1JD9o+XtKpwN8q0DHB9pkVjDsVORX0CGBJ2/tLWoFUAvaagjJ6ag97WjEVFREz+qC2\nlIywDz4cSq1Hl6Krx+X2QmMvafuZvN75Dil+oVH//+IcYFoUSf9ne11Jd5MCr14FHrK9/HROndk6\njiPFK1xO1/LRJfvRN7T8ieRZ2Nv2atnw3xX/2+WIGX1QZ66RtLXt66oWEkyNpJNJ6ZgPM8XjYtLM\nugRXkDqAvSXpr7a/SPX1/6+RNB9wCnAf6XqcU4GOr+Tf32raZ6CKfvTL2f6SpN0BbL8tqe2R5nWm\ncXOal5lWBAYCf3PuCzDTx4sZfVA3muq6C5iLNCN5n4rrutcFSRsCK9j+fY5jmNv2UxXo+Ccpn78t\n9blnYPzJtQxK1zWYEZSat/SblSPvASTdBWwO3Gl7HUnLkQpOfapiaZWRA1g/Q+r2eCdwLzDedlsq\nKcaMPuiCpM1JbrXKylTa7mk9bZYm564PJjVy+T0wG6kD1qd7Oq9NPJnHr8TQ07WWQS1mLLm++zY0\nBZBKqqSTXl2KGQHHAtcDS0i6mPRZ3acCHXVC2bOxL6mg0k9ysaW2EIY+aGVv4ExJr5EC4W4H7rD9\n39JCJN3Smm7S3b5ZjJ2AtUluYWw/r9TEpQreBkZKuoWu68Cl+iKsKekNcp2D/Biq9fxcDbxL6m0+\naTrHto2aFTO6SdJ9wHqkv82hLtyfoYeoe6CSlsaStD4ppqSRkdG7XYOFoQ+6YPsrAJIWI6UG/YpU\niKTYZ0WpheNcwII5x7Sxnjcv8MlSOmrKeNuWZJi81lcVVzGlNWxxbLfti/EjsHgVqajdsDNTihl9\ntVHMqAohknYC/m772rw9n6QdbV9RUEbdvISHAUeTiuQ8JGlZUmnrthBr9EEXcsnZzwCrk5ou3AEM\ns/2PghoOJf0jLAb8hymG/g3gt7bPKKWlbkg6EliBVLnwR6TyqpfY7rH7YBv19AVWzJv/bFcw0ceF\nHKB4S7sqnH0AHbUoZpS1dNezonYxFVWQ00KxPa6d48SMPmjlF8ATpGIbQ20/XVqA7dOA0yQdbPv0\n0uPXGds/lbQF6aZnJeD7tm+qQoukTUhR7k+TbsaWkPSVUul1NeVu4HJJvag2gHR4jv7/LSm1bRxQ\n7Ga9hV7d7KvE9uQI9zOBATnVbw1ge9s/LKxjddIyyvxpUy+T0g8fast4MaMPWpG0KrARqRvYCqSZ\n2l4VadmAqSvjVRFQVAsknWz7qOntK6RlBPBl2//M2yuSvAuDSmupC5KeItUwH+WafLlKWprqihkh\n6XfAWNIyIMBBwPy296lAy22klMPfNGVsjLa9WmEdd5Eq8g3N25sAJ9neoB3jdXenFczCSJqX1BRk\nKZKB7U9FQUWSLgR+SrrhGJJ/BlehpUZ01zBmq+IqErM1jDyAU3e9Yo1+asqzwOiqjbwSe0r6fvbK\njZVUVTrbwcB4UuvrP5KCFQ+qSMuctu9p2VdFT4+5GkYewPatpLikthCu+6CVO5p+zrD9XIVaBgOr\nVP2lWQckHQj8D7CspOaZ2TykPNwqGC7pHKYEee0BDK9IS114ErhV0t/omolQOqr716Qb9M2AE0hr\n9H8l3SwXI6cbXmN705Lj9sArOY+/Ecy6M/BCBTqelPS/QKMF+J6kz05bCEMfdKERMSxpTttvT+/4\nNjMaWIRq/hHrxh9INdN/BHynaf+bVZQ1zRxImpk10umGkQzMrMxT+adv/qmKdXNxmvsBbP83B04W\nxfZESZMk9a9J4aCDgLOBgZL+Q/pb7VmBjq8BxwOXkW46huV9bSHW6IMu5NzOc0nV1paUtCZwQFMD\nk5JahgJrAffQdXa0fWktdUPSwnQthPJMRToWyuO/XMX4QfdI+j9gA+DebPAXAm6sItJd0pWk2g83\n0dRsqGC9he40zQX0sv1mBWP3JrVULubliBl90MovgM+T86NtPyBpo4q0HFfRuLVF0nbAz0iphy+R\nYikeIfU/L6VBpGpn3yTH+UiaCJzuwn3X60Y2qN9m6n70pXuv/5LU0GZhSSeS8uq/V1hDg8vyT+VI\n+n7LNgAlP7dVeDnC0AdTYfvZlp4TlbSItX1b83au8b47cFv3Z8wS/JBUYexm22tL2pTyrsfDSWVM\nhzRq7OeCH2dKOtz2zwvrqRMXk4LOtgW+QWouU8zbIWkJ28/avjhnRWxOSvHbEViulI5mbFfdaKiZ\n5vbF/Uh/p0cq0DEOGCWpiJcjXPdBFyRdSpoxngGsCxwKDLa9W0V61ga+DOxCWk/76yxeMGe47cGS\nHgDWzt2vHrC9ZkEN9wNbtJYxrdI9XBckjbA9SKkffSPe5V7bRYLgJD0KbNla/0LS10jpXMWNvVL/\n+R8xdd39KjrpdUGp8dANtjcpPO5XutvfrpuimNEHrXwDOI1UavY/wI0UToXJ+di7559XSDMk1Shy\nt0rG5mpatwMXS3qJrrOUEszWXa1y2y9LmtXT6xqVAV+QtA3wPKkoSimOAG6UtI3txwAkfYeUEbFx\nQR3N/J601PNzUpW+r1Kf1O45gcVLD1rayxEz+mAyOUjkArepVeIH0DGJFIW6r+3H874n6zADqJoc\nRPQO6YtyD1Kdg4ttv1pQw3221/mgz80KSNqW9NldAjid1J/heNvFegIodaD8Dcldvx/wKWAbV9CY\nKutpeDlG2V69eV8FWkYxpdNhb2Ah4IRSXsKW8aeiXX0SYkYfTCYHiSwlqa/t8RVK+QKwGzBU0vWk\nIhvq+ZTORtLypLKdjZz5ScD5OW5hPqCYoWdK17hWRJNrdlbE9jX54euk2WsVGm6R9FXgVuAuYDPb\n71ahJfNeLgn8mKRvkjyFc1ekZdumxxOAMbZLFsxpjN/wkjbn0bdt1h0z+qALki4AViZF3TcHiVTR\nT3suUjnR3UmFPy4gdXuqtGFIFUi6Bjja9qiW/auTSmduV42yAEDS6fQ8UyuSSibpzaxDwOykpYSJ\nVNi6V9IQUsDbfMAPSF6on9i+u7SWrOcTJI9Lc1nt+wprmKqpTzu9YTGjD1p5Iv/0ouLWjrbfIhWK\n+UP+59wFOIoUNzCrMaDVyAPYHpVrmQfV0lwR8HjSmnRxbNetHSu2780Px5HW5ytD0g+AfUjfcY0b\nM5MmEoWl6NMND13u6dG2uIWY0QfBxwBJj9leYRrPPW57+dKagu7pbrY2KyKpx7iEKgpfSfonsHrF\nS5NIGgT8juTdgNT052vt8izEjD4AQNLV9Ox6nOWr0VXMcElft/3b5p2S9iO1IQ3qQ8yeEuuTmvxc\nAvwf9YizGU1aQnipShG2R5BiXfrn7bYWzokZfQCApEbqzRdI9eUbjUp2JwWsHF6JsAAASQNIlc7G\nM8WwDybVU9/J9otVaQu6MqtnHjTIWTxbkL5D1gCuJbUxbkvP9RnUNBi4kmTwKyurnf+fTwIWs72V\npFWA9W2f25bxwtAHzTQKskxvX1ANuRJeo3f2Q7b/XqWeINEUBAcpN7vREKqyILg6kQvT7A6cQko3\nrKTolaSHSKmHo2hqv91ahbOAjr+R6gscY3tNSX2A+xvphzObcN0HrcwlaVnbTwJIWoY29kkOPhhO\nPayHTvfAoCh1DIKrA9nAb0My8kszpQZ/Vbxt+5dVDS6pT07nW9D2nyUdDWB7Qu4X0RbC0AetHE7q\np/0kaTayFHBAtZKCIPi4kVN1VwOuI83iR1csCWCYpB+R0oebXfel0uvuAdYB3pK0ANkLJGk9Uu2F\nthCu+2Aq8l34wLz5qO33ejo+CIKglVzhslGLo9nQVJnT3503zKW6CzYyMiStQ6qcuBopXmAhYGfb\nD7Zl3DD0QSs5p3NpuhaUuKAyQUEQBB2ApOdITcMg5c3PTrrxeQ+Y2K7CZOG6D7og6UJSO8uRTGlP\na1JVuiAIgo81udnQqnTtpFeqH31vUvnf1lTDOds5aBj6oJXBwCoOV08QBB2GpLNIRnVT4BxgZ9K6\neSleKHhTMZm6tAoM6sNoUh59EARBp7GB7b2B/9o+nlTUZ8WC41dSNChm9EErCwIPS7qHCgtKBEEQ\ntIF38u+3JS1G6vq4aMHxNy841mTC0AetHFe1gCAIgjZxjaT5SIV77iPFH/2251NmHrZfKzVWMxF1\nH/RI7ne+u+2DpntwEATBx4ScRtyv3XXm60Cs0QdTIWltSadIeprUP/qRiiUFQRB8aCQNkbRI0/be\nwJ+BH0iavzplZQhDHwAgaUVJx0p6lFTI4RmSx2fTqupSB0EQzCR+Q2oIhaSNgB+TUoZfB86uUFcR\nwnUfAJOrWA0D9rX9eN73pO1lq1UWBEHw0ZD0gO018+NfAS/bPi5vj7S9VpX62k3M6IMGXwBeAIZK\n+q2kzalH/+ggCIKPSu/cIQ5S5Htz18eOD0rv+DcYzBi2rwCukDQXsANwGLCwpDOBy23fWKnAIAiC\nD88lwG2SXiGl2A0DkLQ8bWwmUxfCdR9ME0mfAHYBvmS7kvzPIAiCmUHuELcocKPtt/K+FYG5C3av\nq4Qw9EEQBEHQwcQafRAEQRB0MGHogyAIgqCDCUMfBMGHRtJESSMljZb0F0kfut2mpPMk7ZwfnyNp\nlR6O3UTSBh9ijKclLdjN/rkl/UbSE5JGSLpV0rr5uXEfdJwgqBNh6IMg+Ci8Y3st26uRCpJ8o/nJ\nppSmD4Tt/Ww/3MMhmwAf2ND3wDnAa8AKtgcBXyU1eAqCjz1h6IMgmFkMA5bPs+1hkq4idULsnUsq\n3yvpQUkHAChxhqR/SroZWLjxQnlGPTg/3lLSfZIekHSLpKVJNxSHZ2/CZyQtJOmveYx7JX06n7uA\npBslPSTpHLqpDSFpOWBd4Hu2JwHYfsr2tS3HzZ3Hv0/SKEk75P1zSbo26xst6Ut5/48lPZzf809n\n7qUOghkn8uiDIPjI5Jn7VsD1edc6wGq2n5K0P/C67SG5kcidkm4E1gZWAlYBBgAPA79red2FSN3F\nNsqvNb/t1ySdBYyz/dN83B+An9u+Q9KSwA3AysCxwB22T5C0DbBvN/JXBUbanjidt/kusJPtN7L7\n/+58M7Ml8LztbbKW/pIWAHYCBtp27pgWBJUQhj4Igo/CHJJG5sfDgHNJLvV7bD+V938OWKOx/g70\nB1YANgIuyQb2eUnN1coarAfc3nitHtp8fhZYRZo8YZ9X0tx5jC/kc6+V9N8P+T4heQNOyrXSJwGf\nJN2gjAJOlXQycI3tYfnG513gXEnXANd8hHGD4CMRhj4Igo/CO611wrOxfat5F3Cw7Rtajtt6Juro\nBaxn+91utEyPh4A1JfWezqx+D2AhYJDt93N3x362/yVpHWBr4IeSbskehE+Ryq3uDHwT2OwDv6sg\nmAnEGn0QBO3mBuBASbPB5E6JcwG3A1/Ka/iLApt2c+7dwEaSlsnnNlqKvgnM03TcjcDBjQ1JjZuP\n24Ev531bAZ9oHcD2E8Bw4HjlOwNJS2dXfzP9gZeykd8UWCofuxjwtu2LgFOAdbI3ob/t64DDgTWn\nd5GCoF3EjD4IgnZzDrA0cF82pC8DOwKXk2a5D5PaIv+j9UTbL+c1/ssk9QJeArYArgYuzQFxBwOH\nAL+S9CDpe+12UsDe8cAlkh4C7srjdMd+wKnA45LeAV4BvtVyzMXA1ZJGkW4MHs37VwdOUeoA+T5w\nIOkm5EpJ/UgejSNm7FIFwcwnSuAGQRAEQQcTrvsgCIIg6GDC0AdBEARBBxOGPgiCIAg6mDD0QRAE\nQdDBhKEPgiAIgg4mDH0QBEEQdDBh6IMgCIKggwlDHwRBEAQdzP8H8yS7QUUUXXkAAAAASUVORK5C\nYII=\n",
            "text/plain": [
              "<Figure size 504x504 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MyGLAMgdlK8-",
        "colab_type": "text"
      },
      "source": [
        "**On Test Data**"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JpnAvdAWEmjP",
        "colab_type": "code",
        "outputId": "657c1bc5-4ec4-4f24-fa76-9e6813cc0136",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "model.evaluate(X_test, y_test, batch_size=1)"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "22/22 [==============================] - 1s 64ms/step - loss: 0.7912 - accuracy: 0.7727\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.791175127029419, 0.7727272510528564]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "g45NUJ-lDElV",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "y_pred = model.predict(X_test, batch_size=1)\n",
        "y_pred = list(np.argmax(y_pred, axis=1))\n",
        "y_true = list(np.argmax(y_test.values, axis=1))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zXoY5-F_dObh",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "for i in range(len(y_pred)):\n",
        "    y_pred[i] = column_names[y_pred[i]]\n",
        "    y_true[i] = column_names[y_true[i]]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sOZDmUA2DZT9",
        "colab_type": "code",
        "outputId": "bbc6c7db-d702-4941-b091-84166d5e2afe",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 408
        }
      },
      "source": [
        "print(pd.DataFrame({'True Class': y_true, 'Predicted Class': y_pred}))"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "           True Class    Predicted Class\n",
            "0      Tedros Adhanom     Tedros Adhanom\n",
            "1      Tedros Adhanom     Tedros Adhanom\n",
            "2         Fadlo Khuri       Hamad Hassan\n",
            "3         Fadlo Khuri      Marcel Ghanem\n",
            "4        Hamad Hassan       Hamad Hassan\n",
            "5        Hamad Hassan       Hamad Hassan\n",
            "6       Anthony Fauci      Anthony Fauci\n",
            "7       Anthony Fauci      Anthony Fauci\n",
            "8       Marcel Ghanem      Marcel Ghanem\n",
            "9       Marcel Ghanem      Marcel Ghanem\n",
            "10         Bill Gates         Bill Gates\n",
            "11         Bill Gates         Bill Gates\n",
            "12     Cate Blanchett     Cate Blanchett\n",
            "13     Cate Blanchett     Cate Blanchett\n",
            "14       Donald Trump       Donald Trump\n",
            "15       Donald Trump       Donald Trump\n",
            "16  Samuel L. Jackson  Samuel L. Jackson\n",
            "17  Samuel L. Jackson  Samuel L. Jackson\n",
            "18       Keanu Reeves        Fadlo Khuri\n",
            "19       Keanu Reeves       Keanu Reeves\n",
            "20    Andrew M. Cuomo        Fadlo Khuri\n",
            "21    Andrew M. Cuomo        Fadlo Khuri\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "f4IY9zJcbBwJ",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 539
        },
        "outputId": "f0761366-2e48-4777-d478-fa36abd7f44b"
      },
      "source": [
        "df_cm = pd.DataFrame(confusion_matrix(y_true, y_pred, labels=column_names), column_names, column_names)\n",
        "plt.figure(figsize=(7,7))\n",
        "sn.heatmap(df_cm, annot=True, cmap='PuBu', cbar=False, square=True)\n",
        "plt.title('Test Confusion Matrix')\n",
        "plt.xlabel('Predicted Class')\n",
        "plt.ylabel('True Class')\n",
        "plt.show()"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfoAAAIKCAYAAADLb43NAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n/Zgw9EEQBEHQj/n/8cOWIUcPRJwAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 504x504 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    }
  ]
}