Yelp_NLP_RNN_RA.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyML6xVut+4balUnQXJuvnuI",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU",
    "gpuClass": "standard"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/EAName/yelp-reviews/blob/main/Yelp_NLP_RNN_RA.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "id": "A4cNq1ya5nVW"
      },
      "outputs": [],
      "source": [
        "from packaging import version\n",
        "\n",
        "import tensorflow as tf\n",
        "import tensorflow_datasets as tfds\n",
        "from tensorflow import keras\n",
        "from tensorflow.keras import layers\n",
        "from tensorflow.keras.layers import TextVectorization\n",
        "\n",
        "\n",
        "import pandas as pd\n",
        "import numpy as np\n",
        "import re\n",
        "import string\n",
        "import matplotlib.pyplot as plt\n",
        "import matplotlib as mpl\n",
        "import seaborn as sns\n",
        "import statistics\n",
        "\n",
        "%matplotlib inline\n",
        "\n",
        "from sklearn.metrics import confusion_matrix, classification_report\n",
        "from sklearn.preprocessing import StandardScaler\n",
        "from sklearn.decomposition import PCA\n",
        "from sklearn.manifold import TSNE\n",
        "from sklearn.ensemble import RandomForestClassifier\n",
        "from sklearn.metrics import mean_squared_error as MSE\n",
        "from sklearn.metrics import mean_absolute_error\n",
        "from sklearn.metrics import accuracy_score\n",
        "from sklearn.metrics import roc_auc_score, accuracy_score, confusion_matrix, silhouette_score\n",
        "from sklearn.metrics import precision_score, recall_score, f1_score, classification_report\n",
        "\n",
        "from sklearn.model_selection import train_test_split\n",
        "\n",
        "from datetime import datetime\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "B9P82kqa5nVX"
      },
      "source": [
        "## Verify TensorFlow version and Keras version"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VegoVCje5nVY",
        "outputId": "249ae531-4dfb-48aa-87aa-d3bc7adc6267"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "This notebook requires TensorFlow 2.0 or above\n",
            "TensorFlow version:  2.9.2\n"
          ]
        }
      ],
      "source": [
        "print(\"This notebook requires TensorFlow 2.0 or above\")\n",
        "print(\"TensorFlow version: \", tf.__version__)\n",
        "assert version.parse(tf.__version__).release[0] >=2"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "QFhc0Jb-5nVZ",
        "outputId": "961a7f11-3858-4863-c60a-e4d6ce68daf0"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Keras version:  2.9.0\n"
          ]
        }
      ],
      "source": [
        "print(\"Keras version: \", keras.__version__)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kdlCk2h65nVZ"
      },
      "source": [
        "## Mount Google Drive to Colab environment"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "metadata": {
        "id": "3_tETfHF5nVZ",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "a68b8a8e-0053-4257-e8b1-ae04986dcfcf"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/gdrive\n"
          ]
        }
      ],
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/gdrive')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4l7SoL-w5nVa"
      },
      "source": [
        "## Load Reviews Dataset"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df =pd.read_csv('/content/gdrive/MyDrive/Data/yelp_reviews_clean_version4_st_louis.csv')"
      ],
      "metadata": {
        "id": "p2KLuijTcWil"
      },
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df = df.iloc[:,3:]\n",
        "df['stars'] = df['stars'].replace([1.0, 2.0, 3.0, 4.0, 5.0], [1, 2, 3, 4, 5])\n",
        "df['stars'] = df['stars'].astype(int)"
      ],
      "metadata": {
        "id": "olWIlW2bYj1D"
      },
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "train, test = train_test_split(df, test_size=0.7, random_state = 42)"
      ],
      "metadata": {
        "id": "VnFOdkX9eoAS"
      },
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "test, val = train_test_split(test, test_size=0.5, random_state = 42)"
      ],
      "metadata": {
        "id": "-lxDrBSCey37"
      },
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df2 = train\n",
        "labels = df2.pop('stars')\n",
        "train_ds = tf.data.Dataset.from_tensor_slices((df2, labels))\n",
        "train_ds = train_ds.batch(32)"
      ],
      "metadata": {
        "id": "cyn122OopHnt"
      },
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df2 = test\n",
        "labels = df2.pop('stars')\n",
        "test_ds = tf.data.Dataset.from_tensor_slices((df2, labels))\n",
        "test_ds = test_ds.batch(32)"
      ],
      "metadata": {
        "id": "Bznxwyt_pHqI"
      },
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df2 = val\n",
        "labels = df2.pop('stars')\n",
        "val_ds = tf.data.Dataset.from_tensor_slices((df2, labels))\n",
        "val_ds = val_ds.batch(32)"
      ],
      "metadata": {
        "id": "MKClyp7EpURz"
      },
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def plot_history(history):\n",
        "  losses = history.history['loss']\n",
        "  accs = history.history['accuracy']\n",
        "  val_losses = history.history['val_loss']\n",
        "  val_accs = history.history['val_accuracy']\n",
        "  epochs = len(losses)\n",
        "\n",
        "  plt.figure(figsize=(16, 4))\n",
        "  for i, metrics in enumerate(zip([losses, accs], [val_losses, val_accs], ['Loss', 'Accuracy'])):\n",
        "    plt.subplot(1, 2, i + 1)\n",
        "    plt.plot(range(epochs), metrics[0], label='Training {}'.format(metrics[2]))\n",
        "    plt.plot(range(epochs), metrics[1], label='Validation {}'.format(metrics[2]))\n",
        "    plt.legend()\n",
        "  plt.show()"
      ],
      "metadata": {
        "id": "yzHGgRJ6JRuD"
      },
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def print_validation_report(test_labels, predictions):\n",
        "    print(\"Classification Report\")\n",
        "    print(classification_report(test_labels, predictions))\n",
        "    print('Accuracy Score: {:.3f}'.format(accuracy_score(test_labels, predictions)))\n",
        "    print('Mean Absolute Error: {:.3f}'.format(mean_absolute_error(test_labels, predictions)))\n",
        "    print('Root Mean Square Error: {:.3f}'.format(np.sqrt(MSE(test_labels, predictions))))\n",
        "    print('Mean Square Error: {:.3f}'.format(MSE(test_labels, predictions, squared = True)))"
      ],
      "metadata": {
        "id": "lyFZe1XEZc_9"
      },
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def plot_confusion_matrix(y_true, y_pred):\n",
        "    mtx = confusion_matrix(y_true, y_pred)\n",
        "    fig, ax = plt.subplots(figsize=(8,8))\n",
        "    sns.heatmap(mtx, annot=True, fmt='d', linewidths=.75,  cbar=False, ax=ax,cmap='Blues',linecolor='white')\n",
        "    #  square=True,\n",
        "    plt.ylabel('true label')\n",
        "    plt.xlabel('predicted label')\n",
        "    ax.xaxis.set_ticklabels(['1', '2', '3', '4', '5'])\n",
        "    ax.yaxis.set_ticklabels(['1', '2', '3', '4', '5'])"
      ],
      "metadata": {
        "id": "4tFlVmj7Z8Gh"
      },
      "execution_count": 27,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mq7Q6zlFRro_"
      },
      "source": [
        "## Preparing Integer Sequence Datasets"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 16,
      "metadata": {
        "id": "AfTnCGIERrpA"
      },
      "outputs": [],
      "source": [
        "max_length = 150\n",
        "max_tokens = 1000\n",
        "text_vectorization = layers.TextVectorization(\n",
        "    max_tokens=max_tokens,\n",
        "    output_mode=\"int\",\n",
        "    output_sequence_length=max_length,\n",
        ")\n",
        "\n",
        "text_only_train_ds = train_ds.map(lambda x, y: x)\n",
        "text_vectorization.adapt(text_only_train_ds)\n",
        "\n",
        "int_train_ds = train_ds.map(\n",
        "    lambda x, y: (text_vectorization(x), y),\n",
        "    num_parallel_calls=4)\n",
        "int_val_ds = val_ds.map(\n",
        "    lambda x, y: (text_vectorization(x), y),\n",
        "    num_parallel_calls=4)\n",
        "int_test_ds = test_ds.map(\n",
        "    lambda x, y: (text_vectorization(x), y),\n",
        "    num_parallel_calls=4)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "test_data = [(example.numpy(), label.numpy()) for example, label in int_test_ds]\n",
        "test_y = []\n",
        "\n",
        "for i in range(len(test)//32):\n",
        "  for j in range(32):\n",
        "    test_y.append(test_data[i][1][j])\n",
        "\n",
        "for j in range(len(test)%32):\n",
        "  test_y.append(test_data[len(test)//32][1][j])\n",
        "\n",
        "test_y = np.array(test_y)"
      ],
      "metadata": {
        "id": "Lpq3bXDLZ1kp"
      },
      "execution_count": 17,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NBL0CkaB5ukv"
      },
      "source": [
        "## Bi-directional RNN\n",
        "\n",
        "When translating in real-time, it would help to have access to worlds towards the end of a sentence, say, as well as earlier words in the sentence. One way to use the later words in the sentence is to feed the words into our RNN backward. So if we create two independent RNNs, we can feed one the words in their forward, or natural order, and the second gets their words in the revser order. This is the idea behind `Bi-directional RNNS`:"
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "5hj8micpzQqo"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qN2wjSjdRrpB"
      },
      "source": [
        "## Sequence Model Built on One-Hot Encoded Vector Sequences"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Hq6dB-zWRrpB",
        "outputId": "d3f97a7e-500f-4adb-ee00-7b58fda1194a"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " input_1 (InputLayer)        [(None, None)]            0         \n",
            "                                                                 \n",
            " tf.one_hot (TFOpLambda)     (None, None, 1000)        0         \n",
            "                                                                 \n",
            " bidirectional (Bidirectiona  (None, 64)               264448    \n",
            " l)                                                              \n",
            "                                                                 \n",
            " dropout (Dropout)           (None, 64)                0         \n",
            "                                                                 \n",
            " dense (Dense)               (None, 4)                 260       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 264,708\n",
            "Trainable params: 264,708\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "inputs = tf.keras.Input(shape=(None,), dtype=\"int64\")\n",
        "embedded = tf.one_hot(inputs, depth=max_tokens)\n",
        "x = layers.Bidirectional(layers.LSTM(32))(embedded)\n",
        "x = layers.Dropout(0.5)(x)\n",
        "outputs = layers.Dense(4, activation=\"softmax\")(x)\n",
        "model = tf.keras.Model(inputs, outputs)\n",
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"SparseCategoricalCrossentropy\",\n",
        "              metrics=[\"accuracy\"])\n",
        "model.summary()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ognTeJM6RrpC"
      },
      "source": [
        "## Training Sequence Model"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "5kCbHrrsRrpC",
        "outputId": "24f0000b-c70f-41f2-9207-29200fffd1b7"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/200\n",
            "3563/3563 [==============================] - 62s 15ms/step - loss: 0.5771 - accuracy: 0.7939 - val_loss: 0.4150 - val_accuracy: 0.8595\n",
            "Epoch 2/200\n",
            "3563/3563 [==============================] - 53s 15ms/step - loss: 0.4418 - accuracy: 0.8511 - val_loss: 0.4024 - val_accuracy: 0.8605\n",
            "Epoch 3/200\n",
            "3563/3563 [==============================] - 53s 15ms/step - loss: 0.4247 - accuracy: 0.8561 - val_loss: 0.3980 - val_accuracy: 0.8620\n",
            "Epoch 4/200\n",
            "3563/3563 [==============================] - 53s 15ms/step - loss: 0.4155 - accuracy: 0.8592 - val_loss: 0.3957 - val_accuracy: 0.8638\n",
            "Epoch 5/200\n",
            "3563/3563 [==============================] - 53s 15ms/step - loss: 0.4074 - accuracy: 0.8616 - val_loss: 0.3998 - val_accuracy: 0.8643\n",
            "Epoch 6/200\n",
            "3563/3563 [==============================] - 53s 15ms/step - loss: 0.4018 - accuracy: 0.8638 - val_loss: 0.3946 - val_accuracy: 0.8643\n",
            "Epoch 7/200\n",
            "3563/3563 [==============================] - 53s 15ms/step - loss: 0.3967 - accuracy: 0.8658 - val_loss: 0.3916 - val_accuracy: 0.8633\n",
            "Epoch 8/200\n",
            "3563/3563 [==============================] - 53s 15ms/step - loss: 0.3935 - accuracy: 0.8676 - val_loss: 0.3953 - val_accuracy: 0.8633\n",
            "238/238 [==============================] - 3s 9ms/step - loss: 0.4050 - accuracy: 0.8570\n",
            "Test acc: 0.857\n"
          ]
        }
      ],
      "source": [
        "callbacks = [\n",
        "    tf.keras.callbacks.ModelCheckpoint(\"one_hot_bidir_lstm.keras\",\n",
        "                                    save_best_only=True)\n",
        "    ,tf.keras.callbacks.EarlyStopping(monitor='val_accuracy', patience=3)\n",
        "]\n",
        "model.fit(int_train_ds, validation_data=int_val_ds, epochs=200, callbacks=callbacks)\n",
        "model = keras.models.load_model(\"one_hot_bidir_lstm.keras\")\n",
        "print(f\"Test acc: {model.evaluate(int_test_ds)[1]:.3f}\")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "im1j5PRE5uk0"
      },
      "source": [
        "## Two ways to obtain word embeddings\n",
        "\n",
        "1. `Learn word embeddings jointly with the main task you care about` (such as document classification or sentiment prediction). In this setup, you start with random word vectors and then learn word vectors in the same way you learn the weights of a neural network.\n",
        "2. Load into your model word embeddings that were precomputed using a different machine learning task than the one you’re trying to solve. These are called `pretrained word embeddings`.\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NrVeqKcsRrpE"
      },
      "source": [
        "## Instantiating An Embedding Layer\n",
        "\n",
        "The Embedding layer takes at least two arguments: the number of possible tokens and the dimensionality of the embeddings (here, 256)."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 18,
      "metadata": {
        "id": "WLTjxbMnRrpE"
      },
      "outputs": [],
      "source": [
        "embedding_layer = layers.Embedding(input_dim=max_tokens, output_dim=256)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vWx26f9C5uk2"
      },
      "source": [
        "The Embedding layer is best understood as a dictionary that maps integer indices (which stand for specific words) to dense vectors"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7Mox7FQp5uk2"
      },
      "source": [
        "The Embedding layer takes as input a rank-2 tensor of integers, of shape `(batch_size, sequence_length)`, where each entry is a sequence of integers. The layer then returns a 3D floating-point tensor of shape `(batch_size, sequence_length, embedding_ dimensionality)`.Again, embedding_ dimensionality is 256 above."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PSfWWNO6RrpE"
      },
      "source": [
        "## Model Leveraging Trained Embedding Layer"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FLuGKCM35uk2"
      },
      "source": [
        "One input is a sequence of integers.\n",
        "1. Encode the integers into binary 20,000-dimensional vectors.\n",
        "2. Add a bidirectional LSTM.\n",
        "3. Finally, add a classification layer."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 20,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 780
        },
        "id": "O6JFON-CRrpF",
        "outputId": "92eff930-821b-42f9-d426-4999d0fd4ab2"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model_1\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " input_2 (InputLayer)        [(None, None)]            0         \n",
            "                                                                 \n",
            " embedding_2 (Embedding)     (None, None, 256)         256000    \n",
            "                                                                 \n",
            " bidirectional_1 (Bidirectio  (None, 64)               73984     \n",
            " nal)                                                            \n",
            "                                                                 \n",
            " dropout_1 (Dropout)         (None, 64)                0         \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 4)                 260       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 330,244\n",
            "Trainable params: 330,244\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "Epoch 1/20\n",
            "2741/2741 [==============================] - 68s 24ms/step - loss: nan - accuracy: 1.1401e-05 - val_loss: nan - val_accuracy: 0.0000e+00\n"
          ]
        },
        {
          "output_type": "error",
          "ename": "KeyboardInterrupt",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-20-22d6f3bce0d8>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     15\u001b[0m     \u001b[0;34m,\u001b[0m\u001b[0mtf\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mkeras\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcallbacks\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mEarlyStopping\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mmonitor\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m'val_accuracy'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mpatience\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m3\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     16\u001b[0m ]\n\u001b[0;32m---> 17\u001b[0;31m \u001b[0mmodel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mfit\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mint_train_ds\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mvalidation_data\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mint_val_ds\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mepochs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m20\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcallbacks\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mcallbacks\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     18\u001b[0m \u001b[0mmodel\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mkeras\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmodels\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mload_model\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"embeddings_bidir_gru.keras\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     19\u001b[0m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34mf\"Test acc: {model.evaluate(int_test_ds)[1]:.3f}\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/keras/utils/traceback_utils.py\u001b[0m in \u001b[0;36merror_handler\u001b[0;34m(*args, **kwargs)\u001b[0m\n\u001b[1;32m     62\u001b[0m     \u001b[0mfiltered_tb\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     63\u001b[0m     \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 64\u001b[0;31m       \u001b[0;32mreturn\u001b[0m \u001b[0mfn\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m*\u001b[0m\u001b[0margs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;34m**\u001b[0m\u001b[0mkwargs\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     65\u001b[0m     \u001b[0;32mexcept\u001b[0m \u001b[0mException\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0me\u001b[0m\u001b[0;34m:\u001b[0m  \u001b[0;31m# pylint: disable=broad-except\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     66\u001b[0m       \u001b[0mfiltered_tb\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0m_process_traceback_frames\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0me\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m__traceback__\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.7/dist-packages/keras/engine/training.py\u001b[0m in \u001b[0;36mfit\u001b[0;34m(self, x, y, batch_size, epochs, verbose, callbacks, validation_split, validation_data, shuffle, class_weight, sample_weight, initial_epoch, steps_per_epoch, validation_steps, validation_batch_size, validation_freq, max_queue_size, workers, use_multiprocessing)\u001b[0m\n\u001b[1;32m   1395\u001b[0m       \u001b[0mlogs\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mNone\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1396\u001b[0m       \u001b[0;32mfor\u001b[0m \u001b[0mepoch\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0miterator\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mdata_handler\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0menumerate_epochs\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1397\u001b[0;31m         \u001b[0mself\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mreset_metrics\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m   1398\u001b[0m         \u001b[0mcallbacks\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mon_epoch_begin\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mepoch\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1399\u001b[0m         \u001b[0;32mwith\u001b[0m \u001b[0mdata_handler\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcatch_stop_iteration\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
          ]
        }
      ],
      "source": [
        "inputs = tf.keras.Input(shape=(None,), dtype=\"int64\")\n",
        "embedded = layers.Embedding(input_dim=max_tokens, output_dim=256)(inputs)\n",
        "x = layers.Bidirectional(layers.LSTM(32))(embedded)\n",
        "x = layers.Dropout(0.5)(x)\n",
        "outputs = layers.Dense(6, activation=\"softmax\")(x)\n",
        "model = tf.keras.Model(inputs, outputs)\n",
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"SparseCategoricalCrossentropy\",\n",
        "              metrics=[\"accuracy\"])\n",
        "model.summary()\n",
        "\n",
        "callbacks = [\n",
        "    tf.keras.callbacks.ModelCheckpoint(\"embeddings_bidir_gru.keras\",\n",
        "                                    save_best_only=True)\n",
        "    ,tf.keras.callbacks.EarlyStopping(monitor='val_accuracy', patience=3)\n",
        "]\n",
        "model.fit(int_train_ds, validation_data=int_val_ds, epochs=20, callbacks=callbacks)\n",
        "model = keras.models.load_model(\"embeddings_bidir_gru.keras\")\n",
        "print(f\"Test acc: {model.evaluate(int_test_ds)[1]:.3f}\")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "M0gOTdNnRrpF"
      },
      "source": [
        "## Understanding Padding and Masking"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zOudeXoa5uk4"
      },
      "source": [
        "One thing that’s slightly hurting model performance here is that our input sequences are full of zeros. This comes from our use of the `output_sequence_length=max_ length` option in TextVectorization (with `max_length equal` to 150): sentences longer than 150 tokens are truncated to a length of 150 tokens, and sentences shorter than 150 tokens are padded with zeros at the end so that they can be concatenated together with other sequences to form contiguous batches."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hM95tdc65uk4"
      },
      "source": [
        "The RNN that looks at the tokens in their natural order will spend its last iterations seeing only vectors that encode padding—possibly for several hundreds of iterations if the original sentence was short. The information stored in the internal state of the RNN will gradually fade out as it gets exposed to these meaningless inputs."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "U93_Szgn5uk4"
      },
      "source": [
        "We need some way to tell the RNN that it should skip these iterations. There’s an API for that: `masking`."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "t1SxQSW85uk4"
      },
      "source": [
        "<div class=\"alert alert-block alert-success\"><b>tf.keras.layers.Masking</b><br>\n",
        "https://www.tensorflow.org/api_docs/python/tf/keras/layers/Masking</div>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HGe0pvoxRrpF"
      },
      "source": [
        "## Model Leveraging Embedding Layer With Masking Enabled"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 21,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "tU6YKNrDRrpF",
        "outputId": "0e57767c-8f9f-443d-932f-131a27953b9a"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model_2\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " input_3 (InputLayer)        [(None, None)]            0         \n",
            "                                                                 \n",
            " embedding_3 (Embedding)     (None, None, 256)         256000    \n",
            "                                                                 \n",
            " bidirectional_2 (Bidirectio  (None, 64)               73984     \n",
            " nal)                                                            \n",
            "                                                                 \n",
            " dropout_2 (Dropout)         (None, 64)                0         \n",
            "                                                                 \n",
            " dense_2 (Dense)             (None, 6)                 390       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 330,374\n",
            "Trainable params: 330,374\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n",
            "Epoch 1/20\n",
            "2741/2741 [==============================] - 93s 32ms/step - loss: 0.9778 - accuracy: 0.5829 - val_loss: 0.8684 - val_accuracy: 0.6253\n",
            "Epoch 2/20\n",
            "2741/2741 [==============================] - 84s 30ms/step - loss: 0.8748 - accuracy: 0.6273 - val_loss: 0.8543 - val_accuracy: 0.6357\n",
            "Epoch 3/20\n",
            "2741/2741 [==============================] - 80s 29ms/step - loss: 0.8371 - accuracy: 0.6434 - val_loss: 0.8379 - val_accuracy: 0.6424\n",
            "Epoch 4/20\n",
            "2741/2741 [==============================] - 80s 29ms/step - loss: 0.8126 - accuracy: 0.6530 - val_loss: 0.8326 - val_accuracy: 0.6443\n",
            "Epoch 5/20\n",
            "2741/2741 [==============================] - 80s 29ms/step - loss: 0.7927 - accuracy: 0.6634 - val_loss: 0.8452 - val_accuracy: 0.6448\n",
            "Epoch 6/20\n",
            "2741/2741 [==============================] - 81s 30ms/step - loss: 0.7756 - accuracy: 0.6716 - val_loss: 0.8353 - val_accuracy: 0.6460\n",
            "Epoch 7/20\n",
            "2741/2741 [==============================] - 79s 29ms/step - loss: 0.7594 - accuracy: 0.6801 - val_loss: 0.8635 - val_accuracy: 0.6432\n",
            "Epoch 8/20\n",
            "2741/2741 [==============================] - 81s 30ms/step - loss: 0.7458 - accuracy: 0.6857 - val_loss: 0.8613 - val_accuracy: 0.6429\n",
            "Epoch 9/20\n",
            "2741/2741 [==============================] - 81s 29ms/step - loss: 0.7313 - accuracy: 0.6931 - val_loss: 0.8619 - val_accuracy: 0.6439\n",
            "3198/3198 [==============================] - 30s 9ms/step - loss: 0.8309 - accuracy: 0.6435\n",
            "Test acc: 0.643\n"
          ]
        }
      ],
      "source": [
        "inputs = tf.keras.Input(shape=(None,), dtype=\"int64\")\n",
        "embedded = layers.Embedding(\n",
        "    input_dim=max_tokens, output_dim=256, mask_zero=True)(inputs)\n",
        "x = layers.Bidirectional(layers.LSTM(32))(embedded)\n",
        "x = layers.Dropout(0.5)(x)\n",
        "outputs = layers.Dense(6, activation=\"softmax\")(x)\n",
        "model = tf.keras.Model(inputs, outputs)\n",
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"SparseCategoricalCrossentropy\",\n",
        "              metrics=[\"accuracy\"])\n",
        "model.summary()\n",
        "\n",
        "callbacks = [\n",
        "    tf.keras.callbacks.ModelCheckpoint(\"embeddings_bidir_gru_with_masking.keras\",\n",
        "                                    save_best_only=True)\n",
        "    ,tf.keras.callbacks.EarlyStopping(monitor='val_accuracy', patience=3)\n",
        "]\n",
        "history = model.fit(int_train_ds, validation_data=int_val_ds, epochs=20, callbacks=callbacks)\n",
        "model = keras.models.load_model(\"embeddings_bidir_gru_with_masking.keras\")\n",
        "print(f\"Test acc: {model.evaluate(int_test_ds)[1]:.3f}\")"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model.evaluate(int_test_ds)[0]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "4d3e0cc9-bcf2-4d4a-b148-a2fa7c69363a",
        "id": "5qcVrneg1wJZ"
      },
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "3198/3198 [==============================] - 33s 10ms/step - loss: 0.8309 - accuracy: 0.6435\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.8309476375579834"
            ]
          },
          "metadata": {},
          "execution_count": 22
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plot_history(history)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 166
        },
        "outputId": "15b191e7-4829-45d4-a305-335e2a4946bb",
        "id": "SVQT-WJI1wJl"
      },
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "error",
          "ename": "NameError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-23-a8489d1127d4>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mplot_history\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mhistory\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
            "\u001b[0;31mNameError\u001b[0m: name 'history' is not defined"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "pred1= model.predict(int_test_ds)\n",
        "pred1=np.argmax(pred1, axis=1)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "12dc5a2f-f83f-47d2-d5c7-7dd2c019b257",
        "id": "zZIluWsC1wJm"
      },
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "3198/3198 [==============================] - 33s 9ms/step\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print_validation_report(test_y, pred1)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "491d10b2-f601-450e-c88b-a2539d21389f",
        "id": "a-leCZFS1wJm"
      },
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Classification Report\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           1       0.68      0.75      0.71     11464\n",
            "           2       0.47      0.31      0.37      9191\n",
            "           3       0.48      0.40      0.43     12304\n",
            "           4       0.53      0.51      0.52     25982\n",
            "           5       0.75      0.83      0.79     43391\n",
            "\n",
            "    accuracy                           0.64    102332\n",
            "   macro avg       0.58      0.56      0.57    102332\n",
            "weighted avg       0.63      0.64      0.63    102332\n",
            "\n",
            "Accuracy Score: 0.643\n",
            "Mean Absolute Error: 0.433\n",
            "Root Mean Square Error: 0.805\n",
            "Mean Square Error: 0.648\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "sns.set(font_scale=1.2)\n",
        "plot_confusion_matrix(test_y,pred1)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 506
        },
        "outputId": "31edf87e-81b9-43a2-a357-4e674df22084",
        "id": "00-txT7M1wJm"
      },
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 576x576 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "SDRunHYSRrpG"
      },
      "source": [
        "## Using Pretrained Word Embeddings"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ivO55Vo15uk6"
      },
      "source": [
        "The rationale behind using `pretrained word embedding`s in natural language processing is much the same as for using pretrained convnets in image classification: *you don’t have enough data available to learn truly powerful features on your own*, but you expect that the features you need are fairly generic—that is, common visual features or semantic features. In this case, it makes sense to reuse features learned on a different problem."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XllLQfXb5uk6"
      },
      "source": [
        "The idea of a dense, low-dimensional embedding space for words, computed in an unsupervised way, was initially explored by Bengio et al. in the early 2000s, but it only started to take off in research and industry applications after the release of one of the most famous and successful word-embedding schemes: the `Word2Vec` algorithm (https://code.google.com/archive/p/word2vec), developed by Tomas Mikolov at Google in 2013. `Word2Vec` dimensions capture specific semantic properties, such as gender."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "riqJwrpr5uk6"
      },
      "source": [
        "There are various precomputed databases of word embeddings that you can download and use in a Keras Embedding layer. `Word2vec` is one of them. Another popular one is called `Global Vectors for Word Representatio`n (GloVe, https://nlp.stanford.edu/projects/glove), which was developed by Stanford researchers in 2014. This embedding technique is based on factorizing a matrix of word co-occurrence statistics. Its developers have made available precomputed embeddings for millions of English tokens, obtained from Wikipedia data and Common Crawl data."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oKNrxDE65uk6"
      },
      "source": [
        "First, let’s download the GloVe word embeddings precomputed on the 2014 English Wikipedia dataset. It’s an 822 MB zip file containing 100-dimensional embedding vectors for 400,000 words (or non-word tokens)."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5P6O5iXN5uk7"
      },
      "source": [
        "<div class=\"alert alert-block alert-success\"><b>GloVe: Global Vectors for Word Representation</b><br>\n",
        "https://nlp.stanford.edu/projects/glove/</div>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 29,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-EjBZq6cRrpG",
        "outputId": "baf909a0-cd7e-4f2c-aca2-64a1a41d8c85"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "--2022-10-29 03:59:15--  http://nlp.stanford.edu/data/glove.6B.zip\n",
            "Resolving nlp.stanford.edu (nlp.stanford.edu)... 171.64.67.140\n",
            "Connecting to nlp.stanford.edu (nlp.stanford.edu)|171.64.67.140|:80... connected.\n",
            "HTTP request sent, awaiting response... 302 Found\n",
            "Location: https://nlp.stanford.edu/data/glove.6B.zip [following]\n",
            "--2022-10-29 03:59:15--  https://nlp.stanford.edu/data/glove.6B.zip\n",
            "Connecting to nlp.stanford.edu (nlp.stanford.edu)|171.64.67.140|:443... connected.\n",
            "HTTP request sent, awaiting response... 301 Moved Permanently\n",
            "Location: https://downloads.cs.stanford.edu/nlp/data/glove.6B.zip [following]\n",
            "--2022-10-29 03:59:16--  https://downloads.cs.stanford.edu/nlp/data/glove.6B.zip\n",
            "Resolving downloads.cs.stanford.edu (downloads.cs.stanford.edu)... 171.64.64.22\n",
            "Connecting to downloads.cs.stanford.edu (downloads.cs.stanford.edu)|171.64.64.22|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 862182613 (822M) [application/zip]\n",
            "Saving to: ‘glove.6B.zip’\n",
            "\n",
            "glove.6B.zip        100%[===================>] 822.24M  5.05MB/s    in 2m 42s  \n",
            "\n",
            "2022-10-29 04:01:58 (5.09 MB/s) - ‘glove.6B.zip’ saved [862182613/862182613]\n",
            "\n"
          ]
        }
      ],
      "source": [
        "!wget http://nlp.stanford.edu/data/glove.6B.zip\n",
        "!unzip -q glove.6B.zip"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "o4Ifl-sJRrpH"
      },
      "source": [
        "## Parsing The GloVe Word-Embeddings File\n",
        "\n",
        "First line of `glove.6B.100d.txt`:\n",
        "\n",
        "`the -0.038194 -0.24487 0.72812 ...`"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 30,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VPo9aWdbRrpH",
        "outputId": "2d02207e-8351-47ce-f5e2-d374f8549d6d"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found 400000 word vectors.\n"
          ]
        }
      ],
      "source": [
        "path_to_glove_file = \"glove.6B.100d.txt\"\n",
        "\n",
        "embeddings_index = {}\n",
        "with open(path_to_glove_file) as f:\n",
        "    for line in f:\n",
        "        word, coefs = line.split(maxsplit=1)\n",
        "        coefs = np.fromstring(coefs, \"f\", sep=\" \") # np.dtype('f') returns dtype('float32')\n",
        "        embeddings_index[word] = coefs\n",
        "\n",
        "print(f\"Found {len(embeddings_index)} word vectors.\")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "evKxdRndRrpH"
      },
      "source": [
        "## Preparing The GloVe Word-Embeddings Matrix\n",
        "\n",
        "Next, let’s build an embedding matrix that you can load into an Embedding layer. It must be a matrix of shape `(max_words, embedding_dim)`, where each entry *i* contains the `embedding_dim`-dimensional vector for the word of index *i* in the reference word index (built during tokenization)."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 31,
      "metadata": {
        "id": "jSCD2_KcRrpI"
      },
      "outputs": [],
      "source": [
        "embedding_dim = 100\n",
        "\n",
        "# Retrieve the vocabulary indexed by our previous TextVectorization layer.\n",
        "vocabulary = text_vectorization.get_vocabulary()\n",
        "\n",
        "# Use it to create a mapping from words to their index in the vocabulary.\n",
        "word_index = dict(zip(vocabulary, range(len(vocabulary))))\n",
        "\n",
        "# Prepare a matrix that we’ll fill with the GloVe vectors.\n",
        "embedding_matrix = np.zeros((max_tokens, embedding_dim))\n",
        "\n",
        "# Fill entry i in the matrix with the word vector for index i.\n",
        "for word, i in word_index.items():\n",
        "    if i < max_tokens:\n",
        "        embedding_vector = embeddings_index.get(word)\n",
        "    if embedding_vector is not None:  # Words not found in the embedding index will be all zeros.\n",
        "        embedding_matrix[i] = embedding_vector"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 32,
      "metadata": {
        "id": "ZE2xRs4QRrpI"
      },
      "outputs": [],
      "source": [
        "embedding_layer = layers.Embedding(\n",
        "    max_tokens,\n",
        "    embedding_dim,\n",
        "    embeddings_initializer=keras.initializers.Constant(embedding_matrix),\n",
        "    trainable=False,\n",
        "    mask_zero=True,\n",
        ")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PUsE9u-g5uk9"
      },
      "source": [
        "We’re now ready to train a new model—identical to our previous model, but leveraging the `100-dimensional` pretrained GloVe embeddings instead of `256-dimensional` learned embeddings."
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# **LSTM**"
      ],
      "metadata": {
        "id": "ZlD3xnCeu-HG"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# **1000 Vocab Size**"
      ],
      "metadata": {
        "id": "Dj3JRXXCwmp1"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "# **Experiment 1 (32 units, bidirectional, dropout = 0.5)**"
      ],
      "metadata": {
        "id": "uNgCgFlUJDWj"
      }
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VOA_ss3iRrpJ"
      },
      "source": [
        "## Model Leveraging Pretrained (GloVe) Embedding Layer"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 33,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "RI8Tj4g1RrpJ",
        "outputId": "25f4836d-49e6-4923-af77-f62e2fc131ce",
        "scrolled": false
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model_3\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " input_4 (InputLayer)        [(None, None)]            0         \n",
            "                                                                 \n",
            " embedding_4 (Embedding)     (None, None, 100)         100000    \n",
            "                                                                 \n",
            " bidirectional_3 (Bidirectio  (None, 64)               34048     \n",
            " nal)                                                            \n",
            "                                                                 \n",
            " dropout_3 (Dropout)         (None, 64)                0         \n",
            "                                                                 \n",
            " dense_3 (Dense)             (None, 6)                 390       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 134,438\n",
            "Trainable params: 34,438\n",
            "Non-trainable params: 100,000\n",
            "_________________________________________________________________\n",
            "Epoch 1/200\n",
            "2741/2741 [==============================] - 85s 28ms/step - loss: 1.1165 - accuracy: 0.5260 - val_loss: 0.9698 - val_accuracy: 0.5841\n",
            "Epoch 2/200\n",
            "2741/2741 [==============================] - 77s 28ms/step - loss: 0.9566 - accuracy: 0.5901 - val_loss: 0.9031 - val_accuracy: 0.6099\n",
            "Epoch 3/200\n",
            "2741/2741 [==============================] - 74s 27ms/step - loss: 0.9175 - accuracy: 0.6062 - val_loss: 0.8860 - val_accuracy: 0.6168\n",
            "Epoch 4/200\n",
            "2741/2741 [==============================] - 77s 28ms/step - loss: 0.8933 - accuracy: 0.6154 - val_loss: 0.8656 - val_accuracy: 0.6257\n",
            "Epoch 5/200\n",
            "2741/2741 [==============================] - 76s 28ms/step - loss: 0.8759 - accuracy: 0.6227 - val_loss: 0.8597 - val_accuracy: 0.6287\n",
            "Epoch 6/200\n",
            "2741/2741 [==============================] - 77s 28ms/step - loss: 0.8638 - accuracy: 0.6285 - val_loss: 0.8504 - val_accuracy: 0.6326\n",
            "Epoch 7/200\n",
            "2741/2741 [==============================] - 76s 28ms/step - loss: 0.8538 - accuracy: 0.6332 - val_loss: 0.8506 - val_accuracy: 0.6360\n",
            "Epoch 8/200\n",
            "2741/2741 [==============================] - 75s 27ms/step - loss: 0.8446 - accuracy: 0.6356 - val_loss: 0.8497 - val_accuracy: 0.6348\n",
            "Epoch 9/200\n",
            "2741/2741 [==============================] - 76s 28ms/step - loss: 0.8369 - accuracy: 0.6400 - val_loss: 0.8530 - val_accuracy: 0.6344\n",
            "Epoch 10/200\n",
            "2741/2741 [==============================] - 75s 27ms/step - loss: 0.8300 - accuracy: 0.6437 - val_loss: 0.8537 - val_accuracy: 0.6339\n",
            "\n",
            "Processing Time: 0:13:25.873067 \n",
            "\n",
            "3198/3198 [==============================] - 30s 9ms/step - loss: 0.8496 - accuracy: 0.6341\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.8496147990226746"
            ]
          },
          "metadata": {},
          "execution_count": 33
        }
      ],
      "source": [
        "inputs = tf.keras.Input(shape=(None,), dtype=\"int64\")\n",
        "embedded = embedding_layer(inputs)\n",
        "x = layers.Bidirectional(layers.LSTM(32))(embedded)\n",
        "x = layers.Dropout(0.5)(x)\n",
        "outputs = layers.Dense(6, activation=\"softmax\")(x)\n",
        "model = tf.keras.Model(inputs, outputs)\n",
        "model.compile(optimizer=\"rmsprop\",\n",
        "              loss=\"SparseCategoricalCrossentropy\",\n",
        "              metrics=[\"accuracy\"])\n",
        "model.summary()\n",
        "\n",
        "callbacks = [\n",
        "    tf.keras.callbacks.ModelCheckpoint(\"glove_embeddings_sequence_model.keras\",\n",
        "                                    save_best_only=True)\n",
        "    ,tf.keras.callbacks.EarlyStopping(monitor='val_accuracy', patience=3)\n",
        "]\n",
        "\n",
        "start = datetime.now()\n",
        "\n",
        "history = model.fit(int_train_ds, validation_data=int_val_ds, epochs=200, callbacks=callbacks)\n",
        "model = keras.models.load_model(\"glove_embeddings_sequence_model.keras\")\n",
        "\n",
        "end = datetime.now()\n",
        "print('\\nProcessing Time:', end - start, '\\n')\n",
        "\n",
        "model.evaluate(int_test_ds)[0]"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model.evaluate(int_test_ds)[0]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "C3ZuesscgZXN",
        "outputId": "833b2161-caae-4c35-dd89-49f26cfb5c4f"
      },
      "execution_count": 34,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "3198/3198 [==============================] - 26s 8ms/step - loss: 0.8496 - accuracy: 0.6341\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.8496147990226746"
            ]
          },
          "metadata": {},
          "execution_count": 34
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plot_history(history)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 271
        },
        "outputId": "e489fc56-cfcf-47a3-8766-717f69c24619",
        "id": "De66j3t3dY-3"
      },
      "execution_count": 35,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1152x288 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "pred1= model.predict(int_test_ds)\n",
        "pred1=np.argmax(pred1, axis=1)"
      ],
      "metadata": {
        "id": "5SNfyl0pdY-4",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "3ad390ce-4ab5-4794-889d-573e2e2c32e3"
      },
      "execution_count": 36,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "3198/3198 [==============================] - 28s 8ms/step\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print_validation_report(test_y, pred1)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "c4c291e5-6d2e-4dd0-a81d-46238dff98c0",
        "id": "mRd3c0V5dY-4"
      },
      "execution_count": 37,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Classification Report\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "           1       0.66      0.74      0.70     11464\n",
            "           2       0.43      0.36      0.39      9191\n",
            "           3       0.44      0.42      0.43     12304\n",
            "           4       0.53      0.47      0.50     25982\n",
            "           5       0.76      0.82      0.79     43391\n",
            "\n",
            "    accuracy                           0.63    102332\n",
            "   macro avg       0.56      0.56      0.56    102332\n",
            "weighted avg       0.62      0.63      0.63    102332\n",
            "\n",
            "Accuracy Score: 0.634\n",
            "Mean Absolute Error: 0.448\n",
            "Root Mean Square Error: 0.825\n",
            "Mean Square Error: 0.680\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "sns.set(font_scale=1.2)\n",
        "plot_confusion_matrix(test_y,pred1)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 506
        },
        "outputId": "5e3ab286-1abc-4df9-dbdf-cce8887ee4ec",
        "id": "kaoH1DvtdY-4"
      },
      "execution_count": 38,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 576x576 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfUAAAHpCAYAAACbXKXxAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdZ2AURQPG8X8SUkkooafQAkgnoUkRQXhFQUBEBKSE3nuTJr333ntRpIMUqQoqKoReA0jvJNSEJCTh7v0QODwDGoQUluf3RXd2bm5m2dxzuzu7Z2M2m82IiIjIG882sTsgIiIir4dCXURExCAU6iIiIgahUBcRETEIhbqIiIhBKNRFREQMIllid+B1CDh3P7G7YFjFsqcE4G7Y40TuiXGldrEDICxKd5fGJxd7GyKiE7sXxub0JFHCoxK3H0bmbP/P63WkLiIiYhAKdREREYNQqIuIiBiEQl1ERMQgFOoiIiIGoVAXERExCIW6iIiIQSjURUREDEKhLiIiYhAKdREREYNQqIuIiBiEQl1ERMQgFOoiIiIGoVAXERExCIW6iIiIQSjURUREDEKhLiIiYhAKdREREYNQqIuIiBiEQl1ERMQgFOoiIiIGoVAXERExCIW6iIiIQSjURUREDEKhLiIiYhAKdREREYNQqIuIiBiEQl1ERMQgFOoiIiIGoVAXERExCIW6iIiIQSjURUREDEKhLiIiYhAKdREREYNQqIuIiBiEQl1ERMQgFOoiIiIGkSyxO/AmMZlMrFs6j192bOLu7SBSpExFkZJlqdW4LU5OzgD8vG0Ds8YNivXansOmkN+vOAB37wSzdPZELp47zfUrl8hTsDC9hk+1qh8Z+Yj5U0Zy6exprlw6R5p0GRk3b3X8DzIJOLh/H98uns+ZU4HcuHGdFm060KR5K8v66Oholi5ZyPq1q7hx/RrpM2SkTj1/ataua6kTHBTEpPGjOHMqkEsXL+BXpBhTZs6zep/ZM6Ywd+a05/Zh3pJl5M1XIH4GmATNmDqZmdOnxipft2kLmTNn4bfdvzBj2hQuX7xIWFgY6dNn4OPKn9CidRvs7R0s9bdt2cz8ubO4cP4CTs5O+PkVoWPXbmTOnCUhh5Nk7d8XwKIF8zgVGMj169do274jLVq1saxv2qgB+wL2xnqdk7Mze/YdilUeHBRErc+rc/t2MFt37CJDxozx2v83wYbv1/LNkkVcuXKZyEeP8PDwpEbNWtT3b4SNjQ1rV69kw/p1/HnmNJGRkWTJmo36/o34pEq157YXHBxE7SfbeMv2pL+NFeovYdPqb9i46htadOlLtpy5uX7lIrPGDSYqKoqmHXpZ6tna2jFp8Xqr17q6pbT8f3RUJMndUlCpRl32/LID0+PHsd7LZDKRLJk9H1Sqzp+BRzl94mj8DSyJCQ97SLbsPlSsVIUJY4bHWj97xhTWrVpBz74DyZnrHY4eOcSIwQNIZm9P9RpfABAZFUmKFCn5sn4jfty+hejo6Fjt1PNvTI2ata3Kxo4cyulTgeTJmz9+BpeEeXh6svCb76zKUqd2ByB5clfq1vMnR86cuCRPzqmTJxk8sB/h4eF06xGz7x89cpie3bvQpl0HPqr0Cffv32PcmJF0aNOStRs2J/h4kqKwsDCy++Sg0idVGT1iWKz14yZMJioqyrJsxkzd2jUpVfq9WHVNJhO9enQjf4EC7Nr5U7z2+03iniYNzVu2IWvWbDg4OHDgwD6GDRmIra0t9Ro0ZO+ePyhXvgKdunQnZcpU/PTjdvr27kEyu2R8VKmyVVsmk4neb9g2Vqi/hDMnDlOgcHGKv1cegHQZPChZriInDu2LVTeVe9oXtpMugwcN23QH4NSxQ9wJvhWrjpOTs+WLwv17d96qUC9VpiylypQFYOrEsbHWb1q/jroNGlGu/P8A8PTy5sSxoyyYM9MS6h4ennTr+TUAhw7u59bNG7HacXFJjotLcstyaEgIv/36C81atcHGxua1jyups7O1I23adM9dV8jXj0K+fpZlDw9P9u8LsDqqPHL4EG5ubjRtEXNWxcvbG/9GTejUrg0hISG4ubnF7wDeAGXeL0uZ92P27YnjxsRanzJVKqvl33/bza2bN/miVp1YdWfNmIa9vT31/Ru9MYGTEEqVLmO17OXtzU87trMvYC/1GjRk2Ejr7e7fqAn79+1l65YfYoX6m7iNdU39JeTK68vpE4e5dP4MALeuX+VwwG/4Fi9tVc9kekznxtVpW7cSQ75qxcE9vyRGdw0rMvIRDo6OVmWOjk7cuH6N69eu/ud2N21Yx+PH0XxS9bNX7eIb6ebNG3xUoSwfVShL21bNOXTwwAvrnj93jt2//kzR4sUtZYV8/QgJCWHr5h8wmUyEPHjAxvXf4+tXWIH+H61Y9h258+Qlf4GCVuV79/zBqpXLGTxs5Fv5BTSuzGYzR48e4dChAxQr/u4L64WEhODs7GxVFrD3D1avWs7goSOBN2cbJ5kjdbPZzL59+yhWrFhid+WFKn9ej6ioR3zdzh8bG3j8+DEffFydmv7Prvdm8spM8859yZw9J9FRkez5ZQdjB3SlWac+lPvo00TsvXGULF2G5UuXULR4CXxy5OT4sSNsWBcz3yA4KIhMHp7/qd21q5ZTrvyHpHZ3f53dfSPkK1CQAUOGkT17Dh4+DGXV8mU0bVifqTNmU6LUsy+tH1Uoy907d4iKiqJmrTp06dbDsi5/gYKMnzSVfn160qdnd6KjoylQsBCTps1IjCG98YKCbrFr54/07NPXqvx2cDC9e3ZnyLCRpEmThnNn/0ykHiZdISEhVCz/PlFRUZjNJlq2bkvd+v7Prbtx/TqOHD5M9x69LWVPt/HgoSNxT5OGs2/QNk4yoR4VFYW/vz8nT55M7K680N5fd7B9wypadOlLluy5uH7lIktmjWfFwunUahQz2SVnnoLkzPPsW3XOPAUJfXCfDSsWKdRfk87dezFq6ED869TAxsaGtOnSUbV6DRbNn4ON7X/7Rn340AHOnf2Trk9O2b9tnp4SfqpwkaLcvHmDhfPnWoX6vIXfEBERTuDJk0wcPwZ3d3dat+sAxBy9Dxs8kLoNGlK23Ac8eHCf6VOn0KVje2bPW4idnV2CjulNt3b1KhwcHKj8SRWr8l49ulG1WnVKlCyVSD1L+pInT86yVWuJCA/n8KGDTJo4jnTp0vPZ519Y1fvpx+0MGtCXAYOGkCdvPkt5757dqFr1zdzGSSbUIeZoPSn7dvZEPvq0Nu9ViLnu4p0tB5GRj5g9fjDV6zbFwcHxua/Lmbcgv+/ampBdNbSUKVMxdNR4oqIiuXvnDmnTpWf1ymUAeHp6/6c216xcRpZs2SlStPi/V35LFCzky47t1vutp5cXAD45cmJrZ8vXPb+iUZNmOLu4MHf2THxy5rSaze3tnYVKH35AwN49b+QHZGIxmUysXrmCylWqkjy5q9W6PX/8zr6AvSycPxd49rlZqWJ5qteoSb8Bse++edvY2tpa7rjI9U5uHjx4wJRJE6xCffOmjfT7uif9BgymSrXqVq+3bOMF1tu48kcx27hv/6S7jRM01PPkyfOP65P6taFHEeHY2FpPQ7C1tcVsBv7hC8mFP0+RJm2GeO7d28fe3oH0GWJuL9m2eSN+hYv+p1Pn9+/f46ftW2nVrtPr7uIbLfDkCTJmzPTC9WaTGZPJRFRUFM5AeHgYtjZ/+/uwe7KcxL+wJzW7f/2Fa9euUvM5E+RWrrW+s+b4saP0/7o302fNJVv27AnVxTeKyWTiUeQjy/KqlcsZOWwwg4eOjDU5DmDlGuttfOzYUQb07c20mUl/GydoqLu4uNC7d2+8vWMfTUVGRtK8efOE7M5LK1KyLJtWfUOGTF5kzfEO169cZMWiGRQqWhIHRycAVi2ZhU+ufGT0ykx0VCR7f/2RnVvW4d+qq1VbF8+eBiA05AER4eGW5Sw+uSx1rl48R3R0NPfv3uZxdJSljmfmbCSzt0+IISeKsLCHXLl8CYDoqCju3A7m9KmTODu74J05CyeOH+X6tau8kzsvd+/c4dvFCzhzKpCZ85ZYtXP6VMylnAf37xMeHmZZzvWO9ZfLTevXAVC56tt7eWTMqOG8X/YDPDw8efgwlNUrV/DH778xfnLMveuLFswjW7bsZM6aFRsbG04cO8aEcaMp+0F5UqSMuV3zg/L/o//XvViyaAFly5XnwYP7TJk4nnTp05O/YKHEHF6SEfbwIZcuxezbUVGRBAcHE3jyJC4uLmTO8uxe/pXLvyNf/gLkyZM3Vhs5c+ayWr539y4AWbNmI316HTxMmzKJwkWK4uXlTXR0FPv372PBvNlUq14DgMWLFjBh7Ch69elHkWLFCA4OAsDe3p6UKWPuPsjxt2189w3axgka6rlz58bZ2ZnixWOf4oyMjEzyp98btO5GcreUfDtnIvduB5MiVWp8i7/HFw2fTZQLD3vIwmmjuHf3Dg4Ojnh4Z6F9r2GW2+Ce6tOu/nOXl/zw7Bah0f06E3zreqw64xesJV0Gj9c+vqTi5InjtG3eyLK8ctm3rFz2LX5FijF9zkIiIyOZO3MaV69cJpm9Pb6FizJrwbex/hD963z+3OU/Dp6wKl+7egUf/K+i5Q/6bRQcFETf3j24e+cOrm5u5Mz1DjPmzKf4uyWAmAf+TBg3hmvXrmJrY0MmD09q16lL3QYNLW1UqfYpERHhLF/6LVMnT8TZ2ZkCBQsxbeYcXF1dX/TWb5Xjx4/RrPGzCVvLln7DsqXfULRYceYuWAzAzZs3+eXnXXzdf2BidfON9vBhKEMH9efWrZs4ODri5eVN+05dLbcFfrtkEY8fP2bIoP4MGdTf8roiRZ/9G7zJbMwJmKSbNm0iZcqUlC5dOtY6k8nEunXr+Oyzl7+dKODc/dfRPXmOYtljjsLuhsV+QI68HqldYiaQhUUl7S+1bzoXexsiYj+DSF4jpyeHieFR/1xP/jvnfzlJm6ChHl8U6vFHoR7/FOoJQ6Ee/xTq8e/fQl0PnxERETEIhbqIiIhBKNRFREQMQqEuIiJiEAp1ERERg1Coi4iIGIRCXURExCAU6iIiIgahUBcRETEIhbqIiIhBKNRFREQMQqEuIiJiEAp1ERERg1Coi4iIGIRCXURExCAU6iIiIgahUBcRETEIhbqIiIhBKNRFREQMQqEuIiJiEAp1ERERg1Coi4iIGIRCXURExCAU6iIiIgahUBcRETEIhbqIiIhBKNRFREQMQqEuIiJiEAp1ERERg1Coi4iIGIRCXURExCAU6iIiIgahUBcRETEIhbqIiIhBKNRFREQMQqEuIiJiEDZms9mc2J0QERGRV5cssTvwOgSFRCd2FwwrnVvMLnL9fmQi98S4MqV0ACA4VPtxfErrmoywSB3DxCcXBxsAIrQrxxunf0ltnX4XERExCIW6iIiIQSjURUREDEKhLiIiYhAKdREREYNQqIuIiBiEQl1ERMQgFOoiIiIGoVAXERExCIW6iIiIQSjURUREDEKhLiIiYhAKdREREYNQqIuIiBiEQl1ERMQgFOoiIiIGoVAXERExCIW6iIiIQSjURUREDEKhLiIiYhAKdREREYNQqIuIiBiEQl1ERMQgFOoiIiIGoVAXERExCIW6iIiIQSjURUREDEKhLiIiYhAKdREREYNQqIuIiBiEQl1ERMQgFOoiIiIGoVAXERExCIW6iIiIQSjURUREDEKhLiIiYhAKdREREYNIltgdeJN8u2geu37azsUL58FsJptPDho2bUmJUmUsdTatX8OwgV/Heu34qXMo9m5Jy/KP27ewZMEcLl88j6OTMwV9C9OmQxe8vLNY6vy8cweL583iwvmzODm7UOHDj2ndvguOTk7xO9BEtHXTelZ+t5hrV68QGRlJxkweVPn0c76o64+NjQ17f9/NgtnTuHL5EuHhYaRLl57yFSvRsFlr7O3tLe1cvniBSWOGc+TQARydnChb/kPadOqGs7OLpc6ZUyeZNWUCgSeP8fjxY7L75KRR89YUfbdUYgw9QX2zaB67ftzOpQvnMZvNZPfJQcNm1vvynt9+Ze6saVy5dJHw8DDSpsvAhx9XonHz1tjbO1jqfb9mJSuWLuHqlcukSpWKTz6tQePmrbG1jX3MEBEeTlP/2lw4d5ZpcxZRyK9Igow3qbh79y7Tpkxk5087uH/vHunSp6dps5bUqFnLUufokcOMGTWcwJMnSJEiJVU/rU7b9p2ws7Ozamv1yuV8t/QbLl44j5OTM75+fkycMiOhh5TkLJg3hx3bt3H+/Dkwm/HJkZMWLVtTusz7ljrR0dEsXjifNatXcv3aNTJmzES9Bg2pU7fec9sMDgqi1ufVuX07mK07dpEhY8aEGs5LU6i/hP379vBJtc/Ikzc/jk7ObFi7kh6d2zJ55gIK+ha21LOzs2P1xh1Wr02RMqXl/48fO8KA3t1o1qo9FSpWIuTBfaZMGE33jm1YunojAHv/2M3XX3WiTYeulClXnps3rjN62CDu379H/yGjEmbAiSBVancaNGlJ5izZsHew58jBA0wYNRRbOztq1qlP8uTJ+bxOfbL55MDFJTlnTp1k7PBBRISH065LDwDCwsLo0rYZPjlzMWXuYkLu32fkkH6EDgmh/9DRAEREhNOtfUv8ihRj0syFJEuWjFXLltCra3sWLf+eTB6eibkZ4t2BgD1UqfYZufPlx8nJmfVrV/JVp7ZMmfVsX07u6kqtL+uT3ScnLsmTczrwJCOHDiA8PJyOXXsC8P3qFYwfM5zuvfpRyK8I586eYdTQAURHR9OybcdY7ztmxGA8vby5cO5sgo43KQgLe0jTRvVJnz4Dw0eOxcPDg6CgIEwmk6XOjRvXad2iCRX+V5F+AwZz6eJFBvTrg9kMHTt3tdSbOnkiq1cup2OXbhTy9SM6Opozp08lxrCSnL17/qD6Z5+TL38BnJydWLNyBe3btmLugsX4FY75Ejl96mRWrVhG3wGDeeed3Bw+fJBBA/phb2/P51/UsmrPZDLRq0c38hcowK6dPyXGkF6KQv0ljJ0002q5Tcdu7Pl9N7t+2m4V6gBp0qZ7YTvHjxzC1c0N/yYtnpR4U6d+I3p2aUdoaAiurm5s3vg9Rd8tSZ36jQDw9MpMq/ad6dO9I81bd8DD0+t1Di3JKF6ytNWyh6c3v+76kUP7A6hZpz75CvqSr6CvZX3GTB4cPrCPQwf2Wcp2bNnI/fv3+HrwSFxd3QDo1L0Pvbq0pUWbjmTy9OLypYvcv3cX/6YtyeaTA4AWbTuzduUyzp45ZfhQHzvZel9u27Ebe37bza4fn+3L+Qv6kv9v2/rggQAO7g+wlP2wYR2VqnxK5arVAfD08ubqlcvMnjYZ/ybNrc6MbFq/ljOnAxk0fCy7f94Zj6NLmhbOn0tERDiTps7AwSHmTMff/45XLFtK8uSu9B80FFtbW3xy5OTWrZtMGD+GFi1b4+ziwuXLl5g3ZyaTp82kVOlnZ1Z8nuzHb7tpM+dYLXfu9hW7d//Cju3bLKH+/bo1NGjYhAr/+xAAL29vjh09wpxZM2KF+qwZ07C3t6e+f6M3ItR1Tf0VmEwmHj4MxdnZ2ar88ePHfPHpR3z6UVnatWjE7l92Wq3PX9CX0JAQdmzbjMlkIiTkAVs2radAIT9LCEU+eoSjg6PV6xwdY5YPH9zH28BsNnPy+FGOHTmEX9Hiz61z8cI59vz+K35FilnKjh0+RL78hSzbEqBYiZLY2tpy9MhBALy8M5PaPQ0/rF/Lo4gIoqOj+H71clKkSEm+AoXid2BJ0NN92elv+/JfXTx/jj92/0rhv/xbPIqMtATUU46OTkREhBN44ril7ML5s0ydOJZBw8bEqv+22LF9G35+RRgzajgfflCGz6pWYvzYUYSHh1vqHDp4gJKlSltduij1XhkiwsMJDDwJwE87tpEsWTLu3b3H59Wr8GH5MrRv05I/z5xO8DG9CUwmEw9DH1p9Tkc+eoSjY+z99tq1q1y7dtVStnfPH6xauZzBw0ZiY2OTYH1+FUkm1G/cuJHYXXhpi+bNIjQkhGqffWEpy5wlG736DWHIyPEMHT2Rd3LnoUfntmxYu8pSJ2/+ggwfO5mxIwbxQUk/Kn1Qkps3rjNi7BRLnRKlyvD77p/5ddePmEwmbt64zsI5MdfLgoOCEm6QiSA0NISPyxbnw9KFadusATVqfcnnta2vddWsUoEPSxemYa1P8StSnNYdu1nW3b4dhHuaNFb1kyWzxy1FSu4EBwPg7OzCpJkL2Lf3dz4uW5yKZYqxYuliRk2aSWp369e+DZ7uy5/W+CLWuuqVylOuhC91a1alcNHitOvU3bKuRKn32Lzxe44cOoDZbObC+bMs+3YRAMFBt4CY6+hf9+hC6/adyZIte8IMKAm6cvkS27dt4WFoKBOmTKNjl+5s3fwDgwf0tdQJDg4iTdq0Vq9LmyZm+enf/eXLlzGZzMyYNplOXboxYdI0nJ2dadqoAXdu3064Ab0h5syaQUjIA2p+UdtSVrrM+3z7zWLOnD6F2WzmyJHDrF0T8xkddCtmv70dHEzvnt0ZPHQEadK8OZ8JCXb6/d69e/Tq1YuAgABy587N119/Te7cuS3rK1euzIEDBxKqO69s9YqlLJ4/mxHjppA+w7NJE38/ZZm/oC/3799jyaK5VKn+ORBzdDl2xGBqfdmA0mXKERLygLkzp9K7ewcmzZiPnZ0dn3xag+vXrzKgz1dERcUcDTVu3oajRw69Md8Y/ysXl+TMWbKSRxHhHDtymNnTJpAmbXo++bSGpc7kmQuJiIjgzOmYyW6p3d1p3KJtnN/jUUQEI4f0I0vW7HTrPYBkyZKxYe1Kendtx/T535IhY6b4GFqStHr5UhbNm83I8db78lPT5iyK2danTjJt0jhSu7vTrFU7ABo1a8W9e3dp37IxZrMZV1c3vviyPnNmTMHmydHm+NHD8cmRkyp/+fd7G5lMJlKmSsWAwcNiJnXmg6ioSL7q2okevb8mZcpUcWrHbDIRHR1F9569KfN+OQAGDxvJx/8ry8YN39OgYeN4HMWbZdnSb5g7eyYTp0y3mtzWo2cfhgzqT63Pq2NjY0O6dOn5rEZN5s2ZZdlve/XoRtVq1SlR8s2aOJtgoT569GgiIyMZM2YMv/32G/Xq1WPGjBkUKxZz2tRsNidUV17Zt4vnM2/mVEaMm2I1o/1F8hf0Y/uWTZblRfNmk80nB42atbaUeXpl5vMq/+PAvr0Ue7ckNjY2NG/dgaYt23E7OIgUKVNx7eplpk0ai6eXd7yMK6mwtbXFyzszAD453yE05AFzpk+yCvVMT65FZvPJgZ2tHUP796JOg8Y4O7uQJk06bt20PvMTHR1FyIP7uD85Ctqx9QfO/XmaCdPnkixZzKz5d/LkY3/AHtavWUGz1h0SYqiJ7ttF85k7cyojx794X3563Te7Tw5sbW0Z1Lcn9Ro2wdnZBQcHB77q3Z8uX/Xhzu1g3N3TsHfP7wCW/XTf3t+5dfMGP23fatVu+5aNKVLsXcZPnR2PI0w60qZLh4eHp9VdGj45cgJw/do1UqZMRdq06bj95GzSU7efHH2nTZfO6r9/vYbu6OiIl3dmrl+/Fq9jeJMsnD+XaVMnM3HK9FjBnDJVKkaPm0hUZCR37twhXfr0rFi2FACvJ/vtnj9+Z1/AXhbOnws8y6hKFctTvUZN+g0YlICjibsEC/VffvmFtWvX4u7uTrly5fD19aVdu3bMnDkTX1/fN+boc86MySz7dhGjJ063uo77T04HniB9hmdHfhHhYdjaWF/5sNyu8rcvN7a2tqRLnwGAbZs34pI8OcVLWE8mMzqTyURkZOSL15tNmEwmoqOiwBnyF/Jl8riRPAwNJbmrKwD79vyOyWSiQEE/IObfwAYbbP7+72BrB2/O98tXMnt6zL48ZlLc92WTyYzJZCIqKoq/Xn5PliyZ5Sh/2+aNeHh68U7uvACMnzo75t/miaCgW3Rp14Le/Ye8Vbe0FS5clIC9e4iOjiZZspiP3ovnzwPg8WRipq9fYTau/x6TyWS5rv7b7l9wcnYmd+48APgVLgrA+fPnLV+4oqIiuXr1Ch99XDlBx5RUTZ08kSWLFjB1+iyKFnv+fBwAewcHyxH8D5s2UqRoMdzd3QFYuXa9Vd3jx47S/+veTJ81l2zZk+5lpAQL9YcPH5LyL7d1Va5cGbPZTMuWLZk3b15CdeOVTBw7nHWrVzBg6GgyZ8nK7eCYa1yOTk6WSVlzZ04lb74CeGfJQmRkFDt3bGXDulV07NbL0k6ZchUYNrAPy75ZyHtlPyDkwQNmTp1I2nTpyZu/IAAPHtxn2+aNFC5aHJPJxM4dW/lm4Vy69+pvCSojmj9rKgV8C+Ph6UV0dDRHDu5n6eJ5VKoSM7t62TcLyZwlG96Zs4CNDadOHmfm5PGUfr8cbili9q8KH33CorkzGdy3B81adyDkwX0mjBrGBx9+bDnCL1aiNDOmjGfUkH7Uqd8Yu2R2rF+9gqtXL1PqySlNI5swJmZfHvj3fdnRCVe3mH156eIFZMmaDe8sWbGxgcATx5k2aSzvlf2AFE+29ZXLFzly6CD5C/oSFvaQDWtXsWPrZkZNmGIJpcxZslq9t7NLzIz4TB6ehr/L4K8aNGrC1i0/MHzIQOr7NyY4+Bbjx46iSrVPLbe8flH7S5Yt/YZBA/rSwL8Rly9fZtqUidT5sr5luxV/twQFChZizMhhODkNwt09DfPnzsJsMlG5SrXEHGKSMGr4UFauWMaI0ePImjWbZS6Co5MTbk/27WNHj3Dt2lXy5MnHnTu3WbRgPqcCT7Jg8VJLOzlz5rJq997duwBkzZqN9E8OtJIiG3MCnfeuUqUKY8aMsbqODrBixQrGjh1LaGgox44d+09tB4VEv44u/qv3iuZ7bnmlKp/SZ8AwACaPG8nPO3dw+3Ywjo6OZMmanTr1GlKuQkWr16xdtZw1K5Zy9cplnJydyZe/IC3bdiL7k9NxDx7cp0enNpw9e4bo6Gh8cuSiQePmvF+uQvwO8m/SucV877t+/8VHyq/TlHEj+e3XXQQH3cLBwQEPTy8qVf2MajVqYWdnxzcL57Bt0wZuXL+Gja0NGTJ68L+PKtwKoAcAACAASURBVFPzy/o4OT07dLx08TyTxgzn6KGDODo6UrbCh7Tp1N3qFquD+wNYMHsaZ8+cwmwykzlbdho0bk6pMuUSZKxPZUoZMws3ODRh9mOA0kVevC9/PTBmX148fzabN63nxrWYbZ0xkwcffvwJtb9sYJklf+niBQb0+YqL589hYwO58+anWat2+D45mnye69euUrNqxQR/+Exa12SERSbuaZg9f/zOpAlj+fPMadKkTcuHFT+mVZv2VjOzjxw+xNjRIwg8eQI3txRUq/5ZrIfP3Llzh3GjR7Br10/YYEOBggXp0q2H5XR+YnFxiDnjGpFwu3IshfK989zyap9+xuBhIwA4sH8fQwYN4MrlS9jb21O4SDE6dOpMzlzPfy1AwN49NGvsn+gPn3H6l0PxBAv1MWPGYGdnR+fOnWOtW7RoEcOGDSMwMPA/tZ1Qof42SuhQfxslRqi/jZJCqBtdUgh1o0syoR6fFOrxR6Ee/xTqCUOhHv8U6vHv30I9ydynLiIiIq9GoS4iImIQCnURERGDUKiLiIgYhEJdRETEIBTqIiIiBqFQFxERMQiFuoiIiEEo1EVERAxCoS4iImIQCnURERGDUKiLiIgYhEJdRETEIBTqIiIiBqFQFxERMQiFuoiIiEEo1EVERAxCoS4iImIQCnURERGDUKiLiIgYhEJdRETEIBTqIiIiBqFQFxERMQiFuoiIiEEo1EVERAxCoS4iImIQCnURERGDUKiLiIgYhEJdRETEIBTqIiIiBqFQFxERMQiFuoiIiEEo1EVERAxCoS4iImIQCnURERGDsDGbzebE7oSIiIi8Oh2pi4iIGESyxO7A6/AwUicb4ktyBxsA7oY9TuSeGFdqFzsAzt4KT+SeGJtPemeu3otM7G4YmmcqB0CfyfHp6Wfyi+hIXURExCAU6iIiIgahUBcRETEIhbqIiIhBKNRFREQMQqEuIiJiEAp1ERERg1Coi4iIGIRCXURExCAU6iIiIgbxwsfE9u3bN86NDB48+LV0RkRERP67F4b6hQsX4tSAjc0/P4dWREREEsYLQ33x4sUJ2Q8RERF5RS99Tf3u3bvoJ9hFRESSnjiF+uPHj5k0aRLFihWjdOnSXLlyBYAxY8bw3XffxWsHRUREJG7iFOqzZ89m7dq19OnTB3t7e0t53rx5WbNmTbx1TkREROIuTqG+Zs0aBgwYQPXq1bG1ffaSXLlyxXlCnYiIiMSvOIX69evX8fHxiVVuZ2dHRETEa++UiIiIvLw4hbqnpyeBgYGxyn/77TeyZ8/+2jslIiIiL++Ft7T9Vd26dRk6dCiOjo5AzD3sP//8M+PHj6dnz57x2kERERGJmziFeoMGDbh37x7t2rUjIiKC5s2b4+joSMuWLalZs2Z891FERETiwMb8EjedR0REcObMGcxmMzly5MDFxSU++xZnDyN133x8Se4Q88TAu2GPE7knxpXaxQ6As7fCE7knxuaT3pmr9yITuxuG5pnKAdBncnx6+pn8InE6Uv+rp6fg/zoLXkRERBJfnEI9MjKScePGsXTpUiIjIzGbzTg6OlKnTh26du2Kg4NDfPdTRERE/kWcQn3IkCFs376dPn36ULhwYQAOHDjApEmTCA8PZ9CgQfHaSREREfl3cQr1jRs3MnbsWMqVK2cpy5EjB+nTp6dr164KdRERkSQgThfG7e3tyZw5c6xyb29vkiV76cvyIiIiEg/iFOpffPEFCxYssPp1NrPZzOLFi3VLm4iISBLxwsPsvn37Wv7fZDKxefNmdu/eTcGCBQE4evQo9+7d46OPPor/XoqIiMi/emGo//2HWvLmzQtAcHAwAJkyZSJTpkxcvnw5/nonIiIicfbCUF+8eHFC9kNERERekZ4gIyIiYhBxnrq+d+9eNmzYwNWrV4mKirJat2jRotfeMREREXk5cTpSX7duHU2aNCE4OJg9e/bg6upKcHAwJ06ceO6tbiIiIpLw4hTqc+bMoVevXkybNg17e3t69erFxo0b+fjjj8mYMWN891FERETiIE6hfvnyZd5//30AHBwcCAsLw8bGhkaNGrFs2bJ47aCIiIjETZxC3c3NjfDwmJ+FTJcuneV2t7CwMEJDQ+OtcyIiIhJ3cZoo5+vrS0BAALly5aJcuXIMHz6cY8eOsWPHDooUKRLffUzSTCYTc2ZOZ8P6dQTduknq1O6UK1+B9h274Pzk9+bXrlnFpvXr+PPMaSIjI8mcJSv1/RtTuUpVSzs7f9zBqhXLCDx5gocPQ/Hw9KJGzVp8Wa8BNjb//Pu5RnNw/z6+XTyfM6cCuXHjOi3adKBJ81ZWdVYuX8rK777h+vVruLq68m6J0rTt1JU0adJa6uzYtplF8+dw6cJ5nJycKehXmHYdu+KdOQsAhw7s47tvFnP82BEe3L9HuvQZ+KhyFRo2aWH4Xx5cMm86386fGat8ztLv8fDKzMXzf/LNvBmc+/MU169e4cPK1enUs79V3bOnA5k1eTSXL54nNDSE1KnT8O57ZfFv1hZXtxSx2o6ICKdT83pcunCO0VPnk6+gX7yNLynYumk9q75bzPVrV4iMjCRjJg8++fRzan7pb/mbjogIZ9HcGfy0bTO3g26RMrU7VavXxL9ZawB27/qR9WtWcObUScIePiSjhydVqtekRu16ljZuXLtK3c8+jvX+9Ro3p2mrDgk34CQiLp/Jf3Xu7J/U//ILoiIjCTh03FIeFHSLCWNGcepUIBcvnKdI0eLMmDM/IYfyn8Qp1Hv06EFYWBgA7dq14+HDh+zcuZMcOXLQu3fveO1gUrd44XwWL5zHgMHDyJM3HxcunGdA395ERkbydf+YH7oJ2PMHZT+oQMcu3UmRMiU7f9xOvz49sEtmx0cfVwZgX8BeChbypVnL1qRJk5YD+wMYMXQQkY8e0ahp88QcYoILD3tItuw+VKxUhQljhsdav2PbZsaPHs5XvfpRrERJbt28waihAxnUtxcTp80G4NjRw/Tt2Y0Wbdrz4UeVeXD/PpPGjaJrh9YsX7sJgCOHDuLp7U3tug3IkDEjpwJPMmroQO7cvk2PPv1jva/RZMjkwdjp1neupEyVGoBHERGky5CJd0uXY83y5z+zwt7env9VqoZPrty4uqXgyqULTBs3nNtBt+g7bHys+tPGDiOTpzeXLpx7/YNJglK7u9OgaUu8M2fD3sGeo4cOMHHUUGxt7fi8Tn0eP35M785tefgwlC49++GdOSsP7t/j/v17ljYOHQggX4FCNGjaktTuaThycD8TRw0lMvIRX/o3tXq/waMnkSdfAcuys3PsAHsbxOUz+anw8HB6dOtMseLv8tuvv1iti4qMJEXKlNT3b8S2rZt5HP04IYfxn8Up1L28vCz/7+TkRP/+xv/Ai6vDBw/wbslSVPgw5nG5Hp5efFzpEwL27rHUGTpitNVrGjRswv59AWzb8oMl1Lv16GVVx8vbm8CTJ9i65Ye3LtRLlSlLqTJlAZg6cWys9UcOHSRHzlx8WiPmdwc8PDyp/nktZk+fYqlz7MhhXN3caNS0JQCeXt7U9W9M905tCQ0JwdXNDf8m1tvVw9OL69eusmDuzLci1G1t7XD/y5mNv8qVJz+58uQHYOvGNc+tkzmbD5mz+ViW02fIRJUatfl23oxYdbf/8D3n/jxFz4Gj2LN712vofdJXrERpq2UPT29+3fUjhw4E8Hmd+mzd9D2nA0+weNVGUrunASCjh6fVa9p27hGrjTOnTrJz+5ZYoZ4iRcoX/nu+TeLymfzUiKGD8PUrTIGChWKFuoenFz16xzwu/eD+fdy8eTP+O/8a6OEzr8i3cGEOHzzI6VOnALhy+TK//vIz7z0JpRcJeRDyr9+kQ0L+vc7bqJBvYS6cP8f+fXsxm83cDg7ip+1bKVXmfUudAoV8CQ0JYfvWHzCZTISEPGDzxu8p6FsYVze3F7YdGhKCs7NzQgwj0QUH3aRBjYo0qFGRvt3acuLooVdqL+jmDXbv3E7BwsWsyi9dOMfcaePpMWAk9vb2r/Qebyqz2czJ40c5fuQQfkWKA/DLT9vJnS8/q5d9Q+2q/6PeZx8zZtgAqyP15wkNCcHpOfvo0H49qF6xDK0a1mbFtwuJjo56zquNL66fyRu+X8uJ48fo+lWv5zXzxnrhkXrFihXjfC13y5YtcX7DkJAQ3J58qJ49e5YLFy7g5+eHu7t7nNtISho0bELko0jq1a6BjY0N0dHRfPb5F7Rp3/GFr9m4/nuOHjkc6+j8r/YF7GHLDxsZPX5SfHT7jVb+w48IDQ2lS/tWREdH8zg6mtJlytKn/xBLnXz5CzJq/BQG9etF/z49eBwdTb4CBRk3KfZR5FPnz53lu28X0bpdp4QYRqJ6J09+OvcaSOas2Ql7+JAf1q3kq3ZNGDRmKoWLlXyptrq29ufs6VNERj6ixHvl6N5vmGVdREQ4w/t1p3GrjnhnycbN61df91CStNDQEGpVqUB0VBQms5mGTVtRo3Y9AK5dvcz1a1exsbGl/7CxRESEM23CKPp178iEmQue+/l7aH8AP279gYEjn13ecHZxoUW7LuQv5IeTkxNHDu5n/swpnDkVSO+BsS9fGV1cPpPPnTvLuDEjmTV3IU5OTonY29fvhaFerVq11/pGgYGBNGvWjNu3b9O4cWP8/Pzo1asX0dHRODo6smDBAvLkyfNa3zMhbN+6heXLvqX/4GG8805uLl44z9hRI5g6eQLtOnSOVX/njzsYMrAv/QYNIU/efM9t88jhQ3Tp2I4WrdtRtlz5+B7CG+fQgX1MnzKB9p26UcivCEG3bjJlwhiGDOjDoGExlzounD/HqGGDqFPXn/fKliPkwQNmT59Cjy7tmTp7AXZ2dlZtXrp4gY5tmvPhR5X5ok69xBhWgipWsozVcv5ChQkOusmqpQtfOtR7DhhFeHgYly+eZ9HsKUwZM4SufWK+YM2YMIqs2XNS8ZPqr63vbxIXl+TMXrySiIhwjh89zJxpE0iTLj2Vq9XAZDKD2UzfIaNJkTIlAN2/HkzrRnU4c+okuXLntWrrxNHD9P2qIw2bt6ZUmXKW8pSpUlOnQWPLco5cuXF2cWH0kH40b9uJdOkzJMhYk4p/+0yOjIykR9dOtGnXkRw5cyV2d1+7F4Z6u3btXusbjRkzhqZNm2JjY8OoUaPIkCEDAQEBPH78mIEDBzJlyhSmTp36Wt8zIYwfM5K69fypUvVTAHLmeoeIR48Y1K8PzVu2wdHR0VJ3yw8b6f91L74eMNhS/+/2BeyhU7vWNG7WgqZ/m/EtMWZMmUi58v+jZu26QMw2d3FxoVVTf5q3aod35iwsnDuL7D45aNKiteV1Xt6Z+bRSBfYH7KF4iVKW8rN/nqFD66a8X7Y8X70F19JfJHe+Quzetf2lX5cuQ8wDqDJnzY57mrR0bd2QWvWb4p0lG4f2/0HwrZv8snOb1Wt6dGiGb+HiDBk3/bX0PamytbXF0zvmqZs+Od8h9MED5k6fROVqNUiTNi1RUZGWQAfI+mSOws0b161C/dD+APp0a0fdhs2o1+jf59jkK+hraedtC/V/+0y+HRzE2T/PMHLYYEYOGwzEXB4xmUwU881Hq7bt3+jP3jg/+/1VHT9+nBkzZmAymRgxYgSfffYZNjY2JEuWjM6dO/Ppp88PuaQuPDwMG1vrqQl2traYzWbMZrOlbPXK5YwaPoSBQ0dYJsf93S8/76RH1060bteBBg2bxGu/32ThEeHY/m2b29rGHHk/3ebh4f9Q5y9lJ44fpVPbFnxcuSqdu/d6624f/Kuzp0+SLv2rPSHSZDIBEBn5CIAhY2dYXdu9E3yLr7u2oXPPgeQvZOxb2p7HZDYRGRkJQIFChTlx9DChoSG4usZckrx86QIAGTN5WF7zx68/M7B3Vxq3bEeteg3j9D5nAk8CvHWBDv/+mZwufQaWr/7eav3On35k5rTJLF2xxuq22DdRgoV6VFQUyZLFvJ2bmxspUjy7j9Xd3f2NfYhNufL/Y/GCeXhnzkzu3Hm4cOE80yZPpNR7ZSzXapYsWsDEcaPp0bsvRYoWIzg4CIi5JShlylQAbNuymT49u9O4aXMqfVLVUsfO1o7Ub+h8g/8qLOwhVy5fAiA6Koo7t4M5feokzs4ueGfOQtlyFVg4fzZ58xXAr0hRbt28wfgxI8iR8x28nhwVlf2gAoP792bpkoWUKfsBIQ8eMH3KBNKlS0/+/AWBmBmtXTu2pvz/KtKwSXPu3A629CFN2nQJP/AENGvyGIqXep8MmTwIexjK5vWrObjvD/oNnwDE/L1eunAWiPmCFBpyn7NnArFPZm+Z8b55/WpcXd3InM0HBwdHLp7/k3nTJ+KTMzfZfGJOa3o9eSbAU08nIWb08CRDJuuZ3kazYNZUCvgWJpOnF9HR0Rw5uJ/vFs3j4yoxlyI+rVmHtSuXMmJAb5q27kBERASTRg+lUOGi5MiVG4CdO7YwrF9PvvRvyv8+/sSyj9ra2pIqdcznwuYNa7G1tSNn7jw4ODhy9NB+Zk4ex/vlPyRDxkyJM/hEFJfP5L+fdj9x/Nhzy089+XJ0/8F9wsLDLMvv5E66l4ptzH89nIxHH374IatXr8bNzY1du3ZRtuyzmYjBwcFUr16dX3/99T+1/TAyQYbwXOFhYcyYPoUft20lKOgW7u5pKFO2HG3ad7QE9icflef6tWuxXlukaDFmz4+5B7h54wbs3xcQq04mDw82bvkxfgfxD5I7xBy53g1LuHs09+/bS9vmjWKV+xUpxvQ5C3n8+DGLF8xh0/p13LxxHTe3FBQuVpw27TtbHeGsWbmMlcuXcvXyZZydnclXoBCt23fCJ0dOAAb1682m9Wuf24c/Dp6Il7E9T2qXmDMIZ2+FJ9h7jhzQk2NHDnD/3l2SJ3clm08uajVoiu+Tmdk3r1+lca1PYr0ufcZMLFjxAwBbN67l+1VLuX71CqbHj0mbPgOl3i/P5182JMWTff/vnrabGA+f8UnvzNV7kQn2flPHj+T3X3cRHHQLBwcHMnl4UanqZ1StUcsyp+PMqZNMnzCaE8eP4JYiJe+WKkOLtp0tp+Q7t27M4QP7YrWdIZMHS9fGTFDeuul7li2Zz/VrV8Ec84WpYuWq1KhdP8EfouSZKub9kvpn8t99v3Y1gwf0tXr4DEDhArmfW//A0cDX3u+4evqZ/CIJFurTp0+ncuXKZMmSJda6xYsX88cff/zna+qJuQMZXWKE+tsmMUL9bZTQof42SgqhbnRJJtTjk3ag+KNQj38K9YShUI9/CvX492+hHueHz5w7d47hw4fTqlUrgoJirvfu2LGDwMDEOw0hIiIiz8Qp1Pft20f16tUJDAzk119/JSIiAoh5eMybeBuaiIiIEcUp1MeNG0fbtm1ZuHCh1WMeS5QowZEjR+KtcyIiIhJ3cQr1wMBAKleOfW91mjRpuHPnzmvvlIiIiLy8OIW6o6MjISEhscovXLjwxj6zXURExGjiFOply5Zl5syZlqdFAdy5c4cJEyZQvryeTS4iIpIUxOmWttu3b+Pv78/Dhw8JCgrCx8eHy5cv4+XlxZIlS0j5l2cXJwbdPhF/dEtb/NMtbQlDt7TFP93SFv9e233qjx49YuPGjRw7dgyTyUT+/PmpVq1agj+x6Hm0A8UfhXr8U6gnDIV6/FOoxz89fEZeiUI9/inUE4ZCPf4p1OPfv4V6nH7QZe3a5z8f+6nq1d/O30oWERFJSuIU6n369LFaNplMmM1mbG1tsbW1VaiLiIgkAXEK9ePHrX+5xmQyceLECUaMGEHnzp3jpWMiIiLycuL87HerF9nakj9/fjp16sTgwYNfd59ERETkP/hPof6Uu7s7Fy5ceE1dERERkVcRp9PvN2/efG7ZlClT8PHxee2dEhERkZcXp1AvW7YsNjbW0+jNZjMeHh5MmDAhXjomIiIiLydOob5o0SKrZVtbW9zd3cmSJQt2dnbx0jERERF5Of8a6lFRUfzyyy/UqVMHT0/PhOiTiIiI/Af/OlHO3t6eJUuWJERfRERE5BXEafZ7sWLFOHjwYHz3RURERF5BnK6pV61alTFjxnDlyhUKFCiAs7Oz1frChQvHS+dEREQk7uL0gy65c+d+cQM2Npw8efK1dupl6ccD4o9+0CX+6QddEoZ+0CX+6Qdd4t9r+UGXHTt2vJbOiIiISPyJU6hfu3YNPz8/kiWzrh4dHc3Bgwc1K15ERCQJiNNEOX9/f+7fvx+rPCQkBH9//9feKREREXl5cQp1s9kc64lyAKGhoTg5Ob32TomIiMjL+8fT77169QJiJsMNGTIER0dHyzqTycTx48fJnz9//PZQRERE4uQfQ/3GjRtAzJH6rVu3sLe3t6yzt7enePHiNG7cOH57KCIiInHyj6E+f/58IOaIvU+fPri6uiZIp0REROTlxWn2+/Dhw+O7HyIiIvKK4jRRTkRERJI+hbqIiIhBKNRFREQMQqEuIiJiEAp1ERERg4jTr7SJiIhI0qcjdREREYOI033qSV14VGL3wLicnzxEUNs4/jzdxjcfaCPHpwwp7Flx6Fpid8PQvvD1AODHwNuJ3BPjKp87zT+u15G6iIiIQSjURUREDEKhLiIiYhAKdREREYNQqIuIiBiEQl1ERMQgFOoiIiIGoVAXERExCIW6iIiIQSjURUREDEKhLiIiYhAKdREREYNQqIuIiBiEQl1ERMQgFOoiIiIGoVAXERExCIW6iIiIQSjURUREDEKhLiIiYhAKdREREYNQqIuIiBiEQl1ERMQgFOoiIiIGoVAXERExCIW6iIiIQSjURUREDEKhLiIiYhAKdREREYNQqIuIiBiEQl1ERMQgFOoiIiIGoVAXERExCIW6iIiIQSjURUREDEKhLiIiYhAKdREREYNQqIuIiBhEssTugJHs3fM7rZo3wdPTi/U/bAPgt92/MGPaFC5dvEBYWBjp02fg48qf0LJ1W+ztHWK1ERwcRO3Pq3P7djBbtu8iQ8aMCT2MJO952/nPP88wY9pkTgUGcuXyJap/9jn9Bw19YRtv+3ZeungeP/+0g4sXzoPZTDafHPg3acm7pd6zqrdmxXesXvEtN65dI7mrK8VKlKJ1+y64p0kLwLABfdi8cV2s9m1sbFi7eSep3dMAMdt78tgR7Pn9VwBKlCpDx269LOuN4PyJw+zeuJzrF85yP/gmFWo14YPPG1jW7/9pE4d+3sbNy+eIjooibSYvSn3yBb5lPrTUCbl7mx8WT+fGxT8JvnaZrHl9adJ3rNX77FixgJ9WLnxuH1oNnY5XjtwAXDh5hO3L53Hjwp/Y2NjyTuESVG7YFhe3lPEw+oRx5vhBtq9dypXzZ7gTdJOq9ZpTuVZjqzrnTx1n5dyJXDp3GhdXN0qWr0y1ei2wtbOz1Ll/J5jlcyZw4sAfAOQrUpJazTuTIpW7pU7QjausWTCVMycOERkRQXoPbyrWqE+x95/9e43r05Yzxw5avX+qNOkYPi/230RCUai/JsHBQfTt3ZOSpUpz6eJFS3ny5K7UredPjpw5cUmenMCTJxk8sC/h4eF079Hbqg2TyUTvHt3IX6AAu3b+lNBDeCO8aDtHhIeTKZMH5cqVZ/GiBf/YhrYzHAjYS+Wqn5E7b36cnJzYsG4VPbu0ZdLM+RQoVBiAn7ZvYdLYEXTt+TVFipck6NZNxg4fxNABvRk7eRYAHbr1pGW7zlZt9+neAScnZ0tgm0wmenZui62NDeOmzMZsNjNu5BB6d+vAtLlLsLGxSdjBx5PIR+Gk98xKodIV2Lhwaqz1544dJE/R0nxUryXOrm6cDNjNqqkjsLOzo0Cp8gBER0Xh4pqCUp98wfE/dvH48eNY7bxXtTbFP6xmVbZh3iRuXPwTT593ALh56TwLhnajxMc1qN68K+EPQ9i0cCrfjOlLswET39ht/ig8nEze2Sj2fkVWzJ0Ya/2doJtM6t8Rv5LlqNeuJ7euXWbx5GGYzWY+a9gGiNkfpw7pjq2NDR0GTsSMme9mjGHmsJ50GznTsm2mD/kK15SpaNdvHC6ubgTs2sr8cQNwT5cRnzwFLO9Z7P2KfN6kvWXZ1jZxT4Ar1F8Dk8lEn57dqf1lPR49emQVNoV8/Sjk62dZ9vDwZP++vewL2BurnVkzpmFvb099/0Zvbdj8k3/azvkLFCR/gYIArFm96h/b0XaG0ZNmWC237tCVPb//yq4fd1hC/ejhg/jkzEWV6jUByOThSbUaXzB35rPAcnV1w9XVzbJ8+eIFThw7wsDhz44u9+39ndOBJ1iyYj2Zs2YD4OuBw2lYpzqHDgTgV6R4vI0zIb3jV4J3/EoAsOXbWbHWf9G+j9Xye1VrceHkYY7+vtMS6qnTZ6RKkw4AXAw8yv3bQbHacXRyxtHJ2bIcERbK6UN7KF+zoSWQjvz2I6nSZeTj+q0s9ao27cS0ni04f/wQ2fP7xWr3TZC/aCnyFy0FwJpF02Kt/3nzGpxcklO/fW9sbW3xyJyde7eDWbNwKpVrN8bRyZnAwwFcPnuK/lOXktErCwCNOvdjcPv6nDl2kFwFChMWGsL1y+dp7T+KLE/OfFSu3Zgf1y/j4p8nrELd3sGRlKmTzhmnRL+mPnDgQEJCQhK7G69k1oxpgA2Nmzb/17rnz51l968/U6z4u1blAXv/YPWq5QweOhJ4M79Fx7eX2c4vou38fCaTibCHD3F2fhYWBX0Lc/H8OQ7u34vZbOZ2cDA7d2yjZOkyL2xn3eoVuLunoUy58payY4cPksnDyxLoANl8cpAufQaOHDr4vGbeGuEPQ3FwdHqlNg7u2orp8WMKl/vYUhYdFUmyv13es3dwBGJOyxvVuZNHyONb3OpoOV/hd4l8FMHlc6cBOHvyKGkyeFgCHcAjc3ZSp0nPnycPA+Di6oZHluzs3bWF8IehmEwmAn7eRuSjCN4pWNTqPQ/v2UX3BpXp37o2CycO4U7QjQQYq4L/nwAAIABJREFU6Ysl2JF6QEDAc8s3bNhAyZIlSZ06NcWKFUuo7rw2AXv/YOXy7/hu5Zp/PKVVscL73L1zh6ioKL6oVYcu3XpY1t0ODqZ3z+4MHjoS9zRpOHv2z4To+hslrtv5n2g7v9ji+bMJDQmh6mc1LWXlKlQkNDSErzq2ITo6msePoyn53vv06Dv4uW1ERkayZeM6qlT/nGTJ7C3lt4ODcU8T+0gmTZq03A6OfST6tjj0yzaunDnBJw3bvlI7AdvXk7d4GZKnSGUpy+VbnN0blrN32/cU+aAyj8IfsnXpbAAe3A1+pfdLyu7fvU32PAWtylI8OYq+f/c2EDP+lH+5dv6snjv379y2LHccNIk5o76mS92K2NrZ4eD4//buPKqqqg/j+PeioODM4IDihIoGpDgkWmam5lxqTjklZpZpr6VWillazhPO4+uQqDkPmaalb1iWE4WWU5kQihMgODLDff/Abt1AywSunp7PWq7l2WfffX9nL/G5Z59zLgXp/854ylbwtPSp16g5JTr0oIRbKeJiLrFjzTImDO3LuzODbXb2nmeh3qtXL0wmE2azOcu+//znP5hMJk6ePJlX5eSI+Pg4Aoe/xZix43F1dbtr32UfrSIxKYmfTp5gRtBUSji78NqgzGW2wOHDaNeuPf4NGuZF2Q+de5nnu9E8Z2/z+jWsXLaYCdNmU7LU7zcMHg37jsXzZjLgP0Op6VebmOho5s+axsQPRvHe2ElZxgnZ8znXr1+jXYfOeVn+Q+nk4X1sWTiV9q++hXvlav94nMhTPxId9Stt+w62avf0rUPbgP/wxcf/5dOlM7HLl48GrTtRuFgJm1/zfRiYzZnX2U12drw5bi6OToU4sn8vS6a+xxtj51iW5Bu1bG95TdkKnlSu7suo/p34dventOr8ok1qz7NQb9q0KUlJSYwbN47Sf7jT2N/fn61bt1KqVKm8KiXH/HL6NDHR0fxn4O/XrTIyMjCbzdSp+Qgfjp9E6zbtAChbzgOAKlWqYpcvHyOHv0VA3344Ojlx8MB+Qg8f4qPlSwAsH3xat3ia9h07Mer9D/L4yB4s9zLPd6N5zurj4GUsWzSPCdNmU7d+A6t9i+fP4skmzejY5QUAPKt64ejkxOv9X6TvKwMp51Heqv8nm9ZRr35D3MuWs2p3cXUl9NCBLO8dF3cFl/v4kPaw+uGb/7Fp/kSe6z8Mvyefua+xDu3ehpt7eSp718qyz79lB+q3aM/Nq3EUcHTCDOz7ZC3Opdzv6z0fZMVKuHA9/opV2/WrcZZ9AEVLuHLyaGiW116/Gk+x2zd3/vTjd4TtD2Hi8m2W13lUrsaZUz/wv0/WEDBkdLbvX6hwUUqXq0Bc9MWcOqR7lmehPnfuXLZs2UL37t0ZMGAAnTs//J/mfXx82bB5m1Xb2jWr+XpvCHPmL6JU6TLZvi4jI4OMjAxSU1NxhCxjHDv2I6NHBTJv4RIqVa6cW+U/NP7pPP+Z5tnakgVzWPfxCibPmEetOlkvfSUlJmIyWZ/V/XaW9+cVt1/Dz/DDke8ZO3lGlnF8avqx/L8LOHc2Eo/yFSz9oy9f4tFaD+cNW//U4T2fsn3ZLJ5/bbjl5rh/KuHmdY4f2EvzF/rdsY/JZKLI7VAK3bMdgBr1nrhj/4dd5RqPcujLnWRkZFj+rZ74/gAOBQricXtFxLOGLzvWLiX6wjlKumeebF08G0F87GWq1KgJQEpSEpD1TnY7OzuyWWy2SEpMIPrCOXzq2G41ME/vfm/fvj3+/v6MHDmSnTt3Mnbs2If20QoARycnqlS1XjpzdnbB3t7e0r5i+VIqVqpMhYoVMZlMnDh2jBnTp/BUk6cpWizzedE/jxEfHw9AxYqVKFny4VvByGl/Z55TU1M4c+YMAIkJt7h27RqnTp3E3t4eT88qgOb5j2ZNm8gnm9fz/tjJeFSoxJXYzOusBQoWsNzN3qjx06xc/l9qePtS068OMdGXmT19Ep5Vq1lWnn6zddM6XFzdaNjoqSzvVfexBlSr/ghj3xvO4LcCwWwmaPJYvH1rUqv2w3cfzZ0kJyUSd+k8AOlpady8FsfFX3/BoaAjLqXL8s329exauYC2fQdT8ZFa3Lh9Bpkvf36cChe1jHPx18x7PRJv3iAlKdGyXaZiFav3C9u7CwC/xi2yrefrT9ZQ5dG65Ld34PTRQ3y+ejGNO/TApXTZnD3wPJSUmEDMxSgA0lPTuB4fx7nwnyng6ETJMuV4smUH9m7fwKq5E2n6XDdiLp5n2+rFPNWmk+WJgeo16+Hh6cWy6WPo2n9I5lL7wmlU8vKm6u2nAipX96VwseKsmDmWdj36U9CpEEf2h3DqaCj93sq8pyTmYhQHvtyJb92GFClegriYS3z68RLMZjMNmraxzQQBJnN2F7nzwKpVq5g/fz7Xrl1j9+7d97X8npiag4Xdp/lzZ7Pj008sX4qy9L+L+HTbVi5cOI+dyUQZ97K0at2WHr1etLrT+I8OHzrIy317PxBfiuJ4+36nB2mOIes8nz8fRZsWTbP0K+Nels8+/1+2Yzwo8/zbHF++nneT/GQ9n2zbW7Z5jsDRmV/ak56ezuoVS9j56SdcvnyRIkWK4lfnMV4Z9IbV6khyUhIdWj9Nx84v0G/A69mOGxsbw6ypEzi4fx8mk4n6DZ7gjbcC8/TLZ0oVtWf9kQu5Nn748SMs/eDNLO0VH6lJv/dnMHVQN67GXL7j/t+827VJtuOPXWv9+OXMN1/E3dOLzoMCs+2/bNxbnD9zitTkZFzdPfBv2YF6TdveyyHds861Mpf2/3fqyl/0/Gd+/vF7gt4dlKW9qo8fQ8ZlPmoZ/tMxNi6ZlfnlM4UK06Bpm+y/fGZxEMe/P4jJBN61G9Clv/WXz5wL/5lPVi7k19MnSU1Jxq1MOZq07UzDZplzGBdzmY9mfMiFs+EkJtykWAkXPGvUpG33fpQsY30JKic9Xf3uPzM2C3WAc+fOERoaSuvWrSlQoMA/HudBCxwjeVBD3UhsEer/Rrkd6pL7oS5/Heo2/fIZDw8PPDw8/rqjiIiI/CU92yAiImIQCnURERGDUKiLiIgYhEJdRETEIBTqIiIiBqFQFxERMQiFuoiIiEEo1EVERAxCoS4iImIQCnURERGDUKiLiIgYhEJdRETEIBTqIiIiBqFQFxERMQiFuoiIiEEo1EVERAxCoS4iImIQCnURERGDUKiLiIgYhEJdRETEIBTqIiIiBqFQFxERMQiFuoiIiEEo1EVERAxCoS4iImIQCnURERGDUKiLiIgYhEJdRETEIBTqIiIiBqFQFxERMQiFuoiIiEEo1EVERAxCoS4iImIQCnURERGDUKiLiIgYhMlsNpttXYSIiIjcP52pi4iIGER+WxeQExJStdiQW5zsTQAkpdm4EAMrePunMCFF/45zk5ODiRMXbtm6DEN7xL0QAI5+g2xciXElhs25636dqYuIiBiEQl1ERMQgFOoiIiIGoVAXERExCIW6iIiIQSjURUREDEKhLiIiYhAKdREREYNQqIuIiBiEQl1ERMQgFOoiIiIGoVAXERExCIW6iIiIQSjURUREDEKhLiIiYhAKdREREYNQqIuIiBiEQl1ERMQgFOoiIiIGoVAXERExCIW6iIiIQSjURUREDEKhLiIiYhAKdREREYNQqIuIiBiEQl1ERMQgFOoiIiIGoVAXERExCIW6iIiIQSjURUREDEKhLiIiYhAKdREREYNQqIuIiBiEQl1ERMQgFOoiIiIGoVAXERExCIW6iIiIQeS3dQFGcujgAQa83JeyZcvxyWefA9CvTy++Cz2cpW9BR0f2Hw7L0h4bG0O35ztw5UosO3eHUKp06Vyv+0H3XehhVixfyk+nTnHx4gUGvj6Y/q++Ztn/Up9ehB4+lOV1BR0dORh6BICYmGimT5nEqVOniPw1grr1HmPRkuV5dQgPhfj4eObNmUnIl3u4dvUqbiVL8lK/V+jYqQsAn2zZxPujArO8bv6ipfg3aAjAqZMnmDp5AhHh4dy4cR0XF1caP9WEga+/QZGiRfP0eB4E16/Fs3rpfA59E8KN69dwdnHj+R59eaZtRwDCDn3Lmo8WcjHqHEmJCTi7laTR0y3o0rs/9vb2AISfPsXSudOIOhvBrZs3KF7ChXoNn6THSwMpVLgIAGuWL2DtR4uyrWHy/GCqVvfOmwPOZS+0qceg7k2oVM6Vgg75ibwQx7LN3zJr5f8A6NmuPos/6JXlda1fnc2XB38CYNGYnvR61j9Ln4yMDCo2CyQm/iblyzjz044PsvSZuHgnY+Z9atXW7qlHeavvM/hUdSclNZ0fT5+n/aB53EpMyYlDvmcK9RwSGxvDe4HD8W/4OOciIy3t02bOJjU11bJtNpvp2a0zDRo+kWWMjIwMRr7zFt6+vnwV8mWe1P0wSEhIoLJnFVq1aceUieOz7J8+409zjJnuXTvR8PHf5zglJYWixYrT68U+fLFrJ+np6XlS+8MiIeEWL/XpScmSpZgwaRru7u7ExMSQkZFh1S9fvnzs3B1i1VasWDHL3+0dHHj2uQ541XiEokWLEhkRwYTxHxAdHc30mXPy4lAeGImJCYwc3A8X15IMeXcCJUuXIe6K9Zw6FipE2+e7U76SJ46OTkT88hPzpo0lOSmRvgOHAWBv78DTLdtRqUp1ChcpwvlzkSyaOZG42BiGfzgNgOe69qbFs52s3n/xrMn8+stPVPF6JO8OOpfFxN1k4uKd/Bx5meSUNB6vXYWZI7qQnp7B3I9DAEhLS6dKy3etXhd3LcHy92FTNjBq1lar/eum9+dWYgox8Tet2ju9sZDQY79atm8mJFvtf7F9AyYP7ciYeZ+y5/1TZGSY8a1alvQMcw4c7T+jUM8BGRkZjBz+Nl1e6E5KcrJVqBcrVtyq74FvvyH68mU6d+mWZZzFC+aR396enr1fVKj/QaMnG9PoycYAzJw+Ncv+YsWt53h/NnNctmw5RowcBUDYd99x+fKlXKz44fPRsiUkJSUya+4CHBwcAHAvWy7bvq6ubnccx9OzCp6eVSzbZcq407VbDxbM+3cFOsCWNR+RnJTEyPEzsb89pyVLu1v1qe5dk+reNS3bJUu7c/zodxw78p2lzaNiZTwqVrZsu5UqQ6vnurD2o4WWNkdHJxwdnSzbt27e4PuD++j64iuYTKYcPzZb2b3/pNX2r+ev8OxTj/Jk3aqWUAe4fOXGHce4fjOJ6zeTLNtVypfksUcr0eOtJVn6xl+7dcexihQqyOShHQmcsYUlG7+xtJ+OjP67h5MrFOo5YPGCeZgwEfDSyyz8i/+81q9bQ/Uaj+Dt62vVfvjQATZtXM/H6zYRfuaX3CzX8NavzZxjH99HbV3KQ2PP7i/w86vD1MkT+HLPbgoXLsyTTzXh1ddex9HR0dIvPT2dti2bkZycTIWKFendpy9PNm5yx3EvXbrI7i92Ue+x+nlxGA+U/V/9jxq+tVg6dxoH932JU6HC1G3YiBf6vEqBgo7ZvibqbATfH/yW2vUfv+O4sdGX2P/Vbnz86t6xT8jn20lPT+fpls/e93E8yOp6V8C/VmUm/XenpS1//nyc2DaaggXsOR0ZzYwVe/js62N3HKNfp8e5FHudT0KOZtm3fHwfHAs6EHkxjnWfhTJvTQhpaZkrLc0aVKdoYUcSk1PZt+ptypUqzk8Rl/lg3qd8E3Ym5w/2b7JZqCcmJhIWFobZbMbPzw8nJ6e/ftED6PChA2xYt5aPN2z6y0/EMTHRfBXyJe8EWi8NXYmNZeTwt/lg3EScXVwU6vchJiaavSH/Y/jts3L5e6LOneXc2UiaNW/BjDnziImOYdL4D4mJjmb8pMzVkQoVKzH6g3FU86pOSkoKX3y+k8GDBvDemLF06Gi99Ptiz278dOokycnJNH7qacZNnGKLw7KpyxeiuHT+HA0aNyVw/AziYmNYPGsS8bGxvPnuOKu+/Tq35Nq1eNJSU2nxbCf6DHgzy3jDB/Uh4vRPpKQkU69hY94cOS5Ln9/s2rYR/0ZPU6x4iRw/LlsrWrggZ3aNw8E+H3YmE+MWfca8j/cCmWfJ/d9fyY8/R1HAwZ6Ozf3YNOtVXh2zio+27M8yloN9fnq0rc+yzd9awhrgVmIygUGb2X80nITEFJ6oU4X3BrSlplc5Xhq1AoDK5VwB+GBQO0YEbeF05GV6P9eAHQtfp363iZwKt81qYJ6F+tChQ5k2LfP6T1RUFAEBAcTExGAymXB2dmb58uV4eHjkVTk5Ij4+npHD32b02PF3XZL8zdZNm3BwcKBVm7ZW7YHD36Jtu+csNxvJP7dl00YcHBxo/ac5lrvLyMigWPHijP5wfOYNWt6QmprC20Pf4J3AdylWrDg1a/lRs5af5TU1a/lx7dpVli9dnCXUJ00NIiEhgYjwM8ydNYNxH77Ph+Mm5fVh2VSGOYMiRYvx+jujyZ/fHrwgLS2VKaPfpt9/3qZI0d/vRRg3awnJSUmE/3KK4EWzKFa8BC8EDLAab9h7E0lMTCAqMoJVS+ayYPp4Bo/IejPXyR+PcO7XM/Qf/E6uH6Mt3LiVTP1uE3Aq6IB/zcp88PqzXIy5xkdb9nPwhwgO/hBh6XvwhwicixViaJ/m2YZ6x2a1cC7mZLV8DnDl6i2CVuyxbP/w83lu3Epm0ZiejJq1lQsx1zDZZT48NmXpF6zflXm55Mip9TSqU5WXOz3B0MkbcuPw/1KehfqXX/5+jTgoKIjatWszblzmJ81Ro0YxY8YMS+g/LM6c/pmY6GgGD3zV0paRkYHZbKZuTW8+HD+RVm3aWdo3bVxHqzbtKFSosNU4hw7s57vDh1ixfCmQeTMdQJsWTWnf8XnefT/rD65klZGRwaYN62ndNuscy925urnh7l7Wcsc1gGeVqgBcvHAhy70hv6lZ04+dO3ZkaS9dugwAlSt74urqRp9eL9D3pVeoVLlylr5GVcLZlZKl3TMD/bbfro3HXL5oFeqlypQFoHwlT/LZ5WPG+Hfp0K0PBf9w6cO1ZOaTMB4VKlPCxZURgwJ4vkcA5cpXsnrfXds2ULZ8RXxq3Xl5/mFmNpsJPxcLwLHTFyhe1InRA9tlG9oAB46G06VlnWz39ev0BLv3nyLywpW/fN8DR8MBKO/uzIWYa1yKuQbAiTMXrfqdCr+IRxnnv308OS3PQv23oAIICwtj7dq15M+f+fbDhg2jY8eOeVVKjvH28WX95k+s2tat+Ziv94Ywe/5CSt3+jw3gm31fc/HCBTp16ZplnD+PcfzYj4weNZK5C//7r/pP8H59s+9rLlw4T6dsbkKUu6tduy6HDx0kLS3N8nMZGZF5xuPuXvaOrzt18gSl/+Kxy99+9lNSku/az2geedSPY2GhpKenkS9f5pxeOJd5E+2fb5j7o4yMDDIyMkhLSwWyv/Zuvn13dWqK9WNTN65fY//ePfToNzAHjuDhYGcyUbDAnaOsVnUPoi7FZ2mvXrk0j9euQtch2T8K+Gd+NTJXks9fvgpguW7uVakUX3932tKvasVS7PvDdl7Ls1D/4/XmpKQk3Nx+X652cXHh+vXreVVKjnF0cqJK1WpWbc7Oztjb22dp37h+Ld4+vlSvkfXxkj/3vRqf+Q+wYsVKlCxZKoerfvgk3LrF2bNngcwl4djYWE6dPImTkxPlK1Sw9Nuwbg3ePr7UyGaOAU6dzLxz9tq1qyQkJFi2q9eokctH8ODr1acvn+/6jAljx9CzdwCxsdEETZtM22efo+jtR9YWzJuNt8+jVKhYkdSUFHZ/sYvNmzbw9vCRlnE2bVxPkSJF8KxSlQIOBfjll5+ZGTSN6jUeoWo1L1sdnk0816U334R8wcKgCTzbpSfxV2JZPj+Ip55pS+Eimc/sb10XTFmPirh7VMBkgl9+OsmKRTOp17Cxpc8X2zdTqHARPCpWxsGhAGcjfmHFollUrlqdCpWrWr3nl7u2AdCkRbu8Pdg88u6rrfkm7AwRUbHY58/HE7WrMLRPM4I/OQDAyFdaE3r8V05HRlPAPj8dmvkR0KFhtkvh/Z5/nIsx19j+Vdab6Hq2q096RgZHTp4jKSWNx/08Gf9mezZ98T3nbn9AiIiKZcOu7wjs34qoS/GcPhvNi881wKtiKXoPX5a7E3EXeRbqSUlJ9O7d2/L3ixcvUqZM5plsXFyc1R22RhN9+TL7vtrLyPfG2LqUh9Lx48foF9Dbsr3241Ws/XgVdes9xpLlwQBcvnyZr7/ay7vv33mOu3Zqn+320eM/5ULVDxcvr+rMnreIWTOm0a1ze1xcXWn+TEtefe11S5+bN28ycfwHXImNpUCBglSsVIlJU4No1ryFpU8+OzuWLF5I1LmzpKenU6p0aZo83ZwXA/piZ/fv+gLLSlWqMWrCLIIXz2ZIvxco7uxCw6ea8UKf3y/XpaWlsWLhTKIvX8BksrM8rtauU3dLHzs7OzasXMKlC1Gkp6fjWrIU9Z9oQvuuvbPM6eefbqJB46ZWS/tGUrRQQWYHdsO9ZDGSUtKIiIrlvdnbWLzh68z9hQsyY3gXSrkUJTE5lZ9/vUzPd5ayZc8Rq3EKFrCne9v6LFi7l/T0jCzvk2E2M+TF5lQs64LJlPno3IyP9jBndYhVv/6jVzJ+cHsWjumJYwF7fvgpitavzObnXy/n1hT8JZP5j+viuWjOHOtHvVq2bEmVKpnPs+7YsYNt27Yxf/78fzR2QqrtHvQ3Oif7zBWWpDQbF2JgBW9/tE5I0b/j3OTkYOLEhVu2LsPQHnEvBICj3yAbV2JciWF3f2w6z0I9NynUc49CPfcp1POGQj33KdRz31+F+r9rPUxERMTAFOoiIiIGoVAXERExCIW6iIiIQSjURUREDEKhLiIiYhAKdREREYNQqIuIiBiEQl1ERMQgFOoiIiIGoVAXERExCIW6iIiIQSjURUREDEKhLiIiYhAKdREREYNQqIuIiBiEQl1ERMQgFOoiIiIGoVAXERExCIW6iIiIQSjURUREDEKhLiIiYhAKdREREYNQqIuIiBiEQl1ERMQgFOoiIiIGoVAXERExCIW6iIiIQSjURUREDEKhLiIiYhAKdREREYNQqIuIiBiEQl1ERMQgFOoiIiIGoVAXERExCJPZbDbbuggRERG5fzpTFxERMQiFuoiIiEEo1EVERAxCoS4iImIQCnURERGDUKiLiIgYhEJdRETEIBTqIiIiBqFQFxERMQiFuoiIiEEo1EVERAxCoS4iImIQCvU8sn37drp3707t2rXx8vKydTmGM2XKFNq0aUPt2rV54oknCAwMJD4+3tZlGc68efNo1qwZderUoX79+rz00kucPHnS1mUZ2sCBA/Hy8uLgwYO2LsUwZs+eTY0aNfDz87P8GTJkiK3LyhH5bV3Av0XRokXp3r07SUlJjBw50tblGE6+fPmYMmUKVatW5fr167z11luMGDGCBQsW2Lo0Q2nVqhU9evSgWLFipKSksHLlSl5++WW++uor7Ox0jpDTtmzZQlJSkq3LMKS6desSHBxs6zJynEI9jzRq1AhAn7ZzyR8/Zbu4uNCrVy+GDh1qw4qMqVKlSlbbdnZ2xMTEcOPGDYoVK2ajqozp0qVLzJgxg9WrV9OkSRNblyMPCYW6GNL+/fupXr26rcswpJCQEIYNG8aNGzcwmUwEBAQo0HOY2WwmMDCQAQMG4O7ubutyDOnYsWP4+/vj6OhI7dq1eeONN/Dw8LB1WfdNoS6Gs2PHDtavX8/KlSttXYohPfXUU4SGhnL16lW2bNlCmTJlbF2S4axevRqz2UzXrl1tXYohtWjRgo4dO+Lu7k50dDTTpk0jICCArVu3UqhQIVuXd18U6mIo27dvZ/To0cyfPx9vb29bl2NoxYsXp3fv3tSrV4/KlStTtWpVW5dkCGfPnmX+/PmsXbvW1qUYVrVq1Sx/L1WqFOPGjaNu3bqEhYXxxBNP2LCy+6dQF8NYv349U6ZMYcGCBdSpU8fW5fwrZGRkkJaWRmRkpEI9h/y2CtKxY0er9tdee422bdsyZswYG1VmXCaTCZPJhNlstnUp902hnkfS09NJS0sjNTUVgOTkZADs7e1113AOWLFiBXPnzmXJkiX4+vrauhzDWrFiBa1atcLNzY24uDiCgoJwcHCgVq1ati7NMFq1akXDhg2t2ho3bszYsWOztMs/s2PHDvz9/XF2dubKlStMnToVZ2dn/Pz8bF3afTOZjfDR5CGwadMmRowYkaV9xYoV1K9f3wYVGYuXlxf58+fHwcHBqn379u260SgHvfbaaxw9epRbt25RuHBhfH19GTRokC515DIvLy/9X5GDXn31VY4cOUJiYiJFixalXr16DB48mAoVKti6tPumUBcRETEIrfuKiIgYhEJdRETEIBTqIiIiBqFQFxERMQiFuoiIiEEo1EVERAxCoS5iQFFRUXh5eREaGprtdl6bPXs2zZs3v2ufp59+mnnz5v3tMXPqmGw9NyI5SaEu8i9QpkwZ9u3bR82aNf9W/9DQULy8vIiKisrlykQkJynURR5QKSkpOTZWvnz5cHNzw97ePsfGFJEHj777XSQP9OrVi3LlyuHi4sL69etJTU2lTZs2vPvuuxQoUMDSx8PDg5IlS7JhwwbMZjPffPMNkZGRTJ06lf3792MymfD19eWdd97By8vLMv6OHTsICgri0qVLeHt78/LLL1u9f1RUFE2bNmXVqlXUrVsXwPKd13v37uXGjRu4u7vz8ssv4+/vT48ePQBo2rQpAI899hjBwcFA5lfvLlq0iPDwcNzc3GjevDmDBw/GyckJyPy9BuPHj+fTTz/Fzs6O1q1bU7Ro0Xues23btrFixQrCw8PJnz8/NWvWZMSIEVSqVMmq3/nz55k9ezZhYWG4ubkxZMgQ2rRpY9kfGxvL1KlTCQkJITk5GS8vL4YOHUq9evXuuSaRB53O1EXyyK5du7h69SqrV69m6tSp7N4rnMaAAAAF7klEQVS9m2nTpln1+eyzz4iLi2P58uUsXbqU2NhYunfvjrOzM6tWrWLt2rVUqlSJ3r17ExcXB8CJEycYOnQoLVu2ZOvWrfTt25dx48bdtZakpCR69uzJqVOnmDp1Kjt27GDUqFE4OjpSpkwZy7Xt9evXs2/fPmbPng1k/g6D0aNHExAQwI4dO5g0aRLffvst77//vmXsadOm8fnnnzNp0iTWrFmDk5MTq1atuuf5SklJYcCAAWzevJlly5ZhZ2fHK6+8kmUFY8qUKTz//PNs2bKFtm3bMmzYME6cOGE5zt69e3Pr1i0WL17Mli1baNy4MQEBAZw5c+aeaxJ54JlFJNf17NnT3KRJE3NaWpqlbc2aNWYfHx/zrVu3LH2eeeYZc3p6uqXPrFmzzJ07d7YaKyMjw9y0aVPzsmXLzGaz2Tx06FBz165drfoEBwebq1WrZj58+LDZbDabz507Z7W9bt06s4+Pj/nixYvZ1nv48GFztWrVzOfOnbNqb9KkiXn16tVWbYcOHTJXq1bNfPXqVfOtW7fMPj4+5rVr11r16dChg7lZs2Z3naMmTZqY586de8f98fHx5mrVqplDQ0OtjikoKMiqX9euXc3Dhg0zm81m88aNG82NGjUyp6amWvXp1auXeezYsVbj/DY3Ig8zLb+L5BFfX1/y5ctn2a5duzYpKSmcPXuW6tWrA+Dt7W31q3h//PFHjh8/nuVXQiYlJREZGQnAmTNn8Pf3t9r/V79P/vjx41SpUoXSpUv/7frj4uI4f/48EydOZPLkyZZ28+3fCRUZGYmDgwMpKSlZ6q1Tpw4hISF/+70ATp48yZw5czh58iTx8fGW9gsXLlgd35/fy8/PjwMHDgCZ8xcbG5tlqT0lJYWCBQveUz0iDwOFusgDxNHR0Wo7IyMDf39/3nvvvSx9ixQpkldlWWoBGDlyZLa/ArR06dJERETkyHslJibSt29f6tSpw4QJE3B1dQWgTZs2pKam3lPNnp6ezJkzJ8s+hboYkUJdJI/8+OOPpKenW87Ww8LCcHBwoHz58nd8jY+PD5s3b6Z06dKWG+r+zNPTk7CwMKu277///q61eHt7s3HjRi5dupTt2fpvv5f+tyAHcHV1pUyZMkRERNClS5dsxy1fvjz29vaEhYVRtWrVv13Pn505c4a4uDjefPNNPD09LWOYs/lN0UeOHKFx48aW7bCwMMtrfHx82Lp1K4ULF8bFxeWeahB5GOlGOZE8cvXqVcaMGcOZM2cICQlh5syZdO3a1XLXeHZ69uxJeno6r732GqGhoURFRREaGkpQUJAlKPv06cORI0cICgoiIiKCL774gqVLl961lrZt2+Lu7s6AAQP49ttvOXfuHPv372fHjh0AuLu7Y2dnx969e7ly5Qo3btwA4I033iA4OJj58+fz888/Ex4ezu7duy0rCU5OTnTr1o0ZM2awZ88ewsPDmTx58j2fwbu7u+Pg4EBwcDBnz55l//79jBs3DpPJlKXvhg0b2LZtGxEREcycOZMjR44QEBAAwLPPPku5cuXo378/+/btIyoqiqNHj7Jw4UJ27959TzWJPAx0pi6SR1q0aEGhQoXo3r07KSkptG7dmmHDht31Na6urqxdu5bp06czaNAgbt68iZubG3Xq1MHNzQ3IPBudNm0aQUFBLFmyhBo1ajBixAgGDhx4x3EdHR1ZuXIlU6ZM4c033yQhIYGyZcvSv39/y/sOGTKERYsWMX78eOrWrUtwcDDt27encOHCLF68mAULFpAvXz48PDysvi1u2LBhpKSk8PbbbwPQunVrevTowc6dO//2XDk7OzNlyhSmT5/Oxo0b8fT0JDAwkD59+mTpO3ToUNatW0dgYCBubm5MmTIFb29vAAoUKEBwcDAzZsxgxIgRxMfHU6JECR599FEaNWr0t+sReViYzNmtZ4lIjurVqxfly5f/y0fNRETuh5bfRUREDEKhLiIiYhBafhcRETEInamLiIgYhEJdRETEIBTqIiIiBqFQFxERMQiFuoiIiEH8H/ZZ5CAbwJ3QAAAAAElFTkSuQmCC\n"
          },
          "metadata": {}
        }
      ]
    }
  ]
}