pneumonia.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "pneumonia.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "# **Kết nối Dr**"
      ],
      "metadata": {
        "id": "qsesm0H9NoUD"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Yqp9nwOqNnFR",
        "outputId": "3381c5bc-56e4-4301-fb17-8fa316b60a35"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# **Thư Viện**"
      ],
      "metadata": {
        "id": "9xpcX1tWNx2L"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import keras\n",
        "from keras.models import Sequential\n",
        "from glob import glob\n",
        "from keras.layers import Dense, Dropout, Flatten,Activation\n",
        "from keras.layers import Conv2D, MaxPooling2D, Conv3D, BatchNormalization\n",
        "from keras.preprocessing.image import ImageDataGenerator, load_img,img_to_array\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "from keras.preprocessing.image import load_img\n",
        "from keras.preprocessing.image import img_to_array\n",
        "from keras.preprocessing.image import array_to_img\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt"
      ],
      "metadata": {
        "id": "nKc1CuHrNrgw"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# **LOAD MODEL**"
      ],
      "metadata": {
        "id": "bnRn-5z-Cq7P"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from keras.models import load_model\n",
        "model = load_model('Hue.h5')"
      ],
      "metadata": {
        "id": "FNvvGGThUkG_"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "dudoan = {0: 'NORMAL',\n",
        "          1: 'PNEUMONIA'}"
      ],
      "metadata": {
        "id": "O9UQFDtf_5N2"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# **Loading Data**"
      ],
      "metadata": {
        "id": "Oe17UB7EN46A"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "test_path=\"/content/drive/MyDrive/pneumonia_xray/test\"\n",
        "train_path=\"/content/drive/MyDrive/pneumonia_xray/train\"\n",
        "val_path=\"/content/drive/MyDrive/pneumonia_xray/val\""
      ],
      "metadata": {
        "id": "BRfy-HOoN72R"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import os\n",
        "x1 =os.listdir( '/content/drive/MyDrive/pneumonia_xray/train/NORMAL')\n",
        "x2 =os.listdir( '/content/drive/MyDrive/pneumonia_xray/train/PNEUMONIA')"
      ],
      "metadata": {
        "id": "w3ZPnINSWvH7"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "sns.set_style('whitegrid')\n",
        "sns.barplot(x=['Normal','Pneumonia'],y=[len(x1),len(x2)])\n",
        "plt.title('Datasets')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 298
        },
        "id": "DcxZrhhWVAW6",
        "outputId": "e86249b1-7189-4b25-a884-9e5c0938d0b7"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0.5, 1.0, 'Datasets')"
            ]
          },
          "metadata": {},
          "execution_count": 15
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# **Model Create**"
      ],
      "metadata": {
        "id": "33wVm7C-OF0J"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "train_datagen = ImageDataGenerator(rescale=1/255,\n",
        "                                        rotation_range=0.2, #xoay hình\n",
        "                                        horizontal_flip=True,\n",
        "                                        vertical_flip=True,\n",
        "                                        zoom_range=0.2)\n",
        "\n",
        "test_datagen = ImageDataGenerator(rescale=1./255)\n",
        "\n",
        "train_set = train_datagen.flow_from_directory(train_path,\n",
        "                                                 target_size=(64, 64),\n",
        "                                                 batch_size=10,\n",
        "                                                 class_mode='categorical')\n",
        "\n",
        "\n",
        "\n",
        "test_set = test_datagen.flow_from_directory(test_path,\n",
        "                                            target_size=(64, 64),\n",
        "                                            batch_size=10,\n",
        "                                            class_mode='categorical')\n",
        "\n",
        "\n",
        "\n",
        "val_set = train_datagen.flow_from_directory(val_path,\n",
        "                                            target_size=(64, 64),\n",
        "                                            batch_size=10,\n",
        "                                            shuffle=True,\n",
        "                                            class_mode='categorical')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9vMaeBzAOR6M",
        "outputId": "4acc1385-df53-4dec-8b62-f18df33b9699"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found 5232 images belonging to 2 classes.\n",
            "Found 624 images belonging to 2 classes.\n",
            "Found 16 images belonging to 2 classes.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model = Sequential()\n",
        "model.add(Conv2D(64, kernel_size=(3,3), activation='relu', input_shape=(64, 64, 3)))\n",
        "model.add(MaxPooling2D(pool_size=(2,2)))\n",
        "\n",
        "model.add(Conv2D(32, kernel_size=(3,3), activation='relu'))\n",
        "model.add(MaxPooling2D(pool_size=(2,2)))\n",
        "\n",
        "model.add(Conv2D(16, kernel_size=(3,3), activation='relu'))\n",
        "model.add(MaxPooling2D(pool_size=(2,2)))\n",
        "\n",
        "model.add(Conv2D(8, kernel_size=(3,3), activation='relu'))\n",
        "model.add(MaxPooling2D(pool_size=(2,2)))\n",
        "\n",
        "model.add(Flatten())\n",
        "model.add(Activation(\"relu\"))\n",
        "model.add(Dense(1024))\n",
        "model.add(Dropout(0.4))\n",
        "model.add(Dense(2))\n",
        "model.add(Activation(\"softmax\"))\n",
        "model.compile(optimizer='adam', loss='binary_crossentropy', metrics=['accuracy'])\n"
      ],
      "metadata": {
        "id": "y3MunGKSOG5w"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "model.summary()"
      ],
      "metadata": {
        "id": "dLgbOTj-HrRk",
        "outputId": "3d8aee5c-a74f-4b52-9183-276bb7555918",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " conv2d (Conv2D)             (None, 62, 62, 64)        1792      \n",
            "                                                                 \n",
            " max_pooling2d (MaxPooling2D  (None, 31, 31, 64)       0         \n",
            " )                                                               \n",
            "                                                                 \n",
            " conv2d_1 (Conv2D)           (None, 29, 29, 32)        18464     \n",
            "                                                                 \n",
            " max_pooling2d_1 (MaxPooling  (None, 14, 14, 32)       0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " conv2d_2 (Conv2D)           (None, 12, 12, 16)        4624      \n",
            "                                                                 \n",
            " max_pooling2d_2 (MaxPooling  (None, 6, 6, 16)         0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " conv2d_3 (Conv2D)           (None, 4, 4, 8)           1160      \n",
            "                                                                 \n",
            " max_pooling2d_3 (MaxPooling  (None, 2, 2, 8)          0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " flatten (Flatten)           (None, 32)                0         \n",
            "                                                                 \n",
            " activation (Activation)     (None, 32)                0         \n",
            "                                                                 \n",
            " dense (Dense)               (None, 1024)              33792     \n",
            "                                                                 \n",
            " dropout (Dropout)           (None, 1024)              0         \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 2)                 2050      \n",
            "                                                                 \n",
            " activation_1 (Activation)   (None, 2)                 0         \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 61,882\n",
            "Trainable params: 61,882\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "history = model.fit_generator(train_set,\n",
        "                    steps_per_epoch=3800//10,\n",
        "                    epochs=100,\n",
        "                    validation_data=test_set,\n",
        "                    validation_steps=2200//10)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "T41oe4enOXrk",
        "outputId": "34e0b7f2-dc43-451d-fafd-a15b252b2139"
      },
      "execution_count": null,
      "outputs": [
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/ipykernel_launcher.py:5: UserWarning: `Model.fit_generator` is deprecated and will be removed in a future version. Please use `Model.fit`, which supports generators.\n",
            "  \"\"\"\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/100\n",
            "380/380 [==============================] - ETA: 0s - loss: 0.4660 - accuracy: 0.7853WARNING:tensorflow:Your input ran out of data; interrupting training. Make sure that your dataset or generator can generate at least `steps_per_epoch * epochs` batches (in this case, 220 batches). You may need to use the repeat() function when building your dataset.\n",
            "380/380 [==============================] - 2178s 6s/step - loss: 0.4660 - accuracy: 0.7853 - val_loss: 0.4600 - val_accuracy: 0.7949\n",
            "Epoch 2/100\n",
            "380/380 [==============================] - 541s 1s/step - loss: 0.3313 - accuracy: 0.8611\n",
            "Epoch 3/100\n",
            "380/380 [==============================] - 199s 523ms/step - loss: 0.3228 - accuracy: 0.8676\n",
            "Epoch 4/100\n",
            "380/380 [==============================] - 120s 315ms/step - loss: 0.3009 - accuracy: 0.8713\n",
            "Epoch 5/100\n",
            "380/380 [==============================] - 98s 258ms/step - loss: 0.2978 - accuracy: 0.8721\n",
            "Epoch 6/100\n",
            "380/380 [==============================] - 90s 237ms/step - loss: 0.2693 - accuracy: 0.8832\n",
            "Epoch 7/100\n",
            "380/380 [==============================] - 93s 245ms/step - loss: 0.2878 - accuracy: 0.8768\n",
            "Epoch 8/100\n",
            "380/380 [==============================] - 88s 232ms/step - loss: 0.2657 - accuracy: 0.8855\n",
            "Epoch 9/100\n",
            "380/380 [==============================] - 89s 233ms/step - loss: 0.2631 - accuracy: 0.8887\n",
            "Epoch 10/100\n",
            "380/380 [==============================] - 88s 231ms/step - loss: 0.2568 - accuracy: 0.8979\n",
            "Epoch 11/100\n",
            "380/380 [==============================] - 89s 235ms/step - loss: 0.2458 - accuracy: 0.8989\n",
            "Epoch 12/100\n",
            "380/380 [==============================] - 92s 241ms/step - loss: 0.2367 - accuracy: 0.9087\n",
            "Epoch 13/100\n",
            "380/380 [==============================] - 91s 240ms/step - loss: 0.2325 - accuracy: 0.9034\n",
            "Epoch 14/100\n",
            "380/380 [==============================] - 88s 231ms/step - loss: 0.2206 - accuracy: 0.9109\n",
            "Epoch 15/100\n",
            "380/380 [==============================] - 89s 233ms/step - loss: 0.2251 - accuracy: 0.9077\n",
            "Epoch 16/100\n",
            "380/380 [==============================] - 90s 237ms/step - loss: 0.2229 - accuracy: 0.9082\n",
            "Epoch 17/100\n",
            "380/380 [==============================] - 89s 235ms/step - loss: 0.2271 - accuracy: 0.9069\n",
            "Epoch 18/100\n",
            "380/380 [==============================] - 92s 241ms/step - loss: 0.2036 - accuracy: 0.9175\n",
            "Epoch 19/100\n",
            "380/380 [==============================] - 89s 234ms/step - loss: 0.1988 - accuracy: 0.9206\n",
            "Epoch 20/100\n",
            "380/380 [==============================] - 89s 235ms/step - loss: 0.2208 - accuracy: 0.9095\n",
            "Epoch 21/100\n",
            "380/380 [==============================] - 89s 235ms/step - loss: 0.2047 - accuracy: 0.9188\n",
            "Epoch 22/100\n",
            "380/380 [==============================] - 90s 237ms/step - loss: 0.1997 - accuracy: 0.9229\n",
            "Epoch 23/100\n",
            "380/380 [==============================] - 88s 233ms/step - loss: 0.2095 - accuracy: 0.9233\n",
            "Epoch 24/100\n",
            "380/380 [==============================] - 90s 238ms/step - loss: 0.1877 - accuracy: 0.9277\n",
            "Epoch 25/100\n",
            "380/380 [==============================] - 92s 242ms/step - loss: 0.1953 - accuracy: 0.9214\n",
            "Epoch 26/100\n",
            "380/380 [==============================] - 90s 236ms/step - loss: 0.1988 - accuracy: 0.9219\n",
            "Epoch 27/100\n",
            "380/380 [==============================] - 91s 239ms/step - loss: 0.1865 - accuracy: 0.9280\n",
            "Epoch 28/100\n",
            "380/380 [==============================] - 90s 238ms/step - loss: 0.1858 - accuracy: 0.9246\n",
            "Epoch 29/100\n",
            "380/380 [==============================] - 92s 242ms/step - loss: 0.1810 - accuracy: 0.9321\n",
            "Epoch 30/100\n",
            "380/380 [==============================] - 89s 235ms/step - loss: 0.1924 - accuracy: 0.9241\n",
            "Epoch 31/100\n",
            "380/380 [==============================] - 91s 240ms/step - loss: 0.1889 - accuracy: 0.9272\n",
            "Epoch 32/100\n",
            "380/380 [==============================] - 91s 239ms/step - loss: 0.1716 - accuracy: 0.9335\n",
            "Epoch 33/100\n",
            "380/380 [==============================] - 91s 239ms/step - loss: 0.1728 - accuracy: 0.9303\n",
            "Epoch 34/100\n",
            "380/380 [==============================] - 92s 242ms/step - loss: 0.1790 - accuracy: 0.9335\n",
            "Epoch 35/100\n",
            "380/380 [==============================] - 92s 242ms/step - loss: 0.1797 - accuracy: 0.9320\n",
            "Epoch 36/100\n",
            "380/380 [==============================] - 94s 248ms/step - loss: 0.1643 - accuracy: 0.9363\n",
            "Epoch 37/100\n",
            "380/380 [==============================] - 92s 243ms/step - loss: 0.1775 - accuracy: 0.9317\n",
            "Epoch 38/100\n",
            "380/380 [==============================] - 93s 246ms/step - loss: 0.1698 - accuracy: 0.9335\n",
            "Epoch 39/100\n",
            "380/380 [==============================] - 91s 238ms/step - loss: 0.1755 - accuracy: 0.9320\n",
            "Epoch 40/100\n",
            "380/380 [==============================] - 90s 238ms/step - loss: 0.1616 - accuracy: 0.9379\n",
            "Epoch 41/100\n",
            "380/380 [==============================] - 92s 242ms/step - loss: 0.1831 - accuracy: 0.9325\n",
            "Epoch 42/100\n",
            "380/380 [==============================] - 90s 238ms/step - loss: 0.1759 - accuracy: 0.9314\n",
            "Epoch 43/100\n",
            "380/380 [==============================] - 94s 246ms/step - loss: 0.1726 - accuracy: 0.9296\n",
            "Epoch 44/100\n",
            "380/380 [==============================] - 89s 235ms/step - loss: 0.1669 - accuracy: 0.9409\n",
            "Epoch 45/100\n",
            "380/380 [==============================] - 89s 235ms/step - loss: 0.1658 - accuracy: 0.9370\n",
            "Epoch 46/100\n",
            "380/380 [==============================] - 90s 238ms/step - loss: 0.1758 - accuracy: 0.9320\n",
            "Epoch 47/100\n",
            "380/380 [==============================] - 91s 238ms/step - loss: 0.1648 - accuracy: 0.9380\n",
            "Epoch 48/100\n",
            "380/380 [==============================] - 91s 240ms/step - loss: 0.1808 - accuracy: 0.9275\n",
            "Epoch 49/100\n",
            "380/380 [==============================] - 89s 235ms/step - loss: 0.1548 - accuracy: 0.9417\n",
            "Epoch 50/100\n",
            "380/380 [==============================] - 91s 239ms/step - loss: 0.1768 - accuracy: 0.9380\n",
            "Epoch 51/100\n",
            "380/380 [==============================] - 91s 239ms/step - loss: 0.1625 - accuracy: 0.9359\n",
            "Epoch 52/100\n",
            "380/380 [==============================] - 91s 240ms/step - loss: 0.1681 - accuracy: 0.9391\n",
            "Epoch 53/100\n",
            "380/380 [==============================] - 90s 237ms/step - loss: 0.1541 - accuracy: 0.9415\n",
            "Epoch 54/100\n",
            "380/380 [==============================] - 88s 231ms/step - loss: 0.1688 - accuracy: 0.9346\n",
            "Epoch 55/100\n",
            "380/380 [==============================] - 93s 244ms/step - loss: 0.1663 - accuracy: 0.9400\n",
            "Epoch 56/100\n",
            "380/380 [==============================] - 93s 244ms/step - loss: 0.1711 - accuracy: 0.9393\n",
            "Epoch 57/100\n",
            "380/380 [==============================] - 92s 241ms/step - loss: 0.1536 - accuracy: 0.9430\n",
            "Epoch 58/100\n",
            "380/380 [==============================] - 92s 241ms/step - loss: 0.1632 - accuracy: 0.9392\n",
            "Epoch 59/100\n",
            "380/380 [==============================] - 93s 245ms/step - loss: 0.1571 - accuracy: 0.9430\n",
            "Epoch 60/100\n",
            "380/380 [==============================] - 96s 253ms/step - loss: 0.1597 - accuracy: 0.9396\n",
            "Epoch 61/100\n",
            "380/380 [==============================] - 92s 243ms/step - loss: 0.1574 - accuracy: 0.9391\n",
            "Epoch 62/100\n",
            "380/380 [==============================] - 92s 243ms/step - loss: 0.1463 - accuracy: 0.9459\n",
            "Epoch 63/100\n",
            "380/380 [==============================] - 91s 241ms/step - loss: 0.1511 - accuracy: 0.9432\n",
            "Epoch 64/100\n",
            "380/380 [==============================] - 92s 242ms/step - loss: 0.1591 - accuracy: 0.9367\n",
            "Epoch 65/100\n",
            "380/380 [==============================] - 90s 238ms/step - loss: 0.1593 - accuracy: 0.9412\n",
            "Epoch 66/100\n",
            "380/380 [==============================] - 91s 239ms/step - loss: 0.1612 - accuracy: 0.9405\n",
            "Epoch 67/100\n",
            "380/380 [==============================] - 89s 233ms/step - loss: 0.1456 - accuracy: 0.9454\n",
            "Epoch 68/100\n",
            "380/380 [==============================] - 88s 231ms/step - loss: 0.1651 - accuracy: 0.9408\n",
            "Epoch 69/100\n",
            "380/380 [==============================] - 90s 238ms/step - loss: 0.1554 - accuracy: 0.9412\n",
            "Epoch 70/100\n",
            "380/380 [==============================] - 88s 231ms/step - loss: 0.1580 - accuracy: 0.9387\n",
            "Epoch 71/100\n",
            "380/380 [==============================] - 103s 271ms/step - loss: 0.1560 - accuracy: 0.9430\n",
            "Epoch 72/100\n",
            "380/380 [==============================] - 90s 238ms/step - loss: 0.1557 - accuracy: 0.9396\n",
            "Epoch 73/100\n",
            "380/380 [==============================] - 94s 246ms/step - loss: 0.1467 - accuracy: 0.9474\n",
            "Epoch 74/100\n",
            "380/380 [==============================] - 91s 240ms/step - loss: 0.1563 - accuracy: 0.9415\n",
            "Epoch 75/100\n",
            "380/380 [==============================] - 92s 243ms/step - loss: 0.1530 - accuracy: 0.9405\n",
            "Epoch 76/100\n",
            "380/380 [==============================] - 89s 235ms/step - loss: 0.1528 - accuracy: 0.9480\n",
            "Epoch 77/100\n",
            "380/380 [==============================] - 91s 238ms/step - loss: 0.1642 - accuracy: 0.9387\n",
            "Epoch 78/100\n",
            "380/380 [==============================] - 91s 239ms/step - loss: 0.1542 - accuracy: 0.9449\n",
            "Epoch 79/100\n",
            "380/380 [==============================] - 89s 233ms/step - loss: 0.1496 - accuracy: 0.9428\n",
            "Epoch 80/100\n",
            "380/380 [==============================] - 92s 241ms/step - loss: 0.1559 - accuracy: 0.9433\n",
            "Epoch 81/100\n",
            "380/380 [==============================] - 90s 236ms/step - loss: 0.1399 - accuracy: 0.9497\n",
            "Epoch 82/100\n",
            "380/380 [==============================] - 92s 241ms/step - loss: 0.1450 - accuracy: 0.9483\n",
            "Epoch 83/100\n",
            "380/380 [==============================] - 89s 235ms/step - loss: 0.1492 - accuracy: 0.9404\n",
            "Epoch 84/100\n",
            "380/380 [==============================] - 93s 243ms/step - loss: 0.1458 - accuracy: 0.9459\n",
            "Epoch 85/100\n",
            "380/380 [==============================] - 91s 238ms/step - loss: 0.1528 - accuracy: 0.9425\n",
            "Epoch 86/100\n",
            "380/380 [==============================] - 90s 236ms/step - loss: 0.1545 - accuracy: 0.9444\n",
            "Epoch 87/100\n",
            "380/380 [==============================] - 91s 240ms/step - loss: 0.1430 - accuracy: 0.9459\n",
            "Epoch 88/100\n",
            "380/380 [==============================] - 89s 233ms/step - loss: 0.1350 - accuracy: 0.9463\n",
            "Epoch 89/100\n",
            "380/380 [==============================] - 92s 242ms/step - loss: 0.1403 - accuracy: 0.9483\n",
            "Epoch 90/100\n",
            "380/380 [==============================] - 91s 240ms/step - loss: 0.1503 - accuracy: 0.9466\n",
            "Epoch 91/100\n",
            "380/380 [==============================] - 90s 236ms/step - loss: 0.1323 - accuracy: 0.9505\n",
            "Epoch 92/100\n",
            "380/380 [==============================] - 89s 234ms/step - loss: 0.1440 - accuracy: 0.9442\n",
            "Epoch 93/100\n",
            "380/380 [==============================] - 91s 240ms/step - loss: 0.1548 - accuracy: 0.9401\n",
            "Epoch 94/100\n",
            "380/380 [==============================] - 89s 234ms/step - loss: 0.1337 - accuracy: 0.9486\n",
            "Epoch 95/100\n",
            "380/380 [==============================] - 90s 237ms/step - loss: 0.1390 - accuracy: 0.9445\n",
            "Epoch 96/100\n",
            "380/380 [==============================] - 88s 232ms/step - loss: 0.1506 - accuracy: 0.9408\n",
            "Epoch 97/100\n",
            "380/380 [==============================] - 90s 236ms/step - loss: 0.1439 - accuracy: 0.9491\n",
            "Epoch 98/100\n",
            "380/380 [==============================] - 90s 236ms/step - loss: 0.1498 - accuracy: 0.9446\n",
            "Epoch 99/100\n",
            "380/380 [==============================] - 89s 233ms/step - loss: 0.1379 - accuracy: 0.9486\n",
            "Epoch 100/100\n",
            "380/380 [==============================] - 89s 233ms/step - loss: 0.1474 - accuracy: 0.9433\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model.save('Hue.h5')"
      ],
      "metadata": {
        "id": "bnyIMXZGOhRj"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plt.plot(history.history[\"loss\"], label =\"Train loss\")\n",
        "plt.show()\n",
        "plt.plot(history.history[\"accuracy\"], label =\"Accuracy\")\n",
        "plt.show()"
      ],
      "metadata": {
        "id": "Wzb-V5J9OdDw"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "predictions = model.predict_generator(val_set, steps=16, verbose=1)"
      ],
      "metadata": {
        "id": "2U3Bo0k_X_ls"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "predictions.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "5R5fJtCvbvgB",
        "outputId": "c4e39db7-2692-4031-e847-5c83688317ef"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(16, 2)"
            ]
          },
          "metadata": {},
          "execution_count": 17
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(test_set.class_indices)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "INdaBehzSaUs",
        "outputId": "caa428f3-cdf0-4dbf-9edf-0c02866508f6"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "{'NORMAL': 0, 'PNEUMONIA': 1}\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "img = load_img(\"/content/drive/MyDrive/pneumonia_xray/test/NORMAL/IM-0003-0001.jpeg\",target_size=(64,64))\n",
        "plt.imshow(img)\n",
        "img = img_to_array(img)\n",
        "img = img.reshape(1,64,64,3)  \n",
        "img = img.astype('float32')\n",
        "img /= 255\n",
        "prediction = model.predict(img).argmax()\n",
        "plt.title(\"Predicted Label for the image is {}\".format(dudoan[prediction]))\n",
        "y_pred = model.predict(img)\n",
        "a = y_pred.max()\n",
        "a = a*100\n",
        "print(\"Accuracy: \",a, '%')\n"
      ],
      "metadata": {
        "id": "8orRxbrWjV96"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "img = load_img('/content/drive/MyDrive/pneumonia_xray/test/PNEUMONIA/person143_bacteria_688.jpeg',target_size=(64,64))\n",
        "plt.imshow(img)\n",
        "img = img_to_array(img)\n",
        "img = img.reshape(1,64,64,3)  \n",
        "img = img.astype('float32')\n",
        "img /= 255\n",
        "prediction = model.predict(img).argmax()\n",
        "plt.title(\"Predicted Label for the image is {}\".format(dudoan[prediction]))\n",
        "y_pred = model.predict(img)\n",
        "a = y_pred.max()\n",
        "a = a*100\n",
        "print(a, '%')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 298
        },
        "id": "OzNwohKXDKTu",
        "outputId": "1faaa825-338d-4ba2-a120-3fdc433b98f8"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "99.99572038650513 %\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAARoAAAEICAYAAACArTsqAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO19ebheVXnv72WIjBKGACEJEOahDwREBhGvQqU4VW6vpY6Flnuxw/XqLbVg7bVatRf79Kl6O2ixDlhUoGKFUlRGERSBQJjJJAkkgRAgzCAIvPePvb+d3/7lrPd8Z9jnOye8v+fJk/V9a+211l57n/W9v3da5u5IJBKJLrHRoCeQSCQ2fORGk0gkOkduNIlEonPkRpNIJDpHbjSJRKJz5EaTSCQ6x0A3GjP7hpl9pi4fY2aLJmhcN7O9xrnP5l7G+1oz+4yZPWJmq0c/w3D8N5rZynHq631mdtl49NUlzOwuM3vjoOfxSsGwG42ZLTez58zsaTN7qP6j2Gq8J+Lu17r7vn3M5xQzu268x6f+f2xm/72r/kcKM9sVwOkADnD3ncepz3HfaHtw92+5+/Fd9D2ecPcD3f3HI7mm3pBfrv8WnjKzRWb2e3Xd7vW6XirXnGtmnxziev53VF2/3Mx+Xa5v3ncaY4G02cHMXjCz5UNce4eZPWtmq83sS2Y2neo/Wfd3En23Sf3d7vXn9X4EzWyret4/6Hft+pVo3uHuWwE4FMBhAP5CG5jZJv0OmhgRdgXwqLuvGemF+Uw6wQP138KrAZwB4CtmdgDVH2Fmrxvuevl3/QjnsIWZ/Rp9fi+AZdzAzE4H8DkAHwWwDYAjAewG4HIzm0ZN1wL4lJltPILx/xuA5wG82cz6+vEbEXVy91UAfgDg14Dml/GPzWwJgCX1d283s1vN7HEz+5mZHdS73swOMbNb6l+D8wFsRnUt8d3M5pjZ98zsYTN71Mz+wcz2B/BlAEfVO+rjddtXmdnfmtn9tdT1ZTPbnPr6qJk9aGYPmNnvj+SeGWb2b/UvwxNm9hMzO1Ca7GBml9f3d42Z7UbX7lfXra1/CU/CMKh/3S4HsEt9v9+ov//NWvR/vJbA9qdrlpvZGWZ2O4BndLMxs5/UxdvqPn+H6k43szX1Wv0efR+ur/Tfkjjrd+SPzGxJvS6fNrM963fjSTO7oPfim9m2ZnZJ/cwfq8uzqa+59bo/ZWZXmNk/mtm5VH9k3e/jZnabBdTISHows8PNbH49n4fM7O+i5wIAXuH7AB4DwBvN3wD47HDXjxH/CuBk+vy7AL7Z+2BmrwbwKQAfcvcfuvuv3H05gJMA7A7g/XTtDwG8IN8Nh5NR/R3e3u91I9pozGwOgLcCYNHtRABHADjAzA4B8DUAHwSwPYB/BnBx/aJOA/B9VIu0HYB/Q7UzDjXOxgAuAXAfqoWZBeA8d78HwB8AuL7+JeiJgWcB2AfAPAB71e0/Ufd1AoA/BfBmAHsDaImmI8QP6j52BHALgG9J/fsAfBrADgBu7dWb2ZaoNoxv19e+G8A/yS/henD3KwC8Bet+BU8xs30AfAfARwDMAHApgP+QX6n3AHgbgOnu/qL0+Ya6eHDd5/n1551R/fLNAnAqgH80s23ruuL69onfAPAaVL+qfwbgbFQv6BxUP1rvqdttBODrqH55dwXwHIB/oH6+DeBGVO/WJwF8oFdhZrMA/CeAz6B6v/4UwIVmNqOP+X0RwBfd/dUA9gRwwXAXmNlGZvZfAUwHcAdV/ROAfUwo0DjjXADvNrON63doKwA3UP3rUP2If48vcvenUb0vb+avAfwfAH9pZpsON3D94/lGVO/2t1BtcsOi343m+7X0cB2AawD8NdX9X3df6+7PATgNwD+7+w3u/pK7n4NKxDqy/rcpgC/UO+x3AdxUGO9wALsA+Ki7P+Puv3T3IfUyZmb1uP+7nsdT9fzeXTc5CcDX3f1Od38G1Qs6Krj719z9KXd/vu7nYDPbhpr8p7v/pK7/OCrJaw6AtwNY7u5fd/cX3X0BgAsB/PYopvE79TiXu/uvAPwtgM1RvVw9/D93X1E/k37xKwB/VT+bSwE8DWDfPta3H/yNuz/p7ncBuBPAZe5+r7s/gWrzPgQA3P1Rd7/Q3Z+tx/ksgP8CNLqq1wL4hLu/UL8PF9MY7wdwqbtf6u4vu/vlAOaj+mHs5973MrMd3P1pd/950HaX+m/hEQB/CeAD7s5GjOfqeZeU+7vUEhf/27KPOTJWAliE6kfzd1H9eDN2APCI/sjUeLCub+DuFwN4GEA/uskPALjd3e8GcB6AA2sBI0S/HP7E+td1KKyg8m4ATjazD9F301BtGg5glbejOO8r9DkHwH2FhVLMALAFgJurvwkAgAHocc5dANzcx5ghainrs6g2hxkAXq6rdgDwRF1u1sLdnzaztfX4u6Hi7o9Tl5tg/RekH+wCugd3f9nMVqCSMnpYsd5Vw+NRWe9nUf1SDre+/eAhKj83xOedAcDMtgDweQAnAOhJU1vXa78LgLXu/ixduwLVuwJUa/zbZvYOqt8UwNV9zO9UAH8FYKGZLQPwKXe/pND2AXefXajr4V8AfFTm0s/1L9ZzZmyKaiNUfBPAKah+YI5BJXH28AgqGr/JEH9DM+t6xV+gkiaHeyd/F8BXgEqVYmbXoKJSC6KLxsO8zRvHCgCfdffp9G8Ld/8Oqp10ltHbiko8HgorAOyq+oUhxgOqRXsOwIE05ja1wg71uHOofWnM4fBeAO9E9SuyDSpKB1R/dD0041hlmdsOwAOo7ucaWZet3P0PRzGPB1D9UfXGsXrcVdRmPEPyh1vf8cTpAPYFcERNY3o0z1A9x+3qzagHfq4rAPyrrPGW7n7WcIO6+xJ3fw8qWvs5AN8dhZTB/b2ASkfyabTfj+FwP9a9Vz3MxdA/jheiosf3uvv9Unc9KibxW/xl/U6+BcCVQ8z5cgBLAfxRaXJWKbn3BvAxq3SVq1GpTd5b+FttMN5+NF8B8AdmdoRV2NLM3mZmW6O6+RcB/C8z29TMfgsVRRoKN6J6sc6q+9jMzI6u6x4CMLunk3D3l+txP29mOwIVXzez36jbXwDgFDM7oH5J/7KP+9ikHrP3b1MAW6N6eI+i+oX/6yGue6uZvb6e26cB/NzdV6DSN+1jZh+o731TM3utkRJ3BLgAwNvM7Lh6XqfX8/rZCPp4CMAe/TTsY33HE1uj2tQeN7PtQM/K3e9DRYU+aWbTrDIJs8RwLoB3mNlv1LqLzawyMAwnfcDM3m9mM+p77UmdL0fX9IF/RaUnOWEE15wP4CNWGQ7MzA4D8PuoKEoLtRrgWAxBd2pK+ikAf29mJ9Tv2+6o3p2VKEstH0elQyvhZFS6xgNQ6evmodKxbY5qAytiXDcad58P4H+gUuA9hmqHPKWuewHVDnsKKpPa70CUVdTPS6heor1Q7fIr6/YAcBWAuwCsNrOeCHhGPdbPzexJAFeg+mWEu/8AwBfq65bW/w+HL6F64Xv/vo5KVL0PleRwN4ChePy3Uf1xrEWl/Hx/PYenAByPSq/xAIDVqH45X9XHXFqo9QHvB/D3qKSNd6ByP3hhBN18EsA5tX5gWOsXgvUdZ3wB1Uv7CKr1/aHUvw/AUag2+8+g+sN8HgDqDf2dAP4clb5hBSrTbj/v+AkA7jKzp1Epht89Qv3Weqjf4U+gkmoZPQsi/+sZRb6C6l37D1R0/JsAPu7uug69Mea7+y8KdX+Dai3+FsCTqJTFKwAcV+sQh7rmp6h+5NeDmW2GSt/59+6+mv4tw/pWsPWv90x8lZiisMpFYqG79yOlJgaIjHVKTBnUdHPP2rR8AioJ5vuDnldieKTnaGIqYWdUdHt7VHT6D2tXgcQkR1KnRCLROTqhTrWme5GZLTWzM7sYI5FITB2Mu0RTO1ctRuXmvBKV9+97ak/C9TBt2jTffPPNe9f2PU5p3tOmTWt9fvnldVZK7Z8/v/TSS015o43a+y/XRX1svfXWxT643VNPPdWq22STTYZsF81/443b/nL8WfvgteKxFNyO71mhY5fqeO76meeo7fqdx4svrvNF22yzzVp1fN2vftX2d3vhhXUGOh7rVa9qGwH58zPPPNOqe/bZZ1EC97nppuv87/S5RPdW6k/f+x133HHIsYD2s1i4cOEj7t5POEYn6EJHcziApe5+LwCY2XmolHZDbjSbb745jjrqqGoy8kfAC6ULzC8ZP8Bddtml1Y5fCH0Q/Pnpp59uyltssUWr3eOPr3Po1TnyxvaGN7yhKfc2zx74D+Gqq9oW9u233754Hc+f73O77dpW0222WRcJoffJf1gzZqx713RNf/nLXzZl3Qx54+QNVft49atf3ZSff75tRS09C/2j5fk+8cQTrTpeg4cffrgp779/2yXpySefbMqrV7fT+Cxfvrwp8yY0d+7cVrs999yzKd90UztaZsGCdaoh/VHh+e+0005NWX8EH3vssaasmzffJ/enm+aHPrTOCX/nnduB1PxOHHHEEaPyiB8vdEGdZqHtAr8Sbfd4mNlpVkXLzudFTCQSGyYGYnVy97NRRfBixowZvuuuVVTAgw8+WLwmogoXXnhhU2ZJBwDOOOOMpqyiKksn/GujEgGL0PrrxWBpSn95WFJ53eva6Ur4F1fFd74fphjPPdf2J+M56/y32mpdtABLT7pWPOdIrOf+lLLwZ5YEdc6lX2wglkJ5/Zk2sHQAtCUrbge0nztLZ7r25557brGOJUqlVfxuslSn78SWW25ZrCvRL11vlvhYWgXWp6SDRBcSzSq0Y1Bmox2Hk0gkXmHoYqO5CcDeViUpmobK7f7iYa5JJBIbMMadOrn7i2b2PwH8CFUqga/VeUgSicQrFJ3oaOrESZcO2xAVn9122yr1iFp7mHer9YF5MFtIdtihldMH733ve5vy7NntQF7m6jNnzmzKqodZu3Ztcf7M3dnSodYv1sOoFYT7X7hwYauO+fn06dOHLANtPQfrYYCybkfNrXzf2n/JtB5ZA9XqxGvF83300Udb7R55ZF26FO2Dr+N7OeaYY1rt+Fnss88+rToeb82aNUNeA7R1O6rv4PtUPRJj1qx1dhDtn9dOnwW/+5EJni1Nqr9hPdWgkbFOiUSic+RGk0gkOsekCKrsiaUqrrNIqiI0gx2TPvjBDw7ZNwDsu287hQqL0GzmvP9+TVi2DiqespmTxVp2JgOARYvWpZVV0zFTBRV32ZTMa6B9MO1RszKL3rzG6kDG4ruavpkucVkdGCNqFtGl0lhq9uU+ef7XXnttqx1TaKVOTC253UUXXdRqx+ZnBc+L6RHQpsb8XFQ1wOuh9Iv757L+jbCTpb6bXDdopESTSCQ6R240iUSic+RGk0gkOsfAdTRm1nBt5ZSs89BgQ9ZlcGCZhiowj1c3debMPLaaMpn79sIlemC374ceWneKyA033NBqx/oV1X/wfbJOBmjzf57vAw880GrH96nhCazP4To1lbLOIDLncp3qUFhn02+Igz53XgNdjyVLljRl1o/tsUc71zrrpRYvXtyqO/jgg5sy68f22qt9HPmKFetC9vS94rXTOtaj8JqyCwXQDu5U0zfrdnhN1X2D302t02czSKREk0gkOkduNIlEonMMnDq5e0Mr2MQMtEV0Nf8xjeB2agZnytWvCVEpBecUUXMueyyzyVZzrLCXqYraPGfNKcIiNFMMjUhmT2TOrQO0o4t5PXQeJXFdwTQtaqd1vK5RFDnTHjUx8zvC89Dnzs9T6df111/flA877LCmrGvKfSod7XmzDwV+N3lNlcKxqV7nWIqeV3rH7z5Td2D993iQSIkmkUh0jtxoEolE5xg4dWKrk4q/LDZrHdMg1uyrCMrtImsP0wsV19nzWL0v2eOXLWGa3IrFWvWKZeuGiuhMb5geKe1hsV89ctk6w+uoVIHFfJ0H1/HYKp4zXYq8erk/tZZwkKmuFffP9EXnwd7dakHj9WDrHVNkoE3vNLCW3zO1mrEFieehz4xprL6bpfc2CqpUqhrR2olGSjSJRKJz5EaTSCQ6R240iUSicwycxLl7o2NRTsncV/UOHCFb0mMAbR6s/bO+hU2SqudhjqzHkJSOOdFE66tWrUubrGbI0hEiQFsPwfesSbzuu2/daRrK91knwfcZJWxSPVUpSliThPEaa/+s9+H71DVlPVKUhIzNxZF3Lj8joP18eb76XPg6db3gthoFz2Oz57GuRymqHmh7sbMuTd0f+FnrGmTiq0Qi8YpCbjSJRKJzDJw6Pf/8801wmYq4bLZW8bSUiEnbMVQMZ/Mri9Cac5epx5VXXtmq40C5e++9tykrHWAqouZcplmaHInpBovaKiZzO/UM5uuiY2QZetYSryvXKR1lj1adB9MBnoeah0v3DKxPb3rQe2H6pdcwveb56+mf/H4odWdKq17gpdNFo6RV0flVTP91LL5vDTzu98jdiUBKNIlEonPkRpNIJDpHbjSJRKJzDFxHw1C9BvNW5eqso+B26rIeue0z2E1d3bzZjKomQ05wxXUjcc2fO3duU9bkSKxrYA6u51yxLkDXivvg69QVgHUGkfs635uOxfOIQhBYRxO5Lui51tyW9V6q4+D5a6jF0qVLm/KBBx7YlNW1gPUfqrfjsdWVgXVMrGvRPlh3FCVsY6ieh3V6ek3qaBKJxCsKo95ozOxrZrbGzO6k77Yzs8vNbEn9fzlpRyKReMVgLNTpGwD+AcA36bszAVzp7meZ2Zn15zOiTty9EbGVbrA43K8oryJolN+Wqci8efOasnrWcgTuypUri/2z6KoibiTGMrVRD9QSFdHES2xyjjxCo4jhKCq7BKUsvB7qTsA0iOevlJnXjnMya/9MG/SemZZEuajZJUHdDnhN9XmymTk6Y4vnqO1KxwxrH/yc9B3mOeo6bhCewe7+EwB6KPU7AZxTl88BcOJo+08kEhsOxltHs5O79zRjqwHsNFQjMzvNzOab2Xx1DEskEhseOrM6ububmRfqzgZwNgBMnz69aaPiI1s0VOzkzyxaqncki8lq3WCrAovXHAAJxMe8smjP81cKNH/+/KasR9Yy/dDjePk4V6ZYmghMPYoZHEQYidPch9KqUtIqpbQs2qsVhNebrYNRPl61avFacVnvny1NSu947TjxlVqdjjjiiGIdP3ftv5QLWL16ee2U3vG7xJRTE3DxderZHB0jPdEYb4nmITObCQD1/2vGuf9EIjEFMd4bzcUATq7LJwO4KGibSCReIRiLefs7AK4HsK+ZrTSzUwGcBeDNZrYEwK/XnxOJxCsco9bRuPt7ClXHjbCfRh+g5lDmusqDWRcTRWwz/1cOy9HiXMfHrgLAnDlzivNg3QCbvtesabNGvk6TSvG9qBmVeTxz7ujMJDWBaiKvHlTfFOleWCfGY2k7Nj+rXo31VsuWLWvKqrtQHRmDzzVij1w1x/Pz1DrWf3A7jTZfsGBBUz700ENbdWzE0P75vrmsujleb9Wn8Lry+uhasbdxdLTwoJGewYlEonPkRpNIJDrHwIMqN9poo0b0VFGbKZGezaOfe1BzaL+Jnm699damrKZSNi9qYByf68TtVAxnmqLzYHOu1vF1PC9tF3lOsylZ588oJQID2vQgog28BvqMmMZGHtvRkbscFMrUQL1imWJwkCYAzJo1a8ixdG34XpTO8RlQeh7UXXfd1ZRLbgFAeU2B9U3+PajbAX/WxHFKswaJlGgSiUTnyI0mkUh0jtxoEolE5xi4juall15qzH4aPcs8Xk2xjCg5Oet99thjj1bdL37xiyH7UJ0Bn6XM5ycBbZ0Bc2TVNzHnVjMkIzpTivVPqqNh/U0UBc/9qQ4lOk+Jx4v0KxH42XBoxU9/+tPifKPnzuuhSbxYP6Fz5LADfhZ77713cR4ausGhIrvvvnurjhOZsb5Gn0uUfJ/fH56/vlf8/unfT569nUgkXlHIjSaRSHSOgctW7Bkcmbcjz2AW69W8zZSCjyfVPlkEVUrB1ym1YRGdxWttx32qZ7Aeb1u6LvKU5rHVPM9zZEqklCKK7Oa23EeU9EmfBYOf31ve8pZW3e23317sn10BeA30mfG7o1SVP7NZnL2VAeCAAw4Ysj+g/Xw5eRbQNnczJR+JFzVfxxRLqWREhaMjjycaKdEkEonOkRtNIpHoHAOnTi+99FIjvmoeXBbXlQ6waBlZJpiWaP7ZO+64oylzjlmlFEzN1Brz2te+timzB6p6abLlSo8QYWqm1iSeC5d33nnnVjuel64Hi/k8DxWto7zGpaDNyGM7ohvcBx+VC7TvTQMROaEVUwoNSuR7Uc9dvi6aL3tU6zE43IeuIwfG8jutybOinMEly6R6++62225NOalTIpF4RSM3mkQi0Tlyo0kkEp1j4DqaF198sUkSpXoN5rfKP1knEekn2JSsPJ51JWwm1MhZ9ijW/tk8yroG1TuwTkJN3yXTMVCOLtZoX76X6IwgNrFGkeJqmub15zo12fJ6a6RxKSFUlCBL9WWsN+H5a1J3Xjc9EpffA9ZxqB6QXR5WrFjRquMEXPqsWY/C7VRHw89FvYZ57fhZaxQ56x1Vx6n6rUEiJZpEItE5cqNJJBKdY+DU6eWXX27EfhUtOVdvdOYTi5lKS1icVJMhn5HDAWkq8jNt0IRWLPaz6K0iP89DTZQsoqvXMN83i/x6fhXPg03YCl4fNefyfapplKkZj60Ujr2L1RO7RAeinL7qrczrz/esNJDXODrGmGmV0lZ+/5SGcGCmuhpwLmOmgVGOZn0WpYMV1TWC3zNN/qWBpoNESjSJRKJz5EaTSCQ6R240iUSicwxcR8MhCKqj4WhaPZOJ+T7rNaKzoRYuXNiqY7M1hyCo/oNNimr2ZfNilDiKP2sSbObjOjaPx/2rDoX70P5ZN8X9Kd9nU3WUSDtKSMb6Ju2j3yRerKPR9eb7ZLOy6vCi89j5Pnm+el46j6XhK5wAjd8doH1mN7tssO4GaOubdK34fqJo+ZIOT/sfNFKiSSQSnWMsR+LOMbOrzexuM7vLzD5cf7+dmV1uZkvq/7cdrq9EIrFhYyzU6UUAp7v7LWa2NYCbzexyAKcAuNLdzzKzMwGcCeCMqKOeaKiemSz+6hGznKeVxff99tuv1Y5FY6UUXMdjs+kSaFO6KO9rZN7m6yLv32gN+D7VRM51OkemTkwblPZwXURFSsf0ap9RIrPoTCNeOzX78jNkL+GRJIQqURF9xzhiW10e2CyulJwTYR100EFDXgO0n6Gud4k6qQmbMwaoK4DOeZAYtUTj7g+6+y11+SkA9wCYBeCdAM6pm50D4MSxTjKRSExtjIsy2Mx2B3AIgBsA7OTuPa3XagA7DdH+NACnjcfYiURi8mPMymAz2wrAhQA+4u4tl1SvTAbrJY5197Pd/TB3P2ys4ycSicmPMUk0ZrYpqk3mW+7+vfrrh8xsprs/aGYzAawp91Bx6x7nVy7NHJyzlgHtzHmRyTYyIbLp8frrr2/K6jrP0MhudlPneaiuhe9N+Tjfp/JqHo9N2qpvYtOs6jXYNSCaB89fdR6l7HvaLuq/NJaalaPIbn6G7GKveh6eh2YF4DXm90Pnwbotdb3g9dcQBJ4zm+AjE7beZ+m8LG3H8y+FLUwGjMXqZAC+CuAed/87qroYwMl1+WQAF41+eolEYkPAWCSaowF8AMAdZnZr/d2fAzgLwAVmdiqA+wCcNLYpJhKJqY5RbzTufh2AUlbw40bTJ5vqgPYRpf2aL1WMve2225qyJpjmI3GjxNxsSo8io1nUVvrC5lz1EGXPUjWBlpIjqWcwjxeJ6FHSqsh8zmPzfeo8SuZ4HY/XW/uIPINLtEdpN1Pr6HjiEq3UsTjSH2hTG01+zn2yyVxdL3h9OFG+zoXH0nlcddVVTVnptGYyGCTSMziRSHSO3GgSiUTnGHhQJUMtLiyuq7cri8NvetObmrJaOphSLFmypFXHyYyiBEUs/qqYfPfddzdlplVqweDPbKkCgLlz5zZltTawtY3rlNrwfesaMB2LLDpR4it+NkzN9Jkx1VEKx/2zR7F6F/OzVisi98/P4rrrrmu14zqlo6VjajXpGHvhKs3ke9P58xyZyt9yyy2tdvzco7XisgZwshc7e7oD8XlnE42UaBKJROfIjSaRSHSO3GgSiUTnGLiOxswafhpFsEbm3OXLlzdlNektWrSoKWsiID7Th3m1epJydK5G6jJnZrOm9hGdI85mSeXgrBtg3YWuVckcCrS9UzkJlOpouE5NvTwe66I0aVXpjG6gf8/Vklcs0NYJsQf3IYcc0mrHLgSqm+M577PPPsWx+J733HPPVh3rjlQ3ws+e9Te6pvyuqhm/5FWtrhH8DFWPGa3jRCMlmkQi0Tlyo0kkEp1jUlCnHg2KjlBVT1s23TEFUs9dNgPPmjWrVcd9srcr54MF2pRLjz9lEZeph5oreWw9X4rnrMeashcrjxUFM6oIzWJ5RF+YpqmYz3SjlEgLKOcF1rZRACd/VsrMdZzHV+kLn9OlZzJpgGRpLDaR63vFVIqpO9B+ZjwPXXteRzWR87Pme1O6G5mw9T0bJFKiSSQSnSM3mkQi0Tlyo0kkEp1j4Doad294vfJN5syqv+HPXOaIbKCt81B3dnYx57KawdkcHemRuKx8nHUcGmGuugEGmyhZB6Rm5VKUtyI6uyk6r6mkL1PdWRS9zW15/ppMjOs01KL0Tmhibl43dvUH2jqa+++/f8j5AW09CWcS0D40xIF1Nty/9sH3qZkF2C2DzeWRW4OGOOy6666YLEiJJpFIdI7caBKJROcYOHUC1jeD9lAyhwLA4Ycf3pSZHqlJj7111TS9ePHipsxemlE0sYquKpb3oKI8R28rtWHqoPfJ9Ibnz8m4gDat0v5LuWmVOjHdi6KVuU69T3n+SglL10WR4hFFZKqjz4wpl94Lj8e0eOXKlShB3x02ras3N/fJZnF1m2AK3a/aIHIZULqe0duJROIVhdxoEolE55gU1Kkk4kXJnNiCdOeddzZl9sQE2hYk1uQDbY09B+gp9WB6pBYSnjsnTtLgzhIFAtpivnqxlo5RiXIXKxUpUT8N7mTRW61rXFc6rhWIPX5LR+6qFzLTKn3uXBdRA14DXQ+eI/LqR8wAABtWSURBVFPOyGta14PfF7VW8bsZHQnDtFAtV0zr1VrK4DXQd0cp4yCREk0ikegcudEkEonOkRtNIpHoHJOHxKFs5gbW1wXwMbIrVqxoypqgiE2KynWZ3/J16o3KXFf1GtwnJzFXPs56kuiYV9VJcD+sC9B2/XoXR+c/8fpHiZh43VS/wn1qH3wvfF10dpPqRvheuA91a+A6TTrO+jO+bo899mi1Y13asmXLWnVs3lbdCCefZw9i1eUsXbq0KbOuCFj/2fSg7w4/93vuuadVx97ug0ZKNIlEonOM5eztzczsRjO7zczuMrNP1d/PNbMbzGypmZ1vZtOG6yuRSGzYGAt1eh7Ase7+tJltCuA6M/sBgD8B8Hl3P8/MvgzgVABfGs0AkfmSjxflPKpqVmaxnE2NQNu8yO00+RTnnFXKwqIsn8Gkx5My1VOaw6K9Ui6eF/evYjjTO6UbvI48X6U9TO+0jscr0Rf9rM+vZPrW+fIaq+dxicLpmvI8lNqUvJzVHMyuEhpoy33q2KtWrWrKHCyp3uJMq/gIX6AcoKs0k9dA10qTgQ0So5ZovEJv9Tet/zmAYwF8t/7+HAAnjmmGiURiymNMOhoz29jMbgWwBsDlAH4B4HF37/1ErQQwa4jrTjOz+WY2P1IAJxKJDQNj2mjc/SV3nwdgNoDDAew3zCW9685298Pc/bDJFPiVSCS6wbiYt939cTO7GsBRAKab2Sa1VDMbwKro2pdffrkxEUcRw9E5Q8zxb7311lY7jurlJOZA+9xi1pPoOUBsAlUdEHNpjqhWfsz6oLVr17bqWKpT3QuPHYVkMD/XM6V4jpFuhOcxkqTjjJJZFmjrGnhOkUk/QhQKEZm3uY7HUvcH1olpiACvAYcjAG39DfehJmxOTHXvvfe26o4//vimzIm09Jkx9Jmxu8WgMRar0wwzm16XNwfwZgD3ALgawLvqZicDuGisk0wkElMbY5FoZgI4x8w2RrVhXeDul5jZ3QDOM7PPAFgA4KvjMM9EIjGFMeqNxt1vB3DIEN/fi0pf0xc22mijxttWkxexqKliIXvvsjn65z//eavdvHnzmrJ6j7I4fPPNNzdlNVeyqVrFehabeU4aAc6mdaU2LMproicGi+FKB5gSKRUpUaLIQzmiInydPpd+zec8JzXLRvSudPaUPjNeR32veDy+T11T9gzW+2SztZq++T75fVG3CabuGtHPmQZ4PaJ80Prcr7jiimLbiUZ6BicSic6RG00ikegckyKosidGq5jMYq2awVkcXrBgQVNmj2GgHQSplIUtVCwaazBaKVES0LZklTyNtU8Vw1nkVQrAFirOf6z9R5Yg7j+yLHFdFLzHz0lpj9KlErhddGRLZIksHWkMxAGipQRf2o77V4sU37d6/PJRuuwRroGZvKYcpAmU332mz0Cbruv8Odh40EiJJpFIdI7caBKJROfIjSaRSHSOSaGj6UH5OPP4o48+ulXHupG77rqrKSuX5gjcO+64o1XHeh7W5XDiIqDt6amewdw/m7pV38S6FtVJsJldPYpLSbyjI4JVj8TjlRKEA+21i6KmWX+jJuFIb8Jtuf/I21XnWIpWjjyl1WWA2/I7oAnfo+TerENRT29+N9n7l5PoA8BrXvOapqxnQ+2zzz5DzkP1QVHi8shLe6IxeWaSSCQ2WORGk0gkOsekok58hpFCTXds1mPxlL0tAeCaa65pykpL+EhSzhcbmWjV7MvBcKtXr27KKuKyOK1UgSmGmrd5TSJv2pLHrLZlahN57kZJq6I+mJaoF2spZ7CuB9dp/yWvYaVHPLb2z2vM96U0kNc08i6eNWtWsY77YLM3AFx55ZVN+cgjj2zV8Vr167E9mc5xUqREk0gkOkduNIlEonPkRpNIJDrHpCJ1qqOJEi+XEobzWTlAOdkS0DZjs3lbx2LTo5pAWb/CHFnN4DxHPb+b+b9G8XLkOF+nJnJeqyhrYSkcQecfRWWXTN1ap5HuJV1DFHoSJeOO5lFKGKZ98FjRGd1R8izVm3CyfHZdUBN2lFhLI72HmrtC1yrN24lE4hWF3GgSiUTnGDh1cvdGxFa6wSK1RlSzWMjeuVwGgPvvv78pH3zwwa06pgdsLldawlHTUR5Wjg5XEZfFcu4PaFMnNQlzHVPLKLpaKQvfJ9NCPd43MiuXTP56n5wESqkIP8+IejC0rpRYKzreVz2UeY2jM6qic6OY2ugRymxqv+yyy5qyHtfMdEzzDvN7HEVvcz5hBf89KW2baKREk0gkOkduNIlEonMMnDqZWUNVlDawtUDFX6YDK1eubMoq4kYeuSxORsmWuE6pGYu8bCHSeZS8RfU69UBlawT3oRY67jOynnBZx4py9XKgJlOzKHlW5PHL9DQ6ZrjfnMFK4SKLFN93lFyt3zzMGsTKlIj7WLx4cavd/vvvX+yDk7KVjhLWz1GytUEjJZpEItE5cqNJJBKdIzeaRCLROQauo2FE5sUDDjigVcd6Aj6/RtuxN6163XKkN4/NkdxAm+uqjoZ1KMyz1bPz4YcfbspRwm3l1fyZ26nOQD8zSkfiqu6C9SbqasCIdDmqH2KUPI/VNB0ddVvSSUR6nujI3eiYYf6sJmx+1pH+kM39qkPh5FacXA1om8j5Hda/kSgp/WRCSjSJRKJzjGmjMbONzWyBmV1Sf55rZjeY2VIzO9/Mpg3XRyKR2PAxVur0YQD3AOhlefocgM+7+3lm9mUApwL40nCd9MQ/Nd2xCKreugwW8yOzrB6XW0oUpEeccsIiTWjFc+b+VJxm0VuDKplmaUAne36y57F6iLIIrbStRA90vXnOmou2dMyrmmWZskTeuhHtKXlsA21vZn4nojOZFEzvuP/IC1nBJnKliyUafvfdd7fa3XLLLU157733btXxexYlJCslNRuq7SAxaonGzGYDeBuAf6k/G4BjAXy3bnIOgBPHOsFEIjH1MRbq9AUAfwagt31vD+Bxd++JFCsBzBrqQjM7zczmm9n8qaLMSiQSo8eoNhozezuANe5+82iud/ez3f0wdz9sMol3iUSiG4xWR3M0gN80s7cC2AyVjuaLAKab2Sa1VDMbwKqRdBqdr606Dz4jJ4puXbhwYVOeOXNmq45Nisz9dR7M3VV/w1yak2JptCyHDGioBX+O9BoM1aFwGEOUcCo6H5x1JZrgi58F96FrFSV25/6jqOkompt1KnydJifneUVz4ueu68afVXfG/aseiXVp7A7B5zgB7TXVPlgvGD2zaK0mE1sYlUTj7h9z99nuvjuAdwO4yt3fB+BqAO+qm50M4KJxmWUikZjSGG8/mjMA/ImZLUWls/nqOPefSCSmIMbsGezuPwbw47p8L4DDR3h98UjU6MhT9shlWnXttde22jGVUvMf0x4WoVUcZbqklIVFXjaHMpUB2iZbTXzF40XJv5h+RRROza0lk3Z0NpRSUJ5HvzmJlYqUopD79f7V/hlKPXhspao8Hq+Brlt0rhN7bOv7y205kjuiMmzqBtrrzx7tURS5QtUNg0R6BicSic6RG00ikegcAw+qNLNGRFVPzP32268pazAjB0TusssuTZmPqAWAtWvXNmU90oKPxGVKxEfbAm2ROhJdWURXCqTzZyxbtqwp67EePB574aoFLTq+hOfP96mUIko4xfMo5f7Vdv1ak0ZyLEjJW1e/57EjChEd78uf9bkwPYqsVUx39Z3gOescS/RO3z+eo9IotlapFXGikRJNIpHoHLnRJBKJzpEbTSKR6BwD19G4e8MtlX/yOThqVmbTI5v/NDKauSknGgLaPJgjnpXvM+dWfQLrOdgkqebnJ554otgHQ3k2z4V1NKpfYZNw5HnM+hUdi3UIkc4jSugdgceLdErR8b4lj9+RJEkvefX2e3Qu0NZ/qH6F9T6sX1E9Dycr1/Xm8Y466qimzOdEAXHi8smElGgSiUTnyI0mkUh0joFTJ4aKriyeah0HnbFXpdIGNgNHx8jy2VDax5w5c5qymr657aOPPtqUNffvjBkzmrImz4qOVy0F10VnTyklYmrJJn31duX10T5K5xip2ZRFfjUX8/yjgMjoPK/Ssb06Vun8pwh6z7ym6jXMz109zvk6XuMlS5a02vFaHXvssa06dtngd1Np4GQKnIyQEk0ikegcudEkEonOkRtNIpHoHAPX0ZhZY5aLuLSai/lsYoYmeWZdieoT+NxsNkOqjmbp0qVNWTkxmxQ5Kpt1MkBbL6P6G07ApaEK3JaTabHeCGjrctQVoGT21Pvk61SfxXoTNtWrfiXS0XBdZN6OXO5LfWgYQHTWOetNuE5dI7h/reP71tCCUkIrft+A9tnbqi/j94x1bCPRyUwm/U1KNIlEonPkRpNIJDrHwKkTQ0V8pktqQvzZz37WlGfPnt2UVZRnMV9N0xxZy7RBPThZDOcEWQCw7777NuUo6RPPX/tgEV2P7eX74fmqKH/fffc1ZTZh67z4XlS0ZuqnR7QylWIaqB6zpeNmFWy21rWKruN15Ov0/SidIaVz5vXQe2Haqn1o0jBGKQe0rumNN97YlI8++uhWHb/7TKt0HqWzyYDJ5SmcEk0ikegcudEkEonOMamok4rMLD7OmzevVcc0gqmTgimGWllWrVp3GozSJcYhhxzSlNXCwOIpe8yq9y9bH5TesfirdeyRWjqWVtupiM5Ujeevlo5Srlsdj2mPiu48R6V3peNylfYwLVFPabZQRZZCXg+1oJW8hpUO8RyVsvC7qe8EU67IknfQQQc1ZaU5/C5xsG7kAa19RMcCTzRSokkkEp0jN5pEItE5cqNJJBKdY1LoaHrcUjks6wlUJ8FcnXUSmrSbTbYaPcs8mD1tOZoaaHNw5eOsU4k8VblddO5S5JHLegI1kXM7NdMyV2fdguqDIm9a7oPnqPNl/U0pkbhCPYP5s+rteF15PaLk3ooooRWDn5Pqs1hXonok7pO9gdXtgKP9NSkbPyd+h/uNRB9qzoNESjSJRKJzjEmiMbPlAJ4C8BKAF939MDPbDsD5AHYHsBzASe7+WKmPRCKx4WM8qNOb3J2jxc4EcKW7n2VmZ9afz+inIzXPseh35513tuo0aLGHhx9+uPWZzZ4qnnIdBzNygCLQpktKe1hMZjE/SkylIm1kRmVRmemLtuN5KR3gzzy2mj+5D6UiTDMjb9TS8cZDzbmfayLKEtHFyARfOhZY2zFN0+fO/XOwq4Lnq8ck831feumlrTrOE8xnlenfSJQzeCTnZXWNLmbyTgDn1OVzAJzYwRiJRGIKYawbjQO4zMxuNrPT6u92cvcH6/JqADvpRWZ2mpnNN7P5kymUPZFIdIOxUqfXu/sqM9sRwOVmtpAr3d3NbL2dxN3PBnA2AGy88ca50yQSGzjGtNG4+6r6/zVm9u8ADgfwkJnNdPcHzWwmgDXD9dPjlsoxmQerKfaYY45pyjfddFNTVl7KvFhN32x+ZVN6pLuIIoujZE5RovVSMiedI5dHkpyc+4z0MHzfrJMBYn0Fg3US+jxL+hXtj+8lOr+K71PXg8eKzumK7oXDXDTyn8MMNHShdC63vsMMDdfg++QobzVv93sG1qAxaupkZlua2da9MoDjAdwJ4GIAJ9fNTgZw0VgnmUgkpjbGItHsBODf611zEwDfdvcfmtlNAC4ws1MB3AfgpLFPM5FITGWMeqNx93sBHDzE948COG40fSqlYHGY86sCwAMPPNCUWXRVcZFN1WpeZLrEorCK4ZGZsGTSjsRYFaEjb0+mByzm6zV8L0rbeI48j+g8In0WPA/2Wo1M3Qpuy5RIqU10nG1k4mco9WOwJy8/W6WSmnuZwVRH12D58uVNmaP4o/VW+sXvCEdvcyI3nf9kxtSYZSKRmNLIjSaRSHSO3GgSiUTnmFTR2wrmtIsXL27VMcdnM6Sei8S6CzWVMphL6xlSPL9IfxPpP0r9KVQ3wm1ZLxOdE60R5iXdURTyofoK7oPH1vuMIrb53qJzs7ndNtts06pjvUkpsh0orxvQfg+4TrMK8j2rGZxDXfSMMe6Ts+ipvmnZsmVNWTNIlsaO1mqDNG8nEolEv8iNJpFIdI5JQZ16Ym8UfaqUgpMGcXLySHyMjmiNKEXJs5bnrlCzI4vNKqJHcyyZviOTeL8is9IephQaNV2ihf0mmALadCyKOmbaoPPg9ec+Io/t6HlyH0r7onXkJFaa0Grt2rVNec2adY7xmgwtSrC/4447Dnmd0q9oHScTUqJJJBKdIzeaRCLROSYFdeqJfCqGs2ipnp7clr0oVfyNEkKVrDhR4qgoUDCieiWrjc653wDAyDqlXqZ8XbRWfJ32z/fJ96KiPI+lnrX9Up3IIsVjR+vN66bPjO+T56v3wl64Og+mj5ykCgAWLlyXxICDMbX/3XbbrSmrpZOTXUXrMRLqOkikRJNIJDpHbjSJRKJz5EaTSCQ6x6TQ0fR4p+oumMOymRBoc3CO0NYzdqJo31K0suqDIh0K8+B+PVXVzBnpLnj+Jc9anfOTTz7ZqivphzSKPDKzc12/ycSjZN+RWTaKyuaxIzN4dKY2j8drFZ1drX1wRLXqSQ488MCmzFHeHNUNtHVYOja/x6VEWkONPVkxNWaZSCSmNHKjSSQSnWNSUKceVGRm8VHpDCeqYrFT8wKzqKniL3vCRsekRl64LF5H3sUs8kfey1G+39IZT0B8XhOPzeUoOZSuQYnO6PfReUql/vv12NbPUZBpdMYW06XoTC2+TvvoN+EXv48rVqxo1bH7hj6zUiI2dRmI3rnJhJRoEolE58iNJpFIdI7caBKJROeYFDqaHjdW0+6qVauaMpsTgbIZVZM+sWu3mgb75bdR+EBJn6D6oH6ji0cbvV06M0n7Z/O8JreKzMp8Xb9u7+oKUDIXq74jCgfhun7vOUpwznrAftcNiEMctthii6bMZ8QffHA7lz+vv47N68NjP/LII612rOeZzEiJJpFIdI7caBKJROcYOHVy92LiKxZro2RRnO83ynWr1KlkLo5MqlHiK+4/OvY2yvva75G7UR+RmTZqxxQjyv3bb5IwXe9S9LnSo4iylI7BjahkdM4Vzyny+h7JO8H3zTRq7ty5rXaLFi0ack7aZ+T+MJlN2oyUaBKJROcY00ZjZtPN7LtmttDM7jGzo8xsOzO73MyW1P9vO16TTSQSUxNjpU5fBPBDd3+XmU0DsAWAPwdwpbufZWZnAjgTwBn9dBZZGCJrTxTkV7JSKKI6DT4sXRcFPTKiIEK9z1JiLW1XohQ6r4gOlKwxOg/2qNa14bEiqhrdS7R2JWoZrUcEnr9aCqNgV26r1/G9sbVUk1vtuuuuTXnbbdu/x3x8849+9KMh+x4Ok4lWjVqiMbNtALwBwFcBwN1fcPfHAbwTwDl1s3MAnDjWSSYSiamNsVCnuQAeBvB1M1tgZv9iZlsC2MndH6zbrAawk15oZqeZ2Xwzmz+G8ROJxBTBWDaaTQAcCuBL7n4IgGdQ0aQGXsme68mw7n62ux/m7oeNYfxEIjFFMBYdzUoAK939hvrzd1FtNA+Z2Ux3f9DMZgJYU+yhRsm8HfFsjjyOvED7TTrOps0oIZSaQJm789jRWIpIj1SafxStHOkrouNsWdega9CvzqPfdmy+jXQj/ertovWI1j5yLWAdU5QILHonSu8p0D6bTJ8FR5hz/5H7g6LfZzERGLVE4+6rAawws33rr44DcDeAiwGcXH93MoCLxjTDRCIx5TFWq9OHAHyrtjjdC+D3UG1eF5jZqQDuA3DSGMdIJBJTHGPaaNz9VgBD6VmOG2E/ANYXcVm8VtGSRdLI1MjtIhM596/mYRZdVYQu0aWRmOpHI+YrReT71jmydyqbpqPAT50Hr0EkkkfJqErnKalnbeka7bOUr1nnwfesfUbJxKLzlCKazHPU/M0Mnv/KlStbdSXKFXkoR/0PGukZnEgkOkduNIlEonPkRpNIJDrHwKO3gbKOJjpbmRP+RCEIkTs76x1YL6PmyqgP1i+MhzkxSgLF84rc3nUdOZwgCqeIUDL/63lKkZ6nZMbX9Y10c6WEZ/rMSq4LWhclHeN70fuMktmXouy1HfevIR+sVxstNgjzdiKRSPSL3GgSiUTnsEGLV2b2MCp/mx0APDJM84lAzqONnEcbU3Ueu7n7jOGbdYOBbzQ9mNn8yRD7lPPIeeQ8xh9JnRKJROfIjSaRSHSOybTRnD3oCdTIebSR82gj5zEKTBodTSKR2HAxmSSaRCKxgSI3mkQi0TkmxUZjZieY2SIzW1qfnDBR437NzNaY2Z303YQfF2Nmc8zsajO728zuMrMPD2IuZraZmd1oZrfV8/hU/f1cM7uhfj7n1/mHOoWZbVznor5kUHOox11uZneY2a29HNcDeC5T/lijgW80ZrYxgH8E8BYABwB4j5kdMEHDfwPACfLdmaiOi9kbwJWQPMgd4UUAp7v7AQCOBPDH9RpM9FyeB3Csux8MYB6AE8zsSACfA/B5d98LwGMATu14HgDwYQD30OdBzKGHN7n7PPJbmejn0jvWaD8AB6Nal0G8p6NH70jaQf0DcBSAH9HnjwH42ASOvzuAO+nzIgAz6/JMAIsGsCYXAXjzIOeC6oyuWwAcgcoDdZOhnldHY89G9cdzLIBLANhEz4HmshzADvLdhD0XANsAWIbacDOIOYzHv4FLNABmAVhBn1fW3w0Kwx4X0yXMbHcAhwC4YRBzqSnLraiSyl8O4BcAHnf3Xoj8RDyfLwD4MwC90OftBzCHHhzAZWZ2s5mdVn83kc9l1McaTSZMho1m0sKrn4sJs/+b2VYALgTwEXdv5YCcqLm4+0vuPg+VVHE4gP26HpNhZm8HsMbdb57IcQO83t0PRUXt/9jM3sCVE/BcRn2s0WTCZNhoVgGYQ59n198NCg/Vx8Sg3+NixgNmtimqTeZb7v69Qc4FALw6dfRqVDRlupn1kr90/XyOBvCbZrYcwHmo6NMXJ3gODdx9Vf3/GgD/jmrzncjnMtSxRodO8BzGjMmw0dwEYO/aqjANwLtRHdkyKEz4cTFWZVH6KoB73P3vBjUXM5thZtPr8uao9ET3oNpw3jUR83D3j7n7bHffHdW7cJW7v28i59CDmW1pZlv3ygCOB3AnJvC5+IZyrNGglUS1MuutABaj0gd8fALH/Q6ABwH8CtUvx6mo9AFXAlgC4AoA203APF6PSvS9HcCt9b+3TvRcABwEYEE9jzsBfKL+fg8ANwJYCuDfALxqgp7PGwFcMqg51GPeVv+7q/duDuC5zAMwv34u3wew7SDe07H8yxCERCLROSYDdUokEhs4cqNJJBKdIzeaRCLROXKjSSQSnSM3mkQi0Tlyo0kkEp0jN5pEItE5/j+6eH7GN+b12gAAAABJRU5ErkJggg==\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}