waveletCNN_Keras_0_6_[Exp_on_10K_data].ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": " waveletCNN_Keras_0.6_[Exp on 10K_data].ipynb",
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "zQyX8ddJk32y",
        "colab_type": "code",
        "outputId": "85d913ac-cc14-4195-f6f8-ef81b190f334",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "source": [
        "import numpy as np\n",
        "from matplotlib import pyplot as plt\n",
        "\n",
        "from keras import backend as K\n",
        "from keras.models import Model\n",
        "from keras.layers import Input\n",
        "from keras.layers import Dense\n",
        "from keras.layers import Conv2D\n",
        "from keras.layers import Lambda\n",
        "from keras.layers import Flatten\n",
        "from keras.layers import Reshape\n",
        "from keras.layers import Dropout\n",
        "from keras.layers import Activation\n",
        "from keras.layers import AveragePooling2D\n",
        "from keras.layers import BatchNormalization\n",
        "from keras.layers.merge import add, concatenate\n",
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "from keras.utils import plot_model"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Using TensorFlow backend.\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EMcF2AfPk_SX",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# batch operation usng tensor slice\n",
        "def WaveletTransformAxisY(batch_img):\n",
        "    odd_img  = batch_img[:,0::2]\n",
        "    even_img = batch_img[:,1::2]\n",
        "    L = (odd_img + even_img) / 2.0\n",
        "    H = K.abs(odd_img - even_img)\n",
        "    return L, H\n",
        "\n",
        "def WaveletTransformAxisX(batch_img):\n",
        "    # transpose + fliplr\n",
        "    tmp_batch = K.permute_dimensions(batch_img, [0, 2, 1])[:,:,::-1]\n",
        "    _dst_L, _dst_H = WaveletTransformAxisY(tmp_batch)\n",
        "    # transpose + flipud\n",
        "    dst_L = K.permute_dimensions(_dst_L, [0, 2, 1])[:,::-1,...]\n",
        "    dst_H = K.permute_dimensions(_dst_H, [0, 2, 1])[:,::-1,...]\n",
        "    return dst_L, dst_H"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WLXXxjIhlCvn",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def Wavelet(batch_image):\n",
        "    # make channel first image\n",
        "    batch_image = K.permute_dimensions(batch_image, [0, 3, 1, 2])\n",
        "    r = batch_image[:,0]\n",
        "    g = batch_image[:,1]\n",
        "    b = batch_image[:,2]\n",
        "\n",
        "    # level 1 decomposition\n",
        "    wavelet_L, wavelet_H = WaveletTransformAxisY(r)\n",
        "    r_wavelet_LL, r_wavelet_LH = WaveletTransformAxisX(wavelet_L)\n",
        "    r_wavelet_HL, r_wavelet_HH = WaveletTransformAxisX(wavelet_H)\n",
        "\n",
        "    wavelet_L, wavelet_H = WaveletTransformAxisY(g)\n",
        "    g_wavelet_LL, g_wavelet_LH = WaveletTransformAxisX(wavelet_L)\n",
        "    g_wavelet_HL, g_wavelet_HH = WaveletTransformAxisX(wavelet_H)\n",
        "\n",
        "    wavelet_L, wavelet_H = WaveletTransformAxisY(b)\n",
        "    b_wavelet_LL, b_wavelet_LH = WaveletTransformAxisX(wavelet_L)\n",
        "    b_wavelet_HL, b_wavelet_HH = WaveletTransformAxisX(wavelet_H)\n",
        "\n",
        "    wavelet_data = [r_wavelet_LL, r_wavelet_LH, r_wavelet_HL, r_wavelet_HH, \n",
        "                    g_wavelet_LL, g_wavelet_LH, g_wavelet_HL, g_wavelet_HH,\n",
        "                    b_wavelet_LL, b_wavelet_LH, b_wavelet_HL, b_wavelet_HH]\n",
        "    transform_batch = K.stack(wavelet_data, axis=1)\n",
        "\n",
        "    # level 2 decomposition\n",
        "    wavelet_L2, wavelet_H2 = WaveletTransformAxisY(r_wavelet_LL)\n",
        "    r_wavelet_LL2, r_wavelet_LH2 = WaveletTransformAxisX(wavelet_L2)\n",
        "    r_wavelet_HL2, r_wavelet_HH2 = WaveletTransformAxisX(wavelet_H2)\n",
        "\n",
        "    wavelet_L2, wavelet_H2 = WaveletTransformAxisY(g_wavelet_LL)\n",
        "    g_wavelet_LL2, g_wavelet_LH2 = WaveletTransformAxisX(wavelet_L2)\n",
        "    g_wavelet_HL2, g_wavelet_HH2 = WaveletTransformAxisX(wavelet_H2)\n",
        "\n",
        "    wavelet_L2, wavelet_H2 = WaveletTransformAxisY(b_wavelet_LL)\n",
        "    b_wavelet_LL2, b_wavelet_LH2 = WaveletTransformAxisX(wavelet_L2)\n",
        "    b_wavelet_HL2, b_wavelet_HH2 = WaveletTransformAxisX(wavelet_H2)\n",
        "\n",
        "\n",
        "    wavelet_data_l2 = [r_wavelet_LL2, r_wavelet_LH2, r_wavelet_HL2, r_wavelet_HH2, \n",
        "                    g_wavelet_LL2, g_wavelet_LH2, g_wavelet_HL2, g_wavelet_HH2,\n",
        "                    b_wavelet_LL2, b_wavelet_LH2, b_wavelet_HL2, b_wavelet_HH2]\n",
        "    transform_batch_l2 = K.stack(wavelet_data_l2, axis=1)\n",
        "\n",
        "    # level 3 decomposition\n",
        "    wavelet_L3, wavelet_H3 = WaveletTransformAxisY(r_wavelet_LL2)\n",
        "    r_wavelet_LL3, r_wavelet_LH3 = WaveletTransformAxisX(wavelet_L3)\n",
        "    r_wavelet_HL3, r_wavelet_HH3 = WaveletTransformAxisX(wavelet_H3)\n",
        "\n",
        "    wavelet_L3, wavelet_H3 = WaveletTransformAxisY(g_wavelet_LL2)\n",
        "    g_wavelet_LL3, g_wavelet_LH3 = WaveletTransformAxisX(wavelet_L3)\n",
        "    g_wavelet_HL3, g_wavelet_HH3 = WaveletTransformAxisX(wavelet_H3)\n",
        "\n",
        "    wavelet_L3, wavelet_H3 = WaveletTransformAxisY(b_wavelet_LL2)\n",
        "    b_wavelet_LL3, b_wavelet_LH3 = WaveletTransformAxisX(wavelet_L3)\n",
        "    b_wavelet_HL3, b_wavelet_HH3 = WaveletTransformAxisX(wavelet_H3)\n",
        "\n",
        "    wavelet_data_l3 = [r_wavelet_LL3, r_wavelet_LH3, r_wavelet_HL3, r_wavelet_HH3, \n",
        "                    g_wavelet_LL3, g_wavelet_LH3, g_wavelet_HL3, g_wavelet_HH3,\n",
        "                    b_wavelet_LL3, b_wavelet_LH3, b_wavelet_HL3, b_wavelet_HH3]\n",
        "    transform_batch_l3 = K.stack(wavelet_data_l3, axis=1)\n",
        "\n",
        "    # level 4 decomposition\n",
        "    wavelet_L4, wavelet_H4 = WaveletTransformAxisY(r_wavelet_LL3)\n",
        "    r_wavelet_LL4, r_wavelet_LH4 = WaveletTransformAxisX(wavelet_L4)\n",
        "    r_wavelet_HL4, r_wavelet_HH4 = WaveletTransformAxisX(wavelet_H4)\n",
        "\n",
        "    wavelet_L4, wavelet_H4 = WaveletTransformAxisY(g_wavelet_LL3)\n",
        "    g_wavelet_LL4, g_wavelet_LH4 = WaveletTransformAxisX(wavelet_L4)\n",
        "    g_wavelet_HL4, g_wavelet_HH4 = WaveletTransformAxisX(wavelet_H4)\n",
        "\n",
        "    wavelet_L3, wavelet_H3 = WaveletTransformAxisY(b_wavelet_LL3)\n",
        "    b_wavelet_LL4, b_wavelet_LH4 = WaveletTransformAxisX(wavelet_L4)\n",
        "    b_wavelet_HL4, b_wavelet_HH4 = WaveletTransformAxisX(wavelet_H4)\n",
        "\n",
        "\n",
        "    wavelet_data_l4 = [r_wavelet_LL4, r_wavelet_LH4, r_wavelet_HL4, r_wavelet_HH4, \n",
        "                    g_wavelet_LL4, g_wavelet_LH4, g_wavelet_HL4, g_wavelet_HH4,\n",
        "                    b_wavelet_LL4, b_wavelet_LH4, b_wavelet_HL4, b_wavelet_HH4]\n",
        "    transform_batch_l4 = K.stack(wavelet_data_l4, axis=1)\n",
        "\n",
        "    # print('shape before')\n",
        "    # print(transform_batch.shape)\n",
        "    # print(transform_batch_l2.shape)\n",
        "    # print(transform_batch_l3.shape)\n",
        "    # print(transform_batch_l4.shape)\n",
        "\n",
        "    decom_level_1 = K.permute_dimensions(transform_batch, [0, 2, 3, 1])\n",
        "    decom_level_2 = K.permute_dimensions(transform_batch_l2, [0, 2, 3, 1])\n",
        "    decom_level_3 = K.permute_dimensions(transform_batch_l3, [0, 2, 3, 1])\n",
        "    decom_level_4 = K.permute_dimensions(transform_batch_l4, [0, 2, 3, 1])\n",
        "    \n",
        "    # print('shape after')\n",
        "    # print(decom_level_1.shape)\n",
        "    # print(decom_level_2.shape)\n",
        "    # print(decom_level_3.shape)\n",
        "    # print(decom_level_4.shape)\n",
        "    return [decom_level_1, decom_level_2, decom_level_3, decom_level_4]\n",
        "\n",
        "\n",
        "def Wavelet_out_shape(input_shapes):\n",
        "    # print('in to shape')\n",
        "    return [tuple([None, 112, 112, 12]), tuple([None, 56, 56, 12]), \n",
        "            tuple([None, 28, 28, 12]), tuple([None, 14, 14, 12])]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "h6dmgkdTlDOk",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "img_batch = K.zeros(shape=(8, 224, 224, 3), dtype='float32')\n",
        "Wavelet(img_batch)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2MpxaRCZlF1J",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def get_wavelet_cnn_model():\n",
        "\n",
        "    input_shape = 224, 224, 3\n",
        "\n",
        "    input_ = Input(input_shape, name='the_input')\n",
        "    # wavelet = Lambda(Wavelet, name='wavelet')\n",
        "    wavelet = Lambda(Wavelet, Wavelet_out_shape, name='wavelet')\n",
        "    input_l1, input_l2, input_l3, input_l4 = wavelet(input_)\n",
        "    # print(input_l1)\n",
        "    # print(input_l2)\n",
        "    # print(input_l3)\n",
        "    # print(input_l4)\n",
        "    # level one decomposition starts\n",
        "    conv_1 = Conv2D(64, kernel_size=(3, 3), padding='same', name='conv_1')(input_l1)\n",
        "    norm_1 = BatchNormalization(name='norm_1')(conv_1)\n",
        "    relu_1 = Activation('relu', name='relu_1')(norm_1)\n",
        "\n",
        "    conv_1_2 = Conv2D(64, kernel_size=(3, 3), strides=(2, 2), padding='same', name='conv_1_2')(relu_1)\n",
        "    norm_1_2 = BatchNormalization(name='norm_1_2')(conv_1_2)\n",
        "    relu_1_2 = Activation('relu', name='relu_1_2')(norm_1_2)\n",
        "\n",
        "    # level two decomposition starts\n",
        "    conv_a = Conv2D(filters=64, kernel_size=(3, 3), padding='same', name='conv_a')(input_l2)\n",
        "    norm_a = BatchNormalization(name='norm_a')(conv_a)\n",
        "    relu_a = Activation('relu', name='relu_a')(norm_a)\n",
        "\n",
        "    # concate level one and level two decomposition\n",
        "    concate_level_2 = concatenate([relu_1_2, relu_a])\n",
        "    conv_2 = Conv2D(128, kernel_size=(3, 3), padding='same', name='conv_2')(concate_level_2)\n",
        "    norm_2 = BatchNormalization(name='norm_2')(conv_2)\n",
        "    relu_2 = Activation('relu', name='relu_2')(norm_2)\n",
        "\n",
        "    conv_2_2 = Conv2D(128, kernel_size=(3, 3), strides=(2, 2), padding='same', name='conv_2_2')(relu_2)\n",
        "    norm_2_2 = BatchNormalization(name='norm_2_2')(conv_2_2)\n",
        "    relu_2_2 = Activation('relu', name='relu_2_2')(norm_2_2)\n",
        "\n",
        "    # level three decomposition starts \n",
        "    conv_b = Conv2D(filters=64, kernel_size=(3, 3), padding='same', name='conv_b')(input_l3)\n",
        "    norm_b = BatchNormalization(name='norm_b')(conv_b)\n",
        "    relu_b = Activation('relu', name='relu_b')(norm_b)\n",
        "\n",
        "    conv_b_2 = Conv2D(128, kernel_size=(3, 3), padding='same', name='conv_b_2')(relu_b)\n",
        "    norm_b_2 = BatchNormalization(name='norm_b_2')(conv_b_2)\n",
        "    relu_b_2 = Activation('relu', name='relu_b_2')(norm_b_2)\n",
        "\n",
        "    # concate level two and level three decomposition \n",
        "    concate_level_3 = concatenate([relu_2_2, relu_b_2])\n",
        "    conv_3 = Conv2D(256, kernel_size=(3, 3), padding='same', name='conv_3')(concate_level_3)\n",
        "    norm_3 = BatchNormalization(name='nomr_3')(conv_3)\n",
        "    relu_3 = Activation('relu', name='relu_3')(norm_3)\n",
        "\n",
        "    conv_3_2 = Conv2D(256, kernel_size=(3, 3), strides=(2, 2), padding='same', name='conv_3_2')(relu_3)\n",
        "    norm_3_2 = BatchNormalization(name='norm_3_2')(conv_3_2)\n",
        "    relu_3_2 = Activation('relu', name='relu_3_2')(norm_3_2)\n",
        "\n",
        "    # level four decomposition start\n",
        "    conv_c = Conv2D(64, kernel_size=(3, 3), padding='same', name='conv_c')(input_l4)\n",
        "    norm_c = BatchNormalization(name='norm_c')(conv_c)\n",
        "    relu_c = Activation('relu', name='relu_c')(norm_c)\n",
        "\n",
        "    conv_c_2 = Conv2D(256, kernel_size=(3, 3), padding='same', name='conv_c_2')(relu_c)\n",
        "    norm_c_2 = BatchNormalization(name='norm_c_2')(conv_c_2)\n",
        "    relu_c_2 = Activation('relu', name='relu_c_2')(norm_c_2)\n",
        "\n",
        "    conv_c_3 = Conv2D(256, kernel_size=(3, 3), padding='same', name='conv_c_3')(relu_c_2)\n",
        "    norm_c_3 = BatchNormalization(name='norm_c_3')(conv_c_3)\n",
        "    relu_c_3 = Activation('relu', name='relu_c_3')(norm_c_3)\n",
        "\n",
        "    # concate level level three and level four decomposition\n",
        "    concate_level_4 = concatenate([relu_3_2, relu_c_3])\n",
        "    conv_4 = Conv2D(256, kernel_size=(3, 3), padding='same', name='conv_4')(concate_level_4)\n",
        "    norm_4 = BatchNormalization(name='norm_4')(conv_4)\n",
        "    relu_4 = Activation('relu', name='relu_4')(norm_4)\n",
        "\n",
        "    conv_4_2 = Conv2D(256, kernel_size=(3, 3), strides=(2, 2), padding='same', name='conv_4_2')(relu_4)\n",
        "    norm_4_2 = BatchNormalization(name='norm_4_2')(conv_4_2)\n",
        "    relu_4_2 = Activation('relu', name='relu_4_2')(norm_4_2)\n",
        "\n",
        "    conv_5_1 = Conv2D(128, kernel_size=(3, 3), padding='same', name='conv_5_1')(relu_4_2)\n",
        "    norm_5_1 = BatchNormalization(name='norm_5_1')(conv_5_1)\n",
        "    relu_5_1 = Activation('relu', name='relu_5_1')(norm_5_1)\n",
        "\n",
        "    pool_5_1 = AveragePooling2D(pool_size=(7, 7), strides=1, padding='same', name='avg_pool_5_1')(relu_5_1)\n",
        "    flat_5_1 = Flatten(name='flat_5_1')(pool_5_1) \n",
        "\n",
        "    fc_5 = Dense(2048, name='fc_5')(flat_5_1)\n",
        "    norm_5 = BatchNormalization(name='norm_5')(fc_5)\n",
        "    relu_5 = Activation('relu', name='relu_5')(norm_5)\n",
        "    drop_5 = Dropout(0.5, name='drop_5')(relu_5)\n",
        "\n",
        "    fc_6 = Dense(2048, name='fc_6')(drop_5)\n",
        "    norm_6 = BatchNormalization(name='norm_6')(fc_6)\n",
        "    relu_6 = Activation('relu', name='relu_6')(norm_6)\n",
        "    drop_6 = Dropout(0.5, name='drop_6')(relu_6)\n",
        "\n",
        "    output = Dense(2, activation='softmax', name='fc_7')(drop_6)\n",
        "\n",
        "    model = Model(inputs=input_, outputs=output)\n",
        "    model.summary()\n",
        "    plot_model(model, to_file='wavelet_cnn_0.5.png')\n",
        "\n",
        "    return model"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XlNN-MeWlQHX",
        "colab_type": "code",
        "outputId": "729d4720-1ed6-47d9-d3c6-21c4f7008202",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "model = get_wavelet_cnn_model()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:4432: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:66: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:190: The name tf.get_default_session is deprecated. Please use tf.compat.v1.get_default_session instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:197: The name tf.ConfigProto is deprecated. Please use tf.compat.v1.ConfigProto instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:2041: The name tf.nn.fused_batch_norm is deprecated. Please use tf.compat.v1.nn.fused_batch_norm instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:4271: The name tf.nn.avg_pool is deprecated. Please use tf.nn.avg_pool2d instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:3733: calling dropout (from tensorflow.python.ops.nn_ops) with keep_prob is deprecated and will be removed in a future version.\n",
            "Instructions for updating:\n",
            "Please use `rate` instead of `keep_prob`. Rate should be set to `rate = 1 - keep_prob`.\n",
            "Model: \"model_1\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "the_input (InputLayer)          (None, 224, 224, 3)  0                                            \n",
            "__________________________________________________________________________________________________\n",
            "wavelet (Lambda)                [(None, 112, 112, 12 0           the_input[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "conv_1 (Conv2D)                 (None, 112, 112, 64) 6976        wavelet[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "norm_1 (BatchNormalization)     (None, 112, 112, 64) 256         conv_1[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "relu_1 (Activation)             (None, 112, 112, 64) 0           norm_1[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv_1_2 (Conv2D)               (None, 56, 56, 64)   36928       relu_1[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv_a (Conv2D)                 (None, 56, 56, 64)   6976        wavelet[0][1]                    \n",
            "__________________________________________________________________________________________________\n",
            "norm_1_2 (BatchNormalization)   (None, 56, 56, 64)   256         conv_1_2[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "norm_a (BatchNormalization)     (None, 56, 56, 64)   256         conv_a[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "relu_1_2 (Activation)           (None, 56, 56, 64)   0           norm_1_2[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "relu_a (Activation)             (None, 56, 56, 64)   0           norm_a[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "concatenate_1 (Concatenate)     (None, 56, 56, 128)  0           relu_1_2[0][0]                   \n",
            "                                                                 relu_a[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv_2 (Conv2D)                 (None, 56, 56, 128)  147584      concatenate_1[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv_b (Conv2D)                 (None, 28, 28, 64)   6976        wavelet[0][2]                    \n",
            "__________________________________________________________________________________________________\n",
            "norm_2 (BatchNormalization)     (None, 56, 56, 128)  512         conv_2[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "norm_b (BatchNormalization)     (None, 28, 28, 64)   256         conv_b[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "relu_2 (Activation)             (None, 56, 56, 128)  0           norm_2[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "relu_b (Activation)             (None, 28, 28, 64)   0           norm_b[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv_2_2 (Conv2D)               (None, 28, 28, 128)  147584      relu_2[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv_b_2 (Conv2D)               (None, 28, 28, 128)  73856       relu_b[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "norm_2_2 (BatchNormalization)   (None, 28, 28, 128)  512         conv_2_2[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "norm_b_2 (BatchNormalization)   (None, 28, 28, 128)  512         conv_b_2[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "conv_c (Conv2D)                 (None, 14, 14, 64)   6976        wavelet[0][3]                    \n",
            "__________________________________________________________________________________________________\n",
            "relu_2_2 (Activation)           (None, 28, 28, 128)  0           norm_2_2[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "relu_b_2 (Activation)           (None, 28, 28, 128)  0           norm_b_2[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "norm_c (BatchNormalization)     (None, 14, 14, 64)   256         conv_c[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "concatenate_2 (Concatenate)     (None, 28, 28, 256)  0           relu_2_2[0][0]                   \n",
            "                                                                 relu_b_2[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "relu_c (Activation)             (None, 14, 14, 64)   0           norm_c[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv_3 (Conv2D)                 (None, 28, 28, 256)  590080      concatenate_2[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "conv_c_2 (Conv2D)               (None, 14, 14, 256)  147712      relu_c[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "nomr_3 (BatchNormalization)     (None, 28, 28, 256)  1024        conv_3[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "norm_c_2 (BatchNormalization)   (None, 14, 14, 256)  1024        conv_c_2[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "relu_3 (Activation)             (None, 28, 28, 256)  0           nomr_3[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "relu_c_2 (Activation)           (None, 14, 14, 256)  0           norm_c_2[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "conv_3_2 (Conv2D)               (None, 14, 14, 256)  590080      relu_3[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv_c_3 (Conv2D)               (None, 14, 14, 256)  590080      relu_c_2[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "norm_3_2 (BatchNormalization)   (None, 14, 14, 256)  1024        conv_3_2[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "norm_c_3 (BatchNormalization)   (None, 14, 14, 256)  1024        conv_c_3[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "relu_3_2 (Activation)           (None, 14, 14, 256)  0           norm_3_2[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "relu_c_3 (Activation)           (None, 14, 14, 256)  0           norm_c_3[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "concatenate_3 (Concatenate)     (None, 14, 14, 512)  0           relu_3_2[0][0]                   \n",
            "                                                                 relu_c_3[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "conv_4 (Conv2D)                 (None, 14, 14, 256)  1179904     concatenate_3[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "norm_4 (BatchNormalization)     (None, 14, 14, 256)  1024        conv_4[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "relu_4 (Activation)             (None, 14, 14, 256)  0           norm_4[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "conv_4_2 (Conv2D)               (None, 7, 7, 256)    590080      relu_4[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "norm_4_2 (BatchNormalization)   (None, 7, 7, 256)    1024        conv_4_2[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "relu_4_2 (Activation)           (None, 7, 7, 256)    0           norm_4_2[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "conv_5_1 (Conv2D)               (None, 7, 7, 128)    295040      relu_4_2[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "norm_5_1 (BatchNormalization)   (None, 7, 7, 128)    512         conv_5_1[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "relu_5_1 (Activation)           (None, 7, 7, 128)    0           norm_5_1[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "avg_pool_5_1 (AveragePooling2D) (None, 7, 7, 128)    0           relu_5_1[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "flat_5_1 (Flatten)              (None, 6272)         0           avg_pool_5_1[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "fc_5 (Dense)                    (None, 2048)         12847104    flat_5_1[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "norm_5 (BatchNormalization)     (None, 2048)         8192        fc_5[0][0]                       \n",
            "__________________________________________________________________________________________________\n",
            "relu_5 (Activation)             (None, 2048)         0           norm_5[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "drop_5 (Dropout)                (None, 2048)         0           relu_5[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "fc_6 (Dense)                    (None, 2048)         4196352     drop_5[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "norm_6 (BatchNormalization)     (None, 2048)         8192        fc_6[0][0]                       \n",
            "__________________________________________________________________________________________________\n",
            "relu_6 (Activation)             (None, 2048)         0           norm_6[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "drop_6 (Dropout)                (None, 2048)         0           relu_6[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "fc_7 (Dense)                    (None, 2)            4098        drop_6[0][0]                     \n",
            "==================================================================================================\n",
            "Total params: 21,490,242\n",
            "Trainable params: 21,477,314\n",
            "Non-trainable params: 12,928\n",
            "__________________________________________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "w6zpGp6flIEz",
        "colab_type": "code",
        "outputId": "25b3b88a-b5f9-42d1-c769-8d46517c3bd7",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 73
        }
      },
      "source": [
        "model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/optimizers.py:793: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JKG4AAEDlWXm",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "!pip install -q kaggle\n",
        "!mkdir -p ~/.kaggle\n",
        "!cp kaggle.json ~/.kaggle/\n",
        "\n",
        "# This permissions change avoids a warning on Kaggle tool startup.\n",
        "!chmod 600 ~/.kaggle/kaggle.json"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RQ7oNY98mc6q",
        "colab_type": "code",
        "outputId": "4965f6df-01ec-45d7-dcdb-71085d2999c6",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 448
        }
      },
      "source": [
        "!kaggle datasets list -s cat"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Warning: Looks like you're using an outdated API Version, please consider updating (server 1.5.6 / client 1.5.4)\n",
            "ref                                                 title                                          size  lastUpdated          downloadCount  \n",
            "--------------------------------------------------  --------------------------------------------  -----  -------------------  -------------  \n",
            "tongpython/cat-and-dog                              Cat and Dog                                   218MB  2018-04-26 10:56:50           9950  \n",
            "crawford/cat-dataset                                Cat Dataset                                     4GB  2018-02-16 23:41:25           6555  \n",
            "rmisra/news-category-dataset                        News Category Dataset                          25MB  2018-12-02 04:09:45           7084  \n",
            "harriken/bias-media-cat                             Bias Media CAT                                 17MB  2017-10-27 22:46:25            283  \n",
            "mrisdal/open-exoplanet-catalogue                    Open Exoplanet Catalogue                      123KB  2017-06-08 19:59:22           2321  \n",
            "chetankv/dogs-cats-images                           Dogs & Cats Images                            435MB  2018-04-19 18:20:08           4978  \n",
            "shayanfazeli/heartbeat                              ECG Heartbeat Categorization Dataset           99MB  2018-05-31 18:47:34           7589  \n",
            "mmoreaux/audio-cats-and-dogs                        Audio Cats and Dogs                            97MB  2017-10-05 09:40:26           3009  \n",
            "hocop1/cat-and-dog-breeds-parameters                Cat and dog breeds parameters                  12KB  2019-01-26 16:39:25            429  \n",
            "lumins/settlers-of-catan-games                      My Settlers of Catan Games                      5KB  2016-08-29 19:52:21            729  \n",
            "lightforever/catalyst                               Catalyst                                       10MB  2019-08-07 19:05:32             18  \n",
            "bharadwaj6/kindle-reviews                           Amazon reviews: Kindle Store Category         525MB  2018-05-22 04:50:16           2563  \n",
            "uciml/ct-slice-localization                         CAT Scan Localization                          18MB  2017-09-06 22:01:03            368  \n",
            "mriganksingh/cat-images-dataset                     Cat images Dataset                              3MB  2019-04-13 20:24:18            369  \n",
            "victorgenin/ices-fish-catch                         Annual Nominal Fish Catches                      0B  2016-05-29 08:12:32           1281  \n",
            "biaiscience/dogs-vs-cats                            Dogs vs Cats                                  817MB  2017-12-05 14:03:33           2301  \n",
            "ulabox/ulabox-orders-with-categories-partials-2017  Ulabox orders with categories' partials 2017  607KB  2018-05-21 20:33:33            611  \n",
            "ppleskov/cute-cats-and-dogs-from-pixabaycom         Cute Cats and Dogs from Pixabay.com             2GB  2019-03-21 04:28:53            250  \n",
            "rahul897/catsdogs                                   cats&dogs                                     217MB  2018-01-05 15:02:32           2251  \n",
            "yufengdev/bbc-fulltext-and-category                 BBC articles fulltext and category              2MB  2018-06-08 05:44:22           1215  \n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CvhyBIfrlb8G",
        "colab_type": "code",
        "outputId": "bedc051b-b843-476e-8507-9cdcae8befa4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 71
        }
      },
      "source": [
        "!kaggle datasets download -d chetankv/dogs-cats-images"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading dogs-cats-images.zip to /content\n",
            " 98% 425M/435M [00:14<00:00, 19.3MB/s]\n",
            "100% 435M/435M [00:14<00:00, 32.1MB/s]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BDA9bhlQmvlX",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "!unzip -qq /content/dogs-cats-images.zip"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gDuuMOgEm-XS",
        "colab_type": "code",
        "outputId": "96e2c956-9217-4ad1-b401-cb23b1dc27d6",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        }
      },
      "source": [
        "import os\n",
        "data_set_path = '/content/dog vs cat/dataset'\n",
        "os.listdir(data_set_path)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['test_set', 'training_set']"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ejmciqicnMnu",
        "colab_type": "code",
        "outputId": "ebacd6d9-2162-4215-da3c-b4afca30a00d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 53
        }
      },
      "source": [
        "train_data_gen = ImageDataGenerator(\n",
        "\t#rescale=1./255,\n",
        "\tshear_range=0.1,\n",
        "\tzoom_range=0.1,\n",
        "\thorizontal_flip=True\n",
        ")\n",
        "val_data_gen = ImageDataGenerator(\n",
        "\t#rescale=1./255\n",
        ")\n",
        "\n",
        "train_img_dir = '/content/dog vs cat/dataset/training_set'\n",
        "val_img_dir   = '/content/dog vs cat/dataset/test_set'\n",
        "\n",
        "# def train & test generators\n",
        "train_generator = train_data_gen.flow_from_directory(\n",
        "\ttrain_img_dir,\n",
        "\ttarget_size=(224, 224),\n",
        "\tbatch_size=32,\n",
        "\tclass_mode='categorical')\n",
        "val_generator = val_data_gen.flow_from_directory(\n",
        "\tval_img_dir,\n",
        "\ttarget_size=(224, 224),\n",
        "\tbatch_size=32,\n",
        "\tclass_mode='categorical')"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 8000 images belonging to 2 classes.\n",
            "Found 2000 images belonging to 2 classes.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "V2740Mdynf0w",
        "colab_type": "code",
        "outputId": "bfc3b627-135e-436e-ac02-82838f4e7ea6",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 805
        }
      },
      "source": [
        "history = model.fit_generator(\n",
        "\ttrain_generator,\n",
        "\tsteps_per_epoch=8000//32,\n",
        "\tepochs=20,\n",
        "\tvalidation_data=val_generator,\n",
        "    validation_steps=2000//32,\n",
        "\tverbose=True)"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/tensorflow/python/ops/math_grad.py:1250: add_dispatch_support.<locals>.wrapper (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.\n",
            "Instructions for updating:\n",
            "Use tf.where in 2.0, which has the same broadcast rule as np.where\n",
            "Epoch 1/20\n",
            "250/250 [==============================] - 151s 605ms/step - loss: 0.8718 - acc: 0.5753 - val_loss: 4.9565 - val_acc: 0.5080\n",
            "Epoch 2/20\n",
            "250/250 [==============================] - 138s 551ms/step - loss: 0.7171 - acc: 0.6630 - val_loss: 0.9513 - val_acc: 0.5210\n",
            "Epoch 3/20\n",
            "250/250 [==============================] - 138s 554ms/step - loss: 0.6609 - acc: 0.6897 - val_loss: 0.6507 - val_acc: 0.7280\n",
            "Epoch 4/20\n",
            "250/250 [==============================] - 138s 553ms/step - loss: 0.6448 - acc: 0.7126 - val_loss: 0.6419 - val_acc: 0.6265\n",
            "Epoch 5/20\n",
            "250/250 [==============================] - 138s 553ms/step - loss: 0.6458 - acc: 0.6949 - val_loss: 0.6266 - val_acc: 0.7055\n",
            "Epoch 6/20\n",
            "250/250 [==============================] - 138s 553ms/step - loss: 0.5991 - acc: 0.7235 - val_loss: 0.7431 - val_acc: 0.6520\n",
            "Epoch 7/20\n",
            "250/250 [==============================] - 139s 554ms/step - loss: 0.5947 - acc: 0.7204 - val_loss: 0.5600 - val_acc: 0.7280\n",
            "Epoch 8/20\n",
            "250/250 [==============================] - 138s 551ms/step - loss: 0.5397 - acc: 0.7424 - val_loss: 0.6106 - val_acc: 0.6660\n",
            "Epoch 9/20\n",
            "250/250 [==============================] - 138s 552ms/step - loss: 0.5087 - acc: 0.7637 - val_loss: 1.6521 - val_acc: 0.5760\n",
            "Epoch 10/20\n",
            "250/250 [==============================] - 138s 551ms/step - loss: 0.4865 - acc: 0.7716 - val_loss: 0.6819 - val_acc: 0.7270\n",
            "Epoch 11/20\n",
            "250/250 [==============================] - 139s 555ms/step - loss: 0.4652 - acc: 0.7877 - val_loss: 0.5327 - val_acc: 0.7535\n",
            "Epoch 12/20\n",
            "250/250 [==============================] - 138s 552ms/step - loss: 0.4501 - acc: 0.7989 - val_loss: 1.9961 - val_acc: 0.5300\n",
            "Epoch 13/20\n",
            "250/250 [==============================] - 138s 552ms/step - loss: 0.4375 - acc: 0.8041 - val_loss: 0.8243 - val_acc: 0.6945\n",
            "Epoch 14/20\n",
            "250/250 [==============================] - 138s 554ms/step - loss: 0.4285 - acc: 0.8086 - val_loss: 0.8058 - val_acc: 0.7455\n",
            "Epoch 15/20\n",
            "250/250 [==============================] - 139s 554ms/step - loss: 0.4157 - acc: 0.8098 - val_loss: 0.5384 - val_acc: 0.7815\n",
            "Epoch 16/20\n",
            "250/250 [==============================] - 138s 552ms/step - loss: 0.4047 - acc: 0.8212 - val_loss: 2.3247 - val_acc: 0.5775\n",
            "Epoch 17/20\n",
            "250/250 [==============================] - 138s 553ms/step - loss: 0.3862 - acc: 0.8270 - val_loss: 0.6734 - val_acc: 0.7070\n",
            "Epoch 18/20\n",
            "250/250 [==============================] - 138s 553ms/step - loss: 0.3744 - acc: 0.8367 - val_loss: 0.4145 - val_acc: 0.8145\n",
            "Epoch 19/20\n",
            "250/250 [==============================] - 137s 549ms/step - loss: 0.3661 - acc: 0.8384 - val_loss: 0.9004 - val_acc: 0.7290\n",
            "Epoch 20/20\n",
            "250/250 [==============================] - 138s 551ms/step - loss: 0.3704 - acc: 0.8414 - val_loss: 1.1758 - val_acc: 0.6295\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MI3d6ONirH6h",
        "colab_type": "code",
        "outputId": "2c7ff91a-14dc-46c3-bd45-d2e3d00d7dab",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 313
        }
      },
      "source": [
        "acc = history.history['acc']\n",
        "val_acc = history.history['val_acc']\n",
        "epochs = range(len(acc))\n",
        "\n",
        "plt.plot(acc, label='training accuracy')\n",
        "plt.plot(val_acc, label='validation accuracy')\n",
        "plt.title('Accuracy curve')\n",
        "plt.xlabel('epochs')\n",
        "plt.ylabel('accuracy')\n",
        "plt.legend()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7f33dee76e48>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 16
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYsAAAEWCAYAAACXGLsWAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAIABJREFUeJzsnXd4VFX6xz8nPYE0ktBDQkJJIPTQ\nBCGAKCigoIhYcQV/YtddV9y1rWXX3XVRd22LBRVFxQq6iIKCivROEgIhECAJJCEkIYHUmfP748yE\nIaTMTOZOCffzPPPMzC3nnkmZ9563fF8hpURHR0dHR6cpvFw9AR0dHR0d90c3Fjo6Ojo6zaIbCx0d\nHR2dZtGNhY6Ojo5Os+jGQkdHR0enWXRjoaOjo6PTLLqx0NHR0dFpFt1Y6HgsQoh1QohiIYS/q+ei\no9Pa0Y2FjkcihIgFLgUkMM3J1/Zx5vVaiqfNV8c90Y2FjqdyK7AJeA+4zXKHECJQCPEvIcQRIUSp\nEGK9ECLQtG+0EGKDEKJECHFMCDHHtH2dEGKuxRhzhBDrLd5LIcQ9QohMINO07RXTGKeFENuFEJda\nHO8thPiTECJLCFFm2h8thHhNCPGvevNdIYR4qKEPKYToK4RYLYQ4JYTIF0L8ybT9PSHEcxbHpQgh\ncizeZwshHhVC7AHOmF5/Xm/sV4QQ/za9DhVCvCOEOC6EyBVCPCeE8Lbi96BzkaAbCx1P5VbgI9Pj\nCiFEB4t9LwJDgEuAdsAfAaMQIgb4DvgPEAUMBHbZcM1rgOFAH9P7raYx2gFLgc+EEAGmfQ8Ds4Er\ngRDgd8BZ4H1gthDCC0AIEQlcZjr/PIQQwcAaYBXQGegB/GjDfGcDVwFhwCfAlaYxMRmC6y2u+x5Q\na7rGIOByYC46OiZ0Y6HjcQghRgMxwDIp5XYgC7jRtM8L9cX8gJQyV0ppkFJukFJWmY5ZI6X8WEpZ\nI6UsklLaYiz+JqU8JaWsAJBSfmgao1ZK+S/AH+htOnYu8LiUcr9U7DYduwUoBSaYjrsBWCelzG/g\nelOAE1LKf0kpK6WUZVLKzTbM999SymNSygop5RFgBzDdtG88cFZKuclkaK8EHpRSnpFSFgAvmeam\nowPoxkLHM7kN+EFKedL0finnXFGRQADKgNQnupHt1nLM8o0Q4g9CiH0mV1cJEGq6fnPXeh+42fT6\nZmBJI8c5dL6on9Ns0+sbObeqiAF8geMm91wJ8F+gfQuurdPK0ANfOh6FKfZwPeAthDhh2uwPhAkh\nBgB7gUogHthd7/RjwLBGhj4DBFm879jAMXUSzab4xB9RK4Q0KaVRCFEMCItrxQOpDYzzIZBqmm8i\n8HUjczpG43f3Ns3XxGfAv4QQXVErjJEW16kCIqWUtY1cT+ciR19Z6Hga1wAGVNxgoOmRCPwK3Cql\nNALvAguFEJ1NgeaRpvTaj4DLhBDXCyF8hBARQoiBpnF3ATOEEEFCiB7AHc3MIxjl4y8EfIQQT6Ji\nE2beBp4VQvQUiv5CiAgAKWUOKt6xBPjC7NZqgG+BTkKIB4UQ/kKIYCHEcIv5XimEaCeE6Ag82NwP\nTkpZCKwDFgOHpZT7TNuPAz+gDEmIEMJLCBEvhBjb3Jg6Fw+6sdDxNG4DFkspj0opT5gfwKvATaY0\n0T+gVhhbgVPA3wEvKeVRlG/+96btu4ABpnFfAqqBfJSb6KNm5vE9KvB8ADiCWs1Yun0WAstQX8Kn\ngXeAQIv97wP9aNwFhZSyDJgITAVOoLKwxpl2L0GtnLJN1/i0mfmaWUrDAfVbAT8gHSgGPgc6WTmm\nzkWA0Jsf6eg4HyHEGJQ7Kkbq/4Q6HoC+stDRcTJCCF/gAeBt3VDoeAq6sdDRcSJCiESgBOXiednF\n09HRsRrdDaWjo6Oj0yz6ykJHR0dHp1laTZ1FZGSkjI2NdfU0dHR0dDyK7du3n5RSRjV3XKsxFrGx\nsWzbts3V09DR0dHxKIQQR6w5TlM3lBBikhBivxDioBBiQQP7uwkh1gohdgoh9gghrjRtjxVCVAgh\ndpkeb2o5Tx0dHR2dptFsZWFStXwNVVSUA2wVQqyQUqZbHPY4SgzuDSFEH2AlEGvalyWlHIiOjo6O\njsvRcmUxDDgopTwkpaxGSSRfXe8YyTmJhFAgT8P56Ojo6OjYiZYxiy6cL3+Qg+oFYMnTwA9CiPuA\nNigZAjPdhRA7UVIJj0spf61/ASHEncCdAN26dbtgAjU1NeTk5FBZWdmCj6HTmggICKBr1674+vq6\neio6Oh6FqwPcs4H3pJT/EkKMBJYIIZKA40A3KWWREGII8LUQoq+U8rTlyVLKRcAigOTk5AsKRnJy\ncggODiY2NhYhRP3dOhcZUkqKiorIycmhe/furp6Ojo5HoaUbKhelx2+mq2mbJXegxNaQUm5E9SGI\nlFJWSSmLTNvNzW162TqByspKIiIidEOhA4AQgoiICH2lqaNjB1oai61ATyFEdyGEH0qXf0W9Y45i\n6hhmkkEIAAqFEFHm/r9CiDigJ3DInknohkLHEv3vQUfHPjRzQ0kpa4UQ96KknL2Bd6WUaUKIZ4Bt\nUsoVKKnot0zN6iUwR0opTYqczwghagAjcJeU8pRWc9XR0dHxFGoMRvJPV3K81PQoqaBtgA83DY/R\n9LqaxiyklCtR6bCW2560eJ0OjGrgvC+AL7ScmzMoKSlh6dKl3H333Tafe+WVV7J06VLCwsIaPebJ\nJ59kzJgxXHbZZY0eo6Oj4zlU1ypDcOJ0JXklFZwwG4RS9TqvtJKT5VXUl/Qb1C3Ms43FxU5JSQmv\nv/56g8aitrYWH5/Gf/wrV65sdJ+ZZ555pkXzcwXNfW4dHU9HSklljZHTlTWUVdZwurKW0xU1lFXW\nmrade2/eX1Re1aghaOvvQ6fQADqGBpDQMYSOoQF0Cg2gU1igeg4NIDhA++w+/b9WQxYsWEBWVhYD\nBw5k4sSJXHXVVTzxxBOEh4eTkZHBgQMHuOaaazh27BiVlZU88MAD3HnnncA5+ZLy8nImT57M6NGj\n2bBhA126dGH58uUEBgYyZ84cpkyZwnXXXUdsbCy33XYb33zzDTU1NXz22WckJCRQWFjIjTfeSF5e\nHiNHjmT16tVs376dyMjI8+Y6f/58tm7dSkVFBddddx1/+ctfANi6dSsPPPAAZ86cwd/fnx9//JGg\noCAeffRRVq1ahZeXF/PmzeO+++6rm3NkZCTbtm3jD3/4A+vWrePpp58mKyuLQ4cO0a1bN/72t79x\nyy23cObMGQBeffVVLrnkEgD+/ve/8+GHH+Ll5cXkyZOZN28eM2fOZMeOHQBkZmYya9asuvc6Oq6i\n1mBkQ1YRX+/KJTO//DxDUGtsWs3b20sQEuBDcIAvwQE+tGvjV2cIOocF0DHUuYbAGi4aY/GXb9JI\nzzvd/IE20KdzCE9N7dvo/hdeeIHU1FR27doFwLp169ixYwepqal1qZvvvvsu7dq1o6KigqFDh3Lt\ntdcSERFx3jiZmZl8/PHHvPXWW1x//fV88cUX3HzzzRdcLzIykh07dvD666/z4osv8vbbb/OXv/yF\n8ePH89hjj7Fq1SreeeedBuf6/PPP065dOwwGAxMmTGDPnj0kJCQwa9YsPv30U4YOHcrp06cJDAxk\n0aJFZGdns2vXLnx8fDh1qvlwUnp6OuvXrycwMJCzZ8+yevVqAgICyMzMZPbs2Wzbto3vvvuO5cuX\ns3nzZoKCgjh16hTt2rUjNDSUXbt2MXDgQBYvXsztt9/e7PV0dLRASkla3mm+2pnLit15FJZVERLg\nw5CYcOKi2hAc4ENIgC/BAb6EBJ4zBiEBvnXGISTQh0Bfb49LtrhojIW7MGzYsPNy/P/973/z1Vdf\nAXDs2DEyMzMvMBbdu3dn4EClfDJkyBCys7MbHHvGjBl1x3z55ZcArF+/vm78SZMmER4e3uC5y5Yt\nY9GiRdTW1nL8+HHS09MRQtCpUyeGDh0KQEiIKrZfs2YNd911V507qV27ds1+7mnTphEYqFpQ19TU\ncO+997Jr1y68vb05cOBA3bi33347QUFB5407d+5cFi9ezMKFC/n000/ZsmVLs9fT0XEkuSUVfL0z\nl6935pJZUI6vt2B8QnumD+rCuIT2+Pt4u3qKmnPRGIumVgDOpE2bNnWv161bx5o1a9i4cSNBQUGk\npKQ0WAPg7+9f99rb25uKiooGxzYf5+3tTW1trdVzOnz4MC+++CJbt24lPDycOXPm2FWL4OPjg9Fo\nBLjgfMvP/dJLL9GhQwd2796N0WgkICCgyXGvvfbauhXSkCFDLjCmOjpaUFpRw6rU43y5I5fNh9Xq\neWhsOM9PT+Kqfp0IC/Jz8Qydi978SEOCg4MpKytrdH9paSnh4eEEBQWRkZHBpk2bHD6HUaNGsWzZ\nMgB++OEHiouLLzjm9OnTtGnThtDQUPLz8/nuu+8A6N27N8ePH2fr1q0AlJWVUVtby8SJE/nvf/9b\nZ5DMbqjY2Fi2b98OwBdfNJ7MVlpaSqdOnfDy8mLJkiUYDAYAJk6cyOLFizl79ux54wYEBHDFFVcw\nf/583QWloynVtUZ+SDvB3R9tZ+jza3j0i70UllXx+4m9+OWRcXx21yXcNDzmojMUoBsLTYmIiGDU\nqFEkJSXxyCOPXLB/0qRJ1NbWkpiYyIIFCxgxYoTD5/DUU0/xww8/kJSUxGeffUbHjh0JDg4+75gB\nAwYwaNAgEhISuPHGGxk1SmUz+/n58emnn3LfffcxYMAAJk6cSGVlJXPnzqVbt27079+fAQMGsHTp\n0rprPfDAAyQnJ+Pt3fiy/O677+b9999nwIABZGRk1K06Jk2axLRp00hOTmbgwIG8+OKLdefcdNNN\neHl5cfnllzv6R6RzkSOlZPuRYh7/ei/D/rqGO5dsZ/OhU9w4rBvL7xnFj78fy30TetItIsjVU3Up\nraYHd3Jysqzf/Gjfvn0kJia6aEbuQVVVFd7e3vj4+LBx40bmz59fF3D3JF588UVKS0t59tlnWzyW\n/ndxcXO6soY9x0rZnVPCzqMl7DpWwsnyKvx9vLi8b0dmDOrC6J6R+HpfHPfSQojtUsrk5o67aGIW\nFytHjx7l+uuvx2g04ufnx1tvveXqKdnM9OnTycrK4qeffnL1VHQ8jOpaI/tPlLHrWDG7jpWy61gx\nWYVn6vbHRbVhTM9IRsZHMCmpo9ukqbojurFo5fTs2ZOdO3e6ehotwpzNpaPTFFJKjp46y65jarWw\n+1gJqXmnqa5VSReRbf0YGB3GNQO7MLBbGP27hBEapBsHa9GNhY6OjsdSeraGz3fk8GtmIbuPlVB8\ntgaAAF8v+nUJ5baRMQyIDmNgdBhdwgI9rrbBndCNhY6Ojsex/0QZ723I5uuduVTUGOjZvi0T+3Rg\nYHQ4A6JD6d0hGJ+LJObgLHRjoaOj4xHUGoys2ZfPexuy2XToFP4+XlwzsAu3XhJD386hrp5eq0c3\nFjo6Om7NqTPVfLL1KB9uPEJeaSVdwgJZMDmBWcnRhLe5+OodXIW+TnMz2rZtC0BeXh7XXXddg8ek\npKRQP024Pi+//HJdcRsoyfOSkhLHTVRHR2NSc0t55LPdjPjbj/xj1X5iI9vw31uG8Msfx3HX2Hj3\nMRRZP0HWWlfPQnP0lYWb0rlzZz7//HO7z3/55Ze5+eab63SWrJE8dyeklEgp8fLS72cuJmoMRlal\nnuD9DdlsO1JMoK83M4d05bZLYunVIbj5AVzBtw+B0QAP7oVWHEDX/xM1ZMGCBbz22mt1759++mle\nfPFFysvLmTBhAoMHD6Zfv34sX778gnOzs7NJSkoCoKKightuuIHExESmT59+njbU/PnzSU5Opm/f\nvjz11FOAEifMy8tj3LhxjBs3DlBSHCdPngRg4cKFJCUlkZSUxMsvv1x3vcTERObNm0ffvn25/PLL\nG9Sg+uabbxg+fDiDBg3isssuIz8/H4Dy8nJuv/12+vXrR//+/evkPlatWsXgwYMZMGAAEyZMOO/n\nYCYpKYns7Gyys7Pp3bs3t956K0lJSRw7dqzBzwdKOv2SSy5hwIABDBs2jLKyMsaMGXNeweHo0aPZ\nvXu31b8vHddRWFbFK2syGfXCT9z38U4Kyqp4/KpENv1pAs9P7+e+hqLkGBRnQ+kxON66/9YunpXF\ndwvgxF7HjtmxH0x+odHds2bN4sEHH+See+4BlLLr999/T0BAAF999RUhISGcPHmSESNGMG3atEbT\n+t544w2CgoLYt28fe/bsYfDgwXX7GpIWv//++1m4cCFr1669oG/F9u3bWbx4MZs3b0ZKyfDhwxk7\ndizh4eFWSaGPHj2aTZs2IYTg7bff5h//+Af/+te/ePbZZwkNDWXvXvUzLi4uprCwkHnz5vHLL7/Q\nvXt3q6TMMzMzef/99+ukT2yRTr/jjjt47733ePnllzlw4ACVlZUMGDCg2WvqtIyKagPFZ6s5W22g\notrA2epaKmrMrw0NvK7lbLWBszUGKqsNlFfVsuNoMTUGyZheUbxwbQwpvdrj5eUBd+nZ68+9zvgW\nOg903Vw05uIxFi5g0KBBFBQUkJeXR2FhIeHh4URHR1NTU8Of/vQnfvnlF7y8vMjNzSU/P5+OHTs2\nOM4vv/zC/fffD0D//v3p379/3b6GpMUt99dn/fr1TJ8+vU6PacaMGfz6669MmzbNKin0nJwcZs2a\nxfHjx6murq6TW1+zZg2ffPJJ3XHh4eF88803jBkzpu4Ya6TMY2JiztPIskU6febMmTz77LP885//\n5N1332XOnDnNXk/HeiqqDRwsKOdAfhmZBeVk5pdxoKCMnOKKC7q7NYaftxeBft4E+noT5Odd9/qm\n4THcMjKG+Ki22n4IR5O9HgLDISoR9n0L4x939Yw04+IxFk2sALRk5syZfP7555w4cYJZs2YB8NFH\nH1FYWMj27dvx9fUlNjbWLklwR0mLm7FGCv2+++7j4YcfZtq0aXVd8GzFUsoczpczt5Qyt/XzBQUF\nMXHiRJYvX86yZcvqFHB1bMMao+DrLYiLbEv/rmFcNziajqH+BPh6E+SnGvsE+pmMQT2j0OpqH7J/\nhZhREDsaVi2AoiyIiHf1rDTh4jEWLmLWrFnMmzePkydP8vPPPwNKort9+/b4+vqydu1ajhw50uQY\nY8aMYenSpYwfP57U1FT27NkDNCwtnpKSApyTR6/vhrr00kuZM2cOCxYsQErJV199xZIlS6z+PKWl\npXTp0gWA999/v277xIkTee211+piIMXFxYwYMYK7776bw4cP17mh2rVrR2xsLN9++y0AO3bs4PDh\nww1eq7HPZymdPnToUMrKyggMDMTHx4e5c+cydepULr300kYbPekoDEbJocJy0vJOk3GizCqj0KtD\nW3p2aEtMRJuLRmivUUqOQskRGHkP9J6sjMW+b2D0g66emSZoaiyEEJOAVwBv4G0p5Qv19ncD3gfC\nTMcskFKuNO17DLgDMAD3Sym/13KuWtG3b1/Kysro0qULnTp1ApTc9tSpU+nXrx/JyckkJCQ0OYa5\nj0NiYiKJiYkMGTIEOF9aPDo6uk5aHODOO+9k0qRJdO7cmbVrz6X1DR48mDlz5jBs2DBAdaEbNGhQ\no9336vP0008zc+ZMwsPDGT9+fN0X/eOPP84999xDUlIS3t7ePPXUU8yYMYNFixYxY8YMjEYj7du3\nZ/Xq1Vx77bV88MEH9O3bl+HDh9OrV68Gr9XY57OUTq+oqCAwMJA1a9bQtm1bhgwZQkhIiN73oh7V\ntUYO5JeRlldKau5p0vJK2Xe8jIoa1UtENwp2YI5XxI6GsG7QaYCKW7RSY6GZRLkQwhs4AEwEcoCt\nwGwpZbrFMYuAnVLKN4QQfYCVUspY0+uPgWFAZ2AN0EtKaWjserpEuQ6o+pSUlBQyMjIaTbtt7X8X\nFdUG0o+fJt1kGFLzSjmQX0aNQf2vt/X3oU/nEJI6h9K3cwhJXUKJi9KNgs18fTfs/w4eyQIvL/j5\nn7D2OXg4A0I6uXp2VuMOEuXDgINSykOmCX0CXA2kWxwjgRDT61Agz/T6auATKWUVcFgIcdA03kYN\n56vj4XzwwQf8+c9/ZuHChRdNfYaUkr25pWw5fIq0vNOk5paSVViO0XQPGB7kS1KXUO4YHUdSF2Ug\nurUL8oxMI3fn8K9qVWH+W0ucoozF/pUw9A7Xzk0DtDQWXYBjFu9zgOH1jnka+EEIcR/QBrjM4lzL\nHqM5pm3nIYS4E7gToFu3bg6ZtI7ncuutt3Lrrbe6ehpO4Wx1LSt25bFk0xHS8k4D0DEkgKQuIVzZ\nr1PdiqFTaICutKoFxUeg9Chcct+5bVEJ0C5euaJ0Y+FwZgPvSSn/JYQYCSwRQiRZe7KUchGwCJQb\nqpFj9H8WnTo8vTPkwYIyPtx0lC925FBWWUtCx2CeuyaJK/p2JCrYv/kBdByDZbzCjBBqdbHxNago\ngcAw18xNI7Q0FrlAtMX7rqZtltwBTAKQUm4UQgQAkVae2ywBAQEUFRURERGhGwwdpJQUFRUREBDg\n6qnYRI3ByA9p+SzZpNRW/by9uLJfR24eEcOQmHD9b9sVZP8KQRHQvl7sK2EK/PYKZP4A/a93zdw0\nQktjsRXoKYTojvqivwG4sd4xR4EJwHtCiEQgACgEVgBLhRALUQHunsAWWyfQtWtXcnJyKCwstP9T\n6LQqAgIC6Nq1q6unYRXHSyv4ePNRPt56jMKyKrqGB/LopASuT+5KRFt9FeEypFQri9jRF2pBdUmG\nth1VCq1uLKxDSlkrhLgX+B6VFvuulDJNCPEMsE1KuQL4PfCWEOIhVLB7jlR+gjQhxDJUMLwWuKep\nTKjG8PX1rase1tHxBIxGyfqDJ/lw0xHW7MtHAuN7t+fmETGM6RWFtx6Ydj0lR5QW1KgHLtzn5QUJ\nV8LuT6CmAnwDnT8/jdA0ZmGqmVhZb9uTFq/TgVH1zzPtex54Xsv56ei4C8Vnqvl8ew4fbT5CdtFZ\nItr4cdfYeGYP60Z0uyBXT0/HksO/qufYSxvenzAFtr2rZMsTrnTevDTG1QFuHZ2LmsKyKhau3s+X\nO3KpqjUyLLYdD03sxaSkjvj7eLt6ejoNkb0egiIhqnfD+2MvBf9QlRWlGwsdHZ2WUGswsmTTERau\nPkBljYHrk6O5ZWQMCR1Dmj9Zx3U0Fa8w4+MHva5QBXuGWvBuHV+zreNT6Oh4EJsPFfHUijQyTpRx\nac9Inp7W1/PUVi9Wig/D6Rzo/lDTxyVOgb3L4OgG6D7GOXPTGN1Y6Og4iYLTlfx15T6+3pVHl7BA\n3rx5MFf07ainvnoSdfUVjcQrzPS4DHwClGy5bix0dHSsocZg5P0N2by8JpPqWiP3je/B3Sk9CPTT\nYxLUVsGSGdB1CEx8xtWzaZ7s9dCmPUQ2LH5Zh18biB8PGf+DyX9vFe1WdWOho6MhG7OKeGpFKgfy\nyxnXO4qnpvYlNrJN8ydeLPzwBBxZDzVnXT2T5pHynB6UNV/+CVOUTtTxXdB5kPbz0xjdWOjoaMCJ\n0kqeX7mPb3bn0TU8kLduTeayxPa6y8mS9OWw5b/g20b1sXZ3Th2CsrzzJT6aotckEF7KFaUbCx0d\nHUuqa428+9th/v1jJrVGyQMTejI/JZ4AX93ldB6nDsPye6HLENU46KfnoLIUAkJdPbPGyW6mvqI+\nbSJUF72Mb2HCE9rNy0noxkJHx0GszzzJUytSySo8w2WJHXhySh+6RegFdRdQWwWfzVGunOsWKzcN\nKCXXTo33j3c52euhbQeI7Gn9OQlTYNWjcPIgRPbQbm5O4OIQ/dfR0ZCc4rPc/dF2bn5nM7VGyeI5\nQ3n7tmTdUDTGD08oA3HNGxAeA+Gxars7u6Ksqa9oiISr1HPGN9rMy4noKwsdHRsxGlXDoZ8yCli3\nv4DdOaUE+Hrx+4m9mDcmTnc5NYU5TjHinnNfpGEx6tmdjUVRFpQdtz5eYSYsGjoNVHGL0c3UZrg5\nurHQ0bGC0ooafs0sZG1GIT8fKOBkeTVCwKDoMP5weS+mD+5Kl7DWIxqnCacOnYtTXPb0ue2BYRAQ\n5t7GwtZ4hSWJU1RM5nQehHR27LyciG4sdHQaQEpJZkE5P2UUsDajgG1HijEYJWFBvoztFcW43u0Z\n0yuKdm38XD1Vz8AyTjHzPSWJYUl4rJsbi/VKejzCjrhDwlRlLDL+B8PmOX5uTkI3Fjo6JiqqDWw8\ndNJkIArJLakAILFTCHeNjWN8QnsGRofrMuH28MPjcHw33PAxhDXQAjk8Fk7sdfq0rEJKtbKIvdS+\n4rqo3srIZHyrGwsdHU9FSsnXu3JZviuPjVlFVNUaCfLzZlSPSO4d34NxvdvTMdSzOuu5HWlfw5ZF\nMPLexlVYw2PVnbfRAF5uFvMpOgjl+bbHK8wIobKiNr4KFcUQGO7Y+TkJ3VjoXLRUVBtY8OUelu/K\nIyYiiBuHd2N8QnuGdW+ny4M7ilOHYMV9qoPchKcaPy48Fow1Kogc6madDM3xipZoPCVOhd9ehgM/\nwIBZjpmXk9GNhc5FydGis9y5ZBv788v4w+W9uDulB166e8mxnBenWHxhnMISy/RZdzMWh3+F4E7Q\nLs7+MToPVjGPjG881ljodRY6Fx3r9hcw9dX1HC+tZPGcodw7vqduKLTAHKe45s2G4xSWuGuthb31\nFfXx8lKpwgd/VO1WPRDdWOhcNBiNkld/yuT297bSKTSAb+4dTUrv9q6eVusk7avm4xSWhHYF4e1+\nxuJkJpwpsC9ltj6JU5RgYtZPLR/LBehuKJ2LgrLKGn6/bDc/pOdz9cDO/G1GP4L89D9/TSjKguVW\nxCks8fZVBsPdjEX2L+rZ3uC2JbGXKu2rfd+eK0j0IDT9bxFCTAJeAbyBt6WUL9Tb/xIwzvQ2CGgv\npQwz7TMA5ly6o1LKaVrOVaf1crCgjDuXbOdI0VmemNKH342KdS/116Ob1Belu/nq7aGmUsUpvLyb\nj1PUxx1rLbLXQ3DnlsUrzHj7KiXaA57ZblUzN5QQwht4DZgM9AFmCyH6WB4jpXxISjlQSjkQ+A/w\npcXuCvM+3VDo2Muq1ONc/epvnK6o4aO5w7ljdHf3MhQ7P4R3J8Hav7p6Jo7hh8fhxB6YbkWcoj7h\nMe5lLMzxiu521lc0RMIUlT5MGnhcAAAgAElEQVR75DfHjOdEtIxZDAMOSikPSSmrgU+Aq5s4fjbw\nsYbz0bmIMBgl/1iVwV0f7qBHh2C+uW80I+IiXD2t89n+Hiy/B5Du9SVpL2lfwda3VJyi92Tbzw+P\nhTOFUFXu8KnZReF+NR9HuKDM9Jig2q1mfOu4MZ2ElsaiC3DM4n2OadsFCCFigO6AZeQnQAixTQix\nSQhxTSPn3Wk6ZlthYaGj5q3j4RSfqWbO4i28vi6L2cOiWfZ/I+gU6ma6TVvegm8egJ5XQJ+roeRY\n8+e4M+Y4Rdeh5+s+2YI5I6rkiIMm1ULq9KAcaCz82kD8BFWAKKXjxnUC7pINdQPwuZTSYLEtRkqZ\nDNwIvCyEiK9/kpRykZQyWUqZHBUV5ay56rgxqbmlTH11PZsPneJvM/rxtxn93a/AbtObsPIP0Psq\nmLVESUGczlV+bE/EMk5x3bvKN28P7pY+m70eQrpCeHfHjps4Rf2+83Y4dlyN0dJY5ALRFu+7mrY1\nxA3Uc0FJKXNNz4eAdYDn9yXU0ZSvduZw7RsbqDVIPv2/EcweZqPP3Bls+I9qhpM41SSo5w+h0SAN\nqnrZE1n9pP1xCkvMX8ruYCwcVV/REL0mqTThfZ7litLSWGwFegohugsh/FAGYUX9g4QQCUA4sNFi\nW7gQwt/0OhIYBaRrOFcdD6bGYOTpFWk89OluBkSH8c19oxnUzQ31d35dqALAfaerDnHmTKEw0z1V\nqQe6ogoyVJxi6Dz74hSWBIaDf4jqmOdqCjPg7EnHuqDMBLWD2FHKFeVBaGYspJS1wL3A98A+YJmU\nMk0I8YwQwjK76QbgEynPc+AlAtuEELuBtcALUkrdWOhcwMnyKm56azPvbcjmd6O689Hc4UQF+7t6\nWhfy8z/gx79Av5kw4+3zXTWhprtxT4xb/PQs+LaBlMdaPpYQ7pMRlb1ePXd3QDFeQyRMgZP7VdGf\nh6Bpoq+UciWwst62J+u9f7qB8zYA/bScm47nk5pbyp0fbKPoTDUvzxrINYMazJ9wLVLCur/Bz3+H\nATfC1a9eqKpqrq8oPer8+bWEnG0qq2fcn6GNgzLNwmOh8IBjxmoJh39R7kFzFz9Hk3AVfPdH2PcN\nXPqwNtdwMO4S4NbRsYkVu/O47s0NSODzuy5xX0Px4zPKUAy6Ba5+rWH5bb8gCIr0rJWFlLDmaTXv\nEXc7btzwWJUNZTQ6bkxbMRpVHYQW8QozoV2h8yCPSqHVjYWOR2EwSv6+KoP7P95JUudQVtw7mn5d\nQ109rQuRElY/AesXQvLvYOq/lZhcY4RFe1bM4tBalVo65hHwb+u4ccNjobZS9Y9wFYUZcLbIMXpQ\nTZEwBXK3q3arHoBuLHQ8htOVNcz7YBtvmOonls4b4Z7xCSlh1WMq82nYnXDVwqYNBSiXh6esLKSE\nNX9RsZbk2x07dlisenZl3EKL+oqGSJyqnj0k0K0bCx2P4FBhOdNf+41fDhTy7NV9+ev0fvj5uOGf\nr9EIKx+BzW8o98zkf1jnygjrBqU5nlGolb4cju+CcX9Sqb+OxB1qLbJ/VYYwXKN4hZmo3hDRU8Ut\nPAA3/G/T0TmfdfsLuPq13zh1ppoldwznlpFuJgRoxmiE/z2kUkkvuR+u+Kv1Pu/QaKitgDMntZ1j\nSzHUwk/PQVQi9L/e8eOHRQPCdcbCaITs37TLgqpP4hSVeXX2lHOu1wJ0Y6HjtkgpWfRLFr97bytd\nwgJZce9oRsa7mb6TGaMBvrlP6T1d+nuY+IxtwVFPyYjavRSKMmHCE9r0yvbxh5AurjMWBelQcUp7\nF5SZhKmqIPPA9865XgvQjYU7cPIg1Fa7ehauoygLqs+ct6myxsDDy3bz15UZTErqyJd3X0J0uyAX\nTbAZjAb4+m6lIDt2AYx/wvYsGnNhnjvHLWoqYN0LSv+ptxUNjezFlVLl5voKZxmLzoOUBLoHZEXp\nxsLVVJ6GNy6BTa+5eiauobIU3hwNqxbUbTpRWsn1/93IVztzeXhiL167cbB7Nyr69iHY8wmMexzG\nPWZfumWoB1Rxb31baRpNeEq7lFI4lz7rCrJ/VbUVLZEtsQUvL+h1ORz6GQw1zrmmnejGwtUUHwZD\nFWSudvVMXEP6CtVqcvencOYk248UM/XV9WQVlLPoliHcP6Gne8YnzNRWwa6PYPBtMPYR+8cJDAe/\ntu67sqgsVXIl8RO09+eHxyqdLGf3qq6rr3BSvMJM/HioLlNptG6MG9+uXSSYdXCObVE6/o7MWXcC\nBqPkrV8PUXy2ms6hgXQOC6RTaABdwgIJC/Jt/ot+7zJoEwVnCtm7/GVmp42iU1gAH80dTq8Owc75\nEC3h5AEw1kL3MS0bRwi1unDXlcWGV5Uvf8KTzR/bUuqkyo+qjCFnUZCmGhM5ywVlpvsYEF6qN3e3\nEc69tg3oxsLVmJfbxho4uhF6TnTtfGzAaJQ89uUelm3LwddbUGM4P+0z0NebTmHKcHQKDaBzmDIm\nyqgE0NmrmIDDv2K89BGydv9C+/1LGBk7iVduGkZYkA3tOF1JvkmyrENSy8dy18K88gLY+JoSQOw8\nUPvrWabPOtNYODteYSYwHDoPhqy1Kh3ZTdGNhaspzga/YDBUw6F1HmMspJQ89799LNuWw/3je/Dg\nZb04eaaK4yWV5JVUkFdqeja9zjhRSGFZ1XljzPX+H4/7Sm7c3A3/8rG877eJd4fm4B3k5H/WlpCf\nCt5+qidFSwmNVitMd+PXf6mq6nGPO+d6rqq1OPyrunZYdLOHOpz4cernXFECgWHOv74V6MbC1RQf\ngXbd1d1F1lpXz8ZqXl6Tybu/Heb2UbE8NLEXQgjaBwfQPjiAAdEN/7FX1RrIL60it6SC46UVjP7p\nL+QYEmjbMYGr+o2FDV/hveVNGDBL2wCqI8lPg6gE8HbAv1JYNFSWQFUZ+LuJC674CGx9BwbdDJEO\nMIjW0CYSfIOcayzM8YrEKc67piXx4+GXf6oAu7my283QA9yupjhb3c3EpSifaZkLNXGs5O1fD/HK\nj5nMHNKVJ67qY3UA2t/Hm24RQYyMj2BG9Bnal++n65jbePu2oUwf3A2G3wV5O93z7rox8tMc44KC\ncxlR7hTkXveCqqcY+6jzrimE89Nn81OVoXZ2cNtM16EqwSHrp+aPdRG6sXAlRqMK4oXHKGMBShrZ\njfl4y1Ge+98+rurXiReu7Y+Xl50rgD3LVFAv6dpz2wbcAAGhsOl1x0xWa84UQfkJ6NDHMeOZ0zXd\nJW6Rnw67P4Zh8yDUyaq+zjYWztKDagxvX2Wo3Ni7oBsLV1Ker9Jmw2Kg0wAICFNxCxspPlONwai9\nptCK3Xn86au9pPSO4qVZA/G211BICXs/g+5jIbjDue1+bWDIHKWV4053141RkKaeO/R1zHh1Kws3\nqeL+6TnlDhvtgn4LZmPhLK2s7PWqrau5kt4VxI9TqfSnDrtuDk2gGwtXYr5zCu+ulvpxY5WxsOEf\n5IvtOQz/649Mf/03Mk6c1mSaAD/uy+fhT3cxNLYdb9w0pGUifse2qCywhrSFhs5Tz1vfsn98Z5Fv\nNhYOckO17aCC5e6wsji2Ffb/T2lcBbVz/vXDY1X9zZlC7a9lNJzrX+FK4ser50PuubrQjYUrMafN\nmtUt41LgdI6Sv2gGo1Hy4vf7+f1nu0nqEkJucQVT/7OeV9ZkUl3r2MYxG7JOMv+jHfTpHMI7tyUT\n6NdCTaC9y8AnQOn51ycsWgX4tr9/gQSI25GfqmpE2rZ3zHheXkoXydWrKnNjozZRMGK+a+ZQlxHl\nhEruE3tV0WFLa2VaSkQPCOnqtnEL3Vi4kuJsQJxzP8SlqOdm7iwqqg3c+/EOXl17kNnDovn0/0by\nw0NjmJzUiZfWHGDaq+tJzS11yBR3Hi1m7vvbiI0I4v3bhxEc4Nv8SU1hqIG0r6D3ZAgIafiYEfNV\nsHH3Jy27ltbkpznOBWXGHWotsn6EI+thzB9dVyTqzPRZc31FzCjtr9UUQihX1OFflLqvm2GVsRBC\nfCmEuEoIoRsXR1J8BII7gW+Aeh/eXQU5m4hbFJRVcsOijXyXeoI/X5nIX6f3w9fbi4i2/vx79iAW\n3TKEU2equfq13/jn9xlU1hjsnt6+46eZs3grUcH+fHjHcMLbOKBQLusn1YWsXxPy1tHDlcDa5jdd\n216zKYwGKNgH7R1sLEK7uXZlYTSqVrBh3VT8yFWYg/3OMhbt4pwfxG+I+HFqlZO309UzuQBrv/xf\nB24EMoUQLwghnFhW2YopOXLuDgrUnUVciioOauDOIj3vNNe8+hsH8sv5781DmDcm7oK01cv7dmT1\nQ2OZPqgLr63NYsp/1rPjaHHjczDUwFsTLriLP1RYzi3vbCHQ15sP7xhO+5AA+z+nJXuWqZqSHpc1\nfowQMHy+ktI45J5Lck4dVoVqWqwsyk8ozSlXkP41HN8N4/4MPi6sovcNVDdSWhsLKSFnC3Qbqe11\nrKV7CiDcMm5hlbGQUq6RUt4EDAaygTVCiA1CiNuFEI36JYQQk4QQ+4UQB4UQCxrY/5IQYpfpcUAI\nUWKx7zYhRKbpcZvtH80DKM6+sBtXXApUlapOZBb8uC+fmW9uwCjhs7tGcnnfjo0OGxrky4szB/De\n7UM5W1XLtW9s4Llv06mobmCVcWgd5G6D1U/WCbflllRw89ubkVLy4dzhjpMGryqH/SuVbERzX0R9\np6uA76Y3HHNtR5Ofqp4dbSzq1GdzHDuuNRhqVAZU+z7Qb6bzr18fZ6TPlhxVK90ug7W9jrW0iVCZ\nkW4Yt7DarSSEiADmAHOBncArKOPRoFyqEMIbeA2YDPQBZgshzktIl1I+JKUcKKUcCPwH+NJ0bjvg\nKWA4MAx4SggRbtMnc3dqq1Sj9rB6xqL7WPVsurOQUvLO+sPM+2AbcVFtWX7vKJK6hFp1iZTe7fn+\noTHcOKwbb68/zORXfmHzoaLzD0r9QmXglOfDtsUUllVx89ubKauq5f3fDaNHewf6rDP+pzJcmnJB\nmfHxU5lRB9dA4QHHzcFR5KepOpGoBMeOW9cEyQWuqF0fwaks1Y9Di8ZGthIWo72xyNuhnju7ibEA\nlRWVs1W1L3AjrI1ZfAX8CgQBU6WU06SUn0op7wMa+zYZBhyUUh6SUlYDnwBXN3GZ2cDHptdXAKul\nlKeklMUogzTJmrl6DCXHAHm+GwqU1EHH/nDoZ2oMRv78dSrPfpvOxD4d+PT/RtDBRndQcIAvz0/v\nx9K5wzFIyaxFm3hyeSrlVbVQUwn7vlVf3t3HYFz/EnPf/pkTpZW8d/tQq42S1exdpnzy0cOtOz75\ndvD2V7ELdyM/TfVP9nWQe86Mq5og1TU2GqaSD9yB8FjVP0NLl1zudnWz5Kj0Z0cQP04pGZsD726C\ntSuLf0sp+0gp/yalPG65Q0qZ3Mg5XQDLv/gc07YLEELEAN0B89rLqnOFEHcKIbYJIbYVFjohH9uR\nlGSr54aawselII9tZv67v7B081HuGhvPGzcNaVEDoEt6RPL9g2O4fVQsSzYd4YqXfiH91y+Vjn7S\nDM6OehSvMwWMPLWct25NZkiMg3PrywtVdWq/61SKqDW0iYT+M1UVcUUTcRdXkJ/quMptS0K6AsL5\nK4stb6keEpdp3NjIFsJjAamt4czdCR37uTY+U5/o4Uoby83iFtYaiz5CiDp1OCFEuBDibgfO4wbg\ncymlTak7UspFUspkKWVyVFSUA6fjBMz54/VXFkBB1EiEoRpD9kb+cV1/FkxOsF9Ww4IgPx+emtqX\nz/5vJP4+XmStfZ9y71AKo0Ywd60P641JPBS0ktExgS2+1gWkfal6DTdUiNcUw+cr19WODxw/J3up\nKlPJCY6OV4D60gru6NyVRUWJUjztcZnrC9Ms0Tp91mhQWUfu5IIC1Yc8ZpTbxS2sNRbzpJR1wWeT\na2heM+fkApZav11N2xriBs65oGw91zMpzlYulrbnB6q3Zp/i6hUGqvHh+QEnuT7Z8XLJybHtWDl/\nMFf47mJ5dTIj//4zG7KKMIxdgH/VKdU+09HsWQYd+kH7RNvO65ikNHO2vOU+uecF+9SzVq4LZzdB\n2vAfVdfijMZGtlBnLDSSvzh5AGrOQJch2ozfEuLHQ9FB95F+wXpj4S0scjRNwevm1m1bgZ5CiO5C\nCD+UQVhR/yAhRAIQDmy02Pw9cLlpBRMOXG7a1nooOaL80xYumS935HDTW5sJaBOMocswOhdt1uzy\nAYdX42esZPjUeSTHhvP3a/sxdsJU1Tbzt1dU5pKjKMpSGVf97cywGTFffXm6S1N7rTKhzDi7MG/X\nUuh9pcrCcSfadlCV/lqtLHJNwW13yYSyJH6cenYjYUFrjcUq4FMhxAQhxATUKmBVUydIKWuBe1Ff\n8vuAZVLKNCHEM0KIaRaH3gB8IuU5QSQp5SngWZTB2Qo8Y9rWejBLk3NOuuPhZbsZHBPGV3dfQmDv\nCZC/V/n6tSD1S2jbkR7Jl/PJnSOZNdRUBDXuTyqVcMsix11r7+eAgKTr7Du/1yT1s3KXQHd+GviH\nnEtzdTSh0VCa65yCxIoSKMuD6GHaX8tWvLxURlSJRpIfudtV47GIntqM3xKiElSdiRvFLaw1Fo8C\na4H5psePwB+bO0lKuVJK2UtKGS+lfN607Ukp5QqLY56WUl5QgyGlfFdK2cP0WGzlPD2H4iMQFkN1\nrZH7P9nJq2sPcn1yVz743XDVUjTOdGdx+GfHX7vyNGSuhr7XXJgi2TUZekyEDf9WvvmWIqXKgood\nbX+FrJc3DPs/1XbWfDfoSvLTVC2CVoHgsGjVZrf8hDbjW3LSlJbs6BRgR6FlrUXeDtUm1tqEC2ci\nhPoOOLROxVbcAGuL8oxSyjeklNeZHv+1NRitY0FFCVSWUBvajbs/2s63e47z6KQE/n5t/3Nqrp0H\nqt4OdkiWN8v+lUoa3bKXhCUpj6nso83/bfm18nYq32tLi7wG3azuAl29upBS9XnQygUFKr0YnBPk\nLsxQz87sdW0L4bHqxsrRUuW1VXAi1T1dUGbix6v/w+O7XT0TwPo6i55CiM+FEOlCiEPmh9aTa7WY\nltX/3WNgzb4Cnr26L/NT4s+X7vDyViqYNkqWW0XqF8rV0XVow/u7DlGunw3/aXlh0N7PVB57n6ZK\nbKwgIAQG3aTcZ2VOuONujNIcVWGvpbEw11o4I25RuB98Ai8sDnUXwmOh6rTjU6dPpKrVmzsGt83E\npahnN8mKsnb9tRh4A6gFxgEfAB9qNanWTvVJld2xMsef565J4paRsQ0fGJeivjBOOdAunz2l/vj6\nTm/ajZKyQGXItORO3mhQhqnn5Y5pQj/sTlWstPWdlo9lL47uYdEQzmyCVJgBkT3do2K7Icx1SI7O\niHLHyu36tI1SNSBaeBfswFpjESil/BEQUsojUsqngau0m1brpbLGwBc//gbAnKtSuHlEE3d05riF\nI4Nc+75RX7iNuaDMdB6kMmQ2vqrcZvZw+GclI2JrbUVjRMSr6uJt76rqc1dgzoSyNQXYFvzbKrFF\nZ60s3DVeAdrVWuRuV/06XNkZzxrixsHRTY7NTrQTa41FlUmePFMIca8QYjqNy3zoNEJljYF5H2yj\ntugw1T7BzBzdzN1puzh1l+nIO4u0L9W41qRJpixQcsn2ri72fAb+odDzCvvOb4jhd8HZk5D6uePG\ntIX8NCWf3VgvDkcRGq19zKKqTBkkd41XwDn3mMONxQ7lgnKXavXGiB+v3GVHNrh6JlYbiwdQulD3\nA0OAm4HWqQSrERXVBu54fyvrD57kso4V+EXFNX+SEKrV6uFfHJMRUV6gxkq61rp/kk4DVDe7ja/Z\n7jOuqVCrmD5THauf1H2M6iGx6U3n9We2pCDdOTpCYd20X1m4eyYUqFVWmyjHGovK0+qzu7MLyky3\nkarWxA3iFs0aC1MB3iwpZbmUMkdKebuU8lop5SYnzK9VcLa6lt+9t5UNWUX887oBdDLmWx9QjDM1\nQ6knWW4X6ctBGqHvDOvPSXlMBRg3vm7btfZ/p3SnrFGYtQUhYMRdqgbF2UJrNZVwMlPb4LYZ88pC\nS4NYuF89u7OxAMenzx7fBUj3Dm6b8Q2AmEvcot6iWWNhSpF1I8EYz8JsKDYfLmLh9QO4blBnFbhs\nQBOqQeoky9e1fDKpX0JUom0CeB2TIHGa6itx1oa6yL2fqaIiLbSG+s2EoAjnp9Ge3K/0rZxhLMKi\nlRSFlgKKhRkqU83av0VX4WhjYa7V6TzIcWNqSdw49bsqda3ikbVuqJ1CiBVCiFuEEDPMD01n1go4\nU1XLnMVb2XL4FC/NGsj0QV1VoZWhqmG12YZoG6U0lVpqLEpz4egGSLLj15byGFSXK3eUNZw9pYr+\nkq7VJsvGNxCG3K76Y5zSSDeoIZyRCWXGGRlRhftV9bK3/WrGTiE8Vv39GmocM17udrWybxPhmPG0\nJn68enZxVpS1xiIAKALGA1NNjylaTao1UF5Vy5zFW9h+pJiXbxjE1QNN1ctmtdmwWOsHixurMiKq\nz9o/ofSv1bMtLigzHfqoau/Nb8KZouaPT/9aBeUclQXVEEPnKkPkSFmS5shPU/7jdlbEm1qKM2ot\nCjPcO7htJjxWregc1T0wb6dnuKDMdOgLbdq7PG5hbQX37Q08fqf15DyVssoabnt3CzuOlvDKDQOZ\nNqDzuZ3m5bQtS/+4cWCoVnIX9pL6hQpYR/aw7/yxC6D6DGz8T/PH7vkMInurJk5aEdJJ1YrsWOK8\njmL5acq/74yahLqVhUbGovqsunFx93gFODZ9trxAGWB3rtyujxBKWPDQOufohTWCtRXci4UQ79Z/\naD05T+S0yVDsPlbCf2YPYkr/zucfUHIEEOfuHK0hZiR4+dq/DC3OVktve1YVZtonKBfW5kVw5mTj\nx5UcVe6u/jO1T0scPl8F0Xct1fY6ZvLTnNdRLShCVVZrtbIoygSk56wswDHGok5p1oNWFqBuGM+e\nVIkdLsJaN9S3wP9Mjx+BEMD1VSJuxunKGm59Zwt7ckp59cbBXNmv04UHFR+BkM6qwYm1+LVR3bPs\nNRapX6rnvtPtO9/M2AVQW6FEBhtjr6n+oaVaUNbQdYhqA7r5Te3vuMoL4EyBc4LboAxtWLR2MQtP\nyYQClSjh5esYY5G3Q/VOdzc59uZwA8lya91QX1g8PgKuBxprp3pRUlpRwy3vbCEtr5TXbxrMpKSO\nDR9YnG2fDk98CpzYY13MoD5pXyodKGuD6o0R1UvJjG95q3Hp9L2fKcPmrAybEfOVFESmxu1O6oLb\nTjIWoG0TpMIM8PJxTvylpXh5q7oTh6wstisD6dem5WM5k+COSunYhXELe7V5ewLtHTkRT6b0bA23\nvLOZ9LxS3rhpCJf3bcRQgHJD2fOlba9k+clMOLG3eXkPaxn7KNRWwm8vX7jvRKoqWnPGqsJM4jTV\nJGf3J9pexxXGIkzDKu7C/dAu3r16TzeFI9JnpTRVbntQvMKS+PEtT3RpAdbGLMqEEKfND+AbVI+L\ni55ag5Fb391MxvEy3rx5CJf16dDEwVVwOs++u+5OA5V0hq3FOalfAgL6XGP7NRsisocqtNv6DpTl\nn79v7zJ1t9qS2IitePsohdyDP0JttXbXKUhXRqlNpHbXqE9oNFScUokFjsZTMqHMOMJYFGern6cn\nVG43RNw4lXZ/1DXSH9a6oYKllCEWj15Syi+0npwnsCX7FLtzSnluehITEpswFGC6S5T2uaG8faD7\npZC1zvqqXilVFlTMKJU95CjG/lFlZ1muLoxG2PuFasvq7Pz13pNVoPvIb9pdIz/VuasKUK4XcFzK\nqJnaKqVk7AnxCjPhsUoFuSVFinkeGtw2E3OJKqJ0UdzC2pXFdCFEqMX7MCGEg25VPZvV6fn4+Xhx\nVUPB7PqUZKtne/35cSlQetR6ueaCdFV1nNTCwHZ9IuJhwA1K/dXcW+LoBjido21tRWN0H6vqHw40\n2enXfgy1UJDhfGOhVfps0UEl++JpKws4V6dkD7k7wNvf+b9HR+EXBN1GuLexAJ6SUpaa30gpS4Cn\ntJmS5yClZHV6PqN7RNLG34oq2LoaCzsDzXEp6tnarKjUL0B4Q2ILGw81xJg/qIra9S+p93uWgW8b\ndZfvbPyClMHY/502WkqnstTy31lps2bqCvMcnBFV1x3Pw1YW0DJXVO4O1R/C29cRM3IN8eOhIM0l\nDcCsNRYNHefmGgHasz+/jJziCiY2FaewpPiIurNp20QAvCkiekBIF+uMhZQqXhE3VkmGOJp2cTBw\nNmxbrP6B07+GxCmuyzLpPUklD5hTQh2JuYeFs+9IgzupGJCjVxYFGSp9NMLOAk1XYL7BKrFzZWGo\nVQKCnuqCMlPX42ad0y9trbHYJoRYKISINz0WAtu1nJgnsDpNBXgnJFiZGFacrfzQ9jaINzdxt0ay\nPG+ncldpGWwe84iSYVg6SynjOlph1hZ6TVLPB75z/Nj56WqFFtnL8WM3hZe3qslxdPpsYQaEd3es\ndLzWBIRCYDv7VxYn90PNWc/NhDLTsb8q2HSBK8rab637gGrgU+AToBK4p7mThBCThBD7hRAHhRAL\nGjnmelNv7zQhxFKL7QYhxC7TY4WV83Qqq/flMzA6jPYhVv7T2Zs2a0lcigryndjT9HGpX6hCpkQN\nJbzCY2HgTerLp03UOTeZKwjprAqt9msQt8hPU4bClkJKRxHazfErC3fvjtcYLcmI8tTK7fp4ean/\ns0Nrnd7PxdpsqDNSygVSymQp5VAp5Z+klE3m85n6YLwGTAb6ALOFEH3qHdMTeAwYJaXsCzxosbtC\nSjnQ9Jhmy4dyBidKK9mTU2q9CwqUG6qlxWpxJsnypu4sjEZI+xp6TFDtObVkzB9Uhka/ma5XL+01\nGXK22Fe42BT5aa4LioY5uDCvtlrFYDwpuG0mPMZ+Y5G3A/xDVG2JpxM/XrUrLkh36mWtzYZaLYQI\ns3gfLoRormR2GHBQSp981nEAAB0PSURBVHlISlmNWpHUj7TOA16TUhYDSCkLrJ+6a1m9T7mgLrfW\nWFSUqNQ/e9JmLWnbXnWKa8pnmbNFZSY5qhCvKcK6wd2bYPwT2l+rOXpdobJ8Mn9w3JiVpSrA7Cpj\nERoNZccdJ8996pDqwa5lD3GtCI9V8if2dI3M3a76V9jrAnYnzHELJ1dzW/uTizRlQAFg+nJvzlHf\nBbC8JcoxbbOkF9BLCPGbEGKTEGKSxb4AIcQ20/YG03SFEHeajtlWWNiI/IRGrEnPJzYiiB7trWxF\nbg7MtdQNBWoZenSTal3aEKlfqlRSZ2UmRcSrjCRX02mgSh5wZNwi33T35uxMKDNh0coAnnZQ45u6\nTChPXFnEKkNn68+iplKtDj09XmEmtItSdXZy3MJaY2EUQnQzvxFCxAKOcJj5oKRDUoDZwFsWK5gY\nKWUycCPwshDigvWjlHKRyTWWHBWlQcZPI5RX1bIxq4iJfTogrFVWtUeavDHiUkyVnA10tjUaVGZS\nz8vBP7jl1/IkvLzU6uLgT46r5i4wy3zY0F3QkTi61qJwPyBU0yNPw9702RN7lZHx1Mrthogfp4pQ\nayqddklrjcWfgfVCiCVCiA+Bn1GxhqbIBSx1uLuatlmSA6yQUtZIKQ8DB1DGAyllrun5ELAOcJse\niD/vL6TaYOSy5iq2LalreuSAlUXMJY1Llh/5TfkzneGCckccXc2dn6YycULqL4qdRF0Vt6OMRYZa\n3brDStBW7DUWnl653RDx45VGW0t63NiItQHuVSiV2f3Ax8DvgUZ8IHVsBXoKIboLIfyAG4D6WU1f\no1YVCCEiUW6pQ6aYiL/F9lGAc6M5TbA6/QThQb4MibEheFycDQFhEBjW7KHN4t8Wooc1bCxSv1DF\ncT0vb/l1PBFHV3Obe1ho3ZujMcxGypErC0/MhAII6apSmG01Frnbla5XSOfmj/UUYkaZbhid54qy\nNsA9F9XH4vfAH4AlwNNNnSOlrAXuBb4H9gHLpJRpQohnhBDm7KbvgSIhRDqwFnhESlkEJKJqO3ab\ntr8gpXQLY1FjMPJTRgHjEzrg421DsMwRabOWxKXA8d2q37UZQw2kr4CEKz3zztER+AWpn40jqrmN\nRhWzcKU8hG+AaqnpiCpuQ61qeuSJ8QpQ2XZh0XYYix1qVeEqg68F/m1VKwAnBrmt/bZ7ABgKHJFS\njkO5hEqaPgWklCtNooPxUsrnTduelFKuML2WUsqHpZR9pJT9pJSfmLZvML0fYHp+x65PpwFbs09x\nurLWtpRZcEzarCVxKYA8X7L80M9KVdOZqq/uSK8rTNXcGS0bp/Socmm5WkvIUVLlxdlKANJTVxZg\nqrWwoYq7slQZyNYUrzATn6LiMY31lnEw1hqLSillJYAQwl9KmQF46O1JyzALB47pZYNUtdGovrwc\nEa8w03kw+AWf74pK+1LJmPeY4LjreCLmau79LcyKMmdCtXexsXBUEyRPzoQyY2thXt5O9dzFbUKe\njiN+vHp2kvSHtcYix5Sl9DWwWgixHGiB/KNnYikcGORnQwFa+Ql1R+dIN5RZstz8h1JbBfu+VRXb\nrqg0difM1dwtjVuYGx65uiYhLFrJlLe0dazZWDhbtsSRhMeqXtRVZdYdb67cbo0ri04DVRzUSXEL\nawPc06WUJVLKp4EngHeAi06iPOOEjcKBZhyZNmtJXIoa+9Rh1fynqhSSLnIXlJlek+HYFjhz0v4x\n8lOVhpK/lbU0WhHaTd1snGlhzWrhfrVK8eSUavPq3FpXVO529TsMaqfdnFyFl7f6Dsj6ySnSHzaX\nM0opf5ZSrjBVZV9UrE7PRwiYkGhjR9m6tNlYx04oLkU9H/5ZZUEFtlPZQDpKhRYJmavtH8OVMh+W\n1EmVt7AJkqd1x2sIW9Nn83a2rpTZ+sSPUxX+Wqgt16MV1L47j9XpJuHAYBvVOkuOAOLcP72jiOwF\nwZ2Vb37/d9Dnas/W6ncknQYqiW97q7lrKpSGkqsqty2pK8xrQUaU0QAnD3h2cBtsMxZlJ1S1d2up\n3G4IJ0p/6MbCSo6XVrA310bhQDPF2cqP7uhYghBqdXFgFdScuXgL8RpCCFVrYm81d2GGktlwVeW2\nJXUrixYEuUuOqiIuT19ZBIarJA5rjEVrjleYCY9R4ohOiFvoxsJK1uxT/uKJtlRtmyl2cCaUJXEp\n6rltB1XZrXOOumru9bafaw5uu8PKIiBUfUG2JH3W7Kbw9JWFENarz+btUEV8nfprPi2Xknw7dB2m\n+WUu+m531rLaVuFAS4qztev1EDcWENB3ugp46ZzDXM29f9W5NENryU8D3yDHJyXYS0ulyltDJpSZ\n8Fgo2Nf8cbnbVSabq7o3OotL7nPKZfSVhRWUVdawMeukbcKBZmqrVADKkWmzlgR3hDnfQkpzUl0X\nIeZq7gN2VHPnp6ovGncxwKEtLMwr3K9iOI6Qm3E14bEqDthUKrGUpuB2K3ZBORndWFjBzwcKqTFI\nJvaxo3d2yTFAanuHGju6dXwJaEGvScpfb0s1t5TukwllxhErC0+PV5gJj1WpxOUnGj+m+LDqKNma\n4xVORjcWVrAmPZ92bfxsEw40Y/atahWz0GmaXleoZ1uqucsL4GyR6yu3LQmNhqrTqomWrUjp2QKC\n9bEmI6qujapuLByFbiya4ZxwYHu8vewQIivJVs9auaF0msaeau78VPXsbisLsG91UZqjsuVa08oC\nmjcWPgHQ3g2y2VoJurFohq2HlXCgTb0rLCnOBm9/1cFNxzXYWs1dlwnlRsYi1NTXwp64RWvJhDIT\nGg2IZozFdujYX687ciC6sWiGH9Lz8bdVONCS4iOqgU1r6P3rqdRVc1vZmzs/TRU7upNEREtWFnUC\ngq3EWPj4QWjXxo2FoVbJ97fmym0XoH+DNYHdwoGWlDhYmlzHdszV3NbGLdwtuA3QJkq5Veyp4i7M\nUD0x3Mn4tZSm1GcLM6C2Qo9XOBjdWDRBxokyckvsEA60pDhbj1e4GiFUoDvrJ5XK3BSGGji53z0q\nty0RQt1N27uyaC3xCjNNFeblblfPeiaUQ9GNRROcEw6001hUFKvmK/rKwvX0mgzV5c335i46qNIy\n3aFyuz6hXW2PWbS2TCgz4bGq13z12Qv35e1QVe/t4pw+rdaMbiyawCwcGBVsp6ZTndqsvrJwOd3H\nnKvmbor/b+/+g+Su6zuOP9+5yyUhP+7y48BAggFMHMPIL6/xBz/EQUPidAJVqlFrwR9lHE0tOrWF\nsQM2tjMqamudTBWVFhULlopGB8VAKSo1kgQDkpBAiIkkhMuSXH5Ak5Dk3v3j8/levll2bze3+93v\n3t7rMbOzu98fu+/73u697/O7GRu3E0NZBGn/jtDltuVKFmeE+z0lpirfvgZOPV/thHWmq1lGTRMH\nJpIPsqqh8lftaO7ex2HUaJg6u1GRVa/rdHixEGbErVarNW4nynWfPXwgrHCoxu26U7Io4771vQDM\nr6m9IkkWs2oPSGqXjOYebF6h3nXhv/D2jsbFVa3OIaxr0WrdZhPlksVzvwM/qvaKDGSaLMxsgZlt\nNLNNZnZ9mWPeZWbrzWydmX0vtf1qM3sq3q7OMs5Sfr6+lzOmjees7hpWSevbEpY9HNtZt7ikBsna\n3IOtcdG7vnkHcg2l+2xhQ1gUa/wQu343q5OmQseEl6+YlzRuq2RRd5klCzNrA5YBC4G5wHvMbG7R\nMbOBG4AL3f1s4Lq4fQpwE/B6YB5wk5kNYa6Nodl/8DArN+8a2sSBaeo221wmTQ/daMu1Wxzog33b\nmrO9AlKLIJ1IsoiN27V8jpuRWenus9sfCd2kJ03PI6qWlmXJYh6wyd03xyVY7wCuKDrmL4Bl7t4H\n4O7JIsOXAyvcfXfctwJYkGGsx0kmDhzyqO2Eus02n1cvhG2r4IXCy/f1rg/3zdgTCsLUJTaq+pKF\ne6hya7XG7UTJZLFGVVAZyTJZnAakP9Xb4ra0OcAcM3vIzFaa2YITOBczu9bMVpvZ6kKhxJd/iFbU\nMnFgor8/1I+rJ1RzmXM54LCpxNrczdwTCsLUFRNPrb5k8WIBDu5pvfaKRFcca5F0WDiwJyyFq8F4\nmci7gbsdmA1cCrwH+IaZVT3Xtrvf4u497t7T3d1dl4AOH+3ngVomDkzs3xH666saqrkMNpq79/FQ\nvz+xiefxOpGpygd6QrVwyeLIgTBLMIT1K0DJIiNZJovtwMzU8xlxW9o2YLm7H3b33wNPEpJHNedm\n4uE4cWBNXWZB3Wab1WCjuXeuD6WKZq7fP5FFkFq1J1SiuEfUwMjt8/OIpuVlmSxWAbPN7Awz6wAW\nA8uLjvkhoVSBmU0jVEttBu4F5pvZ5NiwPT9uy9yKOHHgxbNr7D0y0G32jNqDkvpKRnNvSa3N3d8f\n2iyatQoq0TUT9m0Pk+VVUtgQ1u5u5pJSLYqTxbO/hSlnwbiG9YUZUTJLFu5+BFhC+CP/BPB9d19n\nZkvNbFE87F5gl5mtBx4APuXuu9x9N/BZQsJZBSyN2zKVTBx48ewaJg5M9G0B4nw+0lzOfDO0jzt+\njYs9W8KaD82eLDpnhnEE+3dUPrawMVRBNXNJqRZdcdr2dMlCVVCZybTNwt3vcfc57n6Wu/9j3Haj\nuy+Pj93dP+nuc939te5+R+rcW939VfH2b1nGmXhiR5g4sOZeUBCqoSadCu1DnCpEsjN6XBjNvfFn\nxxpHm71xO3EiYy1acQLBtNFjQ4N/3xbYtyMkUI2vyEzeDdxNpeaJA9P6tqhxu5nNuRz2/iG0U0BM\nFgbdr8k1rIqqXQTpxV2hN1Srtlckku6zz8ZlVNVtNjNKFikrnniO82uZODCtb6u6zTazgdHcsSqq\nd12YpbTjpPxiqkZSrbm3wroWz7d443Zi8qxQit++BqwNpp+Td0QtS8ki2rH3AI9v38fb5tahMfDw\nwVAkVsmieRWP5m7GBY9K6TgJTppWuWTR6t1mE5Nnwb5nYeuvwxoko8flHVHLUrKIkokDa+4yC7E+\n2dVtttklo7n7tsLuzc07crtYNWMtChvD3Emt3sFi8izA4Q+/VhVUxpQsomMTB46v/cU02+zwMCeu\nzf3QV8L9cChZQHVjLQobYNqc1u0JlRj4h8zVuJ0xJQtgX70mDkz0/T7cq82iuU0/N4zm/u13w/Ph\nlCz2bht8XY5WXB2vlPQ/ZOo2myklC+DBjWHiwLpUQUFocGsbAxPq9HqSjWQ099FDMHr88EnuXTPD\nNBcvPl96/4E9oc2s1dsrIHzH2seGcTPN3pNtmFOyIHSZnTq+gwtOr9PIz76toXisZR2b35yF4f6U\nucPn9zWwCFKZHlHPPxnuTx4BfzzNwqjtU8+DthoH0sqgRvzVPXy0nwc27mTB2a+obeLAtL4tw+e/\n1JHuzDeHhuDp5+YdSfW6UutalKqnHyk9oRJXLgslC8nUiE8Whf2HmH3yBOafXcf5c/ZshZmvr9/r\nSXZGj4MP3ze8qgwrLa9a2Bj+eCYD+FqdJg5siBGfLE7tGscPPnph/V7wQB8c3Ktus8PJcKuuGTc5\nlIbKdZ8tbIDuOcOnWk2GBX2a6i3pNqtqKMmK2eDdZ0dKTyhpKCWLetujMRbSAF0zSzdwH9ofShwj\npb1CGkbJot6S6ZJVDSVZKleyKMSeUCpZSJ0pWdRb39ZQpzy2M+9IpJV1zQzrax/af/z2gZ5QShZS\nX0oW9aZus9IInanus2mFDWFAqD6DUmdKFvW2Z6uqoCR7ySpxxT2iChth2mwNUJO6U7Kop/5+2PMH\nNW5L9gZKFkWN3K2+Op7kRsminvbvgKMvqQpAsjfhFGjrOL5k8dKLIXmovUIyoGRRT+o2K40yahRM\nOu34NovnnwJcJQvJhJJFPQ10m52VZxQyUhQvglQYIUupSi4yTRZmtsDMNprZJjO7vsT+a8ysYGZr\n4+3DqX1HU9uXZxln3fRtBexYfbJIljpPP75kUdgAo9rDWuIidZZZlwkzawOWAW8DtgGrzGy5u68v\nOvROd19S4iUOuPt5WcWXib4toWqgvSPvSGQk6JwBLzwHRw5B+5hQspj6KmgbnXdk0oKyLFnMAza5\n+2Z3fwm4A7giw/fLn7rNSiN1Fc0+q55QkqEsk8VpQLoT+La4rdg7zewxM7vLzNL1N2PNbLWZrTSz\nK0u9gZldG49ZXSgU6hj6EPVtVXuFNM7AVOXPwOGDYTlftVdIRvJu4P4xMMvdzwFWALel9r3S3XuA\n9wL/bGZnFZ/s7re4e4+793R3dzcm4nIOH4T9z6rbrDROehGkXZvA+1WykMxkmSy2A+mSwoy4bYC7\n73L3Q/HpN4HXpfZtj/ebgf8BmnuFk6RXikoW0iiTZgAWPnuaE0oylmWyWAXMNrMzzKwDWAwc16vJ\nzKanni4CnojbJ5vZmPh4GnAhUNww3lw026w0WnsHTHxFaLMobAQbFRq4RTKQWW8odz9iZkuAe4E2\n4FZ3X2dmS4HV7r4c+LiZLQKOALuBa+LprwG+bmb9hIT2uRK9qJpLkixUDSWN1DkzjNo+tD90mW0f\nk3dE0qIynW3M3e8B7inadmPq8Q3ADSXO+1/gtVnGVnd7tkL72OG1lrMMf10zYfsaeGGnqqAkU3k3\ncLeOvi1hJlCteyyN1DkT9m6H3U8rWUimNI9xvajbrOShayb0Hw6PlSwkQ/o3uF76tqq9Qhqv8/Rj\nj9VtVjKkZFEPB/rg0F71hJLGS8ZaYGHRI5GMKFnUQ5+mJpecJKO4J8+C0eNyDUVam5JFPajbrORl\nzAQYN1lVUJI5NXDXw8CiR0oWkoOFX1CpVjKnZFEPfVvCf3djO/OOREaic96VdwQyAqgaqh7UE0pE\nWpySRT3s0RgLEWltSha12t8b5uZRe4WItDC1WQzVnmfgoa/Ab78D/UfgzLfkHZGISGaULE7Urqfh\nV1+GR+8ADM5dDBd9Aqa+bG0mEZGWoWRRrd518Msvwbq7oa0Dej4EF34cOmfkHZmISOaULCrZtgZ+\n+UXYeA90TIA3/SW8cQlMODnvyEREGkbJohR32PoQ/OKLsPkBGNsFl94A866Fk6bkHZ2ISMMpWaS5\nw6b74Rc3wzMrYXw3vPXv4Y8+BGMm5h2diEhulCwA+vthw09Cm8SOtTBpBiy8GS54vyZnExFBySJM\n1fG9d0NhQ1jDeNFX4ZzF0N6Rd2QiIk1DyWLSaWGqjks+BXOvhDZdEhGRYpmO4DazBWa20cw2mdn1\nJfZfY2YFM1sbbx9O7bvazJ6Kt6szC7JtNLzv+/Daq5QoRETKyOyvo5m1AcuAtwHbgFVmttzd1xcd\neqe7Lyk6dwpwE9ADOLAmntuXVbwiIlJeliWLecAmd9/s7i8BdwBXVHnu5cAKd98dE8QKYEFGcYqI\nSAVZJovTgGdSz7fFbcXeaWaPmdldZpYsKFztuSIi0gB5zzr7Y2CWu59DKD3cdiInm9m1ZrbazFYX\nCoVMAhQRkWyTxXZgZur5jLhtgLvvcvdD8ek3gddVe248/xZ373H3nu7u7roFLiIix8syWawCZpvZ\nGWbWASwGlqcPMLPpqaeLgCfi43uB+WY22cwmA/PjNhERyUFmvaHc/YiZLSH8kW8DbnX3dWa2FFjt\n7suBj5vZIuAIsBu4Jp6728w+S0g4AEvdfXdWsYqIyODM3fOOoS56enp89erVeYchIjKsmNkad++p\neFyrJAszKwBba3iJacDzdQonC4qvNoqvNoqvNs0c3yvdvWKjb8ski1qZ2epqsmteFF9tFF9tFF9t\nmj2+auTddVZERIYBJQsREalIyeKYW/IOoALFVxvFVxvFV5tmj68itVmIiEhFKlmIiEhFShYiIlLR\niEoWVSzGNMbM7oz7f2NmsxoY20wze8DM1pvZOjP7qxLHXGpme1OLRd3YqPhSMWwxs9/F93/ZKEgL\n/iVew8fM7IIGxvbq1LVZa2b7zOy6omMaeg3N7FYz22lmj6e2TTGzFXFhrxVxSptS52a+AFiZ+G42\nsw3x93e3mXWVOXfQz0KG8X3GzLanfodvL3PuoN/3DOO7MxXbFjNbW+bczK9fXbn7iLgRphx5GjgT\n6AAeBeYWHfNR4Gvx8WLCwkyNim86cEF8PBF4skR8lwI/yfk6bgGmDbL/7cBPAQPeAPwmx9/3c4QB\nR7ldQ+AS4ALg8dS2LwDXx8fXA58vcd4UYHO8nxwfT25QfPOB9vj486Xiq+azkGF8nwH+uorf/6Df\n96ziK9r/JeDGvK5fPW8jqWRRzWJMV3BsmvS7gMvMzBoRnLvvcPdH4uP9hEkVh+MaHlcA3/ZgJdBV\nNGFko1wGPO3utYzqr5m7/4Iw71la+nN2G3BliVMbsgBYqfjc/efufiQ+XUmY9TkXZa5fNWpZfK1q\ng8UX/3a8C/iPer9vHkZSsqhmQaWBY+KXZS8wtSHRpcTqr/OB35TY/UYze9TMfmpmZzc0sMCBn5vZ\nGjO7tsT+Zlm4ajHlv6R5X8NT3H1HfPwccEqJY5rlOn6QUFIspdJnIUtLYjXZrWWq8Zrh+l0M9Lr7\nU2X253n9TthIShbDgplNAP4LuM7d9xXtfoRQrXIu8FXgh42OD7jI3S8AFgIfM7NLcohhUBamxF8E\n/GeJ3c1wDQd4qI9oyv7rZvZpwozQt5c5JK/Pwr8CZwHnATsIVT3N6D0MXqpo+u9S2khKFtUsqDRw\njJm1A53AroZEF95zNCFR3O7uPyje7+773P2F+PgeYLSZTWtUfPF9t8f7ncDdhOJ+WlULV2VsIfCI\nu/cW72iGawj0JlVz8X5niWNyvY5mdg3wx8D7YkJ7mSo+C5lw9153P+ru/cA3yrxv3tevHXgHcGe5\nY/K6fkM1kpJFxcWY4vOk18lVwH+X+6LUW6zf/BbwhLt/ucwxr0jaUMxsHuH318hkNt7MJiaPCQ2h\njxcdthz489gr6g3A3lSVS6OU/Y8u72sYpT9nVwM/KnFMbguAmdkC4G+ARe7+f2WOqeazkFV86Taw\nPynzvtV837P0VmCDu28rtTPP6zdkebewN/JG6KnzJKGXxKfjtqWELwXAWELVxSbgYeDMBsZ2EaE6\n4jFgbby9HfgI8JF4zBJgHaFnx0rgTQ2+fmfG9340xpFcw3SMBiyL1/h3QE+DYxxP+OPfmdqW2zUk\nJK0dwGFCvfmHCO1g9wNPAfcBU+KxPcA3U+d+MH4WNwEfaGB8mwj1/cnnMOkheCpwz2CfhQbF9534\n2XqMkACmF8cXn7/s+96I+OL2f08+c6ljG3796nnTdB8iIlLRSKqGEhGRIVKyEBGRipQsRESkIiUL\nERGpSMlCREQqUrIQyVGcBfcnecchUomShYiIVKRkIVIFM/szM3s4rj3wdTNrM7MXzOyfLKw/cr+Z\ndcdjzzOzlan1ICbH7a8ys/viJIaPmNlZ8eUnmNldcQ2J21MjzD9nYX2Tx8zsizn96CKAkoVIRWb2\nGuDdwIXufh5wFHgfYbT4anc/G3gQuCme8m3gb939HMJI42T77cAyD5MYvokw8hfCDMPXAXMJI3sv\nNLOphKkszo6v8w/Z/pQig1OyEKnsMuB1wKq46tllhD/q/RybKO67wEVm1gl0ufuDcfttwCVxHqDT\n3P1uAHc/6MfmXXrY3bd5mBhvLTCLMD3+QeBbZvYOoOQcTSKNomQhUpkBt7n7efH2anf/TInjhjp3\nzqHU46OEVeqOEGYhvYsw++vPhvjaInWhZCFS2f3AVWZ2Mgysof1KwvfnqnjMe4FfufteoM/MLo7b\n3w886GH1w21mdmV8jTFmdlK5N4zrmnR6mEb9E8C5WfxgItVqzzsAkWbn7uvN7O8Iq5qNIsww+jHg\nRWBe3LeT0K4BYdrxr8VksBn4QNz+fuDrZrY0vsafDvK2E4EfmdlYQsnmk3X+sUROiGadFRkiM3vB\n3SfkHYdII6gaSkREKlLJQkREKlLJQkREKlKyEBGRipQsRESkIiULERGpSMlCREQq+n8qqie0+j9u\ngQAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FjlFssvfrIwO",
        "colab_type": "code",
        "outputId": "f1bfef36-bc05-4740-fd73-4168d19762ad",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 313
        }
      },
      "source": [
        "loss = history.history['loss']\n",
        "val_loss = history.history['val_loss']\n",
        "\n",
        "plt.plot(loss, label='training loss')\n",
        "plt.plot(val_loss, label='validation loss')\n",
        "plt.title('Loss curve')\n",
        "plt.xlabel('epochs')\n",
        "plt.ylabel('loss')\n",
        "plt.legend()"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7f33dede8a58>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 17
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXwAAAEWCAYAAABliCz2AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    }
  ]
}