Bi_LSTM_CNN_DistilBERT.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QZuMKqDSGeDO"
      },
      "source": [
        "**M909 : Deep Learning for NLP |\n",
        "Winter Semester 2023/24** \n",
        "\n",
        "**Assignments 2, 3, 4 : Code and Report**\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Vzh30uW7Es_w"
      },
      "source": [
        "**SENTIMENT ANALYSIS OF THE IMDB DATASET**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rWzpP73UE1H_"
      },
      "source": [
        "The IMDB Dataset consists of 50K movie reviews. The aim of these assignments is to apply different neural network architectures for classifying whether a review is positive or negative.\n",
        "\n",
        "Experiment_setting_1:Freeze embedding layers on LSTM and CNN architectures\n",
        "Experiment_setting_2:Unfreeze embedding layers on LSTM and CNN architectures\n",
        "\n",
        "For hyperparameter tuning the Optuna optimzation framework is used\n",
        "\n",
        "This project consists of 8 parts:\n",
        "* Load the data\n",
        "* Preprocess the data and split\n",
        "* Build the vocabulary\n",
        "* Hyperparameter tuning for the LSTM and CNN model architectures\n",
        "* Training and evaluation of models without Glove pre-trained embeddings (frozen embeddings)\n",
        "* Training and evaluation of models with Glove pre-trained embeddings\n",
        "* Use a model from Hugging Face and fine tune it on the IMDB dataset\n",
        "* Presenting the results from the models and comment on the findings"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ke_U9Ko0w95R"
      },
      "source": [
        "**DOWNLOAD AND IMPORT NECESSARY PACKAGES AND LIBRARIES**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "nDwCNkY9Xl4V"
      },
      "outputs": [],
      "source": [
        "!pip install transformers torchtext==0.6.0\n",
        "!pip install accelerate -U\n",
        "!pip install optuna"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "DFMTsOyIzZZm"
      },
      "outputs": [],
      "source": [
        "import torch\n",
        "from torchtext import data, datasets\n",
        "import random\n",
        "import re\n",
        "import spacy\n",
        "import time\n",
        "import matplotlib.pyplot as plt\n",
        "import pandas as pd\n",
        "import math\n",
        "import numpy as np\n",
        "import nltk\n",
        "import numpy as np\n",
        "from tqdm import tqdm\n",
        "from nltk.corpus import stopwords\n",
        "from nltk.tokenize import word_tokenize\n",
        "from nltk.stem import WordNetLemmatizer\n",
        "import collections\n",
        "from sklearn.utils import resample\n",
        "from sklearn import metrics\n",
        "from sklearn.metrics import precision_recall_fscore_support as score\n",
        "from sklearn.metrics import accuracy_score, f1_score, precision_score, recall_score\n",
        "from sklearn.metrics import classification_report\n",
        "from torch.utils.data import TensorDataset\n",
        "from torchtext.data import Field, TabularDataset, BucketIterator, Iterator\n",
        "import torch.nn as nn\n",
        "import torchtext\n",
        "import torch.nn.functional as F\n",
        "import torch.optim as optim\n",
        "import os\n",
        "from sklearn.model_selection import train_test_split\n",
        "from transformers import EarlyStoppingCallback\n",
        "from transformers import DistilBertTokenizerFast\n",
        "from transformers import DistilBertForSequenceClassification, Trainer, TrainingArguments\n",
        "import optuna"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "WGygZ6DaeYPn"
      },
      "outputs": [],
      "source": [
        "# Download necessary packages and datasets\n",
        "nltk.download('stopwords')\n",
        "nltk.download('punkt')\n",
        "nltk.download('wordnet')\n",
        "stop_words_set = set(stopwords.words('english'))\n",
        "\n",
        "# Download IMDB dataset\n",
        "URL = 'http://ai.stanford.edu/~amaas/data/sentiment/aclImdb_v1.tar.gz'\n",
        "!wget -P 'Data/' $URL\n",
        "!tar xvzf 'Data/aclImdb_v1.tar.gz' -C 'Data/'\n",
        "\n",
        "# Download and unzip Glove pre-trained vectors\n",
        "FILE = 'GLOVE'\n",
        "URL = 'http://nlp.stanford.edu/data/glove.6B.zip'\n",
        "!wget -P $FILE $URL\n",
        "!unzip $FILE/glove*.zip -d $FILE"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "suQnUqZ_xhux"
      },
      "source": [
        "**SET UP GPU FOR TRAINING**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "vnsM_kzK1qgw"
      },
      "outputs": [],
      "source": [
        "# Set up GPU for Training\n",
        "if torch.cuda.is_available():\n",
        "    device = torch.device(\"cuda\")\n",
        "    print(f'There are {torch.cuda.device_count()} GPU(s) available.')\n",
        "    print('Device name:', torch.cuda.get_device_name(0))\n",
        "\n",
        "else:\n",
        "    print('No GPU available, using the CPU instead.')\n",
        "    device = torch.device('cpu')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "j-WdhMDKxlzB"
      },
      "source": [
        "**FUNCTIONS USED**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "yU9CRji5J47M"
      },
      "outputs": [],
      "source": [
        "# A function that calculates the time needed for training and evaluation in each epoch\n",
        "def epoch_time(start_time, end_time):\n",
        "    elapsed_time = end_time - start_time\n",
        "    elapsed_mins = int(elapsed_time / 60)\n",
        "    elapsed_secs = int(elapsed_time - (elapsed_mins * 60))\n",
        "    return elapsed_mins, elapsed_secs"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "qwM-0dtr6XhJ"
      },
      "outputs": [],
      "source": [
        "# A function that reads a set of files in a given directory and saves the lines to a list\n",
        "def load_data(path, file_list, dataset, encoding='utf8'):\n",
        "    for file in file_list:\n",
        "        with open(os.path.join(path, file), 'r', encoding=encoding) as text:\n",
        "            dataset.append(text.read())"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "WbRj03hDBzd9"
      },
      "outputs": [],
      "source": [
        "# A function that opens, reads a file in csv with utf-8 encoding, gets the values, checks and fills empty values and checks for duplicates\n",
        "def read_explore_dataframe(dataframe):\n",
        "\n",
        "  # Count the number of examples in each class\n",
        "  print(dataframe.label.value_counts())\n",
        "  print()\n",
        "\n",
        "  # Get the keys and their corresponding number of values\n",
        "  keys = dataframe.keys()\n",
        "  for key in keys:\n",
        "    df_len = len(dataframe[key].unique()) # the length of the unique values of each column\n",
        "    print('{0:25}{1:10}'.format(key,df_len))\n",
        "\n",
        "  dataframe.dropna(inplace=True) # remove missing values and keep the dataFrame with valid entries in the same variable\n",
        "  dataframe.reset_index(inplace=True,drop=True)\n",
        "\n",
        "  dataframe.duplicated() # check for duplicates\n",
        "  #print('Number of duplicates in the dataframe:', dataframe.duplicated().sum())\n",
        "  print('Dataframe checked for NaN values and duplicates!')\n",
        "\n",
        "  return dataframe"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "aLuSjttVDV_t"
      },
      "outputs": [],
      "source": [
        "# A function that takes as parameter a text and applies preprocessing steps to clean it\n",
        "def text_preprocessing(text):\n",
        "\n",
        "  # Remove tags\n",
        "  TAG_RE = re.compile(r'<[^>]+>')\n",
        "  no_tags = TAG_RE.sub('',text)\n",
        "  # Remove unusual characters\n",
        "  text = re.sub('<[^>]*>', '', no_tags)\n",
        "  # Remove urls\n",
        "  text = re.sub(r'http\\S+','', text)\n",
        "  # Remove emoticons\n",
        "  emoticons = re.findall('(?::|;|=)(?:-)?(?:\\)|\\(|D|P)', text)\n",
        "  # Remove punctuation\n",
        "  no_punct = [word for word in emoticons if word.isalpha()]\n",
        "  # Convert all words to lowercase\n",
        "  text = re.sub('[\\W]+', ' ', text.lower()) + \" \".join(no_punct).replace('-', '')\n",
        "  # Remove numbers\n",
        "  no_numbers_txt = re.sub(r'\\d+', '', text)\n",
        "  # Tokenize\n",
        "  tokens = nltk.word_tokenize(no_numbers_txt)\n",
        "  # Filter out stopwords\n",
        "  no_stopwords = [word for word in tokens if word not in stop_words_set]\n",
        "  # Lemmatize\n",
        "  lemmatizer = WordNetLemmatizer()\n",
        "  lemmas = \" \".join([lemmatizer.lemmatize(token) for token in no_stopwords])\n",
        "\n",
        "  return lemmas"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "OOra4QZADrCV"
      },
      "outputs": [],
      "source": [
        "# A function that gets as input the dataframe and gets the necessary values from the columns after resampling and pre-processing\n",
        "def get_columns(dataframe, feature, label):\n",
        "\n",
        "  # Get only the columns we are interested in\n",
        "  dataset = dataframe[[feature,label]]\n",
        "\n",
        "  # Apply the pre-processing function to the dataframe containing the text (feature column)\n",
        "  dataset[feature] = dataset[feature].apply(text_preprocessing)\n",
        "  print('\\nText done pre-processing!')\n",
        "\n",
        "  # Divide the X and y values\n",
        "  X = dataset[feature]\n",
        "  y = dataset[label]\n",
        "\n",
        "  # Split train dataset to train and validation set\n",
        "  X_train, X_val, Y_train, Y_val = train_test_split(X, y, test_size = 0.20, random_state = 42)\n",
        "  print('The dataset is split into training and validation sets!')\n",
        "\n",
        "  print('Shape of x_train: ', X_train.shape)\n",
        "  print('Shape of y_train: ', Y_train.shape)\n",
        "  print('Shape of x_val:  ', X_val.shape)\n",
        "  print('Shape of y_val:  ', Y_val.shape)\n",
        "\n",
        "  # Concatenate the X,Y data of training and validation data\n",
        "  train = pd.concat([X_train, Y_train], axis=1)\n",
        "  valid = pd.concat([X_val, Y_val], axis=1)\n",
        "\n",
        "  # Save the training and validation data into csv format\n",
        "  train.to_csv(\"/content/train.csv\", index=False)\n",
        "  valid.to_csv(\"/content/valid.csv\", index=False)\n",
        "\n",
        "  return train, valid"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "kJbSz45x8cJL"
      },
      "outputs": [],
      "source": [
        "# A function that creates iterators for the given datasets\n",
        "# In each batch, the sentences are transposed to be descending vertically\n",
        "# Each index represents a token and each column represents a sentence\n",
        "# The batch size specifies the number of columns\n",
        "# The lambda function finds sentences of the same length,\n",
        "#which is beneficial for efficient processing, especially in the context of LSTM models where sequences need to be of the same length.The sorting helps in padding sequences to similar lengths\n",
        "\n",
        "def create_iterators(train_dataset, validation_dataset, test_dataset, batch_size):\n",
        "\n",
        "  train_iterator, val_iterator, test_iterator = data.BucketIterator.splits(\n",
        "  # The Datasets to load Examples from\n",
        "  (train_dataset, validation_dataset, test_dataset),\n",
        "  # The batch size\n",
        "  batch_size = batch_size,\n",
        "  # Function to use for sorting examples\n",
        "  sort_key = lambda x: len(x.Review),\n",
        "  # Repeat the iterator for multiple epochs\n",
        "  repeat = False,\n",
        "  # Shuffle data on each epoch run\n",
        "  shuffle = True,\n",
        "  # Use 'sort_key' to sort examples in each batch: True ensures that sentences within a batch are of similar length, optimizing for efficient processing\n",
        "  sort_within_batch = True,\n",
        "  device = device)\n",
        "\n",
        "  return train_iterator, val_iterator, test_iterator"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "KmsVVRSENegX"
      },
      "outputs": [],
      "source": [
        "# A function that calculates the trainable parameters of the model\n",
        "def count_parameters(model):\n",
        "    return sum(p.numel() for p in model.parameters() if p.requires_grad)\n",
        "\n",
        "# A function that sets seed for reproducibility\n",
        "def set_seed(seed_value):\n",
        "    random.seed(seed_value)\n",
        "    np.random.seed(seed_value)\n",
        "    torch.manual_seed(seed_value)\n",
        "    torch.cuda.manual_seed_all(seed_value)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "jGr8WG1N5Ik5"
      },
      "outputs": [],
      "source": [
        "# ====== METRICS CALCULATION ======\n",
        "\n",
        "# A function that calculates accuracy per batch and returns rounded predictions\n",
        "def binary_accuracy(preds, y):\n",
        "\n",
        "    # Round predictions to the closest integer\n",
        "    rounded_preds = torch.round(torch.sigmoid(preds))\n",
        "    correct = (rounded_preds == y).float() # Convert into float for division\n",
        "    acc = correct.sum() / len(correct)\n",
        "    return acc\n",
        "\n",
        "\n",
        "# A function that calculates all the metrics using the testing set\n",
        "def calculate_metrics(y_true, preds):\n",
        "  print('\\nCalculating metrics...')\n",
        "  # Calculate the accuracy of the model\n",
        "  acc = accuracy_score(y_true, preds)\n",
        "  # Calculate general precision, recall, F1 score of the model and of each class\n",
        "  precision, recall, fscore, support = score(y_true, preds)\n",
        "  print('Accuracy: {}'.format(acc))\n",
        "  print('Precision: {}'.format(precision))\n",
        "  print('Recall: {}'.format(recall))\n",
        "  print('F1 score: {}'.format(fscore))\n",
        "  print('Support: {}'.format(support))\n",
        "\n",
        "  # Print classification report\n",
        "  print(classification_report(y_true, preds))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "lxNRFdJ2Ns7_"
      },
      "outputs": [],
      "source": [
        "# ====== TRAINING (TRAINING SET) ======\n",
        "def train(model, iterator, optimizer, criterion, clip):\n",
        "\n",
        "  training_loss = 0\n",
        "  training_acc = 0\n",
        "\n",
        "  # Set model to training mode\n",
        "  model.train()\n",
        "\n",
        "  # For each batch in the training iterator\n",
        "  for batch in tqdm(iterator):\n",
        "\n",
        "    # Zero the gradients\n",
        "    optimizer.zero_grad()\n",
        "\n",
        "    # Compute the predictions\n",
        "    predictions = model(batch.Review).squeeze(1)\n",
        "\n",
        "    # Compute the loss\n",
        "    loss = criterion(predictions, batch.Label)\n",
        "\n",
        "    # Compute the accuracy\n",
        "    acc = binary_accuracy(predictions, batch.Label.float())\n",
        "\n",
        "    # Use loss to compute gradients\n",
        "    loss.backward()\n",
        "\n",
        "    # Clip the gradient value\n",
        "    nn.utils.clip_grad_norm_(model.parameters(), clip)\n",
        "\n",
        "    # Use optimizer to take gradient step\n",
        "    optimizer.step()\n",
        "\n",
        "    training_loss += loss.item()\n",
        "    training_acc += acc.item()\n",
        "\n",
        "  loss_of_epoch = training_loss / len(iterator)\n",
        "  accuracy_of_epoch = training_acc / len(iterator)\n",
        "\n",
        "  # Return the loss and accuracy, averaged across each epoch\n",
        "  # len of iterator = number of batches in the iterator\n",
        "  return loss_of_epoch, accuracy_of_epoch"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "wOI6_OSxNzI4"
      },
      "outputs": [],
      "source": [
        "# ====== EVALUATION (VALIDATION SET) ======\n",
        "def evaluate(model, iterator, criterion):\n",
        "\n",
        "  eval_loss = 0\n",
        "  eval_acc = 0\n",
        "\n",
        "  # Set model to evaluation mode\n",
        "  model.eval()\n",
        "\n",
        "  # Don't calculate the gradients\n",
        "  with torch.no_grad():\n",
        "\n",
        "    for batch in tqdm(iterator):\n",
        "\n",
        "      # Compute the predictions\n",
        "      predictions = model(batch.Review).squeeze(1)\n",
        "\n",
        "      # Compute the loss\n",
        "      loss = criterion(predictions, batch.Label)\n",
        "\n",
        "      # Compute the accuracy\n",
        "      acc = binary_accuracy(predictions, batch.Label.float())\n",
        "\n",
        "      eval_loss += loss.item()\n",
        "      eval_acc += acc.item()\n",
        "\n",
        "  loss_of_epoch = eval_loss / len(iterator)\n",
        "  accuracy_of_epoch = eval_acc / len(iterator)\n",
        "\n",
        "  # Return the loss and accuracy, averaged across each epoch\n",
        "  # len of iterator = number of batches in the iterator\n",
        "  return loss_of_epoch, accuracy_of_epoch\n",
        "\n",
        "\n",
        "# ====== EVALUATION (TESTING SET) ======\n",
        "def testing(model, iterator):\n",
        "\n",
        "  predictions_all = []\n",
        "  labels_all = []\n",
        "\n",
        "  model.eval()\n",
        "\n",
        "  with torch.no_grad():\n",
        "\n",
        "    for batch in iterator:\n",
        "      predictions = model(batch.Review).squeeze(1)\n",
        "      predictions_all +=  torch.round(torch.sigmoid(predictions)).flatten().cpu().numpy().tolist()\n",
        "      labels_all += batch.Label.flatten().cpu().numpy().tolist()\n",
        "\n",
        "  return predictions_all, labels_all"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uBMUH3WBoUr2"
      },
      "source": [
        "**MODELS**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "-C_Tq8Id1Sug"
      },
      "outputs": [],
      "source": [
        "# ====== Recurrent Neural Network Model with LSTM ======\n",
        "class LSTM_Model(nn.Module):\n",
        "\n",
        "  def __init__(self, vocab_size, embedding_dim, hidden_dim, output_dim, num_layers, d_rate, pad_idx):\n",
        "\n",
        "    super(LSTM_Model, self).__init__()\n",
        "    self.embedding = nn.Embedding(vocab_size, embedding_dim, padding_idx = pad_idx)\n",
        "    self.lstm = nn.LSTM(embedding_dim, hidden_dim, num_layers=num_layers, bidirectional=True, dropout=d_rate)\n",
        "    self.dense = nn.Linear(2 * hidden_dim, output_dim)\n",
        "    self.dropout = nn.Dropout(d_rate)\n",
        "\n",
        "  def forward(self, x):\n",
        "    # Get the word embeddings of the batch\n",
        "    # x: (sentence_length, batch_size)\n",
        "    embedded = self.dropout(self.embedding(x))\n",
        "    # embedded: (sentence_length, batch_size, embedding_dim)\n",
        "\n",
        "    # Propagate the input through LSTM layer/s\n",
        "    _, (hidden, _) = self.lstm(embedded)\n",
        "\n",
        "    # Extract output of the last time step\n",
        "    # Concat the final output of forward direction and backward direction\n",
        "    hidden = self.dropout(torch.cat((hidden[-2,:,:], hidden[-1,:,:]), dim=1))\n",
        "    # hidden: (batch_size, hidden_dim * 2)\n",
        "\n",
        "    output = self.dense(hidden)\n",
        "\n",
        "    return output"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "8n6Se6QT0xUF"
      },
      "outputs": [],
      "source": [
        "# ====== Convolutional Neural Network Model ======\n",
        "class CNN(nn.Module):\n",
        "\n",
        "  def __init__(self, vocab_size, embedding_dim, n_filters, filter_sizes, output_dim, dropout, pad_idx):\n",
        "    super(CNN,self).__init__()\n",
        "    #Embedding layer\n",
        "    self.embedding = nn.Embedding(vocab_size, embedding_dim, padding_idx = pad_idx)\n",
        "    #Conv Network\n",
        "    # in_channels: the number of \"channels\" in the image going into the convolutional layer\n",
        "    # out_channels: the number of filters\n",
        "    # the kernel_size: the size of the filters. Each of the kernel_sizes is going to be [n x emb_dim] where n is the size of the n-grams\n",
        "    self.convs = nn.ModuleList([nn.Conv2d(in_channels = 1, out_channels = n_filters, kernel_size = (fs, embedding_dim)) for fs in filter_sizes])\n",
        "    #Fully  connected +Dropout\n",
        "    self.fc = nn.Linear(len(filter_sizes) * n_filters, output_dim)\n",
        "    self.dropout = nn.Dropout(dropout)\n",
        "\n",
        "  def forward(self, text):\n",
        "    #input text is permuted to have the batch size as the first dimension\n",
        "    text = text.permute(1, 0)\n",
        "    #text is embedded, and dropout is applied\n",
        "    embedded = self.dropout(self.embedding(text))\n",
        "    embedded = embedded.unsqueeze(1)\n",
        "    #Convolution and ReLU activation are applied to each filter size.\n",
        "    conved = [F.relu(conv(embedded)).squeeze(3) for conv in self.convs]\n",
        "    #1D max pooling is performed over the width of each feature map\n",
        "    pooled = [F.max_pool1d(conv, conv.shape[2]).squeeze(2) for conv in conved]\n",
        "    #Pooled feature maps are concatenated and passed through dropout\n",
        "    cat = self.dropout(torch.cat(pooled, dim = 1))\n",
        "    #result is passed through the fully connected layer to get the final output\n",
        "    return self.fc(cat)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "iSl9hO3yX03w"
      },
      "source": [
        "**RUN THE SCRIPT - APPLY FUNCTIONS**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MWoVC-87YDp4"
      },
      "source": [
        "**PART 1: LOAD THE DATA**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "DF4vJNM96YpB"
      },
      "outputs": [],
      "source": [
        "set_seed(42)\n",
        "\n",
        "# Path to dataset location\n",
        "path = '/content/Data/aclImdb'\n",
        "\n",
        "# Create lists that will contain read lines\n",
        "train_pos, train_neg, test_pos, test_neg = [], [], [], []\n",
        "\n",
        "# Create a dictionary of paths and lists that store lines (key: value = path: list)\n",
        "sets_dict = {'train/pos/': train_pos, 'train/neg/': train_neg,\n",
        "             'test/pos/': test_pos, 'test/neg/': test_neg}\n",
        "\n",
        "# Load the data\n",
        "for dataset in sets_dict:\n",
        "        file_list = [f for f in os.listdir(os.path.join(path, dataset)) if f.endswith('.txt')]\n",
        "        load_data(os.path.join(path, dataset), file_list, sets_dict[dataset])\n",
        "\n",
        "# Concatenate positive and negative into datasets, one for the training and one for the testing\n",
        "train_dataset = pd.concat([pd.DataFrame({'review': train_pos, 'label':1}),\n",
        "                          pd.DataFrame({'review': train_neg, 'label':0})],\n",
        "                          axis=0, ignore_index=True)\n",
        "\n",
        "test_dataset = pd.concat([pd.DataFrame({'review': test_pos, 'label':1}),\n",
        "                          pd.DataFrame({'review': test_neg, 'label':0})],\n",
        "                          axis=0, ignore_index=True)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 206
        },
        "id": "olrm854A8VrD",
        "outputId": "7bcb3a3a-8b6a-4bd0-cf7e-1521f7e5f70b"
      },
      "outputs": [
        {
          "data": {
            "text/html": [
              "\n",
              "  <div id=\"df-0a93f74b-1eda-4799-9997-aa33d22bee31\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>review</th>\n",
              "      <th>label</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>PRC which was the lowest of the low actually s...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>Considering it was made on a low budget, THE D...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>this movie is so complex that it can be given ...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>Rented and watched this short (&lt; 90 minutes) w...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>Dressed to Kill starts off with Kate Miller (A...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-0a93f74b-1eda-4799-9997-aa33d22bee31')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-0a93f74b-1eda-4799-9997-aa33d22bee31 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-0a93f74b-1eda-4799-9997-aa33d22bee31');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-d8be46bc-2c3a-4cdd-ba81-7810d1fb5dc5\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-d8be46bc-2c3a-4cdd-ba81-7810d1fb5dc5')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-d8be46bc-2c3a-4cdd-ba81-7810d1fb5dc5 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "text/plain": [
              "                                              review  label\n",
              "0  PRC which was the lowest of the low actually s...      1\n",
              "1  Considering it was made on a low budget, THE D...      1\n",
              "2  this movie is so complex that it can be given ...      1\n",
              "3  Rented and watched this short (< 90 minutes) w...      1\n",
              "4  Dressed to Kill starts off with Kate Miller (A...      1"
            ]
          },
          "execution_count": 18,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "train_dataset.head()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JR9S59ZjYVkI"
      },
      "source": [
        "**PART 2: PREPROCESS THE DATA AND SPLIT**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8iticRxxCCnt",
        "outputId": "95b79eba-791e-4fe0-8712-3d748ec1bb9d"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "------ Preparing the training and validation dataset... ------\n",
            "1    12500\n",
            "0    12500\n",
            "Name: label, dtype: int64\n",
            "\n",
            "review                        24904\n",
            "label                             2\n",
            "Dataframe checked for NaN values and duplicates!\n",
            "\n",
            "Text done pre-processing!\n",
            "The dataset is split into training and validation sets!\n",
            "Shape of x_train:  (20000,)\n",
            "Shape of y_train:  (20000,)\n",
            "Shape of x_val:   (5000,)\n",
            "Shape of y_val:   (5000,)\n"
          ]
        }
      ],
      "source": [
        "# ====== Prepare the datasets ======\n",
        "print('------ Preparing the training and validation dataset... ------')\n",
        "\n",
        "# Opens, reads, checks for duplicates and removes missing values\n",
        "train_df = read_explore_dataframe(train_dataset)\n",
        "\n",
        "# Gets the necessary columns, resamples, preprocesses the texts, splits into training and validation sets and saves them to csv\n",
        "train_dataframe, valid_dataframe = get_columns(train_df, 'review', 'label')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "zWWF8e1J7czI",
        "outputId": "02cdebc5-5eab-46ad-b6bd-e18084ad74df"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "------ Preparing the testing dataset... ------\n",
            "1    12500\n",
            "0    12500\n",
            "Name: label, dtype: int64\n",
            "\n",
            "review                        24801\n",
            "label                             2\n",
            "Dataframe checked for NaN values and duplicates!\n",
            "\n",
            "Text done pre-processing!\n",
            "The testing dataset is ready and saved into csv format!\n"
          ]
        }
      ],
      "source": [
        "# ====== Prepare the datasets ======\n",
        "print('------ Preparing the testing dataset... ------')\n",
        "\n",
        "# Opens, reads, checks for duplicates and removes missing values\n",
        "test_df = read_explore_dataframe(test_dataset)\n",
        "\n",
        "# Apply the function for preprocessing the text\n",
        "test_df['review'] = test_df['review'].apply(text_preprocessing)\n",
        "print('\\nText done pre-processing!')\n",
        "\n",
        "# Divide the text and the labels\n",
        "test_texts = test_df['review']\n",
        "test_labels = test_df['label']\n",
        "\n",
        "# Concatenate the text and labels of the clean dataset into a new dataframe\n",
        "test_dataframe = pd.concat([test_texts, test_labels], axis=1)\n",
        "\n",
        "# Save the testing data into csv format\n",
        "test_dataframe.to_csv(\"/content/test.csv\", index=False)\n",
        "print('The testing dataset is ready and saved into csv format!')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AAJexT2NYgTG"
      },
      "source": [
        "**PART 3: BUILD THE VOCABULARY**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "iop_k6CmNOY4",
        "outputId": "71187121-37da-4637-965b-dfb100826342"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Number of training examples: 20000\n",
            "Number of valid examples: 5000\n",
            "Number of testing examples: 25000\n",
            "\n",
            "Tokens in TEXT vocabulary: 25002\n",
            "Tokens in LABEL vocabulary: 2\n"
          ]
        }
      ],
      "source": [
        "# ====== Load the data & Build the vocabulary ======\n",
        "\n",
        "set_seed(42)\n",
        "\n",
        "# Create the TEXT and LABEL fields\n",
        "spacy_en = spacy.load('en_core_web_sm')\n",
        "TEXT = data.Field(tokenize=lambda text: [tok.text for tok in spacy_en.tokenizer(text)], lower=True, sequential=True)\n",
        "LABEL = data.LabelField(dtype = torch.float, sequential = False)\n",
        "\n",
        "# Loading training, testing and validation data\n",
        "train_data, valid_data, test_data = data.TabularDataset.splits(\n",
        "    path ='/content/', train = 'train.csv',\n",
        "    validation = 'valid.csv', test = 'test.csv', format = 'csv', skip_header = True,\n",
        "    fields = [('Review', TEXT), ('Label', LABEL)])\n",
        "\n",
        "print(f'Number of training examples: {len(train_data)}')\n",
        "print(f'Number of valid examples: {len(valid_data)}')\n",
        "print(f'Number of testing examples: {len(test_data)}')\n",
        "\n",
        "# Build the vocabulary using the train data\n",
        "TEXT.build_vocab(train_data, vectors = torchtext.vocab.Vectors('/content/GLOVE/glove.6B.100d.txt'), max_size = 25000, unk_init=torch.Tensor.normal_)\n",
        "LABEL.build_vocab(train_data)\n",
        "\n",
        "print('\\nTokens in TEXT vocabulary: {}'.format(len(TEXT.vocab)))\n",
        "print('Tokens in LABEL vocabulary: {}'.format(len(LABEL.vocab)))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "SjcviMp-ytmX",
        "outputId": "86a7c0d5-75dc-4faf-f733-d9a396e1c62b"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[('movie', 41450), ('film', 37789), ('one', 22111), ('like', 16614), ('time', 12670), ('good', 12162), ('character', 11308), ('story', 10604), ('even', 10191), ('would', 10000), ('get', 9966), ('make', 9758), ('see', 9631), ('really', 9435), ('well', 8671), ('scene', 8493), ('much', 7778), ('people', 7545), ('bad', 7482), ('first', 7360)]\n",
            "['<unk>', '<pad>', 'movie', 'film', 'one', 'like', 'time', 'good', 'character', 'story']\n",
            "defaultdict(None, {'0': 0, '1': 1})\n"
          ]
        }
      ],
      "source": [
        "print(TEXT.vocab.freqs.most_common(20))\n",
        "print(TEXT.vocab.itos[:10])\n",
        "print(LABEL.vocab.stoi)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9nYB4I4zo7y-"
      },
      "source": [
        "**PART 4: Hyperparameter tuning with Optuna framework**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8J4qvh7cDh4G"
      },
      "source": [
        "#Defining Optuna search space for RNN_LSTM, and picking the chosen parameters as per Otuna best trial:\n",
        "1. The number of layers (i.e number of stacked RNNs): 2-5 (4 was chosen)\n",
        "2. Number of epochs: 5-50 (17 was chosen)\n",
        "3. Hidden neurons: 10-100 (16 neurons were chosen)\n",
        "4. Dropout probability : 0.2-0.7 (0.6 was chosen)\n",
        "5. learing rate: 0.0001 - 0.01 (0.002 was chosen)\n",
        "6. batch size: 16-128 (118 was chosen)\n",
        "7. gradient clipping to avoid exploding gradients: 1-5 (4 was chosen)\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 31,
      "metadata": {
        "id": "7JkhVjzto-U1"
      },
      "outputs": [],
      "source": [
        "def RNN_LSTM(TEXT, data_train, data_valid, data_test, hidden_dim, num_layers, dropout, learning_rate, num_epochs, batch_size, clip, freeze_embeddings=True):\n",
        "    try:\n",
        "        # Set the device\n",
        "        device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
        "\n",
        "        # The number of words in the vocabulary\n",
        "        input_dim = len(TEXT.vocab)\n",
        "        # The dimensions of each Glove vector\n",
        "        embedding_dim = 100\n",
        "        # The number of output layers\n",
        "        output_dim = 1\n",
        "        # The Glove pre-trained embeddings\n",
        "        embeddings = TEXT.vocab.vectors\n",
        "        # Get pad token index from vocab\n",
        "        PAD_IDX = TEXT.vocab.stoi[TEXT.pad_token]\n",
        "\n",
        "        # Instantiate the model\n",
        "        model = LSTM_Model(input_dim, embedding_dim, hidden_dim, output_dim, num_layers, dropout, PAD_IDX)\n",
        "        print(model)\n",
        "        print(f'The model has {count_parameters(model):,} trainable parameters\\n')\n",
        "        model.to(device)\n",
        "\n",
        "        # Replace the initial weights of the embedding layer with the pre-trained embeddings\n",
        "        model.embedding.weight.data.copy_(embeddings)\n",
        "        # Initialize <unk> and <pad> both to all zeros\n",
        "        UNK_IDX = TEXT.vocab.stoi[TEXT.unk_token]\n",
        "\n",
        "        # Setting row in the embedding weights matrix to zero using the token index\n",
        "        model.embedding.weight.data[UNK_IDX] = torch.zeros(embedding_dim)\n",
        "        model.embedding.weight.data[PAD_IDX] = torch.zeros(embedding_dim)\n",
        "\n",
        "        # Initialize the model parameters\n",
        "        set_seed(42)\n",
        "\n",
        "        # Create the iterators\n",
        "        train_iterator, valid_iterator, test_iterator = create_iterators(data_train, data_valid, data_test, batch_size)\n",
        "\n",
        "        # Initialize the optimizer and the loss function\n",
        "        optimizer = torch.optim.Adam(model.parameters(), lr=learning_rate, weight_decay=1e-4)\n",
        "\n",
        "        # Define loss function\n",
        "        loss_function = nn.BCEWithLogitsLoss()\n",
        "        loss_function = loss_function.to(device)\n",
        "\n",
        "        best_valid_loss = float('inf')\n",
        "\n",
        "        # Initialize empty lists to store the values of loss and accuracy for each epoch\n",
        "        train_accuracies = []\n",
        "        train_losses = []\n",
        "        val_accuracies = []\n",
        "        val_losses = []\n",
        "\n",
        "        # Freeze or unfreeze the embeddings based on the parameter\n",
        "        model.embedding.weight.requires_grad = not freeze_embeddings\n",
        "\n",
        "\n",
        "        for epoch in range(num_epochs):\n",
        "            start_time = time.time()\n",
        "\n",
        "            train_loss, train_accuracy = train(model, train_iterator, optimizer, loss_function, clip)\n",
        "            validation_loss, validation_accuracy = evaluate(model, valid_iterator, loss_function)\n",
        "\n",
        "            end_time = time.time()\n",
        "\n",
        "            epoch_mins, epoch_secs = epoch_time(start_time, end_time)\n",
        "\n",
        "            print(f'Epoch: {epoch+1:02} | Epoch Time: {epoch_mins}m {epoch_secs}s | Frozen: {freeze_embeddings}')\n",
        "            print(f'\\tTrain Loss: {train_loss:.3f} | Train Acc: {train_accuracy*100:.2f}%')\n",
        "            print(f'\\t Val. Loss: {validation_loss:.3f} |  Val. Acc: {validation_accuracy*100:.2f}%')\n",
        "\n",
        "            if validation_loss < best_valid_loss:\n",
        "                best_valid_loss = validation_loss\n",
        "                torch.save(model.state_dict(), 'lstm_model_frozen.pt' if freeze_embeddings else 'lstm_model_unfrozen.pt')\n",
        "\n",
        "            # Fill the empty lists with the values of accuracy and loss for training and evaluation sets\n",
        "            train_losses.append(train_loss)\n",
        "            train_accuracies.append(train_accuracy)\n",
        "            val_losses.append(validation_loss)\n",
        "            val_accuracies.append(validation_accuracy)\n",
        "\n",
        "        # Plot the loss vs epochs\n",
        "        plt.plot(train_losses)\n",
        "        plt.plot(val_losses)\n",
        "        plt.title('Train vs Validation Model Loss')\n",
        "        plt.ylabel('Loss')\n",
        "        plt.xlabel('Epoch')\n",
        "        plt.legend(['Train', 'Validation'])\n",
        "        plt.show()\n",
        "\n",
        "        # Plot the accuracy vs epochs\n",
        "        plt.plot(train_accuracies)\n",
        "        plt.plot(val_accuracies)\n",
        "        plt.title('Train vs Validation Model Accuracy')\n",
        "        plt.ylabel('Accuracy')\n",
        "        plt.xlabel('Epoch')\n",
        "        plt.legend(['Train', 'Validation'])\n",
        "        plt.show()\n",
        "\n",
        "        # Evaluate model on the testing set\n",
        "        predictions, labels = testing(model, test_iterator)\n",
        "\n",
        "        # Calculate the evaluation metrics on the testing set\n",
        "        calculate_metrics(labels, predictions)\n",
        "\n",
        "        # Return the last validation loss for optimization\n",
        "        return {'val_loss': val_losses[-1]}\n",
        "\n",
        "    except Exception as e:\n",
        "        print(f\"An error occurred: {str(e)}\")\n",
        "        # Return a large value in case of an error\n",
        "        return {'val_loss': float('inf')}\n",
        "\n",
        "def lstm_objective(trial):\n",
        "    # Set up the trial-specific hyperparameters\n",
        "    hidden_dim = trial.suggest_int('hidden_dim', 10, 100)\n",
        "    dropout = trial.suggest_float('dropout', 0.2, 0.7)\n",
        "    learning_rate = trial.suggest_float('learning_rate', 1e-4, 1e-2)\n",
        "    num_layers = trial.suggest_int('num_layers', 2, 5)\n",
        "    num_epochs = trial.suggest_int('num_epochs', 5, 50)\n",
        "    batch_size = trial.suggest_int('batch_size', 16, 128)\n",
        "    clip = trial.suggest_float('clip', 0.1, 5.0)\n",
        "\n",
        "    result = RNN_LSTM(TEXT, train_data, valid_data, test_data,\n",
        "                                       hidden_dim=hidden_dim,\n",
        "                                       num_layers=num_layers,\n",
        "                                       dropout=dropout,\n",
        "                                       learning_rate=learning_rate,\n",
        "                                       num_epochs=num_epochs,\n",
        "                                       batch_size=batch_size,\n",
        "                                       clip=clip)\n",
        "\n",
        "    # Return the validation loss as the objective value\n",
        "    return result['val_loss']\n",
        "\n",
        "#study = optuna.create_study(direction='minimize')\n",
        "#study.optimize(lstm_objective, n_trials=50)\n",
        "#best_params = study.best_params\n",
        "#print(\"Best Hyperparameters:\", best_params)\n",
        "##Best Hyperparameters: {'hidden_dim': 16, 'dropout': 0.6022217632336822, 'learning_rate': 0.0025565376011509387, 'num_layers': 4, 'num_epochs': 17, 'batch_size': 118, 'clip': 4.281124290951043}\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hjuOve3Sve0o"
      },
      "source": [
        "#Defining Optuna search space for CNN, and picking the chosen parameters as per Optuna best trial:\n",
        "In order for the models' results to be comparable, tha same number of epochs, dropout rate,batch size, gradient clipping, learning rate were maintained like in the LSTM architecture. The only chosen hyperparameters for optuna optimization were the CNN_specific parameters, meaning: the number of filters and the filter sizes\n",
        "\n",
        "•\tOptimizer and loss function: I used the Adam optimizer with weight decay of 1e-4 and the BCEWithLogitsLoss loss function, since it was a binary classification task."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 32,
      "metadata": {
        "id": "vrzFSlsl3PuR"
      },
      "outputs": [],
      "source": [
        "def CNN_EMBEDDINGS(TEXT, data_train, data_valid, data_test, num_filters, filter_sizes,dropout, learning_rate, num_epochs, batch_size, clip,freeze_embeddings=True):\n",
        "    try:\n",
        "        # Set the device\n",
        "        device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')\n",
        "\n",
        "        # The number of words in the vocabulary\n",
        "        input_dim = len(TEXT.vocab)\n",
        "        # The dimensions of each Glove vector\n",
        "        embedding_dim = 100\n",
        "        # The number of output layers\n",
        "        output_dim = 1\n",
        "        # The Glove pre-trained embeddings\n",
        "        embeddings = TEXT.vocab.vectors\n",
        "        # Get pad token index from vocab\n",
        "        PAD_IDX = TEXT.vocab.stoi[TEXT.pad_token]\n",
        "\n",
        "        model = CNN(input_dim, embedding_dim, num_filters, filter_sizes, output_dim, dropout, PAD_IDX)\n",
        "        print(model)\n",
        "        print(f'The model has {count_parameters(model):,} trainable parameters')\n",
        "        model.to(device)\n",
        "\n",
        "        # Replace the initial weights of the embedding layer with the pre-trained embeddings\n",
        "        model.embedding.weight.data.copy_(embeddings)\n",
        "        # Initialize <unk> and <pad> both to all zeros - irrelevant for sentiment analysis\n",
        "        UNK_IDX = TEXT.vocab.stoi[TEXT.unk_token]\n",
        "\n",
        "        # Setting row in the embedding weights matrix to zero using the token index\n",
        "        model.embedding.weight.data[UNK_IDX] = torch.zeros(embedding_dim)\n",
        "        model.embedding.weight.data[PAD_IDX] = torch.zeros(embedding_dim)\n",
        "\n",
        "        set_seed(42)\n",
        "\n",
        "        # Create the iterators\n",
        "        train_iterator, valid_iterator, test_iterator = create_iterators(data_train, data_valid, data_test, batch_size)\n",
        "\n",
        "        # Initialize the optimizer and the loss function\n",
        "        optimizer = torch.optim.Adam(model.parameters(), lr=learning_rate, weight_decay=1e-4)\n",
        "\n",
        "        # Define loss function\n",
        "        loss_function = nn.BCEWithLogitsLoss()\n",
        "        loss_function = loss_function.to(device)\n",
        "\n",
        "        best_valid_loss = float('inf')\n",
        "\n",
        "        # Initialize empty lists to store the values of loss and accuracy for each epoch\n",
        "        train_accuracies = []\n",
        "        train_losses = []\n",
        "        val_accuracies = []\n",
        "        val_losses = []\n",
        "\n",
        "        # Freeze or unfreeze the embeddings based on the parameter\n",
        "        model.embedding.weight.requires_grad = not freeze_embeddings\n",
        "\n",
        "        for epoch in range(num_epochs):\n",
        "\n",
        "            start_time = time.time()\n",
        "\n",
        "            train_loss, train_accuracy = train(model, train_iterator, optimizer, loss_function, clip)\n",
        "            validation_loss, validation_accuracy = evaluate(model, valid_iterator, loss_function)\n",
        "\n",
        "            end_time = time.time()\n",
        "\n",
        "            epoch_mins, epoch_secs = epoch_time(start_time, end_time)\n",
        "\n",
        "            print(f'Epoch: {epoch+1:02} | Epoch Time: {epoch_mins}m {epoch_secs}s | Frozen: {not freeze_embeddings}')\n",
        "            print(f'\\tTrain Loss: {train_loss:.3f} | Train Acc: {train_accuracy*100:.2f}%')\n",
        "            print(f'\\t Val. Loss: {validation_loss:.3f} |  Val. Acc: {validation_accuracy*100:.2f}%')\n",
        "\n",
        "            if validation_loss < best_valid_loss:\n",
        "                best_valid_loss = validation_loss\n",
        "                torch.save(model.state_dict(), f'cnn_model{\"_frozen\" if not freeze_embeddings else \"_unfrozen\"}.pt')\n",
        "\n",
        "            # Fill the empty lists with the values of accuracy and loss for training and evaluation sets\n",
        "            train_losses.append(train_loss)\n",
        "            train_accuracies.append(train_accuracy)\n",
        "            val_losses.append(validation_loss)\n",
        "            val_accuracies.append(validation_accuracy)\n",
        "\n",
        "        # Plot the loss vs epochs\n",
        "        plt.plot(train_losses)\n",
        "        plt.plot(val_losses)\n",
        "        plt.title('Train vs Validation Model Loss')\n",
        "        plt.ylabel('Loss')\n",
        "        plt.xlabel('Epoch')\n",
        "        plt.legend(['Train', 'Validation'])\n",
        "        plt.show()\n",
        "\n",
        "        # Plot the accuracy vs epochs\n",
        "        plt.plot(train_accuracies)\n",
        "        plt.plot(val_accuracies)\n",
        "        plt.title('Train vs Validation Model Accuracy')\n",
        "        plt.ylabel('Accuracy')\n",
        "        plt.xlabel('Epoch')\n",
        "        plt.legend(['Train', 'Validation'])\n",
        "        plt.show()\n",
        "\n",
        "        # Evaluate model on the testing set\n",
        "        predictions, labels = testing(model, test_iterator)\n",
        "\n",
        "        # Calculate the evaluation metrics on the testing set\n",
        "        calculate_metrics(labels, predictions)\n",
        "\n",
        "        # Return the last validation loss for optimization\n",
        "        return {'val_loss': val_losses[-1]}\n",
        "\n",
        "    except Exception as e:\n",
        "        print(f\"An error occurred: {str(e)}\")\n",
        "        # Return a large value in case of an error\n",
        "        return {'val_loss': float('inf')}\n",
        "\n",
        "def cnn_objective(trial):\n",
        "\n",
        "    num_filters = trial.suggest_int('num_filters', 10, 50)\n",
        "    filter_sizes = [trial.suggest_int(f'filter_size_{i}', 3, 7) for i in range(3)]\n",
        "    result = CNN_EMBEDDINGS(TEXT, train_data, valid_data, test_data,\n",
        "                                   num_filters=num_filters,\n",
        "                                   filter_sizes=filter_sizes,\n",
        "                                   dropout=0.6,\n",
        "                                   learning_rate=0.002,\n",
        "                                   num_epochs=17,\n",
        "                                   batch_size=118,\n",
        "                                   clip=4\n",
        "                            )\n",
        "    # Return the validation loss as the objective value\n",
        "    return result['val_loss']\n",
        "\n",
        "# Create a study object and optimize hyperparameters\n",
        "#study = optuna.create_study(direction='minimize')\n",
        "#study.optimize(cnn_objective, n_trials=50)\n",
        "\n",
        "# Get the best hyperparameters\n",
        "#best_params = study.best_params\n",
        "#print(\"Best Hyperparameters:\", best_params)\n",
        "##Best hyperparameters: {'num_filters': 20, 'filter_size_0': 3, 'filter_size_1': 4, 'filter_size_2': 5}"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sk6mHKfUYuqH"
      },
      "source": [
        "**PART 5: TRAINING AND EVALUATION OF MODELS WITHOUT GLOVE PRE-TRAINED EMBEDDINGS (FROZEN EMBEDDINGS)**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 33,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "Md3IVoE2X1br",
        "outputId": "4ee164b6-8190-487f-bd75-4596deb58068"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "LSTM_Model(\n",
            "  (embedding): Embedding(25002, 100, padding_idx=1)\n",
            "  (lstm): LSTM(100, 16, num_layers=4, dropout=0.6, bidirectional=True)\n",
            "  (dense): Linear(in_features=32, out_features=1, bias=True)\n",
            "  (dropout): Dropout(p=0.6, inplace=False)\n",
            ")\n",
            "The model has 2,534,537 trainable parameters\n",
            "\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:02<00:00, 60.92it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 134.95it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 01 | Epoch Time: 0m 3s | Frozen: True\n",
            "\tTrain Loss: 0.658 | Train Acc: 58.20%\n",
            "\t Val. Loss: 0.558 |  Val. Acc: 73.91%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:01<00:00, 89.69it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 166.56it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 02 | Epoch Time: 0m 2s | Frozen: True\n",
            "\tTrain Loss: 0.582 | Train Acc: 70.73%\n",
            "\t Val. Loss: 0.489 |  Val. Acc: 78.30%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:01<00:00, 92.02it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 162.44it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 03 | Epoch Time: 0m 2s | Frozen: True\n",
            "\tTrain Loss: 0.544 | Train Acc: 73.62%\n",
            "\t Val. Loss: 0.447 |  Val. Acc: 80.81%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:01<00:00, 92.96it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 164.07it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 04 | Epoch Time: 0m 2s | Frozen: True\n",
            "\tTrain Loss: 0.512 | Train Acc: 75.94%\n",
            "\t Val. Loss: 0.425 |  Val. Acc: 82.39%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:01<00:00, 89.55it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 169.60it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 05 | Epoch Time: 0m 2s | Frozen: True\n",
            "\tTrain Loss: 0.488 | Train Acc: 77.24%\n",
            "\t Val. Loss: 0.410 |  Val. Acc: 83.06%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:02<00:00, 71.04it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 106.71it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 06 | Epoch Time: 0m 2s | Frozen: True\n",
            "\tTrain Loss: 0.478 | Train Acc: 77.81%\n",
            "\t Val. Loss: 0.465 |  Val. Acc: 79.02%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:02<00:00, 74.27it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 164.88it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 07 | Epoch Time: 0m 2s | Frozen: True\n",
            "\tTrain Loss: 0.470 | Train Acc: 78.41%\n",
            "\t Val. Loss: 0.385 |  Val. Acc: 83.88%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:01<00:00, 91.20it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 154.88it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 08 | Epoch Time: 0m 2s | Frozen: True\n",
            "\tTrain Loss: 0.471 | Train Acc: 78.74%\n",
            "\t Val. Loss: 0.480 |  Val. Acc: 77.63%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:01<00:00, 91.43it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 162.44it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 09 | Epoch Time: 0m 2s | Frozen: True\n",
            "\tTrain Loss: 0.457 | Train Acc: 79.07%\n",
            "\t Val. Loss: 0.380 |  Val. Acc: 83.15%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:01<00:00, 92.71it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 161.28it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 10 | Epoch Time: 0m 2s | Frozen: True\n",
            "\tTrain Loss: 0.463 | Train Acc: 78.82%\n",
            "\t Val. Loss: 0.363 |  Val. Acc: 84.57%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:01<00:00, 92.15it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 151.01it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 11 | Epoch Time: 0m 2s | Frozen: True\n",
            "\tTrain Loss: 0.440 | Train Acc: 80.04%\n",
            "\t Val. Loss: 0.354 |  Val. Acc: 84.97%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:02<00:00, 63.79it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 99.82it/s] \n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 12 | Epoch Time: 0m 3s | Frozen: True\n",
            "\tTrain Loss: 0.442 | Train Acc: 80.23%\n",
            "\t Val. Loss: 0.366 |  Val. Acc: 84.63%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:01<00:00, 88.64it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 164.68it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 13 | Epoch Time: 0m 2s | Frozen: True\n",
            "\tTrain Loss: 0.429 | Train Acc: 80.47%\n",
            "\t Val. Loss: 0.346 |  Val. Acc: 85.79%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:01<00:00, 92.03it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 156.76it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 14 | Epoch Time: 0m 2s | Frozen: True\n",
            "\tTrain Loss: 0.438 | Train Acc: 80.40%\n",
            "\t Val. Loss: 0.351 |  Val. Acc: 85.65%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:01<00:00, 90.87it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 163.24it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 15 | Epoch Time: 0m 2s | Frozen: True\n",
            "\tTrain Loss: 0.429 | Train Acc: 80.60%\n",
            "\t Val. Loss: 0.371 |  Val. Acc: 84.68%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:01<00:00, 90.92it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 167.38it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 16 | Epoch Time: 0m 2s | Frozen: True\n",
            "\tTrain Loss: 0.424 | Train Acc: 80.89%\n",
            "\t Val. Loss: 0.341 |  Val. Acc: 86.25%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:02<00:00, 82.64it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 109.06it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 17 | Epoch Time: 0m 2s | Frozen: True\n",
            "\tTrain Loss: 0.425 | Train Acc: 80.95%\n",
            "\t Val. Loss: 0.339 |  Val. Acc: 85.93%\n"
          ]
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\n",
            "Calculating metrics...\n",
            "Accuracy: 0.85584\n",
            "Precision: [0.88678261 0.82948148]\n",
            "Recall: [0.81584 0.89584]\n",
            "F1 score: [0.84983333 0.86138462]\n",
            "Support: [12500 12500]\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "         0.0       0.89      0.82      0.85     12500\n",
            "         1.0       0.83      0.90      0.86     12500\n",
            "\n",
            "    accuracy                           0.86     25000\n",
            "   macro avg       0.86      0.86      0.86     25000\n",
            "weighted avg       0.86      0.86      0.86     25000\n",
            "\n"
          ]
        },
        {
          "data": {
            "text/plain": [
              "{'val_loss': 0.33934337316557417}"
            ]
          },
          "execution_count": 33,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "# ====== RNN_LSTM_FROZEN_EMBEDDINGS Function ======\n",
        "#Best Hyperparameters: {'hidden_dim': 16, 'dropout': 0.6022217632336822, 'learning_rate': 0.0025565376011509387, 'num_layers': 4, 'num_epochs': 17, 'batch_size': 118, 'clip': 4.281124290951043}\n",
        "\n",
        "RNN_LSTM(TEXT, train_data, valid_data, test_data, hidden_dim=16, num_layers=4, dropout=0.6, learning_rate=0.002, num_epochs=17, batch_size=118, clip=4)\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "wlvn9QeeXycw",
        "outputId": "7154ceb0-0bd6-4a4f-d68c-75113bc4bc98"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "CNN(\n",
            "  (embedding): Embedding(25002, 100, padding_idx=1)\n",
            "  (convs): ModuleList(\n",
            "    (0): Conv2d(1, 20, kernel_size=(3, 100), stride=(1, 1))\n",
            "    (1): Conv2d(1, 20, kernel_size=(4, 100), stride=(1, 1))\n",
            "    (2): Conv2d(1, 20, kernel_size=(5, 100), stride=(1, 1))\n",
            "  )\n",
            "  (fc): Linear(in_features=60, out_features=1, bias=True)\n",
            "  (dropout): Dropout(p=0.6, inplace=False)\n",
            ")\n",
            "The model has 2,524,321 trainable parameters\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:02<00:00, 65.62it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 95.48it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 01 | Epoch Time: 0m 3s | Frozen: True\n",
            "\tTrain Loss: 0.658 | Train Acc: 61.69%\n",
            "\t Val. Loss: 0.569 |  Val. Acc: 75.95%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:03<00:00, 46.99it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 54.85it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 02 | Epoch Time: 0m 4s | Frozen: True\n",
            "\tTrain Loss: 0.578 | Train Acc: 69.67%\n",
            "\t Val. Loss: 0.510 |  Val. Acc: 79.36%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:03<00:00, 56.27it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 93.93it/s] \n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 03 | Epoch Time: 0m 3s | Frozen: True\n",
            "\tTrain Loss: 0.561 | Train Acc: 71.34%\n",
            "\t Val. Loss: 0.493 |  Val. Acc: 80.57%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:01<00:00, 101.38it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 149.60it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 04 | Epoch Time: 0m 1s | Frozen: True\n",
            "\tTrain Loss: 0.551 | Train Acc: 72.79%\n",
            "\t Val. Loss: 0.486 |  Val. Acc: 81.18%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:01<00:00, 100.33it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 148.99it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 05 | Epoch Time: 0m 1s | Frozen: True\n",
            "\tTrain Loss: 0.540 | Train Acc: 73.19%\n",
            "\t Val. Loss: 0.472 |  Val. Acc: 81.51%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:01<00:00, 100.39it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 136.74it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 06 | Epoch Time: 0m 2s | Frozen: True\n",
            "\tTrain Loss: 0.534 | Train Acc: 73.61%\n",
            "\t Val. Loss: 0.461 |  Val. Acc: 81.79%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:01<00:00, 90.65it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 63.72it/s] \n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 07 | Epoch Time: 0m 2s | Frozen: True\n",
            "\tTrain Loss: 0.533 | Train Acc: 73.51%\n",
            "\t Val. Loss: 0.470 |  Val. Acc: 81.93%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:03<00:00, 45.85it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 148.27it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 08 | Epoch Time: 0m 4s | Frozen: True\n",
            "\tTrain Loss: 0.532 | Train Acc: 73.78%\n",
            "\t Val. Loss: 0.470 |  Val. Acc: 81.34%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:01<00:00, 103.32it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 146.16it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 09 | Epoch Time: 0m 1s | Frozen: True\n",
            "\tTrain Loss: 0.527 | Train Acc: 74.32%\n",
            "\t Val. Loss: 0.464 |  Val. Acc: 81.16%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:01<00:00, 102.37it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 144.23it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 10 | Epoch Time: 0m 1s | Frozen: True\n",
            "\tTrain Loss: 0.525 | Train Acc: 74.16%\n",
            "\t Val. Loss: 0.473 |  Val. Acc: 81.06%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:01<00:00, 102.16it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 146.04it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 11 | Epoch Time: 0m 1s | Frozen: True\n",
            "\tTrain Loss: 0.521 | Train Acc: 74.38%\n",
            "\t Val. Loss: 0.459 |  Val. Acc: 81.79%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:01<00:00, 104.57it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 144.91it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 12 | Epoch Time: 0m 1s | Frozen: True\n",
            "\tTrain Loss: 0.522 | Train Acc: 74.36%\n",
            "\t Val. Loss: 0.470 |  Val. Acc: 80.19%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:02<00:00, 82.29it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 105.97it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 13 | Epoch Time: 0m 2s | Frozen: True\n",
            "\tTrain Loss: 0.520 | Train Acc: 74.81%\n",
            "\t Val. Loss: 0.454 |  Val. Acc: 81.60%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:02<00:00, 79.76it/s] \n",
            "100%|██████████| 43/43 [00:00<00:00, 140.03it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 14 | Epoch Time: 0m 2s | Frozen: True\n",
            "\tTrain Loss: 0.519 | Train Acc: 74.80%\n",
            "\t Val. Loss: 0.452 |  Val. Acc: 81.49%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:01<00:00, 99.66it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 142.01it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 15 | Epoch Time: 0m 2s | Frozen: True\n",
            "\tTrain Loss: 0.519 | Train Acc: 74.71%\n",
            "\t Val. Loss: 0.463 |  Val. Acc: 81.06%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:01<00:00, 99.94it/s] \n",
            "100%|██████████| 43/43 [00:00<00:00, 136.34it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 16 | Epoch Time: 0m 2s | Frozen: True\n",
            "\tTrain Loss: 0.514 | Train Acc: 75.26%\n",
            "\t Val. Loss: 0.458 |  Val. Acc: 81.30%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:01<00:00, 99.58it/s] \n",
            "100%|██████████| 43/43 [00:00<00:00, 137.75it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 17 | Epoch Time: 0m 2s | Frozen: True\n",
            "\tTrain Loss: 0.518 | Train Acc: 74.76%\n",
            "\t Val. Loss: 0.460 |  Val. Acc: 80.95%\n"
          ]
        },
        {
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAkAAAAHHCAYAAABXx+fLAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAABuD0lEQVR4nO3dd3gU5d7G8e+m9wLpEHpvoQkC0gQFLGAFEaXYEQsiFo6C2I/1cFBfERTBgiIcUWwgImChKRBARHqHEFo6abvz/jFkISRASJtkc3+ua6/szs7O/maj2ZtnnmIzDMNAREREpApxs7oAERERkfKmACQiIiJVjgKQiIiIVDkKQCIiIlLlKACJiIhIlaMAJCIiIlWOApCIiIhUOQpAIiIiUuUoAImIiEiVowAkUo6GDx9OnTp1rC7DMrt378ZmszFjxgzntokTJ2Kz2Yr0epvNxsSJE0u1ph49etCjR49SPWZFV6dOHYYPH16s15bF70DECgpAIph/1ItyW7p0qdWllpv+/fvj5+dHamrqOfcZMmQIXl5eHDt2rBwru3h///03EydOZPfu3VaX4rR06VLnf1effPJJoft06dIFm81GixYtyrm6kskLuq+//rrVpYick4fVBYhUBB9//HG+xx999BGLFi0qsL1p06Ylep9p06bhcDhKdIzyMmTIEL755hvmzZvH0KFDCzyfkZHB119/Td++falevXqx3+fpp5/mySefLEmpF/T333/z7LPP0qNHjwItcD/++GOZvveF+Pj4MGvWLG677bZ823fv3s3y5cvx8fGxqDIR16YAJAIFvnxWrlzJokWLCmw/W0ZGBn5+fkV+H09Pz2LVZ4X+/fsTGBjIrFmzCg1AX3/9Nenp6QwZMqRE7+Ph4YGHh3V/iry8vCx7b4CrrrqK+fPnc/ToUcLCwpzbZ82aRWRkJA0bNuTEiRMWVijimnQJTKSIevToQYsWLVizZg3dunXDz8+Pf/3rX4AZBq6++mpiYmLw9vamfv36PP/889jt9nzHOLsP0JmXCqZOnUr9+vXx9vbmkksu4Y8//jhvPX/++Sc2m42ZM2cWeG7hwoXYbDa+/fZbAFJTUxk9ejR16tTB29ubiIgIrrjiCtauXXvO4/v6+nLDDTewePFiEhMTCzw/a9YsAgMD6d+/P8ePH2fs2LG0bNmSgIAAgoKC6NevH+vXrz/vOUDhfYCysrJ45JFHCA8Pd77H/v37C7x2z5493H///TRu3BhfX1+qV6/OzTffnO9S14wZM7j55psB6NmzZ4HLmYX1AUpMTOTOO+8kMjISHx8f4uLiCnzOJfndnWnAgAF4e3szZ86cfNtnzZrFwIEDcXd3L/Ca3Nxcnn/+eed71qlTh3/9619kZWXl288wDF544QVq1qyJn58fPXv2ZNOmTYXWkZSUxOjRo4mNjcXb25sGDRrwyiuvlGmLZVE+Z4DPP/+cdu3aERgYSFBQEC1btuS///2v8/mcnByeffZZGjZsiI+PD9WrV+eyyy5j0aJFZVa7VH5qARK5CMeOHaNfv37ccsst3HbbbURGRgLml2xAQABjxowhICCAn3/+mQkTJpCSksJrr712wePOmjWL1NRU7r33Xmw2G6+++io33HADO3fuPGerUfv27alXrx5ffPEFw4YNy/fc7NmzCQ0NpU+fPgDcd999zJ07lwceeIBmzZpx7NgxfvvtNzZv3kzbtm3PWdeQIUOYOXMmX3zxBQ888IBz+/Hjx1m4cCGDBw/G19eXTZs28dVXX3HzzTdTt25dDh8+zHvvvUf37t35+++/iYmJueBncKa77rqLTz75hFtvvZXOnTvz888/c/XVVxfY748//mD58uXccsst1KxZk927d/Puu+/So0cP/v77b/z8/OjWrRsPPfQQkydP5l//+pfzMua5LmeePHmSHj16sH37dh544AHq1q3LnDlzGD58OElJSTz88MP59i/O7+5Mfn5+DBgwgM8++4yRI0cCsH79ejZt2sT777/Phg0bCv18Zs6cyU033cSjjz7KqlWrePnll9m8eTPz5s1z7jdhwgReeOEFrrrqKq666irWrl3LlVdeSXZ2dr7jZWRk0L17dw4cOMC9995LrVq1WL58OePGjePQoUNMmjTpgudxsYr6OS9atIjBgwfTq1cvXnnlFQA2b97M77//7txn4sSJvPzyy9x111106NCBlJQU/vzzT9auXcsVV1xR6rWLizBEpIBRo0YZZ//v0b17dwMwpkyZUmD/jIyMAtvuvfdew8/Pz8jMzHRuGzZsmFG7dm3n4127dhmAUb16deP48ePO7V9//bUBGN9888156xw3bpzh6emZ77VZWVlGSEiIcccddzi3BQcHG6NGjTrvsQqTm5trREdHG506dcq3fcqUKQZgLFy40DAMw8jMzDTsdnu+fXbt2mV4e3sbzz33XIHz/fDDD53bnnnmmXyfdXx8vAEY999/f77j3XrrrQZgPPPMM85thX3uK1asMADjo48+cm6bM2eOARhLliwpsH/37t2N7t27Ox9PmjTJAIxPPvnEuS07O9vo1KmTERAQYKSkpOQ7l+L+7pYsWWIAxpw5c4xvv/3WsNlsxt69ew3DMIzHHnvMqFevnrO+5s2bF/h87rrrrnzHGzt2rAEYP//8s2EYhpGYmGh4eXkZV199teFwOJz7/etf/zIAY9iwYc5tzz//vOHv729s3bo13zGffPJJw93d3VmXYRgFfgeFyftsXnvttXPuU9TP+eGHHzaCgoKM3Nzccx4rLi7OuPrqq89bk8jZdAlM5CJ4e3szYsSIAtt9fX2d91NTUzl69Chdu3YlIyODf/7554LHHTRoEKGhoc7HXbt2BWDnzp0XfF1OTg5ffvmlc9uPP/5IUlISgwYNcm4LCQlh1apVHDx48IK1nMnd3Z1bbrmFFStW5LuslNc/pVevXoD5ubi5mX9O7HY7x44dIyAggMaNG5/3Mlthvv/+ewAeeuihfNtHjx5dYN8zP/ecnByOHTtGgwYNCAkJuej3PfP9o6KiGDx4sHObp6cnDz30EGlpaSxbtizf/sX93Z3pyiuvpFq1anz++ecYhsHnn3+e7/3Prg9gzJgx+bY/+uijAHz33XcA/PTTT2RnZ/Pggw/mu8RY2Oc4Z84cunbtSmhoKEePHnXeevfujd1u55dffinyuRRVUT/nkJAQ0tPTz3s5KyQkhE2bNrFt27ZSr1NclwKQyEWoUaNGoZ1mN23axPXXX09wcDBBQUGEh4c7O1AnJydf8Li1atXK9zjvC/VCnV/j4uJo0qQJs2fPdm6bPXs2YWFhXH755c5tr776Kn/99RexsbF06NCBiRMnFvkLOq+T86xZswDYv38/v/76K7fccouzf4rD4eA///kPDRs2xNvbm7CwMMLDw9mwYUORzv9Me/bswc3Njfr16+fb3rhx4wL7njx5kgkTJjj7reS9b1JS0kW/75nv37BhQ2egy5N3yWzPnj35thf3d3cmT09Pbr75ZmbNmsUvv/zCvn37uPXWW89Zn5ubGw0aNMi3PSoqipCQEGd9eT8bNmyYb7/w8PB8gQ1g27ZtLFiwgPDw8Hy33r17AxTaB6ykivo533///TRq1Ih+/fpRs2ZN7rjjDhYsWJDvNc899xxJSUk0atSIli1b8thjjxV66VDkTApAIhfhzBaHPElJSXTv3p3169fz3HPP8c0337Bo0SJnf4WidCItrKMrmJ1YL2TQoEEsWbKEo0ePkpWVxfz587nxxhvzjawaOHAgO3fu5K233iImJobXXnuN5s2b88MPP1zw+O3ataNJkyZ89tlnAHz22WcYhpFv9NdLL73EmDFj6NatG5988gkLFy5k0aJFNG/evEw70T744IO8+OKLDBw4kC+++IIff/yRRYsWUb169XKbbqAkv7sz3XrrrcTHxzNx4kTi4uJo1qzZefcv6uSRReFwOLjiiitYtGhRobcbb7yx1N7rYkVERBAfH8/8+fPp378/S5YsoV+/fvn6vXXr1o0dO3Ywffp0WrRowfvvv0/btm15//33LatbKj51ghYpoaVLl3Ls2DG+/PJLunXr5ty+a9eucnn/QYMG8eyzz/K///2PyMhIUlJSuOWWWwrsFx0dzf3338/9999PYmIibdu25cUXX6Rfv34XfI8hQ4Ywfvx4NmzYwKxZs2jYsCGXXHKJ8/m5c+fSs2dPPvjgg3yvS0pKyje0uyhq166Nw+Fgx44d+Vp9tmzZUmDfuXPnMmzYMN544w3ntszMTJKSkvLtdzFhoXbt2mzYsAGHw5GvdSLvUmbt2rWLfKyLcdlll1GrVi2WLl3qDM/nqs/hcLBt27Z8HbkPHz5MUlKSs768n9u2baNevXrO/Y4cOVKgdap+/fqkpaU5W3zKw8V8zl5eXlx77bVce+21OBwO7r//ft577z3Gjx/vbAmrVq0aI0aMYMSIEaSlpdGtWzcmTpzIXXfdVW7nJJWLWoBESiivBeDMf/FnZ2fzf//3f+Xy/k2bNqVly5bMnj2b2bNnEx0dnS+I2e32ApeDIiIiiImJKTBs+lzyWnsmTJhAfHx8gbl/3N3dC7R4zJkzhwMHDlz0+eQFssmTJ+fbXthIpMLe96233iow/YC/vz9AgWBUmKuuuoqEhIR8lxVzc3N56623CAgIoHv37kU5jYtms9mYPHkyzzzzDLfffvt564OCn8ebb74J4Bwt17t3bzw9PXnrrbfyfUaFfY4DBw5kxYoVLFy4sMBzSUlJ5ObmXuzpXFBRP+ezZxl3c3OjVatWAM7/fs/eJyAggAYNGhT5v2+pmtQCJFJCnTt3JjQ0lGHDhvHQQw9hs9n4+OOPL/oSSEkMGjSICRMm4OPjw5133pnvX9SpqanUrFmTm266ibi4OAICAvjpp5/4448/8rWcnE/dunXp3LkzX3/9NUCBAHTNNdfw3HPPMWLECDp37szGjRv59NNP87U8FFXr1q0ZPHgw//d//0dycjKdO3dm8eLFbN++vcC+11xzDR9//DHBwcE0a9aMFStW8NNPPxWYmbp169a4u7vzyiuvkJycjLe3N5dffjkREREFjnnPPffw3nvvMXz4cNasWUOdOnWYO3cuv//+O5MmTSIwMPCiz6moBgwYwIABA867T1xcHMOGDWPq1KnOy6+rV69m5syZXHfddfTs2RMw+/qMHTuWl19+mWuuuYarrrqKdevW8cMPPxRolXvssceYP38+11xzDcOHD6ddu3akp6ezceNG5s6dy+7duy+6JQ9g8eLFZGZmFth+3XXXFflzvuuuuzh+/DiXX345NWvWZM+ePbz11lu0bt3a2QLWrFkzevToQbt27ahWrRp//vmnc9oHkXOybPyZSAV2rmHwZw5HPtPvv/9uXHrppYavr68RExNjPP7448bChQsLDL0+1zD4woYLU4Thxnm2bdtmAAZg/Pbbb/mey8rKMh577DEjLi7OCAwMNPz9/Y24uDjj//7v/4p07DzvvPOOARgdOnQo8FxmZqbx6KOPGtHR0Yavr6/RpUsXY8WKFQWGmBdlGLxhGMbJkyeNhx56yKhevbrh7+9vXHvttca+ffsKfCYnTpwwRowYYYSFhRkBAQFGnz59jH/++ceoXbt2vmHehmEY06ZNM+rVq2e4u7vn+72cXaNhGMbhw4edx/Xy8jJatmyZr+Yzz6W4v7szh8GfT2H/3eXk5BjPPvusUbduXcPT09OIjY01xo0bl2/KBcMwDLvdbjz77LPO30uPHj2Mv/76q9DPJzU11Rg3bpzRoEEDw8vLywgLCzM6d+5svP7660Z2dvZFnVveZ3Ou28cff2wYRtE+57lz5xpXXnmlERERYXh5eRm1atUy7r33XuPQoUPOfV544QWjQ4cORkhIiOHr62s0adLEePHFF/PVLXI2m2GU4z9TRURERCoA9QESERGRKkcBSERERKocBSARERGpchSAREREpMpRABIREZEqRwFIREREqhxNhFgIh8PBwYMHCQwMLNX1dkRERKTsGIZBamoqMTExBRbaPZsCUCEOHjxIbGys1WWIiIhIMezbt4+aNWuedx8FoELkTcG+b98+goKCLK5GREREiiIlJYXY2NgiLVmjAFSIvMteQUFBCkAiIiKVTFG6r6gTtIiIiFQ5CkAiIiJS5SgAiYiISJWjPkAiIuLS7HY7OTk5VpchpcDT0xN3d/dSOZYCkIiIuCTDMEhISCApKcnqUqQUhYSEEBUVVeJ5+hSARETEJeWFn4iICPz8/DSxbSVnGAYZGRkkJiYCEB0dXaLjKQCJiIjLsdvtzvBTvXp1q8uRUuLr6wtAYmIiERERJbocpk7QIiLicvL6/Pj5+VlciZS2vN9pSft1KQCJiIjL0mUv11Nav1MFIBEREalyFIBERERcXJ06dZg0aZLVZVQoCkAiIiIVhM1mO+9t4sSJxTruH3/8wT333FO6xVZyGgVWjgzDYP+Jk7i52agR4mt1OSIiUsEcOnTIeX/27NlMmDCBLVu2OLcFBAQ47xuGgd1ux8Pjwl/l4eHhpVuoC1ALUDl66fvNdH11CTN+32V1KSIiUgFFRUU5b8HBwdhsNufjf/75h8DAQH744QfatWuHt7c3v/32Gzt27GDAgAFERkYSEBDAJZdcwk8//ZTvuGdfArPZbLz//vtcf/31+Pn50bBhQ+bPn1/OZ2stBaBy1DgqCID4fUnWFiIiUgUZhkFGdq4lN8MwSu08nnzySf7973+zefNmWrVqRVpaGldddRWLFy9m3bp19O3bl2uvvZa9e/ee9zjPPvssAwcOZMOGDVx11VUMGTKE48ePl1qdFZ0ugZWj1rEhAGw8kEyO3YGnu/KniEh5OZljp9mEhZa899/P9cHPq3S+cp977jmuuOIK5+Nq1aoRFxfnfPz8888zb9485s+fzwMPPHDO4wwfPpzBgwcD8NJLLzF58mRWr15N3759S6XOik7fwOWoXpg/gT4eZOY42JKQanU5IiJSCbVv3z7f47S0NMaOHUvTpk0JCQkhICCAzZs3X7AFqFWrVs77/v7+BAUFOZeZqArUAlSO3NxstI4N4ddtR1m3L4kWNYKtLklEpMrw9XTn7+f6WPbepcXf3z/f47Fjx7Jo0SJef/11GjRogK+vLzfddBPZ2dnnPY6np2e+xzabDYfDUWp1VnQKQOWszakAFL83idsvrW11OSIiVYbNZiu1y1AVye+//87w4cO5/vrrAbNFaPfu3dYWVQnoElg5a10rBID4fSesLURERFxCw4YN+fLLL4mPj2f9+vXceuutVaolp7gUgMpZXM0QAHYcSSf5ZMkWchMREXnzzTcJDQ2lc+fOXHvttfTp04e2bdtaXVaFZzNKc2yei0hJSSE4OJjk5GSCgoJK/fjdXl3C3uMZfHRHB7o10uRUIiKlLTMzk127dlG3bl18fHysLkdK0fl+txfz/a0WIAu0cV4GS7K0DhERkarK8gD0zjvvUKdOHXx8fOjYsSOrV68+7/5JSUmMGjWK6OhovL29adSoEd9//73z+YkTJxZYO6VJkyZlfRoXJW8+IAUgERERa1jaHX727NmMGTOGKVOm0LFjRyZNmkSfPn3YsmULERERBfbPzs7miiuuICIigrlz51KjRg327NlDSEhIvv2aN2+ebxrwoqyTUp7ODECGYWCz2awtSEREpIqxNBm8+eab3H333YwYMQKAKVOm8N133zF9+nSefPLJAvtPnz6d48ePs3z5cuf8BXXq1Cmwn4eHB1FRUWVae0k0iwnCy92N4+nZ7Dt+klrV/awuSUREpEqx7BJYdnY2a9asoXfv3qeLcXOjd+/erFixotDXzJ8/n06dOjFq1CgiIyNp0aIFL730Ena7Pd9+27ZtIyYmhnr16jFkyJALzoaZlZVFSkpKvltZ8vZwp1mM2TlrnYbDi4iIlDvLAtDRo0ex2+1ERkbm2x4ZGUlCQkKhr9m5cydz587Fbrfz/fffM378eN544w1eeOEF5z4dO3ZkxowZLFiwgHfffZddu3bRtWtXUlPPvfTEyy+/THBwsPMWGxtbOid5HnmXwdbtTSrz9xIREZH8LO8EfTEcDgcRERFMnTqVdu3aMWjQIJ566immTJni3Kdfv37cfPPNtGrVij59+vD999+TlJTEF198cc7jjhs3juTkZOdt3759ZX4uGgkmIiJiHcv6AIWFheHu7s7hw4fzbT98+PA5++9ER0fj6emJu/vpNVWaNm1KQkIC2dnZeHl5FXhNSEgIjRo1Yvv27eesxdvbG29v72KeSfHktQD9fTCFrFw73h6lt06MiIiInJ9lLUBeXl60a9eOxYsXO7c5HA4WL15Mp06dCn1Nly5d2L59e74pvrdu3Up0dHSh4QfMNVF27NhBdHR06Z5ACdWq5kc1fy+y7Q42H9LK8CIiIuXJ0ktgY8aMYdq0acycOZPNmzczcuRI0tPTnaPChg4dyrhx45z7jxw5kuPHj/Pwww+zdetWvvvuO1566SVGjRrl3Gfs2LEsW7aM3bt3s3z5cq6//nrc3d0ZPHhwuZ/f+dhstjP6AakjtIiIlI4ePXowevRo5+M6deowadKk877GZrPx1Vdflfi9S+s45cHSYfCDBg3iyJEjTJgwgYSEBFq3bs2CBQucHaP37t2Lm9vpjBYbG8vChQt55JFHaNWqFTVq1ODhhx/miSeecO6zf/9+Bg8ezLFjxwgPD+eyyy5j5cqVhIdXvCUnWseG8PM/ieoHJCIiAFx77bXk5OSwYMGCAs/9+uuvdOvWjfXr19OqVasiH/OPP/7A39+/NMtk4sSJfPXVV8THx+fbfujQIUJDQ0v1vcqK5TMEPvDAAzzwwAOFPrd06dIC2zp16sTKlSvPebzPP/+8tEorc5oRWkREznTnnXdy4403sn//fmrWrJnvuQ8//JD27dtfVPgByrUBoCLPwXe2SjUKzNXEnQpAe45lcDw929piRETEctdccw3h4eHMmDEj3/a0tDTmzJnDddddx+DBg6lRowZ+fn60bNmSzz777LzHPPsS2LZt2+jWrRs+Pj40a9aMRYsWFXjNE088QaNGjfDz86NevXqMHz+enJwcAGbMmMGzzz7L+vXrnUtO5dV79iWwjRs3cvnll+Pr60v16tW55557SEtLcz4/fPhwrrvuOl5//XWio6OpXr06o0aNcr5XWbK8BagqC/b1pF64PzuPpLN+XxI9mxRc/kNEREqJYUBOhjXv7ekHRVj2yMPDg6FDhzJjxgyeeuop51JJc+bMwW63c9tttzFnzhyeeOIJgoKC+O6777j99tupX78+HTp0uODxHQ4HN9xwA5GRkaxatYrk5OR8/YXyBAYGMmPGDGJiYti4cSN33303gYGBPP744wwaNIi//vqLBQsWOJedCg4OLnCM9PR0+vTpQ6dOnfjjjz9ITEzkrrvu4oEHHsgX8JYsWUJ0dDRLlixh+/btDBo0iNatW3P33Xdf8HxKQgHIYm1iQ9l5JJ11e08oAImIlKWcDHgpxpr3/tdB8CpaP5w77riD1157jWXLltGjRw/AvPx14403Urt2bcaOHevc98EHH2ThwoV88cUXRQpAP/30E//88w8LFy4kJsb8LF566SX69euXb7+nn37aeb9OnTqMHTuWzz//nMcffxxfX18CAgIuuOzUrFmzyMzM5KOPPnL2QXr77be59tpreeWVV5z9fUNDQ3n77bdxd3enSZMmXH311SxevLjMA5AugVms9akJEdepH5CIiABNmjShc+fOTJ8+HYDt27fz66+/cuedd2K323n++edp2bIl1apVIyAggIULF15wyac8mzdvJjY21hl+gEKnnpk9ezZdunQhKiqKgIAAnn766SK/x5nvFRcXl68DdpcuXXA4HGzZssW5rXnz5vnm94uOjiYxMfGi3qs41AJksTan+gGt35eEw2Hg5qaV4UVEyoSnn9kSY9V7X4Q777yTBx98kHfeeYcPP/yQ+vXr0717d1555RX++9//MmnSJFq2bIm/vz+jR48mO7v0+pGuWLGCIUOG8Oyzz9KnTx+Cg4P5/PPPeeONN0rtPc6Ut7h5HpvNlm++v7KiAGSxxlGBeHu4kZKZy65j6dQPD7C6JBER12SzFfkylNUGDhzIww8/zKxZs/joo48YOXIkNpuN33//nQEDBnDbbbcBZp+erVu30qxZsyIdt2nTpuzbt49Dhw45Jwg+e2T18uXLqV27Nk899ZRz2549e/Lt4+XlVWAh8sLea8aMGaSnpztbgX7//Xfc3Nxo3LhxkeotS7oEZjFPdzda1jA7j8VrYVQREQECAgIYNGgQ48aN49ChQwwfPhyAhg0bsmjRIpYvX87mzZu59957CywpdT69e/emUaNGDBs2jPXr1/Prr7/mCzp577F3714+//xzduzYweTJk5k3b16+ferUqcOuXbuIj4/n6NGjZGVlFXivIUOG4OPjw7Bhw/jrr79YsmQJDz74ILfffnuBhdCtoABUAbRx9gPSjNAiImK68847OXHiBH369HH22Xn66adp27Ytffr0oUePHkRFRXHdddcV+Zhubm7MmzePkydP0qFDB+666y5efPHFfPv079+fRx55hAceeIDWrVuzfPlyxo8fn2+fG2+8kb59+9KzZ0/Cw8MLHYrv5+fHwoULOX78OJdccgk33XQTvXr14u233774D6MM2AzDMKwuoqJJSUkhODiY5ORkgoKCyvz9vttwiFGz1tKiRhDfPti1zN9PRMTVZWZmsmvXLurWrYuPj4/V5UgpOt/v9mK+v9UCVAHkjQT751AqmTnnv6YqIiIiJacAVAHEBPsQHuhNrsPgrwPJVpcjIiLi8hSAKoAzV4bXumAiIiJlTwGognB2hNZIMBERkTKnAFRBqAVIRKT0aZyP6ymt36kCUAXRqmYINhscSDpJYmqm1eWIiFRqebMLZ2RYtPiplJm83+nZM0hfLM0EXUEEeHvQKCKQLYdTid+bxJXNz73AnIiInJ+7uzshISHONaX8/PycK6tL5WQYBhkZGSQmJhISEpJv/bDiUACqQFrHhpgBaJ8CkIhISeWtVF4eC2tK+QkJCTnvKvRFpQBUgbSpFcLsP/epI7SISCmw2WxER0cTERFBTk6O1eVIKfD09Cxxy08eBaAKJG9CxA37k7A7DNy1MryISIm5u7uX2pemuA51gq5AGkYE4u/lTnq2ne2JaVaXIyIi4rIUgCoQdzcbLWueWhleC6OKiIiUGQWgCqZNrVBA8wGJiIiUJQWgCiZvQkR1hBYRESk7CkAVTJtTAWjr4VTSs3KtLUZERMRFKQBVMBFBPsQE++AwYMN+rQwvIiJSFhSAKqC84fDqByQiIlI2FIAqoDaxZkfodXs1EkxERKQsKABVQGe2AGklYxERkdKnAFQBtYgJxt3NRmJqFoeStTK8iIhIaVMAqoB8vdxpEhUIqB+QiIhIWVAAqqDy5gNSABIRESl9CkAVVN6M0OoILSIiUvoUgCqovBagjQeSybE7rC1GRETExSgAVVD1wvwJ9PEgM8fBloRUq8sRERFxKQpAFZSbm039gERERMqIAlAFpgAkIiJSNhSAKrA2pyZEVEdoERGR0qUAVIHF1QwBYMeRdJJP5lhbjIiIiAtRAKrAqgd4U6uaHwAb9idZW4yIiIgLUQCq4Jz9gPYmWVqHiIiIK1EAquDUEVpERKT0KQBVcM6O0FoZXkREpNQoAFVwzWKC8HJ343h6NvuOn7S6HBEREZegAFTBeXu40zQmCIB1+zQcXkREpDQoAFUCbdQPSEREpFQpAFUCef2AFIBERERKhwJQJZA3EmzTgRSycu3WFiMiIuICFIAqgVrV/Kjm70W23cHmQ1oZXkREpKQUgCoBm81GXM1gAOK1LpiIiEiJKQBVEq1jQwH1AxIRESkNCkCVhDpCi4iIlB4FoEoi7lRH6N3HMjienm1tMSIiIpWcAlAlEezrSb1wfwDWqxVIRESkRBSAKpG84fDrFIBERERKRAGoEtGM0CIiIqVDAagSaVPLHAm2fl8SDodWhhcRESkuBaBKpHFUIN4ebiSfzGHXsXSryxEREam0FIAqEU93N1rWyJsQMcnaYkRERCoxBaBKprX6AYmIiJSYAlAl01oTIoqIiJSYAlAlk9cRevOhFDJztDK8iIhIcSgAVTIxwT6EB3qT6zD460Cy1eWIiIhUSgpAlYzNZlM/IBERkRJSAKqENCO0iIhIySgAVULOGaE1FF5ERKRYFIAqoVaxIdhscCDpJImpmVaXIyIiUukoAFVCAd4eNIoIBNQKJCIiUhwKQJWUOkKLiIgUnwJQJaUJEUVERIpPAaiSymsB2rA/GbtWhhcREbkolgegd955hzp16uDj40PHjh1ZvXr1efdPSkpi1KhRREdH4+3tTaNGjfj+++9LdMzKqFFkIH5e7qRl5bLjSJrV5YiIiFQqlgag2bNnM2bMGJ555hnWrl1LXFwcffr0ITExsdD9s7OzueKKK9i9ezdz585ly5YtTJs2jRo1ahT7mJWVu5uNVjXNleHX7T1hcTUiIiKVi6UB6M033+Tuu+9mxIgRNGvWjClTpuDn58f06dML3X/69OkcP36cr776ii5dulCnTh26d+9OXFxcsY9ZmbWONdcFUz8gERGRi2NZAMrOzmbNmjX07t37dDFubvTu3ZsVK1YU+pr58+fTqVMnRo0aRWRkJC1atOCll17CbrcX+5iVmXNGaA2FFxERuSgeVr3x0aNHsdvtREZG5tseGRnJP//8U+hrdu7cyc8//8yQIUP4/vvv2b59O/fffz85OTk888wzxTomQFZWFllZWc7HKSkpJTiz8tPm1EiwrYdTSc/Kxd/bsl+niIhIpWJ5J+iL4XA4iIiIYOrUqbRr145Bgwbx1FNPMWXKlBId9+WXXyY4ONh5i42NLaWKy1ZkkA8xwT44DNioleFFRESKzLIAFBYWhru7O4cPH863/fDhw0RFRRX6mujoaBo1aoS7u7tzW9OmTUlISCA7O7tYxwQYN24cycnJztu+fftKcGblK28+IF0GExERKTrLApCXlxft2rVj8eLFzm0Oh4PFixfTqVOnQl/TpUsXtm/fjsPhcG7bunUr0dHReHl5FeuYAN7e3gQFBeW7VRanZ4TWSDAREZGisvQS2JgxY5g2bRozZ85k8+bNjBw5kvT0dEaMGAHA0KFDGTdunHP/kSNHcvz4cR5++GG2bt3Kd999x0svvcSoUaOKfExXo5FgIiIiF8/SXrODBg3iyJEjTJgwgYSEBFq3bs2CBQucnZj37t2Lm9vpjBYbG8vChQt55JFHaNWqFTVq1ODhhx/miSeeKPIxXU3LGsG4u9k4nJLFoeSTRAf7Wl2SiIhIhWczDEPrKJwlJSWF4OBgkpOTK8XlsKsn/8qmgym8O6Qt/VpGW12OiIiIJS7m+7tSjQKTwjnnA9JlMBERkSJRAHIBzo7QGgkmIiJSJApALiBvQsSNB5LJtTvOv7OIiIgoALmCemEBBPp4cDLHzpbDqVaXIyIiUuEpALkANzfbGfMBJVlai4iISGWgAOQitDCqiIhI0SkAuQi1AImIiBSdApCLyAtAO46kkZKZY20xIiIiFZwCkIuoHuBNrWp+GAZs2KeV4UVERM5HAciFaGFUERGRolEAciHqCC0iIlI0CkAupPWpCRHj9yWhJd5ERETOTQHIhTSLDsLT3cax9Gz2nzhpdTkiIiIVlgKQC/HxdKdZTDCghVFFRETORwHIxbTRwqgiIiIXpADkYpwdoTUSTERE5JwUgFxMXgDadDCF7FytDC8iIlIYBSAXU7u6H6F+nmTnOth8KMXqckRERCokBSAXY7NpZXgREZELUQByQa1jQwFYt1f9gERERAqjAOSCzpwQUURERApSAHJBrWuGALD7WAYn0rOtLUZERKQCUgByQcF+ntQL9wcgfn+StcWIiIhUQApALqq1JkQUERE5JwUgF9XGOSFikqV1iIiIVEQKQC4qbyTYeq0MLyIiUoACkItqEh2It4cbySdz2HU03epyREREKhQFIBfl6e5GyxrmyvAaDi8iIpKfApAL04zQIiIihVMAcmFtapn9gL5ad4C/DiRbXI2IiEjFoQDkwno3i6BNrRBSMnO57YNVbDqoECQiIgIKQC7N28OdmXd0oHVsCEkZOQx5fxV/H9QK8SIiIgpALi7Ix5OP7uxAnDMErVQIEhGRKk8BqAoI8vHkozs6EFczmBOnQtDmQwpBIiJSdSkAVRHBvp58dGfHM0LQKv5JUAgSEZGqSQGoCskLQa1qBnM8PZtbp61iS0Kq1WWJiIiUOwWgKibY15OP7+hIyxp5IWilQpCIiFQ5CkBVULCfJ5/c2ZEWNYI4dioEbT2sECQiIlWHAlAVlReCmsecDkHbFIJERKSKUACqwkL8vPj0LjMEHU3LZvC0VWxPVAgSERHXpwBUxeWFoGbRQRxNy+KWqQpBIiLi+hSAxBmCmuYLQWlWlyUiIlJmFIDKU2Yy/P01HFpvdSUFhPqbIahJVCBH07IYPG0lO44oBImIiGtSACpPPz0LXwyFNTOsrqRQ1fy9mHX3pTSJCuRIahaDpyoEiYiIa1IAKk8NrzR/blsEhmFtLedwZghKPBWCdioEiYiIi1EAKk91u4G7NyTvgyP/WF3NOVU7dTmsceSpEDRtJbuOpltdloiISKlRACpPXn5Qt6t5f9uP1tZyAdUDvPn07o40igzgcIrZErRbIUhERFyEAlB5O/MyWAUXFuDNrLsvpVFkAAkpmdyiECQiIi5CAai8Neht/ty7whwVVsHlhaCGEWYIGjxtJXuOKQSJiEjlpgBU3qrXh+oNwJELO5daXU2R5IWgBhEBHEo2W4IUgkREpDJTALJC3mWwrRW7H9CZwgO9mXV3R+qH+3MoOZPBU1ey91iG1WWJiIgUiwKQFfIC0PZF4HBYW8tFiAj04bN7LqV+uD8Hk83LYfuOKwSJiEjlowBkhdqdwdMf0g5Dwgarq7koEYE+fHb3pdQL9+dA0klumaoQJCIilY8CkBU8vKFeD/N+JRgNdraIIB8+v/tS6oUpBImISOWkAGSVhleYPyv4fEDnEhFkXg7LC0GDp61k/wmFIBERqRyKFYD27dvH/v37nY9Xr17N6NGjmTp1aqkV5vLyAtD+PyD9mLW1FFPkqRBUN8yf/SfMliCFIBERqQyKFYBuvfVWlixZAkBCQgJXXHEFq1ev5qmnnuK5554r1QJdVnBNiGgOGLBjsdXVFFtkkNknKC8EDZ62kgNJJ60uS0RE5LyKFYD++usvOnToAMAXX3xBixYtWL58OZ9++ikzZswozfpcW6O8WaEr52WwPFHBZgiqU92PfcdPct07v/O/NftxOCrmgq8iIiLFCkA5OTl4e3sD8NNPP9G/f38AmjRpwqFDh0qvOlfnHA7/Ezjs1tZSQlHB5uWwBhEBHEnN4tE567nh3eXE70uyujQREZECihWAmjdvzpQpU/j1119ZtGgRffv2BeDgwYNUr169VAt0aTU7gHcwnDwBB9ZYXU2JRQf78t1Dl/FE3yb4e7kTvy+J6975nbFz1pOYmml1eSIiIk7FCkCvvPIK7733Hj169GDw4MHExcUBMH/+fOelMSkCdw9ocLl5v5JfBsvj7eHOyB71+XlsD25oWwOAuWv2c/nry3hv2Q6ycyvPxI8iIuK6bIZhFKujht1uJyUlhdDQUOe23bt34+fnR0RERKkVaIWUlBSCg4NJTk4mKCiobN8sfhZ8NRKi4+DeX8r2vSywdu8Jnp2/ifX7zYVf64b5M/6aplzeJNLiykRExNVczPd3sVqATp48SVZWljP87Nmzh0mTJrFly5ZKH37KXd7q8IfWQ2qCtbWUgba1Qpl3fxdevakVYQHe7Dqazh0z/mT4h6vZcSTN6vJERKSKKlYAGjBgAB999BEASUlJdOzYkTfeeIPrrruOd999t1QLdHkBERDT1rxfCWeFLgo3NxsD28eyZGx37u1WD093G0u3HKHvpF946fvNpGbmWF2iiIhUMcUKQGvXrqVr164AzJ07l8jISPbs2cNHH33E5MmTS7XAKqGhawyHv5BAH0/GXdWUhaO70bNxODl2g6m/7KTn60v54o99GjYvIiLlplgBKCMjg8DAQAB+/PFHbrjhBtzc3Lj00kvZs2dPqRZYJeQFoB1LwO76rSH1wgP4cEQHPhx+CfXC/Dmals3j/9vA9f/3O2v3nrC6PBERqQKKFYAaNGjAV199xb59+1i4cCFXXml+gScmJpZ9p2FXFNMG/MIgOxX2rrS6mnLTs0kEC0Z3419XNSHA24P1+5O54f+WM2Z2PIdTNGxeRETKTrEC0IQJExg7dix16tShQ4cOdOrUCTBbg9q0aVOqBVYJbm6nO0O7+GWws3l5uHFPt/r8PLY7N7erCcCX6w7Q8/Wl/N/S7WTlVu4JIkVEpGIq9jD4hIQEDh06RFxcHG5uZo5avXo1QUFBNGnSpFSLLG/lOgw+z8a58L87IbwpjKo6rUBnW78viYnfbGLd3iQAalf34+mrm9G7aQQ2m83a4kREpEK7mO/vYgegPHmrwtesWbMkh6lQLAlAJ0/Aq/XAcMDojRBSq3zetwJyOAy+ij/Av3/4h8TULAC6NgzjmWub0SAi0OLqRESkoirzeYAcDgfPPfccwcHB1K5dm9q1axMSEsLzzz+Pw6GZfovFNxRiO5r3q9hlsLO5udm4oW1Nfh7bg/u618fL3Y1ftx2l76Rfee6bv0k+6fodxUVEpGwVKwA99dRTvP322/z73/9m3bp1rFu3jpdeeom33nqL8ePHX/Tx3nnnHerUqYOPjw8dO3Zk9erV59x3xowZ2Gy2fDcfH598+wwfPrzAPnnrlVVoDa8wf7rofEAXK8Dbgyf7NeHHR7rRu2kkuQ6D6b/v4vLXl/L56r3YNWxeRESKyaM4L5o5cybvv/++cxV4gFatWlGjRg3uv/9+XnzxxSIfa/bs2YwZM4YpU6bQsWNHJk2aRJ8+fc47q3RQUBBbtmxxPi6sb0jfvn358MMPnY/zVq+v0BpeCYufg53LICcTPH0u/JoqoE6YP+8Pa8+yrUd47ptN7DiSzpNfbuSTVXuYeG1z2tepZnWJIiJSyRSrBej48eOFdnRu0qQJx48fv6hjvfnmm9x9992MGDGCZs2aMWXKFPz8/Jg+ffo5X2Oz2YiKinLeIiMLrivl7e2db58z1yyrsCJbQGA05J6EPb9ZXU2F071ROAtGd+Ppq5sS6O3BXwdSuGnKCh7+fB2JGjYvIiIXoVgBKC4ujrfffrvA9rfffptWrVoV+TjZ2dmsWbOG3r17ny7IzY3evXuzYsWKc74uLS2N2rVrExsby4ABA9i0aVOBfZYuXUpERASNGzdm5MiRHDt27JzHy8rKIiUlJd/NEjabLoNdgKe7G3d1rceSx3pwyyWx2GzwdfxBer2xjI9W7NZlMRERKZJiBaBXX32V6dOn06xZM+68807uvPNOmjVrxowZM3j99deLfJyjR49it9sLtOBERkaSkFD4wqCNGzdm+vTpfP3113zyySc4HA46d+7sHI0G5uWvjz76iMWLF/PKK6+wbNky+vXrh91e+JwyL7/8MsHBwc5bbGxskc+h1DXsY/7cuhBKNkDPpYUFePPvG1sxf9RltKoZTGpWLhO+3sT1//c7G0+tPC8iInIuxR4Gf/DgQd555x3++ecfAJo2bco999zDCy+8wNSpU4t8jBo1arB8+XLnZIoAjz/+OMuWLWPVqlUXPEZOTg5NmzZl8ODBPP/884Xus3PnTurXr89PP/1Er169CjyflZVFVlaW83FKSgqxsbHlOwzeWUwqvFIXHDnwwBoIa1C+718J2R0Gs1bt4dUFW0jNysXNBkM71WHMlY0I8vG0ujwRESknZT4MHiAmJoYXX3yR//3vf/zvf//jhRde4MSJE3zwwQdFPkZYWBju7u4cPnw43/bDhw8TFRVVpGN4enrSpk0btm/ffs596tWrR1hY2Dn38fb2JigoKN/NMt6BULuzeb+KD4cvKnc3G7d3qsPiR7vTPy4GhwEzlu+m9xvL+Gb9QUo41ZWIiLigYgeg0uDl5UW7du1YvHixc5vD4WDx4sX5WoTOx263s3HjRqKjo8+5z/79+zl27Nh596lQqsjq8KUtIsiHyYPb8MmdHakb5k9iahYPfraOodNXs/toutXliYhIBWJpAAIYM2YM06ZNY+bMmWzevJmRI0eSnp7OiBEjABg6dCjjxo1z7v/cc8/x448/snPnTtauXcttt93Gnj17uOuuuwCzg/Rjjz3GypUr2b17N4sXL2bAgAE0aNCAPn36WHKOFy0vAO35HbLSrK2lErqsYRg/PNyV0b0bOidRvHLSL/z3p21aW0xERIBizgNUmgYNGsSRI0eYMGECCQkJtG7dmgULFjg7Ru/du9e51hjAiRMnuPvuu0lISCA0NJR27dqxfPlymjVrBoC7uzsbNmxg5syZJCUlERMTw5VXXsnzzz9fOeYCAghrCCG1IWkP7PoFmlxldUWVjo+nO6N7N2JA6xpM+Povft12lP/8tJWv4w/w/HUt6NIgzOoSRUTEQhfVCfqGG2447/NJSUksW7bsnKOtKgtL1gI72/ePweqp0G4EXDvJmhpchGEYfLPhEM9/+zdHTq0tNqB1DE9d3ZSIQE02KSLiKsqsE/SZQ8ULu9WuXZuhQ4eWqHg55cx+QOrEWyI2m43+cTEsfrQ7wzrVzjd30Mcr92juIBGRKqjEq8G7ogrRApRzEl6pA7mZMHI5RDa3pg4XtGF/Ek/N+4uNB8z5guJqBvPi9S1pUSPY4spERKQkymUYvJQxT1+o2828r9FgpapVzRC+GtWFZ/s3J8Dbg/X7k+n/9m9MnL+J1EytNC8iUhUoAFVkzstgWhajtLm72RjW2Zw76JpW0c65g3q9sYzvNhzS3EEiIi5OAagiy1sXbO9KOJlkaSmuKjLIh7dvbctHd3SgdnU/ElOzGDVrLcM//IM9xzR3kIiIq1IAqshC60BYYzDssHOJ1dW4tG6Nwlk4uhsP9zLnDlq29QhX/ucX3lqsuYNERFyRAlBFl9cKtFX9gMqaj6c7j1zRiAWju9KlQXWych28sWgr/f77K8u3H7W6PBERKUUKQBVdXj+g7YvA4bC2liqiXngAn9zZkf/e0pqwAG92Hknn1vdX8cjseOc8QiIiUrkpAFV0tTqBVwCkH4FD8VZXU2XYbDYGtK7B4ke7c/ul5txB89YdoNcbS/l4xW4yc3RZTESkMlMAqug8vKBeD/O+RoOVu2BfT56/rgXz7u9C85ggUjJzGf/1Ji554Scen7uelTuP4dBEiiIilY4mQixEhZgI8UxrZsI3D0GN9nD3YqurqbJy7Q4+WbmHab/u4kDSSef2GiG+XN+mBte3rUH98AALKxQRqdou5vtbAagQFS4ApRyCN5sANnhsO/hrIU8rORwGq3cfZ97aA3y38RBpWbnO5+JiQ7ixbQ2uaRVDNX8vC6sUEal6FIBKqMIFIIApl0HCRrj+PYi7xepq5JTMHDuL/j7Ml2v388u2o851xTzcbPRsEsENbWpwedMIvD3cLa5URMT1Xcz3t0c51SQl1fBKMwBt+1EBqALx8XTn2rgYro2L4UhqFvPXH2Teuv38dSCFRX8fZtHfhwn29eTqVtHc2LYGbWuFYrPZrC5bRKTKUwtQISpkC9DelTC9D/iEwGM7wF3ZtSLbejiVL9ce4Kt1B0hIyXRur13dz+wv1KYGtav7W1ihiIjr0SWwEqqQAcieC6/Vh8wkuGMh1LrU6oqkCOwOgxU7jvHluv0s+CuBjOzTw+fb1w7l+rY1uKZlDMF+nhZWKSLiGhSASqhCBiCAuXfAX/+Dro9CrwlWVyMXKSM7l4WbEvhy7QF+336UvNHzXu5u9G4WwfVtatK9UTheHpqdQkSkOBSASqjCBqD1s2HePRDVEu77zepqpAQSkjP5Ov4AX649wJbDqc7t1fy9uLZVNDe0rUmrmsHqLyQichEUgEqowgag9KPwWgPAgDGbISjG6oqkhAzD4O9DKcxbe4Cv4g9yNO30Uhv1wv25sW1N+sfFUCPEFzc3hSERkfNRACqhChuAAKb1ggN/wrWTod0wq6uRUpRrd/Db9qN8ufYAP/6dQGbO6bXfbDYI8PYgyMeTQB+PU7fT9wO8zftB+bbn3zfA2wN3hSgRcWEaBu/KGl5pBqBtPyoAuRgPdzd6NI6gR+MIUjNz+OGvBOatPcDKXccwDEjNzCU1M/fCBzqPAO+8wHR2iPI8FZ7M+35e7vh6uePn5Y6Ppzt+Xh7mNs8ztnu4q1VKRCottQAVokK3AB1YC9N6mgukPr7LXCtMXFpmjp2UzBxnAEp13jd/ppy6n5b3fFb+fVMyc8nOdVz4jYrBx9MNPy+P/MHI0/x55n3z+YIhKu9+g/AAIoJ8yqRGEak61ALkyqJbg38EpCfC3hVQr7vVFUkZ8/E0g0REYPGPkZVrLzRApZzalnbm9qwcMrLtZGTbOZlt52SO+TMjO5eTOfZ8l+Yycxxk5mSXwllCk6hAujcKp1ujcNrXCdXs2SJSphSAKhs3N2h4BcR/al4GUwCSIvD2cMc7wJ2wAO8SH8vhMMxQlHM6IGWcCkiZzvv2Qu7nFro9LTOXXcfS+SchlX8SUnnvl534erpzab1qdDsViOqF+WtEnIiUKgWgysgZgBZBnxetrkaqGDc3G/7eHvh7l96fj+Pp2fy67Qi/bD3KL9uOcCQ1iyVbjrBkyxEAaoT40q1RON0bhdO5QXWCfDRxpIiUjPoAFaJC9wECOJkEr9YDww4Pr4fQOlZXJFJqDMPgn4RUftl6hF+2HeGPXSfItp++7ObuZqNtrRC6NTRbh1rWCFZnbBEBNAy+xCp8AAL48CrY8ztc9Tp0uNvqakTKTEZ2Lit3HjNbh7YeYefR9HzPh/p5clnDcLo1DKNbo3Ai1ZlapMpSACqhShGAfvsP/DTRHBY/ZI7V1YiUm33HM/hl2xF+2XqE5duPkZqVf2qAJlGBZt+hhuFcUledqUWqEgWgEqoUAejwJni3M3j4wBO7wdPX6opEyl2O3UH8viR+2XqEZVuPsPFAMmf+RfPxdKNTverqTC1SRSgAlVClCECGAf9pASn7Ychcs2O0SBVXWGfqM+V1pm5bK4ToYF+igr2JCPIh0NtDwUjEBSgAlVClCEAA34yGNR9Ch3vgqtesrkakQrlQZ+oz+Xm5ExnkQ2SQ96mfPs7HUafuRwR563KaSAWnAFRClSYA/fM9fD4YQmqbo8H0L1iRc8rIzmXVzuMs23qE7YlpJKRkcjgl86KWFwn183SGo6i8wBTsQ2SgD1HBZkgK8/fWqDQRi2gm6Kqibjdw94KkPXB0G4Q3sroikQrLz8uDnk0i6NkkIt/2jOxcDqdkcfhUIDqckklCchaHUzM5nJxp/kzJIjvXwYmMHE5k5PBPQuo538fDzUZ4oHe+FqQaob7E1QyhZc1g/Lz0Z1ekItD/iZWZdwDU7gI7l5izQisAiVw0Py8P6oZ5UDfM/5z7GIZBUkYOh1MzSUjOJDEly9mCZN7Mx0fTssh1GBxKzuRQcmaB47i72WgcGUjrWiG0iQ2hTa1Q6oX5q8VIxAIKQJVdwytPB6DOD1hdjYhLstlshPp7EervRZOoczer59odHEnLKtCitCMxnXX7TnA4JYu/D6Xw96EUZq3aC0CgjwetY08HotaxIYT6a5FjkbKmPkCFqDR9gACO7YC32oKbJzyxC7xLsGKmiJSpQ8knid+bxLp9Sazbe4KNB5LzLS6bp051PzMUnQpETaOD8PJws6BikcpFnaBLqFIFIIDJbeD4Thj0KTS9xupqRKSIcuwOtiSkOgNR/L4kdh5JL7Cfl4cbLWKCaFMrlDa1QmgdG0KNEF8N3Rc5iwJQCVW6APTDE7BqCrQdBv0nW12NiJRAckYO8ftPB6L4fUkkZeQU2C880PtUK5EZiFrVDCGgFBaozbU7yMixczLbTka2+fNkTu4Z9+1n3c8lO9dBdLAv9cL9qR8eQEyIL+7q1yQWUAAqoUoXgLb/BJ/cCIExMOZvDYcXcSGGYbD7WIYzEK3bm8TmQynkOvL/6XazQaPIQNrUCqFJVBB2h8HJM4NMTu4Z9+2F3M/lZI6dHHvJvxK8PNyoW92f+hH+1AsLoF64P/XCzZ9BPp4lPr7IuSgAlVClC0A5mfBKHcg9Cff9BlEtra5IRMpQZo6dTQeTWXeqP1H83iQOJJ0s1fdws5kj5Hy93PHzcsfX0/2s+x74ndrm4WZj/4mT7DiSxp5jGeeccBIgLMD7VEtR/nAUG+qLh7v6OUnJaB6gqsbTB+p1h60LzNFgCkAiLs3H0512tavRrnY157bElEwzDO1LYkdiGt6e7vh6ujlDjK/nqfCS777HOba74+XuVqw+RnaHwf4TGew8ks6OI2nsPJrOziNp7DySTmJqFkfTzNvqXcfzvc7T3Uatan7UDw9wthblhSSNipOyoBagQlS6FiCAPz6A78ZArU5wxwKrqxERKSA1M4ddR9PZecQMRTtO3d91NK3Q0XB5Qv08zVAUdvpSWnV/L3zOCm8lCW7iGtQCVBXlLYa6bxWcPAG+odbWIyJylkAfT1rVNDtsn8nhMDiUksmOxDSzteiMkHQwOZMTGTms2XOCNXtOXPA93Gw4L9H5ermdvu/pdqqFywMfT3d8vdxO3z+jtcznjJawvOf8vNyx2SDHbpBjd5xxM85/P9dBrsMg2+4gJ9cg1+HIdz/H7iA719w315H/frCvJ10bhtOjcTjRwb5l9Bup2hSAXEVILQhvCkc2w46focWNVlckIlIkbm42aoT4UiPEl26NwvM9l5Gde0arUTo7j6ax62g6KSdznB24z+y87TAgPdtOerbdilMpVd9vTACgcWQgPZqE06NRBO3rhOLpAn2lkjNywAbBvtZ1ilcAciUNrzAD0NYfFYBExCX4eXnQPCaY5jHB590vx+4g84xAlDdcP/PMofs5djLPGMafmXPWc9mF7HfqmAAe7jY83d3wcndz3jdvtrN+mvc9Tu1b2H3zOIXf93S3sedYBku3JLJuXxJbDqey5XAq7y3bSaC3B10ahNGzSTjdG0UQFexTHr+GEnE4DHYeTWftqVa8tXtPsC0xjQnXNOOOy+paVpcCkCtpeCUsnwzbF4HDAW6V/18JIiJFkRc8Al1omP1DvRpyIj2bX7YdYdmWIyzdeoTj6dks2JTAgk1m61CTqEB6NomgR6Nw2tauGK1D6Vm5rN+f5Aw8684xl9Xe4xkWVHeaOkEXolJ2ggaw58Cr9SArBe76GWq2s7oiEREpJQ6HwcYDySzZksjSLUdYvz+JM7/BA3086NowjB6NIujeOJzIoLJvHTIMg/0nTrJ27+nWnc2HUrGfNU+Vt4cbcbEhtKsdSrtTM5pXD/Au9Xo0D1AJVdoABPDFUPj7a+j+JPQcZ3U1IiJSRo6lZfHrtqMs2ZLIL1uPcOKsVpZm0UH0bBJOj8YRtIkNKZV5lrJy7fx1ICXf5azE1KwC+8UE+9C2dqgZeGqH0jQ6qFxapxSASqhSB6B1n8DXoyCmLdyzxOpqRESkHNgdBhv2J7FkyxGWbUlk/f7kfM8H+XjQtVE4PRqF071xOBGBRWsdSkzJPKN1J4mN+5MLTHTp4WajeY1g2tUKpW3tENrWCiUmxJqRawpAJVSpA1BqArzR2Lw/dhsERFhbj4iIlLujaVn8svUIS7Yc4ZetR0g+mb91qEWNIHo0iqBnk3Bax4bi7mYj1+7gn4RU1u49Ybbw7D3BvuMFZxiv7u9F29qhtK1ltu60qhmMj6d7eZ3aeSkAlVClDkAA73WDQ+vhuneh9a1WVyMiIhayOwzi9yWx9FTfoY0H8rcOBft60iAigM2HUsg4a/oAm80cht/2VN+ddrVDqV3dr8JONqmJEKu6hleaAWjbjwpAIiJVnLubzdkX59ErG5OYmskvW82+Q7+eah3Km2Qy0NuD1rVCnPvHxYa47AK2CkCuqOGV8MtrsP1nsOeCu37NIiJiigj04aZ2NbmpXU1y7Q7i9yWx51gGzWsE0TAiEHe3itm6U9r0zeiKarQD32pw8jjsXw21O1tdkYiIVEAe7m60r1ON9nWqXXhnF2P9jElS+tzcoUFv8/62H62tRUREpAJSAHJVDa80f25bZG0dIiIiFZACkKtq0AuwweG/IPmA1dWIiIhUKApArsqvGtS8xLyvy2AiIiL5KAC5srzLYPGzICfT2lpEREQqEAUgV9biBvDwMUeCzRoIWWlWVyQiIlIhKAC5sur1Ychc8AqAXcvg4+vhZJLVVYmIiFhOAcjV1e0KQ78GnxCzJWjmNZB+1OqqRERELKUAVBXUbA/DvwP/CEjYCB/208gwERGp0hSAqoqoFjDiBwiqCUe3wod94fguq6sSERGxhAJQVRLWAO5YANXqQdJemN4XEv+xuioREZFypwBU1YTEwogFENEM0hLMy2EH462uSkREpFwpAFVFgZFmn6CYtuaCqTOvhb0rra5KRESk3CgAVVV+1czRYbW7QFaKOUR+x89WVyUiIlIuFICqMp8gc56gBldATgbMGgSbv7W6KhERkTKnAFTVefnBLbOgaX+wZ8MXQ2HDF1ZXJSIiUqYqRAB65513qFOnDj4+PnTs2JHVq1efc98ZM2Zgs9ny3Xx8fPLtYxgGEyZMIDo6Gl9fX3r37s22bdvK+jQqLw8vuOlDaD0EDDt8eQ/8Od3qqkRERMqM5QFo9uzZjBkzhmeeeYa1a9cSFxdHnz59SExMPOdrgoKCOHTokPO2Z8+efM+/+uqrTJ48mSlTprBq1Sr8/f3p06cPmZlaEPSc3D2g/9vQ4R7AgG8fgd//a3VVIiIiZcLyAPTmm29y9913M2LECJo1a8aUKVPw8/Nj+vRzt0DYbDaioqKct8jISOdzhmEwadIknn76aQYMGECrVq346KOPOHjwIF999VU5nFEl5uYG/V6Fy8aYjxdNgJ9fBMOwti4REZFSZmkAys7OZs2aNfTu3du5zc3Njd69e7NixYpzvi4tLY3atWsTGxvLgAED2LRpk/O5Xbt2kZCQkO+YwcHBdOzY8bzHlFNsNuj9DPR6xnz8y6uwYJxCkIiIuBRLA9DRo0ex2+35WnAAIiMjSUhIKPQ1jRs3Zvr06Xz99dd88sknOBwOOnfuzP79+wGcr7uYY2ZlZZGSkpLvVuV1HQNXvW7eX/UuzH8QHHZraxIRESklll8Cu1idOnVi6NChtG7dmu7du/Pll18SHh7Oe++9V+xjvvzyywQHBztvsbGxpVhxJdbhbrjuXbC5wbqP4X93QW621VWJiIiUmKUBKCwsDHd3dw4fPpxv++HDh4mKiirSMTw9PWnTpg3bt28HcL7uYo45btw4kpOTnbd9+/Zd7Km4rta3ws0zwM0TNn0Js2+DnJNWVyUiIlIilgYgLy8v2rVrx+LFi53bHA4HixcvplOnTkU6ht1uZ+PGjURHRwNQt25doqKi8h0zJSWFVatWnfOY3t7eBAUF5bvJGZoNgMGfg4cPbFsIn94MWWlWVyUiIlJsll8CGzNmDNOmTWPmzJls3ryZkSNHkp6ezogRIwAYOnQo48aNc+7/3HPP8eOPP7Jz507Wrl3Lbbfdxp49e7jrrrsAc4TY6NGjeeGFF5g/fz4bN25k6NChxMTEcN1111lxiq6hYW+47UvwCoTdv8LH18HJE1ZXJSIiUiweVhcwaNAgjhw5woQJE0hISKB169YsWLDA2Yl57969uLmdzmknTpzg7rvvJiEhgdDQUNq1a8fy5ctp1qyZc5/HH3+c9PR07rnnHpKSkrjssstYsGBBgQkT5SLV6QLDvoZPboT9f8CMa+D2eRAQYXVlIiIiF8VmGBrffLaUlBSCg4NJTk7W5bDCHP4bPhoA6YlQvYG5qGpwTaurEhGRKu5ivr8tvwQmlVBkM7hjAQTHwrHtML0fHNthdVUiIiJFpgAkxVO9vhmCqjeA5L3wYT+zZUhERKQSUACS4guuCSN+gIjmkHYYZlwFB9ZaXZWIiMgFKQBJyQREwPBvoUZ7c1TYzP6wc6nVVYmIiJyXApCUnF81GPoV1OkK2almB+np/WDTV2DPtbo6ERGRAhSApHR4B8KQOdDmNnDzgL3LYc4w+G8c/PYfyDhudYUiIiJOGgZfCA2DL6GUg/DndPjzQ8g4am7z8IVWA6HjvRDZ3Nr6RETEJV3M97cCUCEUgEpJTib89T9zNfmEjae31+0GHe+DRn3Bzd26+kRExKUoAJWQAlApMwzYuwJWTYHN34DhMLeH1IYO95iXzXxDLC1RREQqPwWgElIAKkNJ++CP92HNDMhMMrd5+kPrwdDhXghvZGV1IiJSiSkAlZACUDnIzoCNX8Cq9yDxjAkU6/cyL4816A1u6qMvIiJFpwBUQgpA5cgwYNcvZhDa8j1w6j/HavXNDtOtbzVHmImIiFyAAlAJKQBZ5Pgu8/LY2o8hK9nc5hVo9hHqcLe5/IaIiMg5KACVkAKQxbLSYP1nZqvQsW2nNtqgUR+zVaheT7DZLC1RREQqHgWgElIAqiAcDtj5sxmEtv14entYYzMIxd0CXv7W1SciIhWKAlAJKQBVQEe3w+qpEP8pZKeZ23yCoe1Qcyh9SC1r6xMREcspAJWQAlAFlpkM8bPMVqETu8xt7t7Qf7LZIiQiIlXWxXx/a5yxVC4+wXDpSHhwLQyeDbU6gz0L5t0LPz4NDrvVFYqISCWgACSVk5sbNO4Lw7+Dro+a25a/BbMGwskkS0sTEZGKTwFIKjc3N+g1AW78wFxwdftP8H5vs8+QiIjIOSgAiWtoeRPcsQCCaphD56ddDtt+sroqERGpoBSAxHXEtIZ7lkJsR3MixVk3m5fF1M9fRETOogAkriUgAoZ9A21uN1ed//FpmHcf5GRaXZmIiFQgCkDiejy8of9b0O9VsLnDhs9hxlWQcsjqykREpIJQABLXZLOZs0Xf/iX4hMCBNTC1B+xfY3VlIiJSASgAiWur1wPuWQLhTSAtAT7sB+s/t7oqERGxmAKQuL5q9eDORdConyZNFBERQAFIqgqfILhlFnQdaz7WpIkiIlWaApBUHW5u0Gs83DT9jEkTe8HRbVZXJiIi5UwBSKqeFjeemjSxJhzbDtN6wbZFVlclIiLlSAFIqqaY1mbn6NhLT02aOBB+n6xJE0VEqggFIKm6AiJg2PzTkyYuGm92kNakiSIiLk8BSKq2ApMmztakiSIiVYACkMiZkyb6hp4xaeKfVlcmIiJlRAFIJE+9HnD3zxDe9NSkiVdB/GdWVyUiImVAAUjkTNXqwV2LoPFV5qSJX90HC5/SpIkiIi5GAUjkbN6BMOhT6PaY+XjF25o0UYpv3x+w4QuFaJEKRgFIpDBubnD503DTh5o0UYrnyBb47Fb4oDd8eTd8ciOkH7O6KhE5RQFI5Hxa3AB3LtSkiVJ0KYdg/kPwf5fClu/A5gYePrBzidm5/mC81RWKCApAIhcWHZd/0sRPb4ZvRutf85JfZgosfh4mt4G1M825pRpfDfevNDvXV6sHyXvhgyth3adWVytS5dkMQ1Pfni0lJYXg4GCSk5MJCgqyuhypKHKz4YfHYc2H5mOfEPMyWbsR4O5haWliodxs87+JZa9AxqlQXPMSuOJ5qN3p9H4nk8yJNrcuMB+3vxP6/hs8vMq9ZBFXdTHf3wpAhVAAkvPasxy+fxwObzQfR7aEfq9AnS7W1lUaHHbz/DKTwJFrPnbYzfuG/YxtZ/48dd846/GZzxv2878uoinE3QoRTaz+BIrOMGDTPFj8HJzYZW6r3gB6PQNNrzXnlzqbwwG/vAZLXwYMMygN/AiCYsq1dBFXpQBUQgpAckH2XPNf/T+/YIYFgBY3wRXPQXANS0srFnsubJwDv74Bxyzs6F2jHbS+1Vyw1jfUujouZNevsGgCHFxrPvaPgB5PQtuh4O554ddv/RG+vAsyk8E/HG6eWTkDdHa6+d/M2o+g/R3QY1zhwU+knCgAlZACkBRZ+jH4+XlYMwMwwNMfuj0KnR4wl9mo6HKzYf1n8NubcGK3uc0nGMKbmEuDuLmDm8epm3v+nzb3wrfn3bed/dozH3uYnYPdPMwWoG0/wbaFZqsQgLuXORdT6yFQ//KKc4nx8Cb4aSJs+9F87OkPXR4yf9/eARd3rOM7YfbtcPgv87O68gW4dGTlCBB5rV8/Pg0pB05vbzkQBryjy3piGQWgElIAkot2aL15WWzfSvNxaF2zf0fjvtbWdS45mbDuY/j9v5C8z9zmV938Ir/kLvCx4L/7tCNmK1T8p2YoyBMQCa0GmWHIqktkyQdgyUtmbRhmYGk/Aro/YS6qW1zZGfDNQ+Z5g9mK2H8yePmXStllInEzfP8Y7P7VfBxSC5pfD8vfNsNsna4w6BPwDbG0TKmaFIBKSAFIisUwzAnvFk0wl9IAaHilGYSq17e2tjzZGWZr1fLJkHpqwdeASOj8kPmFXlG+eA9tgPhZsPGL0x2LofwvkZ1Mgt/+A6umQG6mua1pf7OfT1iD0nkPw4BV78GPT5ktYBHNYdDHFee/mTyZybD0FfOzMOzm0P7LHoEuD4PnqbmyvhgG2WnmcjK3zYXgmlZXLVWMAlAJKQBJiWSlmh1dV/wfOHLMyzmdRkHXsRd/maTUakqDP943Z7VOP2JuC6oBXUZD29vNL7CKKDfbvNwUP6t8L5HlZpmf1y+vwckT5rZanc0+XrGXlO575dmz3AwQ6YngHQw3ToNGfcrmvS6GwwEbZpvBPj3R3NbkGujzEoTWzr/vofXw6UDzHwCB0XDrFxDdqvxrlipLAaiEFICkVBzdBgueNP9lDOYXwhXPQ8ubyq+fR2YyrJoKK985/UUeUgsuG2O2pFSGfkp5yuMSmcMBf801+3Ul7TW3hTWG3hOhcb+y/72lHIIvhsL+1ebjHuOg2+PmzORWOBhvXu7Kq6d6A3PEY4Pe535N0j5zrqwjm8ErAAbOPP/+IqVIAaiEFICk1BiGOe/LgidPdzKu1RmuehWiWpbd+2Ych5XvmpdWspLNbdXqQ9dHodXAoo1UqsjK4hLZjiVmK0fCBvNxYLQZQFoPKd9O2LnZsHCc2QIF0LAP3PBe+Y6KyzhuhsA/P8TZub/743Dp/UXr4HwyCWbfZvYTsrmb/Zra3FbWVYsoAJWUApCUupxMWPEW/Pom5GSYI6DajTAnUvSrVnrvk3bEvMz1x/tmXwwwR3R1HWsu6+HmXnrvVRGUxiWyQxvgp2dgx8/mY69AuGy0+WXv5Vfmp3BO8bPg20fMvkehdc2OxVEtyvY9HXZzFuvFz51uMWxxE1z5/MXPVZSbBV+POt3Bu/uT5lQBlWGUm1RaCkAlpAAkZSZ5P/w4HjZ9aT72DYXLx0O74SULJ6kJ8Ptk+HM65J40t0W2hG5jzU67Vl1CKU8Xe4ksaS/8/KLZvwUD3Dzhkjuh22PgH1bu5Rfq0HqzJSVpr7kob/+3oNXNZfNe+1bD92PN9wSzM/ZVr0Kdy4p/TIfDbEn67U3zceshcO1/K38LpFRYCkAlpAAkZW7Xr/DDE5C4yXwc1RL6vZZ/6YSiSN4Pv00yJ6KzZ5nbYtqY/UbKo89KRXWuS2Qxbc1LZCd2w+qpYM82t7e40WyNq1bPknLPK+M4/O/O0y1Ul95vdsYurRCRlgiLnoH1s8zH3sFw+VPmUh2ldenvz+nw3aPm+mj1epqzX1sx1YK4PAWgElIAknJhzzW/GJa8YHZWBnMiuSuevfDlhhO7zctp8bPMkWYAsR3N4NOgV9UNPmc71yWyPHW6mmGiRltr6isqhx2WvGjOugxQuwvc9CEERhb/mPYcWD3NXJYjK8Xc1uY26DURAsJLXHIBWxfCnOHmJeDIFjBkjpYAOZeTJ8xgGtZI/y9fJAWgElIAknKVfvTUbNIzOd3h9LFTHU7PGqV1dLv5JbhhtjkXC5hf4t0fN3/qj+W55V0i2zjH7CPUbaw5OqkyfWabv4V590F2qtlJe+BHENvh4o+z6xdz4s4jm83HMW3gqtehZvvSrfdsB9bCrEHmcPqgGmYIimxetu9ZmTgcZv+9xc+affgCo82pEBr1hbrdre2TVkkoAJWQApBY4uA680spb8hxtfrmkOOGV5iz7/7yutl3yHCYz9e/3GzxudjLZlK5Hd0Gnw+Bo1vMfkv9/m1eripKkEs+YC5f4eyDVs0c4t/m9vLrJ3ZiN3xyk7nmnHeQ2bm7Xvfyee+KLPEfmP/g6f//be6n/5ED5sST9XqcDkRqPSuUAlAJKQCJZQqbdC6i+em+QgCN+pkddWu2s6ZGsV5WqjnC6u+vzceth8DVb5x7QsvcLHN04C+vnx6F2P5O6Pmv0h2FWFQZx+HzW2HvCjPEDXgb4m4p/zoqgtxsc7bxX183+6R5BZwKpbfBnt9hywJzKo28JWvyRLUy+/k16gvRravGQIciUAAqIQUgsVxmCvzyqjmXT16/lab9zeCjmXUFzDmmlk82F2c1HBAdBwM/Ljg787ZFZof74zvMx7U6wVWvle08VEWRkwlf3WcuqgpmJ/SuYyvXJcmS2v+n2eqT+Lf5uGEfuObNgkuIGIa5z5YfzL5U+/8AzvjqDoiCRlea/ziq173iLGljAQWgElIAkgrjyFbY/DU0vhoim1ldjVREO5fC3DvM0W6+oXDTdPPy6PFdsPBfsOV7c7+ASHMm8lYDK07IcDjMOZiWTzYftx0KV7/p+sPks9Ph5xfMf+BgmAsR93vVHI1YlN9N2hGzc//WBebowLw5v8C8VFa32+lLZRVtPbasVHO28OR95vxW4Y1K9fAKQCWkACQilUrSPvjidrMfmc0Nmg2Af743p0Zw84BLR5r9xSrq0PPV0+CHx82WrAa94eaZ1q2bV9a2L4ZvR59eaqXVLea6av7Vi3e83CzY/ZvZMrT1h9PHzRPV0gxCjfqZnd3L8lKZYUBmkllD0j7zZ/Kpn3n38ybYBHNyzJ7jSrUEBaASUgASkUonJ9OcyHDdx6e31ethtiyEN7asrCL75zuYe6c5kWd0nLmQamCU1VWVnozjZovc+s/Mx8GxcM0kaFiK66QZhjlgYuupfkP7VpPvUpl/xBmXynpcfMg0DHMx5aR9kLz3dNBxhpx95gjFC/EJgZBYaH0bXHrfxdVwAQpAJaQAJCKV1tqPzQkgL7nL7DdWUS53FcX+P81h8hlHIbiWOUy+pAvcWs0wzFF3Pzxhhgds0PFecwb4sm7lSj9q9gHb+gNs/zl/OHH3hrpdT7UO9TUDicNuzip/dquNM+jsPz3T/Pn4h5uLLgfHmscNqX36fnBsmbZEKgCVkAKQiIhFju80h8kf3wE+wXDLrJItx2Gl5APmDNhbfzAfhzc1lzOJvaT8a8nNNkeV5V0qy1ucOU9gtBmY8iZWPSebuW9IrdOBJu9+SG2zz9G5RiOWAwWgElIAEhGxUPox+OwWc04cdy+47l1oeZPVVRWdwwFrpsOiiWari5unOfHmZWPAw8vq6sxWqSNbzrhUtur0/GJuHuYklc4WnLOCTlCNinEO56AAVEIKQCIiFss5CV/eDZu/MR/3nghdRlf8S3pHt8H8h2DvcvNxzUvMVp+IptbWdT7px8yWt6Bos3WnJAszW+xivr81c5KIiFQ8nr7maLBL7zcf/zTRvJxkzz3vyyxjzzEnmny3ixl+PP3NDuh3LKzY4QfMEWixl5iXrypx+LlYpbTUr4iISClzc4e+L5uXXxb+C/78AFIOwk0fVKzJ/g6sNVt9Dm80HzfoDdf8x7xkJBWWWoBERKRi63Q/DJxpTvK39QeYcTUkbDSH/lspOwMWPgXv9zLDj281uH4qDJmr8FMJqAVIREQqvmYDzCUfPrvFnPBxymWAzeygW60+VG8A1c/4GVwL3MvwK27nUvjm4dOjqVreDH3/Df5hZfeeUqrUCboQ6gQtIlJBHd0O3z0CB+MhK+Xc+7l5QmidM4LRqXBUrb65knpxO1OfPAE/Pg3rPjEfB9U01+9q1Kd4x5NSdTHf32oBEhGRyiOsAQz75tSsxEfh2HZzzqBj20/ddpqPczPh2DbzdjZPv1OtRvVOBaQGp1uR/KqdOxz9/TV8NxbSEwGbOdlk72fAO7BMT1nKhgKQiIhUPjYbBISbt9qd8j/ncEDKgTOC0Y5Tt+3mJaucDLPPTl6n5TP5hORvLape3xwdtfwt+Odbc5+wRubQ9lqXlvVZShmqEJfA3nnnHV577TUSEhKIi4vjrbfeokOHDhd83eeff87gwYMZMGAAX331lXP78OHDmTlzZr59+/Tpw4IFC4pUjy6BiYi4KHuOubSDs8XojJCUsv/8r3XzMCcz7DYWPLzLp165KJXqEtjs2bMZM2YMU6ZMoWPHjkyaNIk+ffqwZcsWIiIizvm63bt3M3bsWLp27Vro83379uXDDz90Pvb21n+sIiJVnrvn6T5BnNVvJzsDTuzKH4qO7zAnCQxrBP1egcjmlpQtpc/yAPTmm29y9913M2LECACmTJnCd999x/Tp03nyyScLfY3dbmfIkCE8++yz/PrrryQlJRXYx9vbm6goF1pJWEREypaXnxlwFHKqBEvnAcrOzmbNmjX07t3buc3NzY3evXuzYsWKc77uueeeIyIigjvvvPOc+yxdupSIiAgaN27MyJEjOXbs2Dn3zcrKIiUlJd9NREREXJelAejo0aPY7XYiIyPzbY+MjCQhIaHQ1/z222988MEHTJs27ZzH7du3Lx999BGLFy/mlVdeYdmyZfTr1w+73V7o/i+//DLBwcHOW2xsbPFPSkRERCo8yy+BXYzU1FRuv/12pk2bRljYuSebuuWWW5z3W7ZsSatWrahfvz5Lly6lV69eBfYfN24cY8aMcT5OSUlRCBIREXFhlgagsLAw3N3dOXz4cL7thw8fLrT/zo4dO9i9ezfXXnutc5vD4QDAw8ODLVu2UL9+/QKvq1evHmFhYWzfvr3QAOTt7a1O0iIiIlWIpZfAvLy8aNeuHYsXL3ZuczgcLF68mE6dOhXYv0mTJmzcuJH4+HjnrX///vTs2ZP4+Phzttrs37+fY8eOER0dXWbnIiIiIpWH5ZfAxowZw7Bhw2jfvj0dOnRg0qRJpKenO0eFDR06lBo1avDyyy/j4+NDixYt8r0+JCQEwLk9LS2NZ599lhtvvJGoqCh27NjB448/ToMGDejTR1OVi4iISAUIQIMGDeLIkSNMmDCBhIQEWrduzYIFC5wdo/fu3YubW9Ebqtzd3dmwYQMzZ84kKSmJmJgYrrzySp5//nld5hIRERGggswEXdFoJmgREZHK52K+vy3tAyQiIiJiBQUgERERqXIUgERERKTKUQASERGRKkcBSERERKocBSARERGpciyfB6giypsZQKvCi4iIVB5539tFmeFHAagQqampAFoQVUREpBJKTU0lODj4vPtoIsRCOBwODh48SGBgIDabrVSPnbfS/L59+1xykkWdX+Xn6ueo86v8XP0cdX7FZxgGqampxMTEXHAVCbUAFcLNzY2aNWuW6XsEBQW55H/YeXR+lZ+rn6POr/Jz9XPU+RXPhVp+8qgTtIiIiFQ5CkAiIiJS5SgAlTNvb2+eeeYZl12ZXudX+bn6Oer8Kj9XP0edX/lQJ2gRERGpctQCJCIiIlWOApCIiIhUOQpAIiIiUuUoAImIiEiVowBUjt555x3q1KmDj48PHTt2ZPXq1VaXVGpefvllLrnkEgIDA4mIiOC6665jy5YtVpdVZv79739js9kYPXq01aWUmgMHDnDbbbdRvXp1fH19admyJX/++afVZZUau93O+PHjqVu3Lr6+vtSvX5/nn3++SGsGVUS//PIL1157LTExMdhsNr766qt8zxuGwYQJE4iOjsbX15fevXuzbds2a4otpvOdY05ODk888QQtW7bE39+fmJgYhg4dysGDB60r+CJd6Hd4pvvuuw+bzcakSZPKrb6SKsr5bd68mf79+xMcHIy/vz+XXHIJe/fuLZf6FIDKyezZsxkzZgzPPPMMa9euJS4ujj59+pCYmGh1aaVi2bJljBo1ipUrV7Jo0SJycnK48sorSU9Pt7q0UvfHH3/w3nvv0apVK6tLKTUnTpygS5cueHp68sMPP/D333/zxhtvEBoaanVppeaVV17h3Xff5e2332bz5s288sorvPrqq7z11ltWl1Ys6enpxMXF8c477xT6/KuvvsrkyZOZMmUKq1atwt/fnz59+pCZmVnOlRbf+c4xIyODtWvXMn78eNauXcuXX37Jli1b6N+/vwWVFs+Ffod55s2bx8qVK4mJiSmnykrHhc5vx44dXHbZZTRp0oSlS5eyYcMGxo8fj4+PT/kUaEi56NChgzFq1CjnY7vdbsTExBgvv/yyhVWVncTERAMwli1bZnUppSo1NdVo2LChsWjRIqN79+7Gww8/bHVJpeKJJ54wLrvsMqvLKFNXX321cccdd+TbdsMNNxhDhgyxqKLSAxjz5s1zPnY4HEZUVJTx2muvObclJSUZ3t7exmeffWZBhSV39jkWZvXq1QZg7Nmzp3yKKkXnOr/9+/cbNWrUMP766y+jdu3axn/+859yr600FHZ+gwYNMm677TZrCjIMQy1A5SA7O5s1a9bQu3dv5zY3Nzd69+7NihUrLKys7CQnJwNQrVo1iyspXaNGjeLqq6/O97t0BfPnz6d9+/bcfPPNRERE0KZNG6ZNm2Z1WaWqc+fOLF68mK1btwKwfv16fvvtN/r162dxZaVv165dJCQk5PvvNDg4mI4dO7rs3xww/+7YbDZCQkKsLqVUOBwObr/9dh577DGaN29udTmlyuFw8N1339GoUSP69OlDREQEHTt2PO9lwNKmAFQOjh49it1uJzIyMt/2yMhIEhISLKqq7DgcDkaPHk2XLl1o0aKF1eWUms8//5y1a9fy8ssvW11Kqdu5cyfvvvsuDRs2ZOHChYwcOZKHHnqImTNnWl1aqXnyySe55ZZbaNKkCZ6enrRp04bRo0czZMgQq0srdXl/V6rK3xyAzMxMnnjiCQYPHuwyC4i+8soreHh48NBDD1ldSqlLTEwkLS2Nf//73/Tt25cff/yR66+/nhtuuIFly5aVSw1aDV5K3ahRo/jrr7/47bffrC6l1Ozbt4+HH36YRYsWld/16XLkcDho3749L730EgBt2rThr7/+YsqUKQwbNszi6krHF198waeffsqsWbNo3rw58fHxjB49mpiYGJc5x6oqJyeHgQMHYhgG7777rtXllIo1a9bw3//+l7Vr12Kz2awup9Q5HA4ABgwYwCOPPAJA69atWb58OVOmTKF79+5lXoNagMpBWFgY7u7uHD58ON/2w4cPExUVZVFVZeOBBx7g22+/ZcmSJdSsWdPqckrNmjVrSExMpG3btnh4eODh4cGyZcuYPHkyHh4e2O12q0sskejoaJo1a5ZvW9OmTcttNEZ5eOyxx5ytQC1btuT222/nkUcecckWvby/K1Xhb05e+NmzZw+LFi1ymdafX3/9lcTERGrVquX8m7Nnzx4effRR6tSpY3V5JRYWFoaHh4elf3cUgMqBl5cX7dq1Y/Hixc5tDoeDxYsX06lTJwsrKz2GYfDAAw8wb948fv75Z+rWrWt1SaWqV69ebNy4kfj4eOetffv2DBkyhPj4eNzd3a0usUS6dOlSYNqCrVu3Urt2bYsqKn0ZGRm4ueX/k+fu7u78l6grqVu3LlFRUfn+5qSkpLBq1SqX+ZsDp8PPtm3b+Omnn6hevbrVJZWa22+/nQ0bNuT7mxMTE8Njjz3GwoULrS6vxLy8vLjkkkss/bujS2DlZMyYMQwbNoz27dvToUMHJk2aRHp6OiNGjLC6tFIxatQoZs2axddff01gYKCzn0FwcDC+vr4WV1dygYGBBfoz+fv7U716dZfo5/TII4/QuXNnXnrpJQYOHMjq1auZOnUqU6dOtbq0UnPttdfy4osvUqtWLZo3b866det48803ueOOO6wurVjS0tLYvn278/GuXbuIj4+nWrVq1KpVi9GjR/PCCy/QsGFD6taty/jx44mJieG6666zruiLdL5zjI6O5qabbmLt2rV8++232O1259+datWq4eXlZVXZRXah3+HZgc7T05OoqCgaN25c3qUWy4XO77HHHmPQoEF069aNnj17smDBAr755huWLl1aPgVaNv6sCnrrrbeMWrVqGV5eXkaHDh2MlStXWl1SqQEKvX344YdWl1ZmXGkYvGEYxjfffGO0aNHC8Pb2Npo0aWJMnTrV6pJKVUpKivHwww8btWrVMnx8fIx69eoZTz31lJGVlWV1acWyZMmSQv+fGzZsmGEY5lD48ePHG5GRkYa3t7fRq1cvY8uWLdYWfZHOd467du0659+dJUuWWF16kVzod3i2yjYMvijn98EHHxgNGjQwfHx8jLi4OOOrr74qt/pshlFJp0EVERERKSb1ARIREZEqRwFIREREqhwFIBEREalyFIBERESkylEAEhERkSpHAUhERESqHAUgERERqXIUgEREisBms/HVV19ZXYaIlBIFIBGp8IYPH47NZitw69u3r9WliUglpbXARKRS6Nu3Lx9++GG+bd7e3hZVIyKVnVqARKRS8Pb2JioqKt8tNDQUMC9Pvfvuu/Tr1w9fX1/q1avH3Llz871+48aNXH755fj6+lK9enXuuece0tLS8u0zffp0mjdvjre3N9HR0TzwwAP5nj969CjXX389fn5+NGzYkPnz55ftSYtImVEAEhGXMH78eG688UbWr1/PkCFDuOWWW9i8eTMA6enp9OnTh9DQUP744w/mzJnDTz/9lC/gvPvuu4waNYp77rmHjRs3Mn/+fBo0aJDvPZ599lkGDhzIhg0buOqqqxgyZAjHjx8v1/MUkVJSbsuuiogU07Bhwwx3d3fD398/3+3FF180DMMwAOO+++7L95qOHTsaI0eONAzDMKZOnWqEhoYaaWlpzue/++47w83NzUhISDAMwzBiYmKMp5566pw1AMbTTz/tfJyWlmYAxg8//FBq5yki5Ud9gESkUujZsyfvvvtuvm3VqlVz3u/UqVO+5zp16kR8fDwAmzdvJi4uDn9/f+fzXbp0weFwsGXLFmw2GwcPHqRXr17nraFVq1bO+/7+/gQFBZGYmFjcUxIRCykAiUil4O/vX+CSVGnx9fUt0n6enp75HttsNhwOR1mUJCJlTH2ARMQlrFy5ssDjpk2bAtC0aVPWr19Penq68/nff/8dNzc3GjduTGBgIHXq1GHx4sXlWrOIWEctQCJSKWRlZZGQkJBvm4eHB2FhYQDMmTOH9u3bc9lll/Hpp5+yevVqPvjgAwCGDBnCM888w7Bhw5g4cSJHjhzhwQcf5PbbbycyMhKAiRMnct999xEREUG/fv1ITU3l999/58EHHyzfExWRcqEAJCKVwoIFC4iOjs63rXHjxvzzzz+AOULr888/5/777yc6OprPPvuMZs2aAeDn58fChQt5+OGHueSSS/Dz8+PGG2/kzTffdB5r2LBhZGZm8p///IexY8cSFhbGTTfdVH4nKCLlymYYhmF1ESIiJWGz2Zg3bx7XXXed1aWISCWhPkAiIiJS5SgAiYiISJWjPkAiUunpSr6IXCy1AImIiEiVowAkIiIiVY4CkIiIiFQ5CkAiIiJS5SgAiYiISJWjACQiIiJVjgKQiIiIVDkKQCIiIlLlKACJiIhIlfP/RVgDNFJ1jqEAAAAASUVORK5CYII=",
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\n",
            "Calculating metrics...\n",
            "Accuracy: 0.79928\n",
            "Precision: [0.86476209 0.75373033]\n",
            "Recall: [0.70952 0.88904]\n",
            "F1 score: [0.77948673 0.81581266]\n",
            "Support: [12500 12500]\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "         0.0       0.86      0.71      0.78     12500\n",
            "         1.0       0.75      0.89      0.82     12500\n",
            "\n",
            "    accuracy                           0.80     25000\n",
            "   macro avg       0.81      0.80      0.80     25000\n",
            "weighted avg       0.81      0.80      0.80     25000\n",
            "\n"
          ]
        },
        {
          "data": {
            "text/plain": [
              "{'val_loss': 0.4600273446981297}"
            ]
          },
          "execution_count": 26,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "# ====== CNN_FROZEN_EMBEDDINGS Function ======\n",
        "CNN_EMBEDDINGS(TEXT, train_data, valid_data, test_data,num_filters = 20,filter_sizes = [3,4,5],dropout=0.6, learning_rate=0.002, num_epochs=17, batch_size=118, clip=4)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ltxnWAqKY9EQ"
      },
      "source": [
        "**PART 6: TRAINING AND EVALUATION OF MODELS WITH GLOVE PRE-TRAINED EMBEDDINGS**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "Svla05weJA-M",
        "outputId": "cf11b288-f675-4882-caa7-21fc7066d351"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "LSTM_Model(\n",
            "  (embedding): Embedding(25002, 100, padding_idx=1)\n",
            "  (lstm): LSTM(100, 16, num_layers=4, dropout=0.6, bidirectional=True)\n",
            "  (dense): Linear(in_features=32, out_features=1, bias=True)\n",
            "  (dropout): Dropout(p=0.6, inplace=False)\n",
            ")\n",
            "The model has 2,534,537 trainable parameters\n",
            "\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:02<00:00, 78.10it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 170.86it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 01 | Epoch Time: 0m 2s | Frozen: False\n",
            "\tTrain Loss: 0.646 | Train Acc: 59.72%\n",
            "\t Val. Loss: 0.476 |  Val. Acc: 79.39%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:01<00:00, 86.76it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 166.35it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 02 | Epoch Time: 0m 2s | Frozen: False\n",
            "\tTrain Loss: 0.476 | Train Acc: 78.43%\n",
            "\t Val. Loss: 0.385 |  Val. Acc: 83.83%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:01<00:00, 85.84it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 175.73it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 03 | Epoch Time: 0m 2s | Frozen: False\n",
            "\tTrain Loss: 0.351 | Train Acc: 85.75%\n",
            "\t Val. Loss: 0.305 |  Val. Acc: 87.67%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:01<00:00, 85.63it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 168.63it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 04 | Epoch Time: 0m 2s | Frozen: False\n",
            "\tTrain Loss: 0.284 | Train Acc: 88.96%\n",
            "\t Val. Loss: 0.326 |  Val. Acc: 85.89%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:02<00:00, 67.09it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 113.63it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 05 | Epoch Time: 0m 2s | Frozen: False\n",
            "\tTrain Loss: 0.242 | Train Acc: 91.03%\n",
            "\t Val. Loss: 0.327 |  Val. Acc: 88.34%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:02<00:00, 74.36it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 156.76it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 06 | Epoch Time: 0m 2s | Frozen: False\n",
            "\tTrain Loss: 0.197 | Train Acc: 92.73%\n",
            "\t Val. Loss: 0.330 |  Val. Acc: 88.05%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:02<00:00, 75.43it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 167.91it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 07 | Epoch Time: 0m 2s | Frozen: False\n",
            "\tTrain Loss: 0.175 | Train Acc: 93.91%\n",
            "\t Val. Loss: 0.327 |  Val. Acc: 88.50%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:02<00:00, 65.18it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 171.30it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 08 | Epoch Time: 0m 2s | Frozen: False\n",
            "\tTrain Loss: 0.157 | Train Acc: 94.39%\n",
            "\t Val. Loss: 0.364 |  Val. Acc: 87.08%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:01<00:00, 86.85it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 174.39it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 09 | Epoch Time: 0m 2s | Frozen: False\n",
            "\tTrain Loss: 0.152 | Train Acc: 94.65%\n",
            "\t Val. Loss: 0.418 |  Val. Acc: 86.42%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:02<00:00, 68.22it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 111.12it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 10 | Epoch Time: 0m 2s | Frozen: False\n",
            "\tTrain Loss: 0.130 | Train Acc: 95.45%\n",
            "\t Val. Loss: 0.392 |  Val. Acc: 87.99%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:03<00:00, 51.13it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 165.18it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 11 | Epoch Time: 0m 3s | Frozen: False\n",
            "\tTrain Loss: 0.108 | Train Acc: 96.26%\n",
            "\t Val. Loss: 0.394 |  Val. Acc: 87.63%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:01<00:00, 85.47it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 164.54it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 12 | Epoch Time: 0m 2s | Frozen: False\n",
            "\tTrain Loss: 0.099 | Train Acc: 96.63%\n",
            "\t Val. Loss: 0.434 |  Val. Acc: 87.49%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:01<00:00, 88.20it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 163.83it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 13 | Epoch Time: 0m 2s | Frozen: False\n",
            "\tTrain Loss: 0.093 | Train Acc: 96.81%\n",
            "\t Val. Loss: 0.444 |  Val. Acc: 87.21%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:01<00:00, 85.96it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 168.91it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 14 | Epoch Time: 0m 2s | Frozen: False\n",
            "\tTrain Loss: 0.092 | Train Acc: 96.89%\n",
            "\t Val. Loss: 0.446 |  Val. Acc: 87.54%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:02<00:00, 80.15it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 112.75it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 15 | Epoch Time: 0m 2s | Frozen: False\n",
            "\tTrain Loss: 0.081 | Train Acc: 97.36%\n",
            "\t Val. Loss: 0.459 |  Val. Acc: 87.00%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:02<00:00, 63.94it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 172.74it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 16 | Epoch Time: 0m 2s | Frozen: False\n",
            "\tTrain Loss: 0.072 | Train Acc: 97.67%\n",
            "\t Val. Loss: 0.551 |  Val. Acc: 87.06%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:01<00:00, 86.15it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 174.79it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 17 | Epoch Time: 0m 2s | Frozen: False\n",
            "\tTrain Loss: 0.072 | Train Acc: 97.71%\n",
            "\t Val. Loss: 0.486 |  Val. Acc: 86.90%\n"
          ]
        },
        {
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjcAAAHHCAYAAABDUnkqAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAAByB0lEQVR4nO3dd3gU5d7G8e9uek8gpACB0KQIhI6ACAiKigULIKIgdgUFOXoUFRQbemxYOGLH8iLtiGIDIQIqoDQp0juhJKGlkJC68/4xZCUGQkg2mWRzf65rr+zOzs78JtHk5pmn2AzDMBARERFxE3arCxARERFxJYUbERERcSsKNyIiIuJWFG5ERETErSjciIiIiFtRuBERERG3onAjIiIibkXhRkRERNyKwo2IiIi4FYUbERe5/fbbiY2NtboMy+zZswebzcbUqVOd25555hlsNluJPm+z2XjmmWdcWlPPnj3p2bOnS49Z2cXGxnL77beX6rPl8TMQsYLCjbg9m81WosfixYutLrXCXHvttfj7+5Oenn7WfYYMGYK3tzdHjx6twMrO36ZNm3jmmWfYs2eP1aU4LV682Pnf1RdffHHGfbp164bNZqNly5YVXF3ZFITYV1991epSRM7K0+oCRMrb559/Xuj1Z599xoIFC4psb968eZnO88EHH+BwOMp0jIoyZMgQvv32W+bMmcPQoUOLvJ+Zmck333zDFVdcQc2aNUt9nqeeeorHH3+8LKWe06ZNm5gwYQI9e/Ys0nL2008/leu5z8XX15dp06Zx6623Ftq+Z88eli1bhq+vr0WVibg3hRtxe//8w/L777+zYMGCItv/KTMzE39//xKfx8vLq1T1WeHaa68lKCiIadOmnTHcfPPNN2RkZDBkyJAyncfT0xNPT+t+zXh7e1t2boCrrrqKuXPncuTIEcLDw53bp02bRmRkJE2aNOH48eMWVijinnRbSgSzb0bLli1ZvXo1l1xyCf7+/jzxxBOA+Ye+X79+1K5dGx8fHxo1asRzzz1Hfn5+oWP8s8/N6c3377//Po0aNcLHx4eOHTuycuXKYutZtWoVNpuNTz/9tMh78+fPx2az8d133wGQnp7O6NGjiY2NxcfHh4iICC677DLWrFlz1uP7+flxww03EB8fT3JycpH3p02bRlBQENdeey3Hjh3jkUceoVWrVgQGBhIcHMyVV17JunXrir0GOHOfm+zsbB5++GFq1arlPMf+/fuLfHbv3r088MADNG3aFD8/P2rWrMmAAQMK3X6aOnUqAwYMAKBXr15FbjGeqc9NcnIyd955J5GRkfj6+hIXF1fk+1yWn93prrvuOnx8fJg1a1ah7dOmTWPgwIF4eHgU+UxeXh7PPfec85yxsbE88cQTZGdnF9rPMAyef/556tati7+/P7169WLjxo1nrCMlJYXRo0cTExODj48PjRs35uWXXy7XlsaSfJ8Bpk+fTvv27QkKCiI4OJhWrVrx5ptvOt/Pzc1lwoQJNGnSBF9fX2rWrMnFF1/MggULyq12qfrUciNyytGjR7nyyiu5+eabufXWW4mMjATMP6CBgYGMGTOGwMBAfv75Z8aPH09aWhqvvPLKOY87bdo00tPTuffee7HZbPznP//hhhtuYNeuXWdt7enQoQMNGzZk5syZDBs2rNB7M2bMICwsjL59+wJw3333MXv2bEaOHEmLFi04evQov/32G5s3b6Zdu3ZnrWvIkCF8+umnzJw5k5EjRzq3Hzt2jPnz5zN48GD8/PzYuHEjX3/9NQMGDKBBgwYkJSXx3nvv0aNHDzZt2kTt2rXP+T043V133cUXX3zBLbfcQteuXfn555/p169fkf1WrlzJsmXLuPnmm6lbty579uzh3XffpWfPnmzatAl/f38uueQSHnroId566y2eeOIJ563Fs91iPHnyJD179mTHjh2MHDmSBg0aMGvWLG6//XZSUlIYNWpUof1L87M7nb+/P9dddx1ffvkl999/PwDr1q1j48aNfPjhh6xfv/6M359PP/2Um266iX/961/88ccfTJw4kc2bNzNnzhznfuPHj+f555/nqquu4qqrrmLNmjVcfvnl5OTkFDpeZmYmPXr04MCBA9x7773Uq1ePZcuWMXbsWA4dOsSkSZPOeR3nq6Tf5wULFjB48GB69+7Nyy+/DMDmzZtZunSpc59nnnmGiRMnctddd9GpUyfS0tJYtWoVa9as4bLLLnN57eImDJFqZsSIEcY//9Pv0aOHARhTpkwpsn9mZmaRbffee6/h7+9vZGVlObcNGzbMqF+/vvP17t27DcCoWbOmcezYMef2b775xgCMb7/9ttg6x44da3h5eRX6bHZ2thEaGmrccccdzm0hISHGiBEjij3WmeTl5RnR0dFGly5dCm2fMmWKARjz5883DMMwsrKyjPz8/EL77N692/Dx8TGeffbZItf7ySefOLc9/fTThb7Xa9euNQDjgQceKHS8W265xQCMp59+2rntTN/35cuXG4Dx2WefObfNmjXLAIxFixYV2b9Hjx5Gjx49nK8nTZpkAMYXX3zh3JaTk2N06dLFCAwMNNLS0gpdS2l/dosWLTIAY9asWcZ3331n2Gw2Y9++fYZhGMajjz5qNGzY0FnfhRdeWOT7c9dddxU63iOPPGIAxs8//2wYhmEkJycb3t7eRr9+/QyHw+Hc74knnjAAY9iwYc5tzz33nBEQEGBs27at0DEff/xxw8PDw1mXYRhFfgZnUvC9eeWVV866T0m/z6NGjTKCg4ONvLy8sx4rLi7O6NevX7E1ifyTbkuJnOLj48Pw4cOLbPfz83M+T09P58iRI3Tv3p3MzEy2bNlyzuMOGjSIsLAw5+vu3bsDsGvXrnN+Ljc3l6+++sq57aeffiIlJYVBgwY5t4WGhvLHH39w8ODBc9ZyOg8PD26++WaWL19e6FZPQX+Q3r17A+b3xW43f1Xk5+dz9OhRAgMDadq0abG3vs7khx9+AOChhx4qtH306NFF9j39+56bm8vRo0dp3LgxoaGh533e088fFRXF4MGDndu8vLx46KGHOHHiBEuWLCm0f2l/dqe7/PLLqVGjBtOnT8cwDKZPn17o/P+sD2DMmDGFtv/rX/8C4Pvvvwdg4cKF5OTk8OCDDxa67Xem7+OsWbPo3r07YWFhHDlyxPno06cP+fn5/PLLLyW+lpIq6fc5NDSUjIyMYm8xhYaGsnHjRrZv3+7yOsV9KdyInFKnTp0zdkDduHEj119/PSEhIQQHB1OrVi1nZ+TU1NRzHrdevXqFXhf8sTxXR9K4uDiaNWvGjBkznNtmzJhBeHg4l156qXPbf/7zH/766y9iYmLo1KkTzzzzTIn/+BZ0GJ42bRoA+/fv59dff+Xmm2929gdxOBy88cYbNGnSBB8fH8LDw6lVqxbr168v0fWfbu/evdjtdho1alRoe9OmTYvse/LkScaPH+/sJ1Jw3pSUlPM+7+nnb9KkiTOsFSi4jbV3795C20v7szudl5cXAwYMYNq0afzyyy8kJCRwyy23nLU+u91O48aNC22PiooiNDTUWV/B1yZNmhTar1atWoXCGMD27duZN28etWrVKvTo06cPwBn7XJVVSb/PDzzwABdccAFXXnkldevW5Y477mDevHmFPvPss8+SkpLCBRdcQKtWrXj00UfPeDtP5HQKNyKnnN5SUCAlJYUePXqwbt06nn32Wb799lsWLFjg7B9Qkg6ZZ+o0CmaH0HMZNGgQixYt4siRI2RnZzN37lxuvPHGQiOQBg4cyK5du3j77bepXbs2r7zyChdeeCE//vjjOY/fvn17mjVrxpdffgnAl19+iWEYhUZJvfjii4wZM4ZLLrmEL774gvnz57NgwQIuvPDCcu2Q+uCDD/LCCy8wcOBAZs6cyU8//cSCBQuoWbNmhQ25L8vP7nS33HILa9eu5ZlnniEuLo4WLVoUu39JJz4sCYfDwWWXXcaCBQvO+Ljxxhtddq7zFRERwdq1a5k7dy7XXnstixYt4sorryzUz+ySSy5h586dfPzxx7Rs2ZIPP/yQdu3a8eGHH1pWt1R+6lAsUozFixdz9OhRvvrqKy655BLn9t27d1fI+QcNGsSECRP43//+R2RkJGlpadx8881F9ouOjuaBBx7ggQceIDk5mXbt2vHCCy9w5ZVXnvMcQ4YMYdy4caxfv55p06bRpEkTOnbs6Hx/9uzZ9OrVi48++qjQ51JSUgoNby6J+vXr43A42LlzZ6HWmq1btxbZd/bs2QwbNozXXnvNuS0rK4uUlJRC+51PEKhfvz7r16/H4XAUalUouL1Yv379Eh/rfFx88cXUq1ePxYsXO4Px2epzOBxs3769UKfopKQkUlJSnPUVfN2+fTsNGzZ07nf48OEirUqNGjXixIkTzpaainA+32dvb2+uueYarrnmGhwOBw888ADvvfce48aNc7Zg1ahRg+HDhzN8+HBOnDjBJZdcwjPPPMNdd91VYdckVYtabkSKUfAv99P/pZ6Tk8N///vfCjl/8+bNadWqFTNmzGDGjBlER0cXCln5+flFbtFERERQu3btIkOHz6aglWb8+PGsXbu2yNw2Hh4eRVoqZs2axYEDB877egrC1ltvvVVo+5lG7JzpvG+//XaRIfgBAQEARULPmVx11VUkJiYWutWXl5fH22+/TWBgID169CjJZZw3m83GW2+9xdNPP81tt91WbH1Q9Pvx+uuvAzhHlfXp0wcvLy/efvvtQt+jM30fBw4cyPLly5k/f36R91JSUsjLyzvfyzmnkn6f/zn7td1up3Xr1gDO/37/uU9gYCCNGzcu8X/fUj2p5UakGF27diUsLIxhw4bx0EMPYbPZ+Pzzz8/7tkRZDBo0iPHjx+Pr68udd95Z6F/C6enp1K1bl5tuuom4uDgCAwNZuHAhK1euLNTiUZwGDRrQtWtXvvnmG4Ai4ebqq6/m2WefZfjw4XTt2pUNGzbwf//3f4VaDEqqTZs2DB48mP/+97+kpqbStWtX4uPj2bFjR5F9r776aj7//HNCQkJo0aIFy5cvZ+HChUVmTG7Tpg0eHh68/PLLpKam4uPjw6WXXkpERESRY95zzz2899573H777axevZrY2Fhmz57N0qVLmTRpEkFBQed9TSV13XXXcd111xW7T1xcHMOGDeP999933hJdsWIFn376Kf3796dXr16A2bfmkUceYeLEiVx99dVcddVV/Pnnn/z4449FWtMeffRR5s6dy9VXX83tt99O+/btycjIYMOGDcyePZs9e/acdwscQHx8PFlZWUW29+/fv8Tf57vuuotjx45x6aWXUrduXfbu3cvbb79NmzZtnC1XLVq0oGfPnrRv354aNWqwatUq59QHImdl2TgtEYucbSj46UNyT7d06VLjoosuMvz8/IzatWsb//73v4358+cXGX58tqHgZxoySwmG3BbYvn27ARiA8dtvvxV6Lzs723j00UeNuLg4IygoyAgICDDi4uKM//73vyU6doHJkycbgNGpU6ci72VlZRn/+te/jOjoaMPPz8/o1q2bsXz58iLDrEsyFNwwDOPkyZPGQw89ZNSsWdMICAgwrrnmGiMhIaHI9+T48ePG8OHDjfDwcCMwMNDo27evsWXLFqN+/fqFhjobhmF88MEHRsOGDQ0PD49CP5d/1mgYhpGUlOQ8rre3t9GqVatCNZ9+LaX92Z0+FLw4Z/rvLjc315gwYYLRoEEDw8vLy4iJiTHGjh1baNoBwzCM/Px8Y8KECc6fS8+ePY2//vrrjN+f9PR0Y+zYsUbjxo0Nb29vIzw83Ojatavx6quvGjk5Oed1bQXfm7M9Pv/8c8MwSvZ9nj17tnH55ZcbERERhre3t1GvXj3j3nvvNQ4dOuTc5/nnnzc6depkhIaGGn5+fkazZs2MF154oVDdIv9kM4wK/CeoiIiISDlTnxsRERFxKwo3IiIi4lYUbkRERMStKNyIiIiIW1G4EREREbeicCMiIiJupdpN4udwODh48CBBQUEuXb9FREREyo9hGKSnp1O7du0ii7L+U7ULNwcPHiQmJsbqMkRERKQUEhISqFu3brH7VLtwUzDtd0JCAsHBwRZXIyIiIiWRlpZGTExMiZZJqXbhpuBWVHBwsMKNiIhIFVOSLiXqUCwiIiJuReFGRERE3IrCjYiIiLiVatfnRkRE3Ed+fj65ublWlyEu4u3tfc5h3iWhcCMiIlWOYRgkJiaSkpJidSniQna7nQYNGuDt7V2m4yjciIhIlVMQbCIiIvD399ekrG6gYJLdQ4cOUa9evTL9TBVuRESkSsnPz3cGm5o1a1pdjrhQrVq1OHjwIHl5eXh5eZX6OOpQLCIiVUpBHxt/f3+LKxFXK7gdlZ+fX6bjKNyIiEiVpFtR7sdVP1OFGxEREXErCjciIiJVWGxsLJMmTbK6jEpF4UZERKQC2Gy2Yh/PPPNMqY67cuVK7rnnHtcWW8VptJQLHT2RzeET2TSL0oKcIiJS2KFDh5zPZ8yYwfjx49m6datzW2BgoPO5YRjk5+fj6XnuP9O1atVybaFuQC03LjLvr0Q6vLCQx/63wepSRESkEoqKinI+QkJCsNlsztdbtmwhKCiIH3/8kfbt2+Pj48Nvv/3Gzp07ue6664iMjCQwMJCOHTuycOHCQsf9520pm83Ghx9+yPXXX4+/vz9NmjRh7ty5FXy11lK4cZE2MaEYBqzfn8LRE9lWlyMiUq0YhkFmTp4lD8MwXHYdjz/+OC+99BKbN2+mdevWnDhxgquuuor4+Hj+/PNPrrjiCq655hr27dtX7HEmTJjAwIEDWb9+PVdddRVDhgzh2LFjLquzstNtKReJCvGlWVQQWxLT+XX7Efq3rWN1SSIi1cbJ3HxajJ9vybk3PdsXf2/X/Dl99tlnueyyy5yva9SoQVxcnPP1c889x5w5c5g7dy4jR44863Fuv/12Bg8eDMCLL77IW2+9xYoVK7jiiitcUmdlp5YbF+rR1LzvuWTbYYsrERGRqqhDhw6FXp84cYJHHnmE5s2bExoaSmBgIJs3bz5ny03r1q2dzwMCAggODiY5Oblcaq6M1HLjQj0viOC9Jbv4ZdthHA4Du10TTImIVAQ/Lw82PdvXsnO7SkBAQKHXjzzyCAsWLODVV1+lcePG+Pn5cdNNN5GTk1Pscf65dIHNZsPhcLiszspO4caF2tcPI9DHk6MZOfx1MJXWdUOtLklEpFqw2WwuuzVUmSxdupTbb7+d66+/HjBbcvbs2WNtUVWAbku5kLenna6NzEXcFm/VrSkRESmbJk2a8NVXX7F27VrWrVvHLbfcUq1aYEpL4cbFejaNANTvRkREyu71118nLCyMrl27cs0119C3b1/atWtndVmVns1w5Ri2KiAtLY2QkBBSU1MJDnb9ZHsHUk7S7aWfsdtgzbjLCPX3dvk5RESqs6ysLHbv3k2DBg3w9fW1uhxxoeJ+tufz91stNy5WJ9SPJhGBOAz4bccRq8sRERGpdhRuykGPC8wh4ep3IyIiUvEUbsrB6f1uHI5qdddPRETEcgo35aBjgzD8vDw4nJ7N5sQ0q8sRERGpVhRuyoGPp4eGhIuIiFhE4aac9NRSDCIiIpZQuCknPS4w+92s3nuctKxci6sRERGpPhRuykm9mv40DA8g32GwTEPCRUREKozCTTm6REPCRUREKpzCTTk6vd9NNZsIWkREykHPnj0ZPXq083VsbCyTJk0q9jM2m42vv/66zOd21XEqgsJNObqoYU18PO0cSs1iW9IJq8sRERELXXPNNVxxxRVnfO/XX3/FZrOxfv368zrmypUrueeee1xRntMzzzxDmzZtimw/dOgQV155pUvPVV4UbsqRr5cHFzUsGBKebHE1IiJipTvvvJMFCxawf//+Iu998skndOjQgdatW5/XMWvVqoW/v7+rSixWVFQUPj4+FXKuslK4KWcaEi4iIgBXX301tWrVYurUqYW2nzhxglmzZtG/f38GDx5MnTp18Pf3p1WrVnz55ZfFHvOft6W2b9/OJZdcgq+vLy1atGDBggVFPvPYY49xwQUX4O/vT8OGDRk3bhy5ueao3qlTpzJhwgTWrVuHzWbDZrM56/3nbakNGzZw6aWX4ufnR82aNbnnnns4ceLvuxS33347/fv359VXXyU6OpqaNWsyYsQI57nKk2e5n6GaK1hnauWeY5zIziPQR99yERGXMwzIzbTm3F7+YLOdczdPT0+GDh3K1KlTefLJJ7Gd+sysWbPIz8/n1ltvZdasWTz22GMEBwfz/fffc9ttt9GoUSM6dep0zuM7HA5uuOEGIiMj+eOPP0hNTS3UP6dAUFAQU6dOpXbt2mzYsIG7776boKAg/v3vfzNo0CD++usv5s2bx8KFCwEICQkpcoyMjAz69u1Lly5dWLlyJcnJydx1112MHDmyUHhbtGgR0dHRLFq0iB07djBo0CDatGnD3Xfffc7rKQv9pS1nDcIDqFfDn33HMlm+8yiXtYi0uiQREfeTmwkv1rbm3E8cBO+AEu16xx138Morr7BkyRJ69uwJmLekbrzxRurXr88jjzzi3PfBBx9k/vz5zJw5s0ThZuHChWzZsoX58+dTu7b5vXjxxReL9JN56qmnnM9jY2N55JFHmD59Ov/+97/x8/MjMDAQT09PoqKiznquadOmkZWVxWeffUZAgHnt77zzDtdccw0vv/wykZHm37qwsDDeeecdPDw8aNasGf369SM+Pr7cw43lt6UmT55MbGwsvr6+dO7cmRUrVhS7f0pKCiNGjCA6OhofHx8uuOACfvjhhwqq9vzZbLbTVglXvxsRkeqsWbNmdO3alY8//hiAHTt28Ouvv3LnnXeSn5/Pc889R6tWrahRowaBgYHMnz+fffv2lejYmzdvJiYmxhlsALp06VJkvxkzZtCtWzeioqIIDAzkqaeeKvE5Tj9XXFycM9gAdOvWDYfDwdatW53bLrzwQjw8PJyvo6OjSU4u/7+FlrbczJgxgzFjxjBlyhQ6d+7MpEmT6Nu3L1u3biUiIqLI/jk5OVx22WVEREQwe/Zs6tSpw969ewkNDa344s9Dz6a1+Pz3vc4h4bYSNF+KiMh58PI3W1CsOvd5uPPOO3nwwQeZPHkyn3zyCY0aNaJHjx68/PLLvPnmm0yaNIlWrVoREBDA6NGjycnJcVmpy5cvZ8iQIUyYMIG+ffsSEhLC9OnTee2111x2jtN5eXkVem2z2XA4HOVyrtNZGm5ef/117r77boYPHw7AlClT+P777/n44495/PHHi+z/8ccfc+zYMZYtW+b8hsXGxlZkyaXSpVFNvD3s7D9+kp2HM2gcEWh1SSIi7sVmK/GtIasNHDiQUaNGMW3aND777DPuv/9+bDYbS5cu5brrruPWW28FzD4027Zto0WLFiU6bvPmzUlISODQoUNER0cD8PvvvxfaZ9myZdSvX58nn3zSuW3v3r2F9vH29iY/P/+c55o6dSoZGRnO1pulS5dit9tp2rRpieotT5bdlsrJyWH16tX06dPn72Lsdvr06cPy5cvP+Jm5c+fSpUsXRowYQWRkJC1btuTFF18s9oeQnZ1NWlpaoUdF8/f2pFODGoBGTYmIVHeBgYEMGjSIsWPHcujQIW6//XYAmjRpwoIFC1i2bBmbN2/m3nvvJSkpqcTH7dOnDxdccAHDhg1j3bp1/Prrr4VCTME59u3bx/Tp09m5cydvvfUWc+bMKbRPbGwsu3fvZu3atRw5coTs7Owi5xoyZAi+vr4MGzaMv/76i0WLFvHggw9y2223OfvbWMmycHPkyBHy8/OLfBMiIyNJTEw842d27drF7Nmzyc/P54cffmDcuHG89tprPP/882c9z8SJEwkJCXE+YmJiXHodJVUwJFz9bkRE5M477+T48eP07dvX2Ufmqaeeol27dvTt25eePXsSFRVF//79S3xMu93OnDlzOHnyJJ06deKuu+7ihRdeKLTPtddey8MPP8zIkSNp06YNy5YtY9y4cYX2ufHGG7niiivo1asXtWrVOuNwdH9/f+bPn8+xY8fo2LEjN910E7179+add945/29GObAZFq0LcPDgQerUqcOyZcsKdXj697//zZIlS/jjjz+KfOaCCy4gKyuL3bt3Ozsovf7667zyyiscOnTojOfJzs4ulDrT0tKIiYkhNTWV4OBgF1/V2W1PSueyN37B29POuvGX4+ftce4PiYhIEQV/Bxo0aICvr6/V5YgLFfezTUtLIyQkpER/vy3rcxMeHo6Hh0eRJrekpKSzDj+Ljo7Gy8urUM/r5s2bk5iYSE5ODt7e3kU+4+PjUylmVGwcEUidUD8OpJzk911H6dWsaIdpERERKTvLbkt5e3vTvn174uPjndscDgfx8fFnHLoG5jCzHTt2FOppvW3bNqKjo88YbCoTm8122irhujUlIiJSXiyd52bMmDF88MEHfPrpp2zevJn777+fjIwM5+ipoUOHMnbsWOf+999/P8eOHWPUqFFs27aN77//nhdffJERI0ZYdQnnRUsxiIiIlD9Lh4IPGjSIw4cPM378eBITE2nTpg3z5s1zdjLet28fdvvf+SsmJob58+fz8MMP07p1a+rUqcOoUaN47LHHrLqE89K1UU087Tb2HM1kz5EMYsOrxrBFERGRqsTy5RdGjhzJyJEjz/je4sWLi2zr0qVLkXH7VUWQrxcdYsP4fdcxlmw7rHAjIlIGFo2HkXLkqp+p5csvVDc9LjA7EqvfjYhI6RRM4pqZadFCmVJuCmZjPn3gUGlY3nJT3fRsWouX521h+a6jZOXm4+ulIeEiIufDw8OD0NBQ5xpF/v7+WtbGDTgcDg4fPoy/vz+enmWLJwo3FaxZVBCRwT4kpWWzYvcx5wgqEREpuYIpQypiEUapOHa7nXr16pU5rCrcVLCCVcJnrtrP4q2HFW5ERErBZrMRHR1NREQEubm5VpcjLuLt7V1oIFFpKdxYoGfTCGau2s+SbclAyRZEExGRojw8PMrcP0PcjzoUW6Bb43A87DZ2Hs4g4Zg6xImIiLiSwo0FQvy8aFcvFNCEfiIiIq6mcGORHs6lGBRuREREXEnhxiI9m5rz3SzbeYScPMc59hYREZGSUrixSIvoYMIDfcjMyWfVnmNWlyMiIuI2FG4sYrfbuOSCcED9bkRERFxJ4cZCBbem1O9GRETEdRRuLNS9cTh2G2xNSudgykmryxEREXELCjcWCgvwJi4mFIBfdGtKRETEJRRuLKYh4SIiIq6lcGOxgn43S3ccITdfQ8JFRETKSuHGYq3qhBDm70V6dh5r9h63uhwREZEqT+HGYh52m3NlcA0JFxERKTuFm0qgZ1P1uxEREXEVhZtKoHsTM9xsOpRGclqWxdWIiIhUbQo3lUB4oA+t64YAujUlIiIulJsF66bDyRSrK6lQCjeVhHNIuMKNiIi4yoLxMOde+HIw5OdZXU2FUbipJAr63fy2/Qh5GhIuIiJldXwPrPrYfL5vGSx52dJyKpLCTSURVzeUYF9PUk/msm5/itXliIhIVbf4JXDkQkg98/Uvr8CuJdbWVEEUbioJTw873QuGhGvUlIiIlEXyZrOvDcDAqdBuKGDAV3fDiWQrK6sQCjeVSE/1uxEREVf4+XnAgObXQJ32cMXLUKs5nEgy++A43Lv7g8JNJVLQqXj9/lSOnMi2uBoREamS9q+GLd+BzQ6XjjO3efvDgE/A0w92/gzL3rS2xnKmcFOJRAT70iI6GIBft6v1RkRESuHnZ82vcYOhVtO/t0c0hytPdSqOfw4SVlR8bRVE4aaS6dFU/W5ERKSUdi2BXYvB7gU9Hiv6fruh0PJGMPJh9h1w0j3XNFS4qWQK+t38sv0IDodhcTUiIlJlGAbETzCfd7gDwuoX3cdmg6snQVgDSE2Ab0aan3MzCjeVTLv6YQT5eHIsI4f1B1KtLkdERKqKrT/AgdXg5Q+XPHL2/XyDzf43di+zb86KDyquxgqicFPJeHnY6dY4HNCtKRERKSFHvtmPBuCi+yEwovj9a7eFy0/t/9OTcGhd+dZXwRRuKqGCfjeLt7n/XAQiIuICG2bB4c3gGwpdHyrZZzrfB02vgvwcmDUcstPLtcSKpHBTCRUsxbAuIYXjGTkWVyMiIpVaXg4setF8fvFo8Ast2edsNrhuMgTXgWM74ft/uU3/G4WbSig6xI+mkUE4DPh1xxGryxERkcpszaeQshcCI6HTvef3Wf8acONHYPOA9TNg7bTyqbGCKdxUUhoSLiIi55STYa4ZBXDJo+ZkfeerfhfoNdZ8/sMjcHir6+qziMJNJVUwJHzJtsMaEi4iImf2x3vmkgqh9aHdsNIf5+Ix0LAn5Gaa/W9yT7qsRCso3FRS7WPD8Pf24MiJbDYdSrO6HBERqWxOpsDSSebzXk+Ap3fpj2X3gOvfh4BakLwR5o11RYWWUbippHw8Peja6NSQcC2kKSIi/7TsLchKNRfEbDWg7McLioQb3gdssPoT2Din7Me0iMJNJeYcEr5VQ8JFROQ06Unw+7vm897jzJYXV2h0KVz8sPl87kNwbLdrjlvBFG4qsYJ+N2v2pZB6MtfiakREpNL49TWzf0ydDuZcNa7U60mI6QzZaeb6U3lVb0oShZtKLKaGP41qBZDvMFiqIeEiIgJwfC+s+th83nu8OV+NK3l4msPDfUPh4Jq/16uqQhRuKrkeF5hTaGtIuIiIALD4JXDkmqObGvYon3OExpgT/AEsfwe2zS+f85QThZtKrmC24iXbDmO4ycyRIiJSSslbYP108/ml48v3XM2v/ntSwDn3QeqB8j2fCyncVHKdGtTA18tOYloWW5PcZ90PEREphUXPg+GAZldD3fblf77Ln4PoODh5DL66G/Lzyv+cLqBwU8n5ennQpWFNABbr1pSISPV1YDVs/hawwaXjKuacnj5w0yfgHQh7l8Iv/6mY85aRwk0V0OMCLcUgIlLtxT9nfo27GSKaVdx5azaCqyeZz5f8B3b/UnHnLqVKEW4mT55MbGwsvr6+dO7cmRUrVpx136lTp2Kz2Qo9fH19K7DaitezqdmpeNXeY5zIrhpNgiIi4kK7f4Fdi8DuBT0tmD249QBoeytgwP/uhhOV+x/bloebGTNmMGbMGJ5++mnWrFlDXFwcffv2JTn57BPXBQcHc+jQIedj7969FVhxxYsND6B+TX9y8zUkXESk2jEMWHhqOHaH4RBW35o6rvwPhDeFE4nw9X3gcFhTRwlYHm5ef/117r77boYPH06LFi2YMmUK/v7+fPzxx2f9jM1mIyoqyvmIjIyswIqtcfpCmiIiUo1s/REOrAIvf+j+iHV1eAfAgKng6Qs7FsLyt62r5RwsDTc5OTmsXr2aPn36OLfZ7Xb69OnD8uXLz/q5EydOUL9+fWJiYrjuuuvYuHHjWffNzs4mLS2t0KMqKrg1tWSrhoSLiFQbjnz4+VRfm873mes/WSmyBVzxkvk8/llIWGltPWdhabg5cuQI+fn5RVpeIiMjSUxMPONnmjZtyscff8w333zDF198gcPhoGvXruzfv/+M+0+cOJGQkBDnIyYmxuXXUREualgTb087B1JOsvPwCavLERGRirBhNiRvAt8Q6PaQ1dWY2t8OF94AjjxzeYaTx62uqAjLb0udry5dujB06FDatGlDjx49+Oqrr6hVqxbvvffeGfcfO3YsqampzkdCQkIFV+waft4edG5QA9CQcBGRaiEvBxa/aD7vNgr8wqytp4DNBte8CWGxkLoP5j5o9guqRCwNN+Hh4Xh4eJCUlFRoe1JSElFRUSU6hpeXF23btmXHjh1nfN/Hx4fg4OBCj6qqh/rdiIhUH39+Bsf3QECEeUuqMvENNue/sXuZc++s/NDqigqxNNx4e3vTvn174uPjndscDgfx8fF06dKlRMfIz89nw4YNREdHl1eZlUZBv5s/dh0jM0dDwkVE3FZOJix5xXze499mZ97Kpk47uOzUKK75T8Kh9dbWcxrLb0uNGTOGDz74gE8//ZTNmzdz//33k5GRwfDhwwEYOnQoY8f+Pab/2Wef5aeffmLXrl2sWbOGW2+9lb1793LXXXdZdQkVplGtAOqE+pGT72D5zqNWlyMiIuVlxfvmkOvQetBumNXVnN1FD8AFV0B+NsweDtmVo0+o5eFm0KBBvPrqq4wfP542bdqwdu1a5s2b5+xkvG/fPg4dOuTc//jx49x99900b96cq666irS0NJYtW0aLFi2suoQKY7PZCi2kKSIibuhkCvz2hvm85xPg6W1pOcWy2eC6/0JQbTi6A36wcKj6aWxGNRtXnJaWRkhICKmpqVWy/82CTUnc/dkq6tXwZ8mjPbHZbFaXJCIirhT/HPz6KtRqDvcvBbuH1RWd295lMLWfuahn/3ehzS0uP8X5/P22vOVGzk+XRjXx8rCx71gme45mWl2OiIi40olk+P1d8/mlT1WNYANQv6vZygTw/b/g8DZLy1G4qWICfTzpGFswJPzsS1SIiEgV9OtrkJsBddpDs35WV3N+uo+BBpdAbqbZ/yY3y7JSFG6qIA0JFxFxQyn7YNWppYd6jzf7s1Qldg+44QMIioaWN4KHdX2FFG6qoIIh4ct3HiUrN9/iakRExCUWvwz5OdCgBzTsaXU1pRMUBQ+uNltx7NZFDIWbKuiCyECign3JznPwx+5jVpcjIiJldXgrrJtmPu893tpayqoSzMmjcFMFnT4kXP1uRETcwM/PmyONml0NdTtYXU2Vp3DjKoYBm78zl6avAM5+N1pnSkSkajuwBjbPBWzmCCkpM4UbV1n7fzBjCPz4b8jLLvfTdWsSjpeHjV1HMth8KK3czyciIuXk5+fMr60HQURza2txEwo3rnLhDRAYZfZ2Xz213E8X7OtF72bmLM4zV1XNlc5FRKq93b/Czp/NBSh7jT33/lIiCjeu4u0PPR8zny/5D2Snl/spB3asC8DXfx4gO0+jpkREqhTDgPhTC0+2vx3CYq2sxq0o3LhS29ugRkPIPPL3DJPl6JImtYgM9uF4Zi7xm9WxWESkStk2D/avBE8/uKRyrMnkLhRuXMnD6+/OYEvfgozyXbnb08POje3M1psZK3VrSkSkynA4zDWkAC66z5wfRlxG4cbVWlwPUa0hJx1+e73cTzewQwwAv2w/zMGUk+V+PhERcYG/ZkPyRvAJgW6jrK7G7SjcuJrdDn2eNp+v+ABSyrdFJTY8gE4NamAY8NWa/eV6LhERcYH8XFj0gvm820PgF2ZtPW5I4aY8NOoNsd0hPxuWvFTupytovZm5aj8Oh1Hu5xMRkTJY8xkc3wMBEXDR/VZX45YUbsqDzQa9T7XerJ1mTqtdjq5qFUWgjyf7jmVqOQYRkcosJ9McUQtwyaOVYqkCd6RwU15iOprTaBuOvydoKif+3p5cExcNaM4bEZFKKSXBnANt+mA4kQgh9aD9MKurclsKN+Xp0qfAZofN38L+1eV6qgGnbk39sOEQaVm55XouERE5h9yTsH0hzBsL73SCSS3h21Gwa7H5/mUTwNPH0hLdmafVBbi1iOYQN9hcmmHh0zDsW/OWVTloGxNKk4hAtief4Nt1BxnSuX65nEdERM7AMODwFtgRDzvjYc9Ss99lAZsd6nSAxr2h6ZUQHWddrdWAwk156zkWNsyCPb/CrkXQ6NJyOY3NZmNghxhe+GEzM1cmKNyIiJS3k8fNlpgd8eYSCmkHCr8fXMf8nd+4DzTsoVFRFUjhpryFxkDHu+H3ybBwAjToaQ4XLwfXt6vDy/O2sG5/KlsS02gWFVwu5xERqZYc+eYK3jvjYcdCOLDa7FdZwNMX6nc1w0yj3lCrabm11kvxFG4qQvcx5tC/Q2th8zdw4fXlcprwQB96N49g/sYkZq3az7irW5TLeUREqo3UA6fCTLzZSpOVUvj98KZmmGl8KdTvBl5+VlQp/6BwUxECwqHrg7D4RXO67WZXm0s1lIOBHWKYvzGJOX8e4LErmuHtqT7jIiIllpsFe5eat5l2xMPhzYXf9wmBRj3NlpnGvSGkriVlSvEUbipKlwdgxftwbCf8+QV0GF4up+lxQS0ignxITs8mfnMSV7aKLpfziIi4BYcDjm4/rSPwb5CXddoONqjT3gwyjXqbzz30p7Oy00+oovgEmRM2zXsMlrwMrQeBt7/LT+PpYefG9nV5d/FOZq5KULgRkaopPw/yTppDqnMzT309/XFq27n2ycsq/vOnj2gqEBR9qmXmUmjYC/xrVPz1S5ko3FSkDsNh+WRI3We24lw8ulxOM+BUuFmy7TCJqVlEhfiWy3lERMos85jZH/Hgn3BwrflIPwSOCpyvy8Pb7AjcqLfZfyaiuToCV3EKNxXJ0wd6PQFf32euGN5+WLkMDWxYK5BOsTVYsecY/1uznxG9Grv8HCIi5+1kChxad1qY+dNcY+lcvPzNjrpe/uaIpILnXn7/eBS8f9r+Xr6F9/X0K/p5n2Dw9C7ni5eKpHBT0VoPhKVvmp3Ulr719wriLjagQ11W7DnGzFUJ3N+jEXa7/hUiIhUoOx0Orf87xBz80+xzeCZhDaB2278fNRr8HT48fdWKIudN4aai2T2g93hzfZHf34XO90JQlMtP0691NM/M3cjeo5ms2HOMixrWdPk5REQAyMmAxA2Fg8yR7YBRdN/QeoWDTHScJrcTl1O4sULTK6FuJ9i/wlwd9urXXX4KczHN2kxfmcDMVQkKNyLiGrknIfGvv0PMobXmsgOnT2ZXILgu1G5zWpBpAwH6XSTlz2YYxhmitftKS0sjJCSE1NRUgoMtnMF3z1KYehXYPWHECqjZyOWnWL33ODe+uwxfLzsrnuxDsG/5zK0jIm7s8DZz+ZiCDr/Jm8DIL7pfUPTfAaZ2WzPUBEZUcLHizs7n77dabqwS2w0aXwY7FsCiF+Gmj1x+inb1QmkcEciO5BN8t+4Qt3Su5/JziIibMgxY9jYsGE+R20sBtf5xa6kNBGvaCak8FG6s1Hu8GW7+mg3dRkF0a5ce3lxMsy4v/rCFmasSFG5EpGTycuD7MfDn5+br+hdDvYv+DjPBtdXJVyo1zc1vpejW0PIm83n8s+Vyiuvb1sXTbmNtQgrbktLL5Rwi4kZOHocvbjCDjc0OV7wMt38HvcdB86shpI6CjVR6CjdW6/WE2e9mxwJz2m8XqxXkw6XNzPveM1cmuPz4IuJGju6ED/uYfWy8A2HwdLjoPoUZqXIUbqxWsxG0G2Y+XzjBvM/tYgM7xADw1Z8HyMk7w4gGEZE9S+HD3nB0hznK6Y75cEFfq6sSKRWFm8qgx7/NWTP3r4CtP7r88D2b1qJWkA/HMnL4eUuSy48vIlXc2mnw2XXmLak67eHunyGqpdVViZSawk1lEBQFF91vPo9/FhxnGGZZBp4edm5sVxeAmav2u/TYIlKFORxmi/HX95trObXoD7d/D0GRVlcmUiYKN5VFt1HgG2ouy7B+pssPP6CDGW4Wb00mMTXL5ccXkSomJxNmDTPXuQPo/gjc9Im55IFIFadwU1n4hcLFD5vPF70IedkuPXyjWoF0jA3DYcD/1qj1RqRaS080JxHdPBfsXtB/ijkayq4/CeIe9F9yZdLpHnOWz9R9sOoTlx9+wKmOxbNWJVDNJqaW6mLVJ/DFjbB9Qbl0zncLiRvgg0vNGYf9asCwudBmsNVVibiUwk1l4u0PPR4zn//yirmqrgv1axVNgLcHe45msmL3MZceW8Rya7+E70bDjoXwfzeZHWQPrbO6qspl64/wUV9IOwDhF8Dd8VC/q9VVibicwk1l0/ZWqNEIMo/A8v+69NABPp5c3bo2oI7F4mZ2xMPckebz+t3Awxt2L4H3esBX90JKNZ/jyTBg+WT4cjDkZkCDHnDnT1CjodWViZQLhZvKxsMLLn3KfL7sbcg44tLDD+xodiz+YcMh0rNyXXpsEUscXAszh4Ijz5zxe9h3MHLlqdm/DVg/Hd5uDwuehqxUq6utePm58N3DMP8JwID2t8Ot/wO/MKsrEyk3CjeVUYv+EB0HOenw6+suPXS7emE0qhXAydx8vlt/yKXHFqlwx3bD/w2AnBPQ4BLo/1+zU2xYrLkY7d2LzHWR8rNh6SR4sw38PsVcO6k6OJlifn9WfwLY4PIX4OpJ5j+iRNyYwk1lZLdD76fN5ys/cGmTurmYptmxeOaqat5UL1VbxlGz83BGMkS2hEFfgKdP4X3qtDPXRRo83exjcvIYzHsM/tsZNn7t3p2Oj+2Gjy6HXYvAKwBungZdR2opBakWFG4qq0aXQmx3yM+BxS+59NDXt6uDh93Gn/tS2K7FNKUqysmEaQPh2E4IiYEhs8E35Mz72mzQ9Eq4fzlc/QYERMCxXeYcLx9dDvv+qNjaK8Le5eaIqCNbIag23DEPml1ldVUiFUbhprKy2aDPM+bzddMgeYvLDh0R5Pv3YppqvZGqJj8PZt8BB1aZE1/e+j8Ijj735zw8ocMd8NAac1Sil7+55MnHl8OMW81FI93Buhnw2bVmK1V0G3MphejWVlclUqEqRbiZPHkysbGx+Pr60rlzZ1asWFGiz02fPh2bzUb//v3Lt0Cr1O0Aza4GwwE/P+fSQzsX01xzgNx8LaYpVYRhwPdjYNuP4OkLt8yAWk3P7xg+QdDrCXhwDbQbCjY7bP4WJneCHx51eSf+CuNwwM/Pw5x7zBbf5tfA8B9KFvxE3Izl4WbGjBmMGTOGp59+mjVr1hAXF0ffvn1JTk4u9nN79uzhkUceoXv37hVUqUUuHWf+8t3yHexf5bLD9mxai/BAH45m5BC/ufjvtUilseQ/sOZTwAY3fgj1Lir9sYKj4dq34f5l0ORyc7TVivfhrbZmR/7cky4ru9zlnoT/3WHOjwXQbTQM+Ay8AywtS8Qqloeb119/nbvvvpvhw4fTokULpkyZgr+/Px9//PFZP5Ofn8+QIUOYMGECDRu6+TwNEc0g7hbz+cJnXNYB0svDzo3t6wDmjMUild6az2Dxi+bzq14xWyZcIaI5DJkFQ+dCVGvIToP4Cebw8bXTzBaRyuxEMky9GjbOAbsnXDcZLpugpRSkWrP0v/6cnBxWr15Nnz59nNvsdjt9+vRh+fLlZ/3cs88+S0REBHfeeec5z5GdnU1aWlqhR5XT83FzUrI9v8LOn1122AHtzVtTi7Ymk5SmxTSlEts2H74dbT7v/i/odLfrz9GwB9yzBK5/3+yknHbAXC37/Utg5yLXn88VkjaaHYcL+h/d9rU5EahINWdpuDly5Aj5+flERkYW2h4ZGUliYuIZP/Pbb7/x0Ucf8cEHH5ToHBMnTiQkJMT5iImJKXPdFS40Bjqe+mUeP8Fl/5JsHBFIh/paTFMquf2rYdbtYORD3GDzVm15sdshbhCMXAV9JoBPiLkW0+f9zWHnSRvL79zna9tP5lIKqQnmrOZ3/wwN3Pw2vUgJVal2y/T0dG677TY++OADwsPDS/SZsWPHkpqa6nwkJFTRWzDdx4B3kLlWzqavXXbYgc7FNPdrMU2pfI7uhGkDIDcTGvU2+8hUxDwtXr5w8WgYtRY632+unL1jIUy5GL4ZAWkHy7+G4vzxHnw5yJzoM7Y73LUQajaytiaRSsTTypOHh4fj4eFBUlJSoe1JSUlERUUV2X/nzp3s2bOHa675+16741QrhqenJ1u3bqVRo8L/g/v4+ODj84+JvaqigHDo+qDZ5+Dn583+Bi6YZfSq1tE88+1Gdh/JYOWe43RqUMMFxYq4wIlk+OIGyDxqDmke+GnFz6zrXwOufAk63wMLJ5j/sPjzC9jwP3NCvG6jzNFXrpKfawa53JOnfT39kWneml7zqbl/21uh3xvg6e26GkTcgKXhxtvbm/bt2xMfH+8czu1wOIiPj2fkyJFF9m/WrBkbNmwotO2pp54iPT2dN998s2recjofXR4wR3Mc22n+gu0wvMyHDPTx5OrW0cxctZ+ZqxIUbqRyyD5hLhtwfA+E1jc7/LoyRJyvGg3NcJWwEn56ChJ+N0cmrZ5q9gEKrv2PQJJZOJDkZp1h22lf806978grYUGn5sHqNkozDoucgaXhBmDMmDEMGzaMDh060KlTJyZNmkRGRgbDh5t/uIcOHUqdOnWYOHEivr6+tGzZstDnQ0NDAYpsd0s+QXDJo+b08UtehtaDwNu/zIcd2CGGmav28/36Qzxz7YUE+lj+n4VUZ/m55uzBh9aCf024bQ4ERlhdlSmmoznb75bvzIU4j+2EeY+7/jw2uznJoJffqYf/3199gqD9cGh6hevPK+ImLP8rNmjQIA4fPsz48eNJTEykTZs2zJs3z9nJeN++fdg1pPFvHYbD8smQus9sxbl4dJkP2b5+GA1rBbDrcAbfrTvIzZ3qlb1OkdIwDJj7kNm/xdMPbplZ+fqS2GzmbeELrjBbbv76H2A7cxDx8j17SPHyM6/xTO95eKlFRqQMbEY160WalpZGSEgIqampBAcHW11O6az9Er6+z1xLZ9Q68Asr8yGnLNnJSz9uoV29UL56oJsLihQphfjn4NdXzZaLm79U64SIOJ3P3281iVRFrQdCreaQlQpL33LJIW84tZjmmn0p7EjWYppigZUfmsEG4OpJCjYiUmoKN1WR3QN6jzef//4upOwr8yEjgnzp1bRgMU3NeSMVbPN35rpOAD3HQvth1tYjIlWawk1V1fRKqN8N8k7C/CdccsiBHeoC8NWa/VpMUyrOvj/gf3eaC8S2G2qu2C0iUgYKN1WVzWaur2PzMFc03rGwzIfs1SyC8EAfjpzI4ectWkxTKsDhbeZkdHlZ0KSvOWeLOtKKSBkp3FRlkRdC53vN5z/8G/Kyy3Q4Lw87N7bTYppSQdIOmZP0nTwOddrDgE/Aw/IBnCLiBkoVbhISEti//+9+GStWrGD06NG8//77LitMSqjn4xAYac63sfydMh9uQIeCxTQPk6zFNKW8ZKWZk/QVrIt0y0zwDrC6KhFxE6UKN7fccguLFpmr5CYmJnLZZZexYsUKnnzySZ599lmXFijn4BsClz1nPl/yCqSUrcWlcUQg7euHke8w+N+aAy4oUOQf8nJgxq2QtAECasGt/zOXFxERcZFShZu//vqLTp06ATBz5kxatmzJsmXL+L//+z+mTp3qyvqkJFoPhHpdXda5uKBj8axVCVpMU1zL4YBvHoDdS8ArwFxWoUYDq6sSETdTqnCTm5vrXIxy4cKFXHvttYC59tOhQ4dcV52UjM0G/V491bl4LuyIL9Ph+rWujb+3B7uOZLB673EXFSkCLHwaNswCuycM/Axqt7W6IhFxQ6UKNxdeeCFTpkzh119/ZcGCBVxxhTnZ1sGDB6lZs6ZLC5QSOr1z8Y9l61wc6ONJv1bRAMxYqY7F4iK/vwvLTk06ee3b0KSPtfWIiNsqVbh5+eWXee+99+jZsyeDBw8mLi4OgLlz5zpvV4kFej4OARFwdEeZOxcP7Gh2LP5+wyFOZJd0pWKRs9g4B+aNNZ/3Hg9tbrG2HhFxa6VeWyo/P5+0tDTCwv5e12jPnj34+/sTEVFJVvA9A7dYW6o462bAnHvMxfdGrIDQmFIdxjAMer+2hF1HMnj5xlYM6qjFNKWU9vwGn18P+TnQ8S646lXNZSMi563c15Y6efIk2dnZzmCzd+9eJk2axNatWyt1sKkWCjoX52aWqXOxzWZzDgvXcgxSakmb4MtbzGDT7Gq48j8KNiJS7koVbq677jo+++wzAFJSUujcuTOvvfYa/fv3591333VpgXKeCs1cXLbOxTeeWkxz9d7j7Eg+4cIipVrY/St8eg1kp0JMZ7jxQ3NdNBGRclaqcLNmzRq6d+8OwOzZs4mMjGTv3r189tlnvPWWa1apljKIagmd7jGfl6FzcUSwL72a1gI0Y7GcB8MwV6v/7DrIPAKRrWDwdPDys7oyEakmShVuMjMzCQoKAuCnn37ihhtuwG63c9FFF7F3716XFiil1GvsaZ2LJ5f6MAW3pv635oAW05Rzy06HWcNgwTgw8qH1ILjzJ/CvYXVlIlKNlCrcNG7cmK+//pqEhATmz5/P5ZdfDkBycrJ7dtKtinxD4PJTMxf/UvqZiy9tFkF4oDdHTmSzeOthFxYobufwVvjgUtj0Ddi9zI7D178H3v5WVyYi1Uypws348eN55JFHiI2NpVOnTnTp0gUwW3HattWkXJVG60FQr4vZufinJ0t1CC8POze0M2cs1pw3clYbvzaDzZFtEBQNw3+ATner87CIWKJU4eamm25i3759rFq1ivnz5zu39+7dmzfeeMNlxUkZ2Wynht16mP+a3vlzqQ4zoL0ZbhZtTSY5XYtpymny8+CnceatqJwTUP9iuPcXiNF8VyJinVKFG4CoqCjatm3LwYMHnSuEd+rUiWbNmrmsOHGB0zsX//BoqToXN4kMom29UPIdBl9pMU0pcCIZPu//96zDXR+Eod9AoKaDEBFrlSrcOBwOnn32WUJCQqhfvz7169cnNDSU5557DodDnU4rHRd0Lh7knPNGi2kKkLAS3usBe34F70AY8Clc/jx4eFpdmYhI6cLNk08+yTvvvMNLL73En3/+yZ9//smLL77I22+/zbhx41xdo5TVPzsXp57/pHz9Wkfj5+XBrsMZfL9Bi6NWW4YBKz6AT66E9IMQfgHc/TNc2N/qykREnEq1/ELt2rWZMmWKczXwAt988w0PPPAABw5U3lsXbr/8wtkYhvkHad9yaHGduSLzeXp1/lbeWbSDYF9P5o2+hNqhmrekWsnJhO/HwLovzdfNr4X+/wWfIGvrEpFqodyXXzh27NgZ+9Y0a9aMY8eOleaQUt5c0Ll4VJ8mxNUNIS0rj4dnrCXfodtT1caxXfDR5WawsdnhsufMgKxgIyKVUKnCTVxcHO+8U3TV6XfeeYfWrVuXuSgpJ4U6F/8b8nLO6+NeHnbevLkt/t4e/LH7GFOW7CyHIqXS2TYf3u8JSRvAP9zsNNztIQ3zFpFKq1S3pZYsWUK/fv2oV6+ec46b5cuXk5CQwA8//OBcmqEyqra3pQpkpcLbHSAjGfo8Axc/fN6HmLkqgX/PXo+n3cb/7u9KXEyoy8uUSsCRD0teNh8AdTuaHYdD6lhbl4hUS+V+W6pHjx5s27aN66+/npSUFFJSUrjhhhvYuHEjn3/+eamKlgpyeufiJf8pVefiAe3r0q9VNHkOg1HT/yQjO8/FRYrlMo/BtIF/B5uOd8Ht3yvYiEiVUKqWm7NZt24d7dq1Iz8/31WHdLlq33ID/+hc3B8Gfnreh0jNzOXKN3/hYGoWA9rX5ZUBca6vU6xxaB3MuBVS9oGnL1w9CdoMtroqEanmyr3lRqq4Qp2Lv4adi877ECH+Xrw+qA02G8xavZ/v1h90fZ1S8dZOMzsOp+yDsFi4c4GCjYhUOQo31VVUS3PtHzg1c/H5dS4GuKhhTR7o2QiAsV9t4EDKSVdWKBUpLxu+exi+vh/ysqBJX7hnMURrgICIVD0KN9VZz4KZi7fD76WbuXh0nwuIiwklPSuPh6dreHiVlLrfvE256mPABj2fgMHTwS/M6spERErlvOZKv+GGG4p9PyUlpSy1SEXzC4XLnoWv74Mlr0CrARBS97wO4eVh581Bbej31q+s2HOMdxfvYOSlTcqnXnG9XUtg9nDIPAq+oXDjh9DkMqurEhEpk/NquQkJCSn2Ub9+fYYOHVpetUp5iLsZ6nWB3AyY/2SpDhEbHsCE61oC8MbC7fy577grK3RfDoc5NL8Ui5mWmWHAb5PMhS8zj0JUa7h3iYKNiLgFl46Wqgo0WuoMEjfAe5eA4YDbvoZGvc77EIZh8OCXf/Ld+kPUq+HPD6O6E+jj5osoGgbkZEBWihlSTn+c/Oe2lH98TYWsNODU/35e/mbLiV+Y2aJ2+vN/vvYt2B5mDu23e5xf3VlpZt+aLd+Zr9sMgX6vgZeW0xCRyut8/n4r3Ijpx8fgjylQswncvww8vc/7EKknc7nqzV85kHKSG9vV5bWBFTw83DDAkff3Iz/XnIjOkQeO3FNf809tz/v7tSO38Gey0v4RQlLOHlyMSjDtgU/ImUOQX9ip16c9x4Dv/2WuEO/hDVe+DO2Ha7ZhEan0FG6KoXBzFidT4J0OkHEY+kyAi0eX6jArdh/j5veX4zDg7cFtuSaudunqST0AK94318DKyy4mjJwWZqwKGnbPv0OEb8g/Hv/YVhBACl77BJujk04eN0PUyZTTnh//x+tTj4L3ck6UvubgOjDwc6jbvkyXLiJSURRuiqFwU4y1X5qdi70CYOTKUs9G+9pPW3n75x0E+Xry46ju1A3zL/mHD6yB3/8LG+eYgcUV7J5nfnh4mbd0nNtOvfYN/kcwCT0tmJwhuHj5WdPykZ97WthJKSYUHS+8X0xHc2K+gPCKr1lEpJQUboqhcFMMw4CPr4CE3+HC62HA1FIdJjffwYApy1mbkEKn2Bp8ec9FeNiL+ePvyIetP8LyybBv2d/b618MHYZDYOSpIOJ5Whg57bXzvYL3//Fat1xERKo8hZtiKNycgws6FwPsPZrBVW/+SkZOPv+67AIe7H2G4eHZJ2Dt/8Hv78Lx3eY2uye0vBEuegBqtyn1ZYiIiHvR8gtSelGtoOOpmYt//HepZi4GqF8zgGdPDQ+fFL+dNacPD0/dDwvGw+stzHMc323e4rn4YRi9AW54X8FGRERKzc3H6kqp9HoCNn4FR7aZ/V9K2bn4hnZ1WLztMN+uO8jo6Wv5cYA/AavfN/vTFHT+rdEILrof2twC3gGuuwYREam2FG6kKOfMxffDkv+cmrn4/DsX22w2nr+2OYG7fuT6E18T8OnWv9+M7Q5dRphrGNnVgCgiIq6jcCNn1vpmWP2p2bn4pyfPv3Nxdjr8+X+E/PEuE3P3gB1yDQ8S6/Uj5qpHILqC58AREZFqQ+FGzsxuh36vmp2LN86BdsNK1rk4JQFWvAerP4PsVHObXxh/1OzPgzs6cDKhFj/4NCGmfKsXEZFqTPcD5OzOp3Px/tUwazi8GQfL3jaDTc3G0O91eHgT7Ye/Tt16DczVw2esJS/fUTHXICIi1Y7CjRSv1xMQUOvvzsWnc+TDprnwUV/48FKzE7KRDw0ugcEzYMRK6HgnePvj6WHnzZvbEujjyaq9x5m8aKc11yMiIm5P4UaKV9C5GMzOxakHzP40v78Lb7WFmbeZ/XLsXhA3GO79FYZ9C02vKNJROKaGP8/1vxCAt37ezuq9Wj1cRERcr1KEm8mTJxMbG4uvry+dO3dmxYoVZ933q6++okOHDoSGhhIQEECbNm34/PPPK7Daaqj1zRBzEeRmwP/dZM5PM+9xSNlrLsjY/RF4+C+4fgpEty72UNe3rct1bWqT7zAYPeNP0rNyK+giRESkurA83MyYMYMxY8bw9NNPs2bNGuLi4ujbty/Jycln3L9GjRo8+eSTLF++nPXr1zN8+HCGDx/O/PnzK7jyaqSgc7HNDsmbIDvNXD386jfg4U3QexwERZX4cM/1b0ndMD8Sjp1k/Dcby7FwERGpjixffqFz58507NiRd955BwCHw0FMTAwPPvggjz/+eImO0a5dO/r168dzzz13zn21/EIZrPwIdv8CbYZA4z5lmp9m1Z5jDHzPXD38zZvbcF2b0i3SKSIi1UOVWX4hJyeH1atX06dPH+c2u91Onz59WL58+Tk/bxgG8fHxbN26lUsuueSM+2RnZ5OWllboIaXU8U4Y+ClccHmZJ97rEFuDBy8115t6as5fJBzLdEWFIiIi1oabI0eOkJ+fT2RkZKHtkZGRJCYmnvVzqampBAYG4u3tTb9+/Xj77be57LLLzrjvxIkTCQkJcT5iYjTDSmXx4KWNaV8/jPTsPEZreLiIiLiI5X1uSiMoKIi1a9eycuVKXnjhBcaMGcPixYvPuO/YsWNJTU11PhISEiq2WDkrTw87kwa1IcjHk9V7j/POoh1WlyQiIm7A0hmKw8PD8fDwICkpqdD2pKQkoqLO3kHVbrfTuHFjANq0acPmzZuZOHEiPXv2LLKvj48PPj4+Lq1bXMccHt6S0TPW8lb8dro3Cad9/RpWlyUiIlWYpS033t7etG/fnvj4eOc2h8NBfHw8Xbp0KfFxHA4H2dnZ5VGiVID+betwfds6OAwYNX0taRoeLiIiZWD5bakxY8bwwQcf8Omnn7J582buv/9+MjIyGD58OABDhw5l7Nixzv0nTpzIggUL2LVrF5s3b+a1117j888/59Zbb7XqEsQFnr3uQmJq+LH/+EnGf/2X1eWIiEgVZvnCmYMGDeLw4cOMHz+exMRE2rRpw7x585ydjPft24f9tJE5GRkZPPDAA+zfvx8/Pz+aNWvGF198waBBg6y6BHGBIF8vJg1qy8D3lvP12oP0bBpB/7YaHi4iIufP8nluKprmuanc3ly4nTcWbiPQx5MfR3Unpoa/1SWJiEglUGXmuRH5pxG9GtGhfhgnsvMYNf1PDQ8XEZHzpnAjlYqnh503Tg0PX7Mvhbd/1vBwERE5Pwo3UunE1PDn+etbAvD2z9v5aePZJ3QUERH5J4UbqZSua1OHm9rXxWHAfV+s5tNle6wuSUREqgiFG6m0XrqhFYM7xeAw4Om5G3nh+004HNWq/7uIiJSCwo1UWp4edl68vhWP9m0KwAe/7mbkl2vIys23uDIREanMFG6kUrPZbIzo1Zg3b26Dt4edHzYkcssHv3P0hGakFhGRM1O4kSrhujZ1+PzOTgT7mqOobnx3GXuOZFhdloiIVEIKN1JldG5Yk68e6ErdMD/2HM3k+v8uZfXeY1aXJSIilYzCjVQpjSOCmPNAN1rXDeF4Zi6DP/iDHzYcsrosERGpRBRupMqpFeTD9Hsuok/zCHLyHIyYtoYPf91FNVtJREREzkLhRqokf29P3rutA8O61Mcw4PnvN/PM3I3ka6i4iEi1p3AjVZaH3cYz117IU/2aY7PBp8v3cu/nq8nMybO6NBERsZDCjVRpNpuNu7o3ZPIt7fD2tLNwcxKD3/+dw+kaKi4iUl0p3IhbuKpVNF/e3Zkwfy/W7U/l+v8uZUdyutVliYiIBRRuxG20r1+Drx7oRmxNf/YfP8kN/13GH7uOWl2WiIhUMIUbcSsNwgP46oFutKsXSlpWHrd9tIJv1h6wuiwREalACjfidmoEeDPt7ou4smUUOfkORk1fy+RFOzRUXESkmlC4Ebfk6+XB5FvacXf3BgC8Mn8rT8zZQF6+w+LKRESkvCnciNuy22082a8FE669ELsNvlyRwJ2fruJEtoaKi4i4M4UbcXvDusby3m0d8PWys2TbYQZOWU5SWpbVZYmISDlRuJFq4bIWkcy4pwvhgd5sOpTG9ZOXsjVRQ8VFRNyRwo1UG3Exocx5oBuNagVwMDWLm95dxtIdR6wuS0REXEzhRqqVmBr+/O/+rnRqUIP07DyGfbyC2av3W12WiIi4kMKNVDuh/t58fmcnrmtTmzyHwSOz1jFp4TYNFRcRcRMKN1It+Xh68MbANozo1QiASQu38+js9eTkaai4iEhVp3Aj1ZbdbuPRvs148fpWeNhtzF69n+FTV5CWlWt1aSIiUgYKN1Lt3dK5Hh8O60CAtwdLdxzl5ve0qriISFWmcCMC9GoawYx7uxAe6MOmQ2kMmLKMhGOZVpclIiKloHAjckrLOiHMvq8LdcP82HM0k5umLGN7kubCERGpahRuRE4TGx7A/+7vygWRgSSlZTPgveWsTUixuiwRETkPCjci/xAZ7MvMe7vQtl4oKZm53PLB7/y2XZP9iYhUFQo3ImcQ6u/NF3d2pnuTcDJz8rlj6kp+3HDI6rJERKQEFG5EziLAx5MPh3WgX6tocvIdjJi2hukr9lldloiInIPCjUgxfDw9eGtwWwZ3isFhwONfbWDKkp1WlyUiIsVQuBE5Bw+7jRevb8X9Pc3ZjF/6cQsTf9ys5RpERCophRuRErDZbDx2RTOeuKoZAO8t2cXYrzaQ71DAERGpbBRuRM7DPZc04uUbW2G3wfSVCYyctobsvHyryxIRkdMo3Iicp0Ed6/HfIe3w9rDz41+J3Dl1FRnZeVaXJSIipyjciJTCFS2j+WR4R/y9PfhtxxFu+fAPjmfkWF2WiIigcCNSat0ahzPt7osI8/diXUIKA99bTmJqltVliYhUewo3ImXQJiaUmfd2ISrYl+3JJ7jx3WXsPpJhdVkiItWawo1IGTWJDGL2/V1oEB7AgZSTDJiyjI0HU60uS0Sk2lK4EXGBumH+zLy3Cy2igzlyIoeb3/udFbuPWV2WiEi1pHAj4iK1gnyYfu9FdIqtQXp2Hrd99Ac/b0myuiwRkWpH4UbEhYJ9vfjszk70bhZBdp6Dez5bzdd/HrC6LBGRakXhRsTFfL08mHJbe65vW4c8h8HoGWv5dNkeq8sSEak2FG5EyoGXh53XBsRxe9dYAJ6eu5FJC7dpPSoRkQpQKcLN5MmTiY2NxdfXl86dO7NixYqz7vvBBx/QvXt3wsLCCAsLo0+fPsXuL2IVu93G09e04OE+FwAwaeF2Jny7CYfWoxIRKVeWh5sZM2YwZswYnn76adasWUNcXBx9+/YlOTn5jPsvXryYwYMHs2jRIpYvX05MTAyXX345Bw6oX4NUPjabjVF9mjDh2gsBmLpsD/+atY7cfIfFlYmIuC+bYXE7eefOnenYsSPvvPMOAA6Hg5iYGB588EEef/zxc34+Pz+fsLAw3nnnHYYOHXrO/dPS0ggJCSE1NZXg4OAy1y9SUl//eYB/zVpHvsOgd7MIJg9ph6+Xh9VliYhUCefz99vSlpucnBxWr15Nnz59nNvsdjt9+vRh+fLlJTpGZmYmubm51KhR44zvZ2dnk5aWVughYoX+bevwwdD2+Hjaid+SzNCPVpCWlWt1WSIibsfScHPkyBHy8/OJjIwstD0yMpLExMQSHeOxxx6jdu3ahQLS6SZOnEhISIjzERMTU+a6RUrr0maRfH5nZ4J8PFmx5xg3v/c7yWlaj0pExJUs73NTFi+99BLTp09nzpw5+Pr6nnGfsWPHkpqa6nwkJCRUcJUihXVqUIPp915EeKA3mw6l0eOVxYz7+i/2aE0qERGXsDTchIeH4+HhQVJS4Vlck5KSiIqKKvazr776Ki+99BI//fQTrVu3Put+Pj4+BAcHF3qIWO3C2iHMuq8rLesEczI3n89/30uv1xZzz2erWLH7mIaMi4iUgaXhxtvbm/bt2xMfH+/c5nA4iI+Pp0uXLmf93H/+8x+ee+455s2bR4cOHSqiVBGXaxAewLcjL2baXZ3p1bQWhgE/bUpi4HvL6T95Kd+uO0ieRlWJiJw3y0dLzZgxg2HDhvHee+/RqVMnJk2axMyZM9myZQuRkZEMHTqUOnXqMHHiRABefvllxo8fz7Rp0+jWrZvzOIGBgQQGBp7zfBotJZXV9qR0Pl66m/+tOUBOnhlq6oT6MbxbLIM6xhDk62VxhSIi1jmfv9+WhxuAd955h1deeYXExETatGnDW2+9RefOnQHo2bMnsbGxTJ06FYDY2Fj27t1b5BhPP/00zzzzzDnPpXAjld2RE9l8vnwvn/++l2MZOQAE+Xhyc6cYbu/WgDqhfhZXKCJS8apcuKlICjdSVWTl5jPnzwN8+Osudh42Oxt72G30axXNXd0b0LpuqLUFiohUIIWbYijcSFXjcBgs3pbMh7/uZtnOo87tnRrU4O7uDendLAK73WZhhSIi5U/hphgKN1KV/XUglY9+2212Nj61RlXD8ADuuLgBN7ari5+3ZjwWEfekcFMMhRtxB4mpWUxdtodpf+wlLSsPgDB/L269qD63dalPRNCZ530SEamqFG6KoXAj7iQjO49ZqxL4aOluEo6dBMDbw07/trW58+KGNI0KsrhCERHXULgphsKNuKN8h8FPGxP54NddrNmX4tx+yQW1uLt7Ay5uHI7Npn45IlJ1KdwUQ+FG3N3qvcf56LddzPsrkVPdcmgWFcRd3RtyTVw0Pp7qlyMiVY/CTTEUbqS62Hc0k0+W7WbGygQyc/IBiAjyYUjn+tzUoa7myxGRKkXhphgKN1LdpJ7M5csV+5i6dA+Jp1Ygt9ng4sbhDOwQw2UtIvH1UmuOiFRuCjfFULiR6io338EPGw4xY2VCoflyQvy86N+mNgM6xNCyToiFFYqInJ3CTTEUbkQg4Vgms1bvZ/aqBA6mZjm3N48OZmCHuvRvU4ewAG8LKxQRKUzhphgKNyJ/y3cYLNt5hJmr9jN/Y6JzwU5vDzuXtYhkQIe6dG9SCw/NgCwiFlO4KYbCjciZpWTmMHfdQWauSuCvA2nO7dEhvtzYri43ta9LbHiAhRWKSHWmcFMMhRuRc9t4MJVZq/bz9doDpGTmOrd3blCDgR1iuLJVFP7enhZWKCLVjcJNMRRuREouOy+fhZuSmbkqgV+2H6bgt0WgjyfXxEUzoEMMbWNCNUGgiJQ7hZtiKNyIlM7BlJN8tWY/M1ftZ9+xTOf2xhGBDOxQl+vb1qVWkI+FFYqIO1O4KYbCjUjZOBwGK/YcY+aqBH7YcIisXLMTsofdRq+mEQzsUJdezSLw8rBbXKmIuBOFm2Io3Ii4TnpWLt+tP8TMVQn8edqaVuGB3tzQri4D2telSaQW7xSRslO4KYbCjUj52J6UzqzV+/lqzX6OnMhxbm8cEUjdMD+iQ3yJDvEjKsTX+Tw6xJcAH3VMFpFzU7gphsKNSPnKzXewaEsyM1ftZ9HWZPIdxf+KCfL1pPYZQo/zdagfgQpAItWewk0xFG5EKs7h9Gw2HUrjUMpJDqVmkZiaxaG0LA6lnCQxNYv07LwSHSfIx9MMO6F+RAebwad2qC9RIQUtQr4E+XqV89WIiJXO5++3/jkkIuWmVpAPPYJqnfX99KxcktKyOJhyKvikZnEo9bQglHqStKw80rPzSE8+wfbkE2c9VuCpANQgPICbO8bQq2kEds2sLFItqeVGRCq1jOy8QmHn0KkQlHja89STuUU+1zA8gOEXN+CmdnXx89aq5yJVnW5LFUPhRsT9ZObkkXgqAC3ZdphpK/aRnmXe8gr192JI53oM7RJLZLCvxZWKSGkp3BRD4UbE/WVk5zFrVQIfL93jnHDQy8PGNa1rc8fFDWhZJ8TiCkXkfCncFEPhRqT6yHcYLNiUxMe/7WbFnmPO7V0a1uTOixtwaTP1yxGpKhRuiqFwI1I9rUtI4aPfdvP9hkPO4ekF/XJubFdHC4GKVHIKN8VQuBGp3g6mnOTT5XuY9of65YhUJQo3xVC4ERH4u1/OJ8v2sPeo+uWIVHYKN8VQuBGR0+U7DBZuTuKjX9UvR6QyU7gphsKNiJzN+v1mv5zv1qtfjkhlo3BTDIUbETkX9csRqXwUboqhcCMiJaV+OSKVh8JNMRRuROR8na1fzkUNa3DXxQ3VL0ekAijcFEPhRkTK4kz9chpHBHLPJQ3p36YO3p52iysUcU8KN8VQuBERVzhTv5yoYF/uvLgBgzvXI9BHnY9FXEnhphgKNyLiSulZuXy5Yh8f/babpLRsAIJ9PbmtS31u79qAWkE+Flco4h4UboqhcCMi5SE7L59v/jzIlF92sutwBgDennZual+Xe7o3JDY8wOIKRao2hZtiKNyISHlyOAwWbE5iypKd/LkvBQC7Da5sGc19PRrRqq5GWImUhsJNMRRuRKQiGIbBit3HmLJkJ4u2HnZu79a4Jvf1aMTFjcOx2TTCSqSkFG6KoXAjIhVtS2Ia7y3Zxdx1B50jrC6sHcx9PRpxZcsoPD00wkrkXBRuiqFwIyJW2X88k49+2830FQmczM0HoF4Nf+7u3oABHWLw9fKwuEKRykvhphgKNyJiteMZOXy2fC9Tl+3meGYuADUDvLm9ayxDu8QS4u9lcYUilY/CTTEUbkSksjiZk8/MVQl88Osu9h8/CYC/tweDO9XjzosbUDvUz+IKRSoPhZtiKNyISGWTl+/g+w2HeHfxTrYkpgPgabdxXZs63NejIU0igyyuUMR6CjfFULgRkcrKMAyWbDvMlCU7+X3X32tY9WkewX09GtEhtoaF1YlYS+GmGAo3IlIVrE1IYcrinczflEjBb+kO9cO4r0cjejSthZdGWEk1o3BTDIUbEalKdh4+wQe/7OKrNQfIyXcA4OVho1GtQJpGBXFBZBDNooJoGhVEnVA/zZ0jbqtKhZvJkyfzyiuvkJiYSFxcHG+//TadOnU6474bN25k/PjxrF69mr179/LGG28wevTo8zqfwo2IVEXJaVl8vHQPX67YR+rJ3DPuE+jjyQWRgTSNCqbpqa/NooIIC/Cu4GpFXO98/n5bumztjBkzGDNmDFOmTKFz585MmjSJvn37snXrViIiIorsn5mZScOGDRkwYAAPP/ywBRWLiFgjItiXx69sxmNXNOVAykm2JqazJTGdbUnpbE1MZ+fhE5zIzmPNvhTWnFr2oUCtIB+anWrlaRpltvQ0iQjCz1vz6oh7srTlpnPnznTs2JF33nkHAIfDQUxMDA8++CCPP/54sZ+NjY1l9OjRarkREQFy8x3sPpLBlsR0tiamsTXxBFuT0kg4dvKM+9tsUL+GP02jgmgaGWS29kQFElszQDMmS6VUJVpucnJyWL16NWPHjnVus9vt9OnTh+XLl1tVlohIleTlYeeCSLN1hrjazu0Z2XnO1p2tBV8T0zmakcOeo5nsOZrJ/I1Jzv29Pe00PtWfp+DRuFYgtUP98LCrP49UDZaFmyNHjpCfn09kZGSh7ZGRkWzZssVl58nOziY7O9v5Oi0tzWXHFhGp7AJ8PGlbL4y29cIKbT9yIvvvW1uJ6WxJSmd7UjqZOflsOpTGpkOFf1d6ediICfOnfk1/6tcMoEF4APVr+hNbM4A6YX4avSWViqV9birCxIkTmTBhgtVliIhUKuGBPoQ39qFb43DnNofDYP/xk2xJTGNbUvqpW1zp7D2aSU6+g11HMth1JAM4XOhYHnYbdcP8qF8zgNhT4afga0wNP3w81bdHKpZl4SY8PBwPDw+SkpIKbU9KSiIqKspl5xk7dixjxoxxvk5LSyMmJsZlxxcRcRd2u416Nf2pV9Ofyy/8+/dwvsMgMS2LvUcy2HM0k71HM9hzNIO9RzPZczSDrFwHe49msvdoJr/845g2G9QO8SvU0lO/pj+x4QHUq+GvxUKlXFgWbry9vWnfvj3x8fH0798fMDsUx8fHM3LkSJedx8fHBx8fH5cdT0SkuvGw26gT6kedUD+6Ni78nmEYJKdns/tIxqnQcyr8HDG/ZuTkcyDlJAdSTvLbjqLHjg7xPS30mC0+tYJ88PKw4+Vhx9vTjreHHS9PW+HXHnb1AZKzsvS21JgxYxg2bBgdOnSgU6dOTJo0iYyMDIYPHw7A0KFDqVOnDhMnTgTMTsibNm1yPj9w4ABr164lMDCQxo0bn/U8IiJSPmw2G5HBvkQG+3JRw5qF3jMMgyMncgqHnlNfdx/JID0rj0OpWRxKzSq03ERJ2W0UCTxenjbnc2/PU9s8bHh7euDtYSsUmrw87Hh72Ajx96ZNTAhtY8I0J5CbsDTcDBo0iMOHDzN+/HgSExNp06YN8+bNc3Yy3rdvH3b7353UDh48SNu2bZ2vX331VV599VV69OjB4sWLK7p8EREphs1mo1aQD7WCfIqsi2UYBimZuYVub+05ddsr9WQuOXkOcvMd5OQ7yM1zkJtvOGdoLuAwIDvPQXZe4e1l0SA8gLb1QmlXL4y29UJpGhmkofFVkOUzFFc0zXMjIlI1GYZBnsMoHHzyT3td6Kvh3Kdgu/naIDfv9NDk4GBqFmv2HWfX4Ywi5/T39iCubmihwFMzUF0drFAl5rkRERE5HzabDa9Tt5bKQ0pmDn8mpPDnvhT+3HectftSSM/OY/muoyzfddS5X/2a/s6g065eGE2jgjQUvpJRy42IiMgZ5DsMdh4+wZq9x1mz7zh/7kthe/KJIvv5eXnQqm5IocBTK0itO65WpRbOrGgKNyIiUlqpJ3NZm2C27Kw51cKTnpVXZL+YGn60jQmjXb1Q2tUPo3l0sFp3ykjhphgKNyIi4ioOh8GuIydYszeFPxOOs2ZvCtuS0/nnX1YfTzut64aYs0XHhNKidjAxYf7YNZy9xBRuiqFwIyIi5SktK5f1Cams2ff37azUk7lF9gvw9qBpVBDNo4NpFh1Mi2hzAdNAH3WHPROFm2Io3IiISEUyDINdRzJYs/c4fyaksC7B7LuTc5Yh7PVq+NM8OohmUcE0jw6meXSQWnlQuCmWwo2IiFgt79RaXZsPpbH5UDpbEtPYfCiNpLTsM+6vVh6Fm2Ip3IiISGV1LCOHLYfS2JyYzuZDaacWMS2+lafZqdDTPNr86q6tPAo3xVC4ERGRqiQv38HuIxlsOpTGllOhpyStPM2izdtaTSOD8Du1QKntVOZxfsWGzVb0tfm8YLvN+ZmCyGSz2U57v/DnAHy87EQE+bri8p0UboqhcCMiIu7gfFt5KlK7eqF89UA3lx5TMxSLiIi4uRoB3nRtHE7XxuHObWdq5dl5+AR5+WY7hmGAgeEcqm6c2mY+K3jf7AR9+vunv8Yo2Nvc/vcx/j62t6e1c/oo3IiIiLgJTw87TSKDaBIZxHVWF2MhTZcoIiIibkXhRkRERNyKwo2IiIi4FYUbERERcSsKNyIiIuJWFG5ERETErSjciIiIiFtRuBERERG3onAjIiIibkXhRkRERNyKwo2IiIi4FYUbERERcSsKNyIiIuJWFG5ERETErXhaXUBFMwwDgLS0NIsrERERkZIq+Ltd8He8ONUu3KSnpwMQExNjcSUiIiJyvtLT0wkJCSl2H5tRkgjkRhwOBwcPHiQoKAibzebSY6elpRETE0NCQgLBwcEuPXZl4O7XB+5/jbq+qs/dr1HXV/WV1zUahkF6ejq1a9fGbi++V021a7mx2+3UrVu3XM8RHBzstv/RgvtfH7j/Ner6qj53v0ZdX9VXHtd4rhabAupQLCIiIm5F4UZERETcisKNC/n4+PD000/j4+NjdSnlwt2vD9z/GnV9VZ+7X6Our+qrDNdY7ToUi4iIiHtTy42IiIi4FYUbERERcSsKNyIiIuJWFG5ERETErSjcuMjkyZOJjY3F19eXzp07s2LFCqtLcpmJEyfSsWNHgoKCiIiIoH///mzdutXqssrNSy+9hM1mY/To0VaX4jIHDhzg1ltvpWbNmvj5+dGqVStWrVpldVkuk5+fz7hx42jQoAF+fn40atSI5557rkRr0FRGv/zyC9dccw21a9fGZrPx9ddfF3rfMAzGjx9PdHQ0fn5+9OnTh+3bt1tTbCkVd425ubk89thjtGrVioCAAGrXrs3QoUM5ePCgdQWfp3P9DE933333YbPZmDRpUoXVV1Ylub7Nmzdz7bXXEhISQkBAAB07dmTfvn0VUp/CjQvMmDGDMWPG8PTTT7NmzRri4uLo27cvycnJVpfmEkuWLGHEiBH8/vvvLFiwgNzcXC6//HIyMjKsLs3lVq5cyXvvvUfr1q2tLsVljh8/Trdu3fDy8uLHH39k06ZNvPbaa4SFhVldmsu8/PLLvPvuu7zzzjts3ryZl19+mf/85z+8/fbbVpdWKhkZGcTFxTF58uQzvv+f//yHt956iylTpvDHH38QEBBA3759ycrKquBKS6+4a8zMzGTNmjWMGzeONWvW8NVXX7F161auvfZaCyotnXP9DAvMmTOH33//ndq1a1dQZa5xruvbuXMnF198Mc2aNWPx4sWsX7+ecePG4evrWzEFGlJmnTp1MkaMGOF8nZ+fb9SuXduYOHGihVWVn+TkZAMwlixZYnUpLpWenm40adLEWLBggdGjRw9j1KhRVpfkEo899phx8cUXW11GuerXr59xxx13FNp2ww03GEOGDLGoItcBjDlz5jhfOxwOIyoqynjllVec21JSUgwfHx/jyy+/tKDCsvvnNZ7JihUrDMDYu3dvxRTlQme7vv379xt16tQx/vrrL6N+/frGG2+8UeG1ucKZrm/QoEHGrbfeak1BhmGo5aaMcnJyWL16NX369HFus9vt9OnTh+XLl1tYWflJTU0FoEaNGhZX4lojRoygX79+hX6W7mDu3Ll06NCBAQMGEBERQdu2bfnggw+sLsulunbtSnx8PNu2bQNg3bp1/Pbbb1x55ZUWV+Z6u3fvJjExsdB/pyEhIXTu3Nltf+eA+XvHZrMRGhpqdSku4XA4uO2223j00Ue58MILrS7HpRwOB99//z0XXHABffv2JSIigs6dOxd7a87VFG7K6MiRI+Tn5xMZGVloe2RkJImJiRZVVX4cDgejR4+mW7dutGzZ0upyXGb69OmsWbOGiRMnWl2Ky+3atYt3332XJk2aMH/+fO6//34eeughPv30U6tLc5nHH3+cm2++mWbNmuHl5UXbtm0ZPXo0Q4YMsbo0lyv4vVJdfucAZGVl8dhjjzF48GC3WWzy5ZdfxtPTk4ceesjqUlwuOTmZEydO8NJLL3HFFVfw008/cf3113PDDTewZMmSCqmh2q0KLmUzYsQI/vrrL3777TerS3GZhIQERo0axYIFCyrufnAFcjgcdOjQgRdffBGAtm3b8tdffzFlyhSGDRtmcXWuMXPmTP7v//6PadOmceGFF7J27VpGjx5N7dq13eYaq6vc3FwGDhyIYRi8++67VpfjEqtXr+bNN99kzZo12Gw2q8txOYfDAcB1113Hww8/DECbNm1YtmwZU6ZMoUePHuVeg1puyig8PBwPDw+SkpIKbU9KSiIqKsqiqsrHyJEj+e6771i0aBF169a1uhyXWb16NcnJybRr1w5PT088PT1ZsmQJb731Fp6enuTn51tdYplER0fTokWLQtuaN29eYaMWKsKjjz7qbL1p1aoVt912Gw8//LBbtsQV/F6pDr9zCoLN3r17WbBggdu02vz6668kJydTr1495++cvXv38q9//YvY2Firyyuz8PBwPD09Lf29o3BTRt7e3rRv3574+HjnNofDQXx8PF26dLGwMtcxDIORI0cyZ84cfv75Zxo0aGB1SS7Vu3dvNmzYwNq1a52PDh06MGTIENauXYuHh4fVJZZJt27digzd37ZtG/Xr17eoItfLzMzEbi/868zDw8P5L0h30qBBA6Kiogr9zklLS+OPP/5wm9858Hew2b59OwsXLqRmzZpWl+Qyt912G+vXry/0O6d27do8+uijzJ8/3+ryyszb25uOHTta+ntHt6VcYMyYMQwbNowOHTrQqVMnJk2aREZGBsOHD7e6NJcYMWIE06ZN45tvviEoKMh5Xz8kJAQ/Pz+Lqyu7oKCgIv2HAgICqFmzplv0K3r44Yfp2rUrL774IgMHDmTFihW8//77vP/++1aX5jLXXHMNL7zwAvXq1ePCCy/kzz//5PXXX+eOO+6wurRSOXHiBDt27HC+3r17N2vXrqVGjRrUq1eP0aNH8/zzz9OkSRMaNGjAuHHjqF27Nv3797eu6PNU3DVGR0dz0003sWbNGr777jvy8/Odv3dq1KiBt7e3VWWX2Ll+hv8Ma15eXkRFRdG0adOKLrVUznV9jz76KIMGDeKSSy6hV69ezJs3j2+//ZbFixdXTIGWjdNyM2+//bZRr149w9vb2+jUqZPx+++/W12SywBnfHzyySdWl1Zu3GkouGEYxrfffmu0bNnS8PHxMZo1a2a8//77VpfkUmlpacaoUaOMevXqGb6+vkbDhg2NJ5980sjOzra6tFJZtGjRGf+fGzZsmGEY5nDwcePGGZGRkYaPj4/Ru3dvY+vWrdYWfZ6Ku8bdu3ef9ffOokWLrC69RM71M/ynqjYUvCTX99FHHxmNGzc2fH19jbi4OOPrr7+usPpshlFFp/AUEREROQP1uRERERG3onAjIiIibkXhRkRERNyKwo2IiIi4FYUbERERcSsKNyIiIuJWFG5ERETErSjciEi1Z7PZ+Prrr60uQ0RcROFGRCx1++23Y7PZijyuuOIKq0sTkSpKa0uJiOWuuOIKPvnkk0LbfHx8LKpGRKo6tdyIiOV8fHyIiooq9AgLCwPMW0bvvvsuV155JX5+fjRs2JDZs2cX+vyGDRu49NJL8fPzo2bNmtxzzz2cOHGi0D4ff/wxF154IT4+PkRHRzNy5MhC7x85coTrr78ef39/mjRpwty5c8v3okWk3CjciEilN27cOG688UbWrVvHkCFDuPnmm9m8eTMAGRkZ9O3bl7CwMFauXMmsWbNYuHBhofDy7rvvMmLECO655x42bNjA3Llzady4caFzTJgwgYEDB7J+/XquuuoqhgwZwrFjxyr0OkXERSpsiU4RkTMYNmyY4eHhYQQEBBR6vPDCC4ZhmKvS33fffYU+07lzZ+P+++83DMMw3n//fSMsLMw4ceKE8/3vv//esNvtRmJiomEYhlG7dm3jySefPGsNgPHUU085X584ccIAjB9//NFl1ykiFUd9bkTEcr169eLdd98ttK1GjRrO5126dCn0XpcuXVi7di0AmzdvJi4ujoCAAOf73bp1w+FwsHXrVmw2GwcPHqR3797F1tC6dWvn84CAAIKDg0lOTi7tJYmIhRRuRMRyAQEBRW4TuYqfn1+J9vPy8ir02maz4XA4yqMkESln6nMjIpXe77//XuR18+bNAWjevDnr1q0jIyPD+f7SpUux2+00bdqUoKAgYmNjiY+Pr9CaRcQ6arkREctlZ2eTmJhYaJunpyfh4eEAzJo1iw4dOnDxxRfzf//3f6xYsYKPPvoIgCFDhvD0008zbNgwnnnmGQ4fPsyDDz7IbbfdRmRkJADPPPMM9913HxEREVx55ZWkp6ezdOlSHnzwwYq9UBGpEAo3ImK5efPmER0dXWhb06ZN2bJlC2COZJo+fToPPPAA0dHRfPnll7Ro0QIAf39/5s+fz6hRo+jYsSP+/v7ceOONvP76685jDRs2jKysLN544w0eeeQRwsPDuemmmyruAkWkQtkMwzCsLkJE5GxsNhtz5syhf//+VpciIlWE+tyIiIiIW1G4EREREbeiPjciUqnpzrmInC+13IiIiIhbUbgRERERt6JwIyIiIm5F4UZERETcisKNiIiIuBWFGxEREXErCjciIiLiVhRuRERExK0o3IiIiIhb+X/LYsCtntdFHQAAAABJRU5ErkJggg==",
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\n",
            "Calculating metrics...\n",
            "Accuracy: 0.84252\n",
            "Precision: [0.87194857 0.81740678]\n",
            "Recall: [0.80296 0.88208]\n",
            "F1 score: [0.83603348 0.84851283]\n",
            "Support: [12500 12500]\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "         0.0       0.87      0.80      0.84     12500\n",
            "         1.0       0.82      0.88      0.85     12500\n",
            "\n",
            "    accuracy                           0.84     25000\n",
            "   macro avg       0.84      0.84      0.84     25000\n",
            "weighted avg       0.84      0.84      0.84     25000\n",
            "\n"
          ]
        },
        {
          "data": {
            "text/plain": [
              "{'val_loss': 0.48574831111486566}"
            ]
          },
          "execution_count": 22,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "# === Run the function ===\n",
        "RNN_LSTM(TEXT, train_data, valid_data, test_data, hidden_dim=16, num_layers=4, dropout=0.6, learning_rate=0.002, num_epochs=17, batch_size=118, clip=4, freeze_embeddings=False)\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "7sD_u9K2pCSh",
        "outputId": "790f8f03-d61d-43f2-b2d4-0844bf827751"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "CNN(\n",
            "  (embedding): Embedding(25002, 100, padding_idx=1)\n",
            "  (convs): ModuleList(\n",
            "    (0): Conv2d(1, 20, kernel_size=(3, 100), stride=(1, 1))\n",
            "    (1): Conv2d(1, 20, kernel_size=(4, 100), stride=(1, 1))\n",
            "    (2): Conv2d(1, 20, kernel_size=(5, 100), stride=(1, 1))\n",
            "  )\n",
            "  (fc): Linear(in_features=60, out_features=1, bias=True)\n",
            "  (dropout): Dropout(p=0.6, inplace=False)\n",
            ")\n",
            "The model has 2,524,321 trainable parameters\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:09<00:00, 17.32it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 94.86it/s] \n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 01 | Epoch Time: 0m 10s | Frozen: False\n",
            "\tTrain Loss: 0.634 | Train Acc: 64.20%\n",
            "\t Val. Loss: 0.483 |  Val. Acc: 79.85%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:08<00:00, 20.81it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 108.97it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 02 | Epoch Time: 0m 8s | Frozen: False\n",
            "\tTrain Loss: 0.473 | Train Acc: 78.06%\n",
            "\t Val. Loss: 0.406 |  Val. Acc: 83.67%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:08<00:00, 21.11it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 160.76it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 03 | Epoch Time: 0m 8s | Frozen: False\n",
            "\tTrain Loss: 0.420 | Train Acc: 81.73%\n",
            "\t Val. Loss: 0.379 |  Val. Acc: 84.82%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:08<00:00, 20.57it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 86.64it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 04 | Epoch Time: 0m 8s | Frozen: False\n",
            "\tTrain Loss: 0.389 | Train Acc: 83.70%\n",
            "\t Val. Loss: 0.361 |  Val. Acc: 85.66%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:09<00:00, 18.11it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 50.00it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 05 | Epoch Time: 0m 10s | Frozen: False\n",
            "\tTrain Loss: 0.353 | Train Acc: 85.54%\n",
            "\t Val. Loss: 0.344 |  Val. Acc: 85.71%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:08<00:00, 20.64it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 158.71it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 06 | Epoch Time: 0m 8s | Frozen: False\n",
            "\tTrain Loss: 0.331 | Train Acc: 86.50%\n",
            "\t Val. Loss: 0.341 |  Val. Acc: 86.04%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:08<00:00, 20.59it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 161.93it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 07 | Epoch Time: 0m 8s | Frozen: False\n",
            "\tTrain Loss: 0.313 | Train Acc: 87.48%\n",
            "\t Val. Loss: 0.343 |  Val. Acc: 85.87%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:08<00:00, 20.89it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 104.39it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 08 | Epoch Time: 0m 8s | Frozen: False\n",
            "\tTrain Loss: 0.296 | Train Acc: 88.17%\n",
            "\t Val. Loss: 0.345 |  Val. Acc: 85.45%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:08<00:00, 20.92it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 159.49it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 09 | Epoch Time: 0m 8s | Frozen: False\n",
            "\tTrain Loss: 0.276 | Train Acc: 89.66%\n",
            "\t Val. Loss: 0.346 |  Val. Acc: 85.38%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:08<00:00, 20.51it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 161.93it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 10 | Epoch Time: 0m 8s | Frozen: False\n",
            "\tTrain Loss: 0.257 | Train Acc: 89.89%\n",
            "\t Val. Loss: 0.349 |  Val. Acc: 85.74%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:08<00:00, 21.10it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 112.40it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 11 | Epoch Time: 0m 8s | Frozen: False\n",
            "\tTrain Loss: 0.242 | Train Acc: 90.67%\n",
            "\t Val. Loss: 0.356 |  Val. Acc: 85.23%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:08<00:00, 20.55it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 156.52it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 12 | Epoch Time: 0m 8s | Frozen: False\n",
            "\tTrain Loss: 0.222 | Train Acc: 91.62%\n",
            "\t Val. Loss: 0.366 |  Val. Acc: 85.01%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:08<00:00, 20.50it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 154.33it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 13 | Epoch Time: 0m 8s | Frozen: False\n",
            "\tTrain Loss: 0.215 | Train Acc: 92.21%\n",
            "\t Val. Loss: 0.373 |  Val. Acc: 84.21%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:08<00:00, 21.16it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 162.44it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 14 | Epoch Time: 0m 8s | Frozen: False\n",
            "\tTrain Loss: 0.203 | Train Acc: 92.60%\n",
            "\t Val. Loss: 0.377 |  Val. Acc: 84.73%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:08<00:00, 20.51it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 160.44it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 15 | Epoch Time: 0m 8s | Frozen: False\n",
            "\tTrain Loss: 0.186 | Train Acc: 93.24%\n",
            "\t Val. Loss: 0.399 |  Val. Acc: 84.26%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:08<00:00, 20.45it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 104.95it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 16 | Epoch Time: 0m 8s | Frozen: False\n",
            "\tTrain Loss: 0.186 | Train Acc: 93.14%\n",
            "\t Val. Loss: 0.393 |  Val. Acc: 84.01%\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 170/170 [00:08<00:00, 20.96it/s]\n",
            "100%|██████████| 43/43 [00:00<00:00, 159.53it/s]\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 17 | Epoch Time: 0m 8s | Frozen: False\n",
            "\tTrain Loss: 0.176 | Train Acc: 93.59%\n",
            "\t Val. Loss: 0.393 |  Val. Acc: 84.26%\n"
          ]
        },
        {
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjcAAAHHCAYAAABDUnkqAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAABnY0lEQVR4nO3dd1hT598G8DsJhL03yFJxKyoCdWul1WodXY5aRVs7XNXa+mvtcLa1dlpHtdqhtrWuV6tdWrRaF4oLJ05kqAwR2Ts57x+BQAQjYJITwv25rlyQk5NzvgdacvucZ0gEQRBAREREZCKkYhdAREREpEsMN0RERGRSGG6IiIjIpDDcEBERkUlhuCEiIiKTwnBDREREJoXhhoiIiEwKww0RERGZFIYbIiIiMikMN0Q6Mm7cOAQEBIhdhmgSEhIgkUiwZs0a9ba5c+dCIpHU6v0SiQRz587VaU19+vRBnz59dHpMYxcQEIBx48bV6736+B0QiYHhhkyeRCKp1WPfvn1il2owQ4YMgbW1NXJzc++7z+jRoyGXy3Hnzh0DVlZ3Fy5cwNy5c5GQkCB2KWr79u1T/3f1888/17hP9+7dIZFI0K5dOwNX93AqQuznn38udilE92UmdgFE+vbTTz9pPF+3bh2ioqKqbW/duvVDnWf16tVQKpUPdQxDGT16NH7//Xds27YNY8eOrfZ6QUEBtm/fjgEDBsDFxaXe53n//ffxzjvvPEypD3ThwgXMmzcPffr0qdZy9s8//+j13A9iaWmJ9evX44UXXtDYnpCQgMOHD8PS0lKkyohMG8MNmbx7P1iOHDmCqKioatvvVVBQAGtr61qfx9zcvF71iWHIkCGws7PD+vXraww327dvR35+PkaPHv1Q5zEzM4OZmXh/ZuRyuWjnBoCBAwdix44dyMjIgKurq3r7+vXr4eHhgaCgINy9e1fEColME29LEUHVN6Ndu3Y4ceIEevXqBWtra7z77rsAVB/0gwYNgre3NywsLNCsWTMsWLAACoVC4xj39rmp2ny/atUqNGvWDBYWFggNDcWxY8e01nP8+HFIJBKsXbu22mu7du2CRCLBH3/8AQDIzc3F9OnTERAQAAsLC7i7u+Oxxx7DyZMn73t8KysrPP3009izZw/S09Orvb5+/XrY2dlhyJAhyMzMxFtvvYX27dvD1tYW9vb2eOKJJ3D69Gmt1wDU3OemuLgYb7zxBtzc3NTnuHHjRrX3JiYmYtKkSWjZsiWsrKzg4uKC5557TuP205o1a/Dcc88BAPr27VvtFmNNfW7S09Px0ksvwcPDA5aWlggODq72c36Y311VQ4cOhYWFBTZv3qyxff369Rg+fDhkMlm195SVlWHBggXqcwYEBODdd99FcXGxxn6CIODDDz9EkyZNYG1tjb59++L8+fM11pGVlYXp06fD19cXFhYWaN68ORYtWqTXlsba/JwBYMOGDQgJCYGdnR3s7e3Rvn17fP311+rXS0tLMW/ePAQFBcHS0hIuLi7o0aMHoqKi9FY7NXxsuSEqd+fOHTzxxBMYOXIkXnjhBXh4eABQfYDa2tpixowZsLW1xb///ovZs2cjJycHn3322QOPu379euTm5uLVV1+FRCLBp59+iqeffhrx8fH3be3p0qULmjZtik2bNiEyMlLjtY0bN8LJyQn9+/cHALz22mvYsmULpkyZgjZt2uDOnTs4ePAg4uLi0Llz5/vWNXr0aKxduxabNm3ClClT1NszMzOxa9cujBo1ClZWVjh//jx+++03PPfccwgMDERaWhq+/fZb9O7dGxcuXIC3t/cDfwZVTZgwAT///DOef/55dOvWDf/++y8GDRpUbb9jx47h8OHDGDlyJJo0aYKEhASsWLECffr0wYULF2BtbY1evXrh9ddfx5IlS/Duu++qby3e7xZjYWEh+vTpg6tXr2LKlCkIDAzE5s2bMW7cOGRlZWHatGka+9fnd1eVtbU1hg4dil9//RUTJ04EAJw+fRrnz5/Hd999hzNnztT481m7di2effZZvPnmmzh69CgWLlyIuLg4bNu2Tb3f7Nmz8eGHH2LgwIEYOHAgTp48iccffxwlJSUaxysoKEDv3r1x8+ZNvPrqq/Dz88Phw4cxa9YspKSkYPHixQ+8jrqq7c85KioKo0aNQr9+/bBo0SIAQFxcHA4dOqTeZ+7cuVi4cCEmTJiAsLAw5OTk4Pjx4zh58iQee+wxnddOJkIgamQmT54s3Puffu/evQUAwsqVK6vtX1BQUG3bq6++KlhbWwtFRUXqbZGRkYK/v7/6+fXr1wUAgouLi5CZmanevn37dgGA8Pvvv2utc9asWYK5ubnGe4uLiwVHR0fhxRdfVG9zcHAQJk+erPVYNSkrKxO8vLyErl27amxfuXKlAEDYtWuXIAiCUFRUJCgUCo19rl+/LlhYWAjz58+vdr0//vijetucOXM0ftaxsbECAGHSpEkax3v++ecFAMKcOXPU22r6uUdHRwsAhHXr1qm3bd68WQAg7N27t9r+vXv3Fnr37q1+vnjxYgGA8PPPP6u3lZSUCF27dhVsbW2FnJwcjWup7+9u7969AgBh8+bNwh9//CFIJBIhKSlJEARBmDlzptC0aVN1fW3btq3285kwYYLG8d566y0BgPDvv/8KgiAI6enpglwuFwYNGiQolUr1fu+++64AQIiMjFRvW7BggWBjYyNcvnxZ45jvvPOOIJPJ1HUJglDtd1CTip/NZ599dt99avtznjZtmmBvby+UlZXd91jBwcHCoEGDtNZEdC/eliIqZ2FhgfHjx1fbbmVlpf4+NzcXGRkZ6NmzJwoKCnDx4sUHHnfEiBFwcnJSP+/ZsycAID4+/oHvKy0txdatW9Xb/vnnH2RlZWHEiBHqbY6Ojjh69Chu3br1wFqqkslkGDlyJKKjozVu9VT0B+nXrx8A1c9FKlX9qVAoFLhz5w5sbW3RsmVLrbe+avLXX38BAF5//XWN7dOnT6+2b9Wfe2lpKe7cuYPmzZvD0dGxzueten5PT0+MGjVKvc3c3Byvv/468vLy8N9//2nsX9/fXVWPP/44nJ2dsWHDBgiCgA0bNmic/976AGDGjBka2998800AwJ9//gkA2L17N0pKSjB16lSN2341/Rw3b96Mnj17wsnJCRkZGepHREQEFAoF9u/fX+trqa3a/pwdHR2Rn5+v9RaTo6Mjzp8/jytXrui8TjJdDDdE5Xx8fGrsgHr+/Hk89dRTcHBwgL29Pdzc3NSdkbOzsx94XD8/P43nFR+WD+pIGhwcjFatWmHjxo3qbRs3boSrqyseffRR9bZPP/0U586dg6+vL8LCwjB37txaf/hWdBhev349AODGjRs4cOAARo4cqe4PolQq8dVXXyEoKAgWFhZwdXWFm5sbzpw5U6vrryoxMRFSqRTNmjXT2N6yZctq+xYWFmL27NnqfiIV583KyqrzeauePygoSB3WKlTcxkpMTNTYXt/fXVXm5uZ47rnnsH79euzfvx/Jycl4/vnn71ufVCpF8+bNNbZ7enrC0dFRXV/F16CgII393NzcNMIYAFy5cgU7d+6Em5ubxiMiIgIAauxz9bBq+3OeNGkSWrRogSeeeAJNmjTBiy++iJ07d2q8Z/78+cjKykKLFi3Qvn17zJw5s8bbeURVMdwQlavaUlAhKysLvXv3xunTpzF//nz8/vvviIqKUvcPqE2HzJo6jQKqDqEPMmLECOzduxcZGRkoLi7Gjh078Mwzz2iMQBo+fDji4+OxdOlSeHt747PPPkPbtm3x999/P/D4ISEhaNWqFX799VcAwK+//gpBEDRGSX388ceYMWMGevXqhZ9//hm7du1CVFQU2rZtq9cOqVOnTsVHH32E4cOHY9OmTfjnn38QFRUFFxcXgw25f5jfXVXPP/88YmNjMXfuXAQHB6NNmzZa96/txIe1oVQq8dhjjyEqKqrGxzPPPKOzc9WVu7s7YmNjsWPHDgwZMgR79+7FE088odHPrFevXrh27Rp++OEHtGvXDt999x06d+6M7777TrS6yfixQzGRFvv27cOdO3ewdetW9OrVS739+vXrBjn/iBEjMG/ePPzf//0fPDw8kJOTg5EjR1bbz8vLC5MmTcKkSZOQnp6Ozp0746OPPsITTzzxwHOMHj0aH3zwAc6cOYP169cjKCgIoaGh6te3bNmCvn374vvvv9d4X1ZWlsbw5trw9/eHUqnEtWvXNFprLl26VG3fLVu2IDIyEl988YV6W1FREbKysjT2q0sQ8Pf3x5kzZ6BUKjVaFSpuL/r7+9f6WHXRo0cP+Pn5Yd++fepgfL/6lEolrly5otEpOi0tDVlZWer6Kr5euXIFTZs2Ve93+/btaq1KzZo1Q15enrqlxhDq8nOWy+UYPHgwBg8eDKVSiUmTJuHbb7/FBx98oG7BcnZ2xvjx4zF+/Hjk5eWhV69emDt3LiZMmGCwa6KGhS03RFpU/Mu96r/US0pK8M033xjk/K1bt0b79u2xceNGbNy4EV5eXhohS6FQVLtF4+7uDm9v72pDh++nopVm9uzZiI2NrTa3jUwmq9ZSsXnzZty8ebPO11MRtpYsWaKxvaYROzWdd+nSpdWG4NvY2ABAtdBTk4EDByI1NVXjVl9ZWRmWLl0KW1tb9O7duzaXUWcSiQRLlizBnDlzMGbMGK31AdV/Hl9++SUAqEeVRUREwNzcHEuXLtX4GdX0cxw+fDiio6Oxa9euaq9lZWWhrKysrpfzQLX9Od87+7VUKkWHDh0AQP3f77372Nraonnz5rX+75saJ7bcEGnRrVs3ODk5ITIyEq+//jokEgl++umnOt+WeBgjRozA7NmzYWlpiZdeeknjX8K5ublo0qQJnn32WQQHB8PW1ha7d+/GsWPHNFo8tAkMDES3bt2wfft2AKgWbp588knMnz8f48ePR7du3XD27Fn88ssvGi0GtdWxY0eMGjUK33zzDbKzs9GtWzfs2bMHV69erbbvk08+iZ9++gkODg5o06YNoqOjsXv37mozJnfs2BEymQyLFi1CdnY2LCws8Oijj8Ld3b3aMV955RV8++23GDduHE6cOIGAgABs2bIFhw4dwuLFi2FnZ1fna6qtoUOHYujQoVr3CQ4ORmRkJFatWqW+JRoTE4O1a9di2LBh6Nu3LwBV35q33noLCxcuxJNPPomBAwfi1KlT+Pvvv6u1ps2cORM7duzAk08+iXHjxiEkJAT5+fk4e/YstmzZgoSEhDq3wAHAnj17UFRUVG37sGHDav1znjBhAjIzM/Hoo4+iSZMmSExMxNKlS9GxY0d1y1WbNm3Qp08fhISEwNnZGcePH1dPfUB0X6KN0yISyf2GglcdklvVoUOHhEceeUSwsrISvL29hf/973/Crl27qg0/vt9Q8JqGzKIWQ24rXLlyRQAgABAOHjyo8VpxcbEwc+ZMITg4WLCzsxNsbGyE4OBg4ZtvvqnVsSssX75cACCEhYVVe62oqEh48803BS8vL8HKykro3r27EB0dXW2YdW2GgguCIBQWFgqvv/664OLiItjY2AiDBw8WkpOTq/1M7t69K4wfP15wdXUVbG1thf79+wsXL14U/P39NYY6C4IgrF69WmjatKkgk8k0fi/31igIgpCWlqY+rlwuF9q3b69Rc9Vrqe/vrupQcG1q+u+utLRUmDdvnhAYGCiYm5sLvr6+wqxZszSmHRAEQVAoFMK8efPUv5c+ffoI586dq/Hnk5ubK8yaNUto3ry5IJfLBVdXV6Fbt27C559/LpSUlNTp2ip+Nvd7/PTTT4Ig1O7nvGXLFuHxxx8X3N3dBblcLvj5+QmvvvqqkJKSot7nww8/FMLCwgRHR0fByspKaNWqlfDRRx9p1E10L4kgGPCfoERERER6xj43REREZFIYboiIiMikMNwQERGRSWG4ISIiIpPCcENEREQmheGGiIiITEqjm8RPqVTi1q1bsLOz0+n6LURERKQ/giAgNzcX3t7e1RZlvVejCze3bt2Cr6+v2GUQERFRPSQnJ6NJkyZa92l04aZi2u/k5GTY29uLXA0RERHVRk5ODnx9fWu1TEqjCzcVt6Ls7e0ZboiIiBqY2nQpYYdiIiIiMikMN0RERGRSGG6IiIjIpDS6PjdERGQ6FAoFSktLxS6DdEQulz9wmHdtMNwQEVGDIwgCUlNTkZWVJXYppENSqRSBgYGQy+UPdRyGGyIianAqgo27uzusra05KasJqJhkNyUlBX5+fg/1O2W4ISKiBkWhUKiDjYuLi9jlkA65ubnh1q1bKCsrg7m5eb2Pww7FRETUoFT0sbG2tha5EtK1ittRCoXioY7DcENERA0Sb0WZHl39ThluiIiIyKQw3BARETVgAQEBWLx4sdhlGBWGGyIiIgOQSCRaH3Pnzq3XcY8dO4ZXXnlFt8U2cBwtpUOZ+SW4nVuMlp4PXrGUiIgal5SUFPX3GzduxOzZs3Hp0iX1NltbW/X3giBAoVDAzOzBH9Nubm66LdQEsOVGR6IupKHzgijM3HJa7FKIiMgIeXp6qh8ODg6QSCTq5xcvXoSdnR3+/vtvhISEwMLCAgcPHsS1a9cwdOhQeHh4wNbWFqGhodi9e7fGce+9LSWRSPDdd9/hqaeegrW1NYKCgrBjxw4DX624GG50pK23PQDg/K0c5BeXiVwNEVHjIggCCkrKRHkIgqCz63jnnXfwySefIC4uDh06dEBeXh4GDhyIPXv24NSpUxgwYAAGDx6MpKQkrceZN28ehg8fjjNnzmDgwIEYPXo0MjMzdVanseNtKR3xdrSCj6MVbmYV4mTSXfQMYjMhEZGhFJYq0Gb2LlHOfWF+f1jLdfNxOn/+fDz22GPq587OzggODlY/X7BgAbZt24YdO3ZgypQp9z3OuHHjMGrUKADAxx9/jCVLliAmJgYDBgzQSZ3Gji03OhQW6AwAOHa98aRjIiLSnS5dumg8z8vLw1tvvYXWrVvD0dERtra2iIuLe2DLTYcOHdTf29jYwN7eHunp6Xqp2Rix5UaHQgOcse3UTcQkMNwQERmSlbkMF+b3F+3cumJjY6Px/K233kJUVBQ+//xzNG/eHFZWVnj22WdRUlKi9Tj3Ll0gkUigVCp1VqexY7jRobBAJwBAbHIWSsqUkJuxYYyIyBAkEonObg0Zk0OHDmHcuHF46qmnAKhachISEsQtqgHgp68ONXOzhbONHEWlSpy7lS12OURE1MAFBQVh69atiI2NxenTp/H88883qhaY+mK40SGJRIIu/qrWG/a7ISKih/Xll1/CyckJ3bp1w+DBg9G/f3907txZ7LKMnkTQ5Ri2BiAnJwcODg7Izs6Gvb29zo//3YF4fPhnHCJau+O7yFCdH5+IqLErKirC9evXERgYCEtLS7HLIR3S9ruty+c3W250LDSgfMRUwl0olY0qNxIRERkFhhsda+ttD2u5DNmFpbiSnid2OURERI0Ow42Omcmk6Oyn6nfDIeFERESGx3CjB+pbU+xUTEREZHAMN3oQWj7fzbGETJ2uOUJEREQPxnCjB518nWAukyAluwg37haKXQ4REVGjwnCjB1ZyGdr5OABQtd4QERGR4TDc6EmYekg4ww0REZEhMdzoSUWn4hh2KiYiIjIohhs96RKg6lR87XY+7uQVi1wNERGZgj59+mD69Onq5wEBAVi8eLHW90gkEvz2228PfW5dHccQGG70xNFajpYedgBUsxUTEVHjNnjwYAwYMKDG1w4cOACJRIIzZ87U6ZjHjh3DK6+8oovy1ObOnYuOHTtW256SkoInnnhCp+fSF4YbPapovWG/GyIieumllxAVFYUbN25Ue+3HH39Ely5d0KFDhzod083NDdbW1roqUStPT09YWFgY5FwPi+FGj8IC2amYiIhUnnzySbi5uWHNmjUa2/Py8rB582YMGzYMo0aNgo+PD6ytrdG+fXv8+uuvWo95722pK1euoFevXrC0tESbNm0QFRVV7T1vv/02WrRoAWtrazRt2hQffPABSktLAQBr1qzBvHnzcPr0aUgkEkgkEnW9996WOnv2LB599FFYWVnBxcUFr7zyCvLyKpcdGjduHIYNG4bPP/8cXl5ecHFxweTJk9Xn0ifRw83y5csREBAAS0tLhIeHIyYmRuv+WVlZmDx5Mry8vGBhYYEWLVrgr7/+MlC1dVPRqfj8rRzkF5eJXA0RkQkTBKAkX5xHLSdrNTMzw9ixY7FmzRqNCV43b94MhUKBF154ASEhIfjzzz9x7tw5vPLKKxgzZswDPxcrKJVKPP3005DL5Th69ChWrlyJt99+u9p+dnZ2WLNmDS5cuICvv/4aq1evxldffQUAGDFiBN588020bdsWKSkpSElJwYgRI6odIz8/H/3794eTkxOOHTuGzZs3Y/fu3ZgyZYrGfnv37sW1a9ewd+9erF27FmvWrKkW7vTBTO9n0GLjxo2YMWMGVq5cifDwcCxevBj9+/fHpUuX4O7uXm3/kpISPPbYY3B3d8eWLVvg4+ODxMREODo6Gr74WvB2tIKPoxVuZhXiZNJd9AxyE7skIiLTVFoAfOwtzrnfvQXIbWq164svvojPPvsM//33H/r06QNAdUvqmWeegb+/P9566y31vlOnTsWuXbuwadMmhIWFPfDYu3fvxsWLF7Fr1y54e6t+Fh9//HG1fjLvv/+++vuAgAC89dZb2LBhA/73v//BysoKtra2MDMzg6en533PtX79ehQVFWHdunWwsVFd+7JlyzB48GAsWrQIHh4eAAAnJycsW7YMMpkMrVq1wqBBg7Bnzx68/PLLtfp51ZeoLTdffvklXn75ZYwfPx5t2rTBypUrYW1tjR9++KHG/X/44QdkZmbit99+Q/fu3REQEIDevXsjODjYwJXXnvrWFIeEExE1eq1atUK3bt3Un3NXr17FgQMH8NJLL0GhUGDBggVo3749nJ2dYWtri127diEpKalWx46Li4Ovr6862ABA165dq+23ceNGdO/eHZ6enrC1tcX7779f63NUPVdwcLA62ABA9+7doVQqcenSJfW2tm3bQiaTqZ97eXkhPT29TueqD9FabkpKSnDixAnMmjVLvU0qlSIiIgLR0dE1vmfHjh3o2rUrJk+ejO3bt8PNzQ3PP/883n77bY0fXlXFxcUoLq4cip2Tk6PbC3mA0ABnbDt1kyuEExHpk7m1qgVFrHPXwUsvvYSpU6di+fLl+PHHH9GsWTP07t0bixYtwtdff43Fixejffv2sLGxwfTp01FSUqKzUqOjozF69GjMmzcP/fv3h4ODAzZs2IAvvvhCZ+eoytzcXOO5RCKBUqnUy7mqEi3cZGRkQKFQqJuuKnh4eODixYs1vic+Ph7//vsvRo8ejb/++gtXr17FpEmTUFpaijlz5tT4noULF2LevHk6r7+2wsoX0TyVlIWSMiXkZqJ3cyIiMj0SSa1vDYlt+PDhmDZtGtavX49169Zh4sSJkEgkOHToEIYOHYoXXngBgKoPzeXLl9GmTZtaHbd169ZITk5GSkoKvLy8AABHjhzR2Ofw4cPw9/fHe++9p96WmJiosY9cLodCoXjgudasWYP8/Hx1682hQ4cglUrRsmXLWtWrTw3qk1apVMLd3R2rVq1CSEgIRowYgffeew8rV66873tmzZqF7Oxs9SM5OdmAFQPN3GzhbCNHcZkSZ29mG/TcRERkfGxtbTFixAjMmjULKSkpGDduHAAgKCgIUVFROHz4MOLi4vDqq68iLS2t1seNiIhAixYtEBkZidOnT+PAgQMaIabiHElJSdiwYQOuXbuGJUuWYNu2bRr7BAQE4Pr164iNjUVGRobG3Y8Ko0ePhqWlJSIjI3Hu3Dns3bsXU6dOxZgxY6o1WohBtHDj6uoKmUxW7ReXlpZ2305MXl5eaNGihcYtqNatWyM1NfW+zXYWFhawt7fXeBiSRCJBF39V681x3poiIiKobk3dvXsX/fv3V/eRef/999G5c2f0798fffr0gaenJ4YNG1brY0qlUmzbtg2FhYUICwvDhAkT8NFHH2nsM2TIELzxxhuYMmUKOnbsiMOHD+ODDz7Q2OeZZ57BgAED0LdvX7i5udU4HN3a2hq7du1CZmYmQkND8eyzz6Jfv35YtmxZ3X8YeiARhFqOYdOD8PBwhIWFYenSpQBULTN+fn6YMmUK3nnnnWr7v/vuu1i/fj3i4+Mhlapy2ddff41Fixbh1q3a3WvNycmBg4MDsrOzDRZ0vjsQjw//jENEa3d8FxlqkHMSEZmqoqIiXL9+HYGBgbC0tBS7HNIhbb/bunx+i3pbasaMGVi9ejXWrl2LuLg4TJw4Efn5+Rg/fjwAYOzYsRodjidOnIjMzExMmzYNly9fxp9//omPP/4YkydPFusSaiVUvUL4XSiVomVJIiKiRkHUeW5GjBiB27dvY/bs2UhNTUXHjh2xc+dO9f26pKQkdQsNAPj6+mLXrl1444030KFDB/j4+GDatGk1TlJkTNp628NaLkN2YSmupOehpaed2CURERGZLFHDDQBMmTKl2oyGFfbt21dtW9euXav1/jZ2ZjIpOvs54eDVDMQkZDLcEBER6VGDGi3VkKlvTXEyPyIiIr1iuDGQ0MDKFcJF7MNNRGQy+LfU9Ojqd8pwYyCdfJ1gLpMgJbsIN+4Wil0OEVGDVTHrbUFBgciVkK5VTOtyv1UHakv0PjeNhZVchnY+DjiVlIVjCZnwda7bdN1ERKQik8ng6OioXqPI2toaEolE5KroYSmVSty+fRvW1tYwM3u4eMJwY0BhAc7qcPN05yZil0NE1GBVTPZqiEUYyXCkUin8/PweOqwy3BhQaIAzvt0fjxh2KiYieigSiQReXl5wd3dHaWmp2OWQjsjlco0pYOqL4caAugSoOhVfu52PO3nFcLG1ELkiIqKGTSaTPXT/DDI97FBsQI7WcrTwsAWgmq2YiIiIdI/hxsAql2LgrSkiIiJ9YLgxsLBAhhsiIiJ9YrgxsIqWm/O3cpBfXCZyNURERKaH4cbAvB2t4ONoBYVSwMkk9rshIiLSNYYbEahvTXFIOBERkc4x3Iig4tZUDPvdEBER6RzDjQjCyhfRPJWUhZIypcjVEBERmRaGGxE0c7OFs40cxWVKnL2ZLXY5REREJoXhRgQSiQRd/FWtNxwSTkREpFsMNyKp6FR8nOGGiIhIpxhuRFI5U/FdKJWCyNUQERGZDoYbkbT1toe1XIbswlJcSc8TuxwiIiKTwXAjEjOZFJ39VP1uOCSciIhIdxhuRKS+NcXJ/IiIiHSG4UZEoYGVI6YEgf1uiIiIdIHhRkSdfJ1gLpMgJbsIN+4Wil0OERGRSWC4EZGVXIZ2Pg4AON8NERGRrjDciCxMPSSc4YaIiEgXGG5Epl5Ek52KiYiIdILhRmQh5cswXLudjzt5xSJXQ0RE1PAx3IjMyUaOFh62AFSzFRMREdHDYbgxAqHsd0NERKQzDDdGoGIRTYYbIiKih8dwYwQqWm7O38pBfnGZyNUQERE1bAw3RsDb0Qo+jlZQKAWcTGK/GyIioofBcGMk1LemOCSciIjooTDcGAn1fDfsd0NERPRQGG6MRFj5IpqnkrJQUqYUuRoiIqKGi+HGSDRzs4WzjRzFZUqcvZktdjlEREQNFsONkZBIJOhSPlsxh4QTERHVH8ONEWGnYiIioofHcGNEKjoVH0+8C6VSELkaIiKihonhxoi09baHtVyG7MJSXEnPE7scIiKiBonhxoiYyaTo7Kfqd8Mh4URERPXDcGNk1Itost8NERFRvTDcGJnQwMoRU4LAfjdERER1xXBjZDr5OsFcJkFKdhFu3C0UuxwiIqIGh+HGyFjJZWjn4wCA890QERHVB8ONEQqr6HfDcENERFRnDDdGSL2IJjsVExER1RnDjREKKV+G4drtfNzJKxa5GiIiooaF4cYIOdnI0cLDFgBwLOGuyNUQERE1LAw3RiqU/W6IiIjqheHGSKkX0WS4ISIiqhOGGyNV0XJz/lYO8ovLRK6GiIio4WC4MVLejlbwcbSCQingZBL73RAREdUWw40RU9+a4pBwIiKiWmO4MWLq+W7Y74aIiKjWGG6MWFj5IpqnkrJQUqYUuRoiIqKGgeHGiDVzs4WzjRzFZUqcvZktdjlEREQNAsONEZNIJOhSPlsxh4QTERHVDsONkWOnYiIiorphuDFyFZ2KjyfehVIpiFwNERGR8WO4MXJtve1hLZchu7AUl9NzxS6HiIjI6BlFuFm+fDkCAgJgaWmJ8PBwxMTE3HffNWvWQCKRaDwsLS0NWK1hmcmk6OxX3u+Gt6aIiIgeSPRws3HjRsyYMQNz5szByZMnERwcjP79+yM9Pf2+77G3t0dKSor6kZiYaMCKDa9yEU3OVExERPQgooebL7/8Ei+//DLGjx+PNm3aYOXKlbC2tsYPP/xw3/dIJBJ4enqqHx4eHgas2PBCAytHTAkC+90QERFpI2q4KSkpwYkTJxAREaHeJpVKERERgejo6Pu+Ly8vD/7+/vD19cXQoUNx/vz5++5bXFyMnJwcjUdD08nXCeYyCVKyi3DjbqHY5RARERk1UcNNRkYGFApFtZYXDw8PpKam1vieli1b4ocffsD27dvx888/Q6lUolu3brhx40aN+y9cuBAODg7qh6+vr86vQ9+s5DK083EAwPluiIiIHkT021J11bVrV4wdOxYdO3ZE7969sXXrVri5ueHbb7+tcf9Zs2YhOztb/UhOTjZwxboRpu53w3BDRESkjajhxtXVFTKZDGlpaRrb09LS4OnpWatjmJubo1OnTrh69WqNr1tYWMDe3l7j0RB1qVhEkyOmiIiItBI13MjlcoSEhGDPnj3qbUqlEnv27EHXrl1rdQyFQoGzZ8/Cy8tLX2UahYplGK7dzsedvGKRqyEiIjJeot+WmjFjBlavXo21a9ciLi4OEydORH5+PsaPHw8AGDt2LGbNmqXef/78+fjnn38QHx+PkydP4oUXXkBiYiImTJgg1iUYhJONHC08bAFwSDgREZE2ZmIXMGLECNy+fRuzZ89GamoqOnbsiJ07d6o7GSclJUEqrcxgd+/excsvv4zU1FQ4OTkhJCQEhw8fRps2bcS6BIMJDXDG5bQ8HEvIxIB2tbttR0RE1NhIhEY2cUpOTg4cHByQnZ3d4PrfbI+9iWkbYtGhiQN2TOkhdjlEREQGU5fPb9FvS1HtVcxUfP5WDvKLy0SuhoiIyDgx3DQg3o5W8HG0gkIp4GQS+90QERHVhOGmgQkLLJ/vhkPCiYiIasRw08BU3JqK4WR+RERENWK40aWcFCD9ol5PEVa+iOappCyUlCn1ei4iIqKGiOFGVy7sAL4OBn6fBuhxAFozN1s428hRXKbE2ZvZejsPERFRQ8Vwoyu+YYBEAiQfAeL36e00EolEPVsx15kiIiKqjuFGV+w8gRDVrMrY94leW2/YqZiIiOj+GG50qfs0QGahar25/p/eTlPRqfh44l0olY1qDkYiIqIHYrjRJXsvoIv+W2/aetvDWi5DdmEpLqfn6uUcREREDRXDja51n65qvUmK1lvrjZlMis5+5f1ueGuKiIhIA8ONrhmo9aZyvhvOVExERFQVw40+GKD1JjSwsuWG/W6IiIgqMdzogwFabzr5OsFaLkNqThG+Oxiv8+MTERE1VAw3+qLRerNf54e3ksvw/qA2AIBPd17C6eQsnZ+DiIioIWK40Rd7LyBknOp7PbXejArzxaD2XihTCpj66ynkFpXq/BxEREQNDcONPvWYXt56c1gvrTcSiQQfP90ePo5WSMoswHvbzkHQ4+SBREREDQHDjT7Ze+u99cbByhxLRnWCTCrBjtO3sPnEDZ2fg4iIqCFhuNE3PbfeAECIvxNmPNYCADBn+3lcTc/Ty3mIiIgaAoYbfbP3BkIiVd/rcd6bib2boXtzFxSWKjD111MoKlXo5TxERETGjuHGEHq8AcjkqtabhAN6OYVUKsFXwzvCxUaOuJQcfPL3Rb2ch4iIyNgx3BiCAfreAIC7vSU+Hx4MAFhzOAFRF9L0ch4iIiJjxnBjKBWtN4mH9NZ6AwB9W7pjQo9AAMDMLaeRkl2ot3MREREZI4YbQzFQ6w0A/G9AK7T3cUBWQSmmb4iFgsszEBFRI8JwY0gGar2Rm0mxdFQn2MhlOHo9E8v+vaq3cxERERkbhhtDsvcGOlcZOaVHAa42+Oip9gCAr/dcxtH4O3o9HxERkbFguDG0qq031/XXegMAwzr54JnOTaAUgOkbY3E3v0Sv5yMiIjIGDDeG5uBjsNYbAJg/tC2autogJbsI//u/M1yegYiITB7DjRjUrTcH9d56Y2NhhiWjOkEukyLqQhp+OpKo1/MRERGJjeFGDAZuvWnn44B3nmgFAPjwzzhcuJWj93MSERGJheFGLAZsvQGA8d0D0K+VO0rKlJj660kUlJTp/ZxERERiYLgRi4MP0Hms6nsDtN5IJBJ89lwwPOwtcO12PubtuKD3cxIREYmB4UZMBm69cbaR46sRHSGRABuPJ2PH6Vt6PycREZGhMdyIyaFJZevNf4sMcspuzVwxtW9zAMC7W88i6U6BQc5LRERkKAw3YqtovUk4YJDWGwB4vV8QQgOckFdchqkbTqFUoTTIeYmIiAyB4UZsIrTemMmkWDyyExyszHE6OQuf/3PJIOclIiIyBIYbYyBC642PoxUWPdMBAPDtf/HYf/m2Qc5LRESkbww3xsChCdBpjOp7A7XeAMCAdp544RE/AMCMTadxO7fYYOcmIiLSF4YbY9FzBiA1V7XeJBw02GnfH9QGrTztkJFXjBmbYqFUcnkGIiJq2BhujEXVvjcGmPemgqW5DEtHdYKluRQHrmRg9YF4g52biIhIHxhujIlIrTdBHnaYM7gtAOCzXZcQm5xlsHMTERHpGsONMRGp9QYARob6YlAHL5QpBbz+6ynkFJUa9PxERES6wnBjbHq8IUrrjUQiwcKn26OJkxWSMgvw3rZzEAT2vyEiooaH4cbYOPoCnctHThm49cbe0hxLRnWCTCrB76dvYfPxGwY9PxERkS4w3BijHlX73hwy6Kk7+znhzcdbAADm7DiPq+m5Bj0/ERHRw2K4MUZVW2/+M2zrDQC81qsZejR3RWGpAlPWn0JRqcLgNRAREdUXw42xqmi9ub7f4K03UqkEXw4PhouNHBdTc7HwrziDnp+IiOhhMNwYK5Fbb9ztLfHF8GAAwNroRPxzPtXgNRAREdUHw40xE7H1BgD6tHTHyz0DAQD/+78zSMkuNHgNREREdcVwY8wcfYFOL6i+F6H1BgBm9m+FDk0ckFVQimkbYqHg8gxERGTkGG6MXc8qrTeJhw1+ermZFEtGdoKthRlirmdi6b9XDF4DERFRXdQr3CQnJ+PGjco5UGJiYjB9+nSsWrVKZ4VROUe/ytYbA897UyHA1QYfPdUOALBkzxV8f/A6J/gjIiKjVa9w8/zzz2Pv3r0AgNTUVDz22GOIiYnBe++9h/nz5+u0QEKV1pv/RGm9AYChHX0Q2dUfSgFY8McFzNxyBsVlHCJORETGp17h5ty5cwgLCwMAbNq0Ce3atcPhw4fxyy+/YM2aNbqsjwCjaL0BgLlD2uL9Qa0hlQBbTtzAyFVHkJ5TJFo9RERENalXuCktLYWFhQUAYPfu3RgyZAgAoFWrVkhJSdFddVTJCFpvJBIJJvRsijXjw2BvaYZTSVkYvOwgTnMVcSIiMiL1Cjdt27bFypUrceDAAURFRWHAgAEAgFu3bsHFxUWnBVI5Rz+g02jV9yK23gBArxZu2D6lB5q72yItpxjPfRuNbae4DhURERmHeoWbRYsW4dtvv0WfPn0watQoBAerJnvbsWOH+nYV6UHPNwGpWXnrTbSopQS62mDbpG6IaO2OkjIl3th4Gh//Fceh4kREJDqJUM9hLwqFAjk5OXByclJvS0hIgLW1Ndzd3XVWoK7l5OTAwcEB2dnZsLe3F7ucuvt9GnBiDdC0DzB2u9jVQKkU8EXUJSzfew2AqlVn6chOcLA2F7kyIiIyJXX5/K5Xy01hYSGKi4vVwSYxMRGLFy/GpUuXjDrYmISK1pv4faK33gCqdahm9m+FpaM6wdJciv2Xb2PYN4dwNT1P7NKIiKiRqle4GTp0KNatWwcAyMrKQnh4OL744gsMGzYMK1as0GmBdA9HP6Bjed8bkWYtrsngYG9sea0bvB0scT0jH08tP4R/L6aJXRYRETVC9Qo3J0+eRM+ePQEAW7ZsgYeHBxITE7Fu3TosWbJEpwVSDYys9aZCOx8H7JjaA6EBTsgtLsNLa4/jm31XOeEfEREZVL3CTUFBAezs7AAA//zzD55++mlIpVI88sgjSExM1GmBVAMn/8rWm38XAGXF4tZThautBX6Z8AhGhflBEIBPd17CtA2xKCzhhH9ERGQY9Qo3zZs3x2+//Ybk5GTs2rULjz/+OAAgPT29Xp10ly9fjoCAAFhaWiI8PBwxMTG1et+GDRsgkUgwbNiwOp+zwev5pmrem8RDwKq+QNp5sStSk5tJsfDp9lgwrB3MpBLsOH0Lz317GLeyuKo4ERHpX73CzezZs/HWW28hICAAYWFh6Nq1KwBVK06nTp3qdKyNGzdixowZmDNnDk6ePIng4GD0798f6enpWt+XkJCAt956S317rNFx8gdGrgesXYH088CqPsDhZYBSKXZlamMe8cfPE8LhbCPHuZs5GLLsII4nZIpdFhERmbh6DwVPTU1FSkoKgoODIZWqMlJMTAzs7e3RqlWrWh8nPDwcoaGhWLZsGQBAqVTC19cXU6dOxTvvvFPjexQKBXr16oUXX3wRBw4cQFZWFn777bdana/BDwW/V146sH0KcGWX6nlgL2DYCsChibh1VZGcWYCX1x3HxdRcmMskWDC0HUaG+YldFhERNSB6HwoOAJ6enujUqRNu3bqlXiE8LCysTsGmpKQEJ06cQERERGVBUikiIiIQHX3/jrLz58+Hu7s7XnrppfqWbzps3YHnNwJPfgWYWwPX9wMrugFnt4hdmZqvszW2TuqGge09UaoQ8M7Ws5i9/RxKFcbTykRERKajXuFGqVRi/vz5cHBwgL+/P/z9/eHo6IgFCxZAWYfbIhkZGVAoFPDw8NDY7uHhgdTU1Brfc/DgQXz//fdYvXp1rc5RXFyMnJwcjYfJkUiALi8Crx4AfEKAomzg/14C/m8CUJgldnUAAGu5GZY/3xkzHmsBAFgXnYix38cgM79E5MqIiAwoKxk4vUH1D9HcNICjSfXCrD5veu+99/D999/jk08+Qffu3QGoQsfcuXNRVFSEjz76SKdFVsjNzcWYMWOwevVquLq61uo9CxcuxLx58/RSj9FxbQ68uAvY/xmw/3Pg7GbVUPGnVqhuV4lMIpHg9X5BaOVphzc2xiI6/g6GLDuI7yK7oJWnCdwiJCK6n+I84NBi4PBSoKyocrulA+DaQvPh1hJw9Adk9fqIJtSzz423tzdWrlypXg28wvbt2zFp0iTcvHmzVscpKSmBtbU1tmzZojHiKTIyEllZWdi+XXN5gdjYWHTq1AkymUy9raKlSCqV4tKlS2jWrJnGe4qLi1FcXDlUOicnB76+vqbT5+Z+ko8BW18G7l4HIAG6TQEe/QAwsxC7MgDApdRcvLzuOJIyC2Atl+HL4cEY0M5L7LKIiHRLqQTObAB2zwPyyu9IeLQDSvKBrERAuM/dDpkccG4GuAapwo46/AQBchvD1W9E6tLnpl7hxtLSEmfOnEGLFi00tl+6dAkdO3ZEYWHth/yGh4cjLCwMS5cuBaAKK35+fpgyZUq1DsVFRUW4evWqxrb3338fubm5+Prrr9GiRQvI5XKt5zO5DsXaFOcBu2YBJ1WzScOjHfD0KsCjrbh1lbubX4Ipv57Eoat3AADT+gVhWr8gSKUSkSsjItKBpCPAzneAW6dUz50CgMc/BFo9qepOUFoEZMYDGZeAjCvA7UtAxmXV92VaPkcdfFUh597WHhs31XFNlN7DTXh4OMLDw6vNRjx16lTExMTg6NGjtT7Wxo0bERkZiW+//RZhYWFYvHgxNm3ahIsXL8LDwwNjx46Fj48PFi5cWOP7x40b17hHS9XGxb+AHVOBggzVvwb6zQEemQRI692fXGfKFEp89FccfjyUAADo39YDXw7vCBsLNscSUQOVlQREzQHOb1U9l9sBvd4CHplYu9ZzpRLIuQHcvlwedqqEn4KM+7/P0gFwLW/lcasSfGp7i0sQyvsAPexXAFKZqh4dqsvnd70+QT799FMMGjQIu3fvVs9xEx0djeTkZPz11191OtaIESNw+/ZtzJ49G6mpqejYsSN27typ7mSclJSkHmpO9dRqINCkS+WQ8X/eU301giHjZjIp5gxui9Ze9nh/2znsOp+GZ1YcxuqxXeDrbC1qbUREdVKtX40E6DwWePR91cjW2pJKVesIOvoBQRGarxVklgeey+UtPVdU4eduomowyY0Y1aMqiUy1ZI+2UKJrTcKACVG6P24t1Xuem1u3bmH58uW4ePEiAKB169Z45ZVX8OGHH2LVqlU6LVKXGmXLTQVBAE78COx6DygtUKXqQV8C7Z8VuzIAwInEu3j1pxPIyCuGk7U5lo/ujG7NatdxnIhINDX1qwnoCfT/GPDqYJgaSouAO1crg0/GZVXLz50rmh2YDUUP4Ubvt6Xu5/Tp0+jcuTMUCuNdR6hRh5sKGVdVnY1vnVQ9b/8cMPBzwMpR1LIAICW7EK/+dAJnbmRDJpVg9pNtMLarPyQmfB+ZiBqwxGhV38b79asRm1KpClzKMgCS8ppq+xVaXpc++Bg6vuvCcKMFw005RWn5kPHPVL317ZsAT60EAsVfzqKoVIF3/u8Mfou9BQB4uWcg3h3YmgGHiIxHVhIQNRs4v031XG4H9J4JhL9mNKNSTY1BZiimBk5mDvR9F3jxH8ApUNV5be1g4J/3RV9l3NJchq9GdMTbA1SzXa8+cB3vbjsHhZKTXRGRyIrzgD0LgKVdyoONBAgZB7x+Eug+jcHGSDDcNHa+ocBrB1Wd3iCoOsKtfhRIuyBqWRKJBBP7NMOiZ9pDIgF+jUnCm5tiUcYlG4hIDEolELseWBoCHPgcUBSr+tW8dgAY/HXdOgyT3tVptNTTTz+t9fWsrKyHqYXEYmELDFkKtBigGjKedk61ynjEHCB8oqhDxkeE+sFaboY3Nsbit9hbKChRYOnznWBhJnvwm4mIdCExWjVfTUqs6rlTYHm/mkHG0a+GqqlTn5vx48fXar8ff/yx3gXpG/vcPEC1VcZ7lw8Z9xG1rN0X0jBp/UmUlCnRM8gV344JgbWcc+EQkR6xX41REa1DcUPAcFMLggAc/0E1ZLysUDVk/MmvgHbPiFrWoasZmLD2OApLFQgNcML340Jhb2kuak1EZIKK84CDX6lu0yuKoepXEwn0fY+3n0TEcKMFw00dVBsyPhwY+JmoQ8ZPJGZi3I/HkFtUhvY+Dlj7YhicbbQvuUFEVCtKJXD6V2DPfM35agYsBDzbi1sbMdxow3BTR/cOGbfzVo0I6DxGtMXbzt3MxtgfYpCZX4IWHrb4+aVwuNtbilILEZkI9qsxegw3WjDc1FNyDLD1lfJVxgFYOQFhr6geNoafRfhqei5Gf3cUaTnF8Hexxi8TwtHEics1EFEtCAJQeFfVxzA3BTi5trJfjYU90GsmEP4q+9UYGYYbLRhuHkJpoWoo5OGllSHHzAro9ALQdTLgHGjQcpLuFOD5747gxt1CeDlY4pcJ4WjqZmvQGojIiJSVAPnpQG4akJemurWUlw7kln+teJ6XBihKNN8rkaqmxOj7PmDrJk79pBXDjRYMNzqgVABxO4CDiyubcCVSoO1TQLfXAe+OBislNbsIo787gmu38+FqK8dPL4WjtRd/r0QmQxBUC0LmlQeWe4NL1W2FmXU7tpUTYOuhWjm79//Yr8bIMdxowXCjQ4IAXN8PHPoauLancnvTPqp+OU37GuRedUZeMcZ+H4MLKTlwsDLH2hfD0NHXUe/nJSIdUSpViz6mnFb9g+luQmWYyUuv28KPUnNVYLF1B+w8VV9tPas896h8nbedGhSGGy0YbvQk5QxweAlwbisglK8t5tlBFXLaDANk+p2TJruwFON/jMHJpCzYyGX4flwoHmnqotdzElE9KMpUK1anxJaHmdNA6lmgJE/7+ywcADuPKuHEo4bnnoClo6gTj5L+MNxowXCjZ3cTgejlwMl1qjlyAMDRH+g2Feg4GpDrr9NvfnEZXl53HIev3YGFmRQrx4Sgb0vOSUEkmrIS4HZcZYi5FauaAb2mlhgzK9VtIa9gwK1lZVixdVd9b25l8PLJuDDcaMFwYyD5d4Bj3wEx3wIFd1TbrF2AsFeBsJcBa2e9nLaoVIHJv5zEnovpMJdJsGRkJzzR3ksv5yKiKkqLgPTzqgBTEWbSL1TvuAuoZvr16qAKMl7BgFdHwDUIkHJZFbo/hhstGG4MrKQAiP1FNcIqK1G1zdwa6DRGNcLKyV/npyxVKPHGxlj8cSYFUgnw6bPBeDakic7PQ9RoleQDqecqQ0xKLJAeV3lLuipLh8oAU/HVuSlvHVGdMdxowXAjEkUZELddNcIq9Yxqm0QGtHtaNcLKq4NuT6cUMGvrGWw6fgMAsGBoW4zpGqDTcxA1CsW5qj51VfvIZFxWTep5L2uXKiEmWDVy0tGfk+CRTjDcaMFwIzJBAOL3AYcWq75WaPYo0H06ENhLZ38IlUoB8/+4gDWHEwAAbw9ohYl9munk2EQmq6QASD6qGgl5fT9w61TNLTK2nqrwor61FAzY+zDIkN4w3GjBcGNEbsWqRlid31b5r0CvjuUjrIbq5P67IAj44p/LWLb3KgBgSt/mePPxFpDwDzCRSlkJcPNEZZi5EVO9n4x9kypBpqOqpdXOU4xqqRFjuNGC4cYIZV5XjbA69XPlCCungMoRVjoYJbFi3zUs2nkRADCuWwBmP9kGUikDDjVCSoXq1nBFmEmMBkrzNfex9wECe6taUgN7Ag7ss0biY7jRguHGiOVnADGrVI/Cu6pt1q6qNV5CJzz0CKufohPwwfbzAIDhXZpg4dMdIGPAIVMnCMDti5VhJuGAasbfqqxdyoNML1WocW7K20tkdBhutGC4aQBK8lWtOIeXAdlJqm3m1pVrWDkF1PvQW07cwP+2nIZSAAZ18MJXwztCbsZRG2RCBEE1w29FmLm+X7XeUlUW9kBAj8pA49aao5fI6DHcaMFw04AoSoEL21XLO6hHWElVMx53fx3w7lSvw/51NgXTNpxCqUJAv1buWD66MyzNOb8GNWA5KZphpuIfBRXMrAC/RypbZryC9T5rOJGuMdxowXDTAFWMsDq8BLj2b+X2gJ6qzsfNI+rchL73Ujpe++kEisuU6NbMBavHdoGNBf/YUwNRkKm6vVQRZjIua74uNQeahFa2zDTpwnWUqMFjuNGC4aaBSz2rmhDw7JbK4anubVSdj9s9C5jJa32oI/F38NKaY8gvUaCTnyPWjAuDg7W5ngonqqf8DNWSBWnnVY+U06rnVUmkqtaYijDj1xWQ24hTL5GeMNxowXBjIrKSgSMrgJNrKxfcs/MGHpkIhIwDLGv3u41NzkLkDzHILixFGy97rHspDK62/BcuiaCsWNUCk3ZeM8zkpdW8v3ubyjDj3w2wcjJsvUQGxnCjBcONiSnMAo7/ABxdWfkhYGGvCjiPTATsvR94iLiUHIz5/igy8krQzM0GP4wLhb8L/9VLeiIIQG5qlRBTHmQyLgPKshreIFGNXvJoC3i0AzzaAL7hqgUliRoRhhstGG5MVFkxcGaTql9ORf8DqTnQYbjqlpV7a61vj7+dhxe+O4pb2UWQm0kxoUcgJvVtDlv2w6GHUVqoWnOpohWmIsgUZta8v6VDeYBpWxlm3FoBFraGrZvICDHcaMFwY+KUSuDKLuDQEiDpcOX2oMdVa1gF9Lhv5+ObWYX435bTOHRVtYq5q60F/te/JZ4NacIJ/0g7QQCykqqHmMxrNa/BJJECLkGqAOPZrjLQcPkCovtiuNGC4aYRuXFcNYw87ncA5f+Ze3dShZzWQ2ocCisIAnbHpeOjPy8g4U4BAKCdjz1mP9kWYYEPN4kgNSClhaqJ7gqzVF+Lsmp4nqV6nn9b1TpTnFPzsaxdysNLlRYZt1aAuaWhrobIJDDcaMFw0wjduaZa3iH2F6CsSLXNKQDoOkW1vIPcutpbissUWHc4EUv2XEFusaofxKD2XnjniVbwda6+PxkZpVIVNioCyH0Dyj3PK75XFNf9nFJzwK2l5i0lj3aqvjFsjSF6aAw3WjDcNGLq5R1WV/Z5sHIGwl4Gwl4BbFyrvSUjrxhfRl3GhpgkKAVAbibFyz0DMalPc86LowuCoAqcJQWqUW+lBaoZqkvyNb8vyVetf1SSX77vvc8r3lsAlOQCRTlQt9bVl0Sq6gNj6QBYOqq+Wjne872D6r8ht5aq20x1mIqAiOqG4UYLhhtCSYGqFSd6mWqaegAws1S14nSdDLg0q/aWuJQczP/9AqLjVf1x3OxU/XGe6dxI+uMoyjTDROm9IaPgwa/fL6DU1CdFV8ysNIOItpBy72tyWy5JQGREGG60YLghNaUCiNuh6pdz61TldovyDzgrp8qHtTMES0dcyjHH1rh8XMszR5ZgC3d3T7wyIASdggKN41/tSmVlgCjOU7ViFOepWjZK8oHi3OqtIhXfV9tWpUVEUaL/2s2sVLcIzW1UE9DJrVVfze/9vuK5rWrNMY3vbVTfVwQWzspLZDIYbrRguKFqBAFIOKgaRn7ln3ofRim3hdTKuXowuickaTy3dASUpQ8II3lVtlX5vmJ/9XvL3/Owt2O0kchU4UFuXSVM2NQQMrSFDxvN7yueS7m+FxHdX10+v9lpgEgiAQJ7qh6FWaq+OYV3yx+ZVb7XfJTlZ6IkNwOWZXmQSgRIK8LHvYsWikEiLQ8htqo5UuS2qhBhYVclXFQNKfcLLPd8L5OzcywRGT2GG6KqrBxVj1owK3+cv5GJL/84jquJyXBEHgJsSjC6vR26eADSoqz7hiMU3q3sbyKRAnK78iBiUyWUlIeRioBiYVclqNjW8J7y182tGEKIqNHibSkiHRAEAbvOp+Hjv+KQlKmaHye4iQNmD26DEP/7zI+jVKpuK0nNGUaIiB6AfW60YLghfSouU+DHQwlY9u9V5JXPjzMk2BtvP9EKPo5WIldHRNRwMdxowXBDhpCeW4Qvdl3GphPJEATA0lyKV3o1w2u9m8JazrvBRER1xXCjBcMNGdK5m9mY/8cFxFxXTRroaW+Jt59oiaHBPo1jfhwiIh1huNGC4YYMTRAE7DyXio/+isONu4UAgI6+jpg9uA06+zmJXB0RUcPAcKMFww2JpahUgR8OXcfyf68iv0QBABja0Rsz+7dEEyeuV0VEpA3DjRYMNyS29JwifP7PJWw+cQOCAMikEgwN9sarvZuhpaed2OURERklhhstGG7IWJy7mY1P/r6Ig1cz1Nv6tXLHxD7N0CXgPsPHiYgaKYYbLRhuyNicuZGFlf9dw9/nUlHxf2MXfydM7NMMfVu6s+MxEREYbrRiuCFjFX87D6sPxOP/TtxEiUI1c3FLDzu81qcpnuzgDXMZV6gmosaL4UYLhhsydmk5Rfjh4HX8cjRJPRGgj6MVXu4ZiBGhfrCSc4FJImp8GG60YLihhiK7sBQ/H0nEj4euIyOvBADgbCNHZNcAjO3qDycbucgVEhEZDsONFgw31NAUlSqw5cQNrNofr163ylouw6gwP7zUIxDeXNaBiBoBhhstGG6ooSpTKPHXuVSs3HcNF1JyAABmUgmGdfLBa72bork7h5ETkeliuNGC4YYaOkEQsP9KBlbsu4oj8Znq7Y+18cDEPs046zERmSSGGy0YbsiUnEq6i5X/XcM/F9LUw8jDA53xWp9m6NPCDRIJh5ETkWlguNGC4YZM0dX0PKzafw3bTt1EqUL1v3RrL3u81rspBrX3ghmHkRNRA8dwowXDDZmylOxC9TDygvL1q3ydrfBKz6Z4rosvLM05jJyIGiaGGy0YbqgxyCoowU/RifjxcAIy81XDyF1s5BjfPQBjHgmAg7W5yBUSEdUNw40WDDfUmBSWKLD5RDJW7Y/HjbuFAFTDyJ8LaYLx3QMR4GojcoVERLXDcKMFww01RmUKJf48m4IV+67hYmouAEAiASJae2BCj0CEBTqz8zERGTWGGy0YbqgxEwQBh67ewfcH47H30m319vY+DnipRyAGdfDiGlZEZJQYbrRguCFSuZqei+8PJmDryRsoLlMt1Olpb4nIbgF4PsyP/XKIyKgw3GjBcEOkKTO/BL8cScTa6ERk5BUDAKzMZXiui6pfTiD75RCREWC40YLhhqhmxWUK/H46Bd8diNfol9OvlQcm9AxEOPvlEJGI6vL5bRQ315cvX46AgABYWloiPDwcMTEx991369at6NKlCxwdHWFjY4OOHTvip59+MmC1RKbJwkyGZ0Oa4O9pPbF+QjgebeUOQQB2x6Vh5KojeHLpQWw7dQMl5bewiIiMlegtNxs3bsTYsWOxcuVKhIeHY/Hixdi8eTMuXboEd3f3avvv27cPd+/eRatWrSCXy/HHH3/gzTffxJ9//on+/fs/8HxsuSGqvavpefjx0HX838kbKCpVhRoPewuM7RqA0eF+cLSWi1whETUWDeq2VHh4OEJDQ7Fs2TIAgFKphK+vL6ZOnYp33nmnVsfo3LkzBg0ahAULFjxwX4Yborq7m1+C9TFJWHs4Aem5lf1yngnxwYvdA9HUzVbkConI1DWY21IlJSU4ceIEIiIi1NukUikiIiIQHR39wPcLgoA9e/bg0qVL6NWrV437FBcXIycnR+NBRHXjZCPH5L7NcfDtR/Hl8GC08bJHYakCPx9JwqNf/IeX1hzD4WsZaGRd+IjISJmJefKMjAwoFAp4eHhobPfw8MDFixfv+77s7Gz4+PiguLgYMpkM33zzDR577LEa9124cCHmzZun07qJGiu5mRRPd26Cpzr54Eh8Jr4/GI/dcenYc1H1aONlj5d6BGJwsDfkZkbRpY+IGqEG+dfHzs4OsbGxOHbsGD766CPMmDED+/btq3HfWbNmITs7W/1ITk42bLFEJkgikaBrMxd8FxmKf9/sjTGP+MPSXIoLKTl4c/Np9Fj0L5b9ewV3y9e1IiIyJFH73JSUlMDa2hpbtmzBsGHD1NsjIyORlZWF7du31+o4EyZMQHJyMnbt2vXAfdnnhkg/KvrlrItOQFqOql+OpbkUz4Y0waQ+zeHtaCVyhUTUkDWYPjdyuRwhISHYs2ePeptSqcSePXvQtWvXWh9HqVSiuLhYHyUSUS1V9Ms58L9H8dWIYLT1tkdRqRI/H0lCn8/2Yc72c0jLKRK7TCJqBETtcwMAM2bMQGRkJLp06YKwsDAsXrwY+fn5GD9+PABg7Nix8PHxwcKFCwGo+tB06dIFzZo1Q3FxMf766y/89NNPWLFihZiXQUTl5GZSPNWpCYZ1VPXL+XrPZRyJz8Ta6ERsOJaM0eH+eK1PU7jbWYpdKhGZKNHDzYgRI3D79m3Mnj0bqamp6NixI3bu3KnuZJyUlASptLKBKT8/H5MmTcKNGzdgZWWFVq1a4eeff8aIESPEugQiqkFFv5yuzbri8LUMfPnPZRxPvIsfDl3H+phEjO0agFd7NYWLrYXYpRKRiRF9nhtDY58bInEIgoCDVzPwxT+XEZucBQCwlsswrlsAXu7ZFE42nBCQiO6vQU3iZ2gMN0TiEgQB+y7dxpdRl3H2ZjYAwNbCDC/2CMRLPQLhYMXVyImoOoYbLRhuiIyDIAiIupCGr3ZfQVyKanJNO0szvNyzKcZ3D4CdJUMOEVViuNGC4YbIuCiVAnadT8VXuy/jcloeAMDR2hyv9GqKyK4BsLEQvWsgERkBhhstGG6IjJNSKeCPsylYvPsy4m/nAwCcbeR4rXdTjHkkAFZymcgVEpGYGG60YLghMm4KpYAdp2/i691XkHCnAADgamuBSX2a4flwP1iaM+QQNUYMN1ow3BA1DGUKJbaeuokle67gxt1CAICHvQUm922OEaG+sDBjyCFqTBhutGC4IWpYSsqU+L+TN7B0zxXcylbNcOztYIkpjwbh2ZAmXKCTqJFguNGC4YaoYSouU2DTsWQs23tVvXZVEycrvN4vCE938oGZjCGHyJQx3GjBcEPUsBWVKvBrTBKW772GjDxVyAlwscbr/YIwtKMPZFKJyBUSkT4w3GjBcENkGgpLFPj5SCJW/ncNd/JLAABN3WwwrV8QnuzgzZBDZGIYbrRguCEyLfnFZVgXnYhv919DVkEpAKC5uy2mPtqcIYfIhDDcaMFwQ2SacotKseZQAr47eB3ZhaqQ08zNBlMfDcLgYIYcooaO4UYLhhsi05ZbVIq1h1Uhp6Ilp6mrDab2a47BHbzZ8ZiogWK40YLhhqhxyC0qxbroRKw+EK8OOYGuNpjStzmGdmTIIWpoGG60YLghalzyisuwLjoBq/fH4255yAlwscaUR4MwjCGHqMFguNGC4YaocaroeLz6QDwyy0dX+btYY3Lf5niqkw/MGXKIjBrDjRYMN0SNW35xGX4+kohV++PVQ8j9nK0xpW9zPNWZIYfIWDHcaMFwQ0QAUFBSGXIy8lQhx9fZCpP7NMczIU0YcoiMDMONFgw3RFRVYYkCvxxNxMr/4tUzHvs4WmHKo83xTGeuXUVkLBhutGC4IaKaVIScb/fH43ZuZciZ3Lc5F+gkMgIMN1ow3BCRNkWlCqw/moSV/11DepWQM7FPMzzXpQkszGQiV0jUODHcaMFwQ0S1UbFA54p9lSHH28ESE/s2x3CGHCKDY7jRguGGiOqiqFSBjceSsWLfNaTmFAEAvBwsMbFPMwzv4gtLc4YcIkNguNGC4YaI6qOoVIFNx5Pxzd7KkONpb4nXejfFc118YWNhJnKFRKaN4UYLhhsiehjFZQpsOpaMb/ZdQ0q2KuTYWphhSEdvjAr1Q/smDiJXSGSaGG60YLghIl0oLlNg8/Eb+O5APBLuFKi3t/Oxx8hQPwzt6A07S3MRKyQyLQw3WjDcEJEuCYKAI/GZ+DUmCTvPpaJEoQQAWJnLMDjYCyPD/NDJ1xESiUTkSokaNoYbLRhuiEhf7uaXYOupm/g1JglX0/PU21t62GFUmC+e6tQEDtZszSGqD4YbLRhuiEjfBEHAicS7+DUmGX+cuYXiMlVrjoWZFAPbe2FUmB9CA5zYmkNUBww3WjDcEJEhZReWYnvsTaw/moSLqbnq7c3cbDAy1A/PhDSBs41cxAqJGgaGGy0YbohIDIIg4PSNbGyIScKO07dQUKIAAJjLJOjf1hOjwvzQtakLpFK25hDVhOFGC4YbIhJbXnEZdsTewoZjSThzI1u93d/FGiNCffFsSBO421mKWCGR8WG40YLhhoiMyflb2dgQk4zfTt1EbnEZAMBMKkG/1u4YFeaHnkFukLE1h4jhRhuGGyIyRgUlZfjzTAo2HEvGicS76u0+jlYY3sUXw0ObwMvBSsQKicTFcKMFww0RGbvLabn4NSYJW0/eRHZhKQBAKgH6tnTHyDA/9G3pBjOZVOQqiQyL4UYLhhsiaiiKShXYeS4Vv8Yk4ej1TPX2pq42mNqvOQZ38GbIoUaD4UYLhhsiaoiu3c7DxmPJ2HQ8GVkFqtYchhxqTBhutGC4IaKGLK+4DOuiE7B6fzzu3hNyhgT7sPMxmSyGGy0YbojIFFSEnFX74zVacl7vF4TBwd4MOWRyGG60YLghIlOSV1yGtYcTsPpAlZDjZoPXH2XIIdPCcKMFww0RmaL7hZxp/YLwZAeGHGr4GG60YLghIlOWW1SKddGJDDlkchhutGC4IaLGoCLkrNofr54rp5mbqk8OQw41RAw3WjDcEFFjkltUWn676jpDDjVoDDdaMNwQUWNUU8hp7m6L1/sFYVB7L4YcMnoMN1ow3BBRY5ZTVIq1h1Qdj3OKVAt1MuRQQ8BwowXDDRFRzSEnqDzkDGTIISPEcKMFww0RUaWcolKsOZSA72oIOYPae0HKkENGguFGC4YbIqLq7hdypkUEYWA7hhwSH8ONFgw3RET3l11YHnIOxiO3POTYWZqhnbcD2jdxQDsfB7T3cYC/szUDDxkUw40WDDdERA9WEXK+P1jZklOVnaUZ2nrbo72PKvB0aOLIwEN6xXCjBcMNEVHtlSqUuJyWi3M3s3H2ZjbO3sxBXEoOSsqU1fa1szBDW5/KwNPexwEBLjYMPKQTDDdaMNwQET2cUoUSV9LyqgSebMSl5KD4PoGnTXkLT8VtrUAGHqoHhhstGG6IiHSvVKHE1fQ8nL2ZrQ49F27VHHhsqwae8laepq4MPKQdw40WDDdERIZRplDi6u08nL1RJfCk5KCotHrgsZHL0NZbFXRC/J3Qq4Ur7CzNRaiajBXDjRYMN0RE4ilTKHHtdn61Fp7CUoXGfuYyCR5p6oLH2nigX2sP+DhaiVQxGQuGGy0YboiIjItCKeBaeQvP2ZvZ2H/lNuJv52vs08bLHhFtPPBYaw+087GHRMJbWI0Nw40WDDdERMbv2u087IlLQ9SFNJxIvAtllU8qT3tL9Gvtjog2HujWzAUWZjLxCiWDYbjRguGGiKhhuZNXjL2XbmP3hTTsv3IbBSWVt7Bs5DL0auGGiNYe6NvKHc42chErJX1iuNGC4YaIqOEqKlUgOv4Odl9Iw+64NKTlFKtfk0qALv7OiGjjjojWHmjqZitipaRrDDdaMNwQEZkGQRBw7mYOoi6kIiouHXEpORqvN3WzwWOtPfBYGw908nPiSucNHMONFgw3RESm6cbdAuyJS8fuuDQcib+DUkXlx5uzjRyPtlK16PQMcoWNhZmIlVJ9NLhws3z5cnz22WdITU1FcHAwli5dirCwsBr3Xb16NdatW4dz584BAEJCQvDxxx/fd/97MdwQEZm+nKJS7L+s6qfz78V0jfWx5GZSdG/mgog2Hoho7QEPe0sRK6XaalDhZuPGjRg7dixWrlyJ8PBwLF68GJs3b8alS5fg7u5ebf/Ro0eje/fu6NatGywtLbFo0SJs27YN58+fh4+PzwPPx3BDRNS4lCqUOJaQid0X0hEVl4rkzEKN1zv6OmLMI/54MtiLI6+MWIMKN+Hh4QgNDcWyZcsAAEqlEr6+vpg6dSreeeedB75foVDAyckJy5Ytw9ixYx+4P8MNEVHjJQgCrqTnIaq8Q3JschYqPgVdbS3wwiN+GB3uDzc7C3ELpWrq8vkt6k3HkpISnDhxArNmzVJvk0qliIiIQHR0dK2OUVBQgNLSUjg7O9f4enFxMYqLK3vT5+Tk1LgfERGZPolEghYedmjhYYfJfZsjPbcIm4/fwLroBKTlFGPx7iv4Zu81PBnshRe7B6Kdj4PYJVM9SMU8eUZGBhQKBTw8PDS2e3h4IDU1tVbHePvtt+Ht7Y2IiIgaX1+4cCEcHBzUD19f34eum4iITIO7nSUm922Og28/iiWjOqGTnyNKFEpsPXkTTy49iOEro/H32RSUKaqvh0XGq0F3F//kk0+wYcMG7Nu3D5aWNXcImzVrFmbMmKF+npOTw4BDREQazGVSDAn2xpBgb5xKuosfDyXgr7MpiEnIRExCJnwcrTC2qz9GhvrBwZoLeho7UcONq6srZDIZ0tLSNLanpaXB09NT63s///xzfPLJJ9i9ezc6dOhw3/0sLCxgYcF7p0REVDud/JzQyc8J7w5sjZ+PJGJ9TBJuZhVi4d8XsXj3FTwT4oNx3QLR3J2TBBorUW9LyeVyhISEYM+ePeptSqUSe/bsQdeuXe/7vk8//RQLFizAzp070aVLF0OUSkREjYyngyXe6t8Sh995FJ8+0wGtPO1QWKrAz0eSEPHlfxj7Qwz2XkqHUin6jCp0D9FHS23cuBGRkZH49ttvERYWhsWLF2PTpk24ePEiPDw8MHbsWPj4+GDhwoUAgEWLFmH27NlYv349unfvrj6Ora0tbG0fnKI5WoqIiOpDEAREx9/Bj4cSsDsuTT3KqqmbDcZ3C8DTnZtwckA9alBDwQFg2bJl6kn8OnbsiCVLliA8PBwA0KdPHwQEBGDNmjUAgICAACQmJlY7xpw5czB37twHnovhhoiIHlbSnQKsjU7ApmPJyC1WTRBoZ2mGkaG+GNs1AL7O1iJXaHoaXLgxJIYbIiLSlbziMmw5now1hxOQcKcAgGoBz8fbeGJ89wCEBTpDIuGaVrrAcKMFww0REemaUilg3+V0/HgoAQeuZKi3t/Gyx/juARgc7A1Lc85+/DAYbrRguCEiIn26nJaLHw8lYNupGygqVc2P42Ijx+hwP7zwiD/cuZZVvTDcaMFwQ0REhpBVUIJfY5KxLjoBKdlFAABzmQQRrT3gamsBmVQCM6kEMln5V6m0/KtE/dVcJtV4biareT8zqbTK65rb5WZS+DlbQyZt2LfHGG60YLghIiJDKlMoset8Gn44dB0nEu+KUoOdpRnCApzxSFMXPNLUBW287Rtc2GG40YLhhoiIxHLmRhYOXMlASZkSCqWAMqUAhVKJMqWAMoXmc/Xr99te9f2KqtuUVd4jIL+kTH17rIKdhRnCAp0R3lQVeNp42cNMJurUdw/EcKMFww0RETUmCqWAC7dycCT+Do7E30HM9Uz18PUKdhZmCA10xiNGHHYYbrRguCEiosZMoRQQl1IZdo5ez0RukfGHHYYbLRhuiIiIKtUm7NhamCE0wEndZ6ett+HDDsONFgw3RERE96cZdjIRc/0Ocowg7DDcaMFwQ0REVHu1DTtdqoSddnoIOww3WjDcEBER1Z9CKeBiag6OxGeqbmPFVw87AS7W2Dezr07PW5fPby5fSkRERLUmk0rQ1tsBbb0d8FKPQI2wc7S8z05bHwdRa2S4ISIionq7N+wolUK1DsmGZlyD2ImIiKhBk0olcLA2F7cGUc9OREREpGMMN0RERGRSGG6IiIjIpDDcEBERkUlhuCEiIiKTwnBDREREJoXhhoiIiEwKww0RERGZFIYbIiIiMikMN0RERGRSGG6IiIjIpDDcEBERkUlhuCEiIiKTYiZ2AYYmCAIAICcnR+RKiIiIqLYqPrcrPse1aXThJjc3FwDg6+srciVERERUV7m5uXBwcNC6j0SoTQQyIUqlErdu3YKdnR0kEolOj52TkwNfX18kJyfD3t5ep8c2BqZ+fYDpXyOvr+Ez9Wvk9TV8+rpGQRCQm5sLb29vSKXae9U0upYbqVSKJk2a6PUc9vb2JvsfLWD61weY/jXy+ho+U79GXl/Dp49rfFCLTQV2KCYiIiKTwnBDREREJoXhRocsLCwwZ84cWFhYiF2KXpj69QGmf428vobP1K+R19fwGcM1NroOxURERGTa2HJDREREJoXhhoiIiEwKww0RERGZFIYbIiIiMikMNzqyfPlyBAQEwNLSEuHh4YiJiRG7JJ1ZuHAhQkNDYWdnB3d3dwwbNgyXLl0Suyy9+eSTTyCRSDB9+nSxS9GZmzdv4oUXXoCLiwusrKzQvn17HD9+XOyydEahUOCDDz5AYGAgrKys0KxZMyxYsKBWa9AYo/3792Pw4MHw9vaGRCLBb7/9pvG6IAiYPXs2vLy8YGVlhYiICFy5ckWcYutJ2zWWlpbi7bffRvv27WFjYwNvb2+MHTsWt27dEq/gOnrQ77Cq1157DRKJBIsXLzZYfQ+rNtcXFxeHIUOGwMHBATY2NggNDUVSUpJB6mO40YGNGzdixowZmDNnDk6ePIng4GD0798f6enpYpemE//99x8mT56MI0eOICoqCqWlpXj88ceRn58vdmk6d+zYMXz77bfo0KGD2KXozN27d9G9e3eYm5vj77//xoULF/DFF1/AyclJ7NJ0ZtGiRVixYgWWLVuGuLg4LFq0CJ9++imWLl0qdmn1kp+fj+DgYCxfvrzG1z/99FMsWbIEK1euxNGjR2FjY4P+/fujqKjIwJXWn7ZrLCgowMmTJ/HBBx/g5MmT2Lp1Ky5duoQhQ4aIUGn9POh3WGHbtm04cuQIvL29DVSZbjzo+q5du4YePXqgVatW2LdvH86cOYMPPvgAlpaWhilQoIcWFhYmTJ48Wf1coVAI3t7ewsKFC0WsSn/S09MFAMJ///0ndik6lZubKwQFBQlRUVFC7969hWnTpoldkk68/fbbQo8ePcQuQ68GDRokvPjiixrbnn76aWH06NEiVaQ7AIRt27apnyuVSsHT01P47LPP1NuysrIECwsL4ddffxWhwod37zXWJCYmRgAgJCYmGqYoHbrf9d24cUPw8fERzp07J/j7+wtfffWVwWvThZqub8SIEcILL7wgTkGCILDl5iGVlJTgxIkTiIiIUG+TSqWIiIhAdHS0iJXpT3Z2NgDA2dlZ5Ep0a/LkyRg0aJDG79IU7NixA126dMFzzz0Hd3d3dOrUCatXrxa7LJ3q1q0b9uzZg8uXLwMATp8+jYMHD+KJJ54QuTLdu379OlJTUzX+O3VwcEB4eLjJ/s0BVH93JBIJHB0dxS5FJ5RKJcaMGYOZM2eibdu2YpejU0qlEn/++SdatGiB/v37w93dHeHh4Vpvzekaw81DysjIgEKhgIeHh8Z2Dw8PpKamilSV/iiVSkyfPh3du3dHu3btxC5HZzZs2ICTJ09i4cKFYpeic/Hx8VixYgWCgoKwa9cuTJw4Ea+//jrWrl0rdmk6884772DkyJFo1aoVzM3N0alTJ0yfPh2jR48WuzSdq/i70lj+5gBAUVER3n77bYwaNcpkFptctGgRzMzM8Prrr4tdis6lp6cjLy8Pn3zyCQYMGIB//vkHTz31FJ5++mn8999/Bqmh0a0KTg9n8uTJOHfuHA4ePCh2KTqTnJyMadOmISoqynD3gw1IqVSiS5cu+PjjjwEAnTp1wrlz57By5UpERkaKXJ1ubNq0Cb/88gvWr1+Ptm3bIjY2FtOnT4e3t7fJXGNjVVpaiuHDh0MQBKxYsULscnTixIkT+Prrr3Hy5ElIJBKxy9E5pVIJABg6dCjeeOMNAEDHjh1x+PBhrFy5Er1799Z7DWy5eUiurq6QyWRIS0vT2J6WlgZPT0+RqtKPKVOm4I8//sDevXvRpEkTscvRmRMnTiA9PR2dO3eGmZkZzMzM8N9//2HJkiUwMzODQqEQu8SH4uXlhTZt2mhsa926tcFGLRjCzJkz1a037du3x5gxY/DGG2+YZEtcxd+VxvA3pyLYJCYmIioqymRabQ4cOID09HT4+fmp/+YkJibizTffREBAgNjlPTRXV1eYmZmJ+neH4eYhyeVyhISEYM+ePeptSqUSe/bsQdeuXUWsTHcEQcCUKVOwbds2/PvvvwgMDBS7JJ3q168fzp49i9jYWPWjS5cuGD16NGJjYyGTycQu8aF079692tD9y5cvw9/fX6SKdK+goABSqeafM5lMpv4XpCkJDAyEp6enxt+cnJwcHD161GT+5gCVwebKlSvYvXs3XFxcxC5JZ8aMGYMzZ85o/M3x9vbGzJkzsWvXLrHLe2hyuRyhoaGi/t3hbSkdmDFjBiIjI9GlSxeEhYVh8eLFyM/Px/jx48UuTScmT56M9evXY/v27bCzs1Pf13dwcICVlZXI1T08Ozu7av2HbGxs4OLiYhL9it544w1069YNH3/8MYYPH46YmBisWrUKq1atErs0nRk8eDA++ugj+Pn5oW3btjh16hS+/PJLvPjii2KXVi95eXm4evWq+vn169cRGxsLZ2dn+Pn5Yfr06fjwww8RFBSEwMBAfPDBB/D29sawYcPEK7qOtF2jl5cXnn32WZw8eRJ//PEHFAqF+u+Os7Mz5HK5WGXX2oN+h/eGNXNzc3h6eqJly5aGLrVeHnR9M2fOxIgRI9CrVy/07dsXO3fuxO+//459+/YZpkDRxmmZmKVLlwp+fn6CXC4XwsLChCNHjohdks4AqPHx448/il2a3pjSUHBBEITff/9daNeunWBhYSG0atVKWLVqldgl6VROTo4wbdo0wc/PT7C0tBSaNm0qvPfee0JxcbHYpdXL3r17a/x/LjIyUhAE1XDwDz74QPDw8BAsLCyEfv36CZcuXRK36DrSdo3Xr1+/79+dvXv3il16rTzod3ivhjYUvDbX9/333wvNmzcXLC0theDgYOG3334zWH0SQWigU3gSERER1YB9boiIiMikMNwQERGRSWG4ISIiIpPCcENEREQmheGGiIiITArDDREREZkUhhsiIiIyKQw3RNToSSQS/Pbbb2KXQUQ6wnBDRKIaN24cJBJJtceAAQPELo2IGiiuLUVEohswYAB+/PFHjW0WFhYiVUNEDR1bbohIdBYWFvD09NR4ODk5AVDdMlqxYgWeeOIJWFlZoWnTptiyZYvG+8+ePYtHH30UVlZWcHFxwSuvvIK8vDyNfX744Qe0bdsWFhYW8PLywpQpUzRez8jIwFNPPQVra2sEBQVhx44d+r1oItIbhhsiMnoffPABnnnmGZw+fRqjR4/GyJEjERcXBwDIz89H//794eTkhGPHjmHz5s3YvXu3RnhZsWIFJk+ejFdeeQVnz57Fjh070Lx5c41zzJs3D8OHD8eZM2cwcOBAjB49GpmZmQa9TiLSEYMt0UlEVIPIyEhBJpMJNjY2Go+PPvpIEATVqvSvvfaaxnvCw8OFiRMnCoIgCKtWrRKcnJyEvLw89et//vmnIJVKhdTUVEEQBMHb21t477337lsDAOH9999XP8/LyxMACH///bfOrpOIDId9bohIdH379sWKFSs0tjk7O6u/79q1q8ZrXbt2RWxsLAAgLi4OwcHBsLGxUb/evXt3KJVKXLp0CRKJBLdu3UK/fv201tChQwf19zY2NrC3t0d6enp9L4mIRMRwQ0Sis7GxqXabSFesrKxqtZ+5ubnGc4lEAqVSqY+SiEjP2OeGiIzekSNHqj1v3bo1AKB169Y4ffo08vPz1a8fOnQIUqkULVu2hJ2dHQICArBnzx6D1kxE4mHLDRGJrri4GKmpqRrbzMzM4OrqCgDYvHkzunTpgh49euCXX35BTEwMvv/+ewDA6NGjMWfOHERGRmLu3Lm4ffs2pk6dijFjxsDDwwMAMHfuXLz22mtwd3fHE088gdzcXBw6dAhTp0417IUSkUEw3BCR6Hbu3AkvLy+NbS1btsTFixcBqEYybdiwAZMmTYKXlxd+/fVXtGnTBgBgbW2NXbt2Ydq0aQgNDYW1tTWeeeYZfPnll+pjRUZGoqioCF999RXeeustuLq64tlnnzXcBRKRQUkEQRDELoKI6H4kEgm2bduGYcOGiV0KETUQ7HNDREREJoXhhoiIiEwK+9wQkVHjnXMiqiu23BAREZFJYbghIiIik8JwQ0RERCaF4YaIiIhMCsMNERERmRSGGyIiIjIpDDdERERkUhhuiIiIyKQw3BAREZFJ+X8tes2izZhkUQAAAABJRU5ErkJggg==",
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAkAAAAHHCAYAAABXx+fLAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAABy90lEQVR4nO3dd1xV9f8H8Ne9F+5l740IiApOnJAjR1k4cm/NlW1nZKW5teRnlpnmN7/11bTMnZplaUpquVduQHGByka2rHvP748DV68MGRcOF17Px+M+uPfcc899H0Duy886MkEQBBARERHVIXKpCyAiIiKqbgxAREREVOcwABEREVGdwwBEREREdQ4DEBEREdU5DEBERERU5zAAERERUZ3DAERERER1DgMQERER1TkMQESlGD9+PLy8vKQuQzJ37tyBTCbD+vXrtdsWLFgAmUxWptfLZDIsWLBArzV169YN3bp10+sxazovLy+MHz++Qq+tip8BUW3AAEQGSSaTlel2+PBhqUutNv369YOZmRnS09NL3Gf06NFQKpVISkqqxsrK79q1a1iwYAHu3LkjdSlahw8f1v5ebdy4sdh9OnXqBJlMhubNm1dzdfoTFhYGmUwGExMTpKSkSF0OUZUxkroAoor48ccfdR7/8MMPOHDgQJHtTZo0qdT7fPfdd9BoNJU6RnUZPXo0fv31V+zatQtjx44t8nxWVhZ++eUX9OzZE/b29hV+nzlz5mDmzJmVKfWZrl27hoULF6Jbt25FWuD+/PPPKn3vZzExMcGmTZvw6quv6my/c+cOjh8/DhMTE4kq04+NGzfCxcUFDx8+xI4dO/D6669LXRJRlWAAIoP09IfPyZMnceDAgSLbn5aVlQUzM7Myv4+xsXGF6pNCv379YGlpiU2bNhUbgH755RdkZmZi9OjRlXofIyMjGBlJ96dDqVRK9t4A0Lt3b+zZsweJiYlwcHDQbt+0aROcnZ3RqFEjPHz4UMIKK04QBGzatAmjRo3C7du38dNPP9XYAJSZmQlzc3OpyyADxi4wqrW6deuG5s2b49y5c+jSpQvMzMzw8ccfAxDDQJ8+feDm5gaVSgUfHx8sXrwYarVa5xhPjwEqHBPz+eef49tvv4WPjw9UKhXat2+PM2fOlFrP2bNnIZPJsGHDhiLP7d+/HzKZDL/99hsAID09HdOnT4eXlxdUKhWcnJzw0ksv4fz58yUe39TUFIMGDUJoaCji4+OLPL9p0yZYWlqiX79+SE5OxowZM9CiRQtYWFjAysoKvXr1wsWLF0s9B6D4MUA5OTl477334OjoqH2Pe/fuFXnt3bt38e6778LX1xempqawt7fH0KFDdbq61q9fj6FDhwIAunfvXqQ7s7gxQPHx8Zg4cSKcnZ1hYmICf3//It/nyvzsntS/f3+oVCps375dZ/umTZswbNgwKBSKIq/Jz8/H4sWLte/p5eWFjz/+GDk5OTr7CYKATz75BPXq1YOZmRm6d++Oq1evFltHSkoKpk+fDg8PD6hUKjRs2BBLly6tVIvlsWPHcOfOHYwYMQIjRozA33//XezPUaPR4KuvvkKLFi1gYmICR0dH9OzZE2fPntXZb+PGjQgICICZmRlsbW3RpUsXnRa8ksYnPT3maf369ZDJZDhy5AjeffddODk5oV69egDK9jtVKCUlBe+9957231W9evUwduxYJCYmIiMjA+bm5pg2bVqR1927dw8KhQIhISFl/E6SIWALENVqSUlJ6NWrF0aMGIFXX30Vzs7OAMQ/qBYWFggODoaFhQX++usvzJs3D2lpaVi2bNkzj7tp0yakp6fjrbfegkwmw2effYZBgwbh1q1bJbYatWvXDg0aNMC2bdswbtw4nee2bt0KW1tbBAUFAQDefvtt7NixA5MnT0bTpk2RlJSEo0ePIiwsDG3atCmxrtGjR2PDhg3Ytm0bJk+erN2enJyM/fv3Y+TIkTA1NcXVq1exe/duDB06FN7e3oiLi8N///tfdO3aFdeuXYObm9szvwdPev3117Fx40aMGjUKHTt2xF9//YU+ffoU2e/MmTM4fvw4RowYgXr16uHOnTv45ptv0K1bN1y7dg1mZmbo0qULpk6dipUrV+Ljjz/WdmOW1J356NEjdOvWDZGRkZg8eTK8vb2xfft2jB8/HikpKUU+0Crys3uSmZkZ+vfvj82bN+Odd94BAFy8eBFXr17F//73P1y6dKnY78+GDRswZMgQvP/++zh16hRCQkIQFhaGXbt2afebN28ePvnkE/Tu3Ru9e/fG+fPn8fLLLyM3N1fneFlZWejatSvu37+Pt956C/Xr18fx48cxa9YsxMTEYMWKFc88j+L89NNP8PHxQfv27dG8eXOYmZlh8+bN+OCDD3T2mzhxItavX49evXrh9ddfR35+Pv755x+cPHkS7dq1AwAsXLgQCxYsQMeOHbFo0SIolUqcOnUKf/31F15++eUK1ffuu+/C0dER8+bNQ2ZmJoCy/U4BQEZGBp5//nmEhYXhtddeQ5s2bZCYmIg9e/bg3r17aNWqFQYOHIitW7di+fLlOkF28+bNEASh0q2nVMMIRLXApEmThKd/nbt27SoAENasWVNk/6ysrCLb3nrrLcHMzEzIzs7Wbhs3bpzg6empfXz79m0BgGBvby8kJydrt//yyy8CAOHXX38ttc5Zs2YJxsbGOq/NyckRbGxshNdee027zdraWpg0aVKpxypOfn6+4OrqKnTo0EFn+5o1awQAwv79+wVBEITs7GxBrVbr7HP79m1BpVIJixYtKnK+33//vXbb/Pnzdb7XFy5cEAAI7777rs7xRo0aJQAQ5s+fr91W3Pf9xIkTAgDhhx9+0G7bvn27AEA4dOhQkf27du0qdO3aVft4xYoVAgBh48aN2m25ublChw4dBAsLCyEtLU3nXCr6szt06JAAQNi+fbvw22+/CTKZTIiKihIEQRA++OADoUGDBtr6mjVrVuT78/rrr+scb8aMGQIA4a+//hIEQRDi4+MFpVIp9OnTR9BoNNr9Pv74YwGAMG7cOO22xYsXC+bm5sL169d1jjlz5kxBoVBo6xIEocjPoCS5ubmCvb29MHv2bO22UaNGCf7+/jr7/fXXXwIAYerUqUWOUVj3jRs3BLlcLgwcOLDI79mT51ZSbZ6enjrn+/333wsAhM6dOwv5+fk6+5b1d2revHkCAGHnzp0l1r1//34BgPDHH3/oPN+yZUud3zmqHdgFRrWaSqXChAkTimw3NTXV3k9PT0diYiKef/55ZGVlITw8/JnHHT58OGxtbbWPn3/+eQDArVu3nvm6vLw87Ny5U7vtzz//REpKCoYPH67dZmNjg1OnTuHBgwfPrOVJCoUCI0aMwIkTJ3S6AArHp7z44osAxO+LXC7+81er1UhKSoKFhQV8fX1L7WYrzu+//w4AmDp1qs726dOnF9n3ye97Xl4ekpKS0LBhQ9jY2JT7fZ98fxcXF4wcOVK7zdjYGFOnTkVGRgaOHDmis39Ff3ZPevnll2FnZ4ctW7ZAEARs2bJF5/2frg8AgoODdba///77AIC9e/cCAA4ePIjc3FxMmTJFp4uxuO/j9u3b8fzzz8PW1haJiYnaW48ePaBWq/H333+X+VwK/fHHH0hKStI5j5EjR2pbtwr9/PPPkMlkmD9/fpFjFNa9e/duaDQazJs3T/t79vQ+FfHGG28U6WIs6+/Uzz//DH9/fwwcOLDEunv06AE3Nzf89NNP2ueuXLmCS5cuPXN8IRkeBiCq1dzd3YsdNHv16lUMHDgQ1tbWsLKygqOjo/YPXGpq6jOPW79+fZ3HhR+ozxr86u/vDz8/P2zdulW7bevWrXBwcMALL7yg3fbZZ5/hypUr8PDwQEBAABYsWFDmD+jCZvpNmzYBEMcv/PPPPxgxYoT2w0Oj0eDLL79Eo0aNoFKp4ODgAEdHR1y6dKlM5/+ku3fvQi6Xw8fHR2e7r69vkX0fPXqEefPmacetFL5vSkpKud/3yfdv1KhRkQ/awi6zu3fv6myv6M/uScbGxhg6dCg2bdqEv//+G9HR0Rg1alSJ9cnlcjRs2FBnu4uLC2xsbLT1FX5t1KiRzn6Ojo46gQ0Abty4gX379sHR0VHn1qNHDwAodgzYs2zcuBHe3t5QqVSIjIxEZGQkfHx8YGZmphMIbt68CTc3N9jZ2ZV4rJs3b0Iul6Np06blrqM03t7eRbaV9Xfq5s2bz1yeQC6XY/To0di9ezeysrIAiN2CJiYm2nFpVHswAFGt9uT/DgulpKSga9euuHjxIhYtWoRff/0VBw4cwNKlSwGgTINIixvoCoiDWJ9l+PDhOHToEBITE5GTk4M9e/Zg8ODBOjOrhg0bhlu3bmHVqlVwc3PDsmXL0KxZM/zxxx/PPH7btm3h5+eHzZs3Ayh+/MKSJUsQHByMLl26YOPGjdi/fz8OHDiAZs2aVem0/ylTpuDTTz/FsGHDsG3bNvz55584cOAA7O3tq225gcr87J40atQoXLhwAQsWLIC/v/8zP+wr0/LxNI1Gg5deegkHDhwo9jZ48OByHS8tLQ2//vorbt++jUaNGmlvTZs2RVZWFjZt2lTu709lPD0ZoVBx/571/Ts1duxYZGRkYPfu3dpZca+88gqsra3LfSyq2TgImuqcw4cPIykpCTt37kSXLl2022/fvl0t7z98+HAsXLgQP//8M5ydnZGWloYRI0YU2c/V1RXvvvsu3n33XcTHx6NNmzb49NNP0atXr2e+x+jRozF37lxcunQJmzZtQqNGjdC+fXvt8zt27ED37t2xdu1andelpKToTO0uC09PT2g0Gty8eVOn1SciIqLIvjt27MC4cePwxRdfaLdlZ2cXWXCvPGHB09MTly5dgkaj0WkFKuzK9PT0LPOxyqNz586oX78+Dh8+rA3PJdWn0Whw48YNnYHccXFxSElJ0dZX+PXGjRto0KCBdr+EhIQirVM+Pj7IyMjQtvhU1s6dO5GdnY1vvvmmyM8/IiICc+bMwbFjx9C5c2f4+Phg//79SE5OLrEVyMfHBxqNBteuXUOrVq1KfF9bW9siP/vc3FzExMSUufay/k75+PjgypUrzzxe8+bN0bp1a/z000+oV68eoqKisGrVqjLXQ4aDLUBU5xS2ADz5P9rc3Fz85z//qZb3b9KkCVq0aIGtW7di69atcHV11QliarW6SHeQk5MT3NzcikybLklha8+8efNw4cKFIrNXFApFkf/Rb9++Hffv3y/3+RQGspUrV+psL24mUnHvu2rVqiL/4y9c36UsKxH37t0bsbGxOt2K+fn5WLVqFSwsLNC1a9eynEa5yWQyrFy5EvPnz8eYMWNKrQ8o+v1Yvnw5AGhny/Xo0QPGxsZYtWqVzveouO/jsGHDcOLECezfv7/IcykpKcjPzy/XuWzcuBENGjTA22+/jSFDhujcZsyYAQsLC2032ODBgyEIAhYuXFjkOIV1DxgwAHK5HIsWLSrSCvPkufn4+BQZr/Ttt9+W2AJUnLL+Tg0ePBgXL17UmXVXXE0AMGbMGPz5559YsWIF7O3ty/SfDjI8bAGiOqdjx46wtbXFuHHjMHXqVMhkMvz444/V2sQ/fPhwzJs3DyYmJpg4caJOy0V6ejrq1auHIUOGwN/fHxYWFjh48CDOnDmj87/c0nh7e6Njx4745ZdfAKBIAHrllVewaNEiTJgwAR07dsTly5fx008/6bQ8lFWrVq0wcuRI/Oc//0Fqaio6duyI0NBQREZGFtn3lVdewY8//ghra2s0bdoUJ06cwMGDB4usTN2qVSsoFAosXboUqampUKlUeOGFF+Dk5FTkmG+++Sb++9//Yvz48Th37hy8vLywY8cOHDt2DCtWrIClpWW5z6ms+vfvj/79+5e6j7+/P8aNG4dvv/1W2/16+vRpbNiwAQMGDED37t0BiGN9ZsyYgZCQELzyyivo3bs3/v33X/zxxx9FWmU++OAD7NmzB6+88grGjx+Ptm3bIjMzE5cvX8aOHTtw586dMrfkPXjwAIcOHSoyiL2QSqVCUFAQtm/fjpUrV6J79+4YM2YMVq5ciRs3bqBnz57QaDT4559/0L17d0yePBkNGzbE7NmzsXjxYjz//PMYNGgQVCoVzpw5Azc3N+16Oq+//jrefvttDB48GC+99BIuXryI/fv3l6sVsqy/Ux988AF27NiBoUOH4rXXXkPbtm2RnJyMPXv2YM2aNfD399fuO2rUKHz44YfYtWsX3nnnHYNaEJXKofonnhHpX0nT4J+cjvykY8eOCc8995xgamoquLm5CR9++KF2CuyTU69Lmga/bNmyIsdEGacbC4I4TRiAAEA4evSoznM5OTnCBx98IPj7+wuWlpaCubm54O/vL/znP/8p07ELrV69WgAgBAQEFHkuOztbeP/99wVXV1fB1NRU6NSpk3DixIkiU8zLMg1eEATh0aNHwtSpUwV7e3vB3Nxc6Nu3rxAdHV3ke/Lw4UNhwoQJgoODg2BhYSEEBQUJ4eHhRaY9C4IgfPfdd0KDBg0EhUKh83N5ukZBEIS4uDjtcZVKpdCiRQudmp88l4r+7J6cBl+a4n7v8vLyhIULFwre3t6CsbGx4OHhIcyaNUtnyQVBEAS1Wi0sXLhQ+3Pp1q2bcOXKlWK/P+np6cKsWbOEhg0bCkqlUnBwcBA6duwofP7550Jubm6Zz+2LL74QAAihoaEl7rN+/XoBgPDLL78IgiAut7Bs2TLBz89PUCqVgqOjo9CrVy/h3LlzOq9bt26d0Lp1a0GlUgm2trZC165dhQMHDuic70cffSQ4ODgIZmZmQlBQkBAZGVniNPgzZ84Uqa08v1NJSUnC5MmTBXd3d0GpVAr16tUTxo0bJyQmJhY5bu/evQUAwvHjx0v8vpBhkwlCNf63l4iIyAAMHDgQly9fLrYlk2oHjgEiIiJ6QkxMDPbu3Vvq2C4yfBwDREREBHEm6LFjx/C///0PxsbGeOutt6QuiaoQW4CIiIgAHDlyBGPGjMHt27exYcMGuLi4SF0SVSGOASIiIqI6R/IWoNWrV8PLywsmJiYIDAzE6dOnS9w3Ly8PixYtgo+PD0xMTODv7499+/bp7LNgwQLIZDKdm5+fX1WfBhERERkQSQPQ1q1bERwcjPnz5+P8+fPw9/dHUFBQidexmTNnDv773/9i1apVuHbtGt5++20MHDgQ//77r85+zZo1Q0xMjPZ29OjR6jgdIiIiMhCSdoEFBgaiffv2+PrrrwGI17fx8PDAlClTMHPmzCL7u7m5Yfbs2Zg0aZJ22+DBg2FqaoqNGzcCEFuAdu/ejQsXLlS4Lo1GgwcPHsDS0lKv1+8hIiKiqiMIAtLT0+Hm5lbkAslPk2wWWG5uLs6dO4dZs2Zpt8nlcvTo0QMnTpwo9jU5OTkwMTHR2WZqalqkhefGjRtwc3ODiYkJOnTogJCQkCJXgH76uE9eYuD+/ft6v4oxERERVY/o6GjUq1ev1H0kC0CJiYlQq9VwdnbW2e7s7Ky9iOHTgoKCsHz5cnTp0gU+Pj4IDQ3Fzp07da75EhgYiPXr18PX1xcxMTFYuHAhnn/+eVy5cqXEJfFDQkKKva5NdHQ0rKysKnGWREREVF3S0tLg4eFRpkvgGNQ6QF999RXeeOMN+Pn5QSaTwcfHBxMmTMC6deu0+zx50bqWLVsiMDAQnp6e2LZtGyZOnFjscWfNmoXg4GDt48JvoJWVFQMQERGRgSnL8BXJBkE7ODhAoVAgLi5OZ3tcXFyJay84Ojpi9+7dyMzMxN27dxEeHg4LC4tSL+BoY2ODxo0bl7qcuUql0oYdhh4iIqLaT7IApFQq0bZtW4SGhmq3aTQahIaGokOHDqW+1sTEBO7u7sjPz8fPP/9c6tWYMzIycPPmTbi6uuqtdiIiIjJskk6DDw4OxnfffYcNGzYgLCwM77zzDjIzMzFhwgQAwNixY3UGSZ86dQo7d+7ErVu38M8//6Bnz57QaDT48MMPtfvMmDEDR44cwZ07d3D8+HEMHDgQCoUCI0eOrPbzIyIioppJ0jFAw4cPR0JCAubNm4fY2Fi0atUK+/bt0w6MjoqK0pnGlp2djTlz5uDWrVuwsLBA79698eOPP8LGxka7z7179zBy5EgkJSXB0dERnTt3xsmTJ+Ho6Kj3+tVqNfLy8vR+XKp+xsbGUCgUUpdBRETVhJfCKEZaWhqsra2Rmppa7HggQRAQGxuLlJSU6i+OqoyNjQ1cXFy49hMRkYF61uf3kwxqFlhNURh+nJycYGZmxg9MAycIArKysrQrkHO8GBFR7ccAVE5qtVobfuzt7aUuh/TE1NQUABAfHw8nJyd2hxER1XKSXwzV0BSO+TEzM5O4EtK3wp8px3UREdV+DEAVxG6v2oc/UyKiuoMBiIiIiOocBiCqMC8vL6xYsULqMoiIiMqNAagOkMlkpd4WLFhQoeOeOXMGb775pn6LJSIiqgacBVYHxMTEaO9v3boV8+bNQ0REhHabhYWF9r4gCFCr1TAyevavRlUsLklERLWbRiPgbnIWzFUKOFmaSFYHW4DqABcXF+3N2toaMplM+zg8PByWlpb4448/0LZtW6hUKhw9ehQ3b95E//794ezsDAsLC7Rv3x4HDx7UOe7TXWAymQz/+9//MHDgQJiZmaFRo0bYs2dPNZ8tERHVFOnZeThzJxk/nLiDWTsvY8DqY2g2fz+6f34YP5+7L2ltbAHSA0EQ8ChPXe3va2qs0NvMpZkzZ+Lzzz9HgwYNYGtri+joaPTu3RuffvopVCoVfvjhB/Tt2xcRERGoX79+icdZuHAhPvvsMyxbtgyrVq3C6NGjcffuXdjZ2emlTiIiqnkKW3XCY9IQFpOGsNh0hMWk4d7DR8XurzSSIyNH2iVHGID04FGeGk3n7a/29722KAhmSv38CBctWoSXXnpJ+9jOzg7+/v7ax4sXL8auXbuwZ88eTJ48ucTjjB8/Xnvh2SVLlmDlypU4ffo0evbsqZc6iYhIWunZeQgvCDhhMeLXiNj0EhsCXK1N0MTVCn4ulmjiaoUmrpbwsjeHkULaTigGIAIAtGvXTudxRkYGFixYgL179yImJgb5+fl49OgRoqKiSj1Oy5YttffNzc1hZWWlvcQEERE9m1ojICMnH6bGChgrZJKtUVbeVh2VkRyNnS3RxNUSfi5W2rBjY6as5srLhgFID0yNFbi2KEiS99UXc3NzncczZszAgQMH8Pnnn6Nhw4YwNTXFkCFDkJubW+pxjI2NdR7LZDJoNBq91UlEVNsIgoAb8Rk4FpmIY5FJOHU7CenZ+QAAuUz8W2+qVMDEWAFT4ye+KhUwNZZrt5kU7Geq3U+us83kidc+3ibuk6vWIKIWtOqUBwOQHshkMr11RdUUx44dw/jx4zFw4EAAYovQnTt3pC2KiKiWiE7OwvGbiTh+MwnHbyYhIT2n2P00ApCZq0ZmbvWPMwXEVh1fF0tt0BFbdmpuq0551K5PbdKbRo0aYefOnejbty9kMhnmzp3LlhwiogpKzMjB8ZtJOHFTbOWJSs7Sed7EWI72Xnbo4GOPTj4O8HWxRE6+Btl5ajzKVSM7X/z6KE+N7Dw1svM0Oo8f76N5Yh/x+Ue5j+9n5xU8X/DafI2graE2tOqUBwMQFWv58uV47bXX0LFjRzg4OOCjjz5CWlqa1GURERmE9Ow8nLqVXNDCk4jw2HSd5xVyGVp52KCTjz06+DigjacNVEa6wxpMjBWwNtUdVqBveWoxZAGApUnVvldNIxMEQXj2bnVLWloarK2tkZqaCisrK53nsrOzcfv2bXh7e8PERLoFnEj/+LMloorKzlPj/N2HOH4zCcduJuLSvVSoNbofr01crdDJxx4dG9ojwNseFiq2QehbaZ/fT+N3n4iIqJzy1RpceZCGY5GJOH4zEWfvPEROvu4wAS97M3Rs6ICOPvbo0MAe9hYqiaql4jAAERERPUNpM7UKOVmq0KmhgziOp6ED3G1MJaqWyoIBiIiI6AkZOfm4lZCBmwkZuBmficj4DJy9+xCJGboztaxMjNDBxx4dfRzQqaE9fBwtJFuzh8qPAYiIiOocQRAQk5pdEHIycDMhE7cSxcATm5Zd7GsKZ2p1KujWauZmDYWcgcdQMQAREVGtlZ2nxu3ETG1rzq1EsWXnVkImskpZW8fBQgUfR3P4OFnAx9ECzdys0Lp+0ZlaZLgYgIiIyKAJgoDEjFwx5BSEm8L79x4+QklznY3kMnjam8HH0QI+ThZo4FAQeBwsYG1Wt6aE10UMQEREZDBSsnJx7u5DXI/L0Iacm/EZSHtqQPKTrEyM0NDJAg0cxdacwpad+nZmMK6li/zRszEAERFRjRWbmo3Td5Jx+nYSztx+iIi49GL3k8kAD1szNHA0Lwg5j4OOvbmSg5OpCAYgIiKqEQRBwN2kLJy+nVwQepKLXDICAHwczdHc3fpx0HEyh5e9OUz0eIFoqv0YgKjMunXrhlatWmHFihUAAC8vL0yfPh3Tp08v8TUymQy7du3CgAEDKvXe+joOEdUcGo2AiLh0ncDz9EVB5TKgqZsVArzsEeBti3ZednDggoKkBwxAdUTfvn2Rl5eHffv2FXnun3/+QZcuXXDx4kW0bNmyzMc8c+YMzM3N9VkmFixYgN27d+PChQs622NiYmBra6vX9yKi6pWn1uDy/VScuS2GnTN3kouM3VEq5PD3sEZ7LzsEeNuhradtnbtGFVUPBqA6YuLEiRg8eDDu3buHevXq6Tz3/fffo127duUKPwDg6OiozxJL5eLiUm3vRUT68ShXjX+jH4otPLeT8W9UCh7l6U49N1Mq0NbTFgEFgcffw4ZdWVQtOPy9jnjllVfg6OiI9evX62zPyMjA9u3bMWDAAIwcORLu7u4wMzNDixYtsHnz5lKP6eXlpe0OA4AbN26gS5cuMDExQdOmTXHgwIEir/noo4/QuHFjmJmZoUGDBpg7dy7y8vIAAOvXr8fChQtx8eJFyGQyyGQybb0ymQy7d+/WHufy5ct44YUXYGpqCnt7e7z55pvIyMjQPj9+/HgMGDAAn3/+OVxdXWFvb49JkyZp34uI9C/1UR7+Co9DyB9hGPSfY2i5cD9GfXcKKw7ewPGbSXiUp4atmTFeauqMOX2aYM/kTrg0/2X8ODEQU15shMAG9gw/VG3YAqQPggDkFR2oV+WMzcSpD2VgZGSEsWPHYv369Zg9e7Z2RsT27duhVqvx6quvYvv27fjoo49gZWWFvXv3YsyYMfDx8UFAQMAzj6/RaDBo0CA4Ozvj1KlTSE1NLXZskKWlJdavXw83NzdcvnwZb7zxBiwtLfHhhx9i+PDhuHLlCvbt24eDBw8CAKytrYscIzMzE0FBQejQoQPOnDmD+Ph4vP7665g8ebJOwDt06BBcXV1x6NAhREZGYvjw4WjVqhXeeOONMn3PiKh0giDgfFQKfr34AKduJyM8Nq3ImjsuViYI8LbT3ho6WkDO1ZOpBmAA0oe8LGCJW/W/78cPAGXZx+C89tprWLZsGY4cOYJu3boBELu/Bg8eDE9PT8yYMUO775QpU7B//35s27atTAHo4MGDCA8Px/79++HmJn4vlixZgl69eunsN2fOHO19Ly8vzJgxA1u2bMGHH34IU1NTWFhYwMjIqNQur02bNiE7Oxs//PCDdgzS119/jb59+2Lp0qVwdnYGANja2uLrr7+GQqGAn58f+vTpg9DQUAYgokpKSM/Brn/vYdvZe4iMz9B5ztvBXNudFeBth3q2ppyCTjUSA1Ad4ufnh44dO2LdunXo1q0bIiMj8c8//2DRokVQq9VYsmQJtm3bhvv37yM3Nxc5OTkwMzMr07HDwsLg4eGhDT8A0KFDhyL7bd26FStXrsTNmzeRkZGB/Px8WFlZles8wsLC4O/vrzMAu1OnTtBoNIiIiNAGoGbNmkGheNyc7urqisuXL5frvYhIlK/W4HBEAraejcah8Hjka8SmHlNjBXq1cMGLfs5o720LJ0sTiSslKhsGIH0wNhNbY6R433KaOHEipkyZgtWrV+P777+Hj48PunbtiqVLl+Krr77CihUr0KJFC5ibm2P69OnIzc3VW7knTpzA6NGjsXDhQgQFBcHa2hpbtmzBF198obf3eJKxse7MEZlMBo1GUyXvRVRb3UzIwPaz9/Dz+Xs6U9Rb17fBsHYeeKWlK2dpkUFiANIHmaxcXVFSGjZsGKZNm4ZNmzbhhx9+wDvvvAOZTIZjx46hf//+ePXVVwGIY3quX7+Opk2blum4TZo0QXR0NGJiYuDq6goAOHnypM4+x48fh6enJ2bPnq3ddvfuXZ19lEol1OqSL1BY+F7r169HZmamthXo2LFjkMvl8PX1LVO9RFSyzJx87L0Ug21no3H27kPtdntzJQa1ccfQdh5o7GwpYYVElccAVMdYWFhg+PDhmDVrFtLS0jB+/HgAQKNGjbBjxw4cP34ctra2WL58OeLi4socgHr06IHGjRtj3LhxWLZsGdLS0nSCTuF7REVFYcuWLWjfvj327t2LXbt26ezj5eWF27dv48KFC6hXrx4sLS2hUukuejZ69GjMnz8f48aNw4IFC5CQkIApU6ZgzJgx2u4vIiofcUDzQ2w9E43fLsVor5QulwHdfZ0wtJ0HXvBzgtKIk4epduBvch00ceJEPHz4EEFBQdoxO3PmzEGbNm0QFBSEbt26wcXFpVyrLsvlcuzatQuPHj1CQEAAXn/9dXz66ac6+/Tr1w/vvfceJk+ejFatWuH48eOYO3euzj6DBw9Gz5490b17dzg6OhY7Fd/MzAz79+9HcnIy2rdvjyFDhuDFF1/E119/Xf5vBlEdF5+ejTVHbuLF5Ucw+JsT2Hb2HrJy1fB2MMeHPX1xYtaLWDu+PXo2d2H4oVpFJghPT1qktLQ0WFtbIzU1tcgA3ezsbNy+fRve3t4wMeFgv9qEP1uqK/LUGhwKj8e2s/dwKCIe6icGNPdp6Yph7TzQ3suWs7fI4JT2+f00doEREdURkfEZ2H42Gj+fv4/EjMcDmtsUDmj2d4OFih8LVDfwN52IqBbLyMnH3ksPsO3sPZx7YkCzg4USg9rUw7B29dDQiQOaqe5hACIiqmUEQcDZuw+x7Uw09l5+PKBZIZehu6+jdkCzsYJjeqjuYgAiIqoFBEFAWEw69l+Nxa8XH+BWYqb2uQYO5hjazgOD27jDyYrj24gABqAK49jx2oc/UzI0Go04dX3flVjsvxaL6ORH2ufMlAr0aeGK4e090NaTA5qJnsYAVE6FqwtnZWXB1NRU4mpIn7KyxAvaPr2CNFFNkpuvwYlbSdh/NRZ/Xo3TGcysMpKjS2NHBDVzQc/mLhzQTFQK/usoJ4VCARsbG8THxwMQ16Th/6wMmyAIyMrKQnx8PGxsbHSuH0ZUE2Tl5uNIRAL2X41FaHg80rPztc9ZmhjhRT8nBDVzQVdfR5gp+WedqCz4L6UCCq9UXhiCqHawsbEp9Sr0RNUpJSsXB8Pisf9qLP6+noCc/MfXsXOwUOHlZs4IauaCDg3suUAhUQVIHoBWr16NZcuWITY2Fv7+/li1ahUCAgKK3TcvLw8hISHYsGED7t+/D19fXyxduhQ9e/as8DErQiaTwdXVFU5OTsjLy9PbcUk6xsbGbPkhycWmZuPPa7HYfzUWJ28laxcoBID6dmYIKgg9revbQiFnyzNRZUgagLZu3Yrg4GCsWbMGgYGBWLFiBYKCghAREQEnJ6ci+8+ZMwcbN27Ed999Bz8/P+zfvx8DBw7E8ePH0bp16wodszIUCgU/NImoUm4lZGD/1TjsvxqLC9EpOs/5uVgiqJkLgpq5oImrJbvbifRI0kthBAYGon379tprOGk0Gnh4eGDKlCmYOXNmkf3d3Nwwe/ZsTJo0Sbtt8ODBMDU1xcaNGyt0zOKUZyltIqLyEAQBVx+kYf9VsaXnelyG9jmZDGhT31bb0uNpby5hpUSGxyAuhZGbm4tz585h1qxZ2m1yuRw9evTAiRMnin1NTk5OkWs0mZqa4ujRoxU+ZuFxc3Iez6RIS0ur0DkRERVHrRFw9k6ytqXnfsrj6epGchk6+NgjqJkLXm7qzHV6iKqJZAEoMTERarUazs7OOtudnZ0RHh5e7GuCgoKwfPlydOnSBT4+PggNDcXOnTuhVqsrfEwACAkJwcKFCyt5RkRUF2k0Ah5m5SIhIweJ6blIzMhBYkYOEtJzxG0Zubh6PxVJmbna15gaK9C1sSOCmjvjBV9nWJtx6QWi6ib5IOjy+Oqrr/DGG2/Az88PMpkMPj4+mDBhAtatW1ep486aNQvBwcHax2lpafDw8KhsuURkoIoLNQnpBcGmINQUPk7OzNUZrFwSa1NjvNhEnK7epZEjTJUcP0gkJckCkIODAxQKBeLi4nS2x8XFlTgV2dHREbt370Z2djaSkpLg5uaGmTNnokGDBhU+JgCoVCqoVKpKnhERGYKkjByExaTrJdQ8ydbMGA4WKjhYqOBoKX51sFTCwUKF+nZmaOtpy2tvEdUgkgUgpVKJtm3bIjQ0FAMGDAAgDlgODQ3F5MmTS32tiYkJ3N3dkZeXh59//hnDhg2r9DGJqHbLzdfgf0dvYVVoJB7lqcv0msJQow00BaHG0UIFB0uV+NVCBXsLJcMNkYGRtAssODgY48aNQ7t27RAQEIAVK1YgMzMTEyZMAACMHTsW7u7uCAkJAQCcOnUK9+/fR6tWrXD//n0sWLAAGo0GH374YZmPSUR1z7HIRMz95QpuJYgXCK1vZ4Z6tqYlhhpHSxXszBlqiGozSQPQ8OHDkZCQgHnz5iE2NhatWrXCvn37tIOYo6KiIJc//gOUnZ2NOXPm4NatW7CwsEDv3r3x448/wsbGpszHJKK6IzY1G5/svYbfLsUAABwslPi4dxMMbO3ONXWI6jhJ1wGqqbgOEJFhy1NrsOH4HXx54Doyc9WQy4CxHbzw3kuNYW3KGVdEtZVBrANERFQVTt9OxtzdVxARlw4AaF3fBov7N0dzd2uJKyOimoQBiIhqhYT0HIT8EYad5+8DEAcwz+zlh6FtPSDndbOI6CkMQERk0NQaARtP3sXnf0YgPTsfMhkwMqA+PnjZF7bmSqnLI6IaigGIiAzW+aiHmLv7Cq4+EC9f08LdGosHNEcrDxtpCyOiGo8BiIgMTnJmLj7bF44tZ6IBAFYmRvigpx9GBdSHgt1dRFQGDEBEZDA0GgFbzkTjs/3hSMnKAwAMaVsPM3v5wcGCq7kTUdkxABGRQbh8LxVzfrmCi9EpAAA/F0t8MqA52nnZSVsYERkkBiAiqtFSs/Lw+Z8R2HjqLgQBsFAZIfilxhjbwRNGXKmZiCqIAYiIaiSNRsDP5+/h//4IR1JmLgCgfys3zO7dBE5WJhJXR0SGjgGIiGqcsJg0zN19BWfvPgQANHSywKL+zdDRx0HiyoiotmAAIqIaIz07D18euIENJ+5ArRFgplRg2ouNMKGTN5RG7O4iIv1hACIiyQmCgD0XH+CTvWFISM8BAPRu4YI5fZrCzcZU4uqIqDZiACIiSUXEpmPBnqs4cSsJAODtYI6F/ZqhS2NHiSsjotqMAYiIqlV2nhqnbyfjyPUEHLmegMj4DACAykiOKS80xBtdGkBlpJC4SiKq7RiAiKhKCYKA24mZ2sBz8lYSsvM02uflMuClps6Y06cpPOzMJKyUiOoSBiAi0ruMnHycuJmEI9fjceR6AqKTH+k872Jlgq6NHdHV1xGdfBxgbWYsUaVEVFcxABFRpQmCgPDYdLGVJyIBZ+8mI08taJ9XKuRo720rhp7GTmjsbAGZjNfsIiLpMAARUYWkZOXinxuJ+Lugayu+YPZWIU97M3QraOV5roE9zJT8c0NENQf/IhFRmag1Ai7dS9GO5bkYnQLN40YemBor0MHHvqCVxxFeDubSFUtE9AwMQERUovj0bPx9PRFHrifgnxsJ2iuwF/J1tkRXX0d0aeSIdl62MDHm7C0iMgwMQESkpdEIOHMnGYcLxvJci0nTed7SxAjPN3JA18aO6NLYEa7WXKSQiAwTAxARIT49GzvO3cOW09GISs7Sea5lPWttt1YrDxtegZ2IagUGIKI6SqMRcDQyEZtPR+HAtTjkFwzosVQZ4cUmTujm64TOjRzgYKGSuFIiIv1jACKqY+LTsrH93D1sPh2Few8fr8/Tpr4NRgbUR5+WrpyxRUS1Hv/KEdUBao2Af24kYPPpKBwMi4e6sLXHxAiD29TDiAAP+LlYSVwlEVH1YQAiqsViU7Ox/Ww0tpyJxv2Ux6097TxtMTKgPnq3cIWpkjO3iKjuYQAiqmXUGgF/X0/AptNR+Cv8cWuPtakxBrVxx8iA+mjsbClxlURE0mIAIqolYlIfYduZe9h6JgoPUrO12wO87DAy0AO9mrtynR4iogIMQEQGLF+twZHr4tiev8LjtSsz25gZY3CbehgZ4IGGTmztISJ6GgMQkQG6n/II285EY9vZaMQ80doT6G2HUYH1EdTMha09RESlYAAiMhD5ag0ORYitPYcjHrf22JoZY0jbehgRUB8+jhbSFklEZCAYgIhquHsPs7DtTDS2no1GXNrjK6539LHHiID6CGrmDJURW3uIiMqDAYioBsrN1+BgWBy2nInGPzcSIBS09tiZKzG0oLXHm1dbJyKqMAYgohokMj4dW89EY+f5+0jKzNVu79TQHiMD6uOlpmztISLSBwYgIoll5eZj76UYbD0TjbN3H2q3O1mqMLRdPQxr5wFPe7b2EBHpEwMQkQQEQcDl+6nYciYaey48QEZOPgBAIZfhBT8njGjvga6NHXnldSKiKsIARFSNUrPysPvCfWw5E42wmDTtdk97Mwxv74EhberBycpEwgqJiOoGBiCiKqbRCDh5OwnbzkTj9yuxyM3XAACURnL0bu6C4e3rI9DbDnJogPQY4M4d4OEdIPm2+DU9FlCaASY2gKktYGpTcL+Er8amgEwmwZkSERkOBiCiKhKflo3t5+5h29lo3E3KAgCY4xFedMzCUJ98dLBJg2nmQeDEHeD3O0BKFKDOLfWYZaJQikGptJCk/fpUoDI2rfz7ExEZAAYgIj3Kz8vDiYtXceLsOSTei4AH4hAsi4eXKgE+xgmwyE8B0gFcKOEAciPApj5g6wXYeotfrdyAvCzgUQqQnaL79dHDJ7alAoJaDFEZceKtvBQqMRTZNQCcmjy+OTYBzO3LfzwiohqKAYioPAQByM0AHt4Vu6ceit1Uj+Ju4lH8TVhkP8DzyMfzQNF/XfkFX03tCgKOF2Dn/fi+rRdg5Q7IKzjNXRCAnPSnQtLD4oPT0wEqOxUQNIA6B8iIFW9Rx3WPb+6kG4qcmgKOfoCJVcXqre3U+UDKXSDpJpD/CHBpKf6M2T1JVCMwAJHhys0EbvwJxIcDmnxAkyd+6GjyxMfqp75q7xduzy96v9h91bqvK4ZpwQ0A8qFAmsoVJk4+MHP20W3NsfUETKyr5vshk4lhxMRKbEUqD40GyE0XQ1FWEpAUCcRfA+LDxFvKXSAzHrgdD9w+ovtaq3oFgcjvcShy9AWUdWDqvkYDpN0Xv1/JN4GkW4/vP7xT9PfF1A5waw24twHc2ohfLV0kKZ2orpMJQuEas1QoLS0N1tbWSE1NhZUV/3dbo+RlA5EHgSs/A9f3iV1DEshSWOOW2gG31Y6IEpwQLTjD2r0xnmvTBp3atIRSqZSkriqTkwEkRAAJBYEo/poYPNMflPACmRj4nm4tcmgEGKmqs/LKEwQgI74g4ESKLTpJkUDyLfGWn13ya41MADsfQGEsfs+KG+Nl5a4bitxai+OxiKjcyvP5zQBUDAagGiY/F7h1WAw9Eb8DOY+nj8PGE2jQFTAyFT9k5ApAblxw30i8KYwLthU8Lvb5J76WsG9aLrD49+v4PSIVmQXtPe42phjWzgND2tWDu00dHED86KEYjJ5sLYoPA7ISi99fpgDsfXRDkVNTwMJRDAsKFSCXaO2jRw8Lws3NJ1p0IsVWndz0kl8nNxJb+Ox9APuG4vipwvuWbo/PJz8HiLsK3D8HPPgXuH8eSAgHUMyfYPuGj1uI3NoAri05QJ2oDBiAKokBqAZQ5wN3/hFDT9iv4jiVQlbuQLOBQPNB4odDNYypOB/1EFM2/Yv7KY9gJJchqJkLhrf3QKeGDlDIOaajiIyEJ1qLnrjlpD77tXJjMQwZKQtCkbKExyrxplA9vm+kKts+j1IKAs4TgedRcilFyQAbj4KAUxBu7H3Em3V9MTBXRE4GEHMReHBeDEb3z4vdjUXeXgE4N9UNRU5NK/6+RLUUA1AlMQBJRKMRB95e2Qlc+0W3FcHc6XHoqRdQba0EGo2A7/65hWX7I5CvEeBpb4bVo9qguXsVjeOpzQRBXOeosPussCstIQLIy5S6OpGla9FWHLuCcVzG1bRAZWaS2EL0ZCjKjC+6n5Gp2DL0ZCiyayBdCxpRDWBQAWj16tVYtmwZYmNj4e/vj1WrViEgIKDE/VesWIFvvvkGUVFRcHBwwJAhQxASEgITE/GP04IFC7Bw4UKd1/j6+iI8PLzMNTEAVSNBAO6dFVt6ru0WPyALmdoBTfuLocezU8VnR1VQcmYu3t92AYciEgAAr7R0RcigFrA0Ma7WOmo9QRAHmatzxG6i/BxxXI06V/ya/8T2EvfJLf41OvsXPM7LFgdo2zcE7Bs8Djl2DQCVhdTfjaIEQRxoff98QSg6LwakJ7uCC6msAffWgKu/2P1m7gBYOAHmjuJ/IkxtGZCoVivP57ek7adbt25FcHAw1qxZg8DAQKxYsQJBQUGIiIiAk5NTkf03bdqEmTNnYt26dejYsSOuX7+O8ePHQyaTYfny5dr9mjVrhoMHD2ofGxmxmbhGEQSx2f/Kz8DV3UBq1OPnVNZAk75A84GAd1dx/I0ETt9OxtTN/yI2LRsqIznm922GkQEekHEKs/7JZAXdW0pAZSl1NTWPTAZY1xNvTfuJ2zQasQvvyVAUe0nsYrx1WLwVeyyFGIrMHR/fLJwKthUEJYsnnjO0AetE5SBpMli+fDneeOMNTJgwAQCwZs0a7N27F+vWrcPMmTOL7H/8+HF06tQJo0aNAgB4eXlh5MiROHXqlM5+RkZGcHHh1NIaJ+5aQejZKc6eKaS0AHx7iy09Pi9I+kdXoxHwn8ORWH7gOjQC0MDRHKtHtUETV7YEUg0il4sz6hwaAf7DxW3qPLFL8f558WtGPJCZKHafZSaIg7wFdfkWyVRZ6wYibWvSU48tnBheyeBIFoByc3Nx7tw5zJo1S7tNLpejR48eOHHiRLGv6dixIzZu3IjTp08jICAAt27dwu+//44xY8bo7Hfjxg24ubnBxMQEHTp0QEhICOrXL3ldlJycHOTk5Ggfp6UV07RMFZN4QxzTc3VnwYyXAkamQOMgMfQ0erlGzHBJSM9B8LYL+OeGOPZoUBt3LO7fHOYqtiCSAVAYi11frv7FP5+fK67xVBiIMhKeul/4OFG8r8kXW5RyUsVB4s/i2kpsoWrSH3BoqNdTI6oKkv1lT0xMhFqthrOzs852Z2fnEsfrjBo1ComJiejcuTMEQUB+fj7efvttfPzxx9p9AgMDsX79evj6+iImJgYLFy7E888/jytXrsDSsvj/oYSEhBQZN0SV8PDO49ATe/nxdoUSaPiSGHoa96xR4y2ORyZi2tYLSEjPgamxAosHNMeQtvWkLotIf4yUgJWreHsWjUaceZlZEIyebk16OjDlZgAxF8Rb6CJxhlrT/kCTfuKSB+w6phpIskHQDx48gLu7O44fP44OHTpot3/44Yc4cuRIkW4tADh8+DBGjBiBTz75BIGBgYiMjMS0adPwxhtvYO7cucW+T0pKCjw9PbF8+XJMnDix2H2KawHy8PDgIOjy0GiA638Ax1YC0Scfb5cbAQ26i6HHt3eNW+BNrRHwVegNrPrrBgQB8HW2xNejWqORM5vzicosIx4I3wuE7QFu/627ArZ9QzEINe0vtk4xDFEVMohB0A4ODlAoFIiL0+2LjouLK3H8zty5czFmzBi8/vrrAIAWLVogMzMTb775JmbPng15MbMbbGxs0LhxY0RGltyEq1KpoFJxsF+FqPOAy9uBoyuAxAhxm0wOeD0vhp4m/QAzO0lLLElcWjambv4Xp26L67+MaO+B+X2bwVRZvbPNiAyehRPQboJ4y0oWV2m/9gtw8y+x++zocvFmU78gDA0A3NvW3BlpggCk3itYhqBg4Uq5EeDS4nE3o51Pza2fykSyAKRUKtG2bVuEhoZiwIABAACNRoPQ0FBMnjy52NdkZWUVCTkKhfhhVVJDVkZGBm7evFlknBBVUm4mcP4H4PjXQNo9cZvKCmj/OhD4Vo2/vtHhiHgEb7uI5MxcmCsVWDKoBfq3cpe6LCLDZ2YHtBol3rLTxOv1XfsFuHEASIkCTnwt3izdxBmfTfsB9TtU+zIXOrKSH6/OXRh6ilt76dahx/eNzQGX5uJFbl39xTWZHJuIXY1kECQd3RkcHIxx48ahXbt2CAgIwIoVK5CZmamdFTZ27Fi4u7sjJCQEANC3b18sX74crVu31naBzZ07F3379tUGoRkzZqBv377w9PTEgwcPMH/+fCgUCowcOVKy86xVspKB098Cp/77eOVcC2fguXfF//1V1YU+9SRPrcHyA9fxzeGbAICmrlb4elRrNHCsOeORiGoNEyugxRDxlpspXsfv2h7g+n7xOnKn/yvezB0Bv1fEMOT1fNUuf5H3SBybWLjI5P1z4pICT5MbAc7NHy8yKWjEpQZiLgKxV8TFO6NPiTfta4zFiwK7FAQiV3/xGDVovCM9JmkAGj58OBISEjBv3jzExsaiVatW2Ldvn3ZgdFRUlE6Lz5w5cyCTyTBnzhzcv38fjo6O6Nu3Lz799FPtPvfu3cPIkSORlJQER0dHdO7cGSdPnoSjo2O1n1+tknoPOLEaOLf+8QVIbb2BTtMA/5HVt0puJTxIeYQpm//FubsPAQBjnvPE7D5NYGLMLi+iKqc0F8cBNe0vLkZ567A4Zih8rzig+tz34s3UVhwv2LQ/0KBb5ZbF0KiBxOuPW3XunxOvx/bkGKVCdj5it1zhzaV5ybNTNWqxay/mkjjwO/aSeD87RQxXsZeBC4U7y8RVxV1aPg5FLv6AuX3Fz4v0QvKVoGsirgT9hIQI4NhXwKWtj/9ouLQEOr8n/oGSstm6HA5ei8OMHReRkpUHS5URlg5pid4tyjAbhoiqljpPHDgdtgcI+033EjgqK3G5jCb9gIY9AKVZycfRrph97nHrzoN/xRlqTzN3BNzbFYSdNoBb68qPVRQEIDW6IBRdfByK0h8Uv7+V+1OhqKW42CUHiVeKQV0KoyZiAIJ4eYqjX4r/Oyu8WrXX80Dn6YDPiwbzjzQ3X4Ol+8Kx9uhtAEDLetb4emQb1Lcv5Q8pEUlDowbuHi8IQ7/qXhrH2Axo9JIYhhoHif8hu3/+cTfWg/PFL/BobC4GHPc2j1t3qjNoZCQAsRfFMFQYiorrcgPE1q/CUOTYRFyywNJVHFNpYmMwf3eLEARx+ET6AyA9Fkh7IP5sPTsB3s/r9a0YgCqpzgYgQRBnbRz9UrwSeyG/V4BO0wGP9pKVVhHRyVmYvOk8Lt4Tr0A+sbM3PurpB6URZ24Q1XgaDXD/rDiA+toe3UvmyI0BTV7R18gUgHMz3a4sR9+a11KdnQbEXdENRQlhxXfNFTIyEYNQYSCycNF9XPhVZVm9QSk3Swwz6TFAWswT9wvCTmHoUecWfW3nYKDHfL2WwwBUSXUuAGnU4h+Zo1+K/xgBcQBgy+HiGB9HX2nrq4A/Lsfgw58vIT07H9amxvh8qD9eaur87BcSUc0jCOJYm2t7xNahwpWpbb2fGrfTovRusposP0e8hElhKEqKBNLjxDCRnVL24xibA5bORYPR08HpWQOzNWpxfaf0B7rB5skWnPQYIDu17LWZOYh1WBXU0igIaPJK2V9fBgxAlVRnAlBeNnBxM3B85eNrcxmbAW3HAx0mic3EBiY7T40lv4fhhxN3AQBt6ttg1ag2cLeR/lIbRKQHgiCuNm9iXWPXGNO7vEdi91567OMQov36xC2nHGFEaVkQjArCkcriceBKjxHfT9CU7VjGZgXBxu3x8bT33R6/TzVc59EgFkIkCWWnAWfXASf/87jP3NQWCHhLXMPHQP+o3E7MxORN53H1gXgtt7e7+uD9lxvDWMEuL6JaQyYD7LylrqJ6GZsCtl7irTS5mY/DUEZsMUEpRgw5ueniLSkdSLpR8vFkCnGZE+1YJFfd+4WPVVYGOT6JAaguyYgHTn4DnFn7+H8KVu5Ah8lAm7EGvVbFLxfu4+Odl5GZq4aduRLLh/mjm6+T1GUREVUfpbk45d7ep/T9ctKfaO0pCEa5GeKK3pZuj0OOuWPNGz+lRwxAdcHDO+I1uv7dCKgLrnnm0Fgc2NxiqEGvXJqdp8bCX69i8+loAECAtx1WjmgNF+uavy4REZEkVJbizaGh1JVIigGoNsvNAva+L67hI6jFbe5txZH3vr0N/jo2UUlZePPHswiPTYdMBkzp3hBTX2wEI3Z5ERHRMzAA1Wahi4CLm8T7Pi+Kixd6dTbIvtqn/XMjAZM3/YvUR3lwsFBixfDW6NzIQeqyiIjIQDAA1VbRZ4BTa8T7w3/S+1RDqQiCgG//voWl+8KhEQB/Dxv899W27PIiIqJyYQCqjfJzgD2TAQhAyxG1Jvxk5ebjo58v49eL4tLyw9rVw6L+zXktLyIiKjcGoNron+VAQri46FTPEKmr0Yvo5Cy88YM43sdILsP8vk3x6nOekNWC7jwiIqp+DEC1TXwY8M8X4v3enxnsmj5POnojEZM3n0dKljje5z+j2yLA2/DPi4iIpMMAVJto1MAvk8Vr5DTuBTQbJHVFlSIIAr775xb+74+C8T71rLFmTFu4WnNVZyIiqhwGoNrk1H/FiwcqLYE+Xxj0bK9HuWp89PMl7CkY7zO0bT0sHsDxPkREpB8MQLXFwzvAX4vF+y8vAqzdJS2nMqKTs/Dmj+cQFpMGI7kM8/o2xRiO9yEiIj1iAKoNBAH4dTqQlwV4dgLajJe6ogp7erzP6lFtENjAXuqyiIiolmEAqg0ubAJuHQKMTIC+Kw1yhWdBEPC/f24j5I8waASgZT1rrHm1Ldx4FXciIqoCDECGLiMe2P+xeL/bTIO8tsujXDVm7ryEXy6I432GtK2HTzjeh4iIqhADkKH7/QMgOwVwaQl0mCJ1NeUWnZyFt348h2sF433mvtIUYztwvA8REVUtBiBDFvYbcG03IFMA/b8GFIb14zwWmYjJm87jYVYe7M2V+M9ojvchIqLqYVifmPTYoxTxSu8A0Gkq4OovaTnlIQgC1h69jSW/c7wPERFJgwHIUB2YB2TEAnY+QNePpK6mzB7lqjFr5yXsLhjvM7hNPXw6kON9iIioejEAGaLbfwPnN4j3+60CjA2j5eTeQ3G8z9UHaVDIZZjTpwnGd/TieB8iIqp2DECGJjcL2DNVvN/uNcCrk7T1lNHxyERMemK8z+rRbfAcx/sQEZFEGIAMzeEQ4OFtwNIN6LFQ6mqeSRAErDt2B0t+D4NaI6CFu3g9L3eO9yEiIgkxABmS++eBE1+L919ZDphYSVvPM2TnqTFr52Xs+vc+AGBQa3csGdSC432IiEhyDECGQp0H7JkCCBqg+WDAt5fUFZXq6fE+s3s3wYROHO9DREQ1AwOQoTj2FRB3BTC1A3oulbqaUh2/mYjJm/5FcmYu7MyV+HpUa3T0cZC6LCIiIi0GIEOQcB04UhB6ev4fYOEobT2l2HD8Dhb9dg1qjYDm7lZY82pb1LM1k7osIiIiHQxANZ1GA/w6FVDnAg1fAloOk7qiEh0Kj8f8PVcBAANbuyOE432IiKiGYgCq6c6uBaJOAEoLceBzDR1Dk5qVh5k7LwEAxnbwxMJ+zTjeh4iIaiy51AVQKVKigYMLxPsvzgds6ktaTmkW/nYVcWk5aOBgjlm9mjD8EBFRjcYAVFMJArA3GMjNADwCgfavS11RiQ5ci8PO8/chlwHLhvrDVMluLyIiqtnKHYC8vLywaNEiREVFVUU9VOjyDuDGn4BCKV7uQl4zs+rDzFx8vOsyAOCN5xugraetxBURERE9W7k/VadPn46dO3eiQYMGeOmll7Blyxbk5ORURW11V2Yi8MeH4v0uHwKOvtLWU4oFv15FQnoOfBzN8d5LjaUuh4iIqEwqFIAuXLiA06dPo0mTJpgyZQpcXV0xefJknD9/vipqrHv2zQQeJQNOzYBO06SupkT7rsTglwsPIJcBXwxrxRlfRERkMCrcr9KmTRusXLkSDx48wPz58/G///0P7du3R6tWrbBu3ToIgqDPOuuO6/uBy9sBmRzovwowUkpdUbGSMnIwe9cVAMDbXX3QysNG2oKIiIjKocLT4PPy8rBr1y58//33OHDgAJ577jlMnDgR9+7dw8cff4yDBw9i06ZN+qy19stOA357T7z/3LuAe1tp6ynFvD1XkZSZi8bOFpjWo5HU5RAREZVLuQPQ+fPn8f3332Pz5s2Qy+UYO3YsvvzyS/j5+Wn3GThwINq3b6/XQuuE0IVA2n3A1gvoPlvqakr026UH2HspBgq5DF8MbQWVEbu+iIjIsJQ7ALVv3x4vvfQSvvnmGwwYMADGxsZF9vH29saIESP0UmCdcfc4cOZ/4v2+KwFlzbx8REJ6DubuFru+JnXzQYt61hJXREREVH7lDkC3bt2Cp6dnqfuYm5vj+++/r3BRdU5etnildwBoPQZo0FXaekogCALm7L6Mh1l58HOxxOQX2PVFRESGqdyDoOPj43Hq1Kki20+dOoWzZ8/qpag65+/PgKRIwMIFePkTqasp0Z6LD7D/ahyM5DJ8McwfSqOauTYRERHRs5T7E2zSpEmIjo4usv3+/fuYNGmSXoqqU2IuAUdXiPf7fA6Y2khZTYni07Ix7xfxQqdTXmiEZm7s+iIiIsNV7gB07do1tGnTpsj21q1b49q1a3opqs5Q5wN7JgOCGmjSD2jSV+qKiiUIAj7edRmpj/LQzM0K73b3kbokIiKiSil3AFKpVIiLiyuyPSYmBkZGvLh8uZxcDcRcBEysgd6fS11NiXb9ex8Hw+JhrBC7vowV7PoiIiLDVu5PspdffhmzZs1CamqqdltKSgo+/vhjvPTSS+UuYPXq1fDy8oKJiQkCAwNx+vTpUvdfsWIFfH19YWpqCg8PD7z33nvIzs6u1DElkXQTOLREvB+0BLB0lraeEsSmZmPBHrHra3qPxvBzsZK4IiIiosordwD6/PPPER0dDU9PT3Tv3h3du3eHt7c3YmNj8cUXX5TrWFu3bkVwcDDmz5+P8+fPw9/fH0FBQYiPjy92/02bNmHmzJmYP38+wsLCsHbtWmzduhUff/xxhY8pCUEAfp0G5GcDDboBrUZLXVGxBEHArJ2XkJadj5b1rPFWlwZSl0RERKQXMqEC16zIzMzETz/9hIsXL8LU1BQtW7bEyJEji10TqDSBgYFo3749vv76awCARqOBh4cHpkyZgpkzZxbZf/LkyQgLC0NoaKh22/vvv49Tp07h6NGjFTpmcdLS0mBtbY3U1FRYWVVBi8e59WIAMjYD3jkO2Hnr/z30YNvZaHy44xKUCjn2Tu2MRs6WUpdERERUovJ8fldo0I65uTnefPPNChVXKDc3F+fOncOsWbO02+RyOXr06IETJ04U+5qOHTti48aNOH36NAICAnDr1i38/vvvGDNmTIWPCQA5OTk6V7RPS0ur1LmVKu0B8Odc8f4Lc2ps+HmQ8giLfxUHtQe/3Jjhh4iIapUKj1q+du0aoqKikJubq7O9X79+ZXp9YmIi1Go1nJ11x744OzsjPDy82NeMGjUKiYmJ6Ny5MwRBQH5+Pt5++21tF1hFjgkAISEhWLhwYZnqrhRBAPbOAHLSxOt8Bb5d9e9ZAYIg4KOfLyE9Jx+t69vgjefZ9UVERLVLhVaCHjhwIC5fvgyZTKa96rtMJgMAqNVq/Vb4hMOHD2PJkiX4z3/+g8DAQERGRmLatGlYvHgx5s6dW+Hjzpo1C8HBwdrHaWlp8PDw0EfJuq7tBiL2AnJjoN/XgLxmXkNry5lo/HMjESojOT4f6g+FXCZ1SURERHpV7kHQ06ZNg7e3N+Lj42FmZoarV6/i77//Rrt27XD48OEyH8fBwQEKhaLIlPq4uDi4uLgU+5q5c+dizJgxeP3119GiRQsMHDgQS5YsQUhICDQaTYWOCYhT+62srHRuVSIhAoAMeD4YcG5aNe9RSfceZuGT38Surw+CfOHjaCFxRURERPpX7gB04sQJLFq0CA4ODpDL5ZDL5ejcuTNCQkIwderUMh9HqVSibdu2OgOaNRoNQkND0aFDh2Jfk5WVBblct2SFQmxFEQShQsesVt1mAq+HAs+/L3UlxdJoBHy44xIyc9Vo52mLCZ1q5vgkIiKiyip3F5harYalpTgg1sHBAQ8ePICvry88PT0RERFRrmMFBwdj3LhxaNeuHQICArBixQpkZmZiwoQJAICxY8fC3d0dISEhAIC+ffti+fLlaN26tbYLbO7cuejbt682CD3rmJKr11bqCkr00+koHL+ZBBNjOZax64uIiGqxcgeg5s2b4+LFi/D29kZgYCA+++wzKJVKfPvtt2jQoHyDZYcPH46EhATMmzcPsbGxaNWqFfbt26cdxBwVFaXT4jNnzhzIZDLMmTMH9+/fh6OjI/r27YtPP/20zMek4kUlZSHk9zAAwEc9/eDtYC5xRURERFWn3OsA7d+/H5mZmRg0aBAiIyPxyiuv4Pr167C3t8fWrVvxwgsvVFWt1abK1wGqYTQaASO/O4lTt5MR4G2HLW88Bzlbf4iIyMBU6TpAQUFB2vsNGzZEeHg4kpOTYWtrq50JRoblhxN3cOp2MsyUCnw+xJ/hh4iIar1yDYLOy8uDkZERrly5orPdzs6O4cdA3UnMxP/tE9dImtXLD/XtzSSuiIiIqOqVKwAZGxujfv36VbrWD1UftUbAjO0XkZ2nQUcfe4wO9JS6JCIiompR7mnws2fPxscff4zk5OSqqIeq0ffHbuPs3YcwVyqwdHBLdn0REVGdUe4xQF9//TUiIyPh5uYGT09PmJvrzhY6f/683oqjqnMzIQPL9ovLFszu0xQeduz6IiKiuqPcAWjAgAFVUAZVp8Kur5x8DZ5v5ICRAVVw2Q8iIqIarNwBaP78+VVRB1Wj//1zC/9GpcBSZYSlg1tyADsREdU55R4DRIbtRlw6vjhwHQAwt29TuNmYSlwRERFR9St3C5BcLi+1xYAzxGqufLUGM7ZfRG6+Bt19HTG0bT2pSyIiIpJEuQPQrl27dB7n5eXh33//xYYNG7Bw4UK9FUb699+/b+HivVRYmRghZBC7voiIqO4qdwDq379/kW1DhgxBs2bNsHXrVkycOFEvhZF+hcemYcVBsetrQb9mcLE2kbgiIiIi6ehtDNBzzz2H0NBQfR2O9CivoOsrTy2gRxNnDGztLnVJREREktJLAHr06BFWrlwJd3d+sNZE3xy+iSv302BjZowlg5qz64uIiOq8cneBPX3RU0EQkJ6eDjMzM2zcuFGvxVHlXX2QipWhNwAAC/s1g5Mlu76IiIjKHYC+/PJLnQAkl8vh6OiIwMBA2Nra6rU4qpzcfA1mbL+EfI2Ans1c0M/fTeqSiIiIaoRyB6Dx48dXQRlUFY5GJiAsRuz6WjyAXV9ERESFyj0G6Pvvv8f27duLbN++fTs2bNigl6JIP67eTwMAvODrBEdLlcTVEBER1RzlDkAhISFwcHAost3JyQlLlizRS1GkH+Fx6QAAXxdLiSshIiKqWcodgKKiouDt7V1ku6enJ6KiovRSFOlHeIzYAuTnaiVxJURERDVLuQOQk5MTLl26VGT7xYsXYW9vr5eiqPKy89S4nZgJAPBjCxAREZGOcgegkSNHYurUqTh06BDUajXUajX++usvTJs2DSNGjKiKGqkCIuMzoBEAWzNjOHH8DxERkY5yzwJbvHgx7ty5gxdffBFGRuLLNRoNxo4dyzFANUhYYfeXixVnfxERET2l3AFIqVRi69at+OSTT3DhwgWYmpqiRYsW8PT0rIr6qIIiYjkAmoiIqCTlDkCFGjVqhEaNGumzFtKj8IIA1MSVAYiIiOhp5R4DNHjwYCxdurTI9s8++wxDhw7VS1FUeeHaFiDOACMiInpauQPQ33//jd69exfZ3qtXL/z99996KYoqJzEjB4kZOZDJgMbOFlKXQ0REVOOUOwBlZGRAqVQW2W5sbIy0tDS9FEWVUzj+x9PODGbKCvdyEhER1VrlDkAtWrTA1q1bi2zfsmULmjZtqpeiqHKenAFGRERERZW7eWDu3LkYNGgQbt68iRdeeAEAEBoaik2bNmHHjh16L5DKjzPAiIiISlfuANS3b1/s3r0bS5YswY4dO2Bqagp/f3/89ddfsLOzq4oaqZw4A4yIiKh0FRog0qdPH/Tp0wcAkJaWhs2bN2PGjBk4d+4c1Gq1Xguk8lFrBFyP4wwwIiKi0pR7DFChv//+G+PGjYObmxu++OILvPDCCzh58qQ+a6MKuJOUiZx8DUyNFahvZyZ1OURERDVSuVqAYmNjsX79eqxduxZpaWkYNmwYcnJysHv3bg6AriHCY8TWn8bOFlDIeQkMIiKi4pS5Bahv377w9fXFpUuXsGLFCjx48ACrVq2qytqoAiJiOQOMiIjoWcrcAvTHH39g6tSpeOedd3gJjBosjDPAiIiInqnMLUBHjx5Feno62rZti8DAQHz99ddITEysytqoAgqnwPtxBhgREVGJyhyAnnvuOXz33XeIiYnBW2+9hS1btsDNzQ0ajQYHDhxAenp6VdZJZZCRk4+o5CwA7AIjIiIqTblngZmbm+O1117D0aNHcfnyZbz//vv4v//7Pzg5OaFfv35VUSOVUeH0dydLFezMi16uhIiIiEQVngYPAL6+vvjss89w7949bN68WV81UQUVzgDzc2XrDxERUWkqFYAKKRQKDBgwAHv27NHH4aiCHs8A4/gfIiKi0uglAFHNUDgDjAGIiIiodAxAtYQgCLwIKhERURkxANUSsWnZSH2UB4VchoZOFlKXQ0REVKMxANUShVeAb+BgDpWRQuJqiIiIajYGoFqCM8CIiIjKjgGolgjnDDAiIqIyqxEBaPXq1fDy8oKJiQkCAwNx+vTpEvft1q0bZDJZkVufPn20+4wfP77I8z179qyOU5FMBGeAERERlVmZL4ZaVbZu3Yrg4GCsWbMGgYGBWLFiBYKCghAREQEnJ6ci++/cuRO5ubnax0lJSfD398fQoUN19uvZsye+//577WOVSlV1JyGx3HwNIuMzAHAGGBERUVlI3gK0fPlyvPHGG5gwYQKaNm2KNWvWwMzMDOvWrSt2fzs7O7i4uGhvBw4cgJmZWZEApFKpdPaztbWtjtORxK3EDORrBFiqjOBuYyp1OURERDWepAEoNzcX586dQ48ePbTb5HI5evTogRMnTpTpGGvXrsWIESNgbm6us/3w4cNwcnKCr68v3nnnHSQlJZV4jJycHKSlpencDMnjAdCWkMlkEldDRERU80kagBITE6FWq+Hs7Kyz3dnZGbGxsc98/enTp3HlyhW8/vrrOtt79uyJH374AaGhoVi6dCmOHDmCXr16Qa1WF3uckJAQWFtba28eHh4VPykJhHMBRCIionKRfAxQZaxduxYtWrRAQECAzvYRI0Zo77do0QItW7aEj48PDh8+jBdffLHIcWbNmoXg4GDt47S0NIMKQY9ngHEKPBERUVlI2gLk4OAAhUKBuLg4ne1xcXFwcXEp9bWZmZnYsmULJk6c+Mz3adCgARwcHBAZGVns8yqVClZWVjo3Q8IZYEREROUjaQBSKpVo27YtQkNDtds0Gg1CQ0PRoUOHUl+7fft25OTk4NVXX33m+9y7dw9JSUlwdXWtdM01TWpWHmJSswEAjRmAiIiIykTyWWDBwcH47rvvsGHDBoSFheGdd95BZmYmJkyYAAAYO3YsZs2aVeR1a9euxYABA2Bvb6+zPSMjAx988AFOnjyJO3fuIDQ0FP3790fDhg0RFBRULedUnQq7v9xtTGFlYixxNURERIZB8jFAw4cPR0JCAubNm4fY2Fi0atUK+/bt0w6MjoqKglyum9MiIiJw9OhR/Pnnn0WOp1AocOnSJWzYsAEpKSlwc3PDyy+/jMWLF9fKtYAKB0A3cWXrDxERUVnJBEEQpC6ipklLS4O1tTVSU1Nr/HigWTsvYfPpaEzq7oMPgvykLoeIiEgy5fn8lrwLjConXDsAumYHNSIiopqEAciAaTQCZ4ARERFVAAOQAbv38BGyctVQKuTwdjB/9guIiIgIAAOQQQsrmAHW0MkCRgr+KImIiMqKn5oGTNv9xRlgRERE5cIAZMAK1wBqwgHQRERE5cIAZMB4EVQiIqKKYQAyUNl5atxJzATALjAiIqLyYgAyUDfiMqARADtzJRwtat8K10RERFWJAchAFc4A83OxhEwmk7gaIiIiw8IAZKDCYzj+h4iIqKIYgAxURBxngBEREVUUA5CBYgsQERFRxTEAGaCE9BwkZeZCJgMaOzMAERERlRcDkAEqXADRy94cpkqFxNUQEREZHgYgA8QrwBMREVUOA5ABCuP4HyIiokphADJAhTPA/DgDjIiIqEIYgAxMvlqD63EZAIAmvAQGERFRhTAAGZg7SVnIzdfATKmAh62Z1OUQEREZJAYgA1M4A6yxsyXkcl4Cg4iIqCIYgAwMZ4ARERFVHgOQgSmcAcYAREREVHEMQAamsAvMlzPAiIiIKowByICkZ+fh3sNHANgCREREVBkMQAbkepzY/eVspYKtuVLiaoiIiAwXA5ABCdcOgGb3FxERUWUwABmQcA6AJiIi0gsGIAOinQLPFaCJiIgqhQHIQAiCgLDCGWDO7AIjIiKqDAYgAxGTmo307HwYyWXwcTKXuhwiIiKDxgBkIArX/2ngaA6VkULiaoiIiAwbA5CB4AwwIiIi/WEAMhDaGWAcAE1ERFRpDEAGorALjFPgiYiIKo8ByADk5KtxKyETALvAiIiI9IEByADcjM9EvkaApYkRXK1NpC6HiIjI4DEAGYCIOLH7q4mLFWQymcTVEBERGT4GIANQOADal+N/iIiI9IIByACE8xIYREREesUAZAA4A4yIiEi/GIBquIeZuYhLywEANHZmACIiItIHBqAarrD7q56tKSxNjCWuhoiIqHZgAKrhIrTdX1z/h4iISF8YgGq4x9cAY/cXERGRvjAA1XBhnAFGRESkdzUiAK1evRpeXl4wMTFBYGAgTp8+XeK+3bp1g0wmK3Lr06ePdh9BEDBv3jy4urrC1NQUPXr0wI0bN6rjVPRKoxFwnVeBJyIi0jvJA9DWrVsRHByM+fPn4/z58/D390dQUBDi4+OL3X/nzp2IiYnR3q5cuQKFQoGhQ4dq9/nss8+wcuVKrFmzBqdOnYK5uTmCgoKQnZ1dXaelF1HJWXiUp4bSSA4vezOpyyEiIqo1JA9Ay5cvxxtvvIEJEyagadOmWLNmDczMzLBu3bpi97ezs4OLi4v2duDAAZiZmWkDkCAIWLFiBebMmYP+/fujZcuW+OGHH/DgwQPs3r27Gs+s8grH/zR2toCRQvIfFRERUa0h6adqbm4uzp07hx49emi3yeVy9OjRAydOnCjTMdauXYsRI0bA3NwcAHD79m3ExsbqHNPa2hqBgYFlPmZNUbgAoq8zu7+IiIj0yUjKN09MTIRarYazs7POdmdnZ4SHhz/z9adPn8aVK1ewdu1a7bbY2FjtMZ4+ZuFzT8vJyUFOTo72cVpaWpnPoSpFFLQANeEAaCIiIr0y6H6VtWvXokWLFggICKjUcUJCQmBtba29eXh46KnCyinsAuNFUImIiPRL0gDk4OAAhUKBuLg4ne1xcXFwcXEp9bWZmZnYsmULJk6cqLO98HXlOeasWbOQmpqqvUVHR5f3VPTuUa4ad5IyAXAGGBERkb5JGoCUSiXatm2L0NBQ7TaNRoPQ0FB06NCh1Ndu374dOTk5ePXVV3W2e3t7w8XFReeYaWlpOHXqVInHVKlUsLKy0rlJ7XpcOgQBsDdXwtFSJXU5REREtYqkY4AAIDg4GOPGjUO7du0QEBCAFStWIDMzExMmTAAAjB07Fu7u7ggJCdF53dq1azFgwADY29vrbJfJZJg+fTo++eQTNGrUCN7e3pg7dy7c3NwwYMCA6jqtSovgAohERERVRvIANHz4cCQkJGDevHmIjY1Fq1atsG/fPu0g5qioKMjlug1VEREROHr0KP78889ij/nhhx8iMzMTb775JlJSUtC5c2fs27cPJiYmVX4++hLGGWBERERVRiYIgiB1ETVNWloarK2tkZqaKll32KjvTuL4zSR8NqQlhrWrGYOyiYiIarLyfH4b9Cyw2koQBITFFF4Fnl1gRERE+sYAVAMlpOfgYVYe5DKgkRMDEBERkb4xANVAhev/eDmYw1SpkLgaIiKi2ocBqAYqvAQGu7+IiIiqBgNQDVTYAsQFEImIiKoGA1ANFB7DS2AQERFVJQagGiZfrUFkfAYAoAlbgIiIiKoEA1ANczsxE7lqDcyUCtSzNZW6HCIiolqJAaiGefIK8HK5TOJqiIiIaicGoBqGM8CIiIiqHgNQDRPBGWBERERVjgGohgnjDDAiIqIqxwBUg6Rl5+F+yiMA7AIjIiKqSgxANcj1gu4vFysT2JgpJa6GiIio9mIAqkHCCsf/uLL1h4iIqCoxANUgEQUzwDj+h4iIqGoxANUghZfA4ArQREREVYsBqIYQBOHxFHh2gREREVUpBqAa4n7KI6Tn5MNILkMDBwupyyEiIqrVGIBqiMLWn4ZOFlAa8cdCRERUlfhJW0M8eQ0wIiIiqloMQDVEOC+BQUREVG0YgGqI8BheBJWIiKi6MADVADn5atxKzATAGWBERETVgQGoBoiMz4BaI8DKxAguViZSl0NERFTrMQDVAIULIPq5WkEmk0lcDRERUe3HAFQDRMQVDoBm9xcREVF1YACqAcK0A6A5A4yIiKg6MADVABFcA4iIiKhaMQBJLDkzF/HpOQAYgIiIiKoLA5DEwmPF7q/6dmawUBlJXA0REVHdwAAkscIZYGz9ISIiqj4MQBIrHP/ThAGIiIio2jAASaywC8yXM8CIiIiqDQOQhNQa4fEaQLwEBhERUbVhAJJQVHIWsvM0UBnJ4WVvLnU5REREdQYDkIQKrwDf2NkSCjkvgUFERFRdGIAkFM4FEImIiCTBACShwgHQvAYYERFR9WIAklDhFHheA4yIiKh6MQBJJCs3H3eTswBwBhgREVF1YwCSyPW4DAgC4GChhIOFSupyiIiI6hQGIIkUzgBj9xcREVH1YwCSSLh2/A+7v4iIiKobA5BEHl8CgwGIiIioujEASUAQhMcXQXVlFxgREVF1kzwArV69Gl5eXjAxMUFgYCBOnz5d6v4pKSmYNGkSXF1doVKp0LhxY/z+++/a5xcsWACZTKZz8/Pzq+rTKJf49Bw8zMqDXAY0dLKQuhwiIqI6x0jKN9+6dSuCg4OxZs0aBAYGYsWKFQgKCkJERAScnJyK7J+bm4uXXnoJTk5O2LFjB9zd3XH37l3Y2Njo7NesWTMcPHhQ+9jISNLTLCKsYAC0t4M5TIwVEldDRERU90iaDJYvX4433ngDEyZMAACsWbMGe/fuxbp16zBz5swi+69btw7Jyck4fvw4jI2NAQBeXl5F9jMyMoKLi0uV1l4ZXACRiIhIWpJ1geXm5uLcuXPo0aPH42LkcvTo0QMnTpwo9jV79uxBhw4dMGnSJDg7O6N58+ZYsmQJ1Gq1zn43btyAm5sbGjRogNGjRyMqKqpKz6W8OAOMiIhIWpK1ACUmJkKtVsPZ2Vlnu7OzM8LDw4t9za1bt/DXX39h9OjR+P333xEZGYl3330XeXl5mD9/PgAgMDAQ69evh6+vL2JiYrBw4UI8//zzuHLlCiwtiw8cOTk5yMnJ0T5OS0vT01kWjxdBJSIiklbNGhzzDBqNBk5OTvj222+hUCjQtm1b3L9/H8uWLdMGoF69emn3b9myJQIDA+Hp6Ylt27Zh4sSJxR43JCQECxcurJZzyFNrEBnPGWBERERSkqwLzMHBAQqFAnFxcTrb4+LiShy/4+rqisaNG0OheDxwuEmTJoiNjUVubm6xr7GxsUHjxo0RGRlZYi2zZs1Camqq9hYdHV2BMyqb24mZyFMLMFcq4G5jWmXvQ0RERCWTLAAplUq0bdsWoaGh2m0ajQahoaHo0KFDsa/p1KkTIiMjodFotNuuX78OV1dXKJXKYl+TkZGBmzdvwtXVtcRaVCoVrKysdG5VpXAGmK+LJeRyWZW9DxEREZVM0nWAgoOD8d1332HDhg0ICwvDO++8g8zMTO2ssLFjx2LWrFna/d955x0kJydj2rRpuH79Ovbu3YslS5Zg0qRJ2n1mzJiBI0eO4M6dOzh+/DgGDhwIhUKBkSNHVvv5FSdCO/6H3V9ERERSkXQM0PDhw5GQkIB58+YhNjYWrVq1wr59+7QDo6OioiCXP85oHh4e2L9/P9577z20bNkS7u7umDZtGj766CPtPvfu3cPIkSORlJQER0dHdO7cGSdPnoSjo2O1n19xwrUrQHMANBERkVRkgiAIUhdR06SlpcHa2hqpqal67w7r9H9/4X7KI2x98zkENrDX67GJiIjqsvJ8fkt+KYy6JPVRHu6nPALARRCJiIikxABUjQrH/7hZm8DazFjiaoiIiOouBqBqFBH7eAYYERERSYcBqBqlZefD1FgBPy6ASEREJCkOgi5GVQ6CVmsE5OSrYaY0qEW4iYiIajwOgq7BFHIZww8REZHEGICIiIiozmEAIiIiojqHAYiIiIjqHAYgIiIiqnMYgIiIiKjOYQAiIiKiOocBiIiIiOocBiAiIiKqcxiAiIiIqM5hACIiIqI6hwGIiIiI6hwGICIiIqpzGICIiIiozuFlyYshCAIAIC0tTeJKiIiIqKwKP7cLP8dLwwBUjPT0dACAh4eHxJUQERFReaWnp8Pa2rrUfWRCWWJSHaPRaPDgwQNYWlpCJpPp9dhpaWnw8PBAdHQ0rKys9HrsmoDnZ/hq+zny/AxfbT9Hnl/FCYKA9PR0uLm5QS4vfZQPW4CKIZfLUa9evSp9Dysrq1r5i12I52f4avs58vwMX20/R55fxTyr5acQB0ETERFRncMARERERHUOA1A1U6lUmD9/PlQqldSlVAmen+Gr7efI8zN8tf0ceX7Vg4OgiYiIqM5hCxARERHVOQxAREREVOcwABEREVGdwwBEREREdQ4DUDVavXo1vLy8YGJigsDAQJw+fVrqkvQmJCQE7du3h6WlJZycnDBgwABERERIXVaV+b//+z/IZDJMnz5d6lL05v79+3j11Vdhb28PU1NTtGjRAmfPnpW6LL1Rq9WYO3cuvL29YWpqCh8fHyxevLhM1wyqif7++2/07dsXbm5ukMlk2L17t87zgiBg3rx5cHV1hampKXr06IEbN25IU2wFlXaOeXl5+Oijj9CiRQuYm5vDzc0NY8eOxYMHD6QruJye9TN80ttvvw2ZTIYVK1ZUW32VVZbzCwsLQ79+/WBtbQ1zc3O0b98eUVFR1VIfA1A12bp1K4KDgzF//nycP38e/v7+CAoKQnx8vNSl6cWRI0cwadIknDx5EgcOHEBeXh5efvllZGZmSl2a3p05cwb//e9/0bJlS6lL0ZuHDx+iU6dOMDY2xh9//IFr167hiy++gK2trdSl6c3SpUvxzTff4Ouvv0ZYWBiWLl2Kzz77DKtWrZK6tArJzMyEv78/Vq9eXezzn332GVauXIk1a9bg1KlTMDc3R1BQELKzs6u50oor7RyzsrJw/vx5zJ07F+fPn8fOnTsRERGBfv36SVBpxTzrZ1ho165dOHnyJNzc3KqpMv141vndvHkTnTt3hp+fHw4fPoxLly5h7ty5MDExqZ4CBaoWAQEBwqRJk7SP1Wq14ObmJoSEhEhYVdWJj48XAAhHjhyRuhS9Sk9PFxo1aiQcOHBA6Nq1qzBt2jSpS9KLjz76SOjcubPUZVSpPn36CK+99prOtkGDBgmjR4+WqCL9ASDs2rVL+1ij0QguLi7CsmXLtNtSUlIElUolbN68WYIKK+/pcyzO6dOnBQDC3bt3q6coPSrp/O7duye4u7sLV65cETw9PYUvv/yy2mvTh+LOb/jw4cKrr74qTUGCILAFqBrk5ubi3Llz6NGjh3abXC5Hjx49cOLECQkrqzqpqakAADs7O4kr0a9JkyahT58+Oj/L2mDPnj1o164dhg4dCicnJ7Ru3Rrfffed1GXpVceOHREaGorr168DAC5evIijR4+iV69eElemf7dv30ZsbKzO76m1tTUCAwNr7d8cQPy7I5PJYGNjI3UpeqHRaDBmzBh88MEHaNasmdTl6JVGo8HevXvRuHFjBAUFwcnJCYGBgaV2A+obA1A1SExMhFqthrOzs852Z2dnxMbGSlRV1dFoNJg+fTo6deqE5s2bS12O3mzZsgXnz59HSEiI1KXo3a1bt/DNN9+gUaNG2L9/P9555x1MnToVGzZskLo0vZk5cyZGjBgBPz8/GBsbo3Xr1pg+fTpGjx4tdWl6V/h3pa78zQGA7OxsfPTRRxg5cmStuYDo0qVLYWRkhKlTp0pdit7Fx8cjIyMD//d//4eePXvizz//xMCBAzFo0CAcOXKkWmrg1eBJ7yZNmoQrV67g6NGjUpeiN9HR0Zg2bRoOHDhQff3T1Uij0aBdu3ZYsmQJAKB169a4cuUK1qxZg3HjxklcnX5s27YNP/30EzZt2oRmzZrhwoULmD59Otzc3GrNOdZVeXl5GDZsGARBwDfffCN1OXpx7tw5fPXVVzh//jxkMpnU5eidRqMBAPTv3x/vvfceAKBVq1Y4fvw41qxZg65du1Z5DWwBqgYODg5QKBSIi4vT2R4XFwcXFxeJqqoakydPxm+//YZDhw6hXr16UpejN+fOnUN8fDzatGkDIyMjGBkZ4ciRI1i5ciWMjIygVqulLrFSXF1d0bRpU51tTZo0qbbZGNXhgw8+0LYCtWjRAmPGjMF7771XK1v0Cv+u1IW/OYXh5+7duzhw4ECtaf35559/EB8fj/r162v/5ty9exfvv/8+vLy8pC6v0hwcHGBkZCTp3x0GoGqgVCrRtm1bhIaGardpNBqEhoaiQ4cOElamP4IgYPLkydi1axf++usveHt7S12SXr344ou4fPkyLly4oL21a9cOo0ePxoULF6BQKKQusVI6depUZNmC69evw9PTU6KK9C8rKwtyue6fPIVCof2faG3i7e0NFxcXnb85aWlpOHXqVK35mwM8Dj83btzAwYMHYW9vL3VJejNmzBhcunRJ52+Om5sbPvjgA+zfv1/q8ipNqVSiffv2kv7dYRdYNQkODsa4cePQrl07BAQEYMWKFcjMzMSECROkLk0vJk2ahE2bNuGXX36BpaWldpyBtbU1TE1NJa6u8iwtLYuMZzI3N4e9vX2tGOf03nvvoWPHjliyZAmGDRuG06dP49tvv8W3334rdWl607dvX3z66aeoX78+mjVrhn///RfLly/Ha6+9JnVpFZKRkYHIyEjt49u3b+PChQuws7ND/fr1MX36dHzyySdo1KgRvL29MXfuXLi5uWHAgAHSFV1OpZ2jq6srhgwZgvPnz+O3336DWq3W/t2xs7ODUqmUquwye9bP8OlAZ2xsDBcXF/j6+lZ3qRXyrPP74IMPMHz4cHTp0gXdu3fHvn378Ouvv+Lw4cPVU6Bk88/qoFWrVgn169cXlEqlEBAQIJw8eVLqkvQGQLG377//XurSqkxtmgYvCILw66+/Cs2bNxdUKpXg5+cnfPvtt1KXpFdpaWnCtGnThPr16wsmJiZCgwYNhNmzZws5OTlSl1Yhhw4dKvbf3Lhx4wRBEKfCz507V3B2dhZUKpXw4osvChEREdIWXU6lnePt27dL/Ltz6NAhqUsvk2f9DJ9maNPgy3J+a9euFRo2bCiYmJgI/v7+wu7du6utPpkgGOgyqEREREQVxDFAREREVOcwABEREVGdwwBEREREdQ4DEBEREdU5DEBERERU5zAAERERUZ3DAERERER1DgMQEVEZyGQy7N69W+oyiEhPGICIqMYbP348ZDJZkVvPnj2lLo2IDBSvBUZEBqFnz574/vvvdbapVCqJqiEiQ8cWICIyCCqVCi4uLjo3W1tbAGL31DfffINevXrB1NQUDRo0wI4dO3Ref/nyZbzwwgswNTWFvb093nzzTWRkZOjss27dOjRr1gwqlQqurq6YPHmyzvOJiYkYOHAgzMzM0KhRI+zZs6dqT5qIqgwDEBHVCnPnzsXgwYNx8eJFjB49GiNGjEBYWBgAIDMzE0FBQbC1tcWZM2ewfft2HDx4UCfgfPPNN5g0aRLefPNNXL58GXv27EHDhg113mPhwoUYNmwYLl26hN69e2P06NFITk6u1vMkIj2ptsuuEhFV0Lhx4wSFQiGYm5vr3D799FNBEAQBgPD222/rvCYwMFB45513BEEQhG+//VawtbUVMjIytM/v3btXkMvlQmxsrCAIguDm5ibMnj27xBoACHPmzNE+zsjIEAAIf/zxh97Ok4iqD8cAEZFB6N69O7755hudbXZ2dtr7HTp00HmuQ4cOuHDhAgAgLCwM/v7+MDc31z7fqVMnaDQaREREQCaT4cGDB3jxxRdLraFly5ba++bm5rCyskJ8fHxFT4mIJMQAREQGwdzcvEiXlL6YmpqWaT9jY2OdxzKZDBqNpipKIqIqxjFARFQrnDx5ssjjJk2aAACaNGmCixcvIjMzU/v8sWPHIJfL4evrC0tLS3h5eSE0NLRaayYi6bAFiIgMQk5ODmJjY3W2GRkZwcHBAQCwfft2tGvXDp07d8ZPP/2E06dPY+3atQCA0aNHY/78+Rg3bhwWLFiAhIQETJkyBWPGjIGzszMAYMGCBXj77bfh5OSEXr16IT09HceOHcOUKVOq90SJqFowABGRQdi3bx9cXV11tvn6+iI8PByAOENry5YtePfdd+Hq6orNmzejadOmAAAzMzPs378f06ZNQ/v27WFmZobBgwdj+fLl2mONGzcO2dnZ+PLLLzFjxgw4ODhgyJAh1XeCRFStZIIgCFIXQURUGTKZDLt27cKAAQOkLoWIDATHABEREVGdwwBEREREdQ7HABGRwWNPPhGVF1uAiIiIqM5hACIiIqI6hwGIiIiI6hwGICIiIqpzGICIiIiozmEAIiIiojqHAYiIiIjqHAYgIiIiqnMYgIiIiKjO+X/AjmaD9Qh9CAAAAABJRU5ErkJggg==",
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\n",
            "Calculating metrics...\n",
            "Accuracy: 0.8252\n",
            "Precision: [0.83357952 0.81723115]\n",
            "Recall: [0.81264 0.83776]\n",
            "F1 score: [0.82297659 0.82736825]\n",
            "Support: [12500 12500]\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "         0.0       0.83      0.81      0.82     12500\n",
            "         1.0       0.82      0.84      0.83     12500\n",
            "\n",
            "    accuracy                           0.83     25000\n",
            "   macro avg       0.83      0.83      0.83     25000\n",
            "weighted avg       0.83      0.83      0.83     25000\n",
            "\n"
          ]
        },
        {
          "data": {
            "text/plain": [
              "{'val_loss': 0.3928542864877124}"
            ]
          },
          "execution_count": 30,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "# === Run the function ===\n",
        "CNN_EMBEDDINGS(TEXT, train_data, valid_data, test_data,num_filters = 20,filter_sizes = [3,4,5],dropout=0.6, learning_rate=0.002, num_epochs=17, batch_size=118, clip=4,freeze_embeddings=False)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oS2-uGfXDqo_"
      },
      "source": [
        "**PART 7: USE A PRE-TRAINED MODEL FROM HUGGING FACE AND FINE TUNE IT TO THE IMDB DATASET**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "I_XqypPG4BA6",
        "outputId": "4df40a29-3147-4f76-fe67-99825cb6dc75"
      },
      "outputs": [
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/huggingface_hub/utils/_token.py:88: UserWarning: \n",
            "The secret `HF_TOKEN` does not exist in your Colab secrets.\n",
            "To authenticate with the Hugging Face Hub, create a token in your settings tab (https://huggingface.co/settings/tokens), set it as secret in your Google Colab and restart your session.\n",
            "You will be able to reuse this secret in all of your notebooks.\n",
            "Please note that authentication is recommended but still optional to access public models or datasets.\n",
            "  warnings.warn(\n",
            "Some weights of DistilBertForSequenceClassification were not initialized from the model checkpoint at distilbert-base-uncased and are newly initialized: ['classifier.bias', 'pre_classifier.bias', 'classifier.weight', 'pre_classifier.weight']\n",
            "You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\n"
          ]
        },
        {
          "data": {
            "text/plain": [
              "DistilBertForSequenceClassification(\n",
              "  (distilbert): DistilBertModel(\n",
              "    (embeddings): Embeddings(\n",
              "      (word_embeddings): Embedding(30522, 768, padding_idx=0)\n",
              "      (position_embeddings): Embedding(512, 768)\n",
              "      (LayerNorm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n",
              "      (dropout): Dropout(p=0.1, inplace=False)\n",
              "    )\n",
              "    (transformer): Transformer(\n",
              "      (layer): ModuleList(\n",
              "        (0): TransformerBlock(\n",
              "          (attention): MultiHeadSelfAttention(\n",
              "            (dropout): Dropout(p=0.1, inplace=False)\n",
              "            (q_lin): Linear(in_features=768, out_features=768, bias=True)\n",
              "            (k_lin): Linear(in_features=768, out_features=768, bias=True)\n",
              "            (v_lin): Linear(in_features=768, out_features=768, bias=True)\n",
              "            (out_lin): Linear(in_features=768, out_features=768, bias=True)\n",
              "          )\n",
              "          (sa_layer_norm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n",
              "          (ffn): FFN(\n",
              "            (dropout): Dropout(p=0.1, inplace=False)\n",
              "            (lin1): Linear(in_features=768, out_features=3072, bias=True)\n",
              "            (lin2): Linear(in_features=3072, out_features=768, bias=True)\n",
              "            (activation): GELUActivation()\n",
              "          )\n",
              "          (output_layer_norm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n",
              "        )\n",
              "        (1): TransformerBlock(\n",
              "          (attention): MultiHeadSelfAttention(\n",
              "            (dropout): Dropout(p=0.1, inplace=False)\n",
              "            (q_lin): Linear(in_features=768, out_features=768, bias=True)\n",
              "            (k_lin): Linear(in_features=768, out_features=768, bias=True)\n",
              "            (v_lin): Linear(in_features=768, out_features=768, bias=True)\n",
              "            (out_lin): Linear(in_features=768, out_features=768, bias=True)\n",
              "          )\n",
              "          (sa_layer_norm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n",
              "          (ffn): FFN(\n",
              "            (dropout): Dropout(p=0.1, inplace=False)\n",
              "            (lin1): Linear(in_features=768, out_features=3072, bias=True)\n",
              "            (lin2): Linear(in_features=3072, out_features=768, bias=True)\n",
              "            (activation): GELUActivation()\n",
              "          )\n",
              "          (output_layer_norm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n",
              "        )\n",
              "        (2): TransformerBlock(\n",
              "          (attention): MultiHeadSelfAttention(\n",
              "            (dropout): Dropout(p=0.1, inplace=False)\n",
              "            (q_lin): Linear(in_features=768, out_features=768, bias=True)\n",
              "            (k_lin): Linear(in_features=768, out_features=768, bias=True)\n",
              "            (v_lin): Linear(in_features=768, out_features=768, bias=True)\n",
              "            (out_lin): Linear(in_features=768, out_features=768, bias=True)\n",
              "          )\n",
              "          (sa_layer_norm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n",
              "          (ffn): FFN(\n",
              "            (dropout): Dropout(p=0.1, inplace=False)\n",
              "            (lin1): Linear(in_features=768, out_features=3072, bias=True)\n",
              "            (lin2): Linear(in_features=3072, out_features=768, bias=True)\n",
              "            (activation): GELUActivation()\n",
              "          )\n",
              "          (output_layer_norm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n",
              "        )\n",
              "        (3): TransformerBlock(\n",
              "          (attention): MultiHeadSelfAttention(\n",
              "            (dropout): Dropout(p=0.1, inplace=False)\n",
              "            (q_lin): Linear(in_features=768, out_features=768, bias=True)\n",
              "            (k_lin): Linear(in_features=768, out_features=768, bias=True)\n",
              "            (v_lin): Linear(in_features=768, out_features=768, bias=True)\n",
              "            (out_lin): Linear(in_features=768, out_features=768, bias=True)\n",
              "          )\n",
              "          (sa_layer_norm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n",
              "          (ffn): FFN(\n",
              "            (dropout): Dropout(p=0.1, inplace=False)\n",
              "            (lin1): Linear(in_features=768, out_features=3072, bias=True)\n",
              "            (lin2): Linear(in_features=3072, out_features=768, bias=True)\n",
              "            (activation): GELUActivation()\n",
              "          )\n",
              "          (output_layer_norm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n",
              "        )\n",
              "        (4): TransformerBlock(\n",
              "          (attention): MultiHeadSelfAttention(\n",
              "            (dropout): Dropout(p=0.1, inplace=False)\n",
              "            (q_lin): Linear(in_features=768, out_features=768, bias=True)\n",
              "            (k_lin): Linear(in_features=768, out_features=768, bias=True)\n",
              "            (v_lin): Linear(in_features=768, out_features=768, bias=True)\n",
              "            (out_lin): Linear(in_features=768, out_features=768, bias=True)\n",
              "          )\n",
              "          (sa_layer_norm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n",
              "          (ffn): FFN(\n",
              "            (dropout): Dropout(p=0.1, inplace=False)\n",
              "            (lin1): Linear(in_features=768, out_features=3072, bias=True)\n",
              "            (lin2): Linear(in_features=3072, out_features=768, bias=True)\n",
              "            (activation): GELUActivation()\n",
              "          )\n",
              "          (output_layer_norm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n",
              "        )\n",
              "        (5): TransformerBlock(\n",
              "          (attention): MultiHeadSelfAttention(\n",
              "            (dropout): Dropout(p=0.1, inplace=False)\n",
              "            (q_lin): Linear(in_features=768, out_features=768, bias=True)\n",
              "            (k_lin): Linear(in_features=768, out_features=768, bias=True)\n",
              "            (v_lin): Linear(in_features=768, out_features=768, bias=True)\n",
              "            (out_lin): Linear(in_features=768, out_features=768, bias=True)\n",
              "          )\n",
              "          (sa_layer_norm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n",
              "          (ffn): FFN(\n",
              "            (dropout): Dropout(p=0.1, inplace=False)\n",
              "            (lin1): Linear(in_features=768, out_features=3072, bias=True)\n",
              "            (lin2): Linear(in_features=3072, out_features=768, bias=True)\n",
              "            (activation): GELUActivation()\n",
              "          )\n",
              "          (output_layer_norm): LayerNorm((768,), eps=1e-12, elementwise_affine=True)\n",
              "        )\n",
              "      )\n",
              "    )\n",
              "  )\n",
              "  (pre_classifier): Linear(in_features=768, out_features=768, bias=True)\n",
              "  (classifier): Linear(in_features=768, out_features=2, bias=True)\n",
              "  (dropout): Dropout(p=0.2, inplace=False)\n",
              ")"
            ]
          },
          "execution_count": 28,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "# Define the pre-trained model from Hugging Face and its tokenizer\n",
        "model_name = 'distilbert-base-uncased'\n",
        "tokenizer = DistilBertTokenizerFast.from_pretrained(model_name)\n",
        "model = DistilBertForSequenceClassification.from_pretrained(model_name, num_labels = 2)\n",
        "model.to(device)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "7WplZYNg5UTw"
      },
      "outputs": [],
      "source": [
        "# Create torch dataset\n",
        "class Dataset(torch.utils.data.Dataset):\n",
        "    def __init__(self, encodings, labels=None):\n",
        "        self.encodings = encodings\n",
        "        self.labels = labels\n",
        "\n",
        "    def __getitem__(self, idx):\n",
        "        item = {key: torch.tensor(val[idx]) for key, val in self.encodings.items()}\n",
        "        if self.labels:\n",
        "            item['labels'] = torch.tensor(self.labels[idx])\n",
        "        return item\n",
        "\n",
        "    def __len__(self):\n",
        "        return len(self.encodings['input_ids'])\n",
        "\n",
        "\n",
        "# Define Trainer parameters\n",
        "def compute_metrics(p):\n",
        "\n",
        "  pred, labels = p\n",
        "  pred = np.argmax(pred, axis=1)\n",
        "\n",
        "  accuracy = accuracy_score(y_true=labels, y_pred=pred)\n",
        "  recall = recall_score(y_true=labels, y_pred=pred)\n",
        "  precision = precision_score(y_true=labels, y_pred=pred)\n",
        "  f1 = f1_score(y_true=labels, y_pred=pred)\n",
        "  return {'accuracy': accuracy, 'precision': precision, 'recall': recall, 'f1': f1}"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "FkFUchVEkzX0",
        "outputId": "d7426a41-265c-4fcb-d7bc-c07432c60275"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Training, validation and testing datasets are ready!\n"
          ]
        }
      ],
      "source": [
        "# Get the texts from the training, validation and test data\n",
        "X_train = list(train_dataframe['review'])\n",
        "X_val = list(valid_dataframe['review'])\n",
        "X_test = list(test_dataframe['review'])\n",
        "\n",
        "y_train = list(train_dataframe['label'])\n",
        "y_val = list(valid_dataframe['label'])\n",
        "\n",
        "# Tokenize and pad the data using DistilBert's tokenizer\n",
        "X_train_tokenized = tokenizer(X_train, padding = True, truncation = True)\n",
        "X_val_tokenized = tokenizer(X_val, padding = True, truncation = True)\n",
        "X_test_tokenized = tokenizer(X_test, padding = True, truncation = True)\n",
        "\n",
        "# Create the Datasets\n",
        "train_dataset = Dataset(X_train_tokenized, y_train)\n",
        "val_dataset = Dataset(X_val_tokenized, y_val)\n",
        "test_dataset = Dataset(X_test_tokenized)\n",
        "print('Training, validation and testing datasets are ready!')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 452
        },
        "id": "4JMC3aUp5smJ",
        "outputId": "41d26bf4-58e3-4e46-f200-f791a6d71c3b"
      },
      "outputs": [
        {
          "data": {
            "text/html": [
              "\n",
              "    <div>\n",
              "      \n",
              "      <progress value='1079' max='6000' style='width:300px; height:20px; vertical-align: middle;'></progress>\n",
              "      [1079/6000 12:42 < 58:01, 1.41 it/s, Epoch 0.54/3]\n",
              "    </div>\n",
              "    <table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              " <tr style=\"text-align: left;\">\n",
              "      <th>Step</th>\n",
              "      <th>Training Loss</th>\n",
              "      <th>Validation Loss</th>\n",
              "      <th>Accuracy</th>\n",
              "      <th>Precision</th>\n",
              "      <th>Recall</th>\n",
              "      <th>F1</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <td>500</td>\n",
              "      <td>0.420900</td>\n",
              "      <td>0.307679</td>\n",
              "      <td>0.878200</td>\n",
              "      <td>0.860439</td>\n",
              "      <td>0.904573</td>\n",
              "      <td>0.881954</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>1000</td>\n",
              "      <td>0.324300</td>\n",
              "      <td>0.338971</td>\n",
              "      <td>0.884200</td>\n",
              "      <td>0.874903</td>\n",
              "      <td>0.898211</td>\n",
              "      <td>0.886404</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table><p>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "text/html": [
              "\n",
              "    <div>\n",
              "      \n",
              "      <progress value='3500' max='6000' style='width:300px; height:20px; vertical-align: middle;'></progress>\n",
              "      [3500/6000 42:24 < 30:18, 1.37 it/s, Epoch 1/3]\n",
              "    </div>\n",
              "    <table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              " <tr style=\"text-align: left;\">\n",
              "      <th>Step</th>\n",
              "      <th>Training Loss</th>\n",
              "      <th>Validation Loss</th>\n",
              "      <th>Accuracy</th>\n",
              "      <th>Precision</th>\n",
              "      <th>Recall</th>\n",
              "      <th>F1</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <td>500</td>\n",
              "      <td>0.420900</td>\n",
              "      <td>0.307679</td>\n",
              "      <td>0.878200</td>\n",
              "      <td>0.860439</td>\n",
              "      <td>0.904573</td>\n",
              "      <td>0.881954</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>1000</td>\n",
              "      <td>0.324300</td>\n",
              "      <td>0.338971</td>\n",
              "      <td>0.884200</td>\n",
              "      <td>0.874903</td>\n",
              "      <td>0.898211</td>\n",
              "      <td>0.886404</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>1500</td>\n",
              "      <td>0.324200</td>\n",
              "      <td>0.287436</td>\n",
              "      <td>0.880200</td>\n",
              "      <td>0.937443</td>\n",
              "      <td>0.816302</td>\n",
              "      <td>0.872689</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>2000</td>\n",
              "      <td>0.325200</td>\n",
              "      <td>0.264963</td>\n",
              "      <td>0.894600</td>\n",
              "      <td>0.899839</td>\n",
              "      <td>0.889463</td>\n",
              "      <td>0.894621</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>2500</td>\n",
              "      <td>0.224200</td>\n",
              "      <td>0.310160</td>\n",
              "      <td>0.899200</td>\n",
              "      <td>0.913616</td>\n",
              "      <td>0.883101</td>\n",
              "      <td>0.898099</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>3000</td>\n",
              "      <td>0.232400</td>\n",
              "      <td>0.330476</td>\n",
              "      <td>0.898200</td>\n",
              "      <td>0.901843</td>\n",
              "      <td>0.895030</td>\n",
              "      <td>0.898423</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <td>3500</td>\n",
              "      <td>0.231700</td>\n",
              "      <td>0.302466</td>\n",
              "      <td>0.902800</td>\n",
              "      <td>0.901147</td>\n",
              "      <td>0.906163</td>\n",
              "      <td>0.903648</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table><p>"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "data": {
            "text/plain": [
              "TrainOutput(global_step=3500, training_loss=0.29755510820661274, metrics={'train_runtime': 2545.5598, 'train_samples_per_second': 23.57, 'train_steps_per_second': 2.357, 'total_flos': 4636358952960000.0, 'train_loss': 0.29755510820661274, 'epoch': 1.75})"
            ]
          },
          "execution_count": 32,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "# ------ Train the model ------\n",
        "# Define the Trainer\n",
        "args = TrainingArguments(\n",
        "    output_dir = 'output',\n",
        "    evaluation_strategy = 'steps',\n",
        "    eval_steps = 500,\n",
        "    per_device_train_batch_size = 10,\n",
        "    per_device_eval_batch_size = 8,\n",
        "    num_train_epochs = 3,\n",
        "    seed = 42,\n",
        "    learning_rate = 2e-5,\n",
        "    weight_decay = 0.01,\n",
        "    load_best_model_at_end = True)\n",
        "\n",
        "trainer = Trainer(\n",
        "    model = model,\n",
        "    args = args,\n",
        "    train_dataset = train_dataset,\n",
        "    eval_dataset = val_dataset,\n",
        "    compute_metrics = compute_metrics,\n",
        "    callbacks = [EarlyStoppingCallback(early_stopping_patience = 3)])\n",
        "\n",
        "# Train the pre-trained model\n",
        "trainer.train()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 17
        },
        "id": "SGSd8u4y6B-y",
        "outputId": "677f6326-cdb4-45b5-cbab-b617114a1499"
      },
      "outputs": [
        {
          "data": {
            "text/html": [],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "# ------ Make predictions ------\n",
        "\n",
        "# Load the trained model\n",
        "model_path = 'output/checkpoint-2000'\n",
        "model = DistilBertForSequenceClassification.from_pretrained(model_path, num_labels = 2)\n",
        "\n",
        "# Define the test trainer\n",
        "test_trainer = Trainer(model)\n",
        "\n",
        "# Make prediction\n",
        "raw_pred, _, _ = test_trainer.predict(test_dataset)\n",
        "\n",
        "# Preprocess raw predictions\n",
        "y_pred = np.argmax(raw_pred, axis=1)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "HRF13_WRSv_C",
        "outputId": "cb1002a1-4153-468e-8ed0-1a80f45b55a4"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "[1 0 1 ... 0 1 0]\n"
          ]
        }
      ],
      "source": [
        "print(y_pred)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Jmck1_dpapAT"
      },
      "source": [
        "**PART 8: PRESENTING THE RESULTS/ COMMENT ON THE FINDINGS**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XWFhFX0BHLFX"
      },
      "source": [
        "![image.png](data:image/png;base64,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)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AIv9cMordxDs"
      },
      "source": [
        "DistilBert: These are the results from the best loaded model at 2000 evaluation steps.\n",
        "From the results’ board above, it is evident that the DistilBert model achieved the highest scores of all the models. It also seems that the CNN with the frozen embeddings did not achieve as satisfying results as the other models. Second best model after DistillBERT seems to be the LSTM with the frozen embeddings, since it achieved second best scoring performance in all metrics.\n",
        "From the plotted learning curves above, we can see though that:\n",
        "the CNN model with the frozen embedings seems to be underfitting the train data, since it reports consistently higher accuracy and lower loss in the validation compared to the train data and the two model trained with GloVe also show some overfitting on the train data.\n",
        "\n",
        "This implies that the chosen hyperarameters from Optuna, did not help the models in optimization.\n",
        "1. Undefitting: First it seems that freezing the embeddings with the CNN architecture and the given optimized hyperprameters were not enough for the complexity of our text data. A more complex model is needed.\n",
        "\n",
        "2. Overfitting: It seems that retraining the pretrained GloVe embeddings and using the chosen Optuna hyperparameters confused the model.\n",
        "Trying out other learning rates, dropout rates, or smaller layer sizes might be necessary to improve model generalization and performance.\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7FsH2YVP57jC"
      },
      "source": [
        "**References**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DZHpYcr_6MQJ"
      },
      "source": [
        "Maas, A. L., Daly, R. E., Pham, P. T., Huang, D., Ng, A. Y., & Potts, C. (2011, June 1). Learning Word Vectors for Sentiment Analysis. ACLWeb; Association for Computational Linguistics. http://www.aclweb.org/anthology/P11-1015"
      ]
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "gpuType": "T4",
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "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": "3.9.18"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}