imdb_sentimental_analysis_rnn.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/manasik29/Neural-Networks/blob/main/imdb_sentimental_analysis_rnn.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": {
        "ExecuteTime": {
          "end_time": "2021-10-27T13:27:10.080799Z",
          "start_time": "2021-10-27T13:27:10.076234Z"
        },
        "id": "EmsYBp9pbXul"
      },
      "outputs": [],
      "source": [
        "import warnings\n",
        "warnings.filterwarnings(\"ignore\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-10-27T13:27:11.451381Z",
          "start_time": "2021-10-27T13:27:11.441434Z"
        },
        "id": "kyL5w_4UbXuq"
      },
      "outputs": [],
      "source": [
        "%matplotlib inline\n",
        "\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "import numpy as np\n",
        "import pandas as pd\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-10-27T13:27:12.619150Z",
          "start_time": "2021-10-27T13:27:12.615517Z"
        },
        "id": "HybYo-cpbXu6"
      },
      "outputs": [],
      "source": [
        "import tensorflow as tf\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-10-28T13:03:36.742079Z",
          "start_time": "2021-10-28T13:03:36.736645Z"
        },
        "id": "PgwaPeF4bXu7"
      },
      "outputs": [],
      "source": [
        "np.random.seed(1337)  # for reproducibility\n",
        "import tensorflow as tf\n",
        "from tensorflow.keras.preprocessing import sequence\n",
        "from tensorflow.keras.models import Sequential\n",
        "from tensorflow.keras.layers import Dense, Dropout, Embedding, LSTM, Bidirectional, SimpleRNN, GRU\n",
        "from tensorflow.keras.layers import Flatten\n",
        "from tensorflow.keras.datasets import imdb\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-10-28T13:04:44.307879Z",
          "start_time": "2021-10-28T13:04:37.687671Z"
        },
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DC8JBBMkbXu8",
        "outputId": "2206c3af-059b-475b-f4b3-a84e4151deff"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Loading data...\n",
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/imdb.npz\n",
            "17465344/17464789 [==============================] - 0s 0us/step\n",
            "17473536/17464789 [==============================] - 0s 0us/step\n",
            "25000 train sequences\n",
            "25000 test sequences\n"
          ]
        }
      ],
      "source": [
        "'''Train a Bidirectional LSTM on the IMDB sentiment classification task.\n",
        "'''\n",
        "\n",
        "\n",
        "\n",
        "max_features = 20000        #max numbers of words in our vocablary\n",
        "maxlen = 100  # cut texts after this number of words (among top max_features most common words)\n",
        "\n",
        "print('Loading data...')\n",
        "(X_train, y_train), (X_test, y_test) = tf.keras.datasets.imdb.load_data(num_words=max_features)\n",
        "print(len(X_train), 'train sequences')\n",
        "print(len(X_test), 'test sequences')"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-10-28T13:04:45.016728Z",
          "start_time": "2021-10-28T13:04:44.310948Z"
        },
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-8yiP-sDbXvB",
        "outputId": "b824dbde-9bdd-42b2-d841-90d7215fa582"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "X_train shape(before pad_sequences):  (25000,)\n",
            "X_test shape(before pad_sequences):  (25000,)\n",
            "Pad sequences (samples x time)\n",
            "X_train shape: (25000, 100)\n",
            "X_test shape: (25000, 100)\n"
          ]
        }
      ],
      "source": [
        "\n",
        "\n",
        "print('X_train shape(before pad_sequences): ', X_train.shape)\n",
        "print('X_test shape(before pad_sequences): ', X_train.shape)\n",
        "\n",
        "print(\"Pad sequences (samples x time)\")\n",
        "X_train = tf.keras.preprocessing.sequence.pad_sequences(X_train, maxlen=maxlen)\n",
        "X_test = tf.keras.preprocessing.sequence.pad_sequences(X_test, maxlen=maxlen)\n",
        "print('X_train shape:', X_train.shape)\n",
        "print('X_test shape:', X_test.shape)\n",
        "\n",
        "y_train = np.array(y_train)\n",
        "y_test = np.array(y_test)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vjTzR3OJbXvH"
      },
      "source": [
        "# LSTM"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 17,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-10-28T13:04:45.024328Z",
          "start_time": "2021-10-28T13:04:45.020410Z"
        },
        "id": "G6QO_rlubXvG"
      },
      "outputs": [],
      "source": [
        "layers = 128\n",
        "dropout = 0.2\n",
        "epochs = 3\n",
        "batch_size = 128"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 8,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "5A_jZQn7bXvP",
        "outputId": "5824ddc8-866d-4a60-dca5-f6d34e8f47cd"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "keras.layers.embeddings.Embedding"
            ]
          },
          "metadata": {},
          "execution_count": 8
        }
      ],
      "source": [
        "tf.keras.layers.Embedding"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 9,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-10-27T13:30:42.338348Z",
          "start_time": "2021-10-27T13:30:41.803431Z"
        },
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "bQqxaxZ0bXvR",
        "outputId": "fc895858-1957-43f7-b4a3-1cf238476d14"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " embedding (Embedding)       (None, 100, 128)          2560000   \n",
            "                                                                 \n",
            " lstm (LSTM)                 (None, 128)               131584    \n",
            "                                                                 \n",
            " dropout (Dropout)           (None, 128)               0         \n",
            "                                                                 \n",
            " dense (Dense)               (None, 1)                 129       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 2,691,713\n",
            "Trainable params: 2,691,713\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "model = tf.keras.models.Sequential()\n",
        "model.add(tf.keras.layers.Embedding(max_features, 128, input_length=maxlen))\n",
        "model.add(tf.keras.layers.LSTM(layers))\n",
        "model.add(tf.keras.layers.Dropout(dropout))\n",
        "model.add(tf.keras.layers.Dense(1, activation='sigmoid'))\n",
        "\n",
        "model.compile('adam', 'binary_crossentropy', metrics=['accuracy'])\n",
        "\n",
        "model.summary()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-10-27T13:32:38.079247Z",
          "start_time": "2021-10-27T13:30:46.185946Z"
        },
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "RtLGC5WGbXvX",
        "outputId": "0db8f12e-8abd-43f6-9690-ea24cb245488"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/3\n",
            "157/157 [==============================] - 70s 428ms/step - loss: 0.4557 - accuracy: 0.7755 - val_loss: 0.3593 - val_accuracy: 0.8452\n",
            "Epoch 2/3\n",
            "157/157 [==============================] - 67s 430ms/step - loss: 0.2336 - accuracy: 0.9087 - val_loss: 0.3756 - val_accuracy: 0.8340\n",
            "Epoch 3/3\n",
            "157/157 [==============================] - 66s 422ms/step - loss: 0.1499 - accuracy: 0.9462 - val_loss: 0.4966 - val_accuracy: 0.8404\n"
          ]
        }
      ],
      "source": [
        "history = model.fit(X_train, y_train,\n",
        "                    epochs=epochs,\n",
        "                    batch_size=batch_size,\n",
        "                    validation_split=0.2)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 11,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-10-27T13:35:42.409961Z",
          "start_time": "2021-10-27T13:35:42.404982Z"
        },
        "id": "DOLC_YBMbXvZ"
      },
      "outputs": [],
      "source": [
        "acc = history.history['accuracy']\n",
        "val_acc = history.history['val_accuracy']\n",
        "loss = history.history['loss']\n",
        "val_loss = history.history['val_loss']\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 12,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-10-27T13:37:06.687968Z",
          "start_time": "2021-10-27T13:37:06.313543Z"
        },
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 545
        },
        "id": "_DVUicIkbXvc",
        "outputId": "10a9583f-0770-4bae-d862-de69fb93497b"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXoAAAEICAYAAABRSj9aAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3deZzN9f7A8dfbvowlW8qupOiKDIoSpVAiUVE3pBLlV1K3tEvbLep2KymlfUELqVu3xZpWI0sRl0TRJrJlHd6/P97f4ZhmOTNzZr5nzryfj8d5OOe7nfec+XrP93zen+/nI6qKc865xFUs7ACcc87lL0/0zjmX4DzRO+dcgvNE75xzCc4TvXPOJThP9M45l+A80bscEZH3RKR/rLcNk4isFpFO+XBcFZEjg+dPiMht0Wybi/e5SEQ+yG2cWRy3g4isjfVxXcErEXYALv+JyLaIl+WAXcDe4PUVqvpytMdS1a75sW2iU9XBsTiOiNQHvgdKqmpqcOyXgah/h67o8URfBKhqUtpzEVkNXKaqH6XfTkRKpCUP51zi8KabIiztq7mI3CgivwDPisghIvKOiKwXkT+C57Uj9pklIpcFzweIyFwRGRNs+72IdM3ltg1EZI6IbBWRj0RkrIi8lEnc0cR4l4h8EhzvAxGpFrH+YhFZIyIbROSWLD6fNiLyi4gUj1jWU0QWB89bi8hnIrJJRH4WkcdEpFQmx3pORO6OeP2PYJ+fRGRgum3PEpEFIrJFRH4UkZERq+cE/24SkW0icmLaZxuxf1sRmScim4N/20b72WRFRI4J9t8kIktEpHvEujNFZGlwzHUicn2wvFrw+9kkIhtF5GMR8bxTwPwDdzWBKkA9YBB2TjwbvK4L7AAey2L/NsByoBrwADBBRCQX274CfAlUBUYCF2fxntHEeCFwCVADKAWkJZ4mwLjg+IcH71ebDKjqF8CfwKnpjvtK8HwvcG3w85wInAZcmUXcBDF0CeI5HWgEpK8P/An0AyoDZwFDROScYF374N/Kqpqkqp+lO3YV4D/AI8HP9hDwHxGpmu5n+Mtnk03MJYG3gQ+C/f4PeFlEGgebTMCaASsAxwIzguXXAWuB6sChwM2Aj7tSwDzRu33AHaq6S1V3qOoGVX1DVber6lbgHuCULPZfo6pPqepe4HngMOw/dNTbikhdoBVwu6ruVtW5wLTM3jDKGJ9V1f+p6g5gMtA8WN4beEdV56jqLuC24DPIzKtAXwARqQCcGSxDVeer6ueqmqqqq4EnM4gjI+cH8X2jqn9if9gif75Zqvq1qu5T1cXB+0VzXLA/DCtU9cUgrleBZcDZEdtk9tlk5QQgCfhn8DuaAbxD8NkAe4AmIlJRVf9Q1a8ilh8G1FPVPar6sfoAWwXOE71br6o7016ISDkReTJo2tiCNRVUjmy+SOeXtCequj14mpTDbQ8HNkYsA/gxs4CjjPGXiOfbI2I6PPLYQaLdkNl7YVfv54pIaeBc4CtVXRPEcVTQLPFLEMe92NV9dg6KAViT7udrIyIzg6apzcDgKI+bduw16ZatAWpFvM7ss8k2ZlWN/KMYedxe2B/BNSIyW0RODJaPBlYCH4jIKhEZEd2P4WLJE71Lf3V1HdAYaKOqFTnQVJBZc0ws/AxUEZFyEcvqZLF9XmL8OfLYwXtWzWxjVV2KJbSuHNxsA9YEtAxoFMRxc25iwJqfIr2CfaOpo6qVgCcijpvd1fBPWJNWpLrAuijiyu64ddK1r+8/rqrOU9UeWLPOVOybAqq6VVWvU9WGQHdguIiclsdYXA55onfpVcDavDcF7b135PcbBlfIKcBIESkVXA2encUueYnxdaCbiJwUFE5Hkf3/g1eAa7A/KK+li2MLsE1EjgaGRBnDZGCAiDQJ/tCkj78C9g1np4i0xv7ApFmPNTU1zOTY7wJHiciFIlJCRC4AmmDNLHnxBXb1f4OIlBSRDtjvaGLwO7tIRCqp6h7sM9kHICLdROTIoBazGatrZNVU5vKBJ3qX3sNAWeB34HPgvwX0vhdhBc0NwN3AJKy/f0ZyHaOqLgGuwpL3z8AfWLEwK2lt5DNU9feI5ddjSXgr8FQQczQxvBf8DDOwZo0Z6Ta5EhglIluB2wmujoN9t2M1iU+CniwnpDv2BqAb9q1nA3AD0C1d3DmmqruxxN4V+9wfB/qp6rJgk4uB1UET1mDs9wlWbP4I2AZ8BjyuqjPzEovLOfG6iItHIjIJWKaq+f6NwrlE51f0Li6ISCsROUJEigXdD3tgbb3OuTzyO2NdvKgJvIkVRtcCQ1R1QbghOZcYvOnGOecSnDfdOOdcgou7pptq1app/fr1ww7DOecKlfnz5/+uqtUzWhd3ib5+/fqkpKSEHYZzzhUqIpL+juj9vOnGOecSnCd655xLcJ7onXMuwUWV6EWki4gsF5GVGY0+F0x8sF5EFgaPyyLW9ReRFcEj7ucPdc65RJNtMTYY+nUsNknCWmCeiEwLRvWLNElVh6bbN23AqWRs1L35wb5/xCR655xz2Yrmir41sFJVVwUDG03Ebk+PRmfgQ1XdGCT3D4EuuQvVOedcbkST6Gtx8CQJazl4EoM0vURksYi8LiJpY21Hta+IDBKRFBFJWb9+fZShO+eci0asirFvA/VVtRl21f58TnZW1fGqmqyqydWrZ9jf3znnElZqKkyaBE89lT/HjybRr+Pg2XBqk262mmAOz7Sxw58GWka7r3POFVVbt8K//w1HHgl9+sCzz0J+DD8WTaKfBzQSkQbBjDx9SDdxs4gcFvGyO/Bt8Px94AwROUREDgHOCJY551yR9dNPcNNNULcuDBtm/771FsydC5IPk3Zm2+tGVVNFZCiWoIsDz6jqEhEZBaSo6jTgahHpDqQCG4EBwb4bReQu7I8FwChV3Rj7H8M55+LfN9/Agw/Cyy/D3r3Qqxdcdx20aZO/7xt3wxQnJyerj3XjnEsUqjBjBowZA//9L5QrBwMHwrXXQsPMZv7NBRGZr6rJGa2Lu0HNnHMuEezZA6+9Zgl+wQI49FC4+24YPBiqVi3YWDzRO+dcDG3ZAk8/DQ8/DD/+CEcfba8vugjKlAknJk/0zjkXA2vXwiOPwJNPWrLv0AHGjYOuXaFYyKOKeaJ3zrk8WLTICqyvvgr79sF558H110Nyhq3l4fBE75xzOaQKH35o7e8ffgjly8NVV1lXyXicIM8TvXPORWn3bruDdcwYWLwYDjsM7rsPrrgCDjkk7Ogy54neOeeysXkzjB9vd7GuWwdNm9pdrH37QunSYUeXPU/0zjmXiR9+sOT+1FM2XMFpp1kPms6d8+cO1vziid4559JZsMCaZyZNstd9+tgdrC1ahBtXbnmid845rMD63/9agp8xA5KS4Jpr7FG3btjR5Y0neudckbZrl3WNHDMGliyBWrXggQfg8suhcuWwo4sNT/TOuSLpjz/s5qZHHoGff4ZmzeCFF+CCC6BUqbCjiy1P9M65ImX1ahue4Omn4c8/4fTT4bnn7N/CVGDNCU/0zrkiISXFmmdee82GJOjb1wqsxx0XdmT5zxO9cy5h7dsH771nCX7WLKhY0ZL71VdD7dphR1dwPNE75xLOzp02uceDD8K330KdOvb8ssss2Rc1nuidcwljwwZ44gl49FH49Vdo3twS/nnnQcmSYUcXHk/0zrlCb9Uq+Ne/4JlnYPt26NLFRpA89dTELbDmRMijJMfYl19ao5xzrkj44gs4/3xo1Mi6Sp5/Pnz9tbXLn3aaJ/k0iZPoly2Ddu2gd2/Yti3saJxz+WTfPpg2Ddq3hxNOgA8+gBtusG6Tzz4Lxx4bdoTxJ6pELyJdRGS5iKwUkRFZbNdLRFREkoPX9UVkh4gsDB5PxCrwv2jc2Kotb70FJ51koxE55xLGjh02guQxx0CPHvZfPG26vvvug8MPDzvC+JVtG72IFAfGAqcDa4F5IjJNVZem264CcA3wRbpDfKeqzWMUb1aBWp+pxo3t1rZWrWDqVDjxxHx/a+dc/vn9d3j8cXjsMVi/Hlq2hIkToVcvKOFVxqhEc0XfGlipqqtUdTcwEeiRwXZ3AfcDO2MYX8517gyff259qDp0gBdfDDUc51zurFgBV15pA4rdcQe0bg0zZ8K8eXYt50k+etEk+lrAjxGv1wbL9hOR44E6qvqfDPZvICILRGS2iJyc0RuIyCARSRGRlPXr10cbe+aOPtqqNO3aQb9+MGKEF2mdKyQ++wzOPde+nE+YABdeaIONvfOOXbt5gTXn8lyMFZFiwEPAdRms/hmoq6otgOHAKyLyl9sVVHW8qiaranL16tXzGpKpUgXefx8GD4b777czx4u0zsWlvXthyhS7Nmvb1u5ivflmWLPGxqRp0iTsCAu3aBL9OqBOxOvawbI0FYBjgVkisho4AZgmIsmquktVNwCo6nzgO+CoWAQelZIlrXHv0UftcqBdOztznHNxYft2GDfOvoSfe66NIvnoo1ZgvftuqFkz7AgTQzSJfh7QSEQaiEgpoA8wLW2lqm5W1WqqWl9V6wOfA91VNUVEqgfFXESkIdAIWBXznyIrIjB0qHWsXbPGirSffFKgITjnDvbbb9buXreutcNXqWKDja1YYf9dy5cPO8LEkm2iV9VUYCjwPvAtMFlVl4jIKBHpns3u7YHFIrIQeB0YrKob8xp0rpx+urXbV65st8s9/3woYThXlC1fDldcYQl+1Cj7kj1njvWf6N0bihcPO8LEJKoadgwHSU5O1pSUlPx7g40b7fa56dPhH/+wDrh+djmXb1TtS/SYMXajU6lS0L8/DB9uBVcXGyIyX1WTM1qXOHfGRqtKFWvGufJKGD0aeva06d2dczG1dy+8/rrdynLyyTB3Ltx2m93o9OSTnuQLUtFL9GBF2rFj7fHuu1bmX7067KicSwh//mk3Nx11lI0amXbD0w8/wJ13Qo0aYUdY9BTNRJ/myitt2ve1a61IO3du2BE5V2j98gvcequN/f5//weHHgpvvGHt8kOGQLlyYUdYdBXtRA/QqZMVaatUsSLtc8+FHZFzhcq338Lll0O9enDvvXZT0yefwKefWpdJL4GFzxM92HfMzz+HU06BSy6xgaz37g07KufilirMng1nn203M730Elx6qV29v/mmtYa6+OGJPs0hh1iRduhQGwWzRw/YsiXsqJyLK6mpMGmSjTvToYN9Gb7zTmt/f/xxGxfexR9P9JFKlLDb8saNs7b7tm1t6hrnirht2+Df/4Yjj4Q+fewa6Mkn7R7E22+HWI1c4vKHJ/qMDB5ssxn89JNdusyZE3ZEzoXi559tzJk6dWDYMPt36lRrlx80CMqWDTtCFw1P9Jk59VT7XlqtmhVsJ0wIOyLnCsySJTBwoBVY77/f/gt89hl8/LG1ahbzzFGo+K8rK40aWZG2Y0e47DK47jov0rqEpQozZsCZZ9p0fJMm2XAF//ufjUNzwglhR+hyyxN9dipXhv/8x2aveugh62aweXPYUTkXM3v2wCuv2MxNp50GX31lI0f+8IOVrI44IuwIXV55oo9GiRJWiXrySfjwQ7un+7vvwo7KuTzZutWuXY44Ai66yOZkffppu0n8llugatWwI3Sx4ok+JwYNsiLtr79akXb27LAjci7H1q2DG2+0wup110HDhvD229Yuf+mlUKZM2BG6WPNEn1MdO1qRtkYNq1A99VTYETkXlcWLbdTI+vVtJMkuXeDLL202p27dvMCayPxXmxtHHmlF2k6d7Cp/2DC7k8S5OKNqrY2dO8Nxx9nYM1ddBStXwsSJNsSTS3ye6HOrUiX7vjtsmLXfd+vmRVoXN3bvhhdfhObN4Ywz7Gr+vvtsir6HH4YGDcKO0BUkT/R5UaIE/Otf1nwzfbr1P1u5MuyoXBG2ebM1yzRsCP36WW/gZ5+1AuuIETbShyt6PNHHwmWXwUcfwfr10KYNzJwZdkSuiPnxRxuLr04dmzitcWObauHrr2HAAChdOuwIXZg80cfKKadYZatmTfuu/OSTYUfkioAFC+Dvf7cr+Icftts85s+3L5hdu4JI2BG6eBBVoheRLiKyXERWisiILLbrJSIqIskRy24K9lsuIp1jEXTcatjQ7hM/4wwbL+fqq71I62JO1cbc69QJjj8e3nrLTrVVq+Dll22Zc5GyTfQiUhwYC3QFmgB9RaRJBttVAK4BvohY1gToAzQFugCPB8dLXBUr2gzIw4fbbYVnnQWbNoUdlUsAu3bZvDjNmtnV+rJl8MAD1mzz4INQt27YEbp4Fc0VfWtgpaquUtXdwESgRwbb3QXcD+yMWNYDmKiqu1T1e2BlcLzEVry4/c97+mlrrz/hBFixIuyoXCG1aRP885/WU+aSS6w55vnn7Qr+H/+wUTqcy0o0ib4W8GPE67XBsv1E5Higjqr+J6f7BvsPEpEUEUlZv359VIEXCpdeakXaDRusSDtjRtgRuUJk9Wq49lorsN50kw009v77sGiR9agpVSrsCF1hkedirIgUAx4CrsvtMVR1vKomq2py9USbwaB9eyvSHn64td2PGxd2RC7OzZ8PffvafXmPPQY9e8LChTb6xhlneIHV5Vw0iX4dUCfide1gWZoKwLHALBFZDZwATAsKstntWzQ0aGAzJXfpAldeadMVepHWRdi3zwZJ7dgRkpOta+Tw4fD99/DCC3ZXq3O5FU2inwc0EpEGIlIKK65OS1upqptVtZqq1lfV+sDnQHdVTQm26yMipUWkAdAI+DLmP0VhULGidY+4/noYO9aqaX/8EXZULmS7dsEzz8Df/mY3V3/3nZV3fvzRCq21a4cdoUsEJbLbQFVTRWQo8D5QHHhGVZeIyCggRVWnZbHvEhGZDCwFUoGrVLXoztxRvDiMHg1Nm9oYOSecYMMoHHVU2JG5ArZxIzzxBDzyiA2G2rw5vPQSnH8+lCwZdnQu0Yiqhh3DQZKTkzUlJSXsMPLf3LnW+JqaatP3dOoUdkSuAKxaZTc2TZgA27dba97119vMld727vJCROaranJG6/zO2LCcdBLMm2ffzbt0seYcl7C+/NKu1hs1siv588+3gcbee89mdfIk7/KTJ/ow1a9vRdozz7QC7ZVX2rxuLiHs22f3zrVvb71rP/gAbrjBuk0++6y1yztXEDzRh61CBZgyxTLAuHFWpN24MeyoXB7s3GkDmjZpAj162NyrDz9sBdb77rOets4VJE/08aB4cbj/fru//eOP7fJv+fKwo3I59PvvcNddUK+e1drLl4dXX7WRq6+5xv6mOxcGT/TxpH9/GzJhy5YD3/Vd3Fu50mZtqlsXbr/dZm2aORNSUqBPH5u2wLkweaKPN23bWuWuXj1ru3/0URuu0MWdzz6DXr2sd+zTT8OFF9oE2++8Ax06eIHVxQ9P9PGoXj345BO7g+bqq2HIEC/Sxom9e2HqVGjXzv4mz5wJN98Ma9ZYsm/yl3FdnQuff6mMV0lJ8OabcMstNnTh//5n/e2rVg07soSlan3bN22yx+bNBz//7Tcb733FChvV4pFHYOBAa4t3Lp55oo9nxYpZN40mTWy6wjZt7E7aY44JO7K4tHevlTcyStLRPt+bzX3brVvD5Ml2r5u3vbvCwk/VwuDii20ow3POsWETJk2ym6wSzK5dByfdnCbrrVuzf4+kJBu/vXJlqFQJDjvM/m5WqnTw8vTPK1WyR9my+f85OBdrnugLixNPtDtpu3e3Waseesja7+Ok4qcKf/6Zu6votOc7d2b9HsWK/TUBH3nkwck4q2RdsaJfhbuiyU/7wqRuXRsj5+KLYdgw6+Lx2GMxmYEiNTX6Zo/MlmXX7FGmzF8TcL160SXpypWtLTxO/q45V6h4oi9skpLgjTfgttvg3nutSPvGG+wsXzXX7dKbNsG2bdm/dYUKByfgWrVsIM5omz3KlMn/j8c591ee6OOAqiXaaBPz5s3F2LTpHjbVvJHNs7ezqXoSu7Lpal+8+F8TcaNG2V9FRzZ7FE/sad2dS1ie6GMgNdWScG6vqDdvtgGwslK27MEJ+JBDoEGDilTevYNKHzxF5dQNVL6sN5XaNs0wYZcr580ezhVVRT7Rq1oRMC9JOppmj4oVD75KrlPHRi+MppBYqRKULp3ZkQ+FH3tYkXbcKDhiDPQd5lndObdfwiT6PXtg2bLcJevdu7M+dokSf03ANWtG3+xRoUI+N3vUqWNF2n79bKLRJUvg8cdjUqR1zhV+CZPoN26EZs0yXleu3MEJuGpVOOKI6Ht7lC1bCC6Qy5e3O2dHjrQhFFessKJttWphR+acC1nCJPoqVeD11zNu9igyF7bFisGoUXYH0CWX2G2cb79tXWOcc0VWwiT6kiVtJEEH9O1rX1l69LAbrV591W6ycs4VSVGNXikiXURkuYisFJERGawfLCJfi8hCEZkrIk2C5fVFZEewfKGIPBHrH8BlonVru5O2USM4+2x48EEf7ti5IirbRC8ixYGxQFegCdA3LZFHeEVV/6aqzYEHgIci1n2nqs2Dx+BYBe6iULu2zVjVqxdcfz1ceqkNKOOcK1KiuaJvDaxU1VWquhuYCPSI3EBVt0S8LA/4pWO8KFfOBkG7/XabkbpTJ1i/PuyonHMFKJpEXwv4MeL12mDZQUTkKhH5DruivzpiVQMRWSAis0Xk5IzeQEQGiUiKiKSs9yQUe8WKwZ13wsSJNr9d69bwzTdhR+WcKyAxm2FKVceq6hHAjcCtweKfgbqq2gIYDrwiIhUz2He8qiaranL16tVjFZJL74ILYM4ca7458USb8845l/CiSfTrgDoRr2sHyzIzETgHQFV3qeqG4Pl84DvgqNyF6mKiVSsr0jZubHfTjh7tRVrnElw0iX4e0EhEGohIKaAPMC1yAxFpFPHyLGBFsLx6UMxFRBoCjYBVsQjc5UGtWnZlf955cMMN1ufei7TOJaxs+9GraqqIDAXeB4oDz6jqEhEZBaSo6jRgqIh0AvYAfwD9g93bA6NEZA+wDxisqhvz4wdxOVSunLXZN20Kd9wBK1faHLU1aoQdmXMuxkTj7Gt7cnKypqSkhB1G0fLaa9C/vyX5adMyH0vCORe3RGS+qiZntC5mxVhXiJ13nvW337MH2ra1ZO+cSxie6J1p2dKKtE2a2CTk99/vRVrnEoQnenfA4YfD7NnWDXPECBgwIPsZu51zcc8TvTtY2bLwyis2CuYLL8Cpp8Kvv4YdlXMuDzzRu78SscnHX38dFi2yvveLFoUdlXMulzzRu8z16mVF2n37oF07mDo17Iicc7ngid5l7fjjrUjbtCn07An33edFWucKGU/0LnuHHQazZsGFF8LNN8PFF3uR1rlCxBO9i07ZsvDSS3DPPfDyy9CxI/zyS9hROeei4IneRU/ErujffBMWL7bhjhcsCDsq51w2PNG7nOvZEz75xJ6fdJIlfudc3PJE73KneXP48ksbF6dXL2vS8SKtc3HJE73LvZo1YeZM+Pvf4dZb4aKLYMeOsKNyzqXjid7lTZkydgftfffZsMcdOsDPP4cdlXMugid6l3ciNjbOlCmwZIkVab/6KuyonHMBT/Qudnr0sCJtsWJWpH3jjbAjcs7hid7F2nHHWZG2eXPo3dsGR/MirXOh8kTvYu/QQ61I26+fTVN44YVepHUuRJ7oXf4oXRqee84mMJk0CU45BX76KeyonCuSokr0ItJFRJaLyEoRGZHB+sEi8rWILBSRuSLSJGLdTcF+y0WkcyyDd3FOBG64wUa9XLrUhjv2+YCdK3DZJnoRKQ6MBboCTYC+kYk88Iqq/k1VmwMPAA8F+zYB+gBNgS7A48HxXFHSvTt8+imULAnt28PkyWFH5FyREs0VfWtgpaquUtXdwESgR+QGqrol4mV5IK361gOYqKq7VPV7YGVwPFfUNGtmRdrjj7epCu+804u0zhWQaBJ9LeDHiNdrg2UHEZGrROQ77Ir+6hzuO0hEUkQkZf369dHG7gqbGjVg+nSbi3bkSOjTB7ZvDzsq5xJezIqxqjpWVY8AbgRuzeG+41U1WVWTq1evHquQXDwqXRqeeQZGj4bXXrOmnHXrwo7KuYQWTaJfB9SJeF07WJaZicA5udzXFQUicP31MG0aLF9uRdp588KOyrmEFU2inwc0EpEGIlIKK65Oi9xARBpFvDwLWBE8nwb0EZHSItIAaAR8mfewXULo1g0++8yu8tu3t26YzrmYK5HdBqqaKiJDgfeB4sAzqrpEREYBKao6DRgqIp2APcAfQP9g3yUiMhlYCqQCV6nq3nz6WVxhdOyxVqTt1cva7JcutZusivktHs7Fimic9XxITk7WFO9rXfTs3g1Dhlj7fe/e8PzzUK5c2FE5V2iIyHxVTc5onV82ufhQqhQ8/TQ8+KANhnbyybB2bdhROZcQPNG7+CECw4fDO+/AihVWpP3ii7Cjcq7Q80Tv4s+ZZ1qRtmxZGyPn1VfDjsi5Qs0TvYtPTZtakbZNGxv98rbbYN++sKNyrlDyRO/iV7Vq8OGHcOmlcPfdcN558OefYUflXKHjid7Ft1Kl4Kmn4F//slEwTzoJfvwx+/2cc/t5onfxTwSGDbMi7apVXqR1Loc80bvCo2tX+PxzKF/eirQvvxx2RM4VCp7oXeFyzDFWpD3xRPj73+GWW7xI61w2PNG7wqdqVXj/fRg0CO691+6k3bYt7Kici1ue6F3hVKoUPPEE/Pvf8NZbVqT94Yewo3IuLnmid4WXCFx9Nbz7LqxebUXazz4LOyrn4o4nelf4de5sRdqKFaFDB3jxxbAjci6ueKJ3ieHoo63LZbt20K8fjBjhRVrnAp7oXeKoUsWKtIMHw/33Q8+esHVr2FE5FzpP9C6xlCwJjz8Ojz4K//mPXeGvWRN2VM6FyhO9SzwiMHQovPee9cRp1Qo++STsqJwLjSd6l7hOP93a7StXhlNPtVmrnCuCPNG7xNa4sfXIOflkGDAAbrgB9vq0xa5o8UTvEl+VKtaMc+WVMHq0F2ldkRNVoheRLiKyXERWisiIDNYPF5GlIrJYRKaLSL2IdXtFZGHwmBbL4J2LWsmSMHasPd59F9q2tZusnCsCsk30IlIcGAt0BZoAfUWkSbrNFgDJqtoMeB14IGLdDlVtHjy6x52GXSMAABgsSURBVChu53Lnyivhv/+1icdbtYK5c8OOyLl8F80VfWtgpaquUtXdwESgR+QGqjpTVbcHLz8Hasc2TOdiqFMnK9JWqWJF2ueeCzsi5/JVNIm+FhA5pc/aYFlmLgXei3hdRkRSRORzETknox1EZFCwTcr69eujCMm5PDrqKCvSnnIKXHIJXH+9F2ldwoppMVZE/g4kA6MjFtdT1WTgQuBhETki/X6qOl5Vk1U1uXr16rEMybnMHXKIFWmHDoUHH4QePWDLlrCjci7mokn064A6Ea9rB8sOIiKdgFuA7qq6K225qq4L/l0FzAJa5CFe52KrRAm7i3bcOGu7b9vWpit0LoFEk+jnAY1EpIGIlAL6AAf1nhGRFsCTWJL/LWL5ISJSOnheDWgHLI1V8M7FzODB8MEH8NNP0Lo1zJkTdkTOxUy2iV5VU4GhwPvAt8BkVV0iIqNEJK0XzWggCXgtXTfKY4AUEVkEzAT+qaqe6F18OvVUK9JWq2YF2wkTwo7IuZgQVQ07hoMkJydrSkpK2GG4omzTJrjgArvCv/Zau8mqePGwo3IuSyIyP6iH/oXfGetcepUr28iXV18N//oXnH02bN4cdlTO5ZoneucyUqKEzUf75JPw4Ydw4onw3XdhR+Vcrniidy4rgwZZE86vv1qRdsIE2LAh7KicyxFP9M5lp2NHK9LWqgWXXQY1atiyf//bx8txhYIneueiceSRsGgRpKTAzTfD77/DsGHQoAEcfzyMGgWLF0OcdW5wDrzXjXO5t2IFvPUWTJ0Kn35qSb5hQzjnHHu0beu9dVyByarXjSd652Lhl1/g7bct6X/0EezeDdWrW4+dnj2tX36ZMmFH6RKYJ3rnCtLWrTacwpQp1k1zyxYoXx66dLGkf+aZNs6OczHkid65sOzeDbNmWdJ/6y34+WfrutmhgzXv9OgBtX1Ub5d3nuidiwf79sG8eda8M2UKLF9uy1u1OtCuf8wxIBJunK5Q8kTvXDxatsyS/tSp1n0TbJz8tKTfpg0U845xLjqe6J2Ldz/9dKAHz4wZkJoKNWtC9+7Wrt+xI5QuHXaULo55oneuMNm0ySYwnzrVJkbZtg0qVLAi7jnn2L8VK4YdpYsznuidK6x27rQr/KlT7Yr/t9+gZEk47TRL+t27w2GHhR2liwOe6J1LBHv32jy3acXc776zwu0JJxxo1z/qqLCjdCHxRO9colGFJUsOFHPnz7flxxxjbfrnnAPJyd6DpwjxRO9covvhhwPF3Nmz7eq/Vi3rp9+zJ5xyijX5uITlid65omTjRrsjd8oUu0N3xw6bTOWssyzpd+4MSUlhR+lizBO9c0XV9u029s7UqTBtmo2lX7o0nH76gWJu9ephR+liIM9TCYpIFxFZLiIrRWREBuuHi8hSEVksItNFpF7Euv4isiJ49M/9j+Gcy7Fy5SyZP/OMDbw2axYMGQJff21j69esCe3bw0MPwapVYUfr8km2V/QiUhz4H3A6sBaYB/RV1aUR23QEvlDV7SIyBOigqheISBUgBUgGFJgPtFTVPzJ7P7+id64AqNr4+mnF3EWLbPnf/nagmNu8uRdzC5G8XtG3Blaq6ipV3Q1MBHpEbqCqM1V1e/DycyBtlKbOwIequjFI7h8CXXLzQzjnYkjEEvnIkbBwoXXVfOghG1Xz7rttMpUGDWxylVmz7E5dV2hFk+hrAT9GvF4bLMvMpcB7OdlXRAaJSIqIpKxfvz6KkJxzMdWwIVx7rfXY+eUXa+pp1gyeeMKGXzj0UBgwwK7+t2/P9nAuvpSI5cFE5O9YM80pOdlPVccD48GabtKv37NnD2vXrmXnzp0xidPlrzJlylC7dm1Kene+wql6dbjkEnts22aTo6fdmfv881C2rPXcOecc6NYNqlYNO2KXjWgS/TqgTsTr2sGyg4hIJ+AW4BRV3RWxb4d0+87KaZBr166lQoUK1K9fH/E2w7imqmzYsIG1a9fSoEGDsMNxeZWUBOeea489e2DOnAPt+lOn2lSJ7dsfGFu/Xr3sj+kKXDRNN/OARiLSQERKAX2AaZEbiEgL4Emgu6r+FrHqfeAMETlERA4BzgiW5cjOnTupWrWqJ/lCQESoWrWqf/tKRGlj7Dz6qN2gNW8ejBhh4+9ccw3Ur29t+3fdZb164qzrdlGWbaJX1VRgKJagvwUmq+oSERklIt2DzUYDScBrIrJQRKYF+24E7sL+WMwDRgXLcsyTfOHhv6siQMSGWLj7bvjmG/jf/+CBB6xZ5447rH3/yCPhuutg7ly7U9eFplDcMPXtt99yzDHHhBSRyw3/nRVhaROlT5kC06cfmCg9bWz9007zidLzQZ5vmCrqNmzYQPPmzWnevDk1a9akVq1a+1/v3r07y31TUlK4+uqrs32Ptm3bxiTWWbNm0a1bt5gcy7lcqVkTLr/cxtRfvx4mTbLkPnmyFW+rVYPzzoOXX7ax912+i2mvm0RVtWpVFi5cCMDIkSNJSkri+uuv378+NTWVEiUy/iiTk5NJTs7wj+xBPv3009gE61w8qVgRzj/fHrt2WZ/8tB48r79uE6V37HigmFsrq57bLrcKX6IfNsxu8Iil5s3h4YdztMuAAQMoU6YMCxYsoF27dvTp04drrrmGnTt3UrZsWZ599lkaN27MrFmzGDNmDO+88w4jR47khx9+YNWqVfzwww8MGzZs/9V+UlIS27ZtY9asWYwcOZJq1arxzTff0LJlS1566SVEhHfffZfhw4dTvnx52rVrx6pVq3jnnXcyjXHjxo0MHDiQVatWUa5cOcaPH0+zZs2YPXs211xzDWDt6XPmzGHbtm1ccMEFbNmyhdTUVMaNG8fJJ5+c+8/UufRKl7ZumZ07w9ix8OWXB8bWv+oqe7RufWBs/aOP9jtzY6TwJfo4snbtWj799FOKFy/Oli1b+PjjjylRogQfffQRN998M2+88cZf9lm2bBkzZ85k69atNG7cmCFDhvylv/mCBQtYsmQJhx9+OO3ateOTTz4hOTmZK664gjlz5tCgQQP69u2bbXx33HEHLVq0YOrUqcyYMYN+/fqxcOFCxowZw9ixY2nXrh3btm2jTJkyjB8/ns6dO3PLLbewd+9etvtNMS4/FStmE6accAL88582UfqUKZb4b77ZHkcddWA4htatfaL0PCh8iT6HV9756bzzzqN48eIAbN68mf79+7NixQpEhD179mS4z1lnnUXp0qUpXbo0NWrU4Ndff6V27doHbdO6dev9y5o3b87q1atJSkqiYcOG+/um9+3bl/Hjx2cZ39y5c/f/sTn11FPZsGEDW7ZsoV27dgwfPpyLLrqIc889l9q1a9OqVSsGDhzInj17OOecc2jevHmePhvncuToo+Gmm+yxbt2BsfUffBDuv9/a/dPG1u/YEUqVCjviQsX/ROZB+fLl9z+/7bbb6NixI9988w1vv/12pv3IS5cuvf958eLFSc1gDJFotsmLESNG8PTTT7Njxw7atWvHsmXLaN++PXPmzKFWrVoMGDCAF154Iabv6VzUatWCK6+0O3LXr7ei7cknw0svQZcu1oOnb18r8m7ZEna0hYIn+hjZvHkztYJC0nPPPRfz4zdu3JhVq1axevVqACZNmpTtPieffDIvv/wyYL1xqlWrRsWKFfnuu+/429/+xo033kirVq1YtmwZa9as4dBDD+Xyyy/nsssu46uvvor5z+BcjlWuDBdeaD12fv8d3nnHCrvTp0OfPpb0zzwTxo+3bp0uQ57oY+SGG27gpptuokWLFjG/AgcoW7Ysjz/+OF26dKFly5ZUqFCBSpUqZbnPyJEjmT9/Ps2aNWPEiBE8//zzADz88MMce+yxNGvWjJIlS9K1a1dmzZrFcccdR4sWLZg0adL+Yq1zcaNMGZsl66mn4Oef4eOP4f/+D5YvhyuugMMPh3btYPRoWLEi7Gjjit8wVYhs27aNpKQkVJWrrrqKRo0ace2114YdVob8d+YKTNpE6WnF3LRvo02aWCG3Z09o2TLhe/D4DVMJ4qmnnqJ58+Y0bdqUzZs3c8UVV4QdknPhE4Fjj4XbboP582HNGnjkERta+f77oVUrqFsXhg61Jp9MOkokMr+id/nCf2cuLmzYYBOlT5168ETp3brZ1X6XLhDRqaIw8yt651zRVLUq9OsHb75pxdypUy3Bv/ce9O5twzGkzambwJMeFb5+9M45lxvlyllf/B49bGrEuXMPjKv/9tt2Q9ZJJx24MzeB5lPwK3rnXNFTogR06GA3YH7/PSxYALfeaoOsDR9uUysed9yBOXXjrIk7pzzRO+eKtrSJ0u+8ExYtOjBReqVKNolKixYHJkqfPbtQTpTuiT4KHTt25P33D54Y6+GHH2bIkCGZ7tOhQwfSispnnnkmmzIYjnXkyJGMGTMmy/eeOnUqS5cu3f/69ttv56OPPspJ+Bny4Yydy0TaROlz5lh//QkTDkyU3qGDDcdwySU2TMOOHWFHGxVP9FHo27cvEydOPGjZxIkToxpYDODdd9+lcuXKuXrv9Il+1KhRdOrUKVfHcs7lUI0aMHAgTJtmxdzXX4euXQ8UdatVs/l0X3gBNuZq8rwCUeiKsWGMUty7d29uvfVWdu/eTalSpVi9ejU//fQTJ598MkOGDGHevHns2LGD3r17c+edd/5l//r165OSkkK1atW45557eP7556lRowZ16tShZcuWgPWRHz9+PLt37+bII4/kxRdfZOHChUybNo3Zs2dz991388Ybb3DXXXfRrVs3evfuzfTp07n++utJTU2lVatWjBs3jtKlS1O/fn369+/P22+/zZ49e3jttdc4+uijM/35fDhj56KQlAS9etljzx5rxkkr5k6ZcmCi9J49reBbt27YEe/nV/RRqFKlCq1bt+a9994D7Gr+/PPPR0S45557SElJYfHixcyePZvFixdnepz58+czceJEFi5cyLvvvsu8efP2rzv33HOZN28eixYt4phjjmHChAm0bduW7t27M3r0aBYuXMgRRxyxf/udO3cyYMAAJk2axNdff70/6aapVq0aX331FUOGDMm2eShtOOPFixdz77330q9fP4D9wxkvXLiQjz/+mLJly/LKK6/QuXNnFi5cyKJFi3yUS1c0lSwJnTrBY4/Bjz/aROk33gi//gpXXw316tnduGlz6oZczC10V/RhjVKc1nzTo0cPJk6cyIQJEwCYPHky48ePJzU1lZ9//pmlS5fSrFmzDI/x8ccf07NnT8qVKwdA9+7d96/75ptvuPXWW9m0aRPbtm2jc+fOWcazfPlyGjRowFFHHQVA//79GTt2LMOGDQPsDwdAy5YtefPNN7M8lg9n7FwepE2UnpwM99xjE6W/9ZZd5d9+u92xe8QRB7ptnniiXf0XoKiu6EWki4gsF5GVIjIig/XtReQrEUkVkd7p1u0VkYXBY1qsAi9oPXr0YPr06Xz11Vds376dli1b8v333zNmzBimT5/O4sWLOeusszIdnjg7AwYM4LHHHuPrr7/mjjvuyPVx0qQNdZyXYY59OGPncuGoo+Af/4BPP7Wx9Z980pY9+qgNt3z44Tan7n/+A3n8fx6tbBO9iBQHxgJdgSZAXxFpkm6zH4ABwCsZHGKHqjYPHt0zWF8oJCUl0bFjRwYOHLi/CLtlyxbKly9PpUqV+PXXX/c37WSmffv2TJ06lR07drB161befvvt/eu2bt3KYYcdxp49e/YPLQxQoUIFtm7d+pdjNW7cmNWrV7Ny5UoAXnzxRU455ZRc/Ww+nLFz+eSww2DQoAMTpU+cCKeeamPpd+tmwyyffz688kq+TpQeTdNNa2Clqq4CEJGJQA9gf1cQVV0drNuXDzHGjb59+9KzZ8/9PXDShvU9+uijqVOnDu3atcty/+OPP54LLriA4447jho1atCqVav96+666y7atGlD9erVadOmzf7k3qdPHy6//HIeeeQRXn/99f3blylThmeffZbzzjtvfzF28ODBufq5Ro4cycCBA2nWrBnlypU7aDjjmTNnUqxYMZo2bUrXrl2ZOHEio0ePpmTJkiQlJfkVvXPRqlgRLrjAHrt2wcyZByZKf+01u4mrVy/7YxBj2Q5qFjTFdFHVy4LXFwNtVHVoBts+B7yjqq9HLEsFFgKpwD9VdWoG+w0CBgHUrVu35Zo1aw5a7wNkFT7+O3MuSvv22UTpU6ZYsr/nnlwdJqtBzQqiGFtPVdeJSENghoh8rarfRW6gquOB8WCjVxZATM45Fx8iJ0rPr7eIYpt1QJ2I17WDZVFR1XXBv6uAWUCLHMTnnHMuj6JJ9POARiLSQERKAX2AqHrPiMghIlI6eF4NaEdE235OxNu4+S5z/rtyLr5km+hVNRUYCrwPfAtMVtUlIjJKRLoDiEgrEVkLnAc8KSJLgt2PAVJEZBEwE2ujz3GiL1OmDBs2bPAEUgioKhs2bKBMmTJhh+KcCxSKGab27NnD2rVr89y33BWMMmXKULt2bUqWLBl2KM4VGWEXY/OsZMmSNEigSQCcc64g+Vg3zjmX4DzRO+dcgvNE75xzCS7uirEish5Yk+2GmasG/B6jcGLJ48oZjytnPK6cScS46qlq9YxWxF2izysRScms8hwmjytnPK6c8bhypqjF5U03zjmX4DzRO+dcgkvERD8+7AAy4XHljMeVMx5XzhSpuBKujd4559zBEvGK3jnnXARP9M45l+AKTaIXkWdE5DcR+SaT9SIijwQTmC8WkeMj1vUXkRXBo38Bx3VREM/XIvKpiBwXsW51sHyhiKRktH8+xtVBRDZHTNx+e8S6LCeDz+e4/hER0zfB5PJVgnX5+XnVEZGZIrJURJaIyDUZbFOg51iUMYV1fkUTW4GfY1HGVeDnmIiUEZEvRWRRENedGWxTWkQmBZ/JFyJSP2LdTcHy5SLSOccBqGqheADtgeOBbzJZfybwHiDACcAXwfIqwKrg30OC54cUYFxt094Pm2D9i4h1q4FqIX1eHbBpH9MvLw58BzQESgGLgCYFFVe6bc8GZhTQ53UYcHzwvALwv/Q/d0GfY1HGFNb5FU1sBX6ORRNXGOdYcM4kBc9LAl8AJ6Tb5krgieB5H2BS8LxJ8BmVBhoEn13xnLx/obmiV9U5wMYsNukBvKDmc6CyiBwGdAY+VNWNqvoH8CHQpaDiUtVPg/cF+ByboSvfRfF5ZWb/ZPCquhtImww+jLj6Aq/G6r2zoqo/q+pXwfOt2NwLtdJtVqDnWDQxhXh+RfN5ZSbfzrFcxFUg51hwzmwLXpYMHul7wvQAng+evw6cJiISLJ+oqrtU9XtgJfYZRq3QJPoo1AJ+jHi9NliW2fIwXIpdEaZR4AMRmS82QXpBOzH4KvmeiDQNlsXF5yUi5bBk+UbE4gL5vIKvzC2wq65IoZ1jWcQUKZTzK5vYQjvHsvvMCvocE5HiIrIQ+A27MMj0/FKb8GkzUJUYfF6FYjz6RCAiHbH/iCdFLD5JbeL0GsCHIrIsuOItCF9hY2NsE5EzgalAowJ672icDXyiqpFX//n+eYlIEvYff5iqbonlsXMrmpjCOr+yiS20cyzK32OBnmOquhdoLiKVgSkicqyqZlirirVEuqLPbBLzPE1uHgsi0gx4GuihqhvSluuBidN/A6aQw69jeaGqW9K+Sqrqu0BJsXl9Q/+8An1I95U6vz8vESmJJYeXVfXNDDYp8HMsiphCO7+yiy2scyyazyxQ4OdYcOxN2NSq6Zv39n8uIlICqARsIBafV6yLDvn5AOqTeXHxLA4ulH0ZLK8CfI8VyQ4JnlcpwLjqYm1qbdMtLw9UiHj+KdClAOOqyYEb5loDPwSfXQmsmNiAA4WypgUVV7C+EtaOX76gPq/gZ38BeDiLbQr0HIsyplDOryhjK/BzLJq4wjjHgOpA5eB5WeBjoFu6ba7i4GLs5OB5Uw4uxq4ih8XYQtN0IyKvYlX8amITkd+BFTRQ1SeAd7FeESuB7cAlwbqNInIXMC841Cg9+Ktafsd1O9bO9rjVVUhVG53uUOzrG9iJ/4qq/rcA4+oNDBGRVGAH0EftrEoVkbTJ4IsDz6jqkgzeIr/iAugJfKCqf0bsmq+fF9AOuBj4OmhHBbgZS6RhnWPRxBTK+RVlbGGcY9HEBQV/jh0GPC8ixbGWlMmq+o6IjAJSVHUaMAF4UURWYn+E+gQxLxGRycBSIBW4Sq0ZKGo+BIJzziW4RGqjd845lwFP9M45l+A80TvnXILzRO+ccwnOE71zziU4T/TOOZfgPNE751yC+383QnyuhZE7ZQAAAABJRU5ErkJggg==\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "epochs = range(1, len(acc) + 1)\n",
        "\n",
        "plt.plot(epochs, acc, 'b', label='Training acc',color=\"red\")\n",
        "plt.plot(epochs, val_acc, 'b', label='Validation acc')\n",
        "plt.title('Training and validation accuracy')\n",
        "plt.legend()\n",
        "\n",
        "plt.figure()\n",
        "\n",
        "plt.plot(epochs, loss, 'b', label='Training loss',color=\"red\")\n",
        "plt.plot(epochs, val_loss, 'b', label='Validation loss')\n",
        "plt.title('Training and validation loss')\n",
        "plt.legend()\n",
        "\n",
        "plt.show()\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 13,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-10-27T13:34:13.942783Z",
          "start_time": "2021-10-27T13:33:56.775746Z"
        },
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "BoYt6sddbXve",
        "outputId": "81a0d81e-0588-4f9a-f3aa-3da3679ebbaf"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "782/782 [==============================] - 27s 35ms/step - loss: 0.5182 - accuracy: 0.8327\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.8327199816703796"
            ]
          },
          "metadata": {},
          "execution_count": 13
        }
      ],
      "source": [
        "score = model.evaluate(X_test, y_test)\n",
        "score[1]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rvXKpihRbXvg"
      },
      "source": [
        "# BLSTM"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "layers = 128\n",
        "dropout = 0.2\n",
        "epochs = 3\n",
        "batch_size = 128"
      ],
      "metadata": {
        "id": "s2HfuvkReoDU"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": 14,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-10-28T13:04:56.244802Z",
          "start_time": "2021-10-28T13:04:55.766740Z"
        },
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Pj8Ii42qbXvj",
        "outputId": "352345ed-9035-4564-8429-8111b19b43b1"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_1\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " embedding_1 (Embedding)     (None, 100, 128)          2560000   \n",
            "                                                                 \n",
            " bidirectional (Bidirectiona  (None, 256)              263168    \n",
            " l)                                                              \n",
            "                                                                 \n",
            " dropout_1 (Dropout)         (None, 256)               0         \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 1)                 257       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 2,823,425\n",
            "Trainable params: 2,823,425\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "model = Sequential()\n",
        "model.add(Embedding(max_features, 128, input_length=maxlen))\n",
        "model.add(Bidirectional(LSTM(layers)))\n",
        "model.add(Dropout(dropout))\n",
        "model.add(Dense(1, activation='sigmoid'))\n",
        "\n",
        "model.compile('adam', 'binary_crossentropy', metrics=['accuracy'])\n",
        "\n",
        "model.summary()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 18,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-10-28T13:08:05.415235Z",
          "start_time": "2021-10-28T13:04:56.510388Z"
        },
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "eiwn1ulbbXvl",
        "outputId": "b4125b0b-db01-4c55-8b96-8b976f884507"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/3\n",
            "157/157 [==============================] - 131s 809ms/step - loss: 0.4464 - accuracy: 0.7793 - val_loss: 0.3698 - val_accuracy: 0.8414\n",
            "Epoch 2/3\n",
            "157/157 [==============================] - 123s 782ms/step - loss: 0.2384 - accuracy: 0.9064 - val_loss: 0.3519 - val_accuracy: 0.8466\n",
            "Epoch 3/3\n",
            "157/157 [==============================] - 122s 776ms/step - loss: 0.1492 - accuracy: 0.9480 - val_loss: 0.4018 - val_accuracy: 0.8454\n"
          ]
        }
      ],
      "source": [
        "history = model.fit(X_train, y_train,\n",
        "                    epochs=epochs,\n",
        "                    batch_size=batch_size,\n",
        "                    validation_split = 0.2)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 19,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "5uO9KQxibXvn",
        "outputId": "10e4d098-6323-447c-c0ec-5e9122250443"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "782/782 [==============================] - 45s 57ms/step - loss: 0.4226 - accuracy: 0.8364\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.8364400267601013"
            ]
          },
          "metadata": {},
          "execution_count": 19
        }
      ],
      "source": [
        "score = model.evaluate(X_test, y_test)\n",
        "score[1]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8k7l35WdbXvp"
      },
      "source": [
        "# SimpleRNN"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 20,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "EXVxGwO7bXvq",
        "outputId": "06221dcc-dee0-47c1-d5bc-a26447b8b70b"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_2\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " embedding_2 (Embedding)     (None, 100, 128)          2560000   \n",
            "                                                                 \n",
            " simple_rnn (SimpleRNN)      (None, 128)               32896     \n",
            "                                                                 \n",
            " dropout_2 (Dropout)         (None, 128)               0         \n",
            "                                                                 \n",
            " dense_2 (Dense)             (None, 1)                 129       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 2,593,025\n",
            "Trainable params: 2,593,025\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "model = Sequential()\n",
        "model.add(Embedding(max_features, 128, input_length=maxlen))\n",
        "model.add(SimpleRNN(layers)) \n",
        "model.add(Dropout(dropout))\n",
        "model.add(Dense(1, activation='sigmoid'))\n",
        "\n",
        "model.compile('adam', 'binary_crossentropy', metrics=['accuracy'])\n",
        "\n",
        "model.summary()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 21,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DxgJ_ImTbXvs",
        "outputId": "2779edd2-ba6b-4995-aef2-0c4a99e3b557"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/3\n",
            "157/157 [==============================] - 22s 132ms/step - loss: 0.5834 - accuracy: 0.6723 - val_loss: 0.4860 - val_accuracy: 0.7752\n",
            "Epoch 2/3\n",
            "157/157 [==============================] - 21s 137ms/step - loss: 0.3110 - accuracy: 0.8699 - val_loss: 0.4139 - val_accuracy: 0.8276\n",
            "Epoch 3/3\n",
            "157/157 [==============================] - 20s 130ms/step - loss: 0.1765 - accuracy: 0.9337 - val_loss: 0.4768 - val_accuracy: 0.8016\n"
          ]
        }
      ],
      "source": [
        "history = model.fit(X_train, y_train,\n",
        "                    epochs=epochs,\n",
        "                    batch_size=batch_size,\n",
        "                    validation_split=0.2)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 22,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-10-28T13:08:39.483211Z",
          "start_time": "2021-10-28T13:08:05.420484Z"
        },
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "IEfX1mQtbXvt",
        "outputId": "a552e395-ff9a-4f29-97b3-a6fa62f5896a"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "782/782 [==============================] - 10s 13ms/step - loss: 0.4750 - accuracy: 0.8085\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.8085200190544128"
            ]
          },
          "metadata": {},
          "execution_count": 22
        }
      ],
      "source": [
        "score = model.evaluate(X_test, y_test)\n",
        "score[1]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IsrIXQ2qbXvv"
      },
      "source": [
        "# GRU"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 23,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "b72wpKCebXvw",
        "outputId": "e1faabf4-8acb-4e4e-aec7-11cb452c73e0"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_3\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " embedding_3 (Embedding)     (None, 100, 128)          2560000   \n",
            "                                                                 \n",
            " gru (GRU)                   (None, 128)               99072     \n",
            "                                                                 \n",
            " dropout_3 (Dropout)         (None, 128)               0         \n",
            "                                                                 \n",
            " dense_3 (Dense)             (None, 1)                 129       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 2,659,201\n",
            "Trainable params: 2,659,201\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "model = Sequential()\n",
        "model.add(Embedding(max_features, 128, input_length=maxlen))\n",
        "model.add(GRU(layers)) \n",
        "model.add(Dropout(dropout))\n",
        "model.add(Dense(1, activation='sigmoid'))\n",
        "\n",
        "model.compile('adam', 'binary_crossentropy', metrics=['accuracy'])\n",
        "\n",
        "model.summary()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 24,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "mPMjJLMXbXvw",
        "outputId": "d78a5af7-b413-46e9-d902-21e0df1a02a5"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/3\n",
            "157/157 [==============================] - 58s 358ms/step - loss: 0.5338 - accuracy: 0.7185 - val_loss: 0.3664 - val_accuracy: 0.8368\n",
            "Epoch 2/3\n",
            "157/157 [==============================] - 55s 351ms/step - loss: 0.2660 - accuracy: 0.8914 - val_loss: 0.3634 - val_accuracy: 0.8412\n",
            "Epoch 3/3\n",
            "157/157 [==============================] - 56s 357ms/step - loss: 0.1666 - accuracy: 0.9403 - val_loss: 0.4109 - val_accuracy: 0.8254\n"
          ]
        }
      ],
      "source": [
        "history = model.fit(X_train, y_train,\n",
        "                    epochs=epochs,\n",
        "                    batch_size=batch_size,\n",
        "                    validation_split=0.2)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 25,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "frHiP_FNbXvx",
        "outputId": "1735256d-68e7-407c-ccd3-991b5f60acc2"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "782/782 [==============================] - 21s 27ms/step - loss: 0.4161 - accuracy: 0.8242\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.8241599798202515"
            ]
          },
          "metadata": {},
          "execution_count": 25
        }
      ],
      "source": [
        "score = model.evaluate(X_test, y_test)\n",
        "score[1]"
      ]
    }
  ],
  "metadata": {
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.7.4"
    },
    "colab": {
      "name": "imdb-sentimental-analysis-rnn.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}