Sentiment_Analysis_AI_CourseworkC_LewisWatson.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Sentiment Analysis Coursework - Lewis N Watson - 40432878"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Imports\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pylab as plt\n",
    "from keras.preprocessing.text import Tokenizer\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Dropout, Activation, LSTM\n",
    "from keras.layers.embeddings import Embedding\n",
    "from keras.utils import np_utils\n",
    "from keras.preprocessing import sequence\n",
    "import gensim.downloader as api\n",
    "np.random.seed(7)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "def filter_stopwords(arr):\n",
    "    print(\"Filtering stopwords...\")\n",
    "    stopwords = ['stop', 'the', 'to', 'and', 'a', 'in', 'it', 'is', 'i', 'so', 'that', 'had', 'on', 'for', 'were',\n",
    "                 'was', 'of', 'as', 'an']\n",
    "    newArr = np.array([])\n",
    "\n",
    "    for sentences in arr:\n",
    "        new_sentence = \"\"\n",
    "        word_count = 0\n",
    "        for word in sentences.split():\n",
    "            if word not in stopwords:\n",
    "                if word_count == 0:\n",
    "                    new_sentence += word\n",
    "                    word_count += 1\n",
    "                else:\n",
    "                    new_sentence += \" \" + word\n",
    "                    word_count += 1\n",
    "\n",
    "        if new_sentence != \"\":\n",
    "            newArr = np.append(newArr, new_sentence)\n",
    "\n",
    "    return newArr"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def load_data(filename):\n",
    "    print(\"Loading data...\")\n",
    "    df = pd.read_csv(filename, names=['sentence', 'label'], sep=',')\n",
    "\n",
    "    X_raw = df['sentence'].values\n",
    "    y_raw = df['label'].values\n",
    "\n",
    "    X_raw = filter_stopwords(X_raw)\n",
    "\n",
    "    return X_raw, y_raw"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Loading data...\n",
      "Filtering stopwords...\n",
      "Loading data...\n",
      "Filtering stopwords...\n",
      "Loading data...\n",
      "Filtering stopwords...\n"
     ]
    }
   ],
   "source": [
    "# Load data\n",
    "X_train, y_train = load_data('train.csv')\n",
    "X_test, y_test = load_data('test.csv')\n",
    "X_val, y_val = load_data('val.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load word2vec model\n",
    "word2vec_model = api.load(\"glove-wiki-gigaword-300\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Embedding matrix shape:  (400000, 300)\n"
     ]
    }
   ],
   "source": [
    "embedding_matrix = word2vec_model.vectors\n",
    "print('Embedding matrix shape: ', embedding_matrix.shape)\n",
    "top_words = embedding_matrix.shape[0]\n",
    "mxlen = 50"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Tokenize and to sequences\n",
    "tokenizer = Tokenizer(num_words=top_words,filters='!\"#$%&()*+,-./:;<=>?@[\\]^_`{|}~', lower=True)\n",
    "tokenizer.fit_on_texts(X_train)\n",
    "sequences_train = tokenizer.texts_to_sequences(X_train)\n",
    "sequences_test = tokenizer.texts_to_sequences(X_test)\n",
    "sequences_val = tokenizer.texts_to_sequences(X_val)\n",
    "\n",
    "word_index = tokenizer.word_index"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Pad sequences\n",
    "X_train = sequence.pad_sequences(sequences_train, maxlen=mxlen)\n",
    "X_test = sequence.pad_sequences(sequences_test, maxlen=mxlen)\n",
    "X_val = sequence.pad_sequences(sequences_val, maxlen=mxlen)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Hyperparameters to fiddle with\n",
    "batch_size = 8\n",
    "nb_epoch = 12"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "embedding_layer = Embedding(embedding_matrix.shape[0],\n",
    "                                embedding_matrix.shape[1],\n",
    "                                weights=[embedding_matrix],\n",
    "                                trainable=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:Layer lstm will not use cuDNN kernels since it doesn't meet the criteria. It will use a generic GPU kernel as fallback when running on GPU.\n",
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " embedding (Embedding)       (None, None, 300)         120000000 \n",
      "                                                                 \n",
      " lstm (LSTM)                 (None, 100)               160400    \n",
      "                                                                 \n",
      " dense (Dense)               (None, 1)                 101       \n",
      "                                                                 \n",
      " activation (Activation)     (None, 1)                 0         \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 120,160,501\n",
      "Trainable params: 160,501\n",
      "Non-trainable params: 120,000,000\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "# Build model\n",
    "model = Sequential()\n",
    "model.add(embedding_layer)\n",
    "model.add(LSTM(100, recurrent_dropout=0.1, dropout=0.3))\n",
    "model.add(Dense(1))\n",
    "model.add(Activation('sigmoid'))  # try softmax for categorical_crossentropy\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/12\n",
      "175/175 [==============================] - 31s 158ms/step - loss: 0.7095 - accuracy: 0.5043 - val_loss: 0.6834 - val_accuracy: 0.5700\n",
      "Epoch 2/12\n",
      "175/175 [==============================] - 27s 155ms/step - loss: 0.6556 - accuracy: 0.6157 - val_loss: 0.6802 - val_accuracy: 0.5700\n",
      "Epoch 3/12\n",
      "175/175 [==============================] - 27s 156ms/step - loss: 0.6094 - accuracy: 0.6771 - val_loss: 0.7255 - val_accuracy: 0.5200\n",
      "Epoch 4/12\n",
      "175/175 [==============================] - 31s 177ms/step - loss: 0.5535 - accuracy: 0.7321 - val_loss: 0.7260 - val_accuracy: 0.5600\n",
      "Epoch 5/12\n",
      "175/175 [==============================] - 32s 181ms/step - loss: 0.4739 - accuracy: 0.7764 - val_loss: 0.7470 - val_accuracy: 0.5800\n",
      "Epoch 6/12\n",
      "175/175 [==============================] - 35s 199ms/step - loss: 0.3991 - accuracy: 0.8193 - val_loss: 0.8494 - val_accuracy: 0.5750\n",
      "Epoch 7/12\n",
      "175/175 [==============================] - 30s 174ms/step - loss: 0.3033 - accuracy: 0.8793 - val_loss: 0.9176 - val_accuracy: 0.5450\n",
      "Epoch 8/12\n",
      "175/175 [==============================] - 33s 190ms/step - loss: 0.2241 - accuracy: 0.9186 - val_loss: 1.1609 - val_accuracy: 0.5300\n",
      "Epoch 9/12\n",
      "175/175 [==============================] - 49s 281ms/step - loss: 0.1853 - accuracy: 0.9293 - val_loss: 1.0600 - val_accuracy: 0.5550\n",
      "Epoch 10/12\n",
      "175/175 [==============================] - 31s 178ms/step - loss: 0.1409 - accuracy: 0.9514 - val_loss: 1.1101 - val_accuracy: 0.5500\n",
      "Epoch 11/12\n",
      "175/175 [==============================] - 30s 172ms/step - loss: 0.0955 - accuracy: 0.9686 - val_loss: 1.5027 - val_accuracy: 0.5450\n",
      "Epoch 12/12\n",
      "175/175 [==============================] - 31s 176ms/step - loss: 0.0841 - accuracy: 0.9729 - val_loss: 1.3030 - val_accuracy: 0.5800\n"
     ]
    }
   ],
   "source": [
    "# Compile and train model\n",
    "model.compile(optimizer='adam', loss='binary_crossentropy', metrics=['accuracy'])\n",
    "rnn = model.fit(X_train, y_train, epochs=nb_epoch, batch_size=batch_size, shuffle=True,validation_data=(X_val, y_val))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Evaluating model...\n",
      "\n",
      "Training Accuracy: 99.7143\n",
      "Training Loss:  2.2998\n",
      "\n",
      "Testing Accuracy:  56.0000\n",
      "Testing Loss:  130.0217\n"
     ]
    }
   ],
   "source": [
    "# Evaluate model\n",
    "print(\"Evaluating model...\\n\")\n",
    "loss, accuracy = model.evaluate(X_train, y_train, verbose=False)\n",
    "print(\"Training Accuracy: {:.4f}\".format(accuracy * 100))\n",
    "print(\"Training Loss:  {:.4f}\".format(loss * 100))\n",
    "\n",
    "print(\"\") # New Line...\n",
    "\n",
    "loss, accuracy = model.evaluate(X_test, y_test, verbose=False)\n",
    "print(\"Testing Accuracy:  {:.4f}\".format(accuracy*100))\n",
    "print(\"Testing Loss:  {:.4f}\".format(loss * 100))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_valloss_vs_train_accuracy(training):\n",
    "    plt.plot(training.history['val_loss'], label='Validation Loss')\n",
    "    plt.plot(training.history['accuracy'], label='Training Accuracy')\n",
    "    plt.title('Sentiment Analysis - Validation loss vs Training Accuracy')\n",
    "    plt.ylabel('Value')\n",
    "    plt.xlabel('Epoch')\n",
    "    plt.legend(loc=\"lower right\")\n",
    "    plt.show()\n",
    "\n",
    "def plot_accuracy(training):\n",
    "    plt.plot(training.history['val_accuracy'], label='Validation Accuracy')\n",
    "    plt.plot(training.history['accuracy'], label='Training Accuracy')\n",
    "    plt.title('Sentiment Analysis - Training vs Validation Accuracy')\n",
    "    plt.ylabel('Value')\n",
    "    plt.xlabel('Epoch')\n",
    "    plt.legend(loc=\"lower right\")\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYYAAAEWCAYAAABi5jCmAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/MnkTPAAAACXBIWXMAAAsTAAALEwEAmpwYAABCiUlEQVR4nO3dd3hUVfrA8e+bQgJJqAktlNA7oaOgAooKgoAVWHVBXdva0HVta1u7v3Vdy6os9goWpIiICgiooNRQQyeUBEIIISSkz5zfH/cGJiE9M5kk836eJ09mbpv3Tjnvveeec64YY1BKKaXy+Xk7AKWUUtWLJgallFIFaGJQSilVgCYGpZRSBWhiUEopVYAmBqWUUgVoYihERKaLyOPejqOqichTIvJpJbdR7d47EYkTkZH240dF5N2yLFuB1zlfRHZUNM4SthslIkZEAty97epORK4TkR/dvawqXY1IDCJynoisFJFUETkuIr+JyEA3bHeqiPzqOs0Yc7sx5pnKbrsCsZS5YBaRZSKSIiJBno6rPDzx3tnJ5uMipkeLSLaINC5HfM8bY/7ipriMiHR02fYvxpgu7th2TWZ/Xun2X46I5Lo8/7482zLGfGaMucTdy1ZUdf3deUK1TwwiUh9YALwBNAYigX8C2d6My1tEJAo4HzDAOO9GUyU+Aq4UkZBC028AFhhjjnshJlUM++Ag1BgTCjwPfJH/3BgzOn+5mnYG5M3fnVfeK2NMtf4DBgAnSlnmJiAWSAF+ANq6zDPA7cAu4ATwJiBANyALcADp+a8BfAg8az8eDhwCHgSOAoeBCcBlwE7gOPCoy2v5AQ8De4Bk4EugsT0vyo5lCnAAOAb8w543CsgBcu1YNpawr08AvwGvYBWMrvM+tPfvOyAN+APo4DL/NeAgcBJYB5zvMu8p4FP78XfA3YW2vQm4wn7v/mO/HyeBzUDPIt67cKyEfsJ+n34B/Cr4HdgB/NnluT+QAIwHOgBL7ff7GPAZ0NBl2ThgZOF9tJ/fAOy31/1HoWUHAavs+A8D/wXq2PNW2J/lKfvzmpj/XXHZdjdgmb3+VmBcWT+nQvue/70JsJ+3BObb7+lu4BaXZQcBa+3PJRF4xZ4eDHxq7+cJYA3QrIjXegj4utC014DX7cdTgb12zPuA60r53Aq/33H2a2zCOrAL4MzvJQ3YBlzhsvxU4NfSfssVWNYf+DfW92UfcJfre1yB311r4BsgyX6P/+sy7xassil///q5xNex0HeicLnzEHAE+ARohPV7SsIq5xYArVzWbwx8gPW7SAHm2tO3AJe7LBdo73ffEj+7ivxQq/IPqG+/2R8Bo4FGheaPx/qBdLO/aI8BKwt9QRYADYE29hs7qqgvUzEfUJ79pQi0P+Qk4HMgDOgBZALt7OXvBX4HWgFBwP+AmYV+4O8AdYForB9Ht6J+RCW8H7uBvwL9sRJJs0KxJ2MVEAFYheQsl/nXA03seX+zv3TBhV8fuBb4w2W9aHu7dYBLsZJKQ84k2BZFvHcvANPt9y0Q62hLKvgd+Aew2OX5pfbnEAh0BC623+8IrEL71UKF0VmJAeiOVahfYK/7iv1Z5y/bHzjHfq+isH7c0wp9r1x/2MOxE4Md127gUfs9uxCrYOhSls+p0L7nf2/yE8MK4C2swr6P/T5caM9bBdxgPw4FzrEf3wZ8C9TDKhT7A/WLeK22QAYQZj/3x0qK5wAhWAknfx9aAD1K+dxOv98un0UMVkFa1552DVay88NKsKc4832aytmFfZl+y6UseztWId0Kq8BdTOmJocjfnf0ebcQ6WAqxP5fzXPYtHhiI9VvpiH3QSumJIQ94Ceu7WRfrd3uV/RmGAV9hF/72Ot8BX9j7EwgMs6c/iHXW5lpebi71N1eRH2pV/2EVPh9iZdE8rCOm/A/me+Bml2X97C+36wdwnsv8L4GHi/oyFfMBZQL+9vMwe3uDXZZfB0ywH8cCF7nMa2F/ifILF0PBLL8amFTUj6iY9+E8e3vh9vPtwH2FYn/X5fllwPYStpcCRBd+ffvLnQJ0sp+/DLxlP74Q62zpHAqdARR6754G5uHy5a/E59/G3u9W9vPPgNeKWXYCsMHleRxFJ4YnKJg0Q7DO2kYWs91pwByX5yUlhvOxkq6fy/yZwFPl/ZxcvjcBWAWqA7vgtue/AHxoP16BVc0aXmgbNwErgd5leK9/xT47w0q4e1zenxNYhVPdMn5uBb7T9mdxUynrxADji/p9Uo7fcinLLgVuc5k3khISAyX87oBzsZLOWeti1V7cW8w2S0sMOdgHbcWs3wdIsR+3AJwUOmi257XEOiipbz//GniwtM+u2l9jADDGxBpjphpjWgE9sXb2VXt2W+A1ETkhIiewTrEF61pEviMujzOwjqbKKtkY47AfZ9r/E13mZ7psry0wxyWWWKwfcjM3xTIF+NEYc8x+/rk9zVWx2xeRB0Qk1r6IfwJogFXlU4AxJgvr6ON6EfEDJmOdzmKMWYpVrfImcFREZtjXgQr7F9ZR1o8isldEHi5qh+yWQvkXJ6cXtYwx5gBWoXe9iIRiFf4f2+s3E5FZIhIvIiexqkzO2qcitMSqVst/jVNYR/H5cXUWkQUicsTe7vNl3O7pbRtjnC7T9lP572RL4LgxJq2Y7d4MdAa2i8gaERlrT/8Eq5CaJSIJIvJ/IhJYzGt8jvV5A/zJfp7//kzEOto+LCLfiUjXMsRc2EHXJyLyZxGJcfnN9KTk97k871txyxb47AvHVISSfnetgf3GmLwi1muNVU1WEUn27xAAEaknIv8Tkf3293EF0FBE/O3XOW6MSSm8EWNMAlYV2FUi0hCr1uWz0l68RiQGV8aY7VjZtac96SBW9m/o8lfXGLOyLJtzc3gHgdGFYgk2xsRXNhYRqYtVxTPMLqyOAPcB0SISXdrGReR8rNPKa7GOLBoCqVhJtCgfAdcBFwEZxphVpwM15nVjTH+s6pjOwN/P2hlj0owxfzPGtMe6WHe/iFxUxHLPmzMXJ28vYRc+wromcBWwzxizzp7+PNZ718sYUx+ruqy4fXJ1GOsHBVg/PKzT9XxvYx0ZdrK3+2gZtwtWPW9rO6nma4NVrVAZCUBjEQkrarvGmF3GmMlAU6xqiK9FJMQYk2uM+acxpjswBBgL/LmY1/gKGC4irbCuKX2eP8MY84Mx5mKsI9TtWNWi5XX6ey4ibe1t3AU0sb+TWyj7+1xRh7GqkfK1Lm7BMvzuDgJtirlAfBDrGlhRMrCqhfI1LzS/cHnwN6ALVm1FfawqULDeq4NY34uGxbzWR1i/i2uAVWUpj6p9YhCRriLyN/uLioi0xjqi+d1eZDrwiIj0sOc3EJFryrj5RKCViNRxU7jTgefsLzwiEiEi48sRS1ShwsTVBKyzj+5Yp5F9sKrYfqH4H7mrMKxquCQgQESewLp+UyQ7ETixLtJ9kj9dRAaKyGD7iPMU1gV8Z+H1RWSsiHQUEcFKQI6iliuH2ViF4D+xvuiu+5UOpIpIJEUkqWJ8DYy1m0LXwar6cn3vw7Dq1NPtI+M7Cq2fCLQvZtt/YP3wHxSRQBEZDlwOzCpjbEUyxhzEqhJ6QUSCRaQ31lnCpwAicr2IRNhnKifs1ZwiMkJEetlHlyexqkWK/CyMMUlYF80/wErAsfa2m4nIeLt1WDbWe16ZzxOs6imD9Z1ERG7kzAGfJ30J3CsikXZh+lAJy06g5N/daqxE86KIhNify1B73XeBB0Skv1g65pcNWFVmfxIRfxEZBQwrJeYwrNqJE2I10X4yf4Yx5jBWlfpbItLI/s5d4LLuXKAf1jXQs5p+F6XaJwas+rHBwB8icgorIWzByqAYY+ZgHR3Nsk+xtmCdLpXFUqwWI0dE5FhpC5fBa1jXP34UkTQ71sFlXPcr+3+yiKwvYv4U4ANjzAFjzJH8P6xqnevK0KTtB2AR1vWB/VgFemmn0B8DvbALHlt9rKO8FM606PlXEet2wrqol451UfQtY8zPpbxeseyqjNlYR3qup8L/xPrSp2JdgPumjNvbCtyJdUR8GGt/Drks8gBWVUoa1v5+UWgTTwEf2VUg1xbadg5WIhiN1QLkLax6++1lia0Uk7GuOyQAc4AnjTGL7XmjgK0iko71XZxkjMnEOhr9GispxALLcUn2Rfgcq979c5dpfsD99usexyrICifLcjHGbMM68FiFlWh7YVV7eNo7wI9YraM2AAuxDpocRSxb4u8O64j9cqwLywewvkMTAYwxXwHPYb2PaVgFdH6/m3vt9U7Y25lbSsyvYl2EPoZVriwqNP8GrIS/HavF4LT8GfZ3YDbQjjL+PvKbbyl1FhH5M3CrMeY8b8eilKeIyGhgujGmbakL11B2DUFnY8z1ZVm+JpwxKC+w69z/CszwdixKuZOI1BWRy0QkwK5+fBLr7KtWsquebqYcv2VNDOosIpLfTyCRgtUJStUGglUFmYJVlRSL1Xy51hGRW7CqjL83xqwo83palaSUUsqVnjEopZQqoEYNZAUQHh5uoqKivB2GUkrVKOvWrTtmjIkoy7I1LjFERUWxdu1ab4ehlFI1iojsL+uyWpWklFKqAE0MSimlCtDEoJRSqgBNDEoppQrQxKCUUqoATQxKKaUK0MSglFKqAE0MSilVAbGHT7IkNrH0BWsgTQxKKVUBj87ZzC0fr2Xd/rPuqFnjaWJQSqly2p98ig0HTmCA+7+M4VR2Ubd8rrk0MSilVDnNi0lABF6d2IcDxzN4ZsE2b4fkVpoYlFKqHIwxzN0Qz6CoxozvE8ntwzowa81Bfth6xNuhuY0mBqWUKofN8ansPXaKCX0jAbhvZGd6tKzPI99s5mhalpejcw9NDEopVQ5zNyRQx9+Py3q2AKBOgB+vTuzDqew8Hvp6E7Xh5meaGJRSqowcTsO3mxIY0TWCBvUCT0/v1CyMR0Z35ecdSXz2xwEvRugemhiUUqqMVu45RlJaNhP6RJ4178/nRnF+p3Ce+y6WvUnpXojOfTQxKKVUGc3dkEBYcAAjujY9a56fn/DyNdEEBfpx3xcx5DqcXojQPTQxKKVUGWTlOvhh6xFG92xOcKB/kcs0qx/M81f0YuOhVN5YsquKI3QfTQxKKVUGi2MTSc/OK7IaydVlvVpwZb9I/vvz7hrbK1oTg1JKlcHcDfE0qx/E4PZNSl32n+N60LJh3RrbK1oTg1JKlSLlVA7LdiQxLrol/n5S6vJhwYG8cm3N7RWtiUEppUrx3ebD5DnN6U5tZTGoXePTvaJ/2lazRmHVxKCUUqWYFxNPp6ahdG9Rv1zr5feKfmj2phrVK1oTg1JKleBQSgZr4lKY0DcSkdKrkVzV1F7RmhiUUqoE82ISABgX3bJC63dqFsbDNaxXtMcSg4i8LyJHRWRLKcsNFJE8EbnaU7EopVRFGGOYFxPPgLaNaN24XoW3M6WG9Yr25BnDh8CokhYQEX/gJeBHD8ahlFIVEns4jZ2J6Ywvx0XnotS0XtEeSwzGmBXA8VIWuxuYDRz1VBxKKVVRc2PiCfATxvRqUeltFegVvXS3G6LzHK9dYxCRSOAK4G1vxaCUUsVxOA3zYxIY1jmCxiF13LLN/F7Rb1bzXtHevPj8KvCQMabUcyoRuVVE1orI2qSkJM9HppTyeX/sS+bIyaxy9V0oi6fG9aB5/eBq3Svam4lhADBLROKAq4G3RGRCUQsaY2YYYwYYYwZERERUYYhKKV81b0MCIXX8GdmtmVu3Wz84kP9U83tFey0xGGPaGWOijDFRwNfAX40xc70Vj1JK5cvKdbBwy2Eu7dmcunWKHkm1Mga1a8xtF1TfXtGebK46E1gFdBGRQyJys4jcLiK3e+o1lVLKHZbtOEpaVukjqVbG/Rd3pnuL+jw8exNJadkee52KCPDUho0xk8ux7FRPxaGUUuU1d0MC4aFBDOlQ+kiqFVUnwI/XJvVh7Bu/8tDsTbw3ZUC5e1Z7ivZ8VkopF6mZuSzdfpTLo1sQ4O/ZIjK/V/TS7Uf5fHX16RWtiUEppVx8v/kwOQ6nR6uRXOX3in52QfXpFa2JQSmlXMyNiaddeAi9WzWoktfz8xP+dXU0dQKqT69oTQxKKWU7nJrJH/uOM6FP+UdSrYzmDapXr2hNDEopZZsfk4AxML5PxUZSrYwxvVtwZV+rV/T6A97tFa2JQSmlbHNjEujTuiFR4SFeef2nxlu9ou/7wru9ojUxKKUUsDMxjdjDJ5nghbOFfK69op/9znu9ojUxKKUUMHdDPP5+wtgK3pDHXfJ7Rc9c7b1e0ZoYlFI+z+k0zItJ4LyO4YSHBnk7HK/3itbEoJTyeWv3pxB/IpMJfb17tpCvToAfr07qQ1p2Hg/Nrvp7RWtiUEr5vLkx8dQN9OeS7s29HcppnZuF8fAo7/SK1sSglPJpOXlOFm4+zCU9mhES5LHh4ypk6hDv9IrWxKCU8mnLdyZxIiO3yobAKI8CvaK/3FhlvaI1MSilfNrcmHgah9ThvE7h3g6lSKd7RR88UWW9oqvXeZNSSlWhtKxcFm9LZOLA1gR6eCTVyhjTuwWxhzsyokvV3MFSE4NSymf9sDWR7Dwn46thNVJhD1zapcpeq/qmSKWU8rB5MfG0aVyPfm0aejuUakUTg1LKJx09mcVvu48xvk/LanPntOpCE4NSyifN35iA01AjqpGqmiYGpZRPmheTQK/IBnRsGurtUKodTQxKKZ+zJymdzfGpXrnvQk2giUEp5XPmbYjHT2Ccl0dSra40MSilfIoxhrkxCQzpEE7T+sHeDqda0sSglPIpGw6e4MDxDK1GKoEmBqWUT5m7IZ6gAD9G9aw+I6lWN5oYlFI+I9fhZMGmw4zs1oyw4EBvh1NtaWJQSvmMX3cd4/ipHCb01b4LJdHEoJTyGXNj4mlYL5BhnatmMLqaShODUsonnMrO48etiVzWqwV1ArToK4m+O0opn/DTtkQycx3V8oY81Y0mBqWUT5gbE09kw7oMaNvI26FUex5LDCLyvogcFZEtxcy/TkQ2ichmEVkpItGeikUp5duOpWfzy65jjOvTEj8/HUm1NJ48Y/gQGFXC/H3AMGNML+AZYIYHY1FK+bAFGxNwOI1WI5WRx+7gZoxZISJRJcxf6fL0d6CVp2JRSvm2uTEJdGtRny7Nw7wdSo1QXa4x3Ax8X9xMEblVRNaKyNqkpKQqDEspVdPFHTtFzMETTNAhMMrM64lBREZgJYaHilvGGDPDGDPAGDMgIkLbHyulym5eTAIiME4TQ5l5rCqpLESkN/AuMNoYk+zNWJRStY8xhnkx8Qxu15gWDep6O5waw2tnDCLSBvgGuMEYs9NbcSilaq/N8ansPXZKLzqXk8fOGERkJjAcCBeRQ8CTQCCAMWY68ATQBHjLvhF3njFmgKfiUUr5nrkbEqjj78foXi28HUqN4slWSZNLmf8X4C+een2llG/Lczj5dlMCI7pG0KCujqRaHl6/+KyUUp6wck8ySWnZWo1UAZoYlPIhH62M456ZGzialuXtUDxubkw8YcEBjOja1Nuh1DhebZWklKo6B5IzeO67WHIcTlbuOca/rolmRJfaWWhm5jj4YcsRxvZuSXCgv7fDqXH0jEEpH/H8wlj8/YRPbh5EeGgQN36whqe/3UZ2nsPbobnd4thETuU4GN9X+y5UhCYGpXzAqj3JLNp6hL8O78D5nSKYe+dQppzblvd/28cVb65k99F0b4foVvNi4mleP5hz2jXxdig1kiYGpWo5h9Pw9IJtRDasyy0XtAcgONCff47vybt/HsDh1Ewuf+NXZq0+gDHGy9FWXsqpHJbtSNKRVCtBE4NStdwXaw4Se/gkj1zW9az69pHdm7Fo2gX0a9uQh7/ZzJ2fryc1I9dLkbrHd5sPk+c0jNchMCpME4NStdjJrFz+/eMOBkU1Zkwxnbya1Q/mk5sG89Corvy4NZHRr61gTdzxKo7UfebFxNOpaSjdW9T3dig1liYGpWqxN5bs4nhGDk9c3h17hIEi+fkJdwzvwOw7hhAY4MfE/63iPz/tJM/hrMJoK+/g8QzWxKUwoW9kifurSqaJQalaat+xU3y4Mo5r+reiZ2SDMq0T3boh391zPhP6RvLakl1MmvE7h1IyPByp+8zfmADAuGitRqoMTQxK1VLPfbeNoAB/Hri0S7nWCw0K4JVr+/DapD5sP5LG6Nd+YcGmBA9F6T7GGOZuiGdgVCNaN67n7XBqNE0MStVCv+xKYnHsUe4c0ZGmYcEV2sb4PpEsvOd8OkSEctfnG3jw641k5OS5OVL3OHg8g1d+2smuo+mM1yEwKk17PitVy+Q5nDyzYBttGtfjpvOiKrWtNk3q8dXt5/Lq4p28tWwPa+NSeH1y3zJXTXlSyqkcFmw+zLwN8azdnwLA+Z3CtTWSG2hiUKqW+Xz1AXYmpjP9+v4EBVR+OIhAfz/+fmlXhnYM5/4vNnLFW7/x0Kiu3DS0XZX3E8jKdbAk9ihzNsSzfOdRch2Gzs1CeWhUV8b1aUlkQ70ZjztoYlCqFjmRkcMrP+3k3PZNuLRHM7due0iHcL6/93wenL2JZ7+L5Zddx3j5mmgiwoLc+jqFOZyGP/YmM2dDPIu2HCEtO49m9YO4cWg7JvSJpFuLMG2B5GaaGJSqRV5dvIuTmbmlNk+tqEYhdZhxQ38+++MAzyzYxujXVvDyNdEM98BgfLGHTzJ3QzzzYhI4cjKL0KAARvdszhV9Ixncvgn+2qvZYzQxKFVL7D6axie/72fSoDZ082DnLhHh+nPaMqhdY+6ZuYGpH6zh5vPa8eCoLpWuuko4kcm8mATmxcSz/UgaAX7C8C4RPDa2GyO7NdORUquIJgalaolnFsRSr44/f7u4c5W8XudmYcy9cygvLIzlvV/3sWpPMq9P7kvHpqHl2k5qZi6LthxmzoZ4/th3HGOgX5uGPDO+B2N6t6RxSB0P7YEqjiYGpWqBn7cfZfnOJB4b040moZ6t83eVPxjf+Z0i+PvXG7n8jV958vLuTBzYusSqrOw8B8t2JDEvJp7FsUfJyXPSPjyE+0Z2ZnyflrRtElJl+6DOJjVtNMUBAwaYtWvXejsMpaqNXIeTS19dAQYWTbuAOgHe6Z6UeDKL+7+M4bfdyVzWqzkvXNGbBvXO3GvZ6TSsO5DCnA3xfLfpMKmZuTQJqcPl0S25om8kvVs10IvIHiQi64wxA8qyrJ4xKFXDfbxqP3uTTvHelAFeSwpwZjC+/63Yy79/3MHGg7/w6qQ+NKoXyBz7IvKhlEzqBvpzSY9mTOgbyXkdwwn013621Y0mBqVqsOOncnht8U7O7xTOhdXg3sb5g/Gd26EJ987awDXTV1nTBc7rFMHfLunMJd2bExKkRU91VuZPR0TqGWNqzmhaSvmAV37awakcB0+M9Uzz1IrqYw/GN2PFXuoHBzCuT8sKD82hql6piUFEhgDvAqFAGxGJBm4zxvzV08EppYq3/chJPv/jADec05ZOzcK8Hc5ZQoMCuL+KWkgp9ypL5d5/gEuBZABjzEbgAk8GpZQqmTGGZxZsIyw4kGkjtfBV7lWmqz7GmIOFJjk8EItSqox+2pbIb7uTuW9kJxppO3/lZmW5xnDQrk4yIhII3AvEejYspVRxsvMcPLcwlo5NQ7nunLbeDkfVQmU5Y7gduBOIBOKBPvZzpZQXfPhbHPuTM3h8bHdt6qk8otQzBmPMMeC6KohFKVWKpLRs3li6mwu7NmVY5whvh6NqqbK0SvoAOKt7tDHmJo9EpJQq1r9/3EFWroN/jOnm7VBULVaWawwLXB4HA1cA1f8GsErVMlviU/li7UFuGtqODhHlG6hOqfIoS1XSbNfnIjIT+LW09UTkfWAscNQY07OI+QK8BlwGZABTjTHryxi3Uj7FGMPTC7bRqF4d7rmok7fDUbVcRa5cdQLK0vf+Q2BUCfNH29vqBNwKvF2BWJTyCd9vOcLqfce5/+LONKgbWPoKSlVCWa4xpGFdYxD7/xHgodLWM8asEJGoEhYZD3xsrOFdfxeRhiLSwhhzuEyRK+UjsnIdPL8wlq7Nw5g0sLW3w1E+oCxVSZ7qax8JuHacO2RPOysxiMitWGcVtGnTxkPhKFU9vffrPg6lZPL5XwYToM1TVRUoNjGISL+SVqzK6wHGmBnADLDux1BVr6uUtyWezOLNn3dzSfdmDOkY7u1wlI8o6Yzh3yXMM8CFlXzteMD1vLiVPU0pZfu/RTvIcxhtnqqqVLGJwRgzwsOvPR+4S0RmAYOBVL2+oNQZGw+eYPb6Q9w2rL3e6lJVqTLdj0FEegLdsfoxAGCM+biUdWYCw4FwETkEPAkE2utOBxZiNVXdjdVc9cbyh69U7ZTfPDU8NIi7RnT0djiqohx54MiGvGzIy7L/sl3+Zxd6bi/jyCli2SzocCF0H+/xsMvSKulJrAK+O1ZhPhqrH0OJicEYM7mU+QYdc0mpIs3fmMC6/Sm8dFUvwoK1eapXGQOZKXDiAKQetP6fsP9nHLML7mIKcuOGgaj9gyAgGAKCoEGrym+vDMpyxnA1EA1sMMbcKCLNgE89G5ZSviszx8GL32+nR8v6XN1fm6d6nDFwKsku8A+4JICDZx7npBdcp04oNGgNoREQ3NAqtAOCIaCO/T/YZZr9379O0dMLrFfEOl64M19ZEkOWMcYpInkiUh84SsGLxkopN/rfij0cTs3itUl98ferPrfrrLGcDkg7UuiI3+Vx6iHr6N5VcENo2BqadID2w63HDdtYyaBhG6jbyCsFdlUpqbnqm8BMYLWINATeAdYB6cCqKolOKR+TcCKT6cv3MKZXCwa1a+ztcGqOzBNwbCck7z77yD/1EDjzCi4fEmEV8s16QJfR0KCNVeA3bG1ND67vld2oLko6Y9gJ/AtoCZzCShIXA/WNMZuqIDalfM5Li7bjNPDw6K7eDqX6MQZOxlsJIGmn9f/YTkjaAaeOuiwoENbCKuQjB0CPK+0C3y78G7SCOvW8ths1QUnNVV8DXhORtsAk4H2gLjBTRDKNMbuqKEalfMK6/SnMi0ngrhEdad3YhwuuvBxI2WcV+Md2wLFd9uNdkHvqzHLBDSC8C3S+BMI7W4/DO1lH/AF6u9PKKMuQGPuBl4CXRKQvVoJ4AvD3cGxK+Qyn0/D0t1tpVj+IO4Z38HY4VSPrpFXYH9tpJYD8s4CUfQWrfuq3gojO0O8GOwF0hoguVnVQLa7n96ayNFcNwGqiOgm4CFgGPOXRqJTyMXM2xLPxUCqvXBtNSFCZuhfVDMZYF35dq33yH6e59Gf1C7Qu9DbtarXTD+9sJYMmnSBI7z1R1Uq6+HwxMBmrE9pqYBZwqzHmVHHrKKXK71R2Hi8t2k5064ZM6BPp7XBKlpMBmcchI9n+O279nTUt2Wr7f+oY5GWeWT+ovlXotx9hFfz5VUCN2oK/9teoLko6NHkE+Bz4mzEmpYriUcrnvLVsN0fTspl+Q3/8qqp5qjGQm3GmEM8vyF0L9oxklwLfTgCuhXxhwQ2hXhOo1xjqt4TmvaxmnY2izlQBhTXX6p8aoKSLz5UdJE8pVQJjDD/vOMo7v+xjQp+W9GvTyFMvZDXj3L3Y+kvcZhX4hdvunyZQt6FVyNdtbNXxN4+Geo3OTKvX5EwSqNfESgr+tagKzMfpJ6lUFXM6DT/FJvLfpbvZHJ9K68Z1ecjdzVOz02DfijPJ4MQBa3p4Z+gw4uyC3bXAr9sQ/LRtiS/TxKBUFXE4Dd9vOcx/l+5m+5E02japx/9d1Zsr+kUSWNkb8BgDiVvsRLAEDqyyWvbUCbV67p53H3S4yKrLV6oUmhiU8rA8h5P5GxN48+fd7Ek6RYeIEP4zMZrLe7es3B3ZMo7D3p+tRLB7MaQnWtOb94Ihd0PHkdBqkLbpV+WmiUEpD8nJczJnwyHeWraH/ckZdG0expt/6seons0rNgaS0wHx660ksGcJxK8D47Qu8Ha40EoEHS60LvAqVQmaGJRys6xcB1+tO8T0ZXuIP5FJr8gGzLihPyO7NSt/q6O0I2fOCPb+bLUcQqDVABj2kJUMWvbVawLKrTQxKOUmmTkOPl99gBkr9pB4Mpt+bRry7BU9Gd45AilrE828HDj4x5lrBYmbremhzaDLZdDxIqsPQD0dYE95jiYGpSopPTuPT3/fz7u/7OVYeg7ntG/Mf67tw7kdmpQtIaTsP5MI9i23xv73C4A258LIp6yzgmY9tf2/qjKaGJSqoNTMXD5aGcf7v+3jREYuF3SO4O4LOzIwqgxH804H7Pge/pgOcb9Y0xq2gd7XWomg3QUQFObZHVCqGJoYlCqnlFM5vP/bPj78LY607DxGdmvKXRd2ok/rhqWvnJkC6z+B1e9A6gFrJNCLnoRul0OTjnpWoKoFTQxKlVFSWjbv/rqXT1ft51SOg9E9m3PXhR3p0bJB6SsnboPV/4ONX1jDSkSdD5c+Z1030B7DqprRb6RSpTiSmsX/Vuxh5uoD5OQ5uTy6JXeO6EjnZqVU9TgdsHORVV20b4V1D9/e18Kg26B5z6oJXqkK0MSgVDEOpWQwffkevlxzCIcxXNE3kr8O70D7iFKGgc5MgQ2fwuoZ1lAU9VtZF5H7TdHWRKpG0MSgVCH7k0/x5s+7+WZ9PCJwdf/W/HV4h9LvqnY0Fv74H2z6whq5tO1QuORZ6DJGq4tUjaLfVqVsxhg+WhnH8wu3IwLXn9OW24a1p0WDusWv5HTAzh/s6qLlVnVRr6ut6qIWvasueKXcSBODUlgtjf7+9SYWxyZyYdemvHhlL5rWDy5+hcwTLtVF+6F+pNW6qN8UCGlSZXEr5QmaGJTP+2NvMtO+iOFYejaPj+3OTUOjiu+YlrTDqi7aONOqLmozBC5+GrqO1eoiVWvoN1n5LIfT8MbSXby+ZBdtGtfjmzuG0qtVEU1PnQ7Y9aOVEPb+DP5B0OsaGHwrtIiu+sCV8jBNDMonHU7NZNqsGP7Yd5wr+kbyzISehAYV+jlknoCYz6zqopQ4CGsJFz4O/adCSLgXolaqamhiUD5n8bZEHvh6Izl5Tv59TTRX9W9VcIGjsbDmXYiZCbmnzoxZ1HWs3rBe+QRNDMpnZOc5ePH77XzwWxzdW9TnjT/1pUNEqFVVdHA17Fho/SXvBv86VnXRoFuhZR9vh65UlfJoYhCRUcBrgD/wrjHmxULz2wAfAQ3tZR42xiz0ZEzKN+1NSufumRvYmnCSqUOieGRka4L2L4PfFlq9kzOSwS8Q2p0Pg2+H7hMgNMLbYSvlFR5LDCLiD7wJXAwcAtaIyHxjzDaXxR4DvjTGvC0i3YGFQJSnYlK+afa6Qzw+bwst/VNZODSR7mkfwyvLIC8LghtAp0vO3OsguAzjHilVy3nyjGEQsNsYsxdARGYB4wHXxGCA+vbjBkCCB+NRPiY9K5c3v1yA7FjI3Hqb6Jy7HdZhDW/d/0boMhraDtHrBkoV4snEEAkcdHl+CBhcaJmngB9F5G4gBBhZ1IZE5FbgVoA2bdq4PVBVizjy4MAqjq2bS/bWBTxkjkAgmIh+0OUx6HoZNO2uw1srVQJvX3yeDHxojPm3iJwLfCIiPY0xTteFjDEzgBkAAwYMMF6IU1Vn2WnW3c92LMTs/AHJOkF9E8Bav95kn3MX7YdchdRv6e0olaoxPJkY4oHWLs9b2dNc3QyMAjDGrBKRYCAcOOrBuFRtcDLBbkX0vTWktSMHZ3AjVvkP5JOcHvh3upBnrz2XRiF1vB2pUjWOJxPDGqCTiLTDSgiTgD8VWuYAcBHwoYh0A4KBJA/GpGoqYyBxi5UItn8Hh2Os6Y3awaBb2RJ2Hrf+7MexNCePjOnK1CElDGuhlCqRxxKDMSZPRO4CfsBqivq+MWariDwNrDXGzAf+BrwjIvdhXYieaozRqiJlsa8XsH0BbF9o3QoTgVYD4KInoMsY8hp34vWf9/DGt7to16Qu30ztS89IbVmkVGVITSuHBwwYYNauXevtMJSn5GbB3mUQ+61VVZR53BqbqMMIq0lp51EQ1gyAhBPWsBar445zVb9WPD2+ByGFh7VQSgEgIuuMMQPKsqz+ipT3ZaXCrp+sZLB7MeSkQ1B9q39Bt7HQ8WIIKnjXtB+3HuHvX28iz+HkPxOjuaJvq2I2rpQqL00MyjvSj1rXCrYvgL3LwZkLIU2tm9x0vRzaXQABZ184zsp18MLCWD5atZ+ekfV5Y3I/2oWHeGEHlKq9NDGoqpMSB7ELrGRw4HfAQKMoGHwbdLscWg0EP/9iV9991BrWIvbwSW4a2o6HRnchKKD45ZVSFaOJQXmOMZC41UoEsQsgcbM1vVkvGP6wNVppsx6ldjYzxvD1ukM8MW8rwYF+vD91ABd2bVYFO6CUb9LEoNzL6YRDa2D7t1YySNkHCLQeDJc8ayWDxu2KXT01M5e9SensO3aKvUmn2Hssnd1H09mZmM457Rvz6sS+NG9Qwi03lVKVpolBVV5eDsStsBLBjoWQnmiNVNp+GAy912pNFHbmCD8nz8mB4/kF/yn22Qlgb9Ipkk/lnF7O309o3agu7SNCmTiwDVOHROHvp30TlPI0TQyqYrLTrRZE2xfAzh8hOxUCQ6DTxdDtckzHkRzNDWZPUjp7t55i37Ftp88EDqZk4nCeaSYdHlqH9uGhjOzWjPYRIbSPCKVdeAhtGtejToCfF3dSKd+kiUGVnTGw/zdY9xHEzoe8LEzdxqS0vZSdjYaxWnqzK8XBvmXp7Pvqd07lOE6vGhzoR7vwUHq0bMDl0S1pF34mATSoq6ObKlWdaGKo4fIcTv7Yd5zFsYmkZuSSfxxujMFgleXG5TkGDIb8fo3G5Xn+8hR4bghzpDA0/UeGpX9P87x4TkkIv9YdyRwzkJ9S2uNIsVoGiRwksqFV9TOgbWPr6D88lHYRIbSoH4yfVgMpVSNoYqiBsvMcrNydzPdbDvPTtkRSMnIJDvQjPDTodAMfQRABAUQEsSYWeF54WVzm+eOgb14Ml2b/yOCc3wnAwdaAHswKm8TKoKEQUI/Wjetxf0QIHSJCaBceStsm9QgO1OajStV0mhhqiMwcB8t3JrFoy2GWxB4lLTuP0KAALurWlNE9mzOsc1Pq1nFDoZwaDxs+hQ2fQNpBqNcE+t8B/f5Mj4gu9ACmVf5VlFLVmCaGaiw9O4+l24+yaMthft6eRGaug4b1AhnVszmjezVnaMdw93TwcuTCzh9g/UfWBWXjhPbD4eKnoesYCAiq/GsopWoMTQzVTGpGLj/FJrJoy2FW7DpGTp6T8NAgruwXyeieLRjcvjGB/m5qqXN8L6z/GGI+t5qYhrWA8+6HfjdYPZKVUj5JE0M1cCw9mx+3JvL9lsOs2pNMntPQskEw1w1uw+ieLejftpH72u/nZVuD1a3/yLrBjfhBp0uh35+tQev89SuhlK/TUsBLjqRmsWjLYb7fcoQ1ccdxGmjbpB43n9+O0T1bEN2qgXtvNHN0u5UMNs6EzBRo2AZGPAZ9rwO97aVSyoUmhip08HgG39vJYMOBEwB0ahrKXSM6MqpnC7q1CHNvMsg5BVvnWNVFB/+weiN3HQP9p0C74eCnnceUUmfzmcRw9GQWWxNOEujvR6C/EBjgRx1/PwL8hUB/67HrvEA/67G/n1SqsN59NI1FW47w/ZYjbE04CUCPlvV54JLOjOrZgo5NQ0vZQgUkxFhnB5u/huyT0KQTXPwMRE+G0Aj3v55SqlbxmcSwOu44d32+odzrieCSOIQAl8eBrskk/3GAH3Xs57uPprPraDoAfds05NHLujKqRwvaNKnn7t2zbnaz+SurV/KRTRAQDN0nWGcHbc4tdQRTpZTK5zOJ4byO4cz56xDynIbcPCc5Die5DkOuw2n/nXmck3fmeZ7DSU6B5Zzk5NnznGce5/9lZDrIsx+HhwZx3eA2XNqzOS0a1PXMjiXvgd/ftloW5Z6yhrS+7GXrhjd1G3nmNZVStZrPJIaG9erQt83ZdwSrkYyxWhT9/pbV/8A/EHpeDYP+Ai376dmBUqpSfCYx1Aq5WbDla+sMIXEL1AuHYQ/CgJsLDGutlFKVoYmhJkg/Cmveg7XvwakkaNodxv0Xel0DgXrTGqWUe2liqM6ObLbODjZ/BY4cqyPauX+FdsO0ukgp5TGaGKobpxN2LrKuH8T9AoH1rF7Jg++A8I7ejk4p5QM0MVQX2ekQ8xn8Md0aw6h+JIz8p9XcVFsXKaWqkCYGbztxAP74H6z/xLo9ZuQAuPox6DbOam2klBvk5uZy6NAhsrKyvB2K8rDg4GBatWpFYGDFyw9NDN5gDBxcDb+/aQ1oh0D3cXDOndB6oLejU7XQoUOHCAsLIyoqyr3DrqhqxRhDcnIyhw4dol27dhXejiaGquTIhW3zrOsH8esguAGcexcMuhUatvZ2dKoWy8rK0qTgA0SEJk2akJSUVKntaGKoChnHYd2HsPodSEuAxh2s3snRkyHIA2MlKVUETQq+wR2fsyYGT0raCX+8DTEzIS/TamY69j/WfQ90ZFOlVDWlpZMn5GTA3L/CmwNhw2fQ6yq4/TeYMh+6jNKkoHzOiBEj+OGHHwpMe/XVV7njjjuKXWf48OGsXbsWgMsuu4wTJ06ctcxTTz3Fyy+/XOJrz507l23btp1+/sQTT7B48eJyRF+0ZcuWMXbs2EpvpzryaAklIqNEZIeI7BaRh4tZ5loR2SYiW0Xkc0/GUyVS9sP7l1qD2g29F+7bCuPfhOY9vR2ZUl4zefJkZs2aVWDarFmzmDx5cpnWX7hwIQ0bNqzQaxdODE8//TQjR46s0LZ8hceqkkTEH3gTuBg4BKwRkfnGmG0uy3QCHgGGGmNSRKSpp+KpEnuXwVc3gtMBf/oCOl/q7YiUOss/v93KNvveIO7SvWV9nry8R7Hzr776ah577DFycnKoU6cOcXFxJCQkcP7553PHHXewZs0aMjMzufrqq/nnP/951vpRUVGsXbuW8PBwnnvuOT766COaNm1K69at6d+/PwDvvPMOM2bMICcnh44dO/LJJ58QExPD/PnzWb58Oc8++yyzZ8/mmWeeYezYsVx99dUsWbKEBx54gLy8PAYOHMjbb79NUFAQUVFRTJkyhW+//Zbc3Fy++uorunbtWqb3YubMmTz//PMYYxgzZgwvvfQSDoeDm2++mbVr1yIi3HTTTdx33328/vrrTJ8+nYCAALp3735W8vQWT54xDAJ2G2P2GmNygFnA+ELL3AK8aYxJATDGHPVgPJ5jDKx8Az65AkKbwq0/a1JQykXjxo0ZNGgQ33//PWCdLVx77bWICM899xxr165l06ZNLF++nE2bNhW7nXXr1jFr1ixiYmJYuHAha9asOT3vyiuvZM2aNWzcuJFu3brx3nvvMWTIEMaNG8e//vUvYmJi6NChw+nls7KymDp1Kl988QWbN28mLy+Pt99++/T88PBw1q9fzx133FFqdVW+hIQEHnroIZYuXUpMTAxr1qxh7ty5xMTEEB8fz5YtW9i8eTM33ngjAC+++CIbNmxg06ZNTJ8+vVzvqSd58uJzJHDQ5fkhYHChZToDiMhvgD/wlDFmUeENicitwK0Abdq08UiwFZZzCubfDVtmW53SJrwFQWHejkqpYpV0ZO9J+dVJ48ePZ9asWbz33nsAfPnll8yYMYO8vDwOHz7Mtm3b6N27d5Hb+OWXX7jiiiuoV8+62dW4ceNOz9uyZQuPPfYYJ06cID09nUsvLfngbMeOHbRr147OnTsDMGXKFN58802mTZsGWIkGoH///nzzzTdl2sc1a9YwfPhwIiKsOyVed911rFixgscff5y9e/dy9913M2bMGC655BIAevfuzXXXXceECROYMGFCmV6jKnj7KmgA0AkYDkwG3hGRhoUXMsbMMMYMMMYMyH/Dq4Xj++C9S2DLN3DRE3Dtx5oUlCrG+PHjWbJkCevXrycjI4P+/fuzb98+Xn75ZZYsWcKmTZsYM2ZMhXtnT506lf/+979s3ryZJ598stK9vIOCggDw9/cnLy+vUttq1KgRGzduZPjw4UyfPp2//OUvAHz33XfceeedrF+/noEDB1b6ddzFk4khHnDttdXKnubqEDDfGJNrjNkH7MRKFNXf7iUwYzikHoTrvobz/6YjnipVgtDQUEaMGMFNN910+qLzyZMnCQkJoUGDBiQmJp6uairOBRdcwNy5c8nMzCQtLY1vv/329Ly0tDRatGhBbm4un3322enpYWFhpKWlnbWtLl26EBcXx+7duwH45JNPGDZsWKX2cdCgQSxfvpxjx47hcDiYOXMmw4YN49ixYzidTq666iqeffZZ1q9fj9Pp5ODBg4wYMYKXXnqJ1NRU0tPTK/X67uLJqqQ1QCcRaYeVECYBfyq0zFysM4UPRCQcq2pprwdjqjxj4LdXYcnTENENJn0Kjdt7OyqlaoTJkydzxRVXnL7IGh0dTd++fenatSutW7dm6NChJa7fr18/Jk6cSHR0NE2bNmXgwDNDyDzzzDMMHjyYiIgIBg8efDoZTJo0iVtuuYXXX3+dr7/++vTywcHBfPDBB1xzzTWnLz7ffvvt5dqfJUuW0KpVq9PPv/rqK1588UVGjBhx+uLz+PHj2bhxIzfeeCNOpxOAF154AYfDwfXXX09qairGGO65554Kt7xyNzHGeG7jIpcBr2JdP3jfGPOciDwNrDXGzBeri96/gVGAA3jOGFPiZfkBAwaY/LbNVS47HebdCdvmQo8rrGaodUK8E4tS5RAbG0u3bt28HYaqIkV93iKyzhgzoCzre7TnszFmIbCw0LQnXB4b4H77r3pL3gNfXA9J2+Hip2HIPVp1pJSqlXRIjLLY9RPMvhnED66fDR0u9HZESinlMZoYSmIM/PJvWPosNOtpXU9oFOXtqJRSyqM0MRQnOw3m3mHdL6Hn1TDuDahTz9tRKaWUx2liKMqx3TDrT5C8Cy55Ds69U68nKKV8hiaGwnYsgm9uAb8AuGEutK9cu2allKppvN3zufpwOmHZSzBzonUd4bblmhSUcoPk5GT69OlDnz59aN68OZGRkaef5+TklLju2rVrueeee0p9jSFDhrgrXACmTZtGZGTk6X4HvkbPGACyTsKc22HHd9B7Elz+KgTW9XZUStUKTZo0ISYmBrDunxAaGsoDDzxwen5eXh4BAUUXRQMGDGDAgNKb3q9cudItsQI4nU7mzJlD69atWb58OSNGjHDbtl2VtN/eVj2jqkpJO63rCcf3wqiXYPBtej1B1W7fPwxHNrt3m817wegXy7z41KlTCQ4OZsOGDQwdOpRJkyZx7733kpWVRd26dfnggw/o0qULy5Yt4+WXX2bBggU89dRTHDhwgL1793LgwAGmTZt2+mwiNDSU9PR0li1bxlNPPUV4eDhbtmyhf//+fPrpp4gICxcu5P777yckJIShQ4eyd+9eFixYcFZsy5Yto0ePHkycOJGZM2eeTgyJiYncfvvt7N1rDc7w9ttvM2TIED7++GNefvllRITevXvzySefMHXq1NNDexeO7/HHH6dRo0Zs376dnTt3MmHCBA4ePEhWVhb33nsvt956KwCLFi3i0UcfxeFwEB4ezk8//USXLl1YuXIlEREROJ1OOnfuzKpVq3D3GHK+nRi2fwff3AYBQdbd1aLO83ZESvmMQ4cOsXLlSvz9/Tl58iS//PILAQEBLF68mEcffZTZs2eftc727dv5+eefSUtLo0uXLtxxxx0EBgYWWGbDhg1s3bqVli1bMnToUH777TcGDBjAbbfdxooVK2jXrl2JNwiaOXMmkydPZvz48Tz66KPk5uYSGBjIPffcw7Bhw5gzZw4Oh4P09HS2bt3Ks88+y8qVKwkPD+f48eOl7vf69evZsmUL7dq1A+D999+ncePGZGZmMnDgQK666iqcTie33HLL6XiPHz+On58f119/PZ999hnTpk1j8eLFREdHuz0pgK8mBqcTlr8Iy1+Cln1h4qfQoFXp6ylVG5TjyN6TrrnmGvz9/QFITU1lypQp7Nq1CxEhNze3yHXGjBlDUFAQQUFBNG3alMTExAJjFYE1kF3+tD59+hAXF0doaCjt27c/XRhPnjyZGTNmnLX9nJwcFi5cyCuvvEJYWBiDBw/mhx9+YOzYsSxdupSPP/4YsEZcbdCgAR9//DHXXHMN4eHhgHXfidIMGjTodBwAr7/+OnPmzAHg4MGD7Nq1i6SkJC644ILTy+Vv96abbmL8+PFMmzaN999///R9HdzN9xJD5gmYcxvsXAR9roMxr0BgsLejUsrnhIScGWfs8ccfZ8SIEcyZM4e4uDiGDx9e5Dr5Q2FD8cNhl2WZ4vzwww+cOHGCXr16AZCRkUHdunXLfW/ngICA0xeunU5ngYvsrvu9bNkyFi9ezKpVq6hXrx7Dhw8vcbjw1q1b06xZM5YuXcrq1asLjCLrTr7VKunodnjnQti9GC572RoET5OCUl6XmppKZGQkAB9++KHbt9+lSxf27t1LXFwcAF988UWRy82cOZN3332XuLg44uLi2LdvHz/99BMZGRlcdNFFp+/w5nA4SE1N5cILL+Srr74iOTkZ4HRVUlRUFOvWrQNg/vz5xZ4Bpaam0qhRI+rVq8f27dv5/fffATjnnHNYsWIF+/btK7BdgL/85S9cf/31Bc643M13EsPuxfDuRZB9EqZ8C4Nu0YvMSlUTDz74II888gh9+/b1yM1q6taty1tvvcWoUaPo378/YWFhNGjQoMAyGRkZLFq0iDFjxpyeFhISwnnnnce3337La6+9xs8//0yvXr3o378/27Zto0ePHvzjH/9g2LBhREdHc//91nigt9xyC8uXLyc6OppVq1YVOEtwNWrUKPLy8ujWrRsPP/ww55xzDgARERHMmDGDK6+8kujoaCZOnHh6nXHjxpGenu6xaiTw8LDbnlDhYbeT98DCv1tDWzSIdH9gSlVjOuw2pKenExoaijGGO++8k06dOnHfffd5O6xyW7t2Lffddx+//PJLsctUdtht3zljaNIBbvhGk4JSPuqdd96hT58+9OjRg9TUVG677TZvh1RuL774IldddRUvvPCCR1/Hd84YlPJhesbgW/SMQSlVJjXtIFBVjDs+Z00MSvmA4OBgkpOTNTnUcsYYkpOTCQ6uXGtL3+vHoJQPatWqFYcOHSIpKcnboSgPCw4OPqvTX3lpYlDKBwQGBhbobatUSbQqSSmlVAGaGJRSShWgiUEppVQBNa4fg4gkAfsruHo4cMyN4VQ3tXn/dN9qrtq8fzVp39oaY8o0RneNSwyVISJry9rBoyaqzfun+1Zz1eb9q637plVJSimlCtDEoJRSqgBfSwxn37KpdqnN+6f7VnPV5v2rlfvmU9cYlFJKlc7XzhiUUkqVQhODUkqpAnwmMYjIKBHZISK7ReRhb8fjLiLSWkR+FpFtIrJVRO71dkzuJiL+IrJBRBZ4OxZ3E5GGIvK1iGwXkVgROdfbMbmLiNxnfye3iMhMEanRN1gXkfdF5KiIbHGZ1lhEfhKRXfb/Rt6M0V18IjGIiD/wJjAa6A5MFpHu3o3KbfKAvxljugPnAHfWon3Ldy8Q6+0gPOQ1YJExpisQTS3ZTxGJBO4BBhhjegL+wCTvRlVpHwKjCk17GFhijOkELLGf13g+kRiAQcBuY8xeY0wOMAsY7+WY3MIYc9gYs95+nIZVsNSa+5eKSCtgDPCut2NxNxFpAFwAvAdgjMkxxpzwalDuFQDUFZEAoB6Q4OV4KsUYswI4XmjyeOAj+/FHwISqjMlTfCUxRAIHXZ4fohYVnvlEJAroC/zh5VDc6VXgQcDp5Tg8oR2QBHxgV5W9KyIh3g7KHYwx8cDLwAHgMJBqjPnRu1F5RDNjzGH78RGgmTeDcRdfSQy1noiEArOBacaYk96Oxx1EZCxw1BizztuxeEgA0A942xjTFzhFLamKsOvax2Mlv5ZAiIhc792oPMtYbf9rRft/X0kM8UBrl+et7Gm1gogEYiWFz4wx33g7HjcaCowTkTis6r8LReRT74bkVoeAQ8aY/DO8r7ESRW0wEthnjEkyxuQC3wBDvByTJySKSAsA+/9RL8fjFr6SGNYAnUSknYjUwboINt/LMbmFiAhWHXWsMeYVb8fjTsaYR4wxrYwxUVif2VJjTK056jTGHAEOikgXe9JFwDYvhuROB4BzRKSe/R29iFpyYb2Q+cAU+/EUYJ4XY3Ebn7i1pzEmT0TuAn7Aah3xvjFmq5fDcpehwA3AZhGJsac9aoxZ6L2QVDncDXxmH7DsBW70cjxuYYz5Q0S+BtZjtZzbQA0fPkJEZgLDgXAROQQ8CbwIfCkiN2PdDuBa70XoPjokhlJKqQJ8pSpJKaVUGWliUEopVYAmBqWUUgVoYlBKKVWAJgallFIFaGJQqhARcYhIjMuf23oji0iU6+icSlVHPtGPQalyyjTG9PF2EEp5i54xKFVGIhInIv8nIptFZLWIdLSnR4nIUhHZJCJLRKSNPb2ZiMwRkY32X/6QEP4i8o59r4IfRaSu13ZKqSJoYlDqbHULVSVNdJmXaozpBfwXa+RXgDeAj4wxvYHPgNft6a8Dy40x0VhjIOX3tu8EvGmM6QGcAK7y6N4oVU7a81mpQkQk3RgTWsT0OOBCY8xee+DCI8aYJiJyDGhhjMm1px82xoSLSBLQyhiT7bKNKOAn+8YuiMhDQKAx5tkq2DWlykTPGJQqH1PM4/LIdnnsQK/1qWpGE4NS5TPR5f8q+/FKzty28jrgF/vxEuAOOH3f6gZVFaRSlaFHKkqdra7LSLVg3ZM5v8lqIxHZhHXUP9medjfWXdj+jnVHtvwRUu8FZtgjbzqwksRhlKrm9BqDUmVkX2MYYIw55u1YlPIkrUpSSilVgJ4xKKWUKkDPGJRSShWgiUEppVQBmhiUUkoVoIlBKaVUAZoYlFJKFfD/2UxP/Z3TrbMAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot Charts\n",
    "plot_valloss_vs_train_accuracy(rnn)\n",
    "plot_accuracy(rnn)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}