tweet-emotion-recognition.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "colab_type": "text",
    "id": "sp7D0ktn5eiG"
   },
   "source": [
    "## Tweet Emotion Recognition: Natural Language Processing with TensorFlow\n",
    "\n",
    "---\n",
    "\n",
    "Dataset: [Tweet Emotion Dataset](https://github.com/dair-ai/emotion_dataset)\n",
    "\n",
    "Created for Coursera guided project [Tweet Emotion Recognition with TensorFlow](https://www.coursera.org/learn/tweet-emotion-tensorflow)\n",
    "\n",
    "---\n",
    "\n",
    "## Task 1: Introduction\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Task 2: Setup and Imports\n",
    "\n",
    "1. Installing Hugging Face's nlp package\n",
    "2. Importing libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 302
    },
    "colab_type": "code",
    "id": "5agZRy-45i0g",
    "outputId": "9bf34b7b-174c-49e4-e93c-e883a24423f7"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2023-01-18 18:17:10.649986: I tensorflow/core/platform/cpu_feature_guard.cc:193] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 FMA\n",
      "To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n",
      "2023-01-18 18:17:12.562475: W tensorflow/compiler/xla/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libnvinfer.so.7'; dlerror: libnvinfer.so.7: cannot open shared object file: No such file or directory\n",
      "2023-01-18 18:17:12.562692: W tensorflow/compiler/xla/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libnvinfer_plugin.so.7'; dlerror: libnvinfer_plugin.so.7: cannot open shared object file: No such file or directory\n",
      "2023-01-18 18:17:12.562704: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Cannot dlopen some TensorRT libraries. If you would like to use Nvidia GPU with TensorRT, please make sure the missing libraries mentioned above are installed properly.\n",
      "/home/gustas/anaconda3/envs/tweet_sentiment/lib/python3.10/site-packages/tqdm/auto.py:22: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
      "  from .autonotebook import tqdm as notebook_tqdm\n"
     ]
    }
   ],
   "source": [
    "import random\n",
    "from typing import Dict, List, Tuple\n",
    "\n",
    "import tensorflow as tf\n",
    "import numpy as np\n",
    "\n",
    "from datasets import load_dataset\n",
    "from datasets.arrow_dataset import Dataset\n",
    "from matplotlib import pyplot as plt\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from tensorflow.keras.preprocessing.text import Tokenizer\n",
    "from tensorflow.keras.preprocessing.sequence import pad_sequences\n",
    "\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "yKFjWz6e5eiH",
    "outputId": "79c40ded-2cbb-4fd1-e2ad-88c1629bf988"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using TensorFlow version 2.11.0\n"
     ]
    }
   ],
   "source": [
    "def show_history(h: Dict[str, np.ndarray]) -> None:\n",
    "    epochs_trained = len(h.history['loss'])\n",
    "    plt.figure(figsize=(16, 6))\n",
    "\n",
    "    plt.subplot(1, 2, 1)\n",
    "    plt.plot(range(0, epochs_trained), h.history.get('accuracy'), label='Training')\n",
    "    plt.plot(range(0, epochs_trained), h.history.get('val_accuracy'), label='Validation')\n",
    "    plt.ylim([0., 1.])\n",
    "    plt.xlabel('Epochs')\n",
    "    plt.ylabel('Accuracy')\n",
    "    plt.legend()\n",
    "\n",
    "    plt.subplot(1, 2, 2)\n",
    "    plt.plot(range(0, epochs_trained), h.history.get('loss'), label='Training')\n",
    "    plt.plot(range(0, epochs_trained), h.history.get('val_loss'), label='Validation')\n",
    "    plt.xlabel('Epochs')\n",
    "    plt.ylabel('Loss')\n",
    "    plt.legend()\n",
    "    plt.show()\n",
    "\n",
    "    \n",
    "def show_confusion_matrix(y_true: np.ndarray, y_pred: np.ndarray, classes: List[str]) -> None:\n",
    "    cm = confusion_matrix(y_true, y_pred, normalize='true')\n",
    "\n",
    "    plt.figure(figsize=(8, 8))\n",
    "    sp = plt.subplot(1, 1, 1)\n",
    "    ctx = sp.matshow(cm)\n",
    "    plt.xticks(list(range(0, 6)), labels=classes)\n",
    "    plt.yticks(list(range(0, 6)), labels=classes)\n",
    "    plt.colorbar(ctx)\n",
    "    plt.show()\n",
    "\n",
    "    \n",
    "print(f'Using TensorFlow version {tf.__version__}')"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Task 3: Importing Data\n",
    "\n",
    "1. Importing the Tweet Emotion dataset\n",
    "2. Creating train, validation and test sets\n",
    "3. Extracting tweets and labels from the examples\n",
    "\n",
    "_HuggingFace is now providing a new version of the Emotion dataset with some adjustments that might differ from the guided project from Coursera._"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "0YHOvjAu5eiL",
    "outputId": "8a7beaba-0fae-4d64-9a70-5949a1a2b3d1"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "No config specified, defaulting to: emotion/split\n",
      "Found cached dataset emotion (/home/gustas/.cache/huggingface/datasets/emotion/split/1.0.0/cca5efe2dfeb58c1d098e0f9eeb200e9927d889b5a03c67097275dfb5fe463bd)\n",
      "100%|██████████| 3/3 [00:00<00:00, 874.72it/s]\n"
     ]
    }
   ],
   "source": [
    "dataset = load_dataset('emotion')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DatasetDict({\n",
       "    train: Dataset({\n",
       "        features: ['text', 'label'],\n",
       "        num_rows: 16000\n",
       "    })\n",
       "    validation: Dataset({\n",
       "        features: ['text', 'label'],\n",
       "        num_rows: 2000\n",
       "    })\n",
       "    test: Dataset({\n",
       "        features: ['text', 'label'],\n",
       "        num_rows: 2000\n",
       "    })\n",
       "})"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "z7eCnxU25eiN"
   },
   "outputs": [],
   "source": [
    "train = dataset['train']\n",
    "val = dataset['validation']\n",
    "test = dataset['test']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "oDYXMfZy5eiP"
   },
   "outputs": [],
   "source": [
    "def get_tweets(data: Dataset) -> Tuple[List[str], List[int]]:\n",
    "    tweets = [x['text'] for x in data]\n",
    "    labels = [x['label'] for x in data]\n",
    "    return tweets, labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "jeq3-vSB5eiR"
   },
   "outputs": [],
   "source": [
    "tweets, labels = get_tweets(train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 50
    },
    "colab_type": "code",
    "id": "bHD3Tk0J5eiU",
    "outputId": "fef2a4a6-885e-4b9b-f0fa-703f25d5fb0d"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "('i didnt feel humiliated', 0)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tweets[0], labels[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Task 4: Tokenizer\n",
    "\n",
    "1. Tokenizing the tweets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "cckUvwBo5eif",
    "outputId": "8d71e463-a7a7-4150-f1b4-c4762abc7b95"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[2, 139, 3, 679]]\n"
     ]
    }
   ],
   "source": [
    "tokenizer = Tokenizer(num_words=10000, oov_token='<UNK>')\n",
    "\n",
    "tokenizer.fit_on_texts(tweets)\n",
    "\n",
    "print(tokenizer.texts_to_sequences([tweets[0]]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Task 5: Padding and Truncating Sequences\n",
    "\n",
    "1. Checking length of the tweets\n",
    "2. Creating padded sequences"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "lengths = [len(t.split(' ')) for t in tweets]\n",
    "\n",
    "plt.hist(lengths, bins=len(set(lengths)))\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "Q9J_Iemf5eiq"
   },
   "outputs": [],
   "source": [
    "def get_sequences(tokenizer: Tokenizer, tweets: List[str]) -> np.ndarray:\n",
    "    sequences = tokenizer.texts_to_sequences(tweets)\n",
    "    padded_sequences = pad_sequences(sequences, truncating='post', maxlen=50, padding='post')\n",
    "    return padded_sequences"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "eglH77ky5ei0"
   },
   "outputs": [],
   "source": [
    "padded_train_sequences = get_sequences(tokenizer, tweets)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 101
    },
    "colab_type": "code",
    "id": "iGR473HA5ei7",
    "outputId": "081725b6-e345-4536-ceb6-9f2c6344054f"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([  2, 139,   3, 679,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n",
       "         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n",
       "         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0,\n",
       "         0,   0,   0,   0,   0,   0,   0,   0,   0,   0,   0], dtype=int32)"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "padded_train_sequences[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Task 6: Preparing the Labels\n",
    "\n",
    "1. Creating classes to index and index to classes dictionaries\n",
    "2. Converting text labels to numeric labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "SufT2bpD5ejE",
    "outputId": "3b0ff532-55b7-45b8-8498-040b86029789"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'joy', 'surprise', 'love', 'sadness', 'anger', 'fear'}\n"
     ]
    }
   ],
   "source": [
    "classes = {'anger', 'fear', 'joy', 'love', 'sadness', 'surprise'}\n",
    "print(classes)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "dNLF6rXL5ejN"
   },
   "outputs": [],
   "source": [
    "classes_to_index = dict((c, i) for i, c in enumerate(classes))\n",
    "index_to_classes = dict((v, k) for k, v in classes_to_index.items())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "_08InVyM5ejc",
    "outputId": "fd4b044c-0e32-4c45-86b3-9b9cac4458d1"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'joy': 0, 'surprise': 1, 'love': 2, 'sadness': 3, 'anger': 4, 'fear': 5}"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "classes_to_index"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{0: 'joy', 1: 'surprise', 2: 'love', 3: 'sadness', 4: 'anger', 5: 'fear'}"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "index_to_classes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 265
    },
    "colab_type": "code",
    "id": "rpwzL88I7YSm",
    "outputId": "df513fa4-49a3-435c-f57d-3ad6715a1583"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.hist(list(map(index_to_classes.get, labels)), bins=11)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Instead of a lambda function proposed in the guided project, we just need to convert from a list to a numpy array\n",
    "train_labels = np.array(labels)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Task 7: Creating the Model\n",
    "\n",
    "1. Creating the model\n",
    "2. Compiling the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 286
    },
    "colab_type": "code",
    "id": "OpewXxPQ5eji",
    "outputId": "b0c61e65-fb96-4f58-aa5b-94dfaadd0b3c"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2023-01-18 18:17:20.738825: I tensorflow/compiler/xla/stream_executor/cuda/cuda_gpu_executor.cc:967] could not open file to read NUMA node: /sys/bus/pci/devices/0000:01:00.0/numa_node\n",
      "Your kernel may have been built without NUMA support.\n",
      "2023-01-18 18:17:20.775636: W tensorflow/compiler/xla/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libcudnn.so.8'; dlerror: libcudnn.so.8: cannot open shared object file: No such file or directory\n",
      "2023-01-18 18:17:20.775664: W tensorflow/core/common_runtime/gpu/gpu_device.cc:1934] Cannot dlopen some GPU libraries. Please make sure the missing libraries mentioned above are installed properly if you would like to use GPU. Follow the guide at https://www.tensorflow.org/install/gpu for how to download and setup the required libraries for your platform.\n",
      "Skipping registering GPU devices...\n",
      "2023-01-18 18:17:20.776322: I tensorflow/core/platform/cpu_feature_guard.cc:193] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 FMA\n",
      "To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " embedding (Embedding)       (None, 50, 16)            160000    \n",
      "                                                                 \n",
      " bidirectional (Bidirectiona  (None, 50, 40)           5920      \n",
      " l)                                                              \n",
      "                                                                 \n",
      " bidirectional_1 (Bidirectio  (None, 40)               9760      \n",
      " nal)                                                            \n",
      "                                                                 \n",
      " dense (Dense)               (None, 6)                 246       \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 175,926\n",
      "Trainable params: 175,926\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model = tf.keras.models.Sequential([\n",
    "    tf.keras.layers.Embedding(10000, 16, input_length=50),\n",
    "    tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(20, return_sequences=True)),\n",
    "    tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(20)),\n",
    "    tf.keras.layers.Dense(6, activation='softmax')\n",
    "])\n",
    "\n",
    "model.compile(\n",
    "    loss='sparse_categorical_crossentropy',\n",
    "    optimizer='adam',\n",
    "    metrics=['accuracy']\n",
    ")\n",
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Task 8: Training the Model\n",
    "\n",
    "1. Preparing a validation set\n",
    "2. Training the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "Ff7F3hCK5ejm"
   },
   "outputs": [],
   "source": [
    "val_tweets, val_labels = get_tweets(val)\n",
    "val_sequences = get_sequences(tokenizer, val_tweets)\n",
    "val_labels = np.array(val_labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 34
    },
    "colab_type": "code",
    "id": "hlMKaZ3H5ejr",
    "outputId": "c37c302d-489c-49a3-cec3-8b27e2aaaffc"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "('im feeling quite sad and sorry for myself but ill snap out of it soon', 0)"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "val_tweets[0], val_labels[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 454
    },
    "colab_type": "code",
    "id": "bzBqnWQ-5ejw",
    "outputId": "499ededc-0e2e-4f06-d59a-15350da7da31"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/20\n",
      "500/500 [==============================] - 26s 44ms/step - loss: 1.3490 - accuracy: 0.4692 - val_loss: 1.0676 - val_accuracy: 0.5905\n",
      "Epoch 2/20\n",
      "500/500 [==============================] - 22s 44ms/step - loss: 0.8551 - accuracy: 0.6636 - val_loss: 0.7281 - val_accuracy: 0.7355\n",
      "Epoch 3/20\n",
      "500/500 [==============================] - 21s 43ms/step - loss: 0.4460 - accuracy: 0.8489 - val_loss: 0.4521 - val_accuracy: 0.8610\n",
      "Epoch 4/20\n",
      "500/500 [==============================] - 23s 45ms/step - loss: 0.2394 - accuracy: 0.9235 - val_loss: 0.4108 - val_accuracy: 0.8670\n",
      "Epoch 5/20\n",
      "500/500 [==============================] - 27s 53ms/step - loss: 0.1754 - accuracy: 0.9418 - val_loss: 0.4590 - val_accuracy: 0.8590\n",
      "Epoch 6/20\n",
      "500/500 [==============================] - 24s 47ms/step - loss: 0.1353 - accuracy: 0.9562 - val_loss: 0.4352 - val_accuracy: 0.8765\n",
      "Epoch 7/20\n",
      "500/500 [==============================] - 23s 45ms/step - loss: 0.1068 - accuracy: 0.9656 - val_loss: 0.3699 - val_accuracy: 0.8940\n",
      "Epoch 8/20\n",
      "500/500 [==============================] - 24s 48ms/step - loss: 0.0910 - accuracy: 0.9697 - val_loss: 0.4254 - val_accuracy: 0.8795\n",
      "Epoch 9/20\n",
      "500/500 [==============================] - 24s 48ms/step - loss: 0.0719 - accuracy: 0.9774 - val_loss: 0.3898 - val_accuracy: 0.8985\n",
      "Epoch 10/20\n",
      "500/500 [==============================] - 22s 44ms/step - loss: 0.0610 - accuracy: 0.9799 - val_loss: 0.4856 - val_accuracy: 0.8915\n",
      "Epoch 11/20\n",
      "500/500 [==============================] - 23s 46ms/step - loss: 0.0527 - accuracy: 0.9829 - val_loss: 0.4567 - val_accuracy: 0.8910\n"
     ]
    }
   ],
   "source": [
    "h = model.fit(\n",
    "    padded_train_sequences, train_labels,\n",
    "    validation_data=(val_sequences, val_labels),\n",
    "    epochs=20,\n",
    "    callbacks=[\n",
    "        tf.keras.callbacks.EarlyStopping(monitor='val_accuracy', patience=2)\n",
    "    ]\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Task 9: Evaluating the Model\n",
    "\n",
    "1. Visualizing training history\n",
    "2. Prepraring a test set\n",
    "3. A look at individual predictions on the test set\n",
    "4. A look at all predictions on the test set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABR8AAAISCAYAAABF+LiMAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjYuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8o6BhiAAAACXBIWXMAAA9hAAAPYQGoP6dpAAC1y0lEQVR4nOzdd3hUZd7G8e+U9EZ6AoQQelNAigI2lI4FUcG1sLZVd1VE1FcRRcWCZVVWXXV3FVEXFcG6ijRRQbDQe29JIIUQ0vvMvH+cJCQSIAkzmUxyf67rXDk5c87z/MLI7nDnKSaHw+FARERERERERERExMnM7i5AREREREREREREmiaFjyIiIiIiIiIiIuISCh9FRERERERERETEJRQ+ioiIiIiIiIiIiEsofBQRERERERERERGXUPgoIiIiIiIiIiIiLqHwUURERERERERERFxC4aOIiIiIiIiIiIi4hMJHERERERERERERcQmFjyIiIiIiIiIiIuISbg0fly9fzuWXX07Lli0xmUx8+eWXp33mp59+ok+fPvj6+tKuXTvefvtt1xcqIiIiIiIiIiIidebW8DE/P5+ePXvyxhtv1Or+/fv3M2rUKC644ALWr1/Po48+ysSJE/nss89cXKmIiIiIiIiIiIjUlcnhcDjcXQSAyWTiiy++YMyYMSe95+GHH+brr79m+/btldfuuusuNm7cyC+//NIAVYqIiIiIiIiIiEhtWd1dQF388ssvDBs2rNq14cOH8+6771JaWoqXl9cJzxQXF1NcXFz5vd1uJzMzk/DwcEwmk8trFhEREXE2h8NBbm4uLVu2xGzWEt6eyG63c/jwYYKCgvSZVERERDxOXT6PelT4mJqaSnR0dLVr0dHRlJWVkZGRQWxs7AnPzJgxg6eeeqqhShQRERFpMElJSbRu3drdZXi05cuX89JLL7F27VpSUlJOOxOnqpUrV3LRRRfRo0cPNmzYUKd+Dx8+TFxcXN0LFhEREWlEavN51KPCR+CE3wxXzBo/2W+Mp0yZwuTJkyu/z87Opk2bNiQlJREcHOy6QkVERERcJCcnh7i4OIKCgtxdiserWIP8lltu4eqrr671c9nZ2UyYMIFLL72UtLS0Ovdb8d7pM6mIiIh4orp8HvWo8DEmJobU1NRq19LT07FarYSHh9f4jI+PDz4+PidcDw4O1gc9ERER8WiarnvmRo4cyciRI+v83J133sn111+PxWLhyy+/rPPzFe+dPpOKiIiIJ6vN51GPWiRowIABLFmypNq1xYsX07dv3xrXexQRERERcbb33nuPvXv38sQTT9T6meLiYnJycqodIiIiIs2BW8PHvLw8NmzYULlGzv79+9mwYQOJiYmAMWV6woQJlfffddddHDx4kMmTJ7N9+3ZmzZrFu+++y4MPPuiO8kVERESkmdm9ezePPPIIc+bMwWqt/SSiGTNmEBISUnlovUcRERFpLtwaPq5Zs4bevXvTu3dvACZPnkzv3r2ZNm0aACkpKZVBJEBCQgILFizgxx9/pFevXjz99NO89tprdVqfR0RERESkPmw2G9dffz1PPfUUnTp1qtOzU6ZMITs7u/JISkpyUZUiIiIijYvJUbFjSzORk5NDSEgI2dnZJ11fx+FwUFZWhs1ma+DqxBUsFgtWq1XrYomISK3Y7A5KyuzGYSs/yqocNhvF1b4/8bx/Qhhnt27hshpr83lG6s5kMp1yt+usrCxCQ0OxWCyV1+x2Ow6HA4vFwuLFi7nkkktq1ZfeQxERaYpsNhulpaXuLkOcxMvLq9rnnqrq8lnGozacaQglJSWkpKRQUFDg7lLEifz9/YmNjcXb29vdpYiISDmHw4HN7qDU5jge4NUy6Cst/1pc4zPHz4urnJfW0McJz9vs2Oxn/nvZKSO7uDR8FPcIDg5m8+bN1a69+eabLFu2jPnz55OQkOCmykRERNwvLy+P5ORkmtkYtybNZDLRunVrAgMDz6gdhY9V2O129u/fj8VioWXLlnh7e2u0nIdzOByUlJRw5MgR9u/fT8eOHTGbPWqfJRGRWnM4jCCv1GYvP6qeV/++pMw4L7MfPy+12SkrDwJrbsMYEVhmt1Na/kxJ+TMV5xX3ldnslNhO0m5Zebt2O57w2dTbasbHYsbbWuWw1HBe5Vr7yDP7gCYNJy8vjz179lR+X7EGeVhYGG3atGHKlCkcOnSIDz74ALPZTI8ePao9HxUVha+v7wnXRUREmhObzUZycjL+/v5ERkYqS2kCHA4HR44cITk5mY4dO550BGRtKHysoqSkBLvdTlxcHP7+/u4uR5zEz88PLy8vDh48SElJCb6+vu4uSUTkBEWlNrILS8kqKCWroMQ4Lywlu6CUrMIS4/ofvi8qtZWP6HMYgaDNA5K80zhZ0OdV5ZrPycK/itfLv/eq4R6fyu8tJ33Ny1K1X5M+PDdxa9asYfDgwZXfT548GYA///nPzJ49+4Q1yEVEROREpaWlOBwOIiMj8fPzc3c54iSRkZEcOHCA0tJShY/OppFxTY/eUxFpCA6Hg/wSG1kFRjhYGSYWVv2+pPJ61deLSu0uqakiQPMqD9W8zFXOLWa8LSasFfdYjMDNWuXcy2LGy2rCaq4I5kyVz1Y9r/qcEd4Z58ZzNT9z4nMK+qThXXzxxaecHjZ79uxTPv/kk0/y5JNPOrcoERERD6XPck2Ls95PhY8iIiJ/YLM7yCkfeZhVUEJWYanxfcHxsDC7oPrr2eVhYtkZrBdoNkELf29a+HkR4u9lfPXzooW/d/nX8sPPm2A/LwJ8LNVCwuqhoQmLWWGeiIiIiIi4l8JHERFpMkptdopKbRSVVnw1zgvLzwtLbeQUVh9xmF1YVm00YlZBCTlFZWdUh7fVTGh5SBhSJUg0wsMqQaLf8fMQfy8Cva2YzQoLRURERESk6VD4KCd18cUX06tXL2bOnFmr+w8cOEBCQgLr16+nV69eLq1NRDyDw+GguMxOcamdojIbhSU2isrKA8GK85quVQkPjWtVw8TqgWLVe89k1GFNAn2sJ4w4rBokhvh5EeLnXe31Fv5e+HrVfz0UERERERHxXMpSTqTwsQk43ZS6igXT6+rzzz/Hy8ur1vfHxcWRkpJCREREnfsSkcahuMxGZn4JR/NKOJpfwtG8YjLzSypHDVYEfYWlNiNQrLz+h8Cw4vsym9t2M/bzsuDrZS7/asGn/PsQv4rw0Ji63KJKuFgZJPp5EeznhZdF68WKiIiIiDRFylIajsLHJiAlJaXyfO7cuUybNo2dO3dWXvvjTlOlpaW1+osQFhZWpzosFgsxMTF1ekZEXMtmd3CsoDxMzCsmo0qgmFF+7Wh+Sfn3xeSe4XTjU7GYTfhazfh5W/CxWvDzNsJA3/JzH2v1sNDP24Kv1YyPl6XKNeN+X6+Kw1x+X/nrXhZ8vIxdi7XWoYiIiIiInIyylIaj8PE0HA4HhaU2t/Tt52Wp1T+eq/5HGhISgslkqrx24MABYmNjmTt3Lm+++Sa//vorb731FldccQX33HMPK1asIDMzk/bt2/Poo4/ypz/9qbKtPw4Vbtu2LXfccQd79uxh3rx5hIaG8thjj3HHHXdU9lV1qPCPP/7I4MGDWbp0KQ8//DDbtm2jV69evPfee3Tu3Lmyn2eeeYbXXnuNwsJCxo8fT0REBAsXLmTDhg1O+FMUaVocDgc5hWVk5BdzNK+EzPzi8hCxhKP5xZWjFStGLh4rKKnzyEOL2UR4gDdhAd5EBPoQGuCNf3kY6FMlHKwIEn29qoSJVa75Wi34epsrQ0GNIhQRERERaR6UpShLqUrh42kUltroNm2RW/reNn04/t7OeYsefvhhXn75Zd577z18fHwoKiqiT58+PPzwwwQHB/Ptt99y00030a5dO84999yTtvPyyy/z9NNP8+ijjzJ//nz++te/cuGFF9KlS5eTPjN16lRefvllIiMjueuuu7j11ltZuXIlAHPmzOHZZ5/lzTffZNCgQXzyySe8/PLLJCQkOOXnFmnsHA4HBSW24+Fh+deMvJLy6c9GoJhRHjQezSup17qGof5ehAf6EF4eKIYFeBMe6E14oA8R5UFjeKAPEYHeBPt6adMTERERERGpN2Up1TX3LEXhYzMxadIkxo4dW+3agw8+WHl+7733snDhQubNm3fKvzCjRo3ib3/7G2D8JXz11Vf58ccfT/kX5tlnn+Wiiy4C4JFHHmH06NEUFRXh6+vL66+/zm233cYtt9wCwLRp01i8eDF5eXn1/llF3K3MZudI+ejDjPIpzkfzSipHK1ab9pxfTFGpvc59BPlYCQ+sHhqGB/hUXosINM7DA3wI9ffCqlGHIiIiIiIidaIsxTkUPp6Gn5eFbdOHu61vZ+nbt2+17202G88//zxz587l0KFDFBcXU1xcTEBAwCnbOfvssyvPK4Ykp6en1/qZ2NhYANLT02nTpg07d+6s/AtYoX///ixbtqxWP5eIOzgcDjLySkjMLCD5WAFJmQUkZRaSdKyAxMwCUrKLsNVxdKKP1VwlMPSuHKUYXiVUrBixGBbgrd2URUTktNJzili0LY3h3aOJCvJ1dzkiItKMKEuprrlnKQofT8NkMjltuK47/fEvwssvv8yrr77KzJkzOeusswgICGDSpEmUlJScsp0/Lq5qMpmw2089aqvqMxXrLlR95o9rMTjctTWuSBW5RaWVgaIRLhaQdKyQpMwCko8Vnnb9EqvZVDkq8cQQ0ZuwivPyr/7etVuXREREpLbu+u9a1iVmAXDTefHuLUZERJoVZSnVNfcsxfP/S5B6WbFiBVdeeSU33ngjYPwHvHv3brp27dqgdXTu3Jnff/+dm266qfLamjVrGrQGaZ6Ky2wcOlZYGSgmHSsgObOQxPLzrILSUz5vMkFMsC9xYf7EhfoTF+ZX/tU4jw7y1bqJIiLiVsO7x7AuMYtFW1IVPoqIiDiBspT6UfjYTHXo0IHPPvuMVatWERoayiuvvEJqamqD/4W59957+ctf/kLfvn0ZOHAgc+fOZdOmTbRr165B65Cmx253kJZbZIxezCyoDBWTy0czpuYUnXYX6FB/r8pwsXWYH20qg0Z/Wrbwxceqqc8iItJ4De8ew4zvdvDrvqNkFZTQwt/b3SWJiIh4NGUp9aPwsZl6/PHH2b9/P8OHD8ff35877riDMWPGkJ2d3aB13HDDDezbt48HH3yQoqIixo0bx80338zvv//eoHWI53E4HGQVlJZPi64yPbp8JOOhY4WU2E49jN3Xy1wtUGwd6kdcmD9tys+DfL1O+byIiEhj1jYigC4xQexIzeX77elc3ae1u0sSERHxaMpS6sfkaMyTwl0gJyeHkJAQsrOzCQ4OrvZaUVER+/fvJyEhAV9fLcrtLkOHDiUmJoYPP/zQaW3qvfVMhSU2Y0OXYwUkHi2oMkW6kOTMAnKLy075vMVsomULX+JCjUCxasAYF+pPRKC31lkUEY90qs8z4hka6j2cuXQXM5fuZmi3aP4zoe/pHxAREakH/Zvb/Ro6S6nLZxmNfBS3Kigo4O2332b48OFYLBY+/vhjli5dypIlS9xdmjSQvOIy9qTnsTstl4NHCyp3jE7KLCQjr/i0z0cG+RBXJVCMCzt+Hhvii9ViboCfQkREpHEa3j2GmUt3s3zXEfKLywjw0cd/ERERT+dpWYo+fYhbmUwmFixYwDPPPENxcTGdO3fms88+Y8iQIe4uTZwsp6iUPel57EnLY1daLrvT89iTnsehrMJTPhfkY6V1mD9t/rChS1yoP61D/fHz1rqLIiIiJ9MlJoj4cH8OHi3gp11HGHVWrLtLEhERkTPkaVmKwkdxKz8/P5YuXeruMsSJsgtL2ZOey660PHan5bE7PZc96XmkZBed9JmIQB86RQeSEBFQueZixSjGED8vTY0WERGpJ5PJxIjuMfxr+T4WbklV+CgiItIEeFqWovBRROolq6DECBjTc9mdZoxi3JWWS3ruyadKRwf70DEqiA5RgXSKDqJjdCAdIgMJDdDumyIiIq4yvIcRPi7bkU5xmQ0fq2YNiIiISMNR+Cgip3Q0r5jd6XnGNOm08hGN6XmnXI8xNsSXjtFBdIwKNI7oQDpEBRHip92jRUREGlqv1i2IDvYhLaeYVXuOMrhLlLtLEhERkWZE4aOI4HA4yMgrqZwivSvt+GjGo/klJ32uVQs/OkZXBIxG2NghKpAgX4WMIiIijYXZbGJ49xg++OUgC7ekKnwUERGRBqXwUaQZcTgcHMktPj5dumIDmPRcsgpKT/pcXJgfHaOCyoNGI2RsHxVIoHbMFBER8QgjysPHJdvTeNZmx2oxu7skERERaSaUHIg0QQ6Hg9ScovINX/LYXb679O60XHKKymp8xmSC+DB/OlSGjMa6jO0iA/D31v9UiIiIeLL+CWG08PciM7+E1QeOMaB9uLtLEhERkWZCiYKIB3M4HBzOLjLCxT+MZswtrjlkNJugbXhA9U1fogJpHxmIr5cWoBcREWmKrBYzQ7tGM29tMou2pip8FBERkQaj8FEAuPjii+nVqxczZ84EoG3btkyaNIlJkyad9BmTycQXX3zBmDFjzqhvZ7XTnDgcDr7ccIhnv91x0o1frGYTbSMCKjd96RAdRKfoQBIiArTLpYiISDM0okcM89Yms3BLKtMu64bZbHJ3SSIiIh5NWUrtKHxsAi6//HIKCwtZunTpCa/98ssvDBw4kLVr13LOOefUus3Vq1cTEBDgzDJ58skn+fLLL9mwYUO16ykpKYSGhjq1r6YsPbeIqV9sYcm2NAC8LCYSIgKqr8kYHUjb8AC8rVrPSURERAyDOkQQ4G0hNaeITYey6RXXwt0liYiIuI2ylIaj8LEJuO222xg7diwHDx4kPj6+2muzZs2iV69edfrLAhAZGenMEk8pJiamwfryZA6Hg/9tSmHaV1vIKijFy2Livks7cseF7RUyioiIyGn5elkY3CWKbzalsHBLqsJHERFp1pSlNBwlFqfjcEBJvnsOh6NWJV522WVERUUxe/bsatcLCgqYO3cuY8aM4U9/+hOtW7fG39+fs846i48//viUbbZt27Zy2DDA7t27ufDCC/H19aVbt24sWbLkhGcefvhhOnXqhL+/P+3atePxxx+ntNTYQXn27Nk89dRTbNy4EZPJhMlkqqzXZDLx5ZdfVrazefNmLrnkEvz8/AgPD+eOO+4gLy+v8vWbb76ZMWPG8Pe//53Y2FjCw8O5++67K/tqio7mFXP3R+uY+PF6sgpK6RYbzNf3nM89l3RU8CgiIiK1NqKH8Q+VhVtScNTys6aIiEidKUsBlKVU0MjH0yktgOdauqfvRw+D9+mH61qtViZMmMDs2bOZNm0aJpOxfs+8efMoKSnh9ttv5+OPP+bhhx8mODiYb7/9lptuuol27dpx7rnnnrZ9u93O2LFjiYiI4NdffyUnJ6fG9QuCgoKYPXs2LVu2ZPPmzfzlL38hKCiI//u//2P8+PFs2bKFhQsXVg5pDgkJOaGNgoICRowYwXnnncfq1atJT0/n9ttv55577qn2Pwg//PADsbGx/PDDD+zZs4fx48fTq1cv/vKXv5z25/E0321O4bEvt3A0vwSr2cTdgztwzyUd8LIodBQREZG6ubhzFN5WMweOFrArLY/OMUHuLklERJoiZSnKUqpQetFE3HrrrRw4cIAff/yx8tqsWbMYO3YsrVq14sEHH6RXr160a9eOe++9l+HDhzNv3rxatb106VK2b9/Ohx9+SK9evbjwwgt57rnnTrjvscceY+DAgbRt25bLL7+cBx54gE8//RQAPz8/AgMDsVqtxMTEEBMTg5+f3wltzJkzh8LCQj744AN69OjBJZdcwhtvvMGHH35IWlpa5X2hoaG88cYbdOnShcsuu4zRo0fz/fff1/FPrXE7ll/CxI/X89c56ziaX0Ln6CC+vHsQ9w/tpOBRRERE6iXQx8qFHSMAWLgl1c3ViIiIuJeylIbJUjTy8XS8/I3U3F1911KXLl0YOHAgs2bNYvDgwezdu5cVK1awePFibDYbzz//PHPnzuXQoUMUFxdTXFxc60VQt2/fTps2bWjdunXltQEDBpxw3/z585k5cyZ79uwhLy+PsrIygoODa/0zVPTVs2fParUNGjQIu93Ozp07iY6OBqB79+5YLMd3bI6NjWXz5s116qsxW7ItjSmfbyYjrxizCf56cXsmXtpRu1SLiIjIGRvePYal29NZuDWV+4Z0dHc5IiLSFClLUZZShcLH0zGZajVctzG47bbbuOeee/jnP//Je++9R3x8PJdeeikvvfQSr776KjNnzuSss84iICCASZMmUVJSUqt2a1oPqGI4coVff/2V6667jqeeeorhw4cTEhLCJ598wssvv1ynn8HhcJzQdk19enl5nfCa3W6vU1+NUXZBKU/9byufrz8EQIeoQF6+tic9tSC8iIiIOMmQrtFYzCa2p+Rw8Gg+8eGe8VlXREQ8iLIUZSlVaO5mEzJu3DgsFgsfffQR77//Prfccgsmk4kVK1Zw5ZVXcuONN9KzZ0/atWvH7t27a91ut27dSExM5PDh47+1+OWXX6rds3LlSuLj45k6dSp9+/alY8eOHDx4sNo93t7e2Gy20/a1YcMG8vPzq7VtNpvp1KlTrWv2RD/sTGfYzJ/4fP0hzCa486J2fHPv+QoeRURExKlCA7w5r10YAIu2auq1iIg0b8pSXE/hYxMSGBjI+PHjefTRRzl8+DA333wzAB06dGDJkiWsWrWK7du3c+edd5KaWvsPmkOGDKFz585MmDCBjRs3smLFCqZOnVrtng4dOpCYmMgnn3zC3r17ee211/jiiy+q3dO2bVv279/Phg0byMjIoLi4+IS+brjhBnx9ffnzn//Mli1b+OGHH7j33nu56aabKocJNzU5RaU8PH8Tt7y3mrScYhIiAph310CmjOyKr5emWYuIiIjzjeheseu1wkcREWnelKW4nsLHJua2227j2LFjDBkyhDZt2gDw+OOPc8455zB8+HAuvvhiYmJiGDNmTK3bNJvNfPHFFxQXF9O/f39uv/12nn322Wr3XHnlldx///3cc8899OrVi1WrVvH4449Xu+fqq69mxIgRDB48mMjIyBq3qPf392fRokVkZmbSr18/rrnmGi699FLeeOONuv9heIAVu48w4tXlzF2ThMkEt52fwIKJF9AnPtTdpYmIiEgTNqw8fFyXmEVaTpGbqxEREXEvZSmuZXLUNAm9CcvJySEkJITs7OwTFvAsKipi//79JCQk4Ovr66YKxRUa23ubV1zGcwu289FviQDEh/vz0jU96Z8Q5ubKRKTRKy2EIzvhyA7j3K8F+IaUH1XOLV6na0k82Kk+z4hnaAzv4dg3V7IuMYunr+zOTQPauqUGERFpGhrbv7nFOU71vtbls4w2nBFpYKv2ZvB/8zeRfKwQgD8PiOfhkV3w99ZfRxGpwlYKR/dA+jZI315+bIPM/UAtfm/oFVAllKxyVAsrQ2oOLn1DwKxlH0SauhE9YliXmMXCrakKH0VERMRllHaINJCCkjJe+G4H7/9iLB7bOtSPF685m4HtI9xcmchJFGVDxm7IOQRBsdAiHgKjjJ3rxHnsNjh2oHrAeGSH8WdvL635Gb8wiOoGPkHG+1T1KMk17inNN47cwzW3cTreQacIK2u6VuVe7yAwa2UXkcZuePcYnluwg1/3ZXIsv4TQAG93lyQiIiJNkMJHkQbw+/5MHpy3kcTMAgBuOLcNU0Z1JdBHfwXFzRwOyDkMGbuqH0d2QV4Niylb/SA0HkLbGmHkH899ghr6J/AcDocR5FYEjOk7yoPGnVBWWPMz3kEQ1RWiuhhhY1RX42tA5MlDYFsZFOdAUdaJwWRhDdcqj/LXSo3/naIk1zhykuvxw5rAN7iGsLLFyUdhevmDyVz+c5lq+PrH1zjFa+VfTeaTtHWSa7V6zowCeGkq4sMD6BobzPaUHJZuT+PavnHuLklERESaICUfIi5UVGrjpUU7mbVyPw4HtAzx5YVrzuaCjpHuLk2am7ISOLbfCLqqBY27oSTv5M8FxkBIa8hLg+xkIyQ7ssM4auIfXiWMbFs9nAxp3XzWIcw7cnwEY9Vp08U5Nd9v9YWITtUDxqiuxp9ZXYMuixX8w4yjPspKysPLKoHkSUPLGq6XFQGO4983aacIRy95HAbe4+4CRU5rRPcYtqfksGhrqsJHERERcQmFjzVoZnvwNAvueE/XHjzGQ/M2si8jH4BxfVvz2GXdCPZtJuGLuEfFVOmMXeVB427I2GmsE+iw1fyMyQJh7YzwK7KT8TWiE0R0NEakVSgrgewkY4pw1kHj67GDx88Lj0HBUeM4tLbmfkJa/SGcrHIeEOF5I8oKs8oDxipTptO3Q0FGzfebrRDeoXrAGNXN+PkbyxqLVm+wRhjvR32UFhnhZbXAMuv0oWVJAeAAh90YJYqj+leHvfycU7x2iudcoqIvTvz7ZS9zUZ8izjWiRwyvLt3F8t0Z5BWXaVaGiIicEeUpTYuz3k99uqjCy8sIhQoKCvDz83NzNeJMBQXGNMKK99iVikptvLp0F/9Zvg+7A6KDfXh+7NkM7hLl8r6lmaicKr3zxKCxpqnSFbwDjUAxorPxNbKzETKGJhiB0+lYvSG8vXHUpCi7ehh57GCVoPIg2IohK9E4auIVYIyUrGnUZGg8eAecvkZXKSkw/ryrBozp241p1DUyGbVXBozlIWN4h9r9WXsyL1/jCGyE/5vnqEdoWe01Th6O/jE4rRrcizRinaIDSYgIYH9GPj/uTOeys1u6uyQREfFAFovxi/SSkhLlKU1ISUkJcPz9rS+Fj1VYLBZatGhBeno6AP7+/pg8bRSOVONwOCgoKCA9PZ0WLVqc8V+Y09mYlMUD8zayJ92Yxjr2nFY8cVl3Qvw12lHqoawEMveVT4+uEjTWZqp05QjGKkFjUKxrRxb6hkDs2cbxR3a7MXX7j6MmK77POWxsjpK+zThqEhB58indwa2M6cZnqqyk5h2mjx3gpKPngltVDxijuhp/7t7+Z16POFflWo8iUsFkMjG8ewxv/7SXhVtSFT6KiEi9WK1W/P39OXLkCF5eXpi1+aDHs9vtHDlyBH9/f6zWM/u3lsLHP4iJiQGoDCClaWjRokXle+sKxWU2Xvt+N2//tA+b3UFEoA8zxp7F0G7RLutTmpCKqdJ/XI+xzlOlO0NEh8Y54spshuBY44gfcOLrZcWQlQRZB2oYNXnA+DPKP2IcyatraN9qrI94wkY4bY2v/mHVQ6fKHaa3VR/NeHTPyafL+oeXh4vdjm8AE9nF2DRFRMSDjehhhI8/7EinqNSGr1cjWQZCREQ8hslkIjY2lv3793Pw4EF3lyNOYjabadOmzRkPzFP4+AcVf2GioqIoLS11dzniBF5eXi4d8bjlUDYPztvIjtRcAK7o2ZKnruhOaEATn1opdVOx03HFyMWqQWNe2smf8w6svgZjXadKewqrjxGcRnSo+fXCYyef0p2VCLaS8usHan7eO9AIIYNbQW6K8edeVlTzvT7BxujFyD/sMB2ojaJEpGk6u1UIsSG+pGQXsXJPBpd21S9PRUSk7ry9venYsWPlVF3xfN7e3k4Zxarw8SQsFovLp+iKZyu12fnnD3t4Y9keyuwOwgO8eWZMD0aeFevu0sSdKqdK7/xD0LjbmFZ8MkGxVdZj7NRwU6U9hV+ocbTsdeJrdrsRKNa41uQB47WSPEjbYhwVrL7Gn/Efd5gObqU/cxFpVsxmY+r17FUHWLglVeGjiIjUm9lsxtfX191lSCOj8FGkHran5PDApxvZlpIDwKizYnj6yh6EB/q4uTJpMIVZVdZgrLIe46mmSputx6dK/zFo9A1u0PKbFLPZ2EU7pBXEDzzx9dKi8o1uDhq7dQdEGSFjY9phWkTEzSrCxyXb0yiz2bFatFaXiIiIOIfCR5E6KLPZefunvfzj+92U2hy08Pfi6St7cNnZsdqcqCmqmCp9pOpmL7WZKh1UPoW4c/X1GEPbNq2p0p7Cy9d4HyI7ubsSEZFGq1/bUMICvMnML+H3/ZkM7BDh7pJERESkiVD4KFJLu9NyeWDeRjYlZwMwtFs0z17Vg6ggDSn3eGUlkLn3eLB4ZNfxKdOnnSrd6fhRETRqqrSIiHgYq8XM0K7RzF2TxKKtqQofRURExGkUPoqchs3u4D8r9vHK4l2U2OwE+1p56srujOnVSqMdPU3lVOmd1UPGYwdqOVW6U/XNXzRVWkREmpARPWLKw8c0nri8O2azPueIiIjImVP4KHIKe4/k8eC8jaxPzAJgcOdInr/6bKKDnTTa0W43RtYV51Y5cqqc59VwLdfYPKM4x3jdy+/4Zhx+oeAfVn4eVvP3Xk18pKbDAdnJx0cuVqzHeGQn5Kef/DnvoCpTpKusxxiWABavhqtfRETETQZ2CCfQx0pqThEbk7Po3SbU3SWJiIhIE6DwUaQGNruD91bu56VFOykusxPkY+Xxy7txbZ/WxmjHsuLyYPAPoWBxLpTknnitIiis6d6G5uVfPaysTWDpF9r41iosKzZ2la62HuNOyNhzmqnSLY/vJF01aAyK0VRpERFp1nysFgZ3ieJ/Gw+zcGuqwkcRERFxCoWP0vzkH4WCjJOOMszOOspvOw4SmZfFm6ZCWoWU0T7YgdfKPFhWfq+txLk1ma3gE1R+BB8/9w6s+bpPEPgEGq+XFkLhMSjMNL4WHPvD9+VfC48ZU4tLC4wj51DdavQOLA8jW9QusPQPA98WYDnD/5kpPHY8XKwMGneWT5W2n/zPs+pU6cjORsgYrqnSIiIipzKiewz/23iYRVtSeWREFy0xIyIiImdM4aM0L8v/Dj88e/LQCggBhgFYyi8UA0dOcrNXQPUwsMaQsCJErOG6T7DxnNXX9aPuHA4jbK0aRlYcldf+GFhmGusk4jCmepfkQXZi3fr1Cal9YFmcc+KmL7WeKl0laAxtq6nSIiIi9XBx50i8rWYOHC1gZ1ouXWL0SzsRERE5MwofpflY9wEse9o4921RLQwsNPuzPs3GgTwz+fgRGhrOpb3aExoaXj0orDoS0TvwzEf1NSSTCXxDjIOE2j9nt0NRVh0Dy2NQZOwKTnG2cWQdrH/tJ0yVrthVWlOlRUREnCnAx8qFHSNZuj2NhVtSFT6KiIjIGfOg5ETkDOxaDP+bZJxf+BBc8hgADoeDOb8l8tyC7RSU2PDzsvDoqC6MPTdeOzxWMJuNEYr+YXV7zlZmBJA1Tf+uKbAsOGZshvPHqdIRnYywV0RExAWWL1/OSy+9xNq1a0lJSeGLL75gzJgxJ73/888/56233mLDhg0UFxfTvXt3nnzySYYPH95wRbvYiB4xleHjpCGd3F2OiIiIeDiFj9L0HVoH8/5srHfY83oYPNW4nFXIw/M38fOeDAD6J4Tx0jVnEx8e4M5qmw6LFQLCjUNERKSRys/Pp2fPntxyyy1cffXVp71/+fLlDB06lOeee44WLVrw3nvvcfnll/Pbb7/Ru3fvBqjY9YZ0jcJiNrEjNZcDGfm0jdBnIxEREak/hY/StGXuh4/GGRustBsMV7yGA/h0dSJPf7OdvOIyfL3M/N/wLtw8sK1GO4qIiDQzI0eOZOTIkbW+f+bMmdW+f+655/jqq6/43//+12TCxxb+3gxoF87PezJYtDWVOy9q7+6SRERExIMpfJSmqyAT5lwD+Ucg5iwY9wGZRQ4mf7qaH3caO8j0iQ/lpWvOpl1koJuLFREREU9kt9vJzc0lLOzUy5MUFxdTXFxc+X1OTo6rSzsjw3vE8POeDBYqfBQREZEzZHZ3ASIuUVoIH42Ho3sgJA6unwe+wTz82SZ+3HkEb6uZR0d14dM7Byh4FBERkXp7+eWXyc/PZ9y4cae8b8aMGYSEhFQecXFxDVRh/QzvFo3JBOsTs0jNLnJ3OSIiIuLBFD5K02O3wWe3Q/Lvxs7ON8yH4Fh+2JHOkm1pWM0mPrtrIHdc2B6LplmLiIhIPX388cc8+eSTzJ07l6ioqFPeO2XKFLKzsyuPpKSkBqqyfqKCfTmnTSgAi7elurkaERER8WQKH6VpcThg4SOw4xuweMOfPoGoLhSV2njyf1sBuPX8BM5qHeLmQkVERMSTzZ07l9tuu41PP/2UIUOGnPZ+Hx8fgoODqx2N3YjuMQAs3KLwUUREROpP4aM0Lateh9//bZyP/TfEDwTgP8v3cfBoAdHBPky8tKMbCxQRERFP9/HHH3PzzTfz0UcfMXr0aHeX4zLDy8PH3/Znkplf4uZqRERExFMpfJSmY/N8WPK4cT78Oeh+FQBJmQX888c9ADw6qiuBPtpnSURERAx5eXls2LCBDRs2ALB//342bNhAYmIiYEyXnjBhQuX9H3/8MRMmTODll1/mvPPOIzU1ldTUVLKzs91Rvku1CfenW2wwNruDpdvT3F2OiIiIeCiFj9I07F8OX9xlnJ/3Nxhwd+VLT3+zjaJSO+cmhHFFz5ZuKlBEREQaozVr1tC7d2969+4NwOTJk+nduzfTpk0DICUlpTKIBPjXv/5FWVkZd999N7GxsZXHfffd55b6XW1ED2P04yJNvRYREZF60hAw8Xxp2+CTG8FeCt3GwLBnK1/6cWc6i7elYTGbeHpMD0wmbTAjIiIix1188cU4HI6Tvj579uxq3//444+uLaiRGdEjhleW7GLF7gzyiss0g0RERETqTCMfxbPlHIY510BxNrQZAFf9C8zGf9bFZTae/NrYZOaWgW3pFB3kzkpFREREPE7HqEDaRQRQYrPzw450d5cjIiIiHkjho3iuomz47zWQcwgiOsF1H4GXb+XL76zYz4GjBUQG+XDfEG0yIyIiIlJXJpOJ4eVTrxdu1dRrERERqTuFj+KZykpg7o2QvhUCo+GG+eAfVvly8rECXl+2G4Cpo7oS5OvlrkpFREREPNqI8l2vf9iRTlGpzc3ViIiIiKdR+Ciex+GAr+42NpnxDoQb5kFofLVbnvlmO0WldvonhHFlL20yIyIiIlJfZ7cOITbEl4ISGz/vznB3OSIiIuJhFD6K5/l+Omz+FMxWGPcBxPas9vJPu46wcGuqscnMldpkRkRERORMmEwmhnfX1GsRERGpH4WP4llWvwM/v2KcX/4adLi02stVN5n584C2dI7RJjMiIiIiZ2pE+bqPS7enUWqzu7kaERER8SQKH8Vz7PgWFjxknA+eCr1vOOGWd1bsZ39GPpFBPkwaqk1mRERERJyhX9swwgO8ySoo5ff9me4uR0RERDyIwkfxDEmrYf5t4LDDOX+GCx864ZZDWYW8sWwPAI+O6kKwNpkRERERcQqL2cTQbtEALNyiqdciIiJSewofpfE7uhc+Hg9lhdBxGIx+BWpYx/HZb7dRWGqjf9swxvRq5YZCRURERJqu4eVTrxdtTcVud7i5GhEREfEUCh+lccs7Av+9GgqOQsvecM17YLGecNuK3UdYsNnYZOapK7trkxkRERERJxvYPpwgHyvpucWsT8pydzkiIiLiIRQ+SuNVkg8fjYNj+6FFPFz/KfgEnnhbmZ0nyjeZmTAgnq6xwQ1dqYiIiEiT52O1cEnXKMAY/SgiIiJSGwofpXGylcH8W+HwOvALgxs/h8CoGm999+f97DuST0SgD/cP7dTAhYqIiIg0HyO6G1OvF25JxeHQ1GsRERE5PbeHj2+++SYJCQn4+vrSp08fVqxYccr758yZQ8+ePfH39yc2NpZbbrmFo0ePNlC10iAcDljwAOxaCFZfuH4uRHSo8dbDWYW8vmw3AFNGapMZEREREVe6qHMkPlYziZkFbE/JdXc5IiIi4gHcGj7OnTuXSZMmMXXqVNavX88FF1zAyJEjSUxMrPH+n3/+mQkTJnDbbbexdetW5s2bx+rVq7n99tsbuHJxqRV/h7WzARNc/S7E9T/prc9+u52CEhv92oYy9hxtMiMiIiLiSv7eVi7qFAnAQk29FhERkVpwa/j4yiuvcNttt3H77bfTtWtXZs6cSVxcHG+99VaN9//666+0bduWiRMnkpCQwPnnn8+dd97JmjVrGrhycZkNH8GyZ4zzUS9B18tOeuvPuzP4dnMKZhM8dUUPbTIjIiIi0gBGVOx6vUXho4iIiJye28LHkpIS1q5dy7Bhw6pdHzZsGKtWrarxmYEDB5KcnMyCBQtwOBykpaUxf/58Ro8efdJ+iouLycnJqXZII7V3GXx9r3E+6D7o/5eT3mpsMrMFgAkD2tKtpTaZEREREWkIl3aJxmo2sTMtl31H8txdjoiIiDRybgsfMzIysNlsREdHV7seHR1NamrNv0UdOHAgc+bMYfz48Xh7exMTE0OLFi14/fXXT9rPjBkzCAkJqTzi4uKc+nOIk6RsgrkTwF4GZ10Llz55yttnrdzP3iP5RAR6a5MZERERkQYU4u/FgPbhACzamubmakRERKSxc/uGM3+cKutwOE46fXbbtm1MnDiRadOmsXbtWhYuXMj+/fu56667Ttr+lClTyM7OrjySkpKcWr84QVYizLkWSnKh7QVw5T/BfPL/NFOyC3nte2OTmUdGdiXET5vMiIiIiDSkiqnXWvdRRERETsfqro4jIiKwWCwnjHJMT08/YTRkhRkzZjBo0CAeeughAM4++2wCAgK44IILeOaZZ4iNjT3hGR8fH3x8fJz/A4hzFB6D/14DeakQ1Q3G/xesp36/KjaZ6RMfytje2mRGREREpKEN7RbNY19uYWNSFinZhcSG+Lm7JBEREWmk3Dby0dvbmz59+rBkyZJq15csWcLAgQNrfKagoADzH0bEWSwWwBgxKR6mtAg+uQEydkJQS7hhHvi1OOUjq/Zk8M0mY5OZ6Vd2x2zWJjMiIiIiDS0qyJe+8aEALNbUaxERETkFt067njx5Mu+88w6zZs1i+/bt3H///SQmJlZOo54yZQoTJkyovP/yyy/n888/56233mLfvn2sXLmSiRMn0r9/f1q2bOmuH0Pqw26HL++CgyvBJxhunA8hrU/5SEmZnWlfbwXgxvPi6d4ypCEqFREREZEaDO9ePvVau16LiIjIKbht2jXA+PHjOXr0KNOnTyclJYUePXqwYMEC4uPjAUhJSSExMbHy/ptvvpnc3FzeeOMNHnjgAVq0aMEll1zCCy+84K4fQepryeOw9QswexlTraO7n/aR2av2syc9j/AAbx4Y2rkBihQRERGRkxnePYZnvt3Ob/uPkplfQliAt7tLEhERkUbI5Ghm85VzcnIICQkhOzub4OBgd5fTPP3yJiyaYpyPfQfOvva0j6TlFHHJ338kv8TGi9eczbi+2rVcRESaL32e8XxN5T0c/doKth7O4cWrz2ZcP30+ExERaS7q8lnG7btdSzOz9UtY9KhxPuTJWgWPYGwyk19io3ebFlxzzqmnZ4uIiIhIwxjRXbtei4iIyKkpfJSGc/AX+PwOwAH9/gKDJtXqsVV7M/h642HMJnj6yh7aZEZERESkkRjRwwgff96dQW5RqZurERERkcZI4aM0jCM74ePrwFYMnUfDyBfAdPoQsdRm54mvjE1mbjg3nh6ttMmMiIiISGPRISqQdpEBlNjs/LDziLvLERERkUZI4aO4Xm4q/PcaKMqC1v3g6nfAbKnVo++vOsDu9DzCArx5cJg2mRERERFpTEwmU+XU60Xa9VpERERqoPBRXKs4F+ZcC9mJENYe/jQXvP1r9WhaThGvLtkFwMMjOhPi7+XKSkVERESkHiqmXv+wM52iUpubqxEREZHGRuGjuI6tFD79M6RugoBIuPEzCAiv9ePPLTA2mekV14Jr+2j3RBEREZHG6KxWIbQM8aWgxMaK3RnuLkdEREQaGYWP4hoOB/xvEuz9Hrz84fq5EJZQ68d/3XeUrzYcxqRNZkREREQaNZPJxPDy0Y8LNfVaRERE/kDho7jGjzNgw3/BZIZrZ0OrPrV+tOomM9f3b8NZrbXJjIiIiEhjVrHu49LtaZTa7G6uRkRERBoThY/ifGvfh59eMM5HvwKdhtfp8fdXHWBnWi6h/l48NFybzIiIiIg0dn3bhhEe4E12YSm/7ct0dzkiIiLSiCh8FOfatRi+ud84v/Ah6HtLnR5Pzyli5tLdADw8ogst/L2dXaGIiIiIOJnFbGJY92gAFm5NcXM1IiIi0pgofBTnObQO5v0ZHDboeT0MnlrnJmZ8t4O84jJ6xrVgXF9tMiMiIiLiKYaXT71etDUNu93h5mpERESksVD4KM6RuR8+GgelBdBuMFzxGpjqtknMb/uO8sX6Q+WbzHTXJjMiIiIiZ8Juh53fga20Qbob2D6CIB8rR3KLWZ90rEH6FBERkcZP4aOcufyj8N+rIf8IxJwF4z4Ai1edmiiz2Xnia2OTmT/1b8PZrVu4oFARERGRZuT9y+Dj62DrFw3SnbfVzKVdowDtei0iIiLHKXyUM1NaaHyozdwLIXFw/TzwDa5zMx/8cpAdqbm08PfioWHaZEZERETkjLUfbHxd+Q9wNMw06BE9jKnXC7em4migPkVERKRxU/go9We3wWe3Q/Lv4BsCN34GwbF1biY9t4hXl+wC4P+GdyE0QJvMiIiIiJyxfreDdyCkbYE9Sxukyws7ReLrZSYps5BtKTkN0qeIiIg0bgofpX4cDlj4COz4Bize8KdPILJ+Ixaf/24HucVlnN06hPH9tMmMiIiIiFP4hUKfm43zn2c2SJf+3lYu6hQJwCJNvRYREREUPkp9rXoNfv+3cT723xA/sF7NrD6QyefrKjaZ6YFFm8yIiIiIOM+Au8HsBQd/hqTVDdJl1anXIiIiIgofpe42z4cl04zz4c9B96vq1UyZzc7jX24B4Lp+cfSMa+GkAkVEREQEgOCWcPZ443zlzAbp8pIu0VjNJnal5bH3SF6D9CkiIiKNl8JHqZv9y+GLu4zz8+42fpteT//9tcomM8O7OKlAEREREalm0ETj645v4cgul3cX4ufFwA4RACzS6EcREZFmT+Gj1F7aVvjkRrCXQrcxMOyZejd1JLeYlxcbH34fGt6ZMG0yIyIiIuIakZ2hy2WAA1b9o0G6HNHdmHqtdR9FRERE4aPUTvYhmHMtFGdDmwFw1b/AXP//fCo2mTmrVQjX9WvjxEJFRERE5ASDJhlfN841Pte52NBu0ZhMsDE5m0NZhS7vT0RERBovhY9yekXZRvCYcwgiOsN1H4GXb72bW3swk8/WJQMw/cru2mRGRERExNXi+kH8IGMGy69vury7yCAf+sWHAbBYU69FRESaNYWPcmp2G8y7BdK3QmA03Dgf/MPq3ZyxycxWAMb3jaN3m1BnVSoiIiIip1Ix+nHtbCg85vLuhlfseq2p1yIiIs2awkc5teUvwd7vweoHN8yDFmc2RXrOb4lsS8khxM+L/xvR2UlFioiIiMhpdRwKUd2hJA9Wv+vy7oZ3jwZg9YFMMvKKXd6fiIiINE4KH+Xk9nwPPz5vnF8+E2J7nlFzGXnF/H3xTgAeHN6Z8ECfMyxQRERERGrNZILzJxnnv70Npa5di7F1qD9ntQrB7oCl29Jc2peIiIg0XgofpWbZh+DzvwAO6HMz9LzujJt84bsd5BaV0aNVMNf31yYzIiIiIg2u+1gIaQP5R2DDHJd3N6Ji6rXWfRQREWm2FD7KiWylMO9mKDgKMWfDiBfOuMm1B48xb62xycxTV/TQJjMiIiIi7mCxwsB7jPNVr4OtzKXdDe9uhI8r92SQU1Tq0r5ERESkcVL4KCda8gQk/w4+ITDugzPa2RrAZncw7astAIzr25o+8dpkRkRERMRtet8IfmFw7ABs/8qlXXWICqRDVCClNgc/7Eh3aV8iIiLSOCl8lOq2fQW//tM4v+otCEs44yY/+u0gWw/nEOxr5eERXc64PRERERE5A94BcO5dxvnPM8HhcGl3I7pr12sREZHmTOGjHHd0L3x5t3E+cCJ0GX3mTeYV89IibTIjIiIi0qj0/wt4+UPqJti7zKVdVaz7+OPOIxSW2Fzal4iIiDQ+Ch/FUFoIn06AklxoMxAuneaUZl9YuIOcojK6twzmhnPjndKmiIiIiJwh/zA458/G+cqZLu2qe8tgWrXwo7DUxvLdR1zal4iIiDQ+Ch/FsOBBSNsCAZFwzSyweJ1xk+sSj/HpGmOTmelXapMZERERaXyWL1/O5ZdfTsuWLTGZTHz55Zenfeann36iT58++Pr60q5dO95++23XF+oKA+4GsxX2L4dD61zWjclkqhz9uEhTr0VERJodhY8C6z6E9f8Fk9kIHoNjz7jJqpvMXNNHm8yIiIhI45Sfn0/Pnj154403anX//v37GTVqFBdccAHr16/n0UcfZeLEiXz22WcurtQFWsTBWdca5y4e/VgRPi7dnkapze7SvkRERKRxsbq7AHGz1M3GqEeAwY9CwoVOafbj3xPZciiHIF8rj4zUJjMiIiLSOI0cOZKRI0fW+v63336bNm3aMHPmTAC6du3KmjVr+Pvf/87VV1/toipdaNB9sPFj2PY1ZOyBiA4u6eacNqFEBHqTkVfCr/uOckHHSJf0IyIiIo2PRj42Z0XZxjqPZUXQYSic/4BTms3MLzm+ycywzkRokxkRERFpIn755ReGDRtW7drw4cNZs2YNpaWlJ32uuLiYnJycakejENUVOo0AHLDqNZd1YzGbGNpNu16LiIg0RwofmyuHA766GzL3QUgcjP03mJ3zn8OLC3eQXVhK19hgbji3jVPaFBEREWkMUlNTiY6OrnYtOjqasrIyMjIyTvrcjBkzCAkJqTzi4uJcXWrtDZpkfN34MeS6LhisXPdxaxo2u8Nl/YiIiEjjovCxufr1Ldj+PzB7wbXvGzseOsGGpCzmrkkC4Okru2O16D8xERERaVpMpuqb6DkcjhqvVzVlyhSys7Mrj6SkJJfWWCfxAyDuXLCVGJ8RXWRAu3CCfK1k5BWzPvGYy/oRERGRxkXJUHOU+Bssedw4HzEDWvdxSrMVm8w4HHD1Oa3p29Y5gaaIiIhIYxETE0NqavXRgenp6VitVsLDw0/6nI+PD8HBwdWORuX8+42va2YZS/O4gLfVzJCuxqhRTb0WERFpPhQ+Njf5GTDvZrCXQY+rod/tTmv6k9WJbErO1iYzIiIi0mQNGDCAJUuWVLu2ePFi+vbti5eXl5uqcoKOwyGyCxTnGAGkiwzvXr7u49bUyhGjIiIi0rQpfGxO7Db47HbIPQzhHeHyf8AppgfVxbEqm8xMHtqJyCBtMiMiIiKNX15eHhs2bGDDhg0A7N+/nw0bNpCYmAgY06UnTJhQef9dd93FwYMHmTx5Mtu3b2fWrFm8++67PPjgg+4o33nMZmPnazCmXpcWuaSbizpF4utlJvlYIVsPN5JNd0RERMSlFD42Jz+9CPt+AC9/GP8h+AQ5rekXF+0kq6CULjFB3HRevNPaFREREXGlNWvW0Lt3b3r37g3A5MmT6d27N9OmTQMgJSWlMogESEhIYMGCBfz444/06tWLp59+mtdee42rr77aLfU7VY9rILgV5KXBpk9c0oWft4WLO0UBsGirpl6LiIg0B1Z3FyANZM/38NMLxvllr0JUV6c1vTEpi09WGx/Kp1/ZQ5vMiIiIiMe4+OKLTzn9d/bs2Sdcu+iii1i3bp0Lq3ITqzcMuAcWTYGVr0Hvm8BscXo3I3rEsHBrKgu3pPLAsM5Ob19EREQaF6VEzUF2sjHdGgf0uQV6Xue0pu1VNpkZ27sV/RO0yYyIiIiIxzpnAvi2gMy9sP1/LulicJcovCwmdqfnsSc9zyV9iIiISOOh8LGpKyuBebdAYSbE9oQRzzu1+blrktiYnE2Qj5VHRmmTGRERERGP5hMI/e8wzlfOBBdsChPi58XA9hGApl6LiIg0Bwofm7qlT0Dy7+ATAte+D16+Tmv6WH4JLy7cAcCkoZ2ICnJe2yIiIiLiJufeCVY/OLwe9i93SRcjehi7Xit8FBERafoUPjZlW7+EX980zq96G8ISnNr8S4t3cqx8k5k/D9AmMyIiIiJNQkAEnHOTcb5ypku6GNotGpMJNiVncyir0CV9iIiISOOg8LGpOroXvrrHOB90H3QZ5dTmNyVn8fHvxiYzT13RXZvMiIiIiDQlA+4BkwX2LoPDG5zefESgD/3aGmuFL9qi0Y8iIiJNmRKjpqikAD6dACW5ED8ILpnm1OaNTWa24nDAmF4tObdduFPbFxERERE3C42HHmON85X/cEkXI7obU68Xauq1iIhIk6bwsSla8BCkbYGASLhmFlisTm1+3tokNiRlEehj5dFRXZ3atoiIiIg0EoPuM75u+xIy9zm9+eHl6z6uPpDJkdxip7cvIiIijYPCx6Zm3Yew4b9gMhvBY1CMU5vPKijh+e/KN5kZ0pGoYG0yIyIiItIkxZwFHYaAww6r3nB6861a+HF26xAcDli6Pc3p7YuIiEjjoPCxKUndDAseNM4HT4WEC53exd/LN5npFB3Inwe2dXr7IiIiItKInH+/8XXDHMhLd3rzwyumXmvdRxERkSZL4WNTUZRtrPNYVgQdh8H5k53exZZD2cz5zdhkZvqVPfDSJjMiIiIiTVv8IGjV1/iM+dvbTm9+RPnU61V7M8guLHV6+yIiIuJ+So+aAocDvrrbWIsnJA6u+heYnfvW2u0OHv9qCw4HXNmrJedpkxkRERGRps9kgvMnGeer34HiXKc23z4ykI5RgZTaHPyww/kjK0VERMT9FD42Bb++Cdv/B2YvuPZ98A9zehfz1yazPjGLAG+LNpkRERERaU46j4bwjsZMm7Wznd58xehHTb0WERFpmhQ+errEX2HJNON8xAxo3cfpXWQXlPL8wopNZjoRrU1mRERERJoPs/n4zte//BPKnLszdcW6jz/uSqewxObUtkVERMT9FD56svwMmHcL2Mugx9XQ73aXdPPykp1k5pfQMSqQmwe1dUkfIiIiItKInT0OgmIhNwU2ferUpru3DKZ1qB9FpXZ+2nXEqW2LiIiI+yl89FR2G3x2O+QehohOcPk/jDV5nGxnai7//fUgAE9d2V2bzIiIiIg0R1YfOO9vxvnKf4Dd7rSmTSYTI8pHPy7aqqnXIiIiTY2SJE/104uw7wfw8odxH4BPkNO7cDgcPPPtNuwOGNE9hoHtI5zeh4iIiIh4iD43g08IHN0NOxc4temKdR+Xbk+jpMx5waaIiIi4n8JHT7RnKfz0gnF+2asQ5ZoNYH7Ymc6K3Rl4W8xMGdXFJX2IiIiIiIfwDYZ+txnnP78KDofTmj6nTSiRQT7kFpXxy76jTmtXRERE3E/ho6fJTobP/gI4oM8t0PM6l3RTarPzzDfbAbjl/LbEhwe4pB8RERER8SDn/RUsPnBoDRxc6bRmzWYTw7pFA9r1WkREpKlR+OhJykpg3s1QmAmxPWHE8y7r6sNfDrIvI5+IQG/uGdzBZf2IiIiIiAcJjILeNxjnP890atMVU6+XbEvFZnfeqEoRERFxL4WPnmTpE5C8GnxDjHUevXxd0s2x/BJmLt0FwAPDOhPk6+WSfkRERETEAw28F0xm2LMEUrc4rdnz2oUT7GslI6+EtQePOa1dERERcS+Fj55i65fw65vG+Zi3IbSty7qauXQXOUVldIkJYlzfOJf1IyIiIiIeKKwddLvSOF/5D6c162UxM0RTr0VERJochY+eIGMPfHWPcT7oPugyymVd7U7L5b+/JQIw7fJuWMwml/UlIiIiIh5q0CTj65bP4NhBpzU7orsx9XrR1lQcTtzQRkRERNxH4WNjV1IAn06AklyIHwSXTHNpd898ux2b3cGwbtEMbB/h0r5ERERExEO17AXtBoPDBr+84bRmL+wUiZ+XhUNZhWw5lOO0dkVERMR9FD42dgsehPStEBAF18wCi9VlXf2wM52fdh3By2Li0VFdXdaPiIiIiDQB508yvq77EPIznNKkr5eFwV0iAVi4NcUpbYqIiIh7KXxszNZ9CBvmGAt6X/MuBMW4rKtSm51nvtkGwC2DEmgbEeCyvkRERESkCUi4CGJ7QVkh/P5vpzU7vHzqtdZ9FBERaRoUPjZWKZuMUY8Ag6dCwoUu7W7OrwfZeySf8ABv7rmkg0v7EhEREZEmwGQ6Pvrx939DcZ5Tmr2kSxTeFjN7j+SzJz3XKW2KiIiI+yh8bIyKso11HsuKoOMwOH+yS7vLKijh1aW7AZg8rBPBvl4u7U9EREREmoiuVxi7Xxceg3UfOKXJIF8vBnUIBzT6UUREpClQ+NjYOBzw5d/g2H4IaQNX/QvMrn2bZi7dTXZhKV1igriuXxuX9iUiIiIiTYjZAgMnGue//BNspU5pdnjlrtdpTmlPRERE3EfhY2Pzyz9hxzdg9oJxs8E/zKXd7UnP5cNfDwIw7bJuWMwml/YnIiIiIk1Mzz8ZmyPmJMPm+U5pcki3aMwm2Hwom+RjBU5pU0RERNxD4WNjkvgrLJlmnI+YAa36uLzLZ7/djs3uYGi3aAZ2iHB5fyIiIiLSxHj5wnl/Nc5X/gPs9jNuMiLQh35tjV/Ca/SjiIiIZ1P42FjkHYF5t4DDBj2uhn63u7zLH3em88POI3hZTDw6qqvL+xMRERGRJqrfbeATDEe2w+5FTmlyRI/yqdda91FERMSjKXxsDOw2+Px2yD0MEZ3g8n8Yuwe6UJnNzjPfbgfg5oFtSYgIcGl/IiIiItKE+YZA31uM859nOqXJinUfVx/M5EhusVPaFBERkYan8LEx+OkF2PcjePnDuA/AJ8jlXX70eyJ70vMIC/Dmnks6urw/EREREWnizvsbWLwh6VdjOaEz1LKFHz1bh+BwwJJtmnotIiLiqRQ+utvupfDTi8b55f+AKNdPf84uKOWVJbsAmDy0EyF+Xi7vU0RERESauKAY6Hmdce6s0Y/lU68XbtXUaxEREU+l8NGdspPh878ADuh7K5w9rkG6/cf3u8kqKKVzdBDX9YtrkD5FREREpBkYeB9ggl3fQdq2M25uRPnU61V7MsguLD3j9kRERKThKXx0l7ISmHczFGZCbE8YPqNBut17JI8PfjkAwGOXdcVq0X8CIiIiIuIkER2g6+XG+arXzri5dpGBdIoOpMzuYNkOTb0WERHxREqe3GXJNEhebSzOPe4D8PJtkG6f+3Y7ZXYHQ7pGcUHHyAbpU0RERESakfMnGV83z4OspDNurmL040Ltei0iIuKRFD66w9Yv4Le3jPMxb0No2wbpdvmuI3y/Ix2r2cSjo1y/tqSIiIiINEOt+kDbC8BeBr++ecbNVaz7+NOuIxSUlJ1xeyIiItKwFD42tIzd8NU9xvmgSdBlVIN0W2az88y3xro7fx7YlnaRgQ3Sr4iIiIg0QxWjH9e+DwWZZ9RUt9hg4sL8KCq1s3zXkTOvTURERBqUwseGVFIAn06AkjyIHwSXPN5gXX+8OoldaXmE+nsx8ZKODdaviIiIiDRD7S+FmLOgNB9+/88ZNWUymSqnXn+zKcUZ1YmIiEgDcnv4+Oabb5KQkICvry99+vRhxYoVp7y/uLiYqVOnEh8fj4+PD+3bt2fWrFkNVO0ZcDjg2wcgfRsERME1s8BibZCuswtLeWXxTgAmD+1EiL9Xg/QrIiIiIs2UyWTM8gH4/V/GL+HPwBU9WwGweFsa2QXa9VpERMSTuDV8nDt3LpMmTWLq1KmsX7+eCy64gJEjR5KYmHjSZ8aNG8f333/Pu+++y86dO/n444/p0qVLA1ZdT+s/hI0fgckM17wLQTEN1vXr3+/mWEEpHaMC+VP/Ng3Wr4iIiIg0Y93GQIt4KDgK6/97Rk31aBVM19hgSsrsfLnhkHPqExERkQbh1vDxlVde4bbbbuP222+na9euzJw5k7i4ON56660a71+4cCE//fQTCxYsYMiQIbRt25b+/fszcODABq68jlI2wbcPGueXPAYJFzZY1/uO5DF71QEAHr+sG1aL2we7ioiIiEhzYLHCwHuN81Wvg63+IxZNJhPj+7YGYO7qM99BW0RERBqO25KokpIS1q5dy7Bhw6pdHzZsGKtWrarxma+//pq+ffvy4osv0qpVKzp16sSDDz5IYWHhSfspLi4mJyen2tGgCrOMdR5txdBxOAy6v0G7f27BDsrsDi7pEsWFnSIbtG8RERERaeZ63wj+EZCdCFu/OKOmruzVCm+LmW0pOWw5lO2kAkVERMTV3BY+ZmRkYLPZiI6OrnY9Ojqa1NTUGp/Zt28fP//8M1u2bOGLL75g5syZzJ8/n7vvvvuk/cyYMYOQkJDKIy4uzqk/xyk5HPDV3XBsP4S0gaveBnPD/ZH/vDuDpdvTsJpNPDqqa4P1KyIiIiICgJcfnHeXcb7yH8bn43oKDfBmWHfj3w4a/SgiIuI53D4H12QyVfve4XCccK2C3W7HZDIxZ84c+vfvz6hRo3jllVeYPXv2SUc/Tpkyhezs7MojKakBP6j88k/Y8Q1YvGHcbPAPa7Cuy2x2nv5mGwA3DYinQ1Rgg/UtIiIiIlKp3+3gHQhpW2DP0jNqanw/YyDBlxsOUVRqc0Z1IiIi4mJuCx8jIiKwWCwnjHJMT08/YTRkhdjYWFq1akVISEjlta5du+JwOEhOTq7xGR8fH4KDg6sdDeLgL7BkmnE+/Dlo1adh+i03d00SO9NyaeHvxX2XdmzQvkVEREREKvmFQp+bjfOfZ55RU4PaR9CqhR+5RWUs3FLzbCkRERFpXNwWPnp7e9OnTx+WLFlS7fqSJUtOuoHMoEGDOHz4MHl5eZXXdu3ahdlspnXr1i6tt07yjsD8W8Bhgx7XGL/tbUA5RaW8vHgXAPcP6UQLf+8G7V9EREREpJrz/gZmLzj4MyStrnczZrOJa7XxjIiIiEdx67TryZMn88477zBr1iy2b9/O/fffT2JiInfdZawLM2XKFCZMmFB5//XXX094eDi33HIL27ZtY/ny5Tz00EPceuut+Pn5uevHOFFZEQRGQ0RnuPwfcJJp5K7yxrI9ZOaX0CEqkOvPbdOgfYuIiIiInCCkFZw93jhfOfOMmrq2bxwmE/yy7ygHj+afeW0iIiLiUm4NH8ePH8/MmTOZPn06vXr1Yvny5SxYsID4+HgAUlJSSExMrLw/MDCQJUuWkJWVRd++fbnhhhu4/PLLee2119z1I9SsRRzcughu/Ax8Gnatxf0Z+by3cj8Aj43uipfF7ct6ioiIiIjAoInG1x3fwpFd9W6mVQs/zu8QAcC8NTUvvSQiIiKNh8nhOIMt5zxQTk4OISEhZGdnN9z6jw3oLx+sYcm2NC7uHMnsW/q7uxwRERFxgab+eaY5aLbv4cfXw85vofeNcOU/693MN5sOc89H64kJ9mXlI5dgMTfsTCMREZHmri6fZTQsrglZuSeDJdvSsJhNPDa6q7vLERERERGp7vxJxteNcyHncL2bGdotmlB/L1Jzili+64hzahMRERGXUPjYRNjsDp7+ZhsAN50XT4eoIDdXJCIiIiLyB3H9IX4Q2Evhl/qPfPSxWhjTuxWgjWdEREQaO4WPTcTc1UnsSM0lxM+L+y7t6O5yRERERERqNmiS8XXtbCg8Vu9mxveLA2Dp9jQy8orPvC4RERFxCYWPTUBOUSkvL94JwKQhHQkN8HZzRSIiIiKe5c033yQhIQFfX1/69OnDihUrTnn/nDlz6NmzJ/7+/sTGxnLLLbdw9OjRBqrWw3UcClHdoSQPVr9b72a6xATTs3UIZXYHX6w75MQCRURExJkUPjYB/1y2h6P5JbSLDODG8+LdXY6IiIiIR5k7dy6TJk1i6tSprF+/ngsuuICRI0eSmJhY4/0///wzEyZM4LbbbmPr1q3MmzeP1atXc/vttzdw5R7KZIJB9xnnv70NpYX1bmpc+ejHuWuSaGb7aIqIiHgMhY8e7kBGPrNW7gfg8dHd8LLoLRURERGpi1deeYXbbruN22+/na5duzJz5kzi4uJ46623arz/119/pW3btkycOJGEhATOP/987rzzTtasWdPAlXuwHmMhJA7yj8CGOfVu5vKeLfH1MrMnPY91iVnOq09EREScRkmVh5vx3XZKbQ4u7BTJxZ0j3V2OiIiIiEcpKSlh7dq1DBs2rNr1YcOGsWrVqhqfGThwIMnJySxYsACHw0FaWhrz589n9OjRJ+2nuLiYnJycakezZvGCgfca56teB1tZvZoJ9vVi1FmxAMxbo41nREREGiOFjx5s1d4MFm1Nw2I28djorphMJneXJCIiIuJRMjIysNlsREdHV7seHR1Nampqjc8MHDiQOXPmMH78eLy9vYmJiaFFixa8/vrrJ+1nxowZhISEVB5xcXFO/Tk8Uu8bwS8Mjh2A7V/Vu5nxfY0/y/9tPEx+cf1CTBEREXEdhY8eymZ38PQ32wG44dw2dIoOcnNFIiIiIp7rj7/EdTgcJ/3F7rZt25g4cSLTpk1j7dq1LFy4kP3793PXXXedtP0pU6aQnZ1deSQlaZQe3gFw7p3G+c8zoZ5rNvZPCCMhIoD8Ehvfbk5xXn0iIiLiFAofPdS8NUlsT8kh2NfKpCGd3F2OiIiIiEeKiIjAYrGcMMoxPT39hNGQFWbMmMGgQYN46KGHOPvssxk+fDhvvvkms2bNIiWl5vDLx8eH4ODgaocA/e8AL39I3QR7l9WrCZPJxLV9WwPw6WqFuiIiIo1NncPHtm3bMn369JPu/ieul1tUyt8X7wTgviGdCAvwdnNFIiIiIp7J29ubPn36sGTJkmrXlyxZwsCBA2t8pqCgALO5+sdoi8UCoB2X68o/DM75s3G+cma9m7n6nNaYTbDm4DH2pOc5pzYRERFxijqHjw888ABfffUV7dq1Y+jQoXzyyScUFxe7ojY5iX/+sJeMvBLaRQRw03nx7i5HRERExKNNnjyZd955h1mzZrF9+3buv/9+EhMTK6dRT5kyhQkTJlTef/nll/P555/z1ltvsW/fPlauXMnEiRPp378/LVu2dNeP4bkG3A1mK+xfDofW1auJ6GBfBneOArTxjIiISGNT5/Dx3nvvZe3ataxdu5Zu3boxceJEYmNjueeee1i3rn4fFqT2Eo8WMOvn/QBMHd0Vb6tmzouIiIicifHjxzNz5kymT59Or169WL58OQsWLCA+3vglb0pKSrVZPzfffDOvvPIKb7zxBj169ODaa6+lc+fOfP755+76ETxbizjocY1xfgajH8f1Mzae+WxdMqU2uxMKExEREWcwOc5wbkhpaSlvvvkmDz/8MKWlpfTo0YP77ruPW265pVHuvpyTk0NISAjZ2dkeudbOX/+7lu+2pHJBxwg+uLV/o/wzFhEREdfy9M8zovfwBGnb4K0BgAnuXQvh7evcRKnNzoAZy8jIK+ZfN/VhePcY59cpIiIiQN0+y9R72FxpaSmffvopV1xxBQ888AB9+/blnXfeYdy4cUydOpUbbrihvk3LSfy67yjfbUnFbILHRndT8CgiIiIiTUN0N+g4HHDAyn/Uqwkvi5mrz2kFaOMZERGRxsRa1wfWrVvHe++9x8cff4zFYuGmm27i1VdfpUuXLpX3DBs2jAsvvNCphTZ3NruDp7/ZBsD157ahc0yQmysSEREREXGi8++H3Ytg48cw+FEIqvvIxWv7xvGv5fv4YWc6aTlFRAf7uqBQERERqYs6j3zs168fu3fv5q233iI5OZm///3v1YJHgG7dunHdddc5rUiBz9Yms/VwDkG+Vu4f0snd5YiIiIiIOFf8AIg7F2wl8Otb9WqiQ1QgfeNDsTtg/tpkJxcoIiIi9VHn8HHfvn0sXLiQa6+9Fi8vrxrvCQgI4L333jvj4sSQV1zGi4t2AnDfpR0JD/Rxc0UiIiIiIi4waJLxdc0sKMquVxMVG8/MW5PEGS5vLyIiIk5Q5/AxPT2d33777YTrv/32G2vWrHFKUVLdmz/sISOvmLbh/kwY0Nbd5YiIiIiIuEanERDZBYpzjACyHkafFUuAt4UDRwv4bX+mkwsUERGRuqpz+Hj33XeTlHTiAs6HDh3i7rvvdkpRclxSZgHv/LwfgKmju+FtrfceQSIiIiIijZvZDIPuM85/fQtKi+rcRICPlct7tgS08YyIiEhjUOcka9u2bZxzzjknXO/duzfbtm1zSlFy3PPf7aCkzM6gDuEM6Rrl7nJERERERFyrxzUQ3Ary0mDTJ/VqomLq9YItKeQUlTqzOhEREamjOoePPj4+pKWlnXA9JSUFq7XOm2fLKfy+P5NvN6dgNsFjo7thMpncXZKIiIiIiGtZvWFA+Yyqla+B3VbnJnrHtaBjVCBFpXa+3nDYyQWKiIhIXdQ5fBw6dChTpkwhO/v4AtBZWVk8+uijDB061KnFNWd2u4Pp32wF4Lr+begaG+zmikREREREGsg5fwbfFpC5F7b/r86Pm0wmxpePfvx0jaZei4iIuFOdw8eXX36ZpKQk4uPjGTx4MIMHDyYhIYHU1FRefvllV9TYLH22Lpkth3II8rHywNBO7i5HRERERKTh+ARC/zuM85UzoR67Vl/VuxVeFhObkrPZnpLj3PpERESk1uocPrZq1YpNmzbx4osv0q1bN/r06cM//vEPNm/eTFxcnCtqbHbyi8t4cdFOACZe2pHwQB83VyQiIiIi0sDOvROsfnB4PexfXufHwwN9GNI1GoC52nhGRETEbeq1SGNAQAB33HGHs2uRcm/9uJcjucW0DffnzwPburscEREREZGGFxABvW+E1f8xRj+2u6jOTYzrF8d3W1L5csMhpozqgo/V4vw6RURE5JTqvUPMtm3bSExMpKSkpNr1K6644oyLas6SjxXw7xX7AHh0VFe8rXUenCoiIiIi0jQMvAfWzIK9y+DwBmjZq06PX9gxktgQX1Kyi1iyLY3Lzm7pkjJFRETk5OocPu7bt4+rrrqKzZs3YzKZcJSvv1KxE7PNVvfd6OS457/bQUmZnYHtwxnaLdrd5YiIiIiIuE9oW+h+FWyZDyv/Ade+V6fHLWYT1/RpzevL9jB3dZLCRxERETeo87C6++67j4SEBNLS0vD392fr1q0sX76cvn378uOPP7qgxOZjzYFMvtmUgtkEj1/WrTLQFREREZETJSUlkZycXPn977//zqRJk/j3v//txqrE6c6fZHzd9iVk7qvz49f2Mdal/3lPBsnHCpxXl4iIiNRKncPHX375henTpxMZGYnZbMZsNnP++eczY8YMJk6c6IoamwW73cH0b7YBML5fG7rGBru5IhEREZHG7frrr+eHH34AIDU1laFDh/L777/z6KOPMn36dDdXJ04TcxZ0GAIOO/z8ap0fbxPuz4B24TgcMH9t8ukfEBEREaeqc/hos9kIDAwEICIigsOHDwMQHx/Pzp07nVtdM/LF+kNsSs4m0MfKA8M6ubscERERkUZvy5Yt9O/fH4BPP/2UHj16sGrVKj766CNmz57t3uLEuS58yPi6/r+QuqXOj4/vZ4x+nLcmGbvd4czKRERE5DTqHD726NGDTZs2AXDuuefy4osvsnLlSqZPn067du2cXmBzkF9cxouLdgBw7yUdiAj0cXNFIiIiIo1faWkpPj7G56alS5dWbnzYpUsXUlJS3FmaOFub86DbGGP048JHwFG3AHFEjxiCfK0cyipk5d4M19QoIiIiNapz+PjYY49ht9sBeOaZZzh48CAXXHABCxYs4LXXXnN6gc3Bv37aS1pOMW3C/Ll5UFt3lyMiIiLiEbp3787bb7/NihUrWLJkCSNGjADg8OHDhIeHu7k6cbqh08HiAwdWwI5v6/Sor5eFMb1aATB3dZIrqhMREZGTqHP4OHz4cMaOHQtAu3bt2LZtGxkZGaSnp3PJJZc4vcCm7lBWIf9abiyc/eiorvhYLW6uSERERMQzvPDCC/zrX//i4osv5k9/+hM9e/YE4Ouvv66cji1NSGg8DLzHOF/8GJQV1+nxiqnXi7emcSy/xNnViYiIyEnUKXwsKyvDarWyZUv1dVbCwsK0M3M9vfDdDorL7JzXLozh3aPdXY6IiIiIx7j44ovJyMggIyODWbNmVV6/4447ePvtt91YmbjM+ZMhMAaO7Ydf36rToz1ahdAtNpgSm50vNxxyUYEiIiLyR3UKH61WK/Hx8dhsNlfV06ysPZjJ1xsPYzLB45d1U4ArIiIiUgeFhYUUFxcTGhoKwMGDB5k5cyY7d+4kKirKzdWJS/gEwpAnjPPlf4e89Do9XjH6ce7qJBx1XDdSRERE6qdeaz5OmTKFzMxMV9TTbNjtDqZ/sx2A8X3j6N4yxM0ViYiIiHiWK6+8kg8++ACArKwszj33XF5++WXGjBnDW2/VbVSceJCzr4OWvaEkF5Y9XadHx/RqhbfVzI7UXDYfynZRgSIiIlJVncPH1157jRUrVtCyZUs6d+7MOeecU+2Q2vlq4yE2JmUR6GPlgWGd3V2OiIiIiMdZt24dF1xwAQDz588nOjqagwcP8sEHH2gjxKbMbIYRzxvn6z6ElI21fjTE34sR3WMAbTwjIiLSUKx1fWDMmDEuKKN5KSgp44XvdgJw9+AORAb5uLkiEREREc9TUFBAUFAQAIsXL2bs2LGYzWbOO+88Dh486ObqxKXanAc9roYtn8HCR+Hmb6CWSxiN7xfH1xsP8/WGwzw2uht+3trwUURExJXqHD4+8cQTrqijWfnXT/tIzSkiLsyPWwa1dXc5IiIiIh6pQ4cOfPnll1x11VUsWrSI+++/H4D09HSCg4PdXJ243JCnYMe3cPBn2P41dLuyVo8NaBdOXJgfSZmFfLclhbHntHZxoSIiIs1bnaddy5k5nFXIv5bvBeDRkV3x9dJvWkVERETqY9q0aTz44IO0bduW/v37M2DAAMAYBdm7d283Vycu1yIOBk40zhc/BqVFtXrMbDZxbZ/jG8+IiIiIa9U5fDSbzVgslpMecmovLtxBUamd/glhjOgR4+5yRERERDzWNddcQ2JiImvWrGHRokWV1y+99FJeffVVN1YmDeb8SRDUErIS4dd/1vqxa/q0xmSC3/ZnciAj33X1iYiISN2nXX/xxRfVvi8tLWX9+vW8//77PPXUU04rrClal3iMLzccxmSCaZd1w1TLdWlEREREpGYxMTHExMSQnJyMyWSiVatW9O/f391lSUPxDoAhT8IXd8CKV6DXDRB0+l/wt2zhx4UdI/lp1xE+XZPE/43o4vpaRUREmqk6j3y88sorqx3XXHMNzz77LC+++CJff/21K2psEux2B9P/tw2Aa/u0pkerEDdXJCIiIuLZ7HY706dPJyQkhPj4eNq0aUOLFi14+umnsdvt7i5PGspZ10KrPlCSB98/XevHxvczpl7PX5tMmU3/vYiIiLiK09Z8PPfcc1m6dKmzmmtyvt54mA1JWQR4W3hwWGd3lyMiIiLi8aZOncobb7zB888/z/r161m3bh3PPfccr7/+Oo8//ri7y5OGYjbDiBeM8w1z4PD6Wj02pGs0YQHepOcW89OuIy4sUEREpHlzSvhYWFjI66+/TuvW2imuJgUlZTz/3Q4A/ja4A1HBvm6uSERERMTzvf/++7zzzjv89a9/5eyzz6Znz5787W9/4z//+Q+zZ892d3nSkOL6wVnjAAcsnAIOx2kf8baauap3KwA+XaONZ0RERFylzms+hoaGVlur0OFwkJubi7+/P//973+dWlxT8e/l+0jNKaJVCz9uOz/B3eWIiIiINAmZmZl06XLiWn1dunQhMzPTDRWJWw15Arb/DxJ/ga1fQI+xp31kXN843v15P99vT+dIbjGRQT4NUKiIiEjzUufw8dVXX60WPprNZiIjIzn33HMJDQ11anFNQUp2IW//tBeAR0d1xddLO4KLiIiIOEPPnj154403eO2116pdf+ONNzj77LPdVJW4TUhrY/frH2fAkmnQeSR4+Z3ykc4xQfSMa8HGpCy+WJ/MHRe2b5haRUREmpE6h48333yzC8poul5cuJOiUjv92oYy6qzT77wnIiIiIrXz4osvMnr0aJYuXcqAAQMwmUysWrWKpKQkFixY4O7yxB0GToR1H0B2EvzyBlz40GkfGd83jo1JWcxdncRfLmhXbaCFiIiInLk6r/n43nvvMW/evBOuz5s3j/fff98pRTUV6xOP8cX6Q5hMMO2y7vogIyIiIuJEF110Ebt27eKqq64iKyuLzMxMxo4dy9atW3nvvffcXZ64g7c/DHnKOF/xKuSknPaRy3vG4udlYe+RfNYlHnNxgSIiIs1PncPH559/noiIiBOuR0VF8dxzzzmlqKbA4XAw/ZttAFx9TmvOah3i5opEREREmp6WLVvy7LPP8tlnn/H555/zzDPPcOzYMf1SvDk76xpo3R9K8+H7p057e5CvF6POigVg7mptPCMiIuJsdQ4fDx48SELCiZumxMfHk5iY6JSimoKvNx5mfWIW/t4WHhre2d3liIiIiIg0DyYTjHjeON/4MRxae9pHxveLA+CbTSnkFZe5sjoREZFmp87hY1RUFJs2bTrh+saNGwkPD3dKUZ6usMTGC9/tAOBvF7cnOtjXzRWJiIiIiDQjrfvA2dcZ5wungMNxytv7tQ2lXUQABSU2vt10uAEKFBERaT7qHD5ed911TJw4kR9++AGbzYbNZmPZsmXcd999XHfdda6o0eMczi7E19tCqxZ+3H5BO3eXIyIiIiLS/Ax5Arz8Iek32PLZKW81mUxc29cY/aip1yIiIs5V592un3nmGQ4ePMill16K1Wo8brfbmTBhgtZ8LNc+MpBFky4kKbMAXy+Lu8sRERERaVLGjh17ytezsrIaphBp3IJbwvmT4YdnYMkT0HmUsSHNSVzdpxV/X7yTdYlZ7EnPpUNUUAMWKyIi0nTVeeSjt7c3c+fOZefOncyZM4fPP/+cvXv3MmvWLLy9vV1Ro0fysphpFxno7jJEREREmpyQkJBTHvHx8UyYMMHdZUpjMPAeCImDnGRY9fopb40K8mVw5yhAox9FREScyeRwnGYBlCYmJyeHkJAQsrOzCQ4Odnc5IiIiInWmzzOeT+9hA9ryGcy/1ZiCfc8aCGl10luXbEvjLx+sITzAm1+mXIq3tc5jNURERJqFunyWqfP/m15zzTU8//zzJ1x/6aWXuPbaa+vanIiIiIiIiOt0Hwtx50FpAXz/1ClvHdw5ksggH47ml7BsR1oDFSgiItK01Tl8/Omnnxg9evQJ10eMGMHy5cudUpSIiIiIiIhTmEwwYoZxvmkuJK0+6a1Wi5mrz2kNaOq1iIiIs9Q5fMzLy6txbUcvLy9ycnKcUpSIiIiIiIjTtDoHet1gnC98BOz2k946rq8RPv606wip2UUNUZ2IiEiTVufwsUePHsydO/eE65988gndunVzSlEiIiIiIiJOdek08A6EQ2tgy/yT3tYuMpD+bcOwO2D+Wo1+FBEROVPWuj7w+OOPc/XVV7N3714uueQSAL7//ns++ugj5s8/+f+Ji4iIiIiIuE1QDFwwGb6fDkuegC6jwTugxlvH9Yvj9wOZfLommb9d3AGz2dTAxYqIiDQddR75eMUVV/Dll1+yZ88e/va3v/HAAw9w6NAhli1bRtu2bV1QooiIiIiIiBOcdze0aAO5h2HlP05626izYgj0sZKYWcCv+482YIEiIiJNT53DR4DRo0ezcuVK8vPz2bNnD2PHjmXSpEn06dPH2fWJiIiIiIg4h5cvDH3aOF/5D8iqeVq1v7eVy3u2BOBTbTwjIiJyRuoVPgIsW7aMG2+8kZYtW/LGG28watQo1qxZ48zaREREREREnKvblRA/CMqKYOmTJ71tfL84AL7bkkp2YWkDFSciItL01Cl8TE5O5plnnqFdu3b86U9/IjQ0lNLSUj777DOeeeYZevfu7ao6RUREREREzpzJBCNmACZj45nE32q8rWfrEDpHB1FcZufrjYcbtkYREZEmpNbh46hRo+jWrRvbtm3j9ddf5/Dhw7z++uuurE1ERERERMT5YntC7xuN84UPg91+wi0mk4lr+7YGNPVaRETkTNQ6fFy8eDG33347Tz31FKNHj8ZisbiyLhEREREREde5dBp4B8Hh9bBpbo23jD2nNV4WE5sPZbPtcE4DFygiItI01Dp8XLFiBbm5ufTt25dzzz2XN954gyNHjriyNhERERGRBvHmm2+SkJCAr68vffr0YcWKFae8v7i4mKlTpxIfH4+Pjw/t27dn1qxZDVStOEVgFFz4gHG+9EkozjvhlrAAb4Z2iwbg0zUa/SgiIlIftQ4fBwwYwH/+8x9SUlK48847+eSTT2jVqhV2u50lS5aQm5vryjpFRERERFxi7ty5TJo0ialTp7J+/XouuOACRo4cSWJi4kmfGTduHN9//z3vvvsuO3fu5OOPP6ZLly4NWLU4xXl/g9C2kJcKP79a4y3j+hobz3yx/hBFpbYGLE5ERKRpMDkcDkd9H965cyfvvvsuH374IVlZWQwdOpSvv/7amfU5XU5ODiEhIWRnZxMcHOzuckRERETqTJ9nnOvcc8/lnHPO4a233qq81rVrV8aMGcOMGTNOuH/hwoVcd9117Nu3j7CwsFr1UVxcTHFxceX3OTk5xMXF6T1sDLb/D+beCBYfuHcNtGhT7WWb3cEFLyzjcHYRr/2pN1f0bOmmQkVERBqPunwerdNu13/UuXNnXnzxRZKTk/n444/PpCkRERERkQZXUlLC2rVrGTZsWLXrw4YNY9WqVTU+8/XXX9O3b19efPFFWrVqRadOnXjwwQcpLCw8aT8zZswgJCSk8oiLi3PqzyFnoMtl0PYCsBXDkmknvGwxm7imjzaeERERqa8zCh8rWCwWxowZ0+hHPYqIiIiIVJWRkYHNZiM6Orra9ejoaFJTU2t8Zt++ffz8889s2bKFL774gpkzZzJ//nzuvvvuk/YzZcoUsrOzK4+kJIVYjYbJBCNmgMkMW7+AgyeGzteWT73+eU8GSZkFDV2hiIiIR3NK+CgiIiIi4slMJlO17x0OxwnXKtjtdkwmE3PmzKF///6MGjWKV155hdmzZ5909KOPjw/BwcHVDmlEYs6CcyYY5wsfAbu92stxYf4M6hAOwLy1yQ1dnYiIiEdT+CgiIiIizVZERAQWi+WEUY7p6eknjIasEBsbS6tWrQgJCam81rVrVxwOB8nJCqY81uDHwCcYUjbCxo9OeLli45n5a5Kw2eu9bL6IiEizo/BRRERERJotb29v+vTpw5IlS6pdX7JkCQMHDqzxmUGDBnH48GHy8vIqr+3atQuz2Uzr1q1dWq+4UGAkXPiQcf79dCjOrfby8O4xhPh5cTi7iJ/3ZLihQBEREc+k8FFEREREmrXJkyfzzjvvMGvWLLZv3879999PYmIid911F2Cs1zhhwoTK+6+//nrCw8O55ZZb2LZtG8uXL+ehhx7i1ltvxc/Pz10/hjjDuXdBWDvIS4MVr1R7ydfLwphexk7X2nhGRESk9hQ+ioiIiEizNn78eGbOnMn06dPp1asXy5cvZ8GCBcTHxwOQkpJCYmJi5f2BgYEsWbKErKws+vbtyw033MDll1/Oa6+95q4fQZzF6g3DnjHOf/knHDtQ7eVx/Yyp14u3pZKZX9LAxYmIiHgmk8PhaFYLluTk5BASEkJ2drYW+hYRERGPpM8znk/vYSPmcMAHV8L+n6DrFTD+w2ovX/b6CrYcyuHxy7px2/kJbipSRETEveryWUYjH0VERERERCqYTDBiBpjMsP1rOPBztZfHl2888+nqJJrZOA4REZF6UfgoIiIiIiJSVXR36HOLcb7wEbDbKl+6olcrfKxmdqblsjE5200FioiIeA63h49vvvkmCQkJ+Pr60qdPH1asWFGr51auXInVaqVXr16uLVBERERERJqfwY+CTwikbob1/628HOLnxcgeMQDM1cYzIiIip+XW8HHu3LlMmjSJqVOnsn79ei644AJGjhxZbUHvmmRnZzNhwgQuvfTSBqpURERERESalYAIuPhh43zZ01CUU/lSxcYz/9t4mIKSMndUJyIi4jHcGj6+8sor3Hbbbdx+++107dqVmTNnEhcXx1tvvXXK5+68806uv/56BgwY0ECVioiIiIhIs9PvLxDeAfKPwIq/V14+LyGcNmH+5BWXsWBzqhsLFBERafzcFj6WlJSwdu1ahg0bVu36sGHDWLVq1Umfe++999i7dy9PPPFErfopLi4mJyen2iEiIiIiInJaVm8Y9qxx/utbkLkPALPZxLV9WgPw6RpNvRYRETkVt4WPGRkZ2Gw2oqOjq12Pjo4mNbXm3x7u3r2bRx55hDlz5mC1WmvVz4wZMwgJCak84uLizrh2ERERERFpJjoNh/aXgK0EFj9eefmavq0xm+D3/Znsz8h3Y4EiIiKNm9s3nDGZTNW+dzgcJ1wDsNlsXH/99Tz11FN06tSp1u1PmTKF7OzsyiMpSb+ZFBERERGRWjKZYPhzYLLAjm9g308AxIb4cWGnSECjH0VERE7FbeFjREQEFovlhFGO6enpJ4yGBMjNzWXNmjXcc889WK1WrFYr06dPZ+PGjVitVpYtW1ZjPz4+PgQHB1c7REREREREai2qK/S91Thf9CjYbQCM72vMqvpsbTJlNru7qhMREWnU3BY+ent706dPH5YsWVLt+pIlSxg4cOAJ9wcHB7N582Y2bNhQedx111107tyZDRs2cO655zZU6SIiIiIi0twMfhR8W0DaFlj3PgCXdo0mPMCb9Nxiftx5xL31iYiINFJunXY9efJk3nnnHWbNmsX27du5//77SUxM5K677gKMKdMTJkwwCjWb6dGjR7UjKioKX19fevToQUBAgDt/FBERERERacr8w+DiKcb5smegKBtvq5mrercCYK6mXouIiNTIreHj+PHjmTlzJtOnT6dXr14sX76cBQsWEB8fD0BKSgqJiYnuLFFERERERMTQ7zaI6AQFR+GnFwEY38+Yer1sRzrpuUXurE5ERKRRMjkcDoe7i2hIOTk5hISEkJ2drfUfRURExCPp84zn03vowXYvgTnXgNkL7v4Nwttz1ZsrWZ+YxSMju3DXRe3dXaGIiIjL1eWzjNt3uxYREREREfEYHYdCh6FgL4XFjwHHN575dHUSzWxsh4iIyGkpfBQREREREamL4c+CyQI7F8DeH7isZ0v8vS3sy8hnzcFj7q5ORESkUVH4KCIiIiIiUheRnaH/X4zzhVMItMLos2IBmLtaG8+IiIhUpfBRRERERESkri56GPxC4ch2WDe7cuOZbzelkFtU6ubiREREGg+FjyIiIiIiInXlHwYXP2qcL3uWPlEm2kUGUFhq45tNKe6tTUREpBFR+CgiIiIiIlIffW+FyC5QmIlp+YuVG89o6rWIiMhxCh9FRERERETqw2KF4c8Z57//m2vbFmE1m9iQlMWutFz31iYiItJIKHwUERERERGprw6XQsfhYC8j7OenuKRLFKDRjyIiIhUUPoqIiIiIiJyJ4c+C2Qq7F/HX1gcA+GL9IUrK7O6tS0REpBFQ+CgiIiIiInImIjpC/zsB6LXtRWIDLWTml7B0e5qbCxMREXE/hY8iIiIiIiJn6qL/A78wTBk7md56NaCp1yIiIqDwUURERERE5Mz5tYBLpgJwScq7hJDH8t1HOJxV6N66RERE3Ezho4iIiIiIiDOcczNEdcNSfIznw77B4YDP1ia7uyoRERG3UvgoIiIiIiLiDBYrjJgBwPDCb2lvOsSna5Ow2x1uLkxERMR9FD6KiIiIiIg4S7uLofMozA4bT3nPISmzkF/3HXV3VSIiIm6j8FFERERERMSZhj0DZi/ON23gYvN65q7RxjMiIo1axm744Tn47d9wcBUUZbu7oibF6u4CREREREREmpTw9nDeXbDqdR63/pfLt/Qku6AHIf5e7q5MRESqKimAFX+Hla+BvbT6ayFtIKYHRPc4/jU0Acwax1dXCh9FRERERESc7cKHcGz4mPYFKYy3LeKrjWcxYUBbd1clIiIVdiyA7x6G7ETj+4QLwcsfUrdATrJxPTsRdi44/oxXAER3M4LI6O4Qc5bx1SfIPT+Dh1D4KCIiIiIi4my+IZgueQy+mcQk62fc9dsIhY8iIo1B5n5Y+AjsWmh8H9waRj4PXS4Dk8m4VpAJ6duMIDJts/E1fTuU5kPyauOoKrRteSBZZZRki3iNkiyn8FFERERERMQVzplA2W//IeTIVkZkzGbLoQvo0SrE3VWJiDRPpUWw8h/w8ytQVgRmLxh4D1z4EHgHVL/XPwzanm8cFWxlkLkXUjdD2hZI22qEkrmH4dgB49jxzfH7vYOOj5KsCCSjuoFPYEP8tI2KwkcRERERERFXMFuwjnoe3r+cGyzf89aKn+hx3RXurkpEpPnZvRS+ewgy9xnfJ1wIo/4OkZ1r34bFatwf2RnOuub49fyjx8PItC1GOHlkB5TkQtJvxlHJBGEJxlTt6LOqjJJsc3zUZROk8FFERERERMRVEi4ko/UwIpIX02fHSxSVjMLXW/8MExFpENnJsHAKbP/a+D4wBoY/Cz2udl7YFxAO7S4yjgq2Uji6p/q07bStkJdqBKCZ+2D7/47f7xNcHkhWHSXZ9cQRmR5K/68nIiIiIiLiQmFjXqDkjWUMZBO/LpvLeSNucHdJIiJNW1kJ/Pom/PSisU6jyQLn3gUXPwK+wa7v3+JlhIdRXYFrj1/Pzzhx2vaRHVCcA4m/GEclE4S3P3EtyZDWHjdKUuGjiIiIiIiIC5kj2rGx1fX0O/QB8aufhSHXgtXb3WWJiDRN+1fAtw9Axk7j+7jzYPTLRnjnbgER0H6wcVQoK4GMXeXTtitGSW6B/CPG6Mmje2Dbl8fv9w05MZCM6gpefg3+49SWwkcREREREREXa3n5Yxx56ytibYc49uMbhA6Z7O6SpMLRvbDzO+NI3wYt4iC8A4R3hIiOxsij8A7gE+TuSkXkVHJTYfFjsHme8b1/BAydDj3/1Lh3nbZ6GyFiTA9g/PHreenHR0lWTNvO2AlF2XBwpXFUMJmN/52qGkhG94CQVg3+49RE4aOIiIiIiIiLtYqJ5p2wW7n92Kv4/fIyDJhgjICRhme3QfIa2LnACBwrRkdVKMyElI0nPhcUWx5KdigPJcvPW8QbG1GIiHvYymD1f+CH54zpy5ig761w6ePgF+ru6uovMAo6XGocFcqK4cjOKtO2y8PJgqPG6MmMXbD18/Lno+HBXe6p/Q/0v5AiIiIiIiINIPqi29j6+ed05yCO+bdh6ncrtL0A/MPcXVrTV5IPe38wwsZdC6Eg4/hrZiu0PR86j4K4/pCTAkd3Q8ZuY1Tk0d3G9MfcFOM4sKJ622YvY/fa8PJRkpXBZEcjYPawtdlEPErS7/DNZGO6MkDLc4wp1q3OcW9drmL1gdizjaOCw2GM+vzjtO0W8e6r8w8UPoqIiIiIiDSAYT1a8tevbuUd+5OY9/8I+38ETMY/IhMuhISLIX5Ak9nd1O1yU42gced3sO9HKCs6/ppvCHQcBp1HQochxvcVWvY+sa3CrONBZMbu4+uwHd0LZYXHRxz9kW9IlSncHY6fh7UDb39n/8QizUd+Bix9Atb/1/jetwUMeQLO+TOYLW4trcGZTBAcaxwdhxy/7nC4r6Y/UPgoIiIiIiLSAHysFtr0Hsq1v9i5O2oTl3hvN3Y5TdloHKteN0bRte4H7S6ChIugVR9tTlNbDoexZmPFdOpDa6u/3iIeuow2Asc2A4zdaGvLrwW07mMcVdntkHPICCWP7i0PJsvDyawkY222Q2tPrAUgJK7madwhcY17fToRd7LbYd1sWPoUFGUZ13rdCEOf0lIWf9SIRl2bHI5GFIU2gJycHEJCQsjOziY4uAG2VxcRERFxMn2e8Xx6D5uvHak5jJhpTNt964ZzGNnWBPuXw76fYP9PkJ1U/QGvAGM0ZMJFRiAZfZaCqapspcamCzu/M0LHrMTqr7fuZ4SNnUdBZJeG/cd4aSFk7jOCyKqjJTN2Hw9NamL1hbD2J07hjujg2evXiZypw+uNKdaH1xnfR/cwpli3Oc+9dTVTdfkso/BRRERExMPo84zn03vYvM1YsJ1/Ld9HoI+V/917PgkR5dOsHQ44tv94ELl/ubGJQFV+YZBwgRFGJlxkBFSNaHRLgyjMgj1LjbBx91Iozj7+mtUX2g02AsdOIyAo2m1lnpTDYbyvNYWSmfvAXnryZ/3Dy9eW7FA+jbv8PCzBWAtOpCkqPAbLnoHV7wIO8A6CS6ZCv79osyc3Uvh4CvqgJyIiIp5On2c8n97D5q3UZueG//zG7wcy6RITxBd/G4Sfdw1rlNntkL61PIxcbozwK8mrfk9wa2O9yIpp2sGxDfNDNLRjB2DnQtj5LRxcBfay468FRBpBY+dR0O5iz15L0VYG2YmQUbGmZJWNb3IPn/w5k9mYVl45hbs9RHSCuHMVSorncjhg48ew+PHjm0SddS0MewaCYtxbmyh8PBV90BMRERFPp88znk/voaTlFDH6tRVk5JVwbZ/WvHRtz9M/ZCuFQ+uMUZH7foLk38FWUv2eiE7Hp2i3Pd9zp+na7cYUy4r1G9O3Vn89sosRNnYeZayL2RymohfnQebeE0dLHt1zYihdwT8CzrkJ+twCoY1n51uR00rdAgsehMRfjO8jOsPovxu/bJFGQeHjKeiDnoiIiHg6fZ7xfHoPBWDV3gxufOc37A548eqzGdcvrm4NlBQY/zCvmKJ9eANQ9Z93JojteXxUZJsBjXtUYGmhEaruXGDsUp2Xdvw1kwXiBx6fTh3e3n11NjYOh/FnVbnZTXlAmbKhyp+hydjdu9/t0OHS5rcbsHiOohz48Xn47W1w2MDLHy76Pzjvbm2+1cgofDwFfdATERERT6fPM55P76FU+OcPe3hp0U58rGY+/9tAurcMqX9jhcfgwM/H14zM2FX9dYs3tO5fZSftc+q247Mr5B0xgsad38HeZVBWePw17yDoOMQY3dhhCPiHua9OT2Qrg13fwep3YN+Px6+3aAN9b4XeN2l3YHdwOIyjOYzWrQuHA7Z8BoumQl6qca3rFTD8OWhRx1/MSINQ+HgK+qAnIiIink6fZzyf3kOpYLc7uO391fyw8wjx4f58fc/5hPg5KRDMOQz7Vxyfpp2TXP1170BjNGHFNO2o7q4PRBwOIxStmE6d9DvVRmuGxJXvTj0S4s/XSCdnydgDa2bBhv9CUfkGPRZv6DbGGA0Z17/5bVzUkBwOSF4Nm+fD1s+hOBdangNx/Yx1OVv3h8BId1fpPkd2wYIHjBHcAKEJMOrvxi8fpNFS+HgK+qAnIiIink6fZzyf3kOpKqughNGv/cyhrEKGdYvmXzf1weTsIMjhMHZS3vdj+TTtFVCYWf0e/3BjPbWKMDI0wTmBlK0Mkn41wsadC4w6qortZYxu7DIKonsoBHOlkgIj/Fr9jrGmZoXoHtDvNjhrHPgEuq++piZ9O2yeZ4SOWQdPfW9oghFEVgSSUd2a/vT4knxY/hKsesPY5d3qC+dPhkH3gZevu6uT01D4eAr6oCciIiKeTp9nPJ/eQ/mjjUlZXPv2L5TY7Ewd1ZW/XNjOtR3a7ZC22RhptO8nYwfp0vzq94TEHQ8iEy6s2+6yRTmw93sjcNy1CIqyjr9m8TbarVi/MaSVU34kqaNDa2H1LNgyH8qKjGveQdDzOiOIjOrq3vo81bGDxvThzfOrb5TkFQBdL4Me1xib/ySvNkb+Jv0OR7af2I53oLGZUlz/8tGRfT13A6k/cjhgx7ew8BHITjKudRwOI1+AsAT31ia1pvDxFPRBT0RERDydPs94Pr2HUpMPfz3I419uwfL/7d15eFT1vcfxz8wkmSxkI3sgCWGTQFAgQQyIuFJxqVhUVFSsdkFRwdz2utC6tTW3i5ZrFZS63baCFLWuWMWlgIiAbKIEUFkCmAAJkEkCmSQz5/5xwoSYgESSnJnM+/U888D85kzmO55Wvnz8LXabXvzZGRreqxP3OGyoM8Mo30naq8yZSEdLGtAURmaNkiLimr9+cGfj/o0LzZmVR78/orsZNJ4yTupzjuSM7vCvhBN0aL+0fp606hnzNO0jskaZIeSAS1n+/l2q90kbXzVnOe5c0TRuDzUP+hk8Qeo/7tgHPh0+KO3+tCmM3PWpVFfV8rrEUxrDyMZAMqFf4O0duX+r9PZd0pfvms9jM6Vx/2POfmbWc0AhfDwOGj0AABDo6GcCH/cQrTEMQ9Pnr9Nr675RcrRTb90xWknRTmuKqasxT9I+cnhN6WdqtjejzW4ul+49xpzJuHmhVLah+c/o3sdcSn3KRWZQ0tWXkAY6r9e816ueNmesGh5zPCpZGnaDlHcjB38crdZlzt77/CXp6w+b/nnJJvU6Uxp8pTTwh99vtqLXYy7Z3rWyKZA8Ohg+IjxO6jm8abl2jzz/Dfbra6VlM6Wlj0oetxnMjrpDGv2LY4ey8GuEj8dBowcAAAId/Uz7mzVrlv74xz+qtLRUgwYN0syZMzV69OjvfN+yZcs0ZswY5ebmat26dSf8edxDHEuNu0GXPbFMX+2t1sg+Cfr7zSPksPvBbKBD+6XtSxvDyCVSxZctr7HZzRDklHFm4JjYr/PrRPuo3C2t+T9p9fNS9R5zzGY3Z68Ov1nqfW7gzbhrD/W10leLzCXVW/7dtFxdktKHmoHjoMulmPT2/+ya8sal2iuknavMmcpHnw4vmfcoeVDTzMiM4e23d+vJ+HKRtPCX0oFt5vPsMeaBMkn9ra0LJ4Xw8Tho9AAAQKCjn2lf8+fP1/XXX69Zs2Zp1KhReuqpp/T0009r48aNyszMPOb7KisrNWzYMPXt21d79uwhfES7+WpvlX74+DIdqvPo9nP76r/GnmJ1SS1V7jZDyG1LzACm3wXm8tKoRKsrQ3vy1Juz+1Y9bYbPR8RnS/k3SUOvkyI7cXsAK3g95nffsEDa+Ibkrmx6LaGfGTgOvkJK6NO5dXnqzdnGRweSlSUtr4tKMk/TPrJcO32oFBrROTUe3Gnu67jpTfN5dJr0g99Jg35kfSCKk0b4eBw0egAAINDRz7SvESNGaNiwYZo9e7ZvLCcnR+PHj1dRUdEx33f11VerX79+cjgcevXVVwkf0a5eW7db015cJ0l67sfDdc4pydYWBOzbLH36rLRuruR2mWMOp5T7I2n4T8wlv10lUDIMafcaM3D84pWm2Z+SFJ1u7uGYe4WUdpp/fWdX6VFLtVdIpeslT13za+whUuqpjTMjGwPJ2J7tW0dDnbT8cfMk6/pDks0hnXGLdPbd/rssHG1G+HgcNHoAACDQ0c+0n7q6OkVGRmrBggW6/PLLfePTpk3TunXrtHjx4lbf99xzz2nWrFlavny5fvvb335n+Oh2u+V2u33PXS6XMjIyuIc4rl+/+rn+/skOxUWG6s3bz1TPePZFgx+oqzGXHa96Wir7rGk89VQzhBx8hRQWZV19J2PfZjNw3PBS0xJhydy3ceBl5izHzJGBs+S8vtYMIHc1hpE7VzYPUo+I6XHU3pGnm/fy+x4ytG2J9NZ/SeVbzOeZBdLFj0gpg77/94Bfaks/GtJJNQEAAAB+p7y8XB6PRykpKc3GU1JSVFZW1up7vvzyS919991aunSpQkJOrJ0uKirSgw8+eNL1Irj86pIcrd91UJ/tqtTUuWv1z5+fIWcIh7bAYmFRUt5k8xCa3avNEPLzV8wg8o07pHd/LQ25Rsq/OTD29KvcJX3+shk6Hn1oUmikuXfp4CulPucG5onfoeFS5gjzodvNGZ0HSxpP1G4MJMs+l1y7pY27zRO7JSkk3DxQyrd35OlSt++YfV1VJr0zwzyARzKXe1/wG+m0q/1rdigsQfgIAACAoGf71l+MDMNoMSZJHo9H1157rR588EH173/if6m+5557VFhY6Ht+ZOYjcDzOEIeeuHaYLvnLR1q/86AefqtYD16Wa3VZgMlmk3rmm48fPCyt/Yf06TPSge3SiifNR6/R5mzIARdLjlCrK25SU2EGbRtekko+bhq3h0h9zzcDx1PGBe4MzmOx2aT4LPNx6pXmWF2NucT86OXahw9IOz8xH0fE9zpq78gRUvJAyREieRqklXOkDx+W6qok2cx7fu6vpIg4C74k/BHLrgEAAAIM/Uz7aeuy64MHDyo+Pl4OR9PsM6/XK8Mw5HA49O677+rcc8/9zs/lHqItPti0Rzc9/6kk6bFrhuqHp3XASbpAe/B6pa0fSKueMU+DNrzmeLfUxtmSk6XYHtbU5q6WNi80Zzh+/YHkbWh6LetMc7n4wMu6/gE638UwpIqvzRDySCC5t1jSt6Kj0CipZ54Z5O79whzrkWcusU4f2ullo/Ox5+Nx0OgBAIBARz/TvkaMGKG8vDzNmjXLNzZw4EBddtllLQ6c8Xq92rhxY7OxWbNm6YMPPtBLL72k7OxsRUV990wZ7iHa6o/vbNITH36tyDCHXr9tlPomc2gD/NzBndLq56U1/yfV7DPHbA5zRuHwn0jZYzp+78SGOumr98zAcfPbUsPhptfSTjMPjcn9UfsfuNLV1FZKuz5tOll716dNhw5JUnicdP4DZrgcKPth4qSx5yMAAABwggoLC3X99dcrPz9fBQUFmjNnjkpKSjRlyhRJ5pLp3bt3629/+5vsdrtyc5sve01OTlZ4eHiLcaA93Xl+f63ZcVDLt1boln+s0Wu3jVJkGH+dgx+Ly5DO+7U05i5p0xvmbMgdy6RNb5qPhL5S/k3SkGvNA13ai9djfs6Gl6SNr0m1B5te697bXFKde0Vg7EfpL8Jjpb7nmQ/J/Ge8b7MZRLqrpCGTpKgEa2uEX+NPKwAAAAS1iRMnqqKiQg899JBKS0uVm5urhQsXKisrS5JUWlqqkpISi6tEsAtx2PW/1wzRJY99pC/3VuveVzbozxOHtLo3KeBXQsKk3AnmY89G6dNnpfUvShVfSe/cK73/G/O14TdLPYZ9v88wDKl0nRk4fv6yVFXa9Fq3VPPnD77CXA7M/2dOnt0hpQw0H8AJYNk1AABAgKGfCXzcQ3xfK7ZW6NqnV8jjNfS7y3M1aUSW1SUBbeeuMpdCr3pG2vN503j6UHNJ9qAfSWGR3/1zyr80A8cNC6T9XzeNh8ea+zcOvlLKGmWGZQDaFXs+HgeNHgAACHT0M4GPe4iT8dTir1X09iaFOex6+ZaRGtwz1uqSgO/HMMwDTVY9bZ4+7akzx8PjzKW8+TdJiX2bv6dyt/TFK2boWLquaTwkwtxPcvCV5vLgEGcnfQkgOBE+HgeNHgAACHT0M4GPe4iTYRiGfvb31Vq0cY96xkfordtHKzYy1OqygJNTUy6t/bu5LPvgUVtd9D7bDCEPHzADx+0fyXfyss1hBo2DrzSDRycHMQGdhfDxOGj0AABAoKOfCXzcQ5ysysP1uuQvS7Vz/2GdNyBZf70hX3Y7e9mhC/B6pK/eN2dDfvmufEHj0TILzD0cB46XohI7u0IA4rRrAAAAAOjSYiNCNXtSnn40+2O9v2mvnlqyVbec3cfqsoCTZ3dI/ceajwPbpdXPmzMeI+LMU6pzJ5gnaQMIGISPAAAAABCAcnvE6sEfDtI9r2zQH9/ZpCEZcSrok2B1WUD7ie8lnf+A+QAQsOxWFwAAAAAA+H6uHp6hHw3rIa8h3T5vrfa6aq0uCQCAZggfAQAAACBA2Ww2/XZ8rk5JiVZ5tVu3z1urBo/X6rIAAPAhfAQAAACAABYZFqJZ1w1TVJhDK7bt1yOLtlhdEgAAPoSPAAAAABDg+iR10++vOFWSNPs/X+u9jXssrggAABPhIwAAAAB0AZecmq4bR/aSJBX+c5127j9kbUEAAIjwEQAAAAC6jHsvytGQjDi5aht0ywurVVvvsbokAECQI3wEAAAAgC4iLMSuJyYNU3xkqD7f7dJv3txodUkAgCBH+AgAAAAAXUiPuAjNvHqobDbphRUl+tfaXVaXBAAIYoSPAAAAANDFjOmfpDvO7SdJuveVz7VlT5XFFQEAghXhIwAAAAB0QXec10+j+yXqcL1HU/6xWtXuBqtLAgAEIcJHAAAAAOiCHHabZk4cotSYcG3dV6N7XtkgwzCsLgsAEGQIHwEAAACgi0ro5tQTk4YqxG7TG+u/0d8/2WF1SQCAIEP4CAAAAABdWF5Wd91zUY4k6TdvbtTakgMWVwQACCaEjwAAAADQxd00qpfG5aaq3mNo6gtrdKCmzuqSAABBgvARAAAAALo4m82mP1xxqrITo/RNZa2mz18nr5f9HwEAHY/wEQAAAACCQHR4qGZNGiZniF2Lt+zTEx9+ZXVJAIAgQPgIAAAAAEEiJy1Gvx2fK0l69L0t+ujLcosrAgB0dYSPAAAAABBErszP0MT8DBmGNO3FtSqrrLW6JABAF0b4CAAAAABB5sHLBiknLUYVNXW6be4a1Xu8VpcEAOiiCB8BAAAAIMiEhzo0e9IwRTtD9OmOA/rDvzdZXRIAoIsifAQAAACAINQrMUp/vPI0SdJfl27Tvz8vtbgiAEBXRPgIAAAAAEHqwtxU/XR0tiTplws+0/byGosrAgB0NYSPAAAAABDE/vvCAcrPileVu0G3vLBGtfUeq0sCAHQhhI8AAAAAEMRCHXY9fu0wJUSFqbjUpftf+8LqkgAAXQjhIwAAAAAEudTYcD12zVDZbNL8T3fqn5/utLokAEAXQfgIAAAAANCovokqPL+/JOnXr36ujd+4LK4IANAVWB4+zpo1S9nZ2QoPD1deXp6WLl16zGtfeeUVXXDBBUpKSlJMTIwKCgr0zjvvdGK1AAAAANB1TT2nr84+JUnuBq9ufWG1XLX1VpcEAAhwloaP8+fP1/Tp0zVjxgytXbtWo0eP1rhx41RSUtLq9UuWLNEFF1yghQsXavXq1TrnnHN06aWXau3atZ1cOQAAAAB0PXa7TX++aojSY8O1veKQ7nrpMxmGYXVZAIAAZjMs/JNkxIgRGjZsmGbPnu0by8nJ0fjx41VUVHRCP2PQoEGaOHGi7rvvvlZfd7vdcrvdvucul0sZGRmqrKxUTEzMyX0BAAAAC7hcLsXGxtLPBDDuIfzd2pIDuuqp5ar3GPr1JQN185nZVpcEAPAjbellLJv5WFdXp9WrV2vs2LHNxseOHauPP/74hH6G1+tVVVWVunfvfsxrioqKFBsb63tkZGScVN0AAAAA0NUNzYzXry4eKEkqWlis1Tv2W1wRACBQWRY+lpeXy+PxKCUlpdl4SkqKysrKTuhnPPLII6qpqdFVV111zGvuueceVVZW+h47d3JqGwAAAAB8lxsKsnTJqWlq8Bqa+sJaVVS7v/tNAAB8i+UHzthstmbPDcNoMdaaefPm6YEHHtD8+fOVnJx8zOucTqdiYmKaPQAAAAAAx2ez2fQ/E05V76QolblqNX3+Onm87P8IAGgby8LHxMREORyOFrMc9+7d22I25LfNnz9fN998s/75z3/q/PPP78gyAQAAACBodXOG6Mnr8hQR6tDSL8v12PtfWl0SACDAWBY+hoWFKS8vT4sWLWo2vmjRIo0cOfKY75s3b55uvPFGzZ07VxdffHFHlwkAAAAAQa1/SrSKfjRYkvTYB19q8ZZ9FlcEAAgkli67Liws1NNPP61nn31WxcXFuvPOO1VSUqIpU6ZIMvdrvOGGG3zXz5s3TzfccIMeeeQRnXHGGSorK1NZWZkqKyut+goAAAAA0OWNH9pDk0ZkyjCk6S+u1TcHD1tdEgAgQFgaPk6cOFEzZ87UQw89pCFDhmjJkiVauHChsrKyJEmlpaUqKSnxXf/UU0+poaFBU6dOVVpamu8xbdo0q74CAAAAAASFX18yUIN7xOrAoXpNnbtGdQ1eq0sCAAQAm2EYQbVjsMvlUmxsrCorKzl8BgAABCT6mcDHPUSg2rn/kC5+bKlctQ368aheuv/SQVaXBACwQFt6GctPuwYAAAAABIaM7pF69KohkqTnlm3XW5+VWlsQAMDvET4CAAAAAE7Y+QNTdMvZfSRJ//3Sen29r9riigAA/ozwEQAAAADQJv91QX+NyO6umjqPfvJ/n+qN9d+wByQAoFWEjwAAAACANglx2PWXa4cqOdqpbeU1un3eWo38n/f1+39vUknFIavLAwD4EcJHAAAAAECbJUeH6807ztQd5/VTcrRT5dV1mv2frzXmTx/qhmdX6p0vytTgYTYkAAQ7TrsGAAAIMPQzgY97iK6m3uPV+8V7NXdliZZs2ecbT4lxauLwTF09PEPpcREWVggAaE9t6WUIHwEAAAIM/Uzg4x6iK9tRUaN5K3dqwac7VVFTJ0my26RzB6Ro0ohMndU/SQ67zeIqAQAng/DxOGj0AABAoKOfCXzcQwQDd4NH736xRy+s2KFPtu73jfeIi9A1p2foquEZSo4Ot7BCAMD3Rfh4HDR6AAAg0NHPBD7uIYLNV3urNW9liV5avUuVh+slSSF2m8YOStGkEVkq6J0gO7MhASBgED4eB40eAAAIdPQzgY97iGBVW+/Rwg2lemFFiVbvOOAb75UQqWtHZOqKvAx1jwqzsEIAwIkgfDwOGj0AABDo6GcCH/cQkDaVuTR3RYleWbNb1e4GSVKYw65xg1M1aUSWhveKl83GbEgA8EeEj8dBowcAAAId/Uzg4x4CTWrcDXpj/Teau7JEn+2q9I33S+6ma0dk6kfDeio2ItTCCgEA39aWXsbeSTUBAAAAfmvWrFnKzs5WeHi48vLytHTp0mNe+8orr+iCCy5QUlKSYmJiVFBQoHfeeacTqwW6lihniK4+PVOv33amXr9tlK4enqGIUIe+3FutB9/YqBEPv6dfLlivtSUHFGRzZwCgSyB8BAAAQFCbP3++pk+frhkzZmjt2rUaPXq0xo0bp5KSklavX7JkiS644AItXLhQq1ev1jnnnKNLL71Ua9eu7eTKga7n1J5x+p8Jp2rFjPP0m8sGaUBqtGrrvVqwepcun/WxLn7sI72wYodvmTYAwP+x7BoAACDA0M+0rxEjRmjYsGGaPXu2bywnJ0fjx49XUVHRCf2MQYMGaeLEibrvvvtO6HruIXBiDMPQmpIDemFFid78rFR1DV5JUlSYQ5cN7aFrT89Ubo9Yi6sEgODTll4mpJNqAgAAAPxOXV2dVq9erbvvvrvZ+NixY/Xxxx+f0M/wer2qqqpS9+7dj3mN2+2W2+32PXe5XN+vYCDI2Gw25WV1V15Wd913yUC9vGa3XlixQ1v31WjuihLNXVGi0zLiNGlEpi49NV0RYQ6rSwYAfAvLrgEAABC0ysvL5fF4lJKS0mw8JSVFZWVlJ/QzHnnkEdXU1Oiqq6465jVFRUWKjY31PTIyMk6qbiAYxUWG6eYzs/V+4RjN++kZuvS0dIU6bFq/86D++6XPdPrD7+mB17/Qlj1VVpcKADgKMx8BAAAQ9Gw2W7PnhmG0GGvNvHnz9MADD+i1115TcnLyMa+75557VFhY6HvucrkIIIHvyWazqaBPggr6JKi8eqAWfLpL81aWqGT/IT3/8XY9//F2nd6ru64dkakLc1MVHspsSACwEuEjAAAAglZiYqIcDkeLWY579+5tMRvy2+bPn6+bb75ZCxYs0Pnnn3/ca51Op5xO50nXC6C5xG5O3XJ2H/38rN766KtyvbBih94r3quV2/dr5fb9in8jVFfmZ+ia0zOVnRhldbkAEJRYdg0AAICgFRYWpry8PC1atKjZ+KJFizRy5Mhjvm/evHm68cYbNXfuXF188cUdXSaA72C323RW/yQ9dX2+Pr77XBVe0F9pseE6cKhec5Zs1Tl/+o8mPf2JFm4oVb3Ha3W5ABBUmPkIAACAoFZYWKjrr79e+fn5Kigo0Jw5c1RSUqIpU6ZIMpdM7969W3/7298kmcHjDTfcoP/93//VGWec4Zs1GRERodhYTt0FrJYSE647zuunW8/uo/9s3qe5K0v04ea9WvZVhZZ9VaHEbk5NHN5TVw/PVEb3SKvLBYAuj/ARAAAAQW3ixImqqKjQQw89pNLSUuXm5mrhwoXKysqSJJWWlqqkpMR3/VNPPaWGhgZNnTpVU6dO9Y1PnjxZzz//fGeXD+AYQhx2nT8wRecPTNGuA4f04sqdmv/pTu2rcuuJD7/WrP98rTH9kzRpRJbOOSVJIQ4WBgJAR7AZhmFYXURncrlcio2NVWVlpWJiYqwuBwAAoM3oZwIf9xCwRr3Hq/c27tHclSVa+mW5bzwtNlwTh2fo6uGZSo0Nt7BCAAgMbellCB8BAAACDP1M4OMeAtbbXl6jeStLtGD1Lu2vqZMk2WzSgNQYFfQ2T9M+Pbu7YiNCLa4UAPwP4eNx0OgBAIBARz8T+LiHgP9wN3j078/L9MKKEq3ctr/Za3abNCg9Vmf07q6CPgka3qu7osMJIwGgLb0Mez4CAAAAAIKWM8Shy4b00GVDemhvVa0+2bpfn2yt0CdfV2hreY027K7Uht2V+uvSbXLYbcrtEeubGZmfFa8oJ3+tBoDjYeYjAABAgKGfCXzcQyAwlFXW6pOtFVr+dYWWb61Qyf5DzV4Psdt0as9YFfRJUEHvROVlxSsizGFRtQDQeVh2fRw0egAAINDRzwQ+7iEQmHYfPKxPGoPI5V9XaPfBw81eD3XYNDQjXmf0SdAZvbtrWGa8wkMJIwF0PYSPx0GjBwAAAh39TODjHgJdw879h7T86wpzduTWCpVW1jZ7PSzErmGZcSronaiCPgk6LSNWzhDCSACBjz0fAQAAAADoYBndI5XRPVJXDc+QYRjaUXHINyty+dYK7atyN+4huV9/fk8KD7UrLyvet2fkqT3jFOqwW/01AKBDET4CAAAAAHCSbDabeiVGqVdilK45PVOGYejrfTW+WZGffF2hipo6LfuqQsu+qpAkRYY5lN+ruy+MzE2PUQhhJIAuhvARAAAAAIB2ZrPZ1De5m/omd9N1Z2TJMAx9ubfanBX5dYVWbKvQgUP1WrJln5Zs2SdJ6uYM0fBe8b4DbAamx8hht1n8TQDg5BA+AgAAAADQwWw2m/qnRKt/SrQmj+wlr9fQ5j1VviXaK7ZWyFXboA8379OHm80wMiY8RKdnm4fXFPRJUE5qjOyEkQACDOEjAAAAAACdzG63KSctRjlpMbrpzGx5vIaKS12+A2xWbtsvV22D3iveo/eK90iS4iJDNSL7yDLtRPVP6SabjTASgH8jfAQAAAAAwGIOu025PWKV2yNWPz2rtxo8Xn3xjct3gM2q7ft18FC93vlij975wgwjE6LCNKJ3056RfZIIIwH4H5thGIbVRXSmthwFDgAA4I/oZwIf9xBAW9V7vPpsV6U+2WrOjFy1fb9q673NrkmKduqM3gkq6G0u1c5OjCKMBNAh2tLLED4CAAAEGPqZwMc9BHCy6hq8Wr/roO8Am9UlB1TX0DyMjHaGaEBatAakxmhAWrRy0mJ0Skq0opwsggRwctrSy/BvHAAAAAAAAkxYiF3De3XX8F7ddcd5/VRb79G6nQd9B9isKzmoKneDVm0/oFXbDzR7b1ZCpAakmqGkue9ktDLiIznMBkCHIHwEAAAAACDAhYc6dEbvBJ3RO0F3ypwZubW8WptKq1Rc6lJxWZU2lbq0t8qtHRWHtKPikG/vSEmKDHPolFRzdmROarQGpMXolNRoxYSHWvelAHQJhI8AAAAAAHQxYSF2c7l1aozGD+3hG6+odmtzWZU2lrq0qaxKm8pc2rKnWofqPFpbclBrSw42+zk94iJ8syOPLN/ulRAlB7MkAZwgwkcAAAAAAIJEQjenRvZ1amTfRN9Yg8er7RU12lhqzo7c1DhL8pvKWu0+eFi7Dx7We8VNsyTDQ+06JaUpjDSXb0crLjLMiq8EwM8RPgIAAAAAEMRCHHb1TY5W3+Ro/fC0dN945aF6FZe5fIFkcVmVNpe5VFvv1fpdlVq/q7LZz0mLDTf3kkyL8S3fzk6MUojD3tlfCYAfIXwEAAAAAAAtxEaG+vaRPMLjNbSjosY3O7K4zNxTcteBwyqtrFVpZa0+3LzPd31YiF39krspJy1GAxr3lByQGq2Ebk4rvhIACxA+AgAAAACAE+Kw29Q7qZt6J3XTRYPTfOOu2nptaZwduanUpeJSlzaXVammzqMvvnHpi29czX5OUrTzqMNtzKXbfZK6KSyEWZJAV0P4CAAAAAAATkpMeKjye3VXfq/uvjGv19CuA4cbl26bMyQ3lbm0Y/8h7atya1/VPi3Z0jRLMtRhU5+kplmSAxoPuknq5pTNxgE3QKAifAQAAAAAAO3ObrcpMyFSmQmR+sGgVN94jbtBW/ZUmftIljYGk2UuVdU2NJ7AXdXs5yREhSkzIVLJ0U6lxIQrOdqp5MZfU2LClRITrvjIUAJKwE8RPgIAAAAAgE4T5QzR0Mx4Dc2M940ZhqFvKmt9S7aPLN/eVl6jipo6VdTUHfdnhjpsSo4OV1K0UykxTiVHh/t+TT7qeXxkmOx2QkqgMxE+AgAAAAAAS9lsNvWIi1CPuAidl5PiG6+t9+jLPdXaffCw9lXVao/Lrb2+X93a66pVRU2d6j2Gdh88rN0HDx/3c0LsNiVHO5UUE66UaKeSY5xKORJQHjWbsjshJdBuCB8BAAAAAIBfCg91aHDPWA3uGXvMa+oavCqvNsPIPa5a7XXVNv2+yq09Lrf2VdWqvLpODV5zhuU3lbXH/dwQu01JzZZ3tz6bMiGKkBL4LoSPAAAAAAAgYIWF2JUeF6H0uIjjXlfvMUPKPS5zxuSeKrf2ucxZlHuqarW3cVZlRY0ZUpZW1qr0BELKxG5mOJl0VDiZEuP0BZTJMU4lRDnlIKREkCJ8BAAAAAAAXV6ow6602AilxZ5YSLnX1TR7srXZlBU1bjV4DZW5alXmqpVUecyf6bDblNgtzHdATo+4CPWMN5eZp8dFqEd8hBKiwjg0B10S4SMAAAAAAECjEw0pGzxelVfX+fagPFZQWV7tlsdrNF7j1rFCyvBQcwbn0cFkj/gIpceav6bGhCvEYe+Abwx0LMJHAAAAAACANgpx2JUaG67U2PDjXtfg8aqipk57Gpd4l1Ue1q6Dh7X7gHlAzjcHD2tvlVu19V5t3VejrftqWv05DrtNqY2zJnt8a9bkkcN6IsIcHfFVgZNC+AgAAAAAANBBQhx233LrY3E3eFRWWavdB1oGk0d+bXai9/bWf05CVFjzYPKocLJnfIRiI0JZ2o1OR/gIAAAAAABgIWeIQ1kJUcpKiGr1da/X0L5qt3Y1hpJmOHlI3xys9QWV1e4GVdTUqaKmTp/tan1pd1SYo9VZk+Yy70glRzs5vRvtjvARAAAAAADAj9ntNt/sybys+BavG4Yh1+EG7Tp4qMWsySPPy6vrVFPn0ZY91dqyp7rVzwl12JQW2zyc7HlUSJkWFy5nCEu70TaEjwAAAAAAAAHMZrMpNjJUsZGxGpQe2+o1tfWeZmHkN42/P7LMu8xVq3qPoZL9h1Sy/9AxPys52ukLJtNjwxUXGaa4yFDFRoQqLqLp97GRoYp2hrDMG4SPAAAAAAAAXV14qEN9krqpT1K3Vl9v8Hi1p8qt3QeaZk02LfM+pN0HD6u23mue6F3l1rqdB7/zMx12mxlENj7iIkMVd+R5ZJjiGsdifb82hZehnOzdZRA+AgAAAAAABLkQh913anZrDMPQ/po6c5/Jg4e068Bh7XHV6uChelUertfBw/WqPFSvg4frVHm4XrX1Xnm85nv219S1uZ6oMIfiIsMUExHqCynjIkMbn4d9K8gMNWdgRoQqMszBbEs/Q/gIAAAAAACA47LZbEro5lRCN6cG92x9affRaus9Zih5JJw8VOcLKM2wss732pHrDh6qk6u2QZJUU+dRTV3j6d5tEGK3NS39jmgKJWN9S8PNsW8/jwkPUQizLTsE4SMAAAAAAADaVXioQ+GhDqXEhLfpfR6voaraxjCyMbRsHlAeCSzrfNccCTfrPYYavIbKq+tUXt322ZYRoQ51Cw9RN2eIopwOdXOG+B5R3/59eCuvhYeoW5j5XoLMJoSPAAAAAAAA8AsOu63xEJuwNr3PMAwdrvccNdOyKaA8siz8yNjRQabrcL2q3OZsy8P1Hh2u92hflfukv0d4qP2Eg8voxl+jnA5Fh7e8PtD3vyR8BAAAAAAAQECz2WyKDAtRZFiI0o+xb+Wx1Hu8ch2uV43boyq3+Wu1u17Vbo9q3A2qrm1Qtdt81Lib/76qtkE1dQ3me2obVOfxSpJq672qrf9+MzC/zRli982sjAprmmXZFFI6mgWW3cJDlBDlVEGfhJP+7PZA+AgAAAAAAICgFeqwN+5nefI/q67B2yygbBZatggxPc0DzcYg88h17gYzyHQ3eOVuqFNFGw7u6ZMUpff/6+yT/0LtgPARAAAAAAAAaAdhIXaFhYQpPqpty8ZbU+/xtphpWd04w7LFDMzGX4/MxjzWqeVWIHwEAAAAAAAA/Eyow/699r/0N4G9YyUAAAAAAAAAv0X4CAAAAAAAAKBDED4CAAAAAAAA6BCEjwAAAAAAAAA6BOEjAAAAAAAAgA5B+AgAAAAAAACgQxA+AgAAAAAAAOgQhI8AAAAAAAAAOgThIwAAAAAAAIAOQfgIAAAAAAAAoEMQPgIAAAAAAADoEISPAAAACHqzZs1Sdna2wsPDlZeXp6VLlx73+sWLFysvL0/h4eHq3bu3nnzyyU6qFAAAILAQPgIAACCozZ8/X9OnT9eMGTO0du1ajR49WuPGjVNJSUmr12/btk0XXXSRRo8erbVr1+ree+/VHXfcoZdffrmTKwcAAPB/NsMwDKuL6Ewul0uxsbGqrKxUTEyM1eUAAAC0Gf1M+xoxYoSGDRum2bNn+8ZycnI0fvx4FRUVtbj+rrvu0uuvv67i4mLf2JQpU7R+/XotX778hD6TewgAAAJZW3oZZj4CAAAgaNXV1Wn16tUaO3Zss/GxY8fq448/bvU9y5cvb3H9D37wA3366aeqr69v9T1ut1sul6vZAwAAIBhYHj6yvw4AAACsUl5eLo/Ho5SUlGbjKSkpKisra/U9ZWVlrV7f0NCg8vLyVt9TVFSk2NhY3yMjI6N9vgAAAICfszR8ZH8dAAAA+AObzdbsuWEYLca+6/rWxo+45557VFlZ6Xvs3LnzJCsGAAAIDJaGj48++qhuvvlm/eQnP1FOTo5mzpypjIyMZvvtHO3JJ59UZmamZs6cqZycHP3kJz/RTTfdpD/96U+dXDkAAAC6gsTERDkcjhazHPfu3dtiduMRqamprV4fEhKihISEVt/jdDoVExPT7AEAABAMQqz64CP769x9993Nxr/P/jrPPPOM6uvrFRoa2uI9brdbbrfb97yyslKS2GcHAAAErCN9TJCdG9ghwsLClJeXp0WLFunyyy/3jS9atEiXXXZZq+8pKCjQG2+80Wzs3XffVX5+fqv9aGuO3Dt6UgAAEIja0o9aFj52xP46aWlpLd5TVFSkBx98sMU4++wAAIBAV1VVpdjYWKvLCHiFhYW6/vrrlZ+fr4KCAs2ZM0clJSWaMmWKJHPJ9O7du/W3v/1Nknmy9eOPP67CwkL99Kc/1fLly/XMM89o3rx5J/yZVVVVkuhJAQBAYDuRftSy8PGIzthfp7Cw0Pfc6/Vq//79SkhIOO7nnCyXy6WMjAzt3LmTZTV+hnvjn7gv/on74r+4N/6ps+6LYRiqqqpSenp6h31GMJk4caIqKir00EMPqbS0VLm5uVq4cKGysrIkSaWlpc32JM/OztbChQt155136oknnlB6eroee+wxTZgw4YQ/Mz09XTt37lR0dHSH9aT8e8J/cW/8E/fFf3Fv/BP3xX91xr1pSz9qWfjYmfvrOJ3OZmNxcXHfv/A2Yk8f/8W98U/cF//EffFf3Bv/1Bn3hRmP7evWW2/Vrbfe2uprzz//fIuxMWPGaM2aNd/78+x2u3r27Pm9398W/HvCf3Fv/BP3xX9xb/wT98V/dfS9OdF+1LIDZ47eX+doixYt0siRI1t9T0FBQYvr27q/DgAAAAAAAIDOYelp14WFhXr66af17LPPqri4WHfeeWeL/XVuuOEG3/VTpkzRjh07VFhYqOLiYj377LN65pln9Itf/MKqrwAAAAAAAADgGCzd89GK/XU6i9Pp1P33399iyTesx73xT9wX/8R98V/cG//EfYE/4X+P/ot745+4L/6Le+OfuC/+y9/ujc04kTOxAQAAAAAAAKCNLF12DQAAAAAAAKDrInwEAAAAAAAA0CEIHwEAAAAAAAB0CMJHAAAAAAAAAB2C8LEDzJo1S9nZ2QoPD1deXp6WLl1qdUlBr6ioSMOHD1d0dLSSk5M1fvx4bd682eqy8C1FRUWy2WyaPn261aVA0u7du3XdddcpISFBkZGRGjJkiFavXm11WUGvoaFBv/rVr5Sdna2IiAj17t1bDz30kLxer9WlBZUlS5bo0ksvVXp6umw2m1599dVmrxuGoQceeEDp6emKiIjQ2WefrS+++MKaYhG06En9C/1o4KAn9R/0o/6JftQ/BFI/SvjYzubPn6/p06drxowZWrt2rUaPHq1x48appKTE6tKC2uLFizV16lR98sknWrRokRoaGjR27FjV1NRYXRoarVq1SnPmzNGpp55qdSmQdODAAY0aNUqhoaF6++23tXHjRj3yyCOKi4uzurSg9/vf/15PPvmkHn/8cRUXF+sPf/iD/vjHP+ovf/mL1aUFlZqaGp122ml6/PHHW339D3/4gx599FE9/vjjWrVqlVJTU3XBBReoqqqqkytFsKIn9T/0o4GBntR/0I/6L/pR/xBI/ajNMAyj0z+1CxsxYoSGDRum2bNn+8ZycnI0fvx4FRUVWVgZjrZv3z4lJydr8eLFOuuss6wuJ+hVV1dr2LBhmjVrln77299qyJAhmjlzptVlBbW7775by5YtY5aMH7rkkkuUkpKiZ555xjc2YcIERUZG6u9//7uFlQUvm82mf/3rXxo/frwk878yp6ena/r06brrrrskSW63WykpKfr973+vn//85xZWi2BBT+r/6Ef9Dz2pf6Ef9V/0o/7H3/tRZj62o7q6Oq1evVpjx45tNj527Fh9/PHHFlWF1lRWVkqSunfvbnElkKSpU6fq4osv1vnnn291KWj0+uuvKz8/X1deeaWSk5M1dOhQ/fWvf7W6LEg688wz9f7772vLli2SpPXr1+ujjz7SRRddZHFlOGLbtm0qKytr1g84nU6NGTOGfgCdgp40MNCP+h96Uv9CP+q/6Ef9n7/1oyGd/oldWHl5uTwej1JSUpqNp6SkqKyszKKq8G2GYaiwsFBnnnmmcnNzrS4n6L344otas2aNVq1aZXUpOMrWrVs1e/ZsFRYW6t5779XKlSt1xx13yOl06oYbbrC6vKB21113qbKyUgMGDJDD4ZDH49Hvfvc7XXPNNVaXhkZH/sxvrR/YsWOHFSUhyNCT+j/6Uf9DT+p/6Ef9F/2o//O3fpTwsQPYbLZmzw3DaDEG69x222367LPP9NFHH1ldStDbuXOnpk2bpnfffVfh4eFWl4OjeL1e5efn6+GHH5YkDR06VF988YVmz55Ns2ex+fPn6x//+Ifmzp2rQYMGad26dZo+fbrS09M1efJkq8vDUegHYDX+N+i/6Ef9Cz2pf6If9V/0o4HDX3oBwsd2lJiYKIfD0eK/KO/du7dF2gxr3H777Xr99de1ZMkS9ezZ0+pygt7q1au1d+9e5eXl+cY8Ho+WLFmixx9/XG63Ww6Hw8IKg1daWpoGDhzYbCwnJ0cvv/yyRRXhiF/+8pe6++67dfXVV0uSBg8erB07dqioqIhmz0+kpqZKMv+Lc1pamm+cfgCdhZ7Uv9GP+h96Uv9EP+q/6Ef9n7/1o+z52I7CwsKUl5enRYsWNRtftGiRRo4caVFVkMx0/7bbbtMrr7yiDz74QNnZ2VaXBEnnnXeeNmzYoHXr1vke+fn5mjRpktatW0eTZ6FRo0Zp8+bNzca2bNmirKwsiyrCEYcOHZLd3vyPb4fDIa/Xa1FF+Lbs7GylpqY26wfq6uq0ePFi+gF0CnpS/0Q/6r/oSf0T/aj/oh/1f/7WjzLzsZ0VFhbq+uuvV35+vgoKCjRnzhyVlJRoypQpVpcW1KZOnaq5c+fqtddeU3R0tG8mQGxsrCIiIiyuLnhFR0e32OcoKipKCQkJ7H9ksTvvvFMjR47Uww8/rKuuukorV67UnDlzNGfOHKtLC3qXXnqpfve73ykzM1ODBg3S2rVr9eijj+qmm26yurSgUl1dra+++sr3fNu2bVq3bp26d++uzMxMTZ8+XQ8//LD69eunfv366eGHH1ZkZKSuvfZaC6tGMKEn9T/0o/6LntQ/0Y/6L/pR/xBQ/aiBdvfEE08YWVlZRlhYmDFs2DBj8eLFVpcU9CS1+njuueesLg3fMmbMGGPatGlWlwHDMN544w0jNzfXcDqdxoABA4w5c+ZYXRIMw3C5XMa0adOMzMxMIzw83Ojdu7cxY8YMw+12W11aUPnwww9b/XNl8uTJhmEYhtfrNe6//34jNTXVcDqdxllnnWVs2LDB2qIRdOhJ/Qv9aGChJ/UP9KP+iX7UPwRSP2ozDMPozLATAAAAAAAAQHBgz0cAAAAAAAAAHYLwEQAAAAAAAECHIHwEAAAAAAAA0CEIHwEAAAAAAAB0CMJHAAAAAAAAAB2C8BEAAAAAAABAhyB8BAAAAAAAANAhCB8BAAAAAAAAdAjCRwCwkM1m06uvvmp1GQAAAAhS9KMAOhrhI4CgdeONN8pms7V4XHjhhVaXBgAAgCBAPwogGIRYXQAAWOnCCy/Uc88912zM6XRaVA0AAACCDf0ogK6OmY8AgprT6VRqamqzR3x8vCRzCcrs2bM1btw4RUREKDs7WwsWLGj2/g0bNujcc89VRESEEhIS9LOf/UzV1dXNrnn22Wc1aNAgOZ1OpaWl6bbbbmv2enl5uS6//HJFRkaqX79+ev31132vHThwQJMmTVJSUpIiIiLUr1+/Fs0pAAAAAhf9KICujvARAI7j17/+tSZMmKD169fruuuu0zXXXKPi4mJJ0qFDh3ThhRcqPj5eq1at0oIFC/Tee+81a+Zmz56tqVOn6mc/+5k2bNig119/XX379m32GQ8++KCuuuoqffbZZ7rooos0adIk7d+/3/f5Gzdu1Ntvv63i4mLNnj1biYmJnfcPAAAAAJaiHwUQ8AwACFKTJ082HA6HERUV1ezx0EMPGYZhGJKMKVOmNHvPiBEjjFtuucUwDMOYM2eOER8fb1RXV/tef+uttwy73W6UlZUZhmEY6enpxowZM45ZgyTjV7/6le95dXW1YbPZjLffftswDMO49NJLjR//+Mft84UBAADgV+hHAQQD9nwEENTOOecczZ49u9lY9+7dfb8vKCho9lpBQYHWrVsnSSouLtZpp52mqKgo3+ujRo2S1+vV5s2bZbPZ9M033+i88847bg2nnnqq7/dRUVGKjo7W3r17JUm33HKLJkyYoDVr1mjs2LEaP368Ro4c+b2+KwAAAPwP/SiAro7wEUBQi4qKarHs5LvYbDZJkmEYvt+3dk1ERMQJ/bzQ0NAW7/V6vZKkcePGaceOHXrrrbf03nvv6bzzztPUqVP1pz/9qU01AwAAwD/RjwLo6tjzEQCO45NPPmnxfMCAAZKkgQMHat26daqpqfG9vmzZMtntdvXv31/R0dHq1auX3n///ZOqISkpSTfeeKP+8Y9/aObMmZozZ85J/TwAAAAEDvpRAIGOmY8Agprb7VZZWVmzsZCQEN8m2gsWLFB+fr7OPPNMvfDCC1q5cqWeeeYZSdKkSZN0//33a/LkyXrggQe0b98+3X777br++uuVkpIiSXrggQc0ZcoUJScna9y4caqqqtKyZct0++23n1B99913n/Ly8jRo0CC53W69+eabysnJacd/AgAAALAS/SiAro7wEUBQ+/e//620tLRmY6eccoo2bdokyTz578UXX9Stt96q1NRUvfDCCxo4cKAkKTIyUu+8846mTZum4cOHKzIyUhMmTNCjjz7q+1mTJ09WbW2t/vznP+sXv/iFEhMTdcUVV5xwfWFhYbrnnnu0fft2RUREaPTo0XrxxRfb4ZsDAADAH9CPAujqbIZhGFYXAQD+yGaz6V//+pfGjx9vdSkAAAAIQvSjALoC9nwEAAAAAAAA0CEIHwEAAAAAAAB0CJZdAwAAAAAAAOgQzHwEAAAAAAAA0CEIHwEAAAAAAAB0CMJHAAAAAAAAAB2C8BEAAAAAAABAhyB8BAAAAAAAANAhCB8BAAAAAAAAdAjCRwAAAAAAAAAdgvARAAAAAAAAQIf4f4LCJ26dw+aIAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 1600x600 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_history(h)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_tweets, test_labels = get_tweets(test)\n",
    "test_sequences = get_sequences(tokenizer, test_tweets)\n",
    "test_labels = np.array(test_labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "63/63 [==============================] - 1s 14ms/step - loss: 0.4528 - accuracy: 0.8865\n"
     ]
    }
   ],
   "source": [
    "_ = model.evaluate(test_sequences, test_labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 67
    },
    "colab_type": "code",
    "id": "rh638vHG5ej6",
    "outputId": "feb44736-ee6e-477e-d24a-a2fad5ebaf62"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Sentence: i didn t feel talented at anything i was doing and eventually wasn t putting fully into it\n",
      "Emotion: surprise\n",
      "1/1 [==============================] - 1s 1s/step\n",
      "Predicted Emotion: surprise\n"
     ]
    }
   ],
   "source": [
    "i = random.randint(0, len(test_labels) - 1)\n",
    "\n",
    "print(f'Sentence: {test_tweets[i]}')\n",
    "print(f'Emotion: {index_to_classes[test_labels[i]]}')\n",
    "\n",
    "p = np.argmax(model.predict(np.expand_dims(test_sequences[i], axis=0)), axis=-1)[0]\n",
    "\n",
    "print(f'Predicted Emotion: {index_to_classes.get(p)}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "63/63 [==============================] - 1s 16ms/step\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "((2000,), (2000,))"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "preds = np.argmax(model.predict(test_sequences), axis=-1)\n",
    "preds.shape, test_labels.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x800 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "show_confusion_matrix(test_labels, preds, list(classes))"
   ]
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "name": "Tweet Sentiment Classification: Complete.ipynb",
   "provenance": []
  },
  "kernelspec": {
   "display_name": "tweet_sentiment",
   "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": "3.10.8"
  },
  "vscode": {
   "interpreter": {
    "hash": "87cefcf3401cbf8070ad802961843d4669156efa18d87856a0e10e1ec5a04c33"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}