MLbyAjPle.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "MLbyAjPle.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/plenoi/CMU_DataScience/blob/master/MLbyAjPle.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "H_nfRBHqDh4A",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import pandas as pd\n",
        "import numpy as np"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MEGmUx6qDv_r",
        "colab_type": "text"
      },
      "source": [
        "# 1. Data Gathering"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ORTRQi9KDlbH",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 253
        },
        "outputId": "00156a06-29f2-4f29-aae8-a07ae68b046e"
      },
      "source": [
        "df = pd.read_csv('https://raw.githubusercontent.com/plenoi/Clinic/master/ultima_all_clean.csv')\n",
        "df = df.set_index('hn')\n",
        "df.head()"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>age</th>\n",
              "      <th>parity</th>\n",
              "      <th>hiv</th>\n",
              "      <th>menopaus</th>\n",
              "      <th>disease</th>\n",
              "      <th>surgery</th>\n",
              "      <th>conization</th>\n",
              "      <th>OPDsize</th>\n",
              "      <th>appearance</th>\n",
              "      <th>stage</th>\n",
              "      <th>pchemo</th>\n",
              "      <th>Wardsize</th>\n",
              "      <th>finalhisto</th>\n",
              "      <th>nodeyiel</th>\n",
              "      <th>RHlvsi</th>\n",
              "      <th>depth</th>\n",
              "      <th>size</th>\n",
              "      <th>utmet</th>\n",
              "      <th>vgmargin</th>\n",
              "      <th>vgmet</th>\n",
              "      <th>pelvicme</th>\n",
              "      <th>pmmet</th>\n",
              "      <th>adnmet</th>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>hn</th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "      <th></th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>2631840</th>\n",
              "      <td>52</td>\n",
              "      <td>3</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1</td>\n",
              "      <td>1</td>\n",
              "      <td>0.0</td>\n",
              "      <td>5.0</td>\n",
              "      <td>4.0</td>\n",
              "      <td>5.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>5.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>21.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>6.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2633481</th>\n",
              "      <td>32</td>\n",
              "      <td>2</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0</td>\n",
              "      <td>1</td>\n",
              "      <td>0.0</td>\n",
              "      <td>5.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>5.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>5.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>11.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>5.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2634477</th>\n",
              "      <td>52</td>\n",
              "      <td>2</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>5.0</td>\n",
              "      <td>NaN</td>\n",
              "      <td>5.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>5.0</td>\n",
              "      <td>4.0</td>\n",
              "      <td>35.0</td>\n",
              "      <td>6.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>4.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2633633</th>\n",
              "      <td>38</td>\n",
              "      <td>2</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>4.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>20.0</td>\n",
              "      <td>16.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>3.8</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>2.0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2630496</th>\n",
              "      <td>55</td>\n",
              "      <td>3</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0</td>\n",
              "      <td>0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>5.0</td>\n",
              "      <td>4.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>17.0</td>\n",
              "      <td>9.0</td>\n",
              "      <td>3.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>1.0</td>\n",
              "      <td>0.0</td>\n",
              "      <td>0.0</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "         age  parity  hiv  menopaus  ...  vgmet  pelvicme  pmmet  adnmet\n",
              "hn                                   ...                                \n",
              "2631840   52       3  0.0       0.0  ...    0.0       0.0    0.0     0.0\n",
              "2633481   32       2  0.0       0.0  ...    0.0       1.0    0.0     2.0\n",
              "2634477   52       2  0.0       0.0  ...    0.0       0.0    0.0     0.0\n",
              "2633633   38       2  0.0       0.0  ...    0.0       0.0    0.0     2.0\n",
              "2630496   55       3  0.0       1.0  ...    0.0       1.0    0.0     0.0\n",
              "\n",
              "[5 rows x 23 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 3
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1AU37nMUDz5g",
        "colab_type": "text"
      },
      "source": [
        "# 2. Data Pre-Processing"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Qtghkn56D6vB",
        "colab_type": "text"
      },
      "source": [
        "Remove Missing Value and Unnecessary Column"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rMPpezQTD3zv",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 433
        },
        "outputId": "20da6466-ca5b-4d95-be71-b947aa660f52"
      },
      "source": [
        "df.isnull().sum(axis=0)"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "age             0\n",
              "parity          0\n",
              "hiv             4\n",
              "menopaus        1\n",
              "disease         0\n",
              "surgery         0\n",
              "conization      5\n",
              "OPDsize        17\n",
              "appearance    101\n",
              "stage          24\n",
              "pchemo          1\n",
              "Wardsize      145\n",
              "finalhisto     10\n",
              "nodeyiel       12\n",
              "RHlvsi        366\n",
              "depth         489\n",
              "size          114\n",
              "utmet          98\n",
              "vgmargin       96\n",
              "vgmet          97\n",
              "pelvicme        1\n",
              "pmmet          94\n",
              "adnmet          7\n",
              "dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jjm8XsiWEEkR",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "df_clean_column = df.drop(['appearance','Wardsize','RHlvsi','depth','nodeyiel','vgmargin','pelvicme','adnmet'],axis = 1)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "D2GlGJEOEOzI",
        "colab_type": "text"
      },
      "source": [
        "Remove Row with Missing Value"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-x8E11kOELps",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "df_clean = df_clean_column.dropna(axis = 0)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jQjQYBZSEfm2",
        "colab_type": "text"
      },
      "source": [
        "Check Total Final Data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TctVEqJJEUab",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "db858ccf-f8f6-4412-a87a-45fb16a534eb"
      },
      "source": [
        "df_clean.shape"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(1555, 15)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 7
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0jaxWS0uEp-C",
        "colab_type": "text"
      },
      "source": [
        "Define Class"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Kf_jnmAAEksz",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "y = df_clean['pmmet'].values\n",
        "X = df_clean.drop(['pmmet'],axis = 1).values"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CzWi-07BE51P",
        "colab_type": "text"
      },
      "source": [
        "Check number of sample for each Class"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ffp25p8NEvPo",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "0f89116a-184b-4b9e-ca76-f70356282acb"
      },
      "source": [
        "allclass = np.unique(y)\n",
        "allclassnumber = [sum(y==eachclass) for eachclass in allclass]\n",
        "allclassnumber"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[1278, 277]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mcdC-T48HHuL",
        "colab_type": "text"
      },
      "source": [
        "# Imbalance Problem !!\n",
        "using smote to oversampling the samples in small class to be the same as bigger class \\\n",
        "\n",
        "or using undersampling by randomly remove samples in bigger class to be the same as small class \\\n",
        "\n",
        "https://imbalanced-learn.readthedocs.io/en/stable/over_sampling.html"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XT7ZcQ-NGKks",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 89
        },
        "outputId": "3fff7421-3869-48f4-b4fe-a8a7eef8ad9e"
      },
      "source": [
        "from imblearn.over_sampling import SMOTE\n",
        "sm = SMOTE(random_state=1)\n",
        "X_res, y_res = sm.fit_resample(X, y)\n",
        "\n",
        "allclass = np.unique(y_res)\n",
        "allclassnumber = [sum(y_res==eachclass) for eachclass in allclass]\n",
        "allclassnumber"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/sklearn/externals/six.py:31: DeprecationWarning: The module is deprecated in version 0.21 and will be removed in version 0.23 since we've dropped support for Python 2.7. Please rely on the official version of six (https://pypi.org/project/six/).\n",
            "  \"(https://pypi.org/project/six/).\", DeprecationWarning)\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[1278, 1278]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LN4pt-nrILro",
        "colab_type": "text"
      },
      "source": [
        "Train-Test Split"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1N0O2qMPILB3",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.model_selection import train_test_split\n",
        "X_train, X_test, y_train, y_test = train_test_split(X_res, y_res, test_size=0.2, random_state=1)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IivbCGJ-IvBy",
        "colab_type": "text"
      },
      "source": [
        "Train-Val Split"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZwGWOSyaIvbG",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "X_train, X_val, y_train, y_val = train_test_split(X_train, y_train, test_size=0.2, random_state=1)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7UM1RXvjIRae",
        "colab_type": "text"
      },
      "source": [
        "Normalization"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IjXyojz-ISoZ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from sklearn.preprocessing import MinMaxScaler\n",
        "scaler = MinMaxScaler(feature_range=(0, 1))\n",
        "scaler.fit(X_train)\n",
        "X_train_norm =  scaler.transform(X_train)\n",
        "X_val_norm = scaler.transform(X_val)\n",
        "X_test_norm = scaler.transform(X_test)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HVsFTS_1Jm3D",
        "colab_type": "text"
      },
      "source": [
        "#3. Choose Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zJYAjaraJp76",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        },
        "outputId": "3a89abe6-0dee-4bd1-a8e0-b404522ac1b2"
      },
      "source": [
        "from sklearn.linear_model import LogisticRegression\n",
        "from sklearn.model_selection import GridSearchCV\n",
        "\n",
        "params = {'C': [1, 2, 4, 8, 16]}\n",
        "clf = GridSearchCV(LogisticRegression(random_state=0, solver='liblinear'),params, cv=10)\n",
        "clf.fit(X_train_norm, y_train)\n",
        "print(\"Best params : \" + str(clf.best_params_))\n",
        "print(\"10CV accuracy : \"+str(clf.best_score_*100))"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Best params : {'C': 4}\n",
            "10CV accuracy : 78.47094801223241\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5ng0cf1BKf99",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "d13dcf1f-8a41-4ed9-c973-d6fa519a1420"
      },
      "source": [
        "y_predict = clf.predict(X_test_norm)\n",
        "print(\"Test accuracy : \"+str(sum(y_test == y_predict)/len(y_test)*100))"
      ],
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Test accuracy : 73.6328125\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wPowKobmLE94",
        "colab_type": "text"
      },
      "source": [
        "Plot Confusion Matrix"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "La0TCoilLACV",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 283
        },
        "outputId": "b6c176f0-1178-48b2-f99d-60f7330b0009"
      },
      "source": [
        "#!pip install matplotlib==3.1.0\n",
        "from sklearn.metrics import confusion_matrix\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "target_names = ['negative', 'positive']\n",
        "C = confusion_matrix(y_test,y_predict) \n",
        "C = C / C.astype(np.float).sum(axis=1)*100\n",
        "sns.heatmap(C,annot=True, fmt=\".2f\",cmap=\"GnBu\",xticklabels=target_names, yticklabels=target_names)\n",
        "plt.ylabel(\"True Label\")\n",
        "plt.xlabel(\"Predicted Label\")\n",
        "plt.show()"
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWgAAAEKCAYAAAA/2c+EAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xdCyAr6cX1Hb",
        "colab_type": "text"
      },
      "source": [
        "# Deep Learning"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_NEn52qcX3GM",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 64
        },
        "outputId": "6bc0b968-6457-47ba-fc2f-ec750002c9ec"
      },
      "source": [
        "#!pip install tensorflow==2.0.0-beta1\n",
        "import tensorflow as tf"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "<p style=\"color: red;\">\n",
              "The default version of TensorFlow in Colab will soon switch to TensorFlow 2.x.<br>\n",
              "We recommend you <a href=\"https://www.tensorflow.org/guide/migrate\" target=\"_blank\">upgrade</a> now \n",
              "or ensure your notebook will continue to use TensorFlow 1.x via the <code>%tensorflow_version 1.x</code> magic:\n",
              "<a href=\"https://colab.research.google.com/notebooks/tensorflow_version.ipynb\" target=\"_blank\">more info</a>.</p>\n"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xkMeZSMvYJre",
        "colab_type": "text"
      },
      "source": [
        "Design Deep Learning Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UnqxmVKJYEGE",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 329
        },
        "outputId": "a7ced310-8c17-4a4b-df6c-40c4a16adf70"
      },
      "source": [
        "tf.reset_default_graph() # Clear Model\n",
        "np.random.seed(1) # Fix Randomness\n",
        "tf.random.set_random_seed(1) # Fix Randomness\n",
        "model = tf.keras.models.Sequential([\n",
        "  tf.keras.layers.Dense(16, activation='relu', input_shape=(14,)),\n",
        "  tf.keras.layers.Dense(8, activation='relu'),\n",
        "  tf.keras.layers.Dense(4, activation='relu'),\n",
        "  tf.keras.layers.Dense(2, activation='relu'),\n",
        "  tf.keras.layers.Dense(1, activation='sigmoid')    \n",
        "])\n",
        "model.summary()"
      ],
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "dense (Dense)                (None, 16)                240       \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 8)                 136       \n",
            "_________________________________________________________________\n",
            "dense_2 (Dense)              (None, 4)                 36        \n",
            "_________________________________________________________________\n",
            "dense_3 (Dense)              (None, 2)                 10        \n",
            "_________________________________________________________________\n",
            "dense_4 (Dense)              (None, 1)                 3         \n",
            "=================================================================\n",
            "Total params: 425\n",
            "Trainable params: 425\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0cmPlrLlYMuO",
        "colab_type": "text"
      },
      "source": [
        "Set up Optimizer to 'adam' with is argubly the best one now, the loss function is set to binary_crossentropy (Binary Classification)."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XPb2TuaAYNCE",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model.compile(optimizer='adam', loss='binary_crossentropy', metrics=['accuracy'])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "7seedw5i4eTE",
        "colab_type": "text"
      },
      "source": [
        "Train Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZmiJkPgwYjCV",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "1a5e32fb-a5f3-4884-c610-8851a1f3bdab"
      },
      "source": [
        "train_acc = list()\n",
        "val_acc = list()\n",
        "for i in range(0,500):\n",
        "  history = model.fit(X_train_norm, y_train, epochs= 1, batch_size = 200, validation_data= (X_val_norm, y_val))\n",
        "  tmp_avg = np.mean(history.history['acc'])\n",
        "  tmp_avg_val = np.mean(history.history['val_acc'])\n",
        "  train_acc.append(tmp_avg)\n",
        "  val_acc.append(tmp_avg_val)"
      ],
      "execution_count": 29,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 79us/sample - loss: 0.7486 - acc: 0.4930 - val_loss: 0.7035 - val_acc: 0.5110\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.7134 - acc: 0.4985 - val_loss: 0.6860 - val_acc: 0.5599\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.6969 - acc: 0.5443 - val_loss: 0.6776 - val_acc: 0.5868\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.6867 - acc: 0.5688 - val_loss: 0.6739 - val_acc: 0.6210\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.6808 - acc: 0.5914 - val_loss: 0.6709 - val_acc: 0.6479\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.6757 - acc: 0.6257 - val_loss: 0.6674 - val_acc: 0.6724\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.6706 - acc: 0.6422 - val_loss: 0.6633 - val_acc: 0.6699\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.6653 - acc: 0.6440 - val_loss: 0.6580 - val_acc: 0.6650\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.6597 - acc: 0.6459 - val_loss: 0.6531 - val_acc: 0.6675\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 19us/sample - loss: 0.6544 - acc: 0.6446 - val_loss: 0.6484 - val_acc: 0.6675\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.6491 - acc: 0.6489 - val_loss: 0.6441 - val_acc: 0.6699\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.6438 - acc: 0.6606 - val_loss: 0.6400 - val_acc: 0.6870\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.6384 - acc: 0.6820 - val_loss: 0.6358 - val_acc: 0.6968\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.6335 - acc: 0.7003 - val_loss: 0.6318 - val_acc: 0.6993\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 19us/sample - loss: 0.6289 - acc: 0.7113 - val_loss: 0.6289 - val_acc: 0.7090\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.6244 - acc: 0.7205 - val_loss: 0.6241 - val_acc: 0.7139\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.6198 - acc: 0.7217 - val_loss: 0.6198 - val_acc: 0.7213\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.6156 - acc: 0.7229 - val_loss: 0.6164 - val_acc: 0.7188\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.6121 - acc: 0.7260 - val_loss: 0.6134 - val_acc: 0.7335\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.6083 - acc: 0.7315 - val_loss: 0.6109 - val_acc: 0.7335\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.6044 - acc: 0.7364 - val_loss: 0.6076 - val_acc: 0.7359\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.6006 - acc: 0.7401 - val_loss: 0.6046 - val_acc: 0.7286\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.5969 - acc: 0.7425 - val_loss: 0.6019 - val_acc: 0.7311\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.5935 - acc: 0.7431 - val_loss: 0.5992 - val_acc: 0.7335\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.5900 - acc: 0.7474 - val_loss: 0.5968 - val_acc: 0.7286\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.5868 - acc: 0.7511 - val_loss: 0.5936 - val_acc: 0.7311\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.5833 - acc: 0.7535 - val_loss: 0.5908 - val_acc: 0.7359\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.5800 - acc: 0.7572 - val_loss: 0.5887 - val_acc: 0.7359\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.5767 - acc: 0.7645 - val_loss: 0.5869 - val_acc: 0.7384\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.5742 - acc: 0.7670 - val_loss: 0.5856 - val_acc: 0.7482\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.5705 - acc: 0.7700 - val_loss: 0.5835 - val_acc: 0.7384\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.5684 - acc: 0.7633 - val_loss: 0.5826 - val_acc: 0.7457\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.5652 - acc: 0.7749 - val_loss: 0.5810 - val_acc: 0.7457\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.5625 - acc: 0.7859 - val_loss: 0.5792 - val_acc: 0.7433\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 18us/sample - loss: 0.5597 - acc: 0.7798 - val_loss: 0.5779 - val_acc: 0.7482\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.5566 - acc: 0.7810 - val_loss: 0.5773 - val_acc: 0.7408\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.5545 - acc: 0.7847 - val_loss: 0.5759 - val_acc: 0.7433\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.5520 - acc: 0.7810 - val_loss: 0.5752 - val_acc: 0.7482\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.5498 - acc: 0.7823 - val_loss: 0.5740 - val_acc: 0.7433\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.5483 - acc: 0.7976 - val_loss: 0.5733 - val_acc: 0.7457\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 11us/sample - loss: 0.5458 - acc: 0.7951 - val_loss: 0.5723 - val_acc: 0.7482\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 23us/sample - loss: 0.5430 - acc: 0.7933 - val_loss: 0.5704 - val_acc: 0.7482\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.5412 - acc: 0.8043 - val_loss: 0.5689 - val_acc: 0.7482\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.5390 - acc: 0.8012 - val_loss: 0.5674 - val_acc: 0.7506\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.5372 - acc: 0.8049 - val_loss: 0.5655 - val_acc: 0.7604\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.5355 - acc: 0.8018 - val_loss: 0.5647 - val_acc: 0.7555\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.5329 - acc: 0.8043 - val_loss: 0.5628 - val_acc: 0.7628\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.5313 - acc: 0.8037 - val_loss: 0.5628 - val_acc: 0.7628\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.5288 - acc: 0.8061 - val_loss: 0.5618 - val_acc: 0.7653\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.5270 - acc: 0.8043 - val_loss: 0.5614 - val_acc: 0.7555\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.5242 - acc: 0.8080 - val_loss: 0.5623 - val_acc: 0.7604\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.5230 - acc: 0.8067 - val_loss: 0.5611 - val_acc: 0.7628\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.5205 - acc: 0.8110 - val_loss: 0.5600 - val_acc: 0.7604\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.5184 - acc: 0.8110 - val_loss: 0.5580 - val_acc: 0.7579\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.5168 - acc: 0.8159 - val_loss: 0.5584 - val_acc: 0.7604\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.5146 - acc: 0.8153 - val_loss: 0.5567 - val_acc: 0.7579\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.5144 - acc: 0.8141 - val_loss: 0.5548 - val_acc: 0.7531\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.5111 - acc: 0.8141 - val_loss: 0.5549 - val_acc: 0.7653\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.5098 - acc: 0.8153 - val_loss: 0.5534 - val_acc: 0.7579\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.5076 - acc: 0.8153 - val_loss: 0.5524 - val_acc: 0.7604\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.5063 - acc: 0.8159 - val_loss: 0.5519 - val_acc: 0.7653\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.5045 - acc: 0.8202 - val_loss: 0.5488 - val_acc: 0.7653\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.5027 - acc: 0.8177 - val_loss: 0.5509 - val_acc: 0.7726\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.5011 - acc: 0.8208 - val_loss: 0.5480 - val_acc: 0.7677\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.5001 - acc: 0.8202 - val_loss: 0.5463 - val_acc: 0.7653\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4971 - acc: 0.8239 - val_loss: 0.5450 - val_acc: 0.7702\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4953 - acc: 0.8208 - val_loss: 0.5449 - val_acc: 0.7726\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4940 - acc: 0.8214 - val_loss: 0.5459 - val_acc: 0.7702\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4922 - acc: 0.8232 - val_loss: 0.5440 - val_acc: 0.7628\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 11us/sample - loss: 0.4910 - acc: 0.8239 - val_loss: 0.5423 - val_acc: 0.7726\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4892 - acc: 0.8232 - val_loss: 0.5431 - val_acc: 0.7726\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.4869 - acc: 0.8239 - val_loss: 0.5405 - val_acc: 0.7628\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.4864 - acc: 0.8257 - val_loss: 0.5409 - val_acc: 0.7628\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.4841 - acc: 0.8214 - val_loss: 0.5417 - val_acc: 0.7702\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.4821 - acc: 0.8245 - val_loss: 0.5393 - val_acc: 0.7726\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4808 - acc: 0.8275 - val_loss: 0.5392 - val_acc: 0.7726\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.4789 - acc: 0.8269 - val_loss: 0.5374 - val_acc: 0.7726\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.4772 - acc: 0.8287 - val_loss: 0.5352 - val_acc: 0.7677\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.4753 - acc: 0.8300 - val_loss: 0.5367 - val_acc: 0.7751\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.4739 - acc: 0.8263 - val_loss: 0.5365 - val_acc: 0.7726\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.4725 - acc: 0.8257 - val_loss: 0.5373 - val_acc: 0.7677\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4706 - acc: 0.8269 - val_loss: 0.5337 - val_acc: 0.7726\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.4703 - acc: 0.8300 - val_loss: 0.5320 - val_acc: 0.7677\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4674 - acc: 0.8281 - val_loss: 0.5349 - val_acc: 0.7653\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.4666 - acc: 0.8275 - val_loss: 0.5334 - val_acc: 0.7677\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.4672 - acc: 0.8312 - val_loss: 0.5318 - val_acc: 0.7775\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.4668 - acc: 0.8300 - val_loss: 0.5371 - val_acc: 0.7677\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4649 - acc: 0.8306 - val_loss: 0.5277 - val_acc: 0.7800\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.4615 - acc: 0.8306 - val_loss: 0.5334 - val_acc: 0.7702\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.4598 - acc: 0.8361 - val_loss: 0.5291 - val_acc: 0.7726\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.4609 - acc: 0.8349 - val_loss: 0.5247 - val_acc: 0.7677\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.4587 - acc: 0.8373 - val_loss: 0.5352 - val_acc: 0.7702\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4558 - acc: 0.8367 - val_loss: 0.5272 - val_acc: 0.7751\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4540 - acc: 0.8361 - val_loss: 0.5284 - val_acc: 0.7775\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4526 - acc: 0.8318 - val_loss: 0.5284 - val_acc: 0.7751\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4507 - acc: 0.8361 - val_loss: 0.5265 - val_acc: 0.7775\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.4494 - acc: 0.8367 - val_loss: 0.5263 - val_acc: 0.7775\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4483 - acc: 0.8379 - val_loss: 0.5271 - val_acc: 0.7824\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.4474 - acc: 0.8379 - val_loss: 0.5263 - val_acc: 0.7775\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.4458 - acc: 0.8379 - val_loss: 0.5299 - val_acc: 0.7775\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.4447 - acc: 0.8410 - val_loss: 0.5231 - val_acc: 0.7751\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4430 - acc: 0.8422 - val_loss: 0.5234 - val_acc: 0.7726\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4432 - acc: 0.8434 - val_loss: 0.5216 - val_acc: 0.7824\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.4401 - acc: 0.8446 - val_loss: 0.5292 - val_acc: 0.7751\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.4410 - acc: 0.8434 - val_loss: 0.5223 - val_acc: 0.7751\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4386 - acc: 0.8459 - val_loss: 0.5183 - val_acc: 0.7702\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.4368 - acc: 0.8446 - val_loss: 0.5245 - val_acc: 0.7873\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 18us/sample - loss: 0.4373 - acc: 0.8446 - val_loss: 0.5193 - val_acc: 0.7897\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.4360 - acc: 0.8465 - val_loss: 0.5167 - val_acc: 0.7873\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.4362 - acc: 0.8453 - val_loss: 0.5217 - val_acc: 0.7848\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.4334 - acc: 0.8477 - val_loss: 0.5143 - val_acc: 0.7824\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4313 - acc: 0.8477 - val_loss: 0.5202 - val_acc: 0.7800\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.4301 - acc: 0.8489 - val_loss: 0.5162 - val_acc: 0.7824\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.4293 - acc: 0.8502 - val_loss: 0.5154 - val_acc: 0.7848\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4288 - acc: 0.8453 - val_loss: 0.5178 - val_acc: 0.7848\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.4269 - acc: 0.8495 - val_loss: 0.5116 - val_acc: 0.7873\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4251 - acc: 0.8495 - val_loss: 0.5128 - val_acc: 0.7873\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.4241 - acc: 0.8508 - val_loss: 0.5153 - val_acc: 0.7824\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.4227 - acc: 0.8520 - val_loss: 0.5118 - val_acc: 0.7824\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4224 - acc: 0.8526 - val_loss: 0.5144 - val_acc: 0.7800\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4208 - acc: 0.8465 - val_loss: 0.5148 - val_acc: 0.7848\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 18us/sample - loss: 0.4195 - acc: 0.8483 - val_loss: 0.5127 - val_acc: 0.7848\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.4186 - acc: 0.8495 - val_loss: 0.5110 - val_acc: 0.7800\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.4181 - acc: 0.8502 - val_loss: 0.5111 - val_acc: 0.7848\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.4160 - acc: 0.8514 - val_loss: 0.5063 - val_acc: 0.7873\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.4154 - acc: 0.8526 - val_loss: 0.5041 - val_acc: 0.7897\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4150 - acc: 0.8508 - val_loss: 0.5093 - val_acc: 0.7897\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.4143 - acc: 0.8532 - val_loss: 0.5088 - val_acc: 0.7873\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.4125 - acc: 0.8520 - val_loss: 0.5060 - val_acc: 0.7897\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4108 - acc: 0.8538 - val_loss: 0.5056 - val_acc: 0.7848\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.4098 - acc: 0.8526 - val_loss: 0.5072 - val_acc: 0.7873\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4097 - acc: 0.8538 - val_loss: 0.5099 - val_acc: 0.7897\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.4085 - acc: 0.8532 - val_loss: 0.5043 - val_acc: 0.7824\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.4071 - acc: 0.8526 - val_loss: 0.5074 - val_acc: 0.7848\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4066 - acc: 0.8538 - val_loss: 0.5030 - val_acc: 0.7873\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4063 - acc: 0.8532 - val_loss: 0.5060 - val_acc: 0.7897\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4040 - acc: 0.8550 - val_loss: 0.5077 - val_acc: 0.7897\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.4038 - acc: 0.8532 - val_loss: 0.5044 - val_acc: 0.7922\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.4022 - acc: 0.8526 - val_loss: 0.5025 - val_acc: 0.7897\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.4011 - acc: 0.8550 - val_loss: 0.5045 - val_acc: 0.7922\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.4002 - acc: 0.8581 - val_loss: 0.5034 - val_acc: 0.7873\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4008 - acc: 0.8489 - val_loss: 0.4993 - val_acc: 0.7873\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3995 - acc: 0.8569 - val_loss: 0.4993 - val_acc: 0.7873\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3978 - acc: 0.8538 - val_loss: 0.5016 - val_acc: 0.7922\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.3967 - acc: 0.8544 - val_loss: 0.5000 - val_acc: 0.7873\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.3948 - acc: 0.8557 - val_loss: 0.5022 - val_acc: 0.7873\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3943 - acc: 0.8532 - val_loss: 0.4992 - val_acc: 0.7848\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.3932 - acc: 0.8557 - val_loss: 0.4949 - val_acc: 0.7824\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3939 - acc: 0.8581 - val_loss: 0.4962 - val_acc: 0.7946\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3920 - acc: 0.8575 - val_loss: 0.4979 - val_acc: 0.7971\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3901 - acc: 0.8612 - val_loss: 0.4963 - val_acc: 0.7971\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3885 - acc: 0.8599 - val_loss: 0.4978 - val_acc: 0.7971\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3885 - acc: 0.8575 - val_loss: 0.4952 - val_acc: 0.7824\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3870 - acc: 0.8563 - val_loss: 0.4969 - val_acc: 0.7922\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3864 - acc: 0.8575 - val_loss: 0.4925 - val_acc: 0.7873\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3903 - acc: 0.8569 - val_loss: 0.4958 - val_acc: 0.7922\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3884 - acc: 0.8569 - val_loss: 0.5024 - val_acc: 0.7971\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3856 - acc: 0.8618 - val_loss: 0.4904 - val_acc: 0.7995\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3849 - acc: 0.8599 - val_loss: 0.4959 - val_acc: 0.7995\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.3816 - acc: 0.8624 - val_loss: 0.4905 - val_acc: 0.7922\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3819 - acc: 0.8581 - val_loss: 0.4929 - val_acc: 0.7922\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3805 - acc: 0.8612 - val_loss: 0.4915 - val_acc: 0.7946\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3792 - acc: 0.8636 - val_loss: 0.4928 - val_acc: 0.8020\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3785 - acc: 0.8642 - val_loss: 0.4922 - val_acc: 0.8044\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3791 - acc: 0.8642 - val_loss: 0.4874 - val_acc: 0.7971\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3783 - acc: 0.8599 - val_loss: 0.4867 - val_acc: 0.7946\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3749 - acc: 0.8618 - val_loss: 0.4930 - val_acc: 0.7971\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3747 - acc: 0.8624 - val_loss: 0.4878 - val_acc: 0.7971\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3729 - acc: 0.8624 - val_loss: 0.4880 - val_acc: 0.8044\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3737 - acc: 0.8624 - val_loss: 0.4890 - val_acc: 0.8044\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3749 - acc: 0.8618 - val_loss: 0.4895 - val_acc: 0.7971\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 21us/sample - loss: 0.3796 - acc: 0.8624 - val_loss: 0.4876 - val_acc: 0.7995\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3748 - acc: 0.8636 - val_loss: 0.4924 - val_acc: 0.7971\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3711 - acc: 0.8667 - val_loss: 0.4833 - val_acc: 0.7995\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3676 - acc: 0.8679 - val_loss: 0.4915 - val_acc: 0.7995\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3686 - acc: 0.8661 - val_loss: 0.4877 - val_acc: 0.8068\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.3684 - acc: 0.8654 - val_loss: 0.4875 - val_acc: 0.8117\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3669 - acc: 0.8673 - val_loss: 0.4844 - val_acc: 0.8068\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3648 - acc: 0.8661 - val_loss: 0.4850 - val_acc: 0.7971\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3641 - acc: 0.8679 - val_loss: 0.4830 - val_acc: 0.7995\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3628 - acc: 0.8673 - val_loss: 0.4880 - val_acc: 0.8068\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3622 - acc: 0.8636 - val_loss: 0.4848 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3617 - acc: 0.8667 - val_loss: 0.4859 - val_acc: 0.8020\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3610 - acc: 0.8703 - val_loss: 0.4784 - val_acc: 0.8117\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3607 - acc: 0.8703 - val_loss: 0.4834 - val_acc: 0.7995\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3593 - acc: 0.8703 - val_loss: 0.4847 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.3587 - acc: 0.8740 - val_loss: 0.4857 - val_acc: 0.8020\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3577 - acc: 0.8679 - val_loss: 0.4771 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3571 - acc: 0.8722 - val_loss: 0.4779 - val_acc: 0.8020\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3562 - acc: 0.8740 - val_loss: 0.4820 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3559 - acc: 0.8691 - val_loss: 0.4814 - val_acc: 0.7995\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3553 - acc: 0.8667 - val_loss: 0.4769 - val_acc: 0.8020\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3535 - acc: 0.8709 - val_loss: 0.4870 - val_acc: 0.7995\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3535 - acc: 0.8709 - val_loss: 0.4806 - val_acc: 0.7971\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3524 - acc: 0.8685 - val_loss: 0.4815 - val_acc: 0.7971\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3511 - acc: 0.8716 - val_loss: 0.4795 - val_acc: 0.7995\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3519 - acc: 0.8685 - val_loss: 0.4737 - val_acc: 0.7971\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3543 - acc: 0.8654 - val_loss: 0.4799 - val_acc: 0.7995\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3508 - acc: 0.8728 - val_loss: 0.4726 - val_acc: 0.8020\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3484 - acc: 0.8673 - val_loss: 0.4755 - val_acc: 0.8020\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3466 - acc: 0.8709 - val_loss: 0.4788 - val_acc: 0.8044\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3467 - acc: 0.8709 - val_loss: 0.4736 - val_acc: 0.8044\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3478 - acc: 0.8673 - val_loss: 0.4804 - val_acc: 0.7995\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.3455 - acc: 0.8697 - val_loss: 0.4731 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3474 - acc: 0.8691 - val_loss: 0.4779 - val_acc: 0.7995\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3482 - acc: 0.8642 - val_loss: 0.4708 - val_acc: 0.8068\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3432 - acc: 0.8679 - val_loss: 0.4733 - val_acc: 0.8068\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3433 - acc: 0.8709 - val_loss: 0.4723 - val_acc: 0.8142\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3408 - acc: 0.8722 - val_loss: 0.4740 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3407 - acc: 0.8709 - val_loss: 0.4733 - val_acc: 0.8117\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.3391 - acc: 0.8728 - val_loss: 0.4757 - val_acc: 0.8020\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3392 - acc: 0.8722 - val_loss: 0.4724 - val_acc: 0.8166\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3397 - acc: 0.8685 - val_loss: 0.4789 - val_acc: 0.8044\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3400 - acc: 0.8740 - val_loss: 0.4721 - val_acc: 0.8117\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3430 - acc: 0.8703 - val_loss: 0.4787 - val_acc: 0.7971\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3417 - acc: 0.8703 - val_loss: 0.4689 - val_acc: 0.8117\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.3395 - acc: 0.8685 - val_loss: 0.4725 - val_acc: 0.8044\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3372 - acc: 0.8703 - val_loss: 0.4738 - val_acc: 0.8068\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3344 - acc: 0.8746 - val_loss: 0.4717 - val_acc: 0.8166\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3363 - acc: 0.8697 - val_loss: 0.4712 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3360 - acc: 0.8722 - val_loss: 0.4745 - val_acc: 0.8020\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3319 - acc: 0.8771 - val_loss: 0.4710 - val_acc: 0.8142\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3342 - acc: 0.8722 - val_loss: 0.4762 - val_acc: 0.8020\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3309 - acc: 0.8740 - val_loss: 0.4737 - val_acc: 0.8142\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3310 - acc: 0.8734 - val_loss: 0.4738 - val_acc: 0.8068\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3303 - acc: 0.8771 - val_loss: 0.4697 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3310 - acc: 0.8752 - val_loss: 0.4699 - val_acc: 0.8117\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3300 - acc: 0.8758 - val_loss: 0.4709 - val_acc: 0.7995\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3306 - acc: 0.8752 - val_loss: 0.4672 - val_acc: 0.8142\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3305 - acc: 0.8716 - val_loss: 0.4765 - val_acc: 0.8068\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3278 - acc: 0.8734 - val_loss: 0.4706 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3275 - acc: 0.8765 - val_loss: 0.4673 - val_acc: 0.8142\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3282 - acc: 0.8740 - val_loss: 0.4732 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3255 - acc: 0.8783 - val_loss: 0.4697 - val_acc: 0.8166\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3254 - acc: 0.8789 - val_loss: 0.4712 - val_acc: 0.8020\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3328 - acc: 0.8697 - val_loss: 0.4769 - val_acc: 0.8020\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3286 - acc: 0.8777 - val_loss: 0.4687 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3254 - acc: 0.8789 - val_loss: 0.4678 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3241 - acc: 0.8765 - val_loss: 0.4719 - val_acc: 0.8068\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3237 - acc: 0.8777 - val_loss: 0.4642 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3226 - acc: 0.8813 - val_loss: 0.4660 - val_acc: 0.8068\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3217 - acc: 0.8758 - val_loss: 0.4672 - val_acc: 0.8020\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 18us/sample - loss: 0.3215 - acc: 0.8765 - val_loss: 0.4650 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3207 - acc: 0.8728 - val_loss: 0.4674 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3201 - acc: 0.8771 - val_loss: 0.4693 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3204 - acc: 0.8783 - val_loss: 0.4648 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3193 - acc: 0.8758 - val_loss: 0.4687 - val_acc: 0.8117\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3185 - acc: 0.8820 - val_loss: 0.4657 - val_acc: 0.8117\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3191 - acc: 0.8777 - val_loss: 0.4766 - val_acc: 0.8020\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 20us/sample - loss: 0.3228 - acc: 0.8789 - val_loss: 0.4631 - val_acc: 0.8117\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3197 - acc: 0.8746 - val_loss: 0.4747 - val_acc: 0.7995\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3176 - acc: 0.8740 - val_loss: 0.4661 - val_acc: 0.8166\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3158 - acc: 0.8765 - val_loss: 0.4742 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3168 - acc: 0.8758 - val_loss: 0.4681 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3143 - acc: 0.8789 - val_loss: 0.4698 - val_acc: 0.8117\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3160 - acc: 0.8795 - val_loss: 0.4633 - val_acc: 0.8142\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3178 - acc: 0.8758 - val_loss: 0.4732 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3148 - acc: 0.8783 - val_loss: 0.4656 - val_acc: 0.8142\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3138 - acc: 0.8813 - val_loss: 0.4659 - val_acc: 0.8166\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3126 - acc: 0.8795 - val_loss: 0.4633 - val_acc: 0.8166\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3123 - acc: 0.8777 - val_loss: 0.4639 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3157 - acc: 0.8777 - val_loss: 0.4768 - val_acc: 0.8044\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3184 - acc: 0.8740 - val_loss: 0.4623 - val_acc: 0.8142\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3138 - acc: 0.8820 - val_loss: 0.4703 - val_acc: 0.7971\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3106 - acc: 0.8789 - val_loss: 0.4619 - val_acc: 0.8142\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3110 - acc: 0.8807 - val_loss: 0.4718 - val_acc: 0.8166\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3103 - acc: 0.8820 - val_loss: 0.4623 - val_acc: 0.8117\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3114 - acc: 0.8801 - val_loss: 0.4582 - val_acc: 0.8142\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.3122 - acc: 0.8752 - val_loss: 0.4674 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.3104 - acc: 0.8795 - val_loss: 0.4664 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.3086 - acc: 0.8795 - val_loss: 0.4680 - val_acc: 0.8068\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3092 - acc: 0.8801 - val_loss: 0.4636 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3089 - acc: 0.8789 - val_loss: 0.4688 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3065 - acc: 0.8813 - val_loss: 0.4621 - val_acc: 0.8117\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3073 - acc: 0.8789 - val_loss: 0.4647 - val_acc: 0.8068\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3079 - acc: 0.8801 - val_loss: 0.4611 - val_acc: 0.8117\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3064 - acc: 0.8820 - val_loss: 0.4655 - val_acc: 0.8068\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3064 - acc: 0.8820 - val_loss: 0.4657 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3056 - acc: 0.8838 - val_loss: 0.4667 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3060 - acc: 0.8832 - val_loss: 0.4601 - val_acc: 0.8142\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3056 - acc: 0.8844 - val_loss: 0.4628 - val_acc: 0.8166\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3051 - acc: 0.8844 - val_loss: 0.4740 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3064 - acc: 0.8820 - val_loss: 0.4625 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3055 - acc: 0.8826 - val_loss: 0.4685 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3020 - acc: 0.8856 - val_loss: 0.4635 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3018 - acc: 0.8832 - val_loss: 0.4684 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3021 - acc: 0.8826 - val_loss: 0.4609 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3022 - acc: 0.8832 - val_loss: 0.4705 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3015 - acc: 0.8850 - val_loss: 0.4633 - val_acc: 0.8362\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3013 - acc: 0.8820 - val_loss: 0.4643 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3004 - acc: 0.8832 - val_loss: 0.4636 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3002 - acc: 0.8826 - val_loss: 0.4701 - val_acc: 0.8117\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.3029 - acc: 0.8820 - val_loss: 0.4630 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2994 - acc: 0.8826 - val_loss: 0.4646 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2998 - acc: 0.8844 - val_loss: 0.4621 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3023 - acc: 0.8826 - val_loss: 0.4702 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2977 - acc: 0.8850 - val_loss: 0.4630 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2988 - acc: 0.8838 - val_loss: 0.4700 - val_acc: 0.8068\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2988 - acc: 0.8850 - val_loss: 0.4609 - val_acc: 0.8166\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3000 - acc: 0.8875 - val_loss: 0.4699 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2970 - acc: 0.8899 - val_loss: 0.4615 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2978 - acc: 0.8875 - val_loss: 0.4644 - val_acc: 0.8142\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2976 - acc: 0.8899 - val_loss: 0.4620 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3004 - acc: 0.8844 - val_loss: 0.4640 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2976 - acc: 0.8893 - val_loss: 0.4658 - val_acc: 0.8313\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2952 - acc: 0.8826 - val_loss: 0.4682 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2964 - acc: 0.8813 - val_loss: 0.4631 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2945 - acc: 0.8844 - val_loss: 0.4695 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 18us/sample - loss: 0.2942 - acc: 0.8820 - val_loss: 0.4657 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2940 - acc: 0.8875 - val_loss: 0.4683 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 11us/sample - loss: 0.2953 - acc: 0.8862 - val_loss: 0.4645 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2952 - acc: 0.8869 - val_loss: 0.4685 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2935 - acc: 0.8905 - val_loss: 0.4652 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2958 - acc: 0.8832 - val_loss: 0.4735 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2978 - acc: 0.8844 - val_loss: 0.4613 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 11us/sample - loss: 0.2945 - acc: 0.8838 - val_loss: 0.4628 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2960 - acc: 0.8862 - val_loss: 0.4742 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2946 - acc: 0.8881 - val_loss: 0.4632 - val_acc: 0.8289\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2943 - acc: 0.8832 - val_loss: 0.4710 - val_acc: 0.8166\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2912 - acc: 0.8850 - val_loss: 0.4667 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2927 - acc: 0.8850 - val_loss: 0.4742 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 21us/sample - loss: 0.2916 - acc: 0.8881 - val_loss: 0.4648 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2906 - acc: 0.8881 - val_loss: 0.4801 - val_acc: 0.8117\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2945 - acc: 0.8838 - val_loss: 0.4641 - val_acc: 0.8142\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2904 - acc: 0.8881 - val_loss: 0.4746 - val_acc: 0.8166\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2907 - acc: 0.8862 - val_loss: 0.4642 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2904 - acc: 0.8838 - val_loss: 0.4596 - val_acc: 0.8289\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2927 - acc: 0.8893 - val_loss: 0.4755 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2915 - acc: 0.8869 - val_loss: 0.4625 - val_acc: 0.8289\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2908 - acc: 0.8887 - val_loss: 0.4798 - val_acc: 0.8166\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2903 - acc: 0.8899 - val_loss: 0.4646 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2878 - acc: 0.8911 - val_loss: 0.4653 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2881 - acc: 0.8862 - val_loss: 0.4732 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2902 - acc: 0.8844 - val_loss: 0.4748 - val_acc: 0.8289\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2873 - acc: 0.8911 - val_loss: 0.4681 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2888 - acc: 0.8869 - val_loss: 0.4578 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2877 - acc: 0.8887 - val_loss: 0.4637 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2863 - acc: 0.8856 - val_loss: 0.4699 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2861 - acc: 0.8887 - val_loss: 0.4684 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 18us/sample - loss: 0.2856 - acc: 0.8856 - val_loss: 0.4661 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2875 - acc: 0.8887 - val_loss: 0.4620 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2877 - acc: 0.8887 - val_loss: 0.4818 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2906 - acc: 0.8917 - val_loss: 0.4652 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2859 - acc: 0.8869 - val_loss: 0.4756 - val_acc: 0.8142\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2860 - acc: 0.8856 - val_loss: 0.4664 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 18us/sample - loss: 0.2845 - acc: 0.8844 - val_loss: 0.4620 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2920 - acc: 0.8875 - val_loss: 0.4781 - val_acc: 0.8166\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2877 - acc: 0.8887 - val_loss: 0.4597 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2842 - acc: 0.8893 - val_loss: 0.4701 - val_acc: 0.8166\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2825 - acc: 0.8930 - val_loss: 0.4678 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2826 - acc: 0.8911 - val_loss: 0.4764 - val_acc: 0.8313\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2855 - acc: 0.8905 - val_loss: 0.4622 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2840 - acc: 0.8924 - val_loss: 0.4846 - val_acc: 0.8142\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2865 - acc: 0.8930 - val_loss: 0.4619 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2821 - acc: 0.8905 - val_loss: 0.4820 - val_acc: 0.8142\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2863 - acc: 0.8807 - val_loss: 0.4702 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2820 - acc: 0.8924 - val_loss: 0.4774 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2817 - acc: 0.8905 - val_loss: 0.4659 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2841 - acc: 0.8905 - val_loss: 0.4731 - val_acc: 0.8313\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2831 - acc: 0.8936 - val_loss: 0.4721 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2818 - acc: 0.8899 - val_loss: 0.4710 - val_acc: 0.8313\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2827 - acc: 0.8899 - val_loss: 0.4726 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2799 - acc: 0.8862 - val_loss: 0.4630 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2836 - acc: 0.8881 - val_loss: 0.4758 - val_acc: 0.8166\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2799 - acc: 0.8930 - val_loss: 0.4695 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2802 - acc: 0.8936 - val_loss: 0.4735 - val_acc: 0.8313\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2784 - acc: 0.8960 - val_loss: 0.4752 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2781 - acc: 0.8911 - val_loss: 0.4731 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2775 - acc: 0.8936 - val_loss: 0.4691 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2770 - acc: 0.8917 - val_loss: 0.4757 - val_acc: 0.8337\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2797 - acc: 0.8942 - val_loss: 0.4663 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2824 - acc: 0.8917 - val_loss: 0.4814 - val_acc: 0.8313\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2806 - acc: 0.8887 - val_loss: 0.4638 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2781 - acc: 0.8905 - val_loss: 0.4680 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2771 - acc: 0.8911 - val_loss: 0.4688 - val_acc: 0.8362\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2759 - acc: 0.8942 - val_loss: 0.4849 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2840 - acc: 0.8942 - val_loss: 0.4687 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2783 - acc: 0.8924 - val_loss: 0.4828 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2764 - acc: 0.8960 - val_loss: 0.4649 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2821 - acc: 0.8899 - val_loss: 0.4742 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2747 - acc: 0.8960 - val_loss: 0.4648 - val_acc: 0.8289\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2790 - acc: 0.8911 - val_loss: 0.4843 - val_acc: 0.8142\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2795 - acc: 0.8911 - val_loss: 0.4710 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2771 - acc: 0.8924 - val_loss: 0.4753 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2746 - acc: 0.8979 - val_loss: 0.4672 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2780 - acc: 0.8960 - val_loss: 0.4912 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2780 - acc: 0.8942 - val_loss: 0.4688 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2738 - acc: 0.8960 - val_loss: 0.4825 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2739 - acc: 0.8960 - val_loss: 0.4687 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2774 - acc: 0.8905 - val_loss: 0.4683 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2761 - acc: 0.8942 - val_loss: 0.4799 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2739 - acc: 0.8972 - val_loss: 0.4703 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2748 - acc: 0.8930 - val_loss: 0.4825 - val_acc: 0.8289\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 19us/sample - loss: 0.2762 - acc: 0.8991 - val_loss: 0.4691 - val_acc: 0.8289\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2723 - acc: 0.8985 - val_loss: 0.4770 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2715 - acc: 0.8991 - val_loss: 0.4739 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2700 - acc: 0.8966 - val_loss: 0.4750 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 20us/sample - loss: 0.2712 - acc: 0.9009 - val_loss: 0.4814 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2714 - acc: 0.8985 - val_loss: 0.4781 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2706 - acc: 0.8960 - val_loss: 0.4769 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2707 - acc: 0.8924 - val_loss: 0.4689 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2722 - acc: 0.8930 - val_loss: 0.4759 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.2689 - acc: 0.8960 - val_loss: 0.4698 - val_acc: 0.8313\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2707 - acc: 0.8948 - val_loss: 0.4751 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2705 - acc: 0.8954 - val_loss: 0.4775 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2690 - acc: 0.8997 - val_loss: 0.4760 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2716 - acc: 0.8954 - val_loss: 0.4844 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2719 - acc: 0.8954 - val_loss: 0.4792 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2726 - acc: 0.8954 - val_loss: 0.4841 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2713 - acc: 0.8948 - val_loss: 0.4801 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2701 - acc: 0.8991 - val_loss: 0.4811 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2692 - acc: 0.8985 - val_loss: 0.4764 - val_acc: 0.8289\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2686 - acc: 0.8985 - val_loss: 0.4782 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2684 - acc: 0.8972 - val_loss: 0.4768 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2678 - acc: 0.8972 - val_loss: 0.4786 - val_acc: 0.8166\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2677 - acc: 0.9015 - val_loss: 0.4793 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2674 - acc: 0.9034 - val_loss: 0.4801 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2682 - acc: 0.8997 - val_loss: 0.4765 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2699 - acc: 0.8954 - val_loss: 0.4870 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.2678 - acc: 0.8954 - val_loss: 0.4783 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2683 - acc: 0.8930 - val_loss: 0.4724 - val_acc: 0.8313\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2668 - acc: 0.8985 - val_loss: 0.4909 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2693 - acc: 0.8948 - val_loss: 0.4747 - val_acc: 0.8289\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2662 - acc: 0.9003 - val_loss: 0.4881 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2655 - acc: 0.9003 - val_loss: 0.4802 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2709 - acc: 0.8966 - val_loss: 0.4954 - val_acc: 0.8117\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2703 - acc: 0.8936 - val_loss: 0.4721 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2658 - acc: 0.8985 - val_loss: 0.4838 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2651 - acc: 0.8979 - val_loss: 0.4787 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2660 - acc: 0.8979 - val_loss: 0.4756 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2666 - acc: 0.9015 - val_loss: 0.4929 - val_acc: 0.8289\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2674 - acc: 0.9015 - val_loss: 0.4784 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2670 - acc: 0.8966 - val_loss: 0.4804 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2646 - acc: 0.8954 - val_loss: 0.4820 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2651 - acc: 0.8960 - val_loss: 0.4780 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2670 - acc: 0.8972 - val_loss: 0.4966 - val_acc: 0.8142\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2654 - acc: 0.8954 - val_loss: 0.4720 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2668 - acc: 0.8985 - val_loss: 0.4936 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2649 - acc: 0.8997 - val_loss: 0.4719 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2638 - acc: 0.9028 - val_loss: 0.4931 - val_acc: 0.8142\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2634 - acc: 0.8997 - val_loss: 0.4877 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2626 - acc: 0.9015 - val_loss: 0.4824 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2705 - acc: 0.9009 - val_loss: 0.5006 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2675 - acc: 0.8979 - val_loss: 0.4736 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2677 - acc: 0.9015 - val_loss: 0.4924 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2617 - acc: 0.9021 - val_loss: 0.4788 - val_acc: 0.8289\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2614 - acc: 0.9021 - val_loss: 0.4903 - val_acc: 0.8386\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2629 - acc: 0.8979 - val_loss: 0.4742 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.2644 - acc: 0.8985 - val_loss: 0.4816 - val_acc: 0.8313\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2609 - acc: 0.9021 - val_loss: 0.4821 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2619 - acc: 0.9009 - val_loss: 0.4880 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2627 - acc: 0.9046 - val_loss: 0.4897 - val_acc: 0.8289\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2604 - acc: 0.8997 - val_loss: 0.4899 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2626 - acc: 0.8997 - val_loss: 0.4901 - val_acc: 0.8313\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2617 - acc: 0.9034 - val_loss: 0.4881 - val_acc: 0.8289\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2607 - acc: 0.8966 - val_loss: 0.4784 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2615 - acc: 0.9028 - val_loss: 0.4876 - val_acc: 0.8313\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2599 - acc: 0.9015 - val_loss: 0.4883 - val_acc: 0.8289\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2624 - acc: 0.9034 - val_loss: 0.4836 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2614 - acc: 0.8966 - val_loss: 0.4933 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2597 - acc: 0.9015 - val_loss: 0.4829 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2593 - acc: 0.8991 - val_loss: 0.4979 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2617 - acc: 0.9009 - val_loss: 0.4774 - val_acc: 0.8289\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2629 - acc: 0.9015 - val_loss: 0.5145 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2695 - acc: 0.8942 - val_loss: 0.4773 - val_acc: 0.8166\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2654 - acc: 0.8979 - val_loss: 0.5130 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2645 - acc: 0.8985 - val_loss: 0.4784 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2622 - acc: 0.9040 - val_loss: 0.5083 - val_acc: 0.8289\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2600 - acc: 0.9021 - val_loss: 0.4771 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2576 - acc: 0.8991 - val_loss: 0.4963 - val_acc: 0.8289\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2578 - acc: 0.9021 - val_loss: 0.4852 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 18us/sample - loss: 0.2580 - acc: 0.8972 - val_loss: 0.4894 - val_acc: 0.8362\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2602 - acc: 0.9046 - val_loss: 0.4915 - val_acc: 0.8362\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2581 - acc: 0.9040 - val_loss: 0.4898 - val_acc: 0.8386\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2575 - acc: 0.9040 - val_loss: 0.4839 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2577 - acc: 0.9070 - val_loss: 0.5046 - val_acc: 0.8313\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2586 - acc: 0.9058 - val_loss: 0.4861 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2569 - acc: 0.9028 - val_loss: 0.4934 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2554 - acc: 0.9009 - val_loss: 0.4862 - val_acc: 0.8289\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.2567 - acc: 0.8991 - val_loss: 0.4843 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2543 - acc: 0.9028 - val_loss: 0.4866 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2564 - acc: 0.8997 - val_loss: 0.4852 - val_acc: 0.8142\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 21us/sample - loss: 0.2556 - acc: 0.9021 - val_loss: 0.4953 - val_acc: 0.8362\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2552 - acc: 0.9021 - val_loss: 0.4824 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2561 - acc: 0.9009 - val_loss: 0.4894 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2551 - acc: 0.8985 - val_loss: 0.4831 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2554 - acc: 0.9015 - val_loss: 0.4832 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2558 - acc: 0.8997 - val_loss: 0.4828 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2544 - acc: 0.9058 - val_loss: 0.4907 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2559 - acc: 0.9009 - val_loss: 0.4844 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.2575 - acc: 0.9040 - val_loss: 0.4912 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2540 - acc: 0.9021 - val_loss: 0.4790 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2549 - acc: 0.9021 - val_loss: 0.4927 - val_acc: 0.8313\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2543 - acc: 0.9009 - val_loss: 0.4922 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2535 - acc: 0.9076 - val_loss: 0.4900 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.2526 - acc: 0.9064 - val_loss: 0.4973 - val_acc: 0.8313\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2531 - acc: 0.9058 - val_loss: 0.4886 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2536 - acc: 0.9034 - val_loss: 0.5086 - val_acc: 0.8289\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2569 - acc: 0.9064 - val_loss: 0.4845 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2532 - acc: 0.9015 - val_loss: 0.4942 - val_acc: 0.8313\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "n7ElN8WeZgqc",
        "colab_type": "text"
      },
      "source": [
        "Check Overfitting and Stopped Iteration"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Gq79-_nwZctA",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 663
        },
        "outputId": "3a5b2535-fec7-4dd7-8281-5032438d95f5"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "sns.set()\n",
        "plt.figure(num=None, figsize=(16, 8), dpi=90, facecolor='w', edgecolor='k')\n",
        "plt.plot()\n",
        "plt.plot(train_acc)\n",
        "plt.plot(val_acc)\n",
        "plt.title('model accuracy')\n",
        "plt.ylabel('accuracy')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['Train', 'Test'], loc='upper left')\n",
        "plt.show()"
      ],
      "execution_count": 30,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAABKMAAAJyCAYAAAAYSAOhAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAAN1wAADdcBQiibeAAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjAsIGh0\ndHA6Ly9tYXRwbG90bGliLm9yZy+17YcXAAAgAElEQVR4nOzdd3hUZfbA8e+UtEnPpCckhARCCoQe\nehVFQFcERVdZRcGCXSzr/lhdXd1dcdfFtugqruiuXQQBQTpICwSSEAIJ6b33nin398eEwZgAQZBQ\nzud5eB7mzr3vfe+dXEJOzjmvSlEUBSGEEEIIIYQQQgghLgJ1T09ACCGEEEIIIYQQQlw9JBglhBBC\nCCGEEEIIIS4aCUYJIYQQQgghhBBCiItGglFCCCGEEEIIIYQQ4qKRYJQQQgghhBBCCCGEuGgkGCWE\nEEIIIYQQQgghLhoJRgkhhBBCCCGEEEKIi0aCUUIIIYQQQgghhBDiopFglBBCCCHEOdi+fTvh4eEU\nFBSc03GTJ0/m1Vdf/ZVmJYQQQghx+ZBglBBCCCGEEEIIIYS4aCQYJYQQQgghzspkMtHW1tbT0xBC\nCCHEFUCCUUIIIYS4bP3+97/n5ptvZseOHUyfPp2YmBjuu+8+ampqyM3NZd68eQwaNIibb76Z1NTU\nDsc2Nzfz8ssvM2bMGAYMGMDs2bPZvXt3h30UReGtt95i1KhRDB48mGeeeYaGhoZO82htbWXp0qVM\nmDCB6OhobrzxRnbu3HnO17N69Wpuv/12RowYwfDhw5k3bx7Jycmd9jt48CDz5s1j8ODBDB06lHnz\n5nHs2DHr+4WFhTz55JPExsYSExPDDTfcwNq1awGIi4sjPDycEydOdBhz3rx5PProo53u7ZYtW5gx\nYwYDBw7kyJEjlJWV8dxzzzFlyhQGDhzIddddxz//+c9OgaqWlhaWLl3KpEmTiI6OZvLkyfzjH/8A\nYOnSpUyZMgVFUTocs2rVKqKjo6mqqjrneyeEEEKIy4e2pycghBBCCHE+iouLefPNN3n88cetAabn\nn3+egoICbr31VhYsWMDrr7/Ok08+yfr161GpVAAsWbKEbdu28eSTTxIUFMRXX33F/fffz8qVKxk2\nbBgAH3/8Me+88w73338/w4YNY/Pmzbz22mud5vDoo49y5MgRHnnkEYKCgtiwYQMPPvgg33zzDRER\nEd2+loKCAm666SaCgoJoa2tj/fr13HHHHaxfv55evXoBlmDSPffcQ2xsLH/7299wcHDg8OHDlJaW\nEhkZSWVlJXPnzsXBwYFnn30WPz8/Tpw4QXFx8Tnf28LCQl577TUWLVqEl5cXgYGBVFdX4+bmxnPP\nPYeLiws5OTm89dZbVFdX89JLLwGWIN6iRYtISEhg0aJFREdHU1paSnx8PACzZ89mxYoVHDhwgNjY\nWOv5Vq1axaRJk/Dw8DjnuQohhBDi8iHBKCGEEEJc1mpra/niiy8ICgoCIC0tjRUrVvDqq69y0003\nWfe77777yMrKIjQ0lMzMTNavX89f//pXZs2aBcC4ceO48cYbWb58OStWrMBkMvH+++8zd+5cnnji\nCes+8+fPp7S01Druvn372LFjB5988gkjRowAYOzYseTk5LB8+XLefPPNbl/Lww8/bP272WxmzJgx\nHDlyhDVr1ljfe/311wkPD2fFihXWwNr48eOtx3300Uc0NDSwatUqvL29ARg1alT3b+hP1NTU8NFH\nH3UIqPn6+vLss89aXw8ZMgQHBwf+8Ic/sGTJEmxtbdm9ezd79uzhX//6F1OmTLHue/LzCA0NZciQ\nIaxatcoajMrPzyc+Pp7ly5f/orkKIYQQ4vIhZXpCCCGEuKwFBARYA1EAwcHBAIwcOdK67eT7J4NI\nycnJKIrCtGnTrPuo1WqmTZvGoUOHAEvGVXl5eYdgCsDUqVM7vN67dy9eXl4MGTIEo9Fo/TNq1CiO\nHj16TteSmZnJQw89xOjRo4mIiCAqKors7GxycnIAaGpqIikpiVmzZlkDUT+3f/9+xo0bZw1EnQ8f\nH59OmV2KovDRRx8xffp0Bg4cSFRUFE899RRtbW3W7Kv9+/fj5ubW6d791Jw5c9i0aRONjY2AJSvK\n09OTcePGnfe8hRBCCHFpk8woIYQQQlzWnJ2dO7y2sbHptP3kttbWVgDKysrQ6XQ4ODh0OFav19Pc\n3ExbWxsVFRXWbT/f56eqq6spLy8nKiqq09w0Gk23r6OhoYF77rkHvV7P73//e/z9/bGzs2PJkiXW\nfkx1dXUoioKXl9dpx6mpqWHAgAHdPu+ZeHp6dtq2cuVKli5dysKFCxk+fDguLi4kJyfz0ksvWe9v\nTU3NGecIcP311/PKK6+wYcMGZs+ezerVq/nNb36DViv/PRVCCCGudPLdXgghhBBXHW9vb5qammhu\nbu4QkKqsrMTBwQFbW1trIKaysrLDsT9/7erqio+PD++88855zSkxMZGSkhI+/PBDQkNDrdvr6+ut\nf3dxcUGtVlNeXn7acdzc3M74vp2dHQAGg6HD9traWtzd3c86z40bN3LddddZSxfBktF1LnMA0Ol0\nzJgxg2+//ZaAgACKioq4+eabz3p+IYQQQlz+pExPCCGEEFedAQMGoFKp+OGHH6zbFEXhhx9+YOjQ\noQD4+fnh5eXF1q1bOxy7efPmDq9HjRpFRUUFOp2OAQMGdPrTXS0tLQDY2tpatx0+fJjCwkLra51O\nR0xMDKtXr+60Et1P57N7925rZtfP+fr6Ah0DSMXFxWRlZXV7nj+dI2Bdqe+nc6ipqWH79u1nHGvO\nnDnEx8fz1ltvMWjQoA5BOCGEEEJcuSQzSgghhBBXndDQUGbMmMFLL71EY2MjvXr14quvviIrK4sX\nXngBsJTYLViwgFdffRV3d3eGDRvGpk2bOmUBjRkzhrFjx3LPPfewcOFCwsLCaGhoIDU1ldbWVhYv\nXtytOQ0aNAidTscf//hHFixYQElJCW+//TY+Pj4d9lu8eDHz589nwYIF1lXzEhMTiY6OZtKkSdx9\n992sXr2aO+64gwceeABfX1+ysrJoampi4cKF+Pr6Eh0dzRtvvIGDgwNms5n33nsPNze3bs1z9OjR\nfPLJJwwcOJCgoCDWrl1Lbm5ul/dk8eLFPPTQQ0RGRlJeXk58fLx1xT2AmJgY+vbty6FDhzpsF0II\nIcSVTTKjhBBCCHFVevnll5k1axbvvPMOixYtorCwkHfffZdhw4ZZ97nrrru4//77+fzzz3nkkUdo\nbGzk6aef7jCOSqXi7bffZvbs2axcuZIFCxbwwgsvkJCQYM2y6g5PT0/eeOMNKioqWLRoEStXruTF\nF1+0NmQ/afjw4Xz44Ye0tLTw9NNP88QTT3DgwAFrxpOHhwefffYZERER/OUvf+GBBx7giy++wN/f\n3zrG66+/jr+/P08//TSvv/46ixYtIiQkpFvzfOihh5g5cyZvvPEGixcvxsbGhiVLlnS6J++88w5z\n585l5cqVLFy4kGXLlnVZBjhlyhTs7e2ZMWNGt++VEEIIIS5vKuV0Od5CCCGEEEL8yubMmUNISAiv\nvfZaT09FCCGEEBeJlOkJIYQQQoiLLjk5mf3795OcnMzzzz/f09MRQgghxEUkmVFCCCGEEOKiCw8P\nx8XFhYULF3Lffff19HSEEEIIcRFJMEoIIYQQQgghhBBCXDTSwFwIIYQQQgghhBBCXDQSjBJCCCGE\nEEIIIYQQF40Eo4QQQgghhBBCCCHERXNVr6ZnNiuYTOaensZ502rVGI2X/3UIcTHI8yJE98izIkT3\nyLMiRPfJ8yJE91zuz4pGo0atVp1xn6s6GGUymampaerpaZwXtVqFXu9EXV0zZrP0ohfiTOR5EaJ7\n5FkRonvkWRGi++R5EaJ7roRnxc1Nh1qtOeM+UqYnhBBCCCGEEEIIIS4aCUYJIYQQQgghhBBCiItG\nglFCCCGEEEIIIYQQ4qK5qntGnY2iKJjNJpRLuExTrVbR1taG0Wi8JOtJVSpQqzWoVGduXiaEEEII\nIYQQQoirgwSjTsNkMlFZWYzZbOrpqZxVRYUas/nS7bSvVmvQ6/3QaM7cwEwIIYQQQgghhBBXPglG\nnUZdXRVqtQYPDx/g0s7q0WpVGI2XXlaUhUJNTQV1dVW4u3v19GSEEEIIIYQQQgjRwyQY1QVFMdPa\n2oS7uzdarU1PT+estFo1cOlmRjk7u1FdXYaimFGppE2ZEEIIIYQQQghxNZPIQBdOlrxpNBKruxBO\n3sdLuZRQCCGEEEIIIYQQF4cEo7pwqmH5pV2ed/mw3MdLuRG8EEIIIYQQQgghLg4JRgkhhBBCCCGE\nEEKIi0aCUUIIIYQQQgghhBDiopFglDir1tZWxo4dxsGDcT09FSGEEEIIIYQQQlzmpEP3FWLs2GFn\nfH/+/IXce+/9v2hsOzs71qzZiIuL6y86XgghhBBCCCGEEOIkCUZdIdas2Wj9+/ffr+Xbb7/m/fdX\nWrc5OOg6HWM2m1EUBY1Gc9bx9XrPCzNRIYQQQgghhBBCXNWkTO8Kodd7Wv/odDrUanWnbfv3720v\nt9vPXXfdzqRJo8jJySY7O4unn36MmTOnMnXqeB588B5SUo5ax/55mV5ubg5jxw5j9+5d3H//fKZM\nGcMDD9xDXl5uT12+EEIIIYQQQgghLhOSGdVNK9Yd43B6xUU955C+ntw7M/KCj/vee//i8cefwtPT\nCw8PD3Jzc5gwYRIPPvgoWq2Gb775kqeffowvvliNs7Pzacf54IN3efjhx/H09OLVV1/mtdf+wltv\nvXfB5yuEEEIIIYQQQogrh2RGXYUeeOBhBg8eSq9eQTg6OhEZGc3MmTfRp08oQUG9efzxp7Gx0RIf\nf+aG5fPmzWfYsBH07h3CHXf8jiNHEjGZTBfpKoQQQgghhBBCCHE5ksyobrp3ZiT39vQkLpD+/Ttm\nWzU0NPDBB+8SF7eXqqpKzGYzra2tlJaWnHGc0NAw69/1ek9MJhN1dbW4u3v8KvMWQgghhBBCCCHE\n5U+CUVche3v7Dq/feOPvHDt2lIceegx//wBsbe147LEHMRiMZxxHqz315aNSqQBLU3QhhBBCCCGE\nEEKI05FglCA5OYlZs+YwduwEAKqrq6isvLj9sYQQQgghhBBCCHF1kGCUIDCwF9u2bWHYsFgMBgPL\nl7+JjY1NT09LCCGEEEIIIYQQVyBpYC54/PGn0Wq1LFx4F3/84++ZNm0Gvr5+PT0tIYQQQgghhBDi\nsqUoCuv25rDtcAFmRenp6VxSVIpy9d4Rg8FETU1Tp+1Go5GKikI8PQM69EW6VGm1aozGS7dX0+V2\nP8WVS61Wodc7UVnZgNl81f7TJ8RZybMiRPfIsyJE98nzIn5N3+3OZv+xUm6b0peBofqeno5VflkD\nL3x4AICIYHfumR6B3rVjD+e6pjZUgLPOFrgynhU3Nx02Npoz7iOZUUIIIYQQQgghhLgsFVY0smZP\nNiVVTSz7KonPt6b/allIZkVhR2Ih1fWt3do/NbcaAK1GzfHcap55dy//+DyB4zlVADS1GHhhxQH+\nvDIe01W2GJgEo4QQQgghhBBCiKvY5VQw9fO5frU9A0WB2EgfHO21bDqYz57k4l/l3PGpZXy8MY1P\nN5/o1v6peZZg1COzBzBzdG9cHG1Jyalm2ddHKKpoZN2+XGob26iobSE1r+ZXmfOlSmqmhBBCCCGE\nEEKIq9Teo8Ws3JjG07cNJizQtUfmUFLVhEatwsvN4Yz7fbj+OPFpZQzp50V0iAc1DW0cyazE282B\ne6ZHkFdazyufHGLVrixG9PfBzvbMpWLnKjmrEoCkzEqaWozo7LXklzXg5WaPvW3H8IrZrJCWV4NW\no6Z/kBsD+ui5aWwI3+3J5rs9OSxffZTS6lNtg+KOlRLV2+OCzvdSJplRQgghhBBCCCGuaDkldWw+\nmH/VlUKdjdFk5ttdWRiMZrYlFHR4r7HFwOofs6hrbOvy2OZWI2t2Z1Nc2Xhec2gzmHh5ZTwvfHiA\nnJK6M871wPFSWtpM7D1awr/XHuPL7RkAzJkYio1WTWiAKyMivKltaGPjgbzzmtfPKYrC0ewq61wO\nnyjnSGYlL3x4gJc/PkRDs6HD/vllDTS1GgkLcMFGawmKqdUqbhwTQt9AVworGjGaFK4d3gutRs2h\ntHIMl3Av6AtNglFCCCGEEEIIIa5YmUW1vPq/BD7bms7eoyU9PZ1zoigK/92UxgfrjnW7D5LRZKa0\nuonSqqZOAZKTY7a2mQA4eLyMyjpL/6OEExW0GkzW/VbtyuK7PTn80B7UqaprYemnh9l0MJ82g4m3\nvjnCmt3ZfLsr67yu8VhONU2tRlraTLz+RVKH4JbBaKatfU65pfW0Gc30D3JjwcwIpo0IYtqIIO68\nth9Dw72sx8yeEIpWo2JDXC6lVR0XLEvOquSVT+I7bK+ub6W0qumsfaAKyhupbWjDwc4SWNp7tJgv\ntqUDUFTRyD+/TKSoopHS6ibMZoXj7f2i+ge5dxhHrVaxYGYkDnYaPFzsmDWuDwND9TS3Gjnannl1\nNejxMr3//e9/rFixgvLyciIiIliyZAkDBw7sct+2tjaWL1/OmjVrqKiooF+/fjz33HMMHTr0Is9a\nCCGEEEIIIcSlrrC8gWVfJlmDLBvj8hgzwA+1SvWrnldRFA6fqKC2sRWVSsXAPvpOq6h1R3JWFdsO\nFwIQ3ceDkZG+FJQ3UF7TTFRvD2x/smJZZW0L38flcvB4mTUIpdWoeOneWHw9dNb91u/L5dtdWVw7\nohcp7Zk+fnodxZVNJGVUMCLCh7rGNnYfsfRdOtn3aFdSEal5NaTm1fD9vhzqmgzt79dgVhTUKhVH\nMisI9nHG1cnOer7WNhPHcqvo5e2Ep2vnMryE9HIAgnycyCttYNlXSby8IBa1WsVf/nuIxmYDf7lv\nJBkFtQBE9PZgdLTfae+Zl5sD02KDWLc3l79/nsgf5g3F3dkOs6Lw2ZZ0Sqqa+GxrOo/fEsPGuDxL\nz6n2Y0dG+TDv2nAc7CyhErOikJJdRViAK0ezLYGiqcN6sSOxyNrjqV8vNxRFIb2gliUfxAEQ4OWI\njcaS+9M/uGMw6uQc/3xvLBqNGjtbDbGRPhw+UU7c8VKG9vc+7bVdSXo0GPX999/z17/+lRdffJGY\nmBhWrlzJggUL2LhxIx4enWslly1bxvr163n55ZcJCgpi/fr1LFiwgPXr1+Pv798DVyCEEEIIIYQQ\nAixlW1/vyGTsQD9C/Fx6ejoAfPxDGo0tRqbFBpFZWEt6QS1J6RUM7ud19oPPQ3JWJe98m2x93T/I\njWd+O+ScxjCZzdYyNIBvdmTh7mTHsq+O0Gow4WCnYcKgAG6ZGArAG18nUVDeiEatom+gK82tJgrK\nG9iRUMhtU/oC0NRiZENcLgrww4F8AEL8XLhhdG/e/OYIccdKGRHhw5ZDBdaSsZySeppajNbAld7F\nnsq6Fnw8dNhq1eSXNVBY3khTi4FlXx0hwNOR5+8ehqLAp1vSiTtWSqvBRN9AV567s2MiidmskJRR\ngVql4qnbBvPBumMcyaxkS3wB9rYackvqATiaVUV6ezCqXzf6Wt00rg/lNS3EHSvl758n8Ps7hpBe\nUEtJe0bUkcxKNuzPZdWuLGy0asLay+b2p5SSUVDLE7fG4Kd3ZMP+XL7ZmUWwjzNarSWAOSBUT1OL\nkS2HClABt0/pi5ebPV9uz6CitoW6xjYKyi3ZXbZa9WmfBQ+XU8HJmFA9DnYaMgtrz3ptV4oeDUb9\n5z//Ye7cucyePRuAF198kR07dvDtt99y7733dtr/u+++4+GHH2bcuHEALFq0iC1btvDZZ5+xePHi\nXzQHtbpzRLyrbeL8qdUqubeiR538+pOvQyHOTJ4VIbpHnhUhOtqeUMj2hEISMyr484JYnBxsrO91\n93mpa2zDRqu2Zqacj8LyBtILavFxd+DWyWEcyahk2VdJbIjLY0i4F6puZEel5VXz9jfJtBpN2GjV\n3H9jFANDPc963Pf7LaVtU4YGciitjLS8GmoaWjsEIM5mV1IJRRWNhAe5obPXknCigqWfJaAolmyc\nrKI6Nsbl0dvXGXs7LQXljYT4ubB47iCcdDbUNrTy5Nt72HO0mDmTQrHVatiZVEhzq4mh4V40NBk4\nUVDDTeNCiArxwNFeS3JWJam51Ww/XGANaqXm1ZCUWUFWcR3uznb85f6RxKeWER3iwdZDBeSXNZCW\nV01BhSUAU1jRyFc7MimvaSYpoxI7Gw22WjVZRXUYTGbsfpLNlVlYS12TgYhgd1wcbfntNf1Iyd7P\nun05aDWnugodOF5KRkENGrWKPgGuZ/06UqNi4Q2RNLcaOZJZybKvjljfGx/jz66kIr7akQnAHVP7\nMXloIA3NBv7z/XEOpZXz7poUHr81hvX7cgFLiSCAo4OWUH9XbG00bE8oZNxAP0L8LcGme2ZEApZs\nqo1xeXyzI5Mh4V7daqJub6fl6duHYDKbr5rvLT0WjGprayMlJYUHH3zQuk2tVjN69GgSExO7PMZg\nMGBnZ9dhm729PQkJCb9oDlqtGr3eqcu5VVSo0WpVaLWXR1utS3ueKtRqNe7uOmxtbXt6MkLg7u7Y\n01MQ4rIgz4oQ3SPPirhatLQZ2Xown35BboQFupFbUs+R9HImDAnEWWfL7mRLWVd1fStfbs/k6XnD\nOo1xpuelodnA/y37ER+9jtcfG3/GYFFqbhUFpfWMiPLDxbHrnzFW7c4GYPqYELw8nZnk4cSqH7PI\nKKylvN5ARMjZVy7buyGN+mYDjvZaGpuNfPtjNhOHB595bjlVnMivIdDbiUdvG8LK9cdYtSODo7k1\nzJoYdtZzKorCtvh8Pt9q6Uf0wM0xONhreWjpNkxmhZsnhnH3zEiOZlbyh+V7WLUryxrkuv26/gT3\nspSF6fVOjIz2Y8+RItIK6xk9wI8t8QWoVLBw1kD89I7UN7VZS+rGxASwKS6XpZ9Zfr6ePKwX0X30\npOYl8t2eHBQFhkX44O/ryo2+luykBoOZ7/bkkFZQS2puNdr2srMt8ZZm6IHeTvztobG883US+5KL\nqWwwMCDsVGbTd3stwZ5xgwPR653Q6524fnRv1u3OBkyMjfEn/ngpB1PLMJkVwoPc8fft/op/f1ww\nkhf+vY9j7Vldwb7OLL5zGOW1uzmeU8XQ/t7MmRqOSqVCDzy/YBRL3t1LcmYFL6+Mp6XNxNQRQWQV\n1ZJZUMuQcB+8vJzx8nLmP3+8FhcnOzRdBI3mzYhizjXhaDUqa/Pys/l5bOJK/97SY8Go6upqTCYT\nnp4do8p6vZ7c3Nwujxk9ejQffvghw4YNw9/fn40bN5KQkEBQUNAvmoPRaKaurrmL7UbMZjNGowJc\n+t3stVo1xku4677RqGA2m6mubkKr7XolBiEuBrVahbu7I9XVjZjN3WsAKcTVSJ4VIbpHnhVxOVEU\npVuZQGeydk8O3+y0ZJM4OdhY+xLtOlzAzDG9Kalsoo+/C7UNbexKLCQi2I1RUb5A956XLfH51De1\nUd/URtyRQvoGunW5X5vBxJ/+vY/GFiMadRLBvs6o1Sq83SwZUG5OdrQZTWw9kIdGrWJwqJ7KygYA\nJg8O4KMNqazZmYG3S2SnsfPL6lnzYza3X9MPVydbDhwrQatR84+Hx/C3/x4mu6iOnfF5DOijB6Ch\nycDKjan4eDgwMtIXLzcHPt14HIBrh/eiurqRmD4erNoB2w7mMX6A7xnvsaIo/Of7VHYlFaFSwa2T\nwnDXWX5sf+CmaGobWpkyNJCqqkb83e0Z3NeThPQKyqqb8dPrCPNzsl4rwKgob/YcKWLVtnQOJBdR\nXd/K0HAv7NVQXW3JZKpstXyOU4cG0NzShsFgxsZGzcyRQZjaP6uTDb/7Brh0GN/L2dayElxqGQBD\n+nkRG+nD8tVH0bvY8+StMRhbDfT2cWJfMsSnFOPvbo/BaCYpo4Lthyylgv0CnK3jXjcskG3x+RhN\nZm4eF4LJaGZfiqXxfG9f5w7n746HZkXz6v8Ok1fawLTYIKqrG7lrWjhbDxVww+jeVFV1XA3w7mnh\nLPmghur6VnT2Wn4zujdmReH7/bmMj/HvcP6aauM5zaU7roTvLS4uDtjYnDkI1+MNzM/FkiVL+MMf\n/sDUqVNRq9VEREQwY8YMUlJSfvGYXX24l+MHPnZs5986/NT8+Qu59977z+scL7zwHBqNluef//Mv\nOt5sVi7LeyuuPPK1KET3yLMiRPfIsyIudV9sS2d/SilP3BpDkI9zt48zmc1kF9cT6u+CSqUiKbMC\nsDRnLixvJNDLEYNJIS2/hsI1lp/JrhkaiLuzHUs/TeDjjWn0DXDtUJp2uudFURR2JBRaX+84XEio\nf9cZMAeOl9LYYsTb3YHGZgNZRXUAZBTUciSzktkT+lBZ10Jji5Hh/b1xcrCxnnN4f28+25rOgeOl\n3DYlDEd7mw5jr92TQ3xaOXa2GkZG+dLcamRgqB5brYbrRwazfPVR1u/NIaq3Jatq2+ECDrYHYtbt\nPZVU4eZkS2yED2azQqCXI74eOnJK6imqaOzQTBwsK7E52mtxdbIjMaOCXUlFuDnZ8sBvounXy806\n96Htfa4UxXK/AOZMDOVIZiUms8K0EUGg0GHVvf5B7ni62pNdXEd2cR22WjUzR/Xu8jPQu9gz//qI\nLrdX1rWgUkFEsHuHY7VqNWEBLtZm3iMivBne3xv3eUPxcXfAWWeL2axYP8sTBTU0txp56aODFFda\nAlwD+ujxcLa3jutob8P/zRuKyazg5mTHiAhvazAqLMD1nP+9dbDV8vs7hpBf1kDfQMv99HZz4Pb2\nPlo/H8/d2Y67poXz3ncpzB7fB529JWxy66SwLvf/tVzp31t6LBjl7u6ORqOhoqKiw/bKykq8vLpu\nJqfX63nvvfdoaWmhrq4Ob29vnnjiCQIDAy/GlC9pa9ZstP79++/X8u23X/P++yut2xwcdF0dJoQQ\nQgghhLiCGU1mdiUV0dxq4vUvk3jujiH4eHTvZ4Mvt2WyOT6f+dP7M7SfF1mFdbg62vLSPSNoM5ix\ns9VQWN7Aix/F09BeyjY03AsbrYbrRgSx8UAeH6w7xlO3D0bNmbOysorrKChvJMDTkYq6Fg6klnHb\nNX07BYsAdiYWAXDXdeH07X0ihpYAACAASURBVOVGfZMBs1nh+7hcth8uZOXGNOu+4wd1XOjKwU7L\nyEgfdiYWsT+llClDT/0saTSZSc6yrJi2P6WU5lbLCnxD2oNAQ/t54e3uQGpeDVlFdfTxdyHueCkA\n00YEkZZfQ6vBhFoFM0f3xqa9lYpKpSI20oc1u7PZfaSYOe0Nx40mM6t/zGbD/lycdDY8c/tgvmpv\nWP67af3p16vrzLCf8tM7csukMDILaxkZ1TnrSq1SMXdyGFsPFTCgj57YSJ9z6lsFlubre46W0MfP\npcvPo3+QO6l5NdjZaogJs1Q+hQV0DCQG+Thhq1WTWVjLrqQiiiub6O3rzIxRvRkYqu/yuk6KCvHA\nycGGxmYDYd1oXt4Ve1vtaTPtujIiwoeYMM8O/a3EhdVjwShbW1uioqLYu3cvkydPBsBsNrNv3z7u\nuuuuMx5rb2+Pvb09dXV17N69myeffPJiTPmSptefKnfU6XSo1eoO205KT0/j7bff4OjRJJydXRg7\ndjyLFj2KTmd52Ddt2sjHH6+gqKgQnU5HZGQ0S5cuY/nyt9i6dXP7PhsAePfd/xAdPeAiXJ0QQggh\nhBCXh+LKRlZuTOOG0b2J6kZPol/i0y0nAPjtNf0A+G5PNhW1Lcy/vn+nUrzUvGqaW03Y2Wioa2zr\nsNT9rqQifjiQh8ms4ORgwyM3D7D2DyqpamLbYUvfn22HC3Gw1WJWFKJCPFCpVNamzAFeTtwyMZTP\ntqYzPsbf2h9n1vg+HM2uJDWvhq2HCrhuxKnWKi1tRt5fewwfdx03jQvB1kZjDTBNGRpIbmk9OxOL\n2H64kJGRPtbjnB1tqahtIb2gFm93B8KD3VGrVLg7W+Y879pwBvf1JDmzCgUFL1cHIoPdO92/CYP8\n2ZlYxI7EQmJC9ejstejsbaz3SqNWYTIrHD5RjgqsARa1WsW02CA+3pjGtz9mMXdSWHuGmBO3Tj5z\nL6iRUT6s3ZPD9/tzaW414qfXsSupyLr6XX2TgT+vjKfNaCYi2J2YLgI0p3Pt8F4wvNdp3x8a7s3Q\ncO9uj/dzg/p6sudoyWnHGBimZ83ubGIjvE8bvNFq1PTxt2RQfbsrC4C7r+/frUw9rUbNQ7OiqWsy\n4Hqa/mC/BglE/bp6tExv/vz5PPvss0RFRTFw4EBWrlxJS0sLs2bNAuCZZ57Bx8fHulJeQkIClZWV\nhIeHU1BQwGuvvUZQUBBz5sz51efavON9jDmHf/Xz/JS29xAcJi68YOPV1NTw2GOLuPXW21m8+Fka\nGxtYtuzvvPrqK7z44l8oKSnmlVde4LHHnmL06LHU19eTkBAPwF133UteXg4ajYYnnngGAFfX7keW\nhRBCCCGEuNIZTWbe+y6FvNIG1u7N+VWCURmFtdbm0NfHBqOz07Jubw5Gk8KEQf6dStsS0i2VKLdf\n05fU3Gr2ty91P3FwAJ9tSbfuV1bdzL6UUqbFWoJGX+/IxGRWUKtU5JbUsyHOsjpcdJ/O1zR1eC/6\n9XIjwOtUNouNVs3CG6L404cHWL8vl0lDAqzvbYzLs84rObsSL1cHjmRaVl2LjfSht58zOxOLWLUr\ni1XtgQsAW60a9/asngkx/qi76IEVHaInOuTMgZzevi4E+ziTW1rPM+/uw1arZtGsaJIyLVlRN40L\n4bs9ORiMZkIDXDsEQMYO8OOHA/mkZFfxscGSgRUbefZAj4+7jkdmD2DF+uNs/0k5Yt9AV+6dGcma\nH7PZl1KCCks52Pn297qQhvTz4k/zhxPo1XnxL7Dcz5cWxOLleuaMq7D2lfla2kxEh3icU8loeFDn\noKK4vPVoMGr69OlUVVXx5ptvUl5eTkREBB988AEeHpZ/4IqLi1GrT60S19LSwt///ncKCgpwdnZm\n6tSpLF68GBubzqmCorMvv/yUwYOHcPfdC6zbnnrq98yffwdPP/0HysvLMJvNTJgwCb3eE19fP/r2\ntfy2RaezrISn0Wi7zLgSQgghhBDiardmdzZ5pZbmxun5NVTVtXQqiWozmFi/L5eoEI8zlmEpisK6\nfbnkty8pP6CPnnEx/mzYf6ov0dHsSlx0thhNlr4yccdKOwSjFEUhMb0ClcqS3TI62pem9qXuP9uS\njkat4qGbB+Css+GVjw+RkF7OtNggTuTXcPhEOXoXOyYPDeSr7ZlkF9ehAmuvpJ8L9u0cWOjl7cSQ\ncC8OpZWz50gxc6a6Ul3fwsa4PLSaU72GCssbcXKw4ZZJoTjYaent68L1sUHWXlAnryW3rIHSqiY0\nahVjBvid5dM4s1snh7UH8cykF9Ty4frjqFQqVMC4gf5U1bey/XChtUTvJK1Gza0TQ3lrVTIZBbUA\nxEb4dHGGzmLCPHnp3hGs3ZuDRqViVLQvvX2dUalUzJ/eHw8XO9yd7bq8lz1JpVKdNXAU4Hn2ld8s\nZXKWr9/rRwZfiKmJy1iPNzC/8847ufPOO7t875NPPunwetSoUWzcuLHLfX9tFzJDqadkZKRz4MA+\npk4dZ912svFdUVEBERFRREcP4I475hAbO5rY2FFMnDgFnU76TQkhhBBCiEtbS5sRFafKx36uqcWI\nwWiylqGdTdyxUtbszuaR2QM69K/pSlpeNXuOlrAnuRh7Ww0DQ/UcOF7GgeNlxITpee+7FEZG+jIt\nNohvdmaxOT6fLYcK+PO9I07bv+dIZqW1nAkgPq2cvLIGEtIrUKtUmBWFlOwqnB1OZe0cPF7GbZP7\nom5faj6npJ7q+lb6BrriorPst+imaP75ZRInCmq4d2YEg8I8URQFd2c7MgpqqWtsY83ubABuHh9K\nTJie1T9mYzCa6e3njLPu3Mqkpo8M5lBaORvj8pg1JZxvdmbRZjRz/cggZk8IJSnDcj1RIR5oNacS\nEW6Z1LnsrdVgIjmzEmedDS7nWa4VEexORHsJ38qNqdYywb6Brrg42nLLxFCCfZytqwH+1KC+nvTr\n5caJ/BrCAlzxdHPo9nndnOyYd214p+1ajZrZE0J/4dVcHsICXHG01xLo5UT/IKmyudr1eDBKXDzN\nzU1MmTKV+fPv6/Set7cPWq2Wd975gKSkBOLi9vHxx//hww//zYoVn0hJnhBCCCGEuGSZzGb+8skh\nqutbefr2wR2yOLKL69iwP5fEjEpUKvjDnUM7ZJ5U1DZzIr+GkZG+1iBOU4uR/20+QUOzge/25HD/\njVGnPfehtHLe+TYZsJSmzZ8egbebAweOl7E/pYS4Y6XklTaQV5pBUUUju5OLAWhuNZ5q7v2zkiyT\n2cyX1kbW4TjZ2/DBumNsPWQpz5s5Opjv9+eSkl1lbSjdy9uJ/LIGUvOqiWzPXjp8ohyAwX1PZffY\n2mh4+vbB1Def6r+jUqkY1NeT7YcLWbMnm+O51Xi7ORAb6YNarWJEf2/2HC0h6izlb10J8XOhf5Ab\nqXk13PfXLZRVNeHkYMOMkb1Rq1Qd5nY2djYahvX/5b2PTmfu5DCO51ZTVt1snY+9rZbxMf5d7q9S\nqbhzaj/e/S6F60cGdbmP6MzBTstf7x+FjVZ9SZUhip6hPvsu4krRr184WVmZBAQEEhjYq8MfW1vL\nNyK1Ws3gwUN54IGHWbnyU6qrq4iPPwiAVmuDyWTqyUsQQgghhBBXMKPJzCc/pBF3rPScjjuYWkZB\neSONLUZe/zKJ0mrLkvEZhbW8+ulh4tPKUanAYDTz/rpjtBlO/Z/26x2ZfLDuOB//kGatGli/P4eG\nZgMAB46XUl7TfNpz70y09P+ZPaEP/3x4LMP7exPk44Svh468sgZyS+sJ8HLERqu2BqIWzowk0MuR\n1Lwaa/+nn9qVaFltLCLYnQkx/gzr782iWdFo1Coc7bVcNyKIvoFuNLYYKatpJtDLiWuGWVaF259S\nSpvBxJrd2Xy/PxcVMLhfxzYbarWqUyPowX0t+2w/bLme62KDrMG5WeP7MGVooPUc52rGqN4AlFU1\nEejlxP2/iUJnf+nkRdjbannk5gFMGRJ42gDUzwV6O/HygthzCqYJcHKwkcbgApBg1FVlzpzbKC4u\n5uWXXyAtLZWCgnz27PmRv//9rwAkJSXw3/9+RGrqcUpKitm8eSMGg4GgIEu038/Pn/T0NAoK8qmp\nqcFoNPbk5QghhBBCXPFKq5pobDH09DQ6MBhNfL41ncff2k1mUW2H97KL63hk2S4S2jNyztX2w4Vs\nTyjksy0nMJnN3TpGURQ27Lc01+4f5EZdYxsvr4zn442pLPsyiTaDmZvGhvDWY+OICvGgqKKRr3dm\nWo8vqmgEYFdSEZ9uTudgahmbDxZga6Nm4iB/FAV+OJDX5bnrGts4llONi86GabFB1gCLSqWyrgJn\nZ6vh0dkDefCmaGxt1EwY5M+oaF/uuyEKjVrF6h+zqGtsAyCnpI6Pf0jjy+2ZqLBk7JzMIBkY6smL\n94xgye+G4WCn7dBIPLqPB0P7eaPVWAJeD/xjJ2t2Z2Or1XDvzAh83M/edqN/kDsOdpYggYvOhjHR\np8rTPFzsuWNqP2up37mKCvHgmd8O5s3FE3l5Yexp+071pAAvJ+64tt8lFSQT4komwairiJ+fP//6\n1wfU19fzyCP3c/fdt/Pee2/j7m75ZuDk5Ex8/AEWL36Y3/52Dl988SlLlrxI376WmuabbpqDl5c3\n8+f/lpkzryE19XhPXo4QQgghRI+qaWhlza5Mmlstv6Cra2pj3d4cmi5Q8Ki4spE/rojj/bXHLsh4\nF0JhRSN/XnmITQfzqWts65TVs+9oCY0tRtbuzTnnsRtbDHy3x9KrqK7JQGpuTbeOS8muIr+sgWBf\nZ566bTCThgTQ0mZiR2IRTa1Gpo8M5saxIdjaaLhnegSO9lq2xBdQUduMWVEorW7GzkaDo72WrYcL\nWL76KEaTmWkjgrh5Qii2Nmp+PFLM/zaf4JudmTS1nPqF7MHUMsyKwvAIHzTqjj9ajR/kT3gvNxbM\niMTLzYFBYZ68+eg47prWH7Bk1pyc65rd2RxKK+PPK+PZkVCI0WTmxrEhnZpG+3s64uNhCSwN+EnJ\n3IAQD3T2WmaODsbH3QEfdweiQzz40/zhjI7uXqNvrUbNgD6WMa8Z1gvbC5y9Etnbg5CfrfQnhLh6\nSdj3CjR79lxmz57b5Xt9+oSydOk/u3wvNDSMZcv+ddpxPT09eeON5RdkjkIIIYQQlzOjycwbXyWR\nXVzPkQhv7rsxindXHyU1r4Y2o5mbx/ehrKaZlRtSuWVSKL19Xc75HBvi8jCaFFJzqzGazB2aO//a\nEtLL2Xa4kPnX98fDxR5FUdiRUMjn2zIwGM30D3Iju7iehPRyWttM1qbhqXnVgKVxdm5JPcG+zp3m\n3mYwdegZ09xqpLnVyIb9eTS2GPF0taeitoW4Y6VEhZw+g6a1zcTh9HLWtQe+po8MRq1WMe/acH4z\nNoRDqWUowKTBAdZj3J3tGB7hw46EQvJKG1CrVBiMZvr4u3D39f3ZlVSEyazgaK9l+shgbG00TBwU\nwKaD+dZ+TSazwq3tzbXjjlvKCWMjO6+m5uZkx7N3DOmw7ecBnhvHhLA3uYSdiUX8eKQIFLhlUijj\nBvrj5HDmFcMDvBzxdnegudVIWKCbdbwbx4Sc8bgzmTMxlF7eTkwd1usXjyGEEN0hwSghhBBCCCHO\n0do9OWQX1wNw4HgZza0mUvMsmTyJ6eXcPL4Pmw7kcTy3mq3xBdw7M/Kcxq+ub2Xf0RIA2oxmckvr\nCT1DVkl+WQNqteq0y6uXVjdxIt8yPx93Hf16WYIXNQ2tlFY1ER7k3mH/dXtzyS6uY8X64yy+bRD/\n23SC7QmFaNQq5kwMZdqIID5Yf4z9KaUkZlQQG+lDXVMbBeWN1jF2JhURVuHCxxvTuGlcH6bFBpFd\nXMff/neYa4YGcsukMIorG3nxo4O0GSwleXa2Gp6cO4jnVxzg0Ily5l0Xjo22cxAuJbuKD9Yfo7bB\nUt4W6u/C0H6neve46GyZNKTr/kaBXpZ7VFDegH17EM3HXUeglxO/vaZfp/1nTwglsrcHLW2WhuM7\nEgqZOSqYplYjGQW1eLraE+p/7sFGsPTPmTm6N19uz8Bsgt9dF87EnwTPzkSlUvHsb4dgMpu7vEe/\nhKerg7W/kxBC/JokGCWEEEIIIcQ5yCisZd2+HOxtNdw/awBvfpFIclYlWo0KJwcbCsobKalqIj61\nDLBkCymKctrVo0xmM+v35RLs40xMmKWJ9OaD+ZjMCk4ONjQ0G8goqD1tMColu4o3vk7C3lbLPx4a\n0ykw0WowsfTTBKrrWwFQqeAvC0fi46Hj3dVHOVFQy9zJYVw3wtIntLq+leziOgCO51bz108OkVlU\nh4ujLY/NGUiInyXwEhvhw/6UUuKOlRIb6cOJ9mDc0H5eHMmqZG9yMbsSizArCqt/zGJEhDefbU3H\nYDSz6WA+Ywf68eW2DNoMZgK8HNHZaZk4OABfDx0xoXoOnSgnOauSIf06NohetSvLmg01NNyL8TH+\nRPZ2tzbbPpuTAbuC8kac27OPfDwcTru/jVbNwFBL+dqxnGp2JRWx9VABmUWWexQb6XNeK4NNGRpI\neW0zvX2cGdfN5tknuTvb/eLzCiFET5JglBBCCCGEEN3U0mbkg7XHUBS4/Zp+XDMimMz8GtbtzWHO\nxDDqm9pYvy+XT7ecoK7J0juqsq6V8toWvN26DngknKhg9Y+WXknjY/xx1tmwLaEArUbNndf24901\nKaQX1HLdiM7HZhbW8vaqZIwmhYZmA0ezKhn8s+DNpgN5VNe3EhboipO9DYkZFexKKmJUtC8nCiwN\nyL/YloHOTsu4GH8S0y3Nx6P7eJCaW01mUR06Oy2L5w6il7eTddyoEA8c7bUkZ1XS0GzgeHuJ3pBw\nL2xtNOxLKUEFhPdyIy2/hte/TKKoohGtRo3RZObtVckUVzbh5WbP83cN7xBEi4304dCJcjbG5dGv\nl5u1ZC0tr5p1e3PQ2Wm5+/r+DOvvfc6fYWD7NRSWN+DRHszx9Th7g2+AabFB/JhUxLftn5efXmcN\n4v1SNlo1864NP68xhBDiciMNzIUQQgghxGXpaFYluSX1F/Wcn2/NoKymmcF9PRkfY2kMPWdiKP94\naAzXDu/FoL6e7XOrAkDvYgl2pOZWn3bMhPbgj1ajZldSEev35dJmMDNzdDAxYZ5o1CoyCmpQFKXD\ncTUNrbzx9RFaDSai23sr7T9W2mGf2oZWvt+fh0atYsGMCOZdF45apWJ3cjHbDhcCMKSfFxq1io82\npnIiv4aE9AoApscGc+e14QR4OvL4LTEdAlEn5zu8vzcms8K6vTnWa+wf5M602CA8Xe2ZNy2cR+cM\nxFlnY1217qFZ0ehd7CmubGq/f2GdsrliwvT46XVkFNbywocHSM2txqwofL41A4A7r+v3iwJRAI72\nNrg721Fa1UxBeQNAt1abA0vQaki4Jdind7Fj8dxBZ+3tJIQQojPJjOrCqSxb5Uy7iW6z3MfzyF4W\nQgghxFXop8GXn5dBZRfX8fqXSTjYaXjpnlj0rvadjt14IA9FsTS2PpdznjzX0exK4lJKmTmmNz7u\nOnYfKWZXUhEuOhvumta/w5xOlkuF+Lng6mRLbUMbNlo1syeE8u+1x0jNq2Z8FyVYRpOZI5mWEr+X\nF4xgZ2IRDnZaRkb54OlqyaQK9nUmq6iOsupm60pqiqLw4ffHaWg2MGlwAHMnh/HYW7tJyqigudXI\nzsQikrMqqWlopdVg4trhvfBuD7jEhOlJSK9gR3sPqN9NCycpQ89/vk/l/bXHqGloxdFeS99ervQP\ndu9y3iddNyKIg6llbDqYD4CPhw53Zzvcne1Y+uBo6343jQ3hk00nGBiqJybMk5Y2E+99l0JYgCvD\nwr06jWuj1bDkd8P476YT7Esp4bXPEhgQqie3tJ4QPxdiIzo3DD8XAZ6OVNe3ktZeWujtfvoyvZ+7\nfUpf9C72TBoSgIeL/dkPEEII0YkEo7qgVmtQqdQ0NNTi5OQKXOpRFBVG46UaOFNoaKhFpVKjVl/Y\n5WGFEEIIcWXJL2vA3dkOJwcb4lPL+O/mE9Q1tqECpg7vxW1T+gKWQMwX2ywZMs2tJlasP8ZTtw9G\n/ZPg0Nc7MtkQlwdYgjlRvU+/KhuA2azw8Q9pHM+t4pGbB2Jvp2H56qM0t5qIP1FORJA7iRkVqID5\n0yNwcbTtchy1SsXgME92JBYRE6pnYKgnKpUlM+pkoKugrIG/f57AtNhggn2caGwxMqCPHm93Hbe0\nr9L2U2EBrmQV1bHpYD7pBTWYzAqBXk4czarCT6/j1slh2NpoGNLXi30pJSz7Kon09vI7ABdHW2aO\n7m19PWFQgDX7aWi4Fy46W8YO8CM5s5L4NEuWliUj6+xFFD4eOh6/NYa/f5ZIq8FERJBbl/tNHByA\nm5Md/drfHxHhjYOdhmBfl9P2W3Kw07Lwhkii+3jwyQ9pHMmsBOC2KWHn1aMJsNy/7CpMZgVXJ1sc\n7Lr/Y5GHi731a1EIIcQvI8GoLqhUKjw8vKmqKqOlpfHsB/QwtVqN2Wzu6WmclkqlxsPD+7z/0yCE\nEEKIK9emg/l8vjUdjVpFkI+ztYG23sWe+uY2Nh3Mp4+/CyMifEhIr+BEfg0BXo6oUJGaV8PX2zOZ\nPioYk1nhhwN5bIzLs/Ym+mJrBn+aP5yUnCrKqi0ldj/NaFEUhf9tPsGupCIA/vFFInpXe5pbTYQG\nuJBVVEdiRgVuTrYsmBlJ5FkCW1OG9SK/vIHpo4LR2WsJ9nEmp6SesupmvN0d+GxrOnVNBr7ekUm/\nXpam5IP7eZ52vL6Bbmw6mM/2hELrtuLKJjRqFffdEIWdjeUXfrGRPuxLKSG9oBYHOw0PzxqAh6s9\nLrqOwZboEA/0LnZU1rUyoT3rSaVS8btp/UkvrKW2oY3BfU8/n58L9Xfl0dkDWLUr67QNuFUqVYde\nViqVioGh3TvHqChfQgNc+WJrOv6ejvQN7DrgdS4CvE6tOtjdEj0hhBAXjgSjTsPGxg5v70DMZhPK\npZp0BKjVKtzddVRXN2E2X3oTValOZppJIEoIIYToDrNZQaXqXJZ2MZnM5i6zYowmM1rNhWs52tpm\nosVgIjG9nM+3pmOjVWOrVZNdbFm5bcGMCKL76EnJqeIfnyfyyQ9p1DcZ2BCXC8DcSWG4Odnx0sp4\nNh7IY3N8PooCZkXBzkbDU7cN4ovtGWQU1PLKJ4esAa5PN58gPMiNkVG+uDvbsSupiENp5bg42tI/\nyI0Dx8uobWyjl7cTz/52CNnFdaRkVzFlaCDOuq4zon4qwNOR/5s3zPq6f7A7OSX17DlaQh9/F47n\nVmNvq6GlzURqe5nYoLAzBKN6uaLVqNBo1My7th99/F05mFpGgKcjwb7O1v0ie7vj6mRLc4uRx+bE\n0K9X10EbtVrFfTdGkVfaQP9gd+t2JwcbnrglhiOZlQzu27l07kwienvwf2cJ0p0PbzcHHpk98IKN\nF+h1qv+V7xlW0hNCCPHrUCk/74R4FTEYTNTUNPX0NM6LWq1Cr3eisrLhkgxGCXEpkedFiO6RZ+X8\nGU1m0vJqiOztfk5BpeLKRl78z0FuHBtyTn2OuqOiphmTWbH2HDqd/25KIz61jOfvHm7NHqqub+XD\n74+TUVDLolnRDOij7/Z5zYpCRkEtoQEuHQJcccdK+fD74xiMluxujVrFI7MHEtnbndySenz1Ohzt\nTzWG/nTLCbbEF1hfD+vvzaKbogHIK61nR2IRB4+XotGoGRHhzcRBAfh7OpJZVMsrHx8CwM3JlmH9\nvTmUVk51fWuHeTo52PDUbYMI9HLiPxuOczS7isVzB3UIWvxcd5+V7OI6Xvn4EGZFwVarps1o5tE5\nA9mZUEhSZiWh/i783++Gnfb4k9forLO19qY6nfKaZszd+Jyvdm0GEw++vhNFgVsmhnL9BX7eRGfy\nvUWI7rkSnhU3Nx02Nmdu0yOZUUIIIYQQF9jXOzLZdDCfe2dEMGaAX7ePO3i8jDajmY1xeUwd1qvT\nCmO/VEOzgRc/OkirwcRjc2KICuk6g0VRFOLTyqlrMvDNziwW3hBJSk4V761JoaHZAMA7q5JZfNug\nbpdKfb41nS3xBYwd4Mc9MyIAOJJZwQfrjmFWFEL8nNFq1EyLDWJgqCXIFRrg2mmcORNCaWoxotWo\nGRXlQ9+fZP0E+Tjzu+vC+d114Z2OC/V3Zc7EUCpqW7h5fB+cHGy4bUpf0vNr2JdSQn2TgWHh3gzu\n54m9reW/xvfOiMSsKB16UJ2PED8Xfn/HEP69NoWK2hYigt2JCdXTx8+F/24+wfiBZ/8aCfJxPus+\nAF5ukuXTHbY2GrzddZRWNUngTggheoAEo4QQQgghLqCGZgM7Ei29fQ6mlp1TMOpodpV1jIT0ckb8\nghXDGlsMaDVqax8hgLV7cmhsMQLw9qpknrptUJcBn6q6Vuoa2wDYl1JCbz9nvt6RicFoZsrQQNyc\nbPlmZxbLvjrCn+8dgYeLPWZFobSqCZNJoaaxlbhjpRSUNzJ5SAAezvbWbKbdycXEhOkxmf+fvbuO\nj+s88/7/OWdAI2aWBZZMMkOMSZzYYWybNNCmmNJun91tt7vb3efXLj1daJfaLbdJU0yaht04aMcQ\nM1u2ZFmSxcww0tCB3x9ndKSxZFuGxJDr/XrlFWnmnDP3GZB1vrru6zb5xcYT6IbJZ+6azQ0LzrxS\n23hul4PP3VN63s8HTFxNT1UUZuUnMys/+Qx7cMmCqFEleYn842eWs7einSWzrF6aCbFuu7pLvP/m\nFCTTM+CnMGtqQZ8QQohLR8IoIYQQQoizMEyT6qZ+AiHDrtwZ1Tvop7Z1kCUz01FVK7zYfLCZYMia\nelZR34c/qNkVN2cz7A9xqnUAVVEwTJNtR1pZOiudI9U9lOQmkBh39ulZZad6ePtAExX1veRnxPPN\nTy9DVRQ6+kZ451AzHreDdUvyeG1PA9997ihf/9gS8jIip6DVhnsqJca5GfAGeWZTNWAtZX/rddOs\ncx4KsOVQCzuPtXHvYmf8IgAAIABJREFUmiJe39PAC9tqJ4znqdcq7f5S65fksflQMz/dUIGmG/Yx\npxpEXStiPE5uXpJ3uYchwh5ZV8K9qwvPOfVRCCHEpXfpOlAKIYQQQlxjjlR38/Uf7+LbTx/mu88d\npaXLa9+3p7ydbz65lx+9fJyNu+sBqyH35oPNqIrCzLxENN3geG3vGY/f1OnlyY0V1LcPcqK+D9OE\nlXMzSU2I4kRDH//y64P88KVj/HRD+VnH2dA+xPeeO0p5XS+mCQ0dQxyt7gbguS2n0A2Tu1cV8OBN\nxdy5Mp9hv8Z/PXuEzn5fxHFqWwcA+PAN08lNs1Ybu29NoR1EAdyy1ApT9lR0oBsGmw82owALilNZ\nNiudz941h689vIjUBA+abrBqbhYfv20m65bkoukGiXFuvvbIoohjCnE5uF0OCaKEEOIykcooIYQQ\nQly1ugd8xHpcEcvWjzc4HMTtUqdUmXQ6TTf41RuVDAwHiY9xMTQSoq5tiNz0ODYfbOZ3b1ehAE6H\nwoad9cwuSGbnsTa8vhArSzNZPDOdquYBDld3s2x2RsSxTdNk08FmnttyCk03ONHQx/Qca9rcguJU\nMpKieXlHHfXtQwBUNvbT3jtC1iS9bUzT5Nl3qjGBB9ZOJyc1lu+/eIzX9jYQ1AwOVXWRmhDFrcus\n8OfBtcUM+zS2H23lb3+yG0WBnLRYvvHJZdS2WpVRJbmJLCxJo6lziLmnrZCWnRpLfmYcjR1eNu5q\noN8bZP70VL7y0YUR2/3TZ5dTUd/LwhKrmuyR9TOYnZ/M7IJk4qJdCCGEEOKDS8IoIYQQQlyVuvp9\nfOOJvZQWJPMXpwUhYPVd+sYTe1EV+LvHlk6pSXFNywBdfT5WzcviaE03A8NBFhSnsnJuJj/bUEFj\n5xCQzZ7ydgD+/MEF9A76+c1bVfz7bw9hArEeJ/euKSQpLgqnQ6HsVDcnG/uoaxvixoU5xHic7Kno\n4JlN1ThUxQ52egc7UYDSwhRm5ydzsqmf4txEVAU27Kxn+5FWHlpXAlgB1M5j7TidCqqiUNnYT3Zq\nDHesyEdRFHLTYjnVMkhTxwkAPnvXHNzhHlKKovDJ22fhcqgcqu7CH9Rp6Rpm25FWGtqHiI5ykpUa\ng6ooJBZNvmreitJMGju8vLKzDoC1iyZOt4vxOCNCOKdDnRDKCSGEEOKDScIoIYQQQlyV9lR0ENIM\nyk710O8NkHRaT6Uth5rtFeD+8/eH+fCN0zlS0wPA8tkZLCxJxeUca/I97A/xveeOMuzXUFWFncfa\nALhpUS4ZydYKZc2dXjTdoKHDS6zHafeQOnqqh7JTPczOT+Lz9861p/7MKUjhWG0P3376MGBVcj12\n2yzeOWQ19f6zB+ZTnJvI3z+5j76hAEU5CXbV0F8/uhiAvqEAr+5qYMexNj5843RcTpUdx9p46rXK\niPN96OYSHKrVgeGOFfk8ufEEQc3gtuumMee06iZVVfj4bTP5+G0zaewY4h+f2s9L79YS1AxK8xLP\n2bx7+exMnttyCtO0+kud3ktLCCGEEOJspGeUEEIIIa54p1oHGAmvBjdqX0UHACaw/0RnxH3BkM6m\ng80oijXtrWcwwBOvnuBAZScHKjv50cvH+davD9rNtCFyxblfv1lJeV0vyfFRzC9OITMlGpdTpbHD\nS2OHFUgV5SSgKAqKovDlD8/j6x9bzF89sjiiB83aRTkoCszIS8TtUtlR1kZlQx+nWgbJTIlh/vRU\nYj0uPnf3HKJcDtbMy5pw7snxUSwsScXrC7G/soOufh9Pb6pGUWDpzHTcLpXFM9IiAqEVpZnkpcdS\nlJ3AA2unn/W5zc+MZ15RCoGgDsD0nIRzvh6piR5m5FnTCm9YkG03KhdCCCGEmAqpjBJCCCGuYIGQ\nzht7G1lZmhkxzSyk6by+p5F501OnFB5ciUb8IV7f28i6JXlnbSJ8oqGP/3jmMCvnZvKFe+cCVoVS\nS/cwcdEuvL4Qe090sH5ZHm/ua0RVFAIhnaGREMvnZPC5e0p5+u0qhv0aK0ozUYBXdtTR2Ollb0UH\na+Zn09k3wuaDzUS5HCybnc7OY9Y0vBsWZNvVRnnpsdS1DXGg0gq+pmePPe8up4NZ+ckTxr5kZjpP\n/M3NKIrC05uq2HSgmR++dAyAtQtzUMIVSHMKU/jhV2+0V+Q73dpFuRyu7uaJV08QHeUkENS5Z3UB\nH7mxGMMwURTsY4E1Je6fPrsc0+SMxxzvzhX5HK/rDZ9X4jm3B/jIjdN5c18TtyyVRuRCCCGEOD/y\nZywhhBDiCvb2/iZe2VHHC9tO2bfphsGPXy7n5R11/Oatkxf9GCHNIBjSL/o4Z2KaJgPeAKZpRtz+\n9oFmNu5u4Nl3qs+6/5bDLQAcPNmFL2BVLu09YVVF3bEin4ykaGpbB/nJK+U8t+UUz75Tw8vvWr2M\n7lxRgNOh8sk7ZvMnH5rHkpnpLJ6ZzqO3zADgjb2N6IbB05uq0Q2Tu1bm88nbZ5GXHofLqXLDgrFe\nSNMy4gHYEZ6+N9pw/FxGQ6Lbr8vHoSoM+zUcqsLq+ZFVUGcLjeZPT+HR9TPITI7GF9AozIrnvjVF\n9n7KJNPqFEWZUhAFMLsgmZK8RKJcDopzpxZuzspP5s8fXEBCrHtK2wshhBBCjJLKKCGEEOIKFQzp\nbDrQBEBFfR+6YaAqCk+9VsmRmm4AGtqHzrjK2lT9xzOHGfKF+NbnlttVQBej3xugsrGPpTMzcDoU\nnn67ms2HmslOjWH1vCxuX56P06FyuKoLgP2VnXzkxhEykieew8Bw0N4upBkcru5i1dws9oan6C2f\nk0EgqPPHXfUcqOwkIcbFitIsDld3MWtaEgVZ8ZOOcea0JIpzEjjVOsi3f3eYmpYBMpKiuW15Pi6n\ng797bAnDvhCpiR57n2kZcQB2H6rzrUhLTfSwfE4mu8vbWTornYSYqYc4iqJw63XTuGVZHq3dw6Qk\neC7p1DhFUfjqRxfiC2jEn8e4hBBCCCEuhIRRQgghxBVq5/F2Bkes4GMkoFHXOsSwP8Su4+2kJEQx\nryiV7Udb2VvRwf3XF53zeBt31xMb7eKmRbn2bQPeADUtAwDUNA9MOtXsfByu6uKp1yvx+kJMy2ik\nJDeRLYdbcDoU2npGeGFbLSHN4PoF2TR2egEwTXhzXxOfuH3WhOPtOtaGbpj2inN7KzoxTege8FOS\nl0haYjQr52byx131REc5+MuHF5GfGW9XPp2JoijcubKAH7x4jJqWARJj3fzlwwuJCq84Fx3lJDoq\n8tek/Mw4++uM5Gi70fj5eGDtdFQF7llTeN77jo47Nz3u3BtegMnOWQghhBDivSC/cQghhBBXANM0\naegYsptIA7y5txGAlaWZ7Kno4HhdD40dVoDzyLoZTMuIs8Oo+9YUoigKgaDOv/zmIO29IygKrJmf\nzSdum8nxul5e2FaLokBJTiJ54Sqf6uYB+/EOV3czKz+ZX2w8we7ydkwTslNj+L+PLSXK7eBcNh9s\n5ndvVwGQmRJDU6eXpk4v0VFOvv6xxRimybd+dZDNB5txOa2qnhsX5rC3ooMdx9pYPCONGI+Lwux4\nVEXBME22HW0F4HN3l/Ifvz9MeV0v1c39KAo8uLYYgOzUWL7y0YWkJXrISYud8nO+aEYa03MS6Ozz\n8bWHF01amTVe3rgQ6EL7dKUkeHj8ntIL2lcIIYQQ4lohYZQQQghxBdhR1sZTr1dOuH1OQTJ3ry5k\nT0UHe8o76BrwkRDrZtGMNJwOlYKseBrah2js8FKQFc++yg6au7xEuRyYpsnWwy3kZ8Sx+WAzYFUh\n/WFLDX/58CIAuyoK4FBVF6vnZbH9aCtOh4rTodDU6eXdslZuWTaNxo4hqpsHWFSSFjF9DWBwOMgL\n207hUBW+/JH5LJieyut7G9hf2cnHb51JfqY1XW5FaQa7yzt4ZYfV0+n6+dl43A7e2t/Ef//hKACL\nZ6Txpx+ex+aDLXT2+ZiZZ4Vny2ZnsOVQC/6gzt2rCpg5Lcl+/PEryU2Vqih8/WOLAcUOx84mOspJ\nRlI0nf0+irKvzqbxQgghhBBXAgmjhBBCiMvMMEw27mkArB5Io32bHA6FO5bnk50aQ3J8FJ39PsBa\n4W20X9CKOZk0tA+x63g7BVnxbDtiVRJ95aMLcDpU/u23h/jNmycxgRl5iQwOBzle18vx2h7mTU+l\nurkfgOT4KLoH/PwyHIh95t5S8lJj+Icn9/HmviaWzsrgv589wuBIiN+9XcWikjS+dP9c3OFpbS/v\nqMMf1Ll12TQWlaQBcPeqQu5eVRhxrneuKGB3eQeabpIQ62Z6TgJZqTGENAN/UKe6uZ/D1d12HyeX\nU+XBm0oAWD03iy2HWsjPjJvStMSpcDnPXfE1XkleIl39PmZf5HRGIYQQQogPMgmjhBBCiIswNBKk\nZ9BPYdbZK2VCmk518wDFOYkTprwdrOqis8/H7PwkvnT/vEn3n1eUwrtl1ipuNywcW+FtRWkmL71b\nyzuHmsnPjKO2dZCslBhmTktCURTuWV3Ahp31ADyyfga9gwF++NIxnt1SQ3FuIo0dXhJi3dy0OJeX\nttdS3z5EXLSL25YXMOz1M68oheN1vfzrbw4yOBKiJDeR7gEfR2q6OVTVxcq5WbR0D7PtSAsxUU7u\nPUcvpLyMOBYUp1J2qodFJamoqkJctMvuF9Uz4OfffneQmpYBHKrCn35oHiV51qp1xbmJ/O3Hl5CT\nFntJm3efj0fWz2D90jy7mbkQQgghhDh/l+c3OSGEEOIacKSmm//v53v5518esFe9G9Xc5eXld2sZ\nHAkS0nS++1wZ//n7I3zl+zt48tUKRvwaYPWKei1cFXXXyoIzPtb86dY0tNLCZDKSou3bk+OjePCm\nYnTD5MmNJwBYuygHRVEAuGd1IWvmZ/HhG4ooyk5gycw0ZuYl0tI1zG/ePIlumMzITWTJjDT7mOuX\n5uEJN7K+MzymnkE/GUnR/OXDC/n0nXMA2BNe0e7l7bWYJty7pnBKTb0/elMxcwuTufW6/An3pSZ6\n+NrDi1hQnMqX7p/HwpK0iPtnTku6oMbhl0pctEum6AkhhBBCXCSpjBJCCCEuwOt7Gnhu6ykAHKrC\n05uqifE4WT0vm5buYb7z9GG8vhBbj7SSkxpDZWM/KQlReH0hdh5vJzbaxSPrZ1DR0EdD+xD5GXHM\nLUo54+MtmZnOx26ZYU+BG2/90jyO1nRTUd+H06GwZn62fZ/TofL43WMNsxVF4eH1M/h/vzpgh0kz\n8hLJSYslLz2WnsEAtyzLs7efnZ9ESW4idW2DfO7eUjxuJ6WFycRFuyiv66WmeYBDVV0kxrpZt2Rs\nlb6zyU2P42uPLD7j/aMNyYUQQgghxLVJwighhBDiPHX2+3jp3VrcLpUv3jcXBYUfvHiMJ149wbYj\nrXT1+/D6QhRkxtPQMcTgcJDM5Gj+9rGlmKbJ13+ym21HWrlndSHPb7ECrbtWFdjVTJNRVYVblk2b\n/D5F4fG7S/nO04dYNCPtnJVDRdkJrJybyZ7ycBgVntL3V48uRtMM4mPc9raKovAXH12A1xciM7za\nnNOhct3sDLYcbuGHLx3DBG69btp5918SQgghhBAfTBJGCSGEEOfpha2n0HSTD60uZPGMdAC+dP9c\n/rClhupma3W6GxZk8+k7Z3Ooyuqt9OEbi0iMtUKe6+dns+VwC//zh6M0dAxRlJ3AdbMzLmpMyfFR\n/NsXV015+wduLObgyS6cDsXuf5QwLoQaL9bjItYTGXCtKM1ky+EWBoaDREc5uGnR1KqihBBCCCGE\nkDBKCCHENa2jb4QdZW3sP9HJwHAQgBWlGXbfo/NV0zLA/spOkuLc3L58rOfRstkZLJ2VTnXzAF39\nPlbOzURRFJbOSmfprPSIY9y+Ip+tR1qoaxsE4JH1JWetinovpCZ6+JtHF2PCBTUDL8lLJDk+ir6h\nADctyiXGI79SCCGEEEKIqZEG5kIIIa5ZTZ1e/v7JfWzc3UBnvw+nQ0HTDbYfbaOzb+S8j1fZ0MeP\nXz4OwEduLCbKFTktTVEUZk5LYs38bBzqmf+JzUiKtiuhls5KZ0Ze0nmP5VIozk2kJDfxgvZVFYX7\n1hRSlJ3AbcsnNiIXQgghhBDiTOTPmEIIIa5Kpmna1US+gMZv36pi6ax0lsy0qpBCms7P/1hOSDNY\ntySXO1bkk5YYzYYddby8o47tR9vCq9AZlNf1sbeig64BX8RjJMdFsXxOBnHRLnYeb2dnWRsmsGZe\nFqvnZV3U+B+6uYSEGDd3rTrzCnpXurWLclkr0/OEEEIIIcR5kjBKCCHEZWOYprWSXGbcWSuJTtfR\nN8K3fnWAVfOyeHT9DJ5+u4rd5e1UNfWxqCQNVVV4aXsdzV3DzClI5mO3zkQNB1fXL8jmlZ117DjW\nxvqlefz3s0do6R4+42Ptr+y0v/a4HXzitlmsusggCiAlwcPHbp150ccRQgghhBDiaiNhlBBCiMtm\nw446NuysZ0ZeIp+/t5S0xOgp7bejrI1hv8amA8109Po4VtsDQM9ggON1vSTEunhzXyPRUU4ev3uO\nHUSBFQItLE7jSE03//jUPoZGQhTnJLB+aR4z8pIY37qprm2IvRXt+II6S2elc93sjAmNvIUQQggh\nhBDnR8IoIYQQl4UvoPH2gWYAqpsH+Idf7OcvH1pI8Tl6GJmmyd6KDgCioxx2ELV2UQ7bjrSy7UgL\nI34NE3jwpmJSEjwTjnHjohyO1HQzNBJidn4SX31oIS6nY8J2KQmeCc3HhRBCCCGEEBdHGpgLIYS4\nLLYeacEX0Fg1N5Pbl0/DF9D46YZyfAHtrPvVtg7SPeCnJDeRr350EcnxUdy7upCP3zqT+BgXh6u7\nOdnUT05aLDcuzJ70GAump1KUHc+MvET+7IEFkwZRQgghhBBCiPeGVEYJIYSYsgOVnfQM+rn9AldP\na+r08s6hZkoLU3h7fxMAd68qJDs1hu5+PwerunhmczWfvWvOGY+xJ1wVtaI0k5K8RP7zT1fbjczX\nzM/mjb2NADx0c/EZ+1CpqsI3P3VdRBN0IYQQQgghxPtDwighhBBT0j3g42d/rEDTDa6bnUFKgoeQ\npuML6iTEuM+5vy+g8f0Xyuge8LPtSCsAi2ekkZMWC8An75hFTcsAO8rayE6J4fYV+ew61s5b+xvJ\nSolhRWkmyfEe9ld2oioK183OAIgIk9YuymHzwWbmFCQzf3rqOcckQZQQQgghhBDvPwmjhBBCTMmL\n22vRdAOAxk4vKQkentx4gqM1Pfz9p5eRnRp71v2f2VRN94Cf0sJkYqKcNHV6+dAN0+3742PcfO6e\nUr73fBnPbT3FlsMtdA/4AWjuGubAyS5727lFKSTETgzAMpNj+PaXVhHrcUrQJIQQQgghxBVKwigh\nhBA20zSBiRVDdW2D7CnvsL9v6hhiQXEqx2p7CYR0nt96ij97YIF9jHfL2qhpHuCR9SXEeFwcPNnJ\njmNtJMa5+dL984iLnnxFurlFKXzzU8v46YZyWruHSU3w8Nm7ZjPs1zhS000wpONwqNxxlmmCSXFR\nF/s0CCGEEEIIId5DEkYJIYQAwOsL8b3njqKqCn/x4AJiPFZgZBgmz2yqBmD1vCx2HW+nsdNLR++I\n3Wz8cHU3Jxv7yEiO4bdvneRwdTcAmm7w8LoSfvXGSQA+e9ecMwZRo6ZlxPH3n1rGsdoe5hSkEOOx\n/qlaFp6WJ4QQQgghhLi6SRglhBDXGE032HKohYKseGZOSyKkGew83oZTVVkyM90OdwzDZHd5OyHN\nYG5RCj955Th1bUMAfPe5Mr728CKi3A7e2NdITcsABZnxPLJ+BruOt9PU6aW2dRCArJQY2ntH+MGL\nxxgJaJgmZCRHE9IM9lR0UNc2iNcXYt2S3Cn1cQJwuxwsnSXhkxBCCCGEENciCaOEEOIyM00T3TBx\nOiZf+W0yIU3H5XQA0DcU4FdvVDIrP4nbrpvGL147wZ7yDhRg/bI8qhr7aez0AvCbt06ysCSNZbPS\needQC1VN/RHHnZGXiGGY1LQM8J1nDnHd7Exe2l6Ly6ny+XtLiYt2kZboobPPR0V9LwAfuqGIt/c3\ncap1kMRYN6vnZXHP6kLq2wb5j98foaPPR1ZKDB+9ueTSPGFCCCGEEEKIq5qEUUIIcZkM+0NsOdTC\n7vJ22ntH+LuPL6UkL/Gs+2i6wSs76nh9TyM3L87l0Vtn8OTGCirq+yg71cPmg830DgZIjo/CH9TY\ndKAZgJnTkshIjubgyU4OVFr/AeRnxFGcm8j+yk4ykqP5iwcXYpgm//n7w9S1DdmVUg+vK7FXvcvP\njKd7wG83FJ+ek8BXHlpIe88IRdkJqKrVb2pOYQofvqGIrUda+fy9pUS5HO/J8yiEEEIIIYS4ukgY\nJYQQl0FlQx8/f7WCvqGAfdueivazhlEdfSP8bEMFdW3W9LjNh5qp7xjkVMsg2akxOFSV5i4vCTEu\n/vrRxThVhee3naIgK57br8tHVRUeu3UmR0/1cLiqi6yUGO5cWYDLqfLYbTOBscbl3/zUMk409LG3\nooNYj4t1S/PsceRnxHGoqouQZpAQ6yY1wYOiKBTnThz7vWuKuGd1oaxsJ4QQQgghhLBJGCWEEJeY\nbhj4gzqxnomNukcrm17b3YAJrCzN5MaFOXznmcMcr+2d9HimabLreDu/fbuKQFCnOCeBu1YV8OSr\nJzjVMohDVfjifXPJTo1h1/F2Zhckk5kcA8CX7p8XcSy3y8F1szO47rRm4KeHRQ5VZV5RKvOKJvZ4\nmpYRZ389PTvhnEGTBFFCCCGEEEKI8SSMEkKIS6ihfYifbChnwBvgnx9fTlpiND0DfvZXdmKaJgdO\ndlLXNoTH7eATt89i1dwswKo2auz00tE3YgdJw/4Qb+5rYk95O90DfhQF7l1dyH3XF+JQVb7yUTe/\nfKOSW5blkZ8ZD8DaRbnv+TlOyxwXRuUkvOePJ4QQQgghhLi2SBglhBCXgGGavLmvkRe31aIbJgB7\nyju4Z3UhT71+gor6Pnvb4twEvnDvXNKTou3b5k5PobHTy/HaXjKXxnCy0ZrG1ztoTeMryUvkwbXF\nzJyWZO9TkpfItz634n06wzGpCR5iopyMBDQJo4QQQgghhBDnTcIoIYS4QIGgTk3LAIZp8sbeRk40\n9OF0qNy5Yhqv721kb0UHy+dkUFHfR3J8FHevKiAu2sXSWek41MiV8+YXpfL6nkaO1/bgcqr86vVK\nTGBFaSYfuXF6RHB1uSmKwoKSVE409FGULWGUEEIIIYQQ4vxIGCWEEBfAH9T4p6f209Hns2/LTY/l\ni/fNJS89jmO1vTR3eXlmUzUAaxfmsG5J3pkOR0leIlFuB+X1vZTV9qCqCp++czar52VdkT2XHr97\nDoZh4nLKCnlCCCGEEEKI8yNhlBBCnEY3DFRFiQiBDNPk+S2nGBgO8NC6Gby0vZaOPh+FWfHkZ8aT\nnuTh1mXTcLuscGbl3Eye3+rl6KkeFAWuX5B91sd0OlTm5CdzpKYbRYEv3jeXZac1Gb+SOFQVh3ru\n7YQQQgghhBDidBJGCSHEOCP+EP/ym4MYhsnn7i2lOCcR0zR55u1qNh9qBqDsVA/Dfo2EWDdfeWgh\nCTHuCcdZPieD57eeAmBhcRopCZ5zPvbqeVlU1PfysVtnXtFBlBBCCCGEEEJcDAmjhBAfaCFN5819\nTSTHR7F6Xha/e7uKtp4RAP7tN4dYOisdf1DnWG0PCTEu5hensvNYOwCfuXP2pEEUQFpiNCV5idQ0\nD3DjopwpjWXZ7AyWzExHVa+8aXlCCCGEEEIIcalIGCWE+MBq6fLy0w3lNHcNA/DOoWbq2oZIjo/i\njhX5vLDtFPsrOwGIjnLy1YcWUZAVz6q5WfiDOgtL0s56/MfvmkNNywALi1OnPCYJooQQQgghhBDX\nOgmjhBAfOKZp8s6hZp59p4aQZjA7P4neoQB1bUOA1Zy7tDCF5bMzaO+1qqRy0mKJD1dBlRamTOlx\nMlNiyEyJeW9OQgghhBBCCCGuUhJGCSE+UAa8Ab73fBlHqrtxqAofvbmY25fnEwzpbNzdQFqixw6b\nEuOiSIyLuswjFkIIIYQQQohri4RRQogPjON1PTz56gn6hgJkpsTwxftKKcxKAMDjdvLA2uLLPEIh\nhBBCCCGEuPZJGCWE+EBoaB/iv39/FMM0uXFRDo+um0GU23G5hyWEEEIIIYQQHzgSRgkhrnmmafLs\nO9UYpsln7pnL2gVZGIZ5uYclhBBCCCGEEB9I6uUegBBCvNeOnuqhsrGfnNQY7r9x+uUejhBCCCGE\nEO85Y7gPreHI5R7Ge8Lob0NrKrvcwxAXQSqjhBDXlKGRIG/sa8Q7ErJvq6jvBeCh9TNwOCSDF0II\nIYQQ177A7mfQavcR88A/40jNv9zDuWSMkQFGNvwrZsBL7GPfQ41OuNxDEhdAwighxBXtZGMfT2+q\npiQvkVWlWRTnJqAoyqTbltf18sTGCga8wQn3zS1MZmFx6ns9XCGEEEIIIa4IelcdAEZv8zUTRpmm\niX/7LzD9Q9b33h6QMOqqJGGUEOKKEtJ0jtX2MrcohZBm8NMN5fR7gzR1etlyqIX0JA+r5mZx18oC\n3C4Hpmmy9XAL24+20dBh/aO0Zn4WK+dm2cdUgcLsM4dYQgghrk1aayUYGs68eZd7KEKcl1D9IdTY\nZBzpRZd7KOIqZYb8mENdABiDXZd5NBdPazuJ1nAYc2QAvfGofbs53A/pF35cMzBMqHoXrjk3oThc\nl2CkYqokjBJCXFFefreO1/c2kp0aQ2qCh35vkBWlmczMS2R3eQc1LQNs2FmPL6Dz6C0z2HW8nd+8\nVQVAdmoM919fxPI5mZf5LIQQQlwJ/Jt/jBnyEffpn6CoMk1bXB1Mvxf/299HTZlG7AP/fLmHI65S\nRl/L2NeDHZdxJJdGYPtTGAPt1jeKA2f+ArSGwxgj/Rd13OCJrQT3PQeKinvu+kswUjFVEkYJIa4Y\npmly4GQnAG2h90DfAAAgAElEQVQ9I7T1jJCa4OETt80ixuPk5iV5tHYP8w+/2Meu4208sHY6Ww9b\n/9B++cPzWDIzXaqfhBBCAGBqQUzfgPX1cA9K/EX86VyI95Ex1A2mieHtudxDEVcxvbfZ/toY7LyM\nI7l4pmlieHvBHY3nxs/gSMlHb68KV0pdXBhlDvcBYPQ0XoqhivMgfyISQlwxWrqH6er3U5AVz6Pr\nZ5CfGccX75tLjGcsN89Ji2XprHSG/Rov76jjVOsg2akxEkQJIYSIYI4M2F8bA1f3hZi4NpimQeDw\nq+i9TWfdzhgOh1CBYUxtYh9MIabCGBdGmVd5GEVwBPQgamwqrunLUZOyUGKTAC4+jPJ7gcjwTrw/\nJIwSQlwxDldZ89mXzEzn1uum8Y+fWU5JXuKE7dYuzAHgjb2N9vcSRAkhhBhv/AXK1V4VIK4NettJ\ngvufJ7j/xbNuZ3p7x74eF6oKcT7sMMrhxPQNYgZ9l3dAF8EIfw5GAygAJSYpfN/FhlFWz1mjrwXT\nNC7qWOL8SBglhLhiHK7uBmDxjLSzbjerIJmM5GgAnA6F1fOz3/OxCSGEuLoYEWHU1d8vRVz9jH6r\n3805K6PGTc+72Att8cFl9DaD6sCRPdv6/ioO5Uf/uKDEjP2RejSMuujKqIBVGUXIb63MJ943EkYJ\nId5Xrd3DvLqrnmF/KOL23kE/9e1DpCd5yE2LPesxVEWxq6OWzsogLlpWvhBCCBFp/AXK5ZqiYvgG\n8W//xVV9EXg1ClXtJHBk4wXvr3fV49/+S0wtAIAx0G59Hxw5636Gfwj/9qfs0GnC/eFQ1Bzqxgz6\nMA0D/57fozWVRWw3/oL4Yi+0r1SmaRDY+we0hsOXeyg2vbcF/7YnMQPDl3soF80YGcD0D6EmZaMm\nWX+0NQY70DtP4X/3l5gh/9n3H+rGt/XnYw3DR2/3D+Hb+nP0rvr3auiTGv0cqDHJ9m2KJw4Uh7Wa\nHqC1VBDY+4fzrm4anaYHkVMbL3iso+/tpmOn3W4SOPQKoVN7z7hvsOIdgmWvX/QYrhbSwFwIcUEM\n00TXDVxOx5TvN02Tn24op6nTy9YjLdx/fRHVzQNUNfUTCOoALJ4xtd5PtyzLQ1EUVs2VlfOEEEJM\nFDFN7zL1jAqVvUGocju4Y/GsfPiyjOGDxjQM/Dt/CyEfrpKVqHGp532MwJ7fo7dV4siZjatkJcFj\nbxGq3IqaXoh7zk1n3C907C1CldvA6caz+uMTxzbYZX9t9LVg6iFCZW+g1R0k9pHv2L//GBHT9K7N\nMMrorCV49DWUmhRi8xeiKJe/RiK4/3m0hsOoKdNwz7/tcg/nooyGKmpKHmqC9buyMdhJ6Pgm9PYq\n1OQ83PNuOeP+oYp30Kp2YvQ0EvOhv0dxWH/4DZVvRqvaid5aSeyD/w/FHfPenwxgDE9SGaWoKDGJ\nmMN9mIZO4MCLGB01OAsW48iaMeVjj07TAyuQdBYsvrixdtUTPPoaalMZzmnzx27vrid44CWU2GRc\nxSsmjkMLEtj1O5ToRDyL7rqoMVwtLv+nXghx1TnZ2Mff/HgX33xiH4GQPuk2P9tQzle+v5OqprFf\nosrremnq9BLldtA7GOCp1yrZUdZGZ5+PgeEgHreDNVOccudyOrhjRT6JcVGX5JyEEEJcW0YvXsC6\nCDNN8319fNM0CNXssb6Wyqj3jdHXDCGrN47eXn3++w/3obedtL4OX9Db/x8483RP0zTt1/tMlXDj\np4vqvc32+MyhLozOU2PH+gBURuntVQCYw70X9DpdamZg2K5Q0zsu/3gultE3PozKAEBvOWE/76FT\ne866/+hrYvQ0ETzwEhD5Hje9PVbo+z4Zm6aXFHG79b2J6Ru0P6fn04jc1AIwbpGAS1EZNToN1+ht\niaiys5+74b5Jq+/0rjowdBxpBRc9hquFhFFCiCkzTZMNO+v4ztOH6R0M0NnvY0dZ24Ttyut62Xei\nE19A43vPH6Wh3fqLw2t7GgB4/K45fOG+UlbNzeKzd83hB1+5gSf+5mZ+8NUbmZYR976ekxBCiGuT\nfRHvigY9+L5f1OsdNZjDVoXLe92zyjRN/LufIXDwlYs+ltZwGN+mH9rNjrW2k/je+t9zT1Eb7GLk\njf854xS198v4YONCQgXt1F7ACi713mZM07QvLs8WKhpddZhDVuWTOUloZZpmRIWe0dcSMb7RqTum\nrkWuBDk8tfdt8PjbBPY9N6VtL4TeVc/Iq99m+MV/YOSVfzln36tzHm/c66TV7L6oYwXLN+Hf9buL\nCpxDdQfA0O2xWa97i/WeHuo6x96TM0N+fJt+RKju4AWP60KNhiqOlDzUxNEwqnzs/o6aiPPSuxvw\nvfk9DG8Ppha0ghFXNLiiCR59Ha3tJEZPA+ZAO2pyLkpsMlr1rnOem6mH8G39OaEpvMaBfc8TOPzH\nse8PbSBwwGr2PzZNLzKMUsMNzfWOGghPPTw9UNIay/Bt+tGk/ddGp+gpiZmT7nshxo5hoodDZtMw\nwj9bLKOBWWDfcwQOv2rdFv5MnE9V19VOwighxFl5fSEGh62/GGzc3cDL79YR5XZw35pCAN7c14hu\njM3NNgyTZ9+xfpjOK0rBF9D5j2cO8+TGCiob+8lMjmbJzHRWlmbx+XtLuX5BNjEeF6qqoMqKeEII\nIS6R0YuX0V/s3+++TVrNWOXBe12ZFap4h9CxNwke3oBpXPhqUKZpEtj7HFrtfvviMbDvObT6Q2iN\nR88+hhNb0BuPEjq5/YIf/1KICKMuoOJmfD8Xo6/FChTDwdzZ3kPjL7aNoa4Jr4M50g96ENzWAixG\nTyN6ew04XKAoaKf2YhoG5kgfYKJ44sf2OwfTNAnsf5HgkY3vSb8jM+jD9/b30VtPYHQ3oHdUEzz6\nxoUfzzSt8EB1gOpAqz2AaWgXdizDILD/BULH38YcuPAg1A4KnFGYI/2YQ90ED2+w3tPVFxaWhU7t\nRavdF+5j9P5VZpqmgdZcDiioaYUocWkwbhqkI3+hPT57rJXb0BoOEzz2Fnp3AxgajuxZeK7/BGDi\n3/IzQhVbAXDNugHPDZ+y9jv+1lnHojUcQavaSfAcQbkx0k/wyKsED7yIMdKPMdxH8MBLBA9tsPqr\njVZGxU5WGQV687igrW8sUDIGu/Bt/hFa7T78W5+Y8DqMhlGOpByU6ASM/jZM/cLei2OP32J/Pfoz\nSG8/GTl1vLc5fM4bCR54AWNkwK5ac2TNvKjHv5pIGCWEOKNhf4hvPLGXr/5gB9/69QFe3F5LlNvB\nXz2ymA/dMJ15RSl0D/jZf2Lsl7PtR1tp7hpmZl4iX31oIXesyMcX1Nh5zPoF4fYV+aiqhE5CCCHe\nW8ZIPzjc9pSHyapV3iumoaHV7gfFgRKfDloQ0zdw7h0vgNHfRmDPs+FvdMzhC18NyuhtwuhvBayL\nc2OoC6OjxrrvHGGefdF1CSoLLoZVbaSgeOIxepvOazl7Y6Ado6sOJTELJT4dc6gLvWNs+tyZQkWr\n6mEfKApKQmb4deiN2Gb0+XPmzgXVaV14hnw4Mopx5JRi+gatsCfcL0pNL7KOPZUwarjXnpponPa4\nl4J/1+8wvT04i5YR+7H/Bocbrf4g5rjpTefDGGi3mmunF+HIm48Z8EaECed1rL4WOyzUwhfz532M\nkX70lhMoMUm4Zl5vHav5mN1c/UKrZUYDLnOwA6O7/oKOcSH09mrM4V4c2TNRY5JQHE6UcO80NTmH\nqOUftcZXMy54DZ+jdmqvPU3VkTUDZ8kqnNOvw/T2EKrcCig4i1fgmLYAJTYFva0KY7jvjGMZDeWN\ngXYM3+BZxwyAaaLV7rc+T+EKRaOvxa5qUqITI/YbDaO05rFm4aMVjaZh4N/yM6tiyuFCbz5OqGJz\nxP52ZZQnDjUlD0x9QtP28zX+/TIaMI0+145pC+xzijznfVZA63ChfoCm6UkDcyEEYP2V6vTG4X/c\nWc/gcBCnQ6G2dRCnQ+HPPzKf6TkJANy5soDjdb28/G4dLqdKQ8cQG3dZU/EeWjcDRVF46OYSbrtu\nGvsqOhjyhbh+ij2hhBBCiKkw9RD+rU/Yv9g7C5cQteIhCAyjxKejJo417z2T4PFNaE1lRN/6ZRTn\nWC/CUM0eQie2EH37V1DCFS1ToTWWYfqHrAs2lwdtqAtjoGPCFBPDN4jv9f/C9I010EVRcC+6G3fp\nOkwtiG/TD3HmzsU9/zZM08D35v9i9DSOnX9wxKq4cXkg5McY6ESNT8e/51lMbzeem7+I4pjar/zj\nq7n0tpMR1S+jU8y01hME9j6H54ZPjQV9o9N6GNdfabAT36YfEbX0/otuCDxVhrcH09uDmjINNSnb\nusDrPIUzb96k24dOvkuoaiee9V9CjUkiVLUTAFfxCoyeRrShroipNWgBTN8AxsgA/nd/heemz+FI\nzrGqHnwDOHLnosQkog12YAx2osan2buOTvFTk7JRBzsweqxpbo6sGagJGegt5Win9uDImWPdnpKH\n3lqJMdJvTcPc9iSKJw7Pykcmnve4KXOmtwdSphEsex29vQbPLX+CojoJlm9CazxK9K1/huJ0E6ra\nQeDAS2CaqHGpeNZ9ATU+fexUm8rw7/gNGBrmcB9KTBKeGz6N4onDWbDQCgwaj+Kafh1gBQD+7b8g\natlHIp5vwz+E/+0f4Jq9lqhZa6xt28IVIJklONIK0RuPEKrZgzNcsTOeaZoEtj8F7mg8qx6dcP/4\nqY56ew3MXkuw7A30tpN4bv0yijr23g+e2IpWuw/PTZ9HjR1blW00+HAWr8CRUUyoYjPBQxvsXkIX\nEkaNBlyjQjV7cIQDxosR2Pc8xkA7nnVfivhcB45sRG89QfQtX7Y/x86SVfb9akIG+lAXzuKV1tS9\nlDyM3ib0vhbUpJyxqagj/YTKNwFWhY6iKHiu/xTD7dWYI/1WwBV+7pzFKwiVvY52ai/uBXdY5z3c\nh3/rz3FOX46reDla01hFpd5Rg1q4ZNLz0sOht/Vc7YZxoa/e22yFslGxKE53xH6jDc3tPmuKCoFh\nTN8AWs0e9I5q1LQCPNd/ipEN/0Jg1zMEj7yGEhVD9O1/YTcvVzzxqK5o9JYKjK46HCl5U31JIhi+\nQUzfIGpyLoa3F72zDtPvJVS3H1QHUUs/xEhTmfWeGv/ePLIRgiPWc+744KwSLpVRQgg6+kb4+k92\n8+SrFfZtnX0jbD7YTJTbwbe/tJq/eXQx3/jkMuYUptjbzM5PorQwmc5+Hz986Tiv7mogyu3gC/eV\n2oEVQFJcFLctz+eBtcU4HfJjRwghxKUTPPCSNcVpuBdzuJdQxTt28KTGJluVKpwjjCp7Hb2pDK1+\nbJl5a+rTC+htJ+2+H1NhBkcIhBv7uuastcOwyfoN6Y1HMbobrLGP9Fn/eXusqSmGgdZ4BL3xKMGy\n162eQ70t6I1HIrZHC+AsWoZ7/u32eZqGQaj8bbTa/QQPTa2PlGka9rQdZ/FKgIgqgtHxa9W7MLpq\n8b/zE6v5L9jTesCq0jEDw4Rq9mB01+Pb8nMM74VXa52P8T1XRqe6nG2qXujku+htlfi3/QK9p4ng\n0ddBdeKaucaqkgB7euLotDljsJPQia0YXbVo4SbQo9UPzoJFYyuXnVaJN/q9mpCBmjx2oevImoGz\naCk4nITqDtjbKfFp1oV2YBhjoA2tagehsjcnrULRe8emBY1WVgXL30GrP4jefBxT1wgceAm96Zg9\n1mD5ZkxvD+ZwH3pHNf4tP4+YWhg88hrmUBfmcD9KdAKem7+A4rH6ejpLrPfHaOhhakH8m3+M0VlL\n6MTWiLFpVTvQ204SOPiyXVWm2a/TTCuodLrR6g/Z76eI5623idDJ7YSOvzVplZs+rhpK76jC1AIE\nDr6M1nAYo2dchUpnLYEdv0ZvqQhP1xo719HG0q7iFfa03sjVONsx9dCExz6bsYBrJTA2DfNimKZp\nhYp1BwgefHnsdi1A8NAG9Obj+Hf82q7KdBUts7dxTr8ONSkb16wbrO8LlwKgt1RETEWF8LmrTjts\nVjxxeNZ9ESU2Gff8O+ztXOH3gd2Y2zSsPwy0VBDY8WsC+14AXYOoWOuxzlK5Zt8XFYvRWYsRDrch\nHDhqwQlhPoAakxzxvSMchBq9zQTLrZ9fnps+hyNjOlGrPw4OB+ZwL0ZvM1pzeURl1GiIGjjw0gVP\nd7VXMUydhiOrBPQgI29+FwLD1muQVggON3pv01iQGhU7blr5B2eKHkgYJcQHXt9QgP/6/RG6B/zs\nOt5O94D1j9Hz22rRDZO7VuSTHB/F7IJk8jPjI/ZVFIWvfHQhX/7wfJbNzmDZ7Az+8bPLWVmadTlO\nRQghxAeM1nbSChBc0cQ++p84i5aBaaDVHwKsKRxqglXtcaYm4qPVNEBEFUxEQ+qzTC85nX/HbzCH\ne3FOX46zYIm9ktVkK7GNXoB5bvsz4j//FPGffwpHbinmSD96+0l7aoc5bIVUo9u7F99rbx//+aeI\nvvX/oCZl2edpDnZYF4FA8MirU+qdpHfUYHp7cGTPwr3wTvt2NTkHVIf9/I1OwzP62wjs/UP4PCKP\nr/e1jl1cBkcmXPy/VyLDqBnh2858ATz6muhNZYy8+u9gaEQtf9AKjEYrI8Ihm7NgEWBN97RXGjtt\ntT01ZdoZ32+jYagy/tgoODJLUNwxOKcthKCP0Ml3rWPFptgVKFrtgfD2ZjjkOO08xlXumN4eqwm6\ntxuAUM1e9JbjEL641turMEN+jO4GlJgk4h7/OWpqAXp7FcGy163jhVcUVGJTiPv8k8R94n9x5pba\nj+GctgDc0WhNR63wdf8Ldp8cvb0qYipjaPQ9PNhpV8/ZvXEyS1BcUTgLloAWQGuY2JfMrtYzzUlD\nYeu1UFBiUzAHOqznz25kHa72CQXwbfkpmAa4Y9BbygmFgwpjsBOjqxYlIRM1vcgKsMMVbUpUnFWp\nZhoY/RMX7Dmb0dXq3Atux5Ez2/5MXwzTP2SfW/DoRntaotZwFMJBnlazGzPgxTFtnh0eArjn3ETs\nQ/9mv6cc2bMA67UwwmGms2Cx1ccLcKQXRVQhOXPmEPfx/8FZOFblqKbmoyZmYXTXYwy0EyrfbDVJ\nd8eAadhhdtSyD1uPNa76KeK8QgGM7kaU6ATcpevGHnP6cmu/8BTO01fSs24bm7anRCfgzJkNQKhy\nO+ZQF2rGdBwp06znoHQd8Z/9GZ51XwSsz//4yihn/gKcJaswh3utqsALYP8sSM7DkRnuV9hRgxKd\nQNSqj6GoqvUzNeizpgRHJ0acsyOr5IIe92olYZQQ15DeQT/fefoQR2u6J9zXNxTgjb2N/PtvD/Kz\nDeUcqe7m7QNN/PvvDtI94CchxoUJvHu0jZrmAQ5UdpIcb1U0nY3TobJ0Vjp/+qF5/OmH5pGRNPVp\nDEIIIS6MqQUZeeVfzrp6mu+t7+Pf9uT7OKqLo/e14n326wQrt016v6mHGH7xH+3lxM3giNUPBBPP\nmsdQ49PsAEKr3Q9YFypKdCI4o6wLj0kqLyJW9Woqs/8iHho3Ze30MCpUux/vM3+N1hF5caw1laHV\n7EaJTcZzw6dQFAVlNIyapDLKrhDJHFs9yVm8wnqMindOm+JSbf8lfbLVlkYrcszBTjswUuLTwDQZ\n2fCvDD3xOMMv/zNmeMWwCWOpGa2KWmFdZCZZ0+qdJaut/km+QczgiBU6uDwoUXGEyjejjau2ceTO\ntc61p9HqteRwoybnobeeIHRi8tfVGOxi+PlvEji0YdL7R5mGzsjG7+Df8Wv7Nt/bP8D3zk8jniPr\n+ZlpBT4uD3rnKfucA4c2MPzKtzC1IGbIb/XxcseA4oDAMI6cObjm32Y9n+Om6ShRcfZrpHfV2w2S\nRyuSRi/mrZXLRl+HyBXY7Gq9xEwcKbn2YyjumPDzbFWZjPaaUuJS7Qvt0fczjIUcEcce37DZ24M5\n1G1Pc9LqDxGqfHfsOWqvRu+sBdOwwiCHE8+6L4DDRfDAi+jdDfaKgs7i5SjKxMtFxeHCWbgMdA3v\nr/4PoWNvgjsGJTHTep+Ez9UYaI/olRSq2UOov8OaspqYhRptVdG7Sqz3/Omr6pmmGdFk+/TQ0/D2\nhqdl5uLMs957gQMvjW0/umLZwZcxBzpw5C8k5p6vg+ogsPcP6H2tY1VRJSvsdhWjr7VVyWJVB002\nVc/ob8f79NcIVe2IvH2wE6MzHHClFY6rJNs74RiTMbUgI6//F0NPPM7QE4/je+v71u2jVXOeeDCt\npuJm0Gc/b6656+1juMI/R87EkTEdFNV6P4yuvJc9C0fefOvrKazopiiKfW7Df/i/BHb9DlQHMfd8\n3e6NpEQn4Jpzs9XDratu0j5jelctmLpVKRc+HoB7yf2gOsaal08WRo2bbqmm5NlVh1qtFdq6xk1V\ntLcL/0w2BzvHKqOirODOs+YxlNgUtFN7GPr54ww99aWIitnJGAPtDD//DQJHXrM/i46UvIjn0LP2\ns/b7ffzPFkfWDLsSFSL/LfggkDBKiGvIH3fVU9nYz6/eqCQQHPtl81htD1//yW7+sKWGquYB9lR0\n8L8vlPHMpmq6+v0sKknj7x5bigK8W9bK78Or4X3kxulEuRyX6WyEEEKcid5Rg95RbU8TOp3hG0Sr\nP0ioaud5Ty+5HExdw//OTzEHOiL6Fo1n9LdhdNcTKt9EqGYP/p3hpsrTr8M5YzUwNsVhtCJCiUm2\nLpimzYeQn8CeP0w47ugFrhKdCIZOqO7AhGW4J4RRVTswh7rwbf4xxrjpLVrTcQCiln0EJTw15Uw9\nqwzfIOZAO8q4i3LAmlqjOqwAQtfshr16e7VdBeLInPjX87EKrE77wtm98G5cC+6wQg1Fweistau9\nJjwPrVZ/G2fhEhRFwb30Q6iZJbhm3WAfW2sut6bLpOQRdeOnAfBve9IKgcLT2yBcYRby4cgsxrP2\nM9ZzVrl1wmOahoF/688xepsIHngRrfn4pGMbHZ/eUmFVPGhBDG8vWt0BqxLE78UM+jB6m1BiU1Dj\nUlFUB460QtCCdgVU6OS71pL23Q326+HILCHq+k/gyJ6N56bP2eGLmphp93RRU/Lspd8j3hfhi1lj\noB0lJslqgmxPCx2rjDJN0/re6UaJTsSRPRtHbinucPAFWP2SXJ6x1zMuxb74tt/P0YkYXXURDZZN\nQ8Poa7NXSzO9vZFVWVoArf4gON12ODf6Wo9+XhzJuVafNUO3Vk2r3gWAa9xF8unc825BSchEiU5A\nScwket0XceZb1WOjoeBoVZRrzs2gKIRq9tL16o+s8y1ebh/LkTcfomLRmo5FTI8ywtV6Y5+ByCq3\n8eGjPb0pMAzhUGm0WktrsColPddbfc7cyz4Megj/lp+O9Vcad66uOTehphXimn8rjuTc8GswMYzS\nmo9jenvwv/sr9L5W+/bTA67Rz3Sobv+UVmoL7H8BvekYONxgYjWLD/rs96xr1g04i5ZhDnXj3/4L\ntKZjEBVL1MpHcS/7CI6cOfY0vDNRXB7U1HyrYqvxCGC9z6MW34OaVoAz/Fk+F9esG1GT86z3QVwq\nnhs+jSOtAM/az+LImol7yf3WZzGzBAzdro4bTx8XyjuSc3HNvhHX7JtwpOTaoTiAGjtJGOWJs8Jk\nrGqk8UEPioIz3NNsvPGf0bHKKCuMUqJiiV7/J6jJuSieWAj5CZ5lxUDT0PFt+Zk1LXDfc2gNY8+l\nI7MER+5c3Evutz8bQEQ/KkfWDBwp4XMuXW//u/FBIWGUENeIAW/AXrGu3xvkzX1Wc9Pq5n5++OIx\nNN3glqV5fPNTy/jCfaUsm53BLcus7//sgflkpsQwd3oK/d4gta2D5GfGsWqeTLcTQogrkT1NyNsz\n6epe9oWTaWCMu0i61C7V1KvgwZcxeqwFMM7ULNgc13fIv+1JtOqdVlPl6z9lVzSoqfkwrgG5Gq4s\niVrzmFXJU7EZraks4rijF7RRKx4CQKvejd5cZlXNhMOB8atAmaZhTzcxBrvoefuX9n2jfxVX0wvt\n26zKLLd14TO+KW/HxKoosC6GnOGqgohx1R0MV4GMVdNE7OeJA3eMVZXRO/bXec/KR4h77Lt2hcBk\nFVqm34vR32pNIQsHIK7iFcTe/w3UmEQ7UBud/uhIycNVtAznzOutqoXAMI70IrtB8/hpWGr6dNSk\nbCsA6rd+TzH1EGZgODyFsMqubvBvfcKuVDjdaLCBoaF310c2ru6otqZwmWZENcLohanR22wtDx8O\n4vTe5rFKpYR03HNuIubev0UNrzgGoKhO+0JYTckdq3gKhMenKIBJqP4gmMbYRXBUrPU6DIytvGcG\nvBD0oSZkWtVyLg8xd/+N3b8HQHG6xwIEl8eqNBpXCaLEp+Eqvdl6Lqp2YgaGMU3DCtoMDUdGMaBg\neHvGBW3jKu4KlliBgBYkVLndun/cc+Waux5H7lxr5bKeRiskPcuqXo60AuIe+TZxj32XuIe/jTN/\noR2Sjk7VG63Ycc+/HUfOHMyRfvwNx1FTp+FefO//z957h8dxnWffvzMz27DoIAACBAkQYKdIkRKb\nRIoiVW3JUmzLlovkIhclLom/90scO/78Xo4TJ/abLy5x4iSWq2IrtiVHsixbxZJFdVJdLGLvAEkU\nohBt28yc94+zM7sLLAorSOj8rosXsbtnZs4MZnYx997P/WT2zbSUYOPa2Ade9Z/3RJ3g0pvSQtp+\npJsRc/zzLKssE9LCkjBwu1pwB08gT7Sp4PhClXsaXHwD5tQ56XPyKEbFDMyy2szyNXOJvvtvMUtr\nMdIlXp57KPs9zxf90sKWdGy13/tyBS7/mk4M4KQF15HeO+0j25XTLBQleus/Ys3IdF/zSz1LqlWY\nfEGpEq1dm8DMSxGmReiSmyl4xxcQgVDe9WcztJTVE1Ci7/6qL8KNhVFYTvS9X1PnwQe/6Z/TRkEp\nBTd/iWDarZXZVua6la6DTAzkdO8DCK/9GOG02J3jUMznjBKG7yA0yqep95J0swmzdkHenCn/Gu1t\n969nL8tvDRcAACAASURBVBPOm0f0vf9A9PbvIKJlqpHAkKw2/z3std8qF1y0DJDqi4tAWDkbzQAF\nN37eL1P0j1mOGDUns89rPjR8rpMcLUZpNBc4x0/EsB2Xx19pwXZcls2rwjIFj7x4mHv+sJtv37uZ\npO1y6/pZfPDaOcysKWbVgql8+p0X8cFr1GPvj/h1SzIfPO+7ajbGkO56Go1Gozk/8G/E7aSfBZNN\ntqDjuQPONMntG+j/6Wewj+48rfU4PUdJbv69+gM+Wo6M9+EOnhg2zgtmFpFiSLu9wus+kZOLIgxT\nlZ94j9Mih1FQmnHyPPMTP0g4201jzboMUVCKc2wnsUe/A6hv/SHXGeX2HIPEgBKFIiX0vfEEdvpm\nyu1qAWFilGS+zRdCKGdRMpYRMsjclFl5ymG8UhURKVHzKqxQ4hijl88YJdXgJLHTzhejPPO57ge5\nn8gTpD5K+R9kXFdemLd3MxW+/DY/X8ecOlttI6sTlNeNyw9E37cJp30/A/f8v/Tf/RmSr9wPZoDI\nDZ/HmrNaiRXP3T1MYJVOCvvgK/5jp3V3bnB1654sASyPGNXdgtuTEWXdrhb/OHguibz7nV7eKJ/u\ni4r+vk1Xnd88Z4031v99O0m/vMgLY/aO40h45WpGtFytJ+tG2qye7TuVkq8/RP/dn2Hwwa/htu9X\ny0ypRxSUIAe6fdEvsPAqPzw6MCsrnDt2AqyQEm/TCGEQXveJzPimlcO6LI9FRnDYq7KITrRiTGnA\nKJ3ql59iWhRc9afDOoZ553xy51NI6SITA6rUShhYTSvSQlrC70IoXVs59dLHRhRX+4JCYO4VGKVT\nkYM92GnXT/Z5IQyD8LpP+mKzNYoDzCirAQRu9xFVnptVtuWLQ4UVuMcPkXztQXVudecRuGZlrgG3\nv5OBX36B2JP/mXOuS9f1S6vDaz6MES3zzysloGZC8EW4kPCVH8usf5R9GIns612ECn0H2tnAL3NN\n52ZJO8nAvX9D/92fUTlTVhBjyvBokLHEKPW8mrdZPh0hhJ8RNVKpon+N2kn/szL7cyQzzkjnVkm/\n7A9UeXX/z/8f9R722oNgBSl4xxf9c9worxv12vH3yQpiVEwfcdxbAS1GaTQXEImkwzObjxJPqm+F\nNm5r5a//YyN/+b3n+eOrLZiG4ANXz+aaS6eTSDn88bUWEimHd66ZydtWjp79BLC4qYIFDWWsXzqN\n+fVlY47XaDQazblHum5OEGy+TmU5GTKn0JZ8PHilWPENd51y5yEAt+MgSElg/jo/VDefgOY5o4LL\nbyEw9wpCl9+W00LeI7sbUXa4bWDmMozqWSoM3AvjznLTCMMgtOK9GBX1qhPS9MUEL7pGbTtLjPJF\npOmLCV36J+pYHNqcaeldWpPTch1y85yGriefAGTVL8WatYrQylsRhpEzZlQxyhM7koPqm/ksB5Vf\nxpcnyN3J6m421noBP5dFBCNErv40Zt1FBOZekQnnVa9iVjcBGZEltWcjsSe/j4z3YZTWYlTOJLz+\nTsyyWiVsFVZg738ZO10m5mE3b4FkDFFU6c/Xac1cA5kSxtzjY2Y5o5wcgbbFPweMkpEFouDCq7Hq\nlxKYuSxzAwsQLCCQLjNz0mJsdumNkZUTJlNxP+fMGqGtvT/faQsJzFvrB8hn33ybU2djlE4lcNG1\nGBXTEdEy3Pb9xDf+t9pmeR2isBykm3G6lNYSWnkr1uzVmNMX5QgyZlUjwsiNYjCiZUSu+jP1+0y7\nsE4Go6AUUVyF23M0necGwYuuBZQwYDVcQuWNn8bMcwNuTp2LUTFDdeTb+hjx53+OjPdhzboMI1Kc\nJWao33Pytd8iT7QqwbOwQpWWLnsXgXlrMWvm+edoavsG//jlzLW4ksj6P8WcvpjAvCsYCWGFECVV\nqhzv6R+DnfRFWS/DKXL95yAQJvnG70i8+CtguDhk1S8BK4R98DX1ntnXgb13E6kdG/wxbvcR5X6s\nnuULKRlB9cgwAdWavpjgilux5qzBrJk34j6MxFDh9mTFx5PBqJwJoajq3pcYUCWZve2IcFHaKXcz\nwrCGLWeOQ4wKLrpelQumnXyBxderBhJZpaDD5uPlRsV6lcsxNNxtCtkdA9NB/MkY8Q3fV19IlE1T\n72FX/RlGSTXhNR9Odza9Lu+6/P2IlBBYdL0q586zz28l3tp7r9FcYPxu40F+v/EQ2/Z38vEbF3Dv\nU+oPsYGYjSslaxbXUFYU4qbVDdiOS2VZhBXzqymJBkdfcRrLNPir9y8de6BGo9FoJgy3q9nvqAQq\nI4Yh5TTZN97OSGVvjg3SzemYNF6kY6sQZFTgcvz5nxNJdyiCtEBmpyAQ8rs3jbg/aceTUVSJTDsb\n3K5myOrelTOuZCrBeVeOuL6cMq0hNy9mxQzctr04XS0YpTU5ZT4AgTmr/dwjIO2gEkPEKG+ZOZhp\nIcNu3YORFsay3UgeXoi507oHEYwiXQf3+EFEuAhRMrwkXlhBIlf9WdY+zfEdOKO1/s523uRkp5AR\nXfIHqQ93FeWuN9c9lJN5UtVIwQ1/lRlbVod7/BBGRaac0CiZilE503cIWfVLCV/3Fzk3vyJYQHj9\nncQe+gbx53+OWTMXI+268vY9tPzdKqPq2G6w4+rm1LFVDo1hQCDsl1WpuajfhdPVgsgqwXO6WvyM\nJTGKW8msnqWEhqzj4Ha1qPJDX1CR/n7749IlfU7bXlK7nkP2tiuBMZ1tNhLCMAmvzbhdRDRbjEqX\n81x+GwDu4AkGf/1lP/PGLK/DKazAbd/vZ0wZxVWYU+ohfb2YVU1qv6U74nlkTV+kMtZOEXPqbOze\ndpXn1nCJv88iECb6ts9RVFFIZ+fwUkxhGITX38ngA39L4sX71HtTtJzw5R/01wtgH3kTo7iK5OsP\nQSCczvlS51F2VzKjvA72v+QHqOcVfRuW5nSHG3GfyuqwT7RBSuXDOd0tSNfF7TuOCBdhVswgfPlt\n6txMl+F5Aqy/f1YIq2Ep9t5NqlNhcRWyr5PExl9i1c5X70dph6JVOz93P0iXmva2g2nlnBehJTeM\nOf+RUJ0DK1XXuTzvW2cSVYq5nNTOp0gdeMU/TqErPqJKNEeaY/Z1NYIYFZi1yheNAAINlxIYIzPL\nu0Yh7QrLE9QPYExpQJRU43bsxz66k9SuZ5B9x7FmLiN8zWdy38NCUSLXfnbU7YJyZoUv+8CY494K\naGeURnOBYDsuz25RbWVf2dXBN+99gxP9SVbMr+Jbn12dLrtTH7SRkMUHr53Dtcumj1uI0mg0Gs2F\ngZ+54eUZDXFGSemq7l5WUIVW5xGjpJ1k8Lf/wMCvvpiTwTJe3M5D4CQxqmepzkN7N/qlaqm9Gxn4\n779k4N4vMnDP/yKV1QksH57jySgs97uM5Z1z1rjRUDfcQoX/DslWyr6xg9zg3HwIw0CEC5HxXr+c\nJrPMLIyyGoxwIU77/kwp1hARCDI3PolNv2Tg3i8y+Ov/D1xHdTMbhxvBm58Xzj0S2WLU0MyX7A5S\n2Ug7idtxEBEqzAkLzkYUTfFDob2g7hHnmt7/ocfUc3qIcBGhtXfk3W+rZq5yBaViqjMXyolgH3oD\nAmGshksxKxuVKCCl6ppXPQtcG+ykOp5G5vZGBCOqxLG3Q7ngAAIRld3TcQAQGGm31XjwhCtz6mxV\nipkOTkaILEdY5lgnX7oPe/ezqq37CPs8Gv7Nd7AgZ/3qtRJCa+/IPC6bliO4iUgxIpjb4VgEQr57\nZDzd0k4F/1yNFBO64qMntc9meR2h5e+FdJ5SeN0n/EBnT0hzDm8m9th3QErCl31wxNLHnPKuSLFf\npnoqZMo161Rnzq4jyIFOcG0/2N6as8Z3vplT5+S9Tn3BxLCIXPfnBC/9E3CSfofIvKWm6SB9p+MA\nMtGvSvRGEE5OBS/nK1vEPVtYnkNy59PpazqSk4+XD1FYoa5Zcp2up0v2eZOdFzVs+0L45bGx330D\ne88LKqvwJM9tTX60M0qjuUDYvPc4vQNJaqdEOXZ8gL0tJ7BMwS1XNlEcDbJs3ug5BBqNRqOZHPjf\nns+8FHv3834reA/Z1wl2AqOqCZnoR55oQyYGcrr0JF653xdP5EC3X/407jlklaoJ0yLx4r2q3K5m\nbqY7XXEVsrcd+/AbBPJ0NPLwxDRRWOHP0ekaXqbnDnQBIqeVdz5EMELo8ttAmMNuFrwbLrerBena\nSqAY4qYZtr5IsXKfpGK4dlK5CMpqEeFC9Q133VwG977qt3c384hRVsMlOEfeRCYGsyZjElxy46j7\nkpl3HcGlN42ZLyKyvu0fKoqJYIFqr97bgXRdX7Rxjh8E11bC4gg3V8K0lKjTdzyv2JaNNWc1TncL\nwcVvy3k+MG8tTmczgXlrc7oHDiW47F2kdj2LfXgLbrxPdRVzUlhNKxBWELN6Vo6jTdopvxtYPlHR\nKK/D6e9U2UqGiTVjcbrENO6HDI+X4IL1yHg/gXlXIkwLo3QqbvcRlVeU5TC06pdiNa5Q5atmQHUo\nG2WfR0KEogRXvFc5WPIIEIGGS5CrPoB0bUQwkhvAPoJIE1p+C/bhzZhZ7pszSaBpJU7bHoLz1p3S\nPgcWXYtMDmIUTcHKckeKQIjQyvf6HSvN2vlYc0cur8tx71XPPi3hIDD3Cty+DkJLbyK24Qe4Hfv9\n8kxP1BBCKMExFB1xXmbdIgILr1GuyvLpGKW1pHY+jXN0B25/Z1a3zKbMfqeD9DNut1MX1fIRXPIO\n5VoapaTtTGFOnYsoKPVzzqymlWM6c4UQhFffhowPnJKLd8T15ohRI4vrAIEFV6kSymQMYQUJLr1p\nzGU040OLURrNBcLTb6jgzfesa2LfkRP8fuMhrl0+ncrSyBhLajQazfmBlFLlYaRLb04Gt79Lfbtt\nnts/XaSdQCZjI5YHTARO6x4QgkDjcuzdzw9zRmV3UpPxfuwTbTjdRzErpuN0Hkb2dpDa8lhmfH/X\nqO4Q6aSQgz05Y7LzeWQ6R8gr//L+D6/9GLHffSMn2yfv+r3yu8IK5WQKRNQf/tJFDvSoP/oNU4lm\nBSXjytgILrwm7/Oe88rpblFByHYSs+6iHDfNUESkGLqPIGO9OOnw5GzRIzx9HoN7X/W7FmaXlXgY\nkWIi13xmzHmPOAchCC2/Zcxxo5XpQVogbN+HHOjyg8fHyovKrLsaZxxilBEpJrLuk8O3HSwgsn74\n88PGmQGsxuWktj+p8qPSrdI9d4I5dQ5sftj/WdpJf9l8bh+zfDqOF7xeWoM5pV6JUeSW6owHo7iK\nyLqPZx6X1+F2HxnmQhPhQiLXfPqk1j0SoTEEy+Di6zPbjWZcgyOJFlbdRXmz1s4UIhjJ+/sf9/LC\nGNZ9zCO4+O0EF799fOspmqI6a9qJ03aBGUVTiKy/E1DvIW7Hfr+zZPZxNsJFhK/8eN51QLoMc/Xt\nOY8DTStJbn6Y5JZHM90ys744gPR5lhajRisrPRXM8mmYo8z5TCIMA6tppeoWCDmldaMRmLPmjM8l\np0xvDGHJKCgZV/md5uTRZXoazQTQN5hk6/5OEilnXOOPHh/gzQNdlBWFWNRYzrvWNvLF2y7h3Wsb\nx15Yo9FozhPsfZsY+MVfYR96/aSWc3uOMfCLz5N46b6zNLORiT/zEwZ++QXcdC7LROMOdCMHujDK\n63zRY6gY5XjfoJfXZToxHd3BwL1fIvbbfyT+1A8A6Zf9yDwB6B5SusQe+xcGfvlF360kpVTuLKE6\n13k3Y14wtpsO9TWrGjFKa5C9bXm742X2qTPTyl4IlV1iJ0jtepaBX3ye+HP/pbqSSTenDOlUEMEC\n1bHvRHumE9cYN6oi7e5wY31+cHz2MuHpWQ6TQBhRdHpzPB1EpEQdS8PMW3KXHawNqrW6nS6jHOs4\neDdv5jkp50mHBu/YgNPypsrlSbtkVFmRSDva6jArG8C0QBiqlGvovLODxbOuCQCj6PRu7D1H3fnS\nESvbGTVW577JjhCGn4N0JksSvfPHTmceGcUn5yodin+uv/lHIL8onHPOXuC/V79cN1J81tx54yG7\nO6Z2OU0c2hml0UwAP31kJ6/vOU4oaLJ8bhW3rGvKyXbae+QEz7xxFNt1OdGfZOfhbiRwxeIazPS3\nt3Omnz/f0ms0Gs148AKvndY9WPXjb5Zgt+4G6agW3ecw9FO6LvahzaqdePdRjHSnt4nEz00qqVEh\ntkL4ziIPzxllpJ1RAMlXHwTpYFQ2YpRUqxse6ZB4/ud+MHg+Utse94Nm7T3PY668FdnbprrGVTYi\nrJB/M+b2tiMdG9l/HBEtUyVVU2fj9hzDaduDkSekViYHIRlTZW/pMhqzvA63bS+JZ/8LpIPTvAV3\n3lq1T6cpRnnHxRnoIrXrGbW9EfKiPDwxSsZO4HYeVuuY0uC/HqxpAsNSpW5l085onsvJIoRQzgzX\nyVt+5glKbm87TFtA8o3f4x4/iFE5M6+Qk01w8dtVCdIoJZdnCjOdReamnWhW4wq/85vf0j4YST9n\nEl73SbXPgdCwdWUHM2eLuDB6J73xEFywHpKDBBZefVrrOVOIrDy10923yUBo1Qdw2/diVJ65L2/9\nkl4npR6fpLsu3/qM0lrcHuWszO/uyzqHT3N7E41ROZPQZR9QXUeHdHM8l3jdMd2ullEzozRnFy1G\naTTnmP5Yii37OglaBtKVPLf1GFv2Hee26+ZSXRbhtd0dPPTCQdI5qQCUFAa5fOFUbrysfuQVazQa\nzQQj7QRuzzHMrBv1nNfToke+bl6j4Ykr8kQbbqw3J4fE7TmmOpKlv9l0T7SqsN8xskrcgW5wnWEl\ng26sV7VsLq3B7T7id08amss0UbgD3QB+uZooKEUOdiNdx//D3u1WDiajvA7SYhTSQRRVUnDj5/1Q\nY8+h5glcMt6PjPdjlKrubk5Xi3KjGSa4Dql9LxJc8Z6cEj1Itz4vKEX2HVe/Wyn9b+/N6tmkdj6D\n07onb8ckTwjLKS/yxAKp3MMy1uvns4gxwsvHg1leh9O8RQV5j+CmySYjRvWqY2uY/jECMKwgZmUD\nTtvenJvGiWK0fC7fGXWiDafjgBIpzSCR9XeOWqqolq0cV6ngmUAIg8CsVSTT5XjWkHKewJBMHs9t\nkQ8/aFw6mOV1KnMsWADJwdMKtQaV6RRaeetpreNMIsJFYAbASZ3xbKELEWvqbDjDQe3GsMYAp3kO\nCYE1axXJV+4HRso9y+oQeYE7o4QQBBddP/bAc4DXHVM7oyYOLUZpNOeYV3a247iStRfX8p51Tdz3\n1D6eev0I//Gbbf6YUMDk1vVNTKssJGAZ1FcXYRi6Y4NGozm/Sbx8P6mtjxG54fNYdQuHve6Vk52q\nGAXKVWXMVC2b3cEeBv7nf2PNWELk2s8ikzEG7v9bzIoZFNz8pVHXGXv4n5HxPqK3fSsngyj26Ldx\nu1qIvv+f/JDk7LlPNDJd7iYKVIi3KKxADnQjB3sQhRU4XUeUQFdQihEuQgajKjfFSRJef2dOdy2v\n5M1NC22xDXfhHNtJ4W3fRoSiJF97EByb0MpbsQ+9gdO6G6d1D6k9LwC55SRGSTXOYA/O0e3qcfoG\nzRvjdwAcuj9+h7ys8qIpM9T8ouWY0xZg734Oe9+mYeNOlZySlyn1ed002fhlel0qN8oorxuWW2XV\nzsVp24tRMeO053c28cWo7hbiG94A6RC67LYRu+hNJFZajBKFFTmBzieLMC2M8lrczmaM8ukIITCn\n1OMc3ZEjKk4GfLdHz9ELXrQ4XzEKShDhItXUIBiBIflOp0IgLUaJaHneUmQRLVfbsZOnXaqsyeC9\n73mfp5pzjxajNJpzzIvbVZbGygXVREIWH75+Losay3lhayuOKykIW9x0eQPV5QVjrEmj0WjOL5yW\nrQCkdj+XV4ySWWKUlHLc3Y08pw+o1tcBT4zqagHHTrdoT7uiUnGctj2q600wf4MH6djpsGmJ29mM\nWTlTrbvriN9hzt7/kr9eNffzwxklB3sAdUMEYETLcVFimYiUEN/wfZAuwUXXASowNnLNp8F1lUsg\nC0/Ykf2dKgeqdQ/YSdy+DsxQFPeEEg2tOWsgEMZp3U18w10qZLesDmtGpiW3KKqCY7uwm9U54IXs\niuIqRKQY9/ghpJ1AWLnCT3YnPQ+zejahy2/DnLYAt+cY9u7n/HDwM+GMMoZ02RpzfFiJUc7RHepx\nnoDy0JIbIVRIYO7a057f2cTrtuekf0/m9MUE5q+fyCmNiFkxg/D6OzFKqk+79DG89mO4ve3+OR+6\n/Dbc44eGBY9PBsJr78AdPKHdHmcRo7xOiZnF1afVpc9fX3EV4Ws+oxyvedYnhFAB2k5qQkvbJhvB\nxW9DRIqw8rh2NecGLUZpNOeQ7r4Eu5t7KCsKMauuxH9+6exKls4+vQBEjUZzfiDj/TgdB7CmLzq5\n5aSLc3gLZt3Ck2o1fr4g4/2+YGAffA1pJ0BK7CPbseqXgGOrb5IBUnFkrBciRWqfpy1AWEEliDRv\nwayd54sWbqxXjQ1FITGggrPT+CHM/V1IJ5VxXEmJ074vp2OUc/wQGBZm+TTkQDegaqGd1t2+GOW5\nbwBSe19ExjKh22fbGSWli3NoM+5gNxgmgYZL/ZtJ++hOjLJajEgxblqMEtGMMwrUMUge3ozbeRiz\nZi6BRW/z123NuDj/RoMFYIVw+ztVGaJXkpjehhzsAWEiwoVYjctJPH+PEhQNk/BVd+acp14+jS/Y\nlGTanZvVs7EPvorTvh+rdr6/r+a0Bbmd9NIIIQhedK36OZR7Q31GnFGlNSAMkO64go1FROWJeJku\nebvUhaLj7vI1kYhQoXJzJGOIkMpeOhM302eLwOzLz8h6zMqZ/nUOqlTTHKMr4IWKWT0LLVecXTJi\n1Jlzn41WXgtgTWDY92RFhAvPm5LBtyq6m55Gcw442NrLfRv2cvejO5HAyvnVGOfxH38ajebUiT/3\nX8Qe+SZ285aTWs7es5HYY98h9eaTZ2lmZxevyxgAdgL74OvEnvge8T98F/vAq8OcRbK3ndSuZ9U+\n73hKraN5C7FHv61ybNJ4JXrW9EWIcBFuh3LYQHa5n1R5RekubpBbFuZ0NjP4m78n9ui31HJZwpI3\nTkpJaq8So0SkGLdjP7K/088yOtvOqORrDxH7w7+QeO6/SDzzExLp/BCn8zCx332DxKZfqXl4YpTn\njEo7hVJ7N5F842EIhAmv++SY+T+QLukprIBkLOd4uQM9SNdBxnrT39QbGOEizOlK3AsuuwVzSDma\nn5tiJ3Mfk1Wqd0SV8Nm7n1f7+vKvs5xR+R1PRkFJTq7PmShREWZAZbAIY5xiVEnO4wtZxBBC+Jk3\nobV3YBToZigazcni5SIa50FGnEZzIaOdURrNWUZKyfd/u522rkH/uVULdaikRjMZkckY9qHXAEjt\neQFr+uIxlsjgd5rrOnxW5na28RxL5oyLcQ5vJvHCPb4Tyj1+EDEkV8PtbffFCc9R5Turju7AK+bK\n7gyHncp12GSJT25vW04WlTcfaSeJP/l9cG1VjpaM+eWCoMQoKaUSn/o6MKpnYU1bqPKSAKt+Cand\nz59VZ5TTvl9tzwoSvOg6km/8DqfjoHrNK0FMnxeZMj0lInjijHP4DQDCqz80LJR9NERhOfQcxT68\n2X9ODp5QbjQkIkusCK/5CE7rbqymFcPWM9QhkP3YalhKYtMvVAD6snf7mVP2vhf9cr7RHE/m1NnY\nvW1gBoY5pU6V8FV3qvyncYgxYkgY/oV+Axpe+zFkXzvWjCUTPRWN5oLEmnUZYSFOqiusRqMZjhaj\nNJqzzO7mHtq6BplRXcgNq+opLwozo1q3ENVoJiP2wdfAsdM/v543I8dDJgdJ7X2RwJzLEVYIt9vr\nGJcuPZMu9u7nseqXIsKFSCmx927EnLbAv4FO7X8Zs7IBo+jclPm6J9pwu49iNQz/A9xz1oSWvZvB\ntr2ZkjxUVzajRAUFq+5vPbi9bf4yXoC2///xQ8hUAhEI+WKUWaZCo+2Dr+K07sGqnY/b25E1t3bV\nHQ1ACJy2fUjXIfHy//jHFpQI5mZ1xpOxE8i+Dt8VFWhaiVmXEaPMqbNxju7E7Tmak0Pl9nWQ2rNR\ndY4rLMeafTnCMJHxfuzmLViNy3PK2GQyRmrnU8hUctixS+15AaRLaNUHCMxfR/LNP+J2H0FK199/\n94TK2ZIDPapbVlDlCmaLONbMZVgnWdZkFFbgAE5LpomGHOz2g9K9bCo1thxjSFcz/7XizDkowkU5\neV1GcRVGVSNu+36cQ2/4nfFkrBcZU+eJV3aYD3PqbOzdzyEKK85YSZlZNg3GmRckAiGwgsr1FQgj\nCscv9p2PmGW1UFY70dPQaC5YhGGcsRJSjeatjC7T02jOMk+/ob7pv3bZdFbMr87JitJoNJOL1N6N\nAIiiSlWqduiNvOOklMSe/D6J5+4mtf0plZXkiQ5pQcU+8Crxp39E4tXfACocPL7hLr9cy2nfR/yJ\n75HY+IuzvVs+8ef+i9gf/gW351ju/jgpnI79EIpiVEwn0KicM8El7wAEbleL7ywy07kXztGdKqeI\nTLC571iSLk77PjUuyxnllVQ5bXuUUJPlhPKdUcLErJkHdoLU1sdIbX0MggVYDenQ8952fzte+Vdq\n9/Okdj0LwsBqXIFZWotROVOta+pcv4TMzSrVS7x4L8lX7if56gPEn/4RyVd/g3QdBh/7jvo9Dfm9\nJLc/SWLTr0i++sCwf7K3DXPGxQTmr1NlVOXTwE6o0kOvk6CdUM6uRD+ioNQXZYyiKWBYiIJSwld8\n9KTFGm/fssVDOXgCOditXh9FJMpZTyiq2sqTCcnOJtCkRKz4sz8BpLpG1NYQkZJRc9Ks2vmqpG4C\nw6a9Uj2jbNp5nbGk0Wg0Gs2FgnZGaTRngd8+f4A9zT284/IGXtnVTiRksWyebrGr0Uxm3FgvzpHt\niEgxocveT/wP/4q9dxOBppXDxqZ2PIWTLotyWnepsqfEAJB26qTiuMcPqteP7VL/p90k3mPbe3z8\nTBBtNgAAIABJREFU0FndLw8pJY43p+OHclrBux0HwbExpy1ECIPQyluxGi7BrFtIav/LyN62jMOp\ndh723o04rbszy6e7ueVkObXtwaydpzrpBcLKFVNQCmYQp22vCiF3kmCFwE7gdrUgB3sQxdWYNfNw\nju4g8eK9AITXfBgZ78M++KoSrdKiUmDWKpKvPei7oIKX3Ow7gSLX/QVyoFu5gQrLcUiLZeXT1LE4\ntguESWj1bUqYeuN3yjmWzs5KbX8Sa8YSv+Oc93sLXvoujKECjxXEarjEFznM8jrctr04XS0ZMYpM\n6WF2aZkIRSm4+Usq2+kUumfllMeZFjg27mAPbtoZJU4iU0gUVyLjfXmdelbTChKbfpEu/4PwlR8j\n9si3VHeoMXKgjOIqCv7ky2ekk96pIiJFyL6OCzovSqPRaDSa8wntjNJozjAdPTEefO4Abx7s5v/8\n9+vYjuTyhVMJBXRvE41mMmPvfwmki9W4Amv6xRCKYjdvQaZFJg/3RBuJTb9Q3byEidO6B7erOXdM\nb3vGKdXVgkwMZIK2B7pw+zt9MUflIA1ytpGDPb5g5nYfAVQwePyFe/ywbc+5JIIRFTguDP/m3T7y\nphpTNUuVmWWTikNyMCck3Gndg+w7DnYCo7wOIQTCtDCrZkIyhn3wdbU+z2nVqkQgo6QqJ5TamrWK\nwKxVfqC2PJFxRllZQqFROZPgJTdnHkfLMKsa1f6kxRKvvE/2tqu8oSn1BBdcRfjy21TnwP0vgRkk\ntPJ9AMSf/hEy3q+6x7XtBTNAcMmNBOatzf03axXCCma2XaaOmXNke2654zH1OxfRXIHIrGo85S5z\n2UKQWdUEQiAHe9IdBzNB6ePBO8ZGHmeUUVDq/648wdDr8meMQ2QyqxonNGxbhFVuVL5OehqNRqPR\naE4eLUZpNGeAZzYf5bfPHcB1JY+9dBgpYU5dCaahvuVeu0RnM2g0kx2vJM9qWokwLQINl4DrYKdb\n3Xsktz8JdpLg0pswKhuUY+fAq+pFQxmW3d72LEeMxD620w+yBuWyye5e53YdOXs75m+jOetnNbfE\ni78ite1xnPQ+WtMWDlvOD3tOxtTjooqccGtPeHFPtCHjfYiiKWAp91Nyy6MAmFPq/fFeZ7bU7ufU\n44rpKmBaOmp9xVVKRApEENFywqs/5D8P6WPb3wmBMEZpjRIXrCCR9XcijPyGcU/o8UQsP6w9LXpZ\nc9ZgzVwGQOiy9xO8+O1Yc65Axk6Q3LFBBbMnBzGrGhHm2KZ0T/Cw97+UfkIt47Qqd9XJuJXG3FaW\nGGVUzEBESlSZXlqMMgrGV6YH+OKdWdmY9/XAnDXq/7lrEEIQmLvW3+75jhJVxbi672k0Go1Goxkb\nXaan0YyDnv4ElmlQGBmeabHtQCc/fUSVy7R2D/Larg6CAYPP3rKY3oEkfYNJpledme4/Go3m/ES6\nLk7bPrCCyrmDctqw61kl3KSFCsgEfQdmX460E7jt+/zwbHPaApzmLSrAO6tkLbX1cXBtMIPgJEnt\nfMZ3KYHKVTrbN8nZgpfT1ZLe571KyLn6U4hICWblzGHL5ThJQlFEIIxRXIXbfQRRWIFZPQu3u8U/\nLkZRJRRV4hzdQWr7k4hQIcGlN/mrMKvVfnpljEZJNUZxNU66/MsorkIEwkRv+aoq70t38RNFU1Sw\neechSMX97J/IDZ8HJzlqCLzvjPLEqLQrzXeCCUH46k/jnmhV4dBA8OK3Y+9+Fnvvi/4cvLmPhecm\n80ra/PMi3W3wTIpR2ZlQRnkdonUPcrAHJy0+nowzKrDgKsyaeZgV0/O+bs26jILSWl98smYspuA9\nX8vrpDrfCF5yM9acyzFL9ZdLGo1Go9GcCbQzSqPJw6u72jlwTN0EDMRT/O8fvsjX7n6FlK2+ed+y\nr5NnNh+lrXuQH/9eOQIiIYtNb7aRtF3WXlxLYSRA7ZQoc2eM/1tljUZzdpDSJfH6Q6QOvjrm2NTu\n54k98T1iT3yP5JZHxrV+t7sFUjHMqibfXeOJMNmZP9JO4B4/hIgUI4qzysnsBABWwyXq4cHXgEzA\ntnNMCd6BeVfkPPZez+4WdyZIHXiFtvv/mcE//BuJVx7ICVgHkH0dOO17IRXHrGrCql/qu2KGki1G\neeVYIu1SMqfO9nOAPIFHFFb47ieA0NqP5pRnmdVNQCZAWhRXI0qynFZeqVhxFUakODPOtFQXtLRD\nyxOYjIKSMbsRGtF0yLcvRqWdUVnikjAMX4gC1bHMqJiO291CasfT/v6OBxEuzBGcvPPCn8+ZFKPM\ngHKWoUQwb7ueE268AeYAwjBHFKJAiXZmZQPCyPz5aZbXjRpefr4grKAWojQajUajOYNoMUqjGULv\nQJJ/f2Ab3/rVGwzEUzz52hEG4jbtPTEef6WFfUdO8N1fb+Gnj+zkb76/iZ7+JKsvmsrf3HYJBSGL\ngGVw/fLzv+RAo3krkXrzjyRf/h8Sz/0MKeWI46RrE3/ubuz9L2Pvf5nEpl/hDvaMuf6MODHLf87r\n/JUTQN2+H6SDWT1b3ZhnO2VCUcwaJcJ4mUyBphUQjPhDrKZVORk/wQXrhm3jdHHa9zH4h+8xsGMj\nqf0vk3ztQdyO/f42jLT7KbV9g9rPMQQWo7haBWMDIqrm7u2nVX+JXybmh3MXVvih34G5awlkucpA\nBXbnCFzFVb4ABeQIU8PnkiVaRccfhu0JMm5/F268D7fnGKK42g87Hwkr3UHO7TwEiJzzYyyy99Gq\nW6hccd58znB2klk9CxEpxiirywhdjg3CPKVQdI1Go9FoNJqx0GKURjOEY50DSGAgbnP/M/v54yvN\nCAGmIfj9xoP84KHtuFKyuKmCaNhiankBH7hmDnVVhfzDJ1fyt3csp6IkPNG7odFo0jjdR/yuanKw\nB9nXMeJYt7MZ7CRm9WysxuVq+bTQNOo2PDEqy9EjQlFEtBy3tw1pJ/OOMyLFiJKp6rnyurRDJ+P6\nMSqmZwQMw8KcUp8jYFmzV4NpqbK5UUS28SJTCWIb7gLpUr7+dkLp8rjUnhdwe45AKOrnQtkHXh62\nz/kQhomRdpR4wpNVfwnR275NoGmF74zyStJEYTlmVRPRD36T0NqP5l2nL4BZISWi+CKTGNXllC1G\nnUxnNmEFEZFi5EBXVone2MJSdidFo2yaX643HnwxKhBGFE7JnXv0zIpR4av+jOitX0cEQjnrFgUl\nCKH/VNRoNBqNRnPmmfC/MO655x6uuuoqFi1axK233sqWLVtGHGvbNt/61re46qqrWLx4Mddddx0/\n+tGPzuFsNW8FjnVmulJteO0IvYMpls2tYt3SacQSDu09MZbOnsLn3rOYf/mLK/j7T6ygIKy+9S8p\nDFFTMf6bDY1Gc3aR0iW+4QeqfXxEuVhGE5d8oaHuIqwZS8Yc7y/XtgflfGnKed4orwMpcXuO5a4/\nu9tb+mejbJoqmcoSSYzyOl98MitnqlIhf3wtRqQYo3QaJAaQfceJP/9zUvteGnO+Q0lufYzBh77O\n4G++ijzRhlW/hJLL3klw0bUghMqocmwlmHkiiWODEKoD2xgYaZeY5+oSQmCk3UZDu8B5j43CihGF\nEP8YlFSpdaWFGlFYPmrJl5FdzneS3edEYQW4DokX/js9h9FFOACjaApGWkw82Uwvz1nndRLMmfsZ\ndkYJK5jJ1irIFaM0Go1Go9FozgYTKkY9/PDDfP3rX+czn/kMDzzwAHPnzuUTn/gEXV1decffdddd\n3HfffXzlK1/h4Ycf5nOf+xzf/e53+c1vfnOOZ66ZzHhiVH11kf/cDavquXl1A9GwRXE0yEfeNk/d\nHBgC05hwTVej0YyA29mMe/wgRkU9ocs+AIwlRmW6pHnigVc+NuI2+juR/Z0YFXWIYEHOa2ZWbpQf\ncm4GMaZkSnk9B5bf5t4LczYtjOJqlRdkWJlx0xeDFcKaqR574lDsj/9O6s0niD/5fZx0uPd4SO1/\nmcTGX+Ac24XbfRRRVEnkyo+p97iCEszaBeCk1LbK6jDKM5lARvl0RFYZ4UhY9UvV/zXDBRxVApeV\nATUOx5JZu0CVNaZdWkbZNERBad5ufjnbyi7nO0kxyhMFZX8nmIExt+URXHgNgN9tb9zbq50HVtA/\nL7ycLcwADDnPziTZQteZFr00Go1Go9FoPCa0m95PfvIT3ve+93HLLbcA8NWvfpWnnnqKBx54gI9/\n/OPDxm/evJlrr72WK6+8EoC6ujruv/9+tmzZwjvf+c5zOnfN5KW1S4lRt103h188sZup5VHqpyph\n6u8+vhLDEBRHg6OtQqPRnCd44pJVfzFmzVz13AjikpRSjReGcvtYQUSkRHW2S8URgfzlt/nCrD08\nR5DT1YzRPV2FnNfM80POQYlLhR//ISKdq2QUV+Ec2Y5RWqsCocvrKLzjP8ALRi+upPAj34O0EG6W\nT8MG3I4DIEyQDvEn76Lg3X+LsEZ/r3IHe4g/+1MAwtd8RoVLF5RiBDLLBWatwjnyptp2+TSM0mow\nTHCdcXeHCzStwJp5Sc5+ewgzgCgoQaazuYzo2CKRUVBC4Yf+Re0vIAIhoh/8Jggx+nJZYtTJOqNC\nl32A4OLrQbqqBHOcglBg1iqshkvG/F0Mm2tRJYUf/Xd/H333V0EpYoz9PB1yyvROIrxco9FoNBqN\n5mSYMDEqmUzy5ptv8qlPfcp/zjAMLr/8ct544428yyxdupRf//rXHDx4kIaGBrZu3cq2bdu4/fbb\nT3kehnH2/qA7F3jzv9D343yitWsAAdRPLeIrd6zIeU1nQV3Y6Otl8uD2dxF7+se4/Z0gDMKX/okK\n+x46rj0tRtXMwSoqRxRVqnDw5ADGkGBmt7cDGTuBWTkTM6SudbNmNvb+V5DHD2AfP4xz/KByDVlB\nkjueIrH1ceTgCX8bQ88ta4pyEcnuI7jHdo04DiNTWmaWVJMCzIq6zDgjOPL4rO5lkav/lNSuZ7Gb\nt5J86T4iazKfj9J1iD/3M0S4iNDydwOQePpHkBggeNG1hGZljl/2tRJsvJT4s3eDa2NVTMe0Ahil\nNbhdLVi1efZlJIxRyucKK3AGexDhQv/Yn/T68ghdQxGlmWwps6gccVLvBQKKp5zE+CyCoVNbLmsf\nrdKpJAAjWnp238OyBKizvq1JgP5c0WjGj75eNJrx8Va5ViZMjOru7sZxHKZMyf3DrqKigkOHDuVd\n5s4776S3t5frr78ey7KQUvLFL36R9evXn9IcLMugomJydIkpK9M5RWeCRMrh+Ik4lWURaqfqrIzJ\nir5eLmykdDn26D9jN2/1n3N2PMHUFVcNG9vftheEQdX8izFCBTj18+nf1kF08AgF0y7NGdt35BX6\ngOjMhf5ng9W0iM79r+Bs+R3xQ8odFCktI7pgNSeeuRukC4AZLaFq0QrM6BCBq2Q2/cLA7TxEMl06\nV7H4MsKjfPYkF63k6OaHKb94LdFxfEY50YtpebaC6OxlTFl5NfaCpbT84H+R3PY45Ysuo6BRlXl1\nP/drkukOeMV19biJGHbzVgIV06i94Q6MwHDBpKwsCmVRuPQ6Yvs3Uzl3IUYghLHoCvo2/1Htc8Hp\nf47a5VUMtO8jUFJ51j+X7dmqXG5K1YVVguZELyL5Qg2F81ZQdhaPkSyL0CcMkC5FVdUUT5K/k842\n+nNFoxk/+nrRaMbHZL9WJrRM72R55JFHePTRR/nOd75DY2MjW7du5Rvf+AY1NTVce+21J70+23bp\n7Y2dhZmeOwxDUFYWpbt7ANc9/U5KbzVe2NbKzJoiP3S8ub0PKaGqLEJnZ/8Ez05zptHXy+QgseUx\n4ge3YpTXUfgnX6LvF18gcXQfx9u6ckqh3L5OnL5OjIoZdPe70N+PXd4IPEP3ns3EyufmrDe2V4lb\ndkmDf/3bRfUAvhCFFaT35Yfp2/YcSJfI2jsIzL4MTIueuAnx4e8bRkm1H2AeXHg1A+FaBkZ7fzEr\nKPrIvxEH4uN8H4p+8JsIIdLzDhJe81EGH/832h78LoW3/gNubwcDz96r8oYcm45HfqCENMMktO5O\nuntTQCoz5yHXirj0fRRc+r7MuPlvIzr/bfTEgNjpv1emgkr8dyOlZ/29N3D1nwNckO/xBbd+A5ez\nP3cRKUYO9jDoRkhdgMfpXKI/VzSa8aOvF41mfEyGa6W4OEIgYI46ZsLEqLKyMkzT5Pjx4znPd3Z2\nUlmZvy3zP/3TP/GpT32Kt7/97QDMnTuXgwcP8oMf/OCUxCjggv3lDsV15aTZl3PF0eMD3PXbNykt\nDPJ3H19JYSTA0eMqL2pqeYE+npMYfb1MPKmDr6quZKlE5slAiMi1f45Z2TDicm5vO/EX7wXDIrz+\nT5GBAszqWdiHXifVfgArq8NZKl0WZ06d7f++jSrV2SzVsp3AMtfPV3KObEcm1ZcTonqWP16U14EV\nBDuJNecKrBmLiT/xPWSsF6vhEsy5a5Hp/B45wjlllNfh9hzDKJlKcOWtZ+3ckzKzXnPmMqzZq7H3\nPE/fPX8Jrg2uQ2jNbci+4yQ3PwxAcPktiIr6Eed0rq4VES33/9fX5sQjCkpVhlekRP8+xon+XNFo\nxo++XjSa8THZr5UJawMWDAZZuHAhL7zwgv+c67ps3LiRJUuW5F0mHo9jmrnqmmmauK57VueqmZx0\n9Kgbz57+JD//wy6klBzrHACgpvzsdSrSaN7quH0dxDf8ANnfiXRS6l8qjuzvxG7eMuqyqV3PgmMT\nXHKDn5Xkd71r3Z0zNl+wuFFWi1EyFbfjAKmdT5N87SHs/S8hUzEwLayGS3M6iAnDIrDgKsypcwhf\n/kECjcsJLnkHRvUsQld8dFxB0lbTKozSGsJX/xnCOsXsoFMgvPo2FdouJQiTwLx1BOavJ7js3Vj1\nS7GaVhG8+IZzNp/RsGZcjFFWe9Id5zRnB6txOcaUBozSmomeikaj0Wg0mknKhJbp3XHHHXzhC19g\n4cKFLF68mLvvvpt4PM673vUuAP76r/+a6upq/vIv/xKA9evX8x//8R9UV1fT2NjIli1buOeee/J2\n3tNoxqKrL+PIeGlHO4ubKmjtTDujKiZ3fa5GM1FI1yW+4QeQihNc8g5CK94DgH14M7FHv43s7xp5\nWSlJ7d0EQGDOFf7zntjkiU8evhiV5ZYSwiC8/k4GH/yacma5Nlghou/5e79b2VDCq96f8zi04j2c\njKQUmHkpgZmXjj3wDCOCBRTc9DfDXzAtItd/7pzPZzSMkmqi7/3HiZ6GJk1oyY2Eltw40dPQaDQa\njUYziZlQMeqGG26gq6uL7373u3R0dDB//nx++MMfUl6u7PrHjh3DMDLmrS9/+ct85zvf4Stf+Qqd\nnZ1UV1dzxx138MlPfnKidkFzAdPdFwdgzeIant9yjJ8+spOCsOpcVFOhnVEazdkgtesZnNbdGBX1\nBC99p/+8KFTv++5A54jLuh37kX0dGNWzMIoz5dxGZQOYFk7bXqR0EcLA7e/C7WpBFFZgpNftYVY1\nErzkZpKv/gaA0OUfHFGI0mg0Go1Go9FoNGeeCQ8wv/3227n99tvzvvazn/0s53FhYSFf/vKX+fKX\nv3wupqaZ5HT1KmfUZQuqmTm1iJ/9YTe9A0kiIZOSaHCMpTUazangHNsJQGjlexFm5iPIKKwAQPaP\nLEb5rqimlTnPCzOAWdmI07pbZTOV1hB/+keAHDbWI7j0Jty+TkQ4SmDu2tPZJY1Go9FoNBqNRnOS\nTLgYpdFMFF29yhlVVhxmfkM5/XGbB57Zz7QphePKgdFoJjNSuiBBpN2p0nUA4T8ePl6CdBHG6F0z\n3N52AIzS2pznRbAAAhHc/i6klAghkK6btX0Xe99LIARW44ph6zWnzsZp3Y3TvAXn8BacI29ilNXm\nuK9ytmeYRNbpEm+NRqPRaDQajWYi0GKU5i1LdzozqqxIpb+847J6pk2JMlWHl2ve4shUgsHffQMc\nm4KbvwRmgNjv/g/uYA/RW/8RYQaGLZN4/ufY+16k4L1fywkAH7buE+1gBhDR4WOMwgrc7hZIDmJ3\nHyX20NcJX/WnBJpW4hzbiYydwJy2EKOgZNiyXm5UYtOv0iszCa//U4SlXY4ajUaj0Wg0Gs35hhaj\nNG9JpJR09SWIhi1CAeXkEEJwyZzKMZbUaCY/iU2/xO04AED8hZ9jFJThtKkwcLt5K4GGS3LGy2Qs\n3eUuib33RYKLr8+7XpkYQCb6McpqEWK4w0oUlkN3C25/F87hzSBdktseJ9C0Etsr0Zu1Ku+6zWkL\nsBqX4/Z3IYRB4KJrMKfUn/Ix0Gg0Go1Go9FoNGcPLUZp3pIMxG1StqtdUJpxIe0kIBHWyfRQO7u4\nA93IeB+YFkZJjV9a6sZ6EaHosHI5Ge+HQDgnpynn9VQCt7cNt7OZ1I4NiHARGCb27ueBTNmqvXfT\nMDHKPvQ6OEkAUvs2jShG+SV6xdV5XzcKy3FQuVFOV4tapm0vbs8xUgdeAcPCGrJtD2EFiVzzmbyv\naTQajUaj0Wg0mvMLLUZp3pJ4eVHlReePuKA5P5GJAQZ+83eqZO1dX8GIFE/0lEjte4n4k/8J0gUg\nMH8d4Ss+in10B7GH/1k9Xv0hf7zT2czgg1/DmnExkWs+PWx97mAPgw98FTnQ7T8XWnsHwgwQe+Sb\ngCS05sMkXrwX+9AbyFQcEQhn5pN2LWFauB0HcE+0YpRMHb6dtBglRuhcJ9Ih5m5/J273Ef/5+NM/\nhuQgVsOliFB0nEdJo9FoNBqNRqPRnK/kT6LVaCY5Xie9suLwGCM1b3Xiz/8ceaIN2d9J4pmfqKDu\nCcTt7yT+7E9ASswZFyPCRaR2PEVq5zPEN/wAXIfU7hfSbi7l6opv+D7YCezmLUjXzVmflJL40z9G\nDnRjTGnAql9K6IqPEmi4BGv6IsLrPkFo1fsJzF+vXElOEvvga5nl4/04LW8iwkUEF70NgNS+F/PP\n/UQbAMYIYpTXUc/tPors64C04OWVCFqz8nfG02g0Go1Go9FoNBcW2hmleUvS3aedUZqxSe17EXvv\nRkS0HEwL+9Dr2LueJTBv7UmtR7oOTutucB1EuOiUs4ykdIk/9UNIxggsfhvhVe/Hbt5K7JFvEn/m\nx2qQMCEVw27eQmDmMhKv3I+bLnkjFcftas7ZfmrHBpzmLRilNRTc/DfDShEDc9Zkfm5ahb3nBVI7\nn0akHWLOke0gHazG5QTmrCb5xu9I7dmIWdWEUTgFozTjkPLL9Eryl+l5zii7eQsAVu183L4ONf9A\nGGvGklM6bhqNRqPRaDQajeb8QotRmrckXUM66Wk0+fA6s4XXfQJhBRn87T8Q3/jfmLXzRnT3DEVK\nl9ij38Zp2eY/F7n5S1hT55z0fFJbH8c5ugOjvI7Q8lsAsKYvIrDwalJv/hERLSe07F3En/4R9t5N\niFCU1JbHIBjBaliGvftZnNY9vhjl9hwjsfGXILzOc6NfD2bdAkS4COfYLmLHduW8ZjWtxCitwZhS\nj3v8ELGH/xkQhK/7LIGGS9Wx8DOjRnJGlatxfR3qcXkdRnUTyZd+jVW/VHfG02g0Go1Go9FoJgla\njNJMGmIJmwefO8B1y6dTPkb5nVemN9Y4zVsXaSeQA12I4mqsaQsACC55B8nXHyK24S4KbvoSwhi7\n0jm17Qmclm2I4iqMkmqc5q3Yu587aTHK6Woh8fJ9YFhKODID/muhle/DKKzAnHExRlElvPDf2Iff\nwGnfD0jCaz6MCBcpMaptD1x0DdK1iW24C5wkwWXvxqxsGHMOwrAIr/sk9sFXcp8vrsZM7094zUdI\n7XwGaSew924k8cxPlUuqoFQ5o4TpO6CGrT9ahgpLV6WQRnmdckPZKQLzrjyp46XRaDQajUaj0WjO\nX7QYpZk0bNrexh9ebiaetPno2+fnvNbdl6AkGsQwRPpxukyvWDujNPmRgycAMKKl/nPBS/8Eu2Ub\nbttekpt/T2jpTfmXTQ5iH96CTMVJvHQvGBaRaz+LUVhB/88+R+rAq4RWfyhHUAKV72Qf3gxOCqwQ\nVv3FCMNCOimV++TYhFa+D7Nies5ywgoSvPgG/7E1U7mg5EAXVtNKArMuQyZjIAROq8pfSr72EG7H\nAYzqWQSX3Dju42LNWIw1Y/GIr5tVjZhVjQDEgxFS258k/vSPiVzzaeRgD6K4elinP38/DAtRUIIc\n7AHAKKtDBEKElr1r3PPTaDQajUaj0Wg05z86wFwzaTjWOQDAtgNdOSHTu5t7+Py/v8CvntzrP+eX\n6RVqMUqTHzctiIiCjBglDIvI+jvBDJJ85Tc4HQeHLSftJIMPfo34k/9J4tmfKgFp+bsxK2YgQlEl\n5CQGcJq3DVs2uflh4k98j/iGu4g//q+ktv0RgNSu53A7mzFr5hFYfP2Ycw/MWqXmGy0nvObD6udg\nBKN8OnKgi9S+l0i+/hAEwkTW3zmiOHS6hFa9D1EyFad5C4kX7wXAKBm9vNF3TRkmRmn+bCmNRqPR\naDQajUZzYaPFKM2kobVrEFAleMc6B/3nH3rhIK6UbHj9CCf6E0gp6e5LUBgJEAycnZtwzYWPHBgu\nRgEYpTWEVr0PpEN8w/f9rnUeiZfuw+0+ilHZSHDJOwhf+XEC6S5zAFaTEopS+zblbk9KUnteACCw\n4Co1Zq96bO/dCEBw+S0IMfbbtjltAeG1HyNyw18hQtHM81NnAyiXlXQJXfaBcWdfnQrCCinxTpik\ntj8JgFE0+va8jnpGaS3C0OZdjUaj0Wg0Go1mMqLFKM2koTVLgNq2vxOAQ619vHmgCwDbcXni1Rb6\nYilStqs76U1SpJSkDr6KG+sdc6x9bBfuibb86/FKxYaIUaDEInP6IhUAnnb8ANgtb5La9jiEokSu\n+3NCK95DYO4VOdlSVv3FEAhjH3yd5JZHSe3dhHQd3I4DyN52jKomQqs/hFFai3v8EHbLmzituxGF\nFZjVs8Z1DIQQBOatxSyrzXnerFZiFK6DVb+UwNyT6wp4KphVjQQvvdl/PLYzSoWYG+XTzuqj/LFT\nAAAgAElEQVS8NBqNRqPRaDQazcShxSjNpCBlu3SeiCNUJBTb0gLUoy8dBuCGVfWYhuDJ147wqz+q\ncr2KEh1ePhlxmjcT/8O/knj27lHHubFeYr/7J2J//Pe8r3tilIgOF6OEEISv/DgiVEjqzSewW7Yh\n4/3En/4hAOErPoIRLcu7XmGFsBouBSdJYtMviT/5nyRf/h9Se5VTKjBrFUIIrFkrAfx1es+fDmbN\nXBACESkmtPaO017feAkueQdGVROgcqBGwyitAcCsbDzr89JoNBqNRqPRaDQTg66B0EwK2rsHkcC8\n6aUcautnV3MPW/Z18tKONgojAW5a3UDvQJLnth5j45utlESD3LS6YaKnrTkLpPaokjb78GZkYiCn\nTC0bt/MwSAf3+OG849yBbmB4mZ6HUVBKaO1HiT/+b8Sf+iFm5UzkQDfW7MsJNK4YdY6hVe9TId+O\nTeK135Dc/AgEwiAEVuNyAAJNq0i+8gAyPQ+vvO90MKJlRG74PEa0HCNSfNrrGy/CMCl42//CProD\nM92ZcCQCc1YjggVY9UvP0ew0Go1Go9FoNBrNuUY7ozSTgtauGAC1U6IsaCgjZbt8577NSAk3XlZP\nKGBy42X1lESDXDKnkq9+fAUNU8/dzbhmZNzBHpI7nkJK97TXJVMJ7EOvp1dsYx94deTtdrV4S+G0\n7R32uoylu+mNIEYBBGYuw5qzBjnYg33odURhBeHVt485TyNSTHDh1QQXX0949YcACakYZu18f3tG\nSTVG5Uz1c1ktRvnojqLxYk1bgFE69Yys62QQ4UICjcvHdGMJw1LjTP1diUaj0Wg0Go1GM1nRYpRm\nUtDapTrpVZcXsLhRBSCHgiafeMd8rls+3X/tW59dzWffvYjiguCEzVWTS/K1h0g8+9OMiHQa2Ide\nBzuJKFFd2FL7XhxxrNN1JPNz655hr48UYD6U8OW3IYoqAUF43ScRwYKTmrM16zKstJMqMHt1zmve\nY2v2mnNWUqfRaDQajUaj0Wg0Zxv91bNmUuB10qspL2BBQzlCCObOKKWyNJIzTt/Qn3+4PUfV/8cP\nQ8Olp7WuVLrrXHjV+4k/ezfO0e24gz153U1ud4v/s9M2XIxyB3vACqnyuVEQwQjRd30FN3YCs+zk\nQ7eFEITXfxJn/jrM2vk5rwUWXoVRPg1z6tyTXq9Go9FoNBqNRqPRnK9oZ5RmUtCWLtOrLi/AMARr\nFtcME6I05wdO9xESb/zeL8tze9vV/10toy3mI+P9JF77LW6/6pgokzESL/8Psad+hNOyDREqxJy+\nSLmNpCT+5PeJP/MTnI4DmXW4Lm7XEQhEIFiA074f6diZ1+0kJAcR0dJxCZgiXHhKQpS/vBnAmrZg\n2LaEMLBq5+d049NoNBqNRqPRaDSaCx19h6OZFLR2DWKZBhXFukPe+U7y5f8h+dJ9OEe2I50Usl91\nPnS6xxajpJTENtxF8pX7iT32XaRjE3/uv0i+/hD27mfBdbBmX6Zyh+asBgTO0R2kdj7N4MP/7IeS\nu73t4CQxy+swq2eBk8I9fjCznXQnvdHyojQajUaj0Wg0Go1Gc2roMj3NBU9/LEV/LMW0KVEMQ5fh\nne84nc0AuJ3NiMJyQAIgT7Qj7QTCCo24bGrHBpzmLenlDxF7+P/HObYLEf2/7N15mFwFnf/7zzmn\ntu7qvdPpTtJJIBuEkI2wQxKI0UFQHEDDIImIKzgjekfRO9c7AuoMPs9cr44bXmSR4YcoPyAqEEEh\nhEUSIEAIJiRk3zvdSe9L1amqc+4fp6s6ne4kXUl1V3X1+/U8PnafOqfqW2mPf3ye7/d7KhRa+DnJ\n9HlPqZNkjZqowk9+X25Xq+JbVyu2+RV1rbpf7rI7UiN6ZkWtjKIKJfasV+LgFi+YUveInk68LwoA\nAAAAkD7CKAx7yX1RNRXpLY7GqYltfkVOR5MCcz8+4F1cbiwit61BktcJZZaNOfJVOU37ZXU/Qe5o\nTstBRdf8TjIshT50iyIv3a/Egc2SpNBlX5Bv3Fl9rrG6n0BnVU9R4uA2xfe8p9a3nlXisDfiZ1bU\npp5SZ7/3V8X3b5LvtHNkBLwRT8IoAAAAAMg8xvQw7NUd7g6jKgmjhorrJBR57RHZa59UbPPLA77O\naep5gp3TuFdO60HvF18gdexY7A3PS3FbgblXyT/pPIUu/YxkmArMuarfIOpIhi+g0KIvSaalxhf+\nJ/XkPrOiVlbV6TJCxXI7GpXY/a6if/tfcprrvNcJowAAAAAg4+iMwrB3sMkLo6rLCaOGitO4R4pF\nJEnR134r35gzZZZWn/C6xBFhk9O0X06LF0b5xs9SfMdaJRr3yt/Pda7jKL7tDckw5D9rkSTJP/Vi\n+SbMlhEMD6hma9RpCp57jaJvPK5Ew07vWEWtDF9A4et/KKejUfZbf1R8x1rFNr0kSTIKSwf03gAA\nAACAgSOMwrA3EjujnLZDir72iALnf7Lfp7g5nc2Krvmd/NMula/27NRxNxFT5MV75bQdlgwpcPaH\n5Z9yUZ/rXcdR9PXfK1G3pddxX+0MBc+7LnXcCFfI7WhU16pfq/Dq78gwDMU2v6JE/TYFL1kmw7R6\n15UKowwpEVN8z3ve+552juI71spp3KtE3RZF1z2t0KWfkVlUKUlKHNgkt6tF1rgZvbqVBhpEJQXn\nXCXjwAZF9rwvI1yeut4IhmUFw/KfuUDxHWvldj+pzwiXp/X+AAAAAIATY0wPw95I3BkV2/SS4rve\nkf3OU31ec11XkVX3Kb51jbqe/4Wc7mBFkhIHtyq+/U05Ddvl1G+X/e6Kft/fXv+sYu895513xH/s\nd56S0344FUaF5n9WZkWtnINb5TTs8MKuNb9T7P1VSux/v8/7JsMoa8w0r9bu/VG+8bMkw5RzeLe6\nXvilErvflf3351PXxbeukST5J19wMv9cKYZpqurqr8ooKJVvwuw+r1vjzpIRKu45n84oAAAAAMg4\nwigMa47j6mBTl4oK/Coq6G/AKz8lw6D4zrflxqK9XotteEGJvX+XTJ9kdymy6j65riOpJwwKzPmY\njGCRnKYDcp1E7/c+vFv22icky6fCf/x3hZf+t8JL/1v+mf/gvf/W15Wo+0AyDFk1U+U/c2H38TVK\n7Pm7FO3o/n11n7qdpn2Sacl3+rmpY0ZhmYxQkcyyGrmRNrkdTd532/a6XNeRm4gptmOtZPrkO33e\nKf/b+cuqVbz0/1Vo/mf7vGaYPvkmnZf63SykMwoAAAAAMo0wCsPOwaZOPbDifTW3R3W4NaJ4wlF1\nRUG2yxoyrhNXon6790vcTi3jliSntV7R138vmZYKP/5/yqwYr8T+9xXbsNJ7vTuMMkdP8p4i58RT\ne5skyXUdRV68V3ISCp6/RNboyTILS2UWlso/7VJJUmzDX+V2NsusnCAjUOCFN4ah+PY3FNv6Wuq9\n4jvekhu3e2rrapXb1SqzbIzMUaeljpslo73/LveeamcUVcqsmuQtFK/boviudZLdKd+EWWmP5R2L\nYR17Qtk35cLuH4KSP5SRzwMAAAAA9CCMwrDiuq4eXLFJr64/oFXv7NPBETii5xzaLSVsKeB951j3\nCJskxTa/IiViCsz6qKzqKQotuFmSFN/+hiQp0f00O6ui1guj1PsJdk7TATmNe2WOmij/2Yt7fa5Z\nUSuzfGyqc8mqnuodLyyTNfYsuZ3Nim9/U/KHZE2YLcUiiu9+t+e9k0FYea2s8rGp40Z3GOWbdK6M\ncLlCi74s/5kLvO+zcaWiq38rSakOrMFmVU+RNW6GfKefK8MwhuQzAQAAAGAkIYzCsPLOlkP6YE+z\nJGnTriYdGIFhVKLuA0lS4OzFUjCsxN735Eba5bpuKphKBjdm1WneOQ3b5cZtLxDyBWUUj+oJo5r2\n9nlvX+1MGUbv/3swDEO+yRemfrdqpqV+PnKXk2/iXPnP8MKk+LbXU8dTYVRFrYxgWEa4wvu9O4zy\nTzpfRTf+WL6aafKffq5kWN6oXkeTfN1PzhsKhmGq8KrbVXD5F4fk8wAAAABgpCGMwrARTzj63y9u\nlSRZpqFt+1u152C7pJEWRnn7oqyx0+U//TzJSSi2dbW3QLytQeboyTJLqiR5wYpVPUVKxBXf+bYU\ni8gsH+cd76czKnGw+71rpvT72f4pR4ZRU1M/+06f5+2o6j7HN36mFChQfPc6OZ1eeJgcLUx+bjIM\nM0ur+3yOESqSNd57CqBRVKnQJUvT+jcCAAAAAOQuwigMG8+v3auDTV06+/QKzZk6SgnH1Zub6yWN\nnDDKdV0vMDIsWVWT5J9+mWQYir75hOx1z0jqHRhJPR1Msfdf9H5PhkDl4yRJiSPDqGTQVT1V/TFL\nRss//XL5pl4iM9yz3NsIhhWYc6Ws8bNk1c6Q4Qt43VGJuCIvP6j4vo2Kb1sjBQq9cExe95ZZPUXW\nuLP6/azA7CtlltcqtOgWGYGR8fcFAAAAgJHg2Ft8gRyyt6FdT768XT7L0PWLpmjT7ma9tblBUTsh\nQ9Lo8uwuMHddV5IGdceQG7fltBz0loBXTZLhD8qqOk2B2VfJXve04jvfkgxDvknn97ou2cGUOLBZ\nUk9HkhEokFFUKbe1QW4sKtfu9Dqryscdd1F4aP5N/R4Pnnttn9/ju9cpsftdde1/37v20s/ICBVJ\nkvynz5P/OE/H8405Q75P/eB4/yQAAAAAgGGIzijkvFjc0a+f2qh4wtG1CyZrXFWRzpzY05VTWRqS\n32dlrT7Xiatrxf+jzuV3yXWdU3qv2Ad/U/v/fFVO84Hex3esVfuDt6jziX+X1HtELjDvH2WOmugd\nH3uWzMLSXtdao05LjdBJPWFUz8+unKZ9R4zo9d8VlS7DH1TB5V+WDFOK2/JNvrBP1xYAAAAAYOQh\njELOW/n2Xu2pb9eZE8r0kfPHS5LGVhaqJByQlP0RPfvtp5TYt0HOoZ1y2w+f0nvFd70jN9Km2I63\neh2PbXhBch0ZpTUyqybJf8alqdcMy6eCRbfKGjdDgbkf7/Oehi/gLTLvdmQYdeTeqBON6J0Ma/Qk\nhRZ+Tr5J5yl06bKMvS8AAAAAYPhiTA857+87GiVJn7xsiszuMTjDMHTmhDK98X59VsIoNxGT29Ek\np6VO9jtPpY47jXtlFlf1nOe6UiImwxcY0Ps6rd4OrGSXkiQ5HU1K7N8kI1yu8JL/7POUO0kyy2pU\neNXtx3xfX8002Qe3yigokVlQ0nNddxiVqN+qRMMOSb2fkpcJ/mmXyj/t0hOfCAAAAAAYEQijkNMc\nx9W2fS0K+i1NrCnq9dp5Z47WG+/X64wJ5ce4enC4rqvO5d+T07gndcysmiSnYbsSjXvlmzjXO89x\n1PXcT+Q07FDhJ74js7TmhO+bCqPqtsh1HRmGqfi2NyS58k06v98gaiCS3U7JpeWpusu9MCq26WVJ\nklFYJqN41El9BgAAAAAAA8GYHnLa3oZ2ReyEJo0tkWX2/p/rvDNG6ye3Xapzpg1teOK2H/KCqGBY\n1tjpCpx7rYLnf1KS5DTuS51nr39WiT3r5Uba1PXivXKdxPHfN9ImxSLdF3fKafL2RsW2rZEk+adc\ndNI1W+Nnyn/Woj5jfGbFOPnPWiRr7HRZY6creP4nB3UJOwAAAAAAdEYhp23Z2yJJmlpb2u/rJYUD\nG387VW60QzJMGYGC1G4l/5SLFLpkqSTJ6Wr1/rtxryQpcWiX7LVPSJZfZmm1nPrtst95WsF5n/De\nz3XkRtp7jcy5LQd7fWbi4BYZlk9Oww4ZpdWpJeUnw7B8Cl36mb7HDbPf4wAAAAAADBY6o5DTtuxt\nliRNOUYYNRRcJ6GOx/9dnU//UK7r9iz6PmK3kllQIqOgRE7zAbmJuKJ/+1+Sk1DwgiUq+PC/SL6A\n7Lf/qET9drmuq67n/lsdj/wfiu/bmHqP5IhecnQuUfeB7L//RZLkn3whHUsAAAAAgLxAGIWc5bqu\ntuxtkWFIk8dmL4xyWg/K7WiUc2iXnEO7UsvFrZreT50zK2olN6H4nvVeV1NJtfwzPiSztEbBC2+Q\nXEddL/5/st/9sxK735WchCKrfu11XaknjPJNvViSFN/5tvcUvWBY/umXDd0XBgAAAABgEBFGIWcd\nbo2oqS2q8VVFKghmb6L0yD1QsY0vyGncJ6N4lMxw78XpyY4me+2TkiT/lAtSC8f90y+TNWG23JaD\nst94TDIsWWOny+1oUuTV//E+p9Ub07NGn+4tGu/eHxW69KY+nwUAAAAAwHBFGIWctTW1L6osq3Uk\n90BJUmzzq5Lc1NPpjmRV1PY63zf5wtRrhmEotOBmGaFiSVLg3GtU8JGvyiiqVHzb64of2CynpXtM\nr6Q6NQLom3KR/JPPH5TvBQAAAABANrDAHDlr3dZDkqSp47M3oif1hEtGqNh74p36juhJ3WN6yZ8r\nJ8gqH9v79cIyFXzsW0rUbZH/zMtkmKYCcz+u6Cu/UXzLarmt9ZLlkxEuU2Dux2QUj1LgrEWD+M0A\nAAAAABh6dEYhJzU0d+nNTfUqKvBr9uRRWa0l0eSFUYE5V6WO9RtGlY+T5C0ZP7Ir6khWxXgFzlok\nw+we3zv9XMm0FNu2Rm60XWbJaBmGKbOoUsE5V8kIFGT42wAAAAAAkF2EUchJz72xW64rfWherYIB\nK2t1uPGo3JZ6GUWV8p8x3+tcChZ1B0+9Gf6gzLIxkmHIP+WCAb2/ESqSVTsztR/KLKnOaP0AAAAA\nAOQaxvSQc1o7bb26/oACflMfmld74gsGkdO0X5Irs6JWRjCsgitvl2H5UovJjxb60C1yu9pkFlUO\n+DP8Uy5UYvc6SZJRMjoTZQMAAAAAkLMIo5Bznl2zW3bc0eJ5tSoq8Ge1luS+KKu7E8o35ozjnm9V\nTkj7M3wT50q+gBS3ZRJGAQAAAADyHGN6yClb9jbruTd3KxSwdMUF6Qc7mZboDqOOXE6eaYY/KN/p\n53mfcxJhFgAAAAAAwwmdUcgZXdG47nt6o1xX+vTiaaooCWW7pFRn1GCGUZIUumSpElMvkq+fxegA\nAAAAAOQTOqOQM5a/sl0NzRHNm1alS2bWZLscSZLTtE8yTG8x+SAyAgXy1Z49qJ8BAAAAAEAuIIxC\nTmjttPXyuv0K+E0t+4czZBhGtktSbPsbcjubveXlVnZ3VwEAAAAAkC8Io5ATXli7V3bc0YLZY1US\nDmS7HDkdTYq88pAkKXjB9VmuBgAAAACA/MHOKGRdxI5r5dt7ZZmG/uG8Yy/wdjqaZL/3FwXPuVpG\noOCY57mxiOy3/yT/9Mt6PZ3OTcRlv/1H+U6bJ6vqtJ7jTkL2umfkNO3v+azGvVK0Q/6zPyxf7YxT\n+4IAAAAAACCFMApZ9/K7B9QRieuiGTWqLD320nL7vecUW/+szKJKBc5efOzzNrwg+90VcrraVHDZ\n51PH49tel/3OU4rvf1/hT/zfPeeve1r22uV93sesGK/g+Z86yW8FAAAAAAD6QxiFrHJcVyvf9p5Y\n99ELjt0VJUnO4T3efzfuOe558a1rJEmJg1t6HY9t8447B7fKaWuQWVylRP122W/9UfIFVLD4n2UE\nCr2TDUPmqInsigIAAAAAIMMIo5BVm3c1qb6pS1NrS1U7uui45zqNXmiVaNrn/d5ar+jqR+XGozL8\nBQqce61k9IRVbkudnK5WmQUlcrpaldi7IfVesW2vKzDjw+p68V7JdRS88Ab5JswepG8JAAAAAACS\nCKOQVS+96+1pWjhn7HHPcyJtcrtavJ8b98p1XcU2vaT4rndS5ySa9so3YY73i+WXEjElDm6Redo8\nxbe/KbmOrLHTldj/vuJbX5fb3ii3pU7WhNnyT79sUL4fAAAAAADojafpIWtaO229/UGDCoM+nXvG\n6OOem+yKkiTFInLbDylR543hFVz5TVkTZsttOajYe89JkgKzrpCk1Dnxba9LkoLnXiuzvFZO4x7F\nNq6UESpWaMHNMgwj018PAAAAAAD0gzAKWdHSYesPL29XPOHq4rNrFPBbxz2/J4zyQqNEw04lGrZL\nwbCscWcptOBzMkLFkiRz9CT5p17inVe3RU5rvRJ1H8goqpRZPUW+KRek3je44LMyC8sy/wUBAAAA\nAEC/CKMw5P706g594+d/06p1++WzDC2cO+6E1ziN3p4oa9x0SVLsg1elRFxW9RQZhimzsFShyz4v\n+YIKnP1hGaXVMkLFcg7tVOSl+yVJ/jMWyDAM+adeIiNYJP/ZH5b/tHmD90UBAAAAAEAf7IzCkNqw\no1F/eHWHQgFLl84ao/mzxmrcqPAJr0s0eZ1RvskXKLFvoxJ71kuSrJqpqXN8E+ao6OZfpUburJpp\niu98S4kDm2VW1Cow50pJkllUofBnfsZoHgAAAAAAWUAYhSHT3hXT/c9slCR97srpOvfM4++JSnJd\n1xvT8wXlmzBHUe+gJC9wOtKRAZNVM0XxnW9Jpk+hy78sw/L3ex4AAAAAABg6jOlhyPzuhS1qbrd1\n8dk1Aw6iJMltPyzFIjIrxsksLJVRUOK9YPpkjTrtmNf5TjtXRnGVgpcslVU5/hSrBwAAAAAAmUBn\nFIZEfVOnVm+oU3GhX59ePO3EFxwhubzcqqiVJJkVtUrs2yiz6jQZvsAxrzNLqlR0w3+dfNEAAAAA\nACDj6IzCkHjujT1yXWnxvFoVhtLLQBP12yRJZnl3GFXuLTy3qqce8xoAAAAAAJCbCKMw6Fo7bL36\n3gEF/ZYuP6c2rWsTTftkr39WMi35xs+UJPmnXSpz9GT5z5w/GOUCAAAAAIBBxJgeBt3za3crHo/r\nw+dNVFGB/8QXSHKjHXITcUVW3islYgqc/ymZZWMkSdaoiQr/478PZskAAAAAAGCQEEZhUMXjcU3Y\n+JC+W9qo4rl3D+ia6LoVst94LPW7VTNNgVkfHawSAQAAAADAECKMwqDa/8ofdIa1R5JUENknqeS4\n57uuo9jGFyRJ5qiJMgrLFbp0mQyTiVIAAAAAAPIBYRQyzk3E5LTUy+1oVOmWFanjibot8o2d3nOe\n40h2p4xQUc85B7fJbT8sq2aaCq/+v4a0bgAAAAAAMPgIo5BxXSt+pMSBTZK8Dflv25N0TmC7EnUf\npM5xXVeRv/5M8b0bFP7Uf8gsqZIkxbeukST5plw45HUDAAAAAIDBx+wTMsp1XSUatkuWT+0VZ+rP\nnbO1cczHJSvgdT05jiQptnGl4rvekRK2Yltf8651Eopvf0MyLPkmnZfNrwEAAAAAAAYJYRQyK9oh\nxW2ZpWO03PcxPRuZrfNmjJM1epIU65LTtFdO8wFF1/xeMrz/+cW3vu6FWPs2yo20yaqdITNUnOUv\nAgAAAAAABgNhFDLKaT/s/RCu0PpthxT0W5o1uVJWzVRJUmL/JnW9eK+UsBWY948yKyfKad4v5/Bu\n2e89J0nyT74gW+UDAAAAAIBBxs4oZJTb3ihJanLCsmOOzj2jSgG/pXh3GBVd+6QUi8isnqLAnKtk\nWH5FD+9SZOWv5DQfkFlaw4geAAAAAAB5jM4oZFSyM2p3u1+SNHeqt5jcqp4iyZBiEckfUsHlX5Jh\nWvJNPt+7rvmAZFgKLfqyDF8gK7UDAAAAAIDBRxiFjEqGUZsaDJmGoZmTKyVJRqBQZuUESVLwohtk\nloyWJJlFlbLGnCFJCsz7hKyq07NQNQAAAAAAGCqM6SGj3A5vTG9/JKgzJpSpqMCfei208GY5jXvl\nm3pJr2uC829SYv8m+c9cOKS1AgAAAACAoUcYhYxKdkY1OWFdMXVUr9esUafJGnVan2ussrGyysYO\nRXkAAAAAACDL0h7TW7lypRzHGYxakAec9sNyXEMtTqHmThl14gsAAAAAAMCIknYY9c///M9asGCB\n/uu//kvbtm0bjJowTLlOXG5Hs1qcAp09uUqjygqyXRIAAAAAAMgxaYdRf/3rX7VkyRL9+c9/1sc+\n9jFdf/31euyxx9Te3j4Y9WEYaT98SIZcNblhfeqyydkuBwAAAAAA5KC0w6ja2lrddtttWrlypR54\n4AFNmDBBd999ty699FLdfvvtWrNmzWDUiRznuq5efu3vkiR/ySiNqyrKckUAAAAAACAXndIC84su\nukgXXXSRDh48qH/913/VU089paefflpjx47VsmXLtHTpUvl87EjPd62dtn6zYpOs3Tu1oEiqPW1C\ntksCAAAAAAA56pSSojfeeENPPvmknnvuOfn9ft14441avHixXnnlFf30pz/Ve++9px/96EeZqhU5\naMOORv1txTOyolGND9uSpFBZVZarAgAAAAAAuSrtMGrfvn1avny5/vCHP2jfvn06//zz9f3vf18f\n+chHFAgEJHkdU3PnztXtt9+e8YKRO55+baf++PIW/bB8lQL+hFQ+XmqSzKLKbJcGAAAAAAByVNph\n1OLFizV69Ghdc801uu666zR+/Ph+z5syZYpmzpx5ygUiN9mxhP7wyg5NDjYpYCS8g017JElGUUUW\nKwMAAAAAALks7TDqV7/6lebPny/TPP7u89NPP10PP/zwSReG3LbvUIcc19UFo1qlLknBsBTtkERn\nFAAAAAAAOLa0n6Y3b948HTp0qN/X6uvr1dHRccpFIfftPtgmSZpo1UuSCi7/ooySahnFVV4wBQAA\nAAAA0I+0O6O+853vqLi4WD/4wQ/6vPbzn/9cbW1t+vGPf5yR4pC7dte3S3JVEdkrmT5ZY89S+Lq7\nJEmGYWS3OAAAAAAAkLPS7oxau3atLrvssn5fW7Bggd58881TrQnDwJ76do02W+WLd8qsOk2GLyDD\nH5LhD2W7NAAAAAAAkMPSDqPa2toUCvUfOASDQbW2tp5yUchtjutqT327zgh545q+mmlZrggAAAAA\nAAwXaYdREydO1KpVq/p97aWXXtKECRNOtSbkuIbmLkXthM4KN0qSrOqpWa4IAAAAAAAMF2nvjFq2\nbJnuuOMO+f1+XXvttaqqqlJDQ4OWL1+u3/72t7rzzjsHoUzkkj0H2yVJE4w6SZJZMzAC6JEAACAA\nSURBVCWb5QAAAAAAgGEk7TBqyZIlOnTokO6991795je/SR0PBoP6+te/riVLlmSyPuSg3fXtqrGa\nVRRvklkxXmaoONslAQAAAACAYSLtMEqSvvKVr2jZsmV655131NzcrLKyMs2dO1fFxYQSI8Geg206\nJ7BDkuSbfEGWqwEAAAAAAMPJSYVRklRcXKwFCxZkshYME7vr23R1dxjlJ4wCAAAAAABpOOkwau3a\ntdq5c6ei0Wif12688cZTKgq5qzMSV0nnPo0qbZdZPUVmSVW2SwIAAAAAAMNI2mHUoUOH9NnPflZb\nt26VYRhyXVeSZBhG6hzCqPxV19ipc4I7JdEVBQAAAAAA0meme8EPf/hDFRUV6aWXXpLrunrssce0\ncuVKfe1rX9PEiRP13HPPDUadyBEHDrVrbmCnXBnyTTo/2+UAAAAAAIBhJu0w6s0339TnPvc5VVX1\njGeNHTtWt9xyi66++mrdddddGS0QuaW5vk6lZpe6wmNlFpZmuxwAAAAAADDMpB1Gtba2qqKiQqZp\nqqioSIcPH069NnfuXL399tsZLRC5JXZojyTJKK/NciUAAAAAAGA4SjuMqq2tVX19vSRpypQpeuqp\np1KvvfjiiyorK8tcdcg5VtsBSVJB9cQsVwIAAAAAAIajtMOohQsX6m9/+5sk6dZbb9Vf/vIXLViw\nQIsWLdLDDz+spUuXZrxI5IaE46g46gWRhFEAAAAAAOBkpP00vW9+85upnxcuXKhHH31Uzz//vCKR\niC6++GItXLgwowUidxxqiajabJIkmRWM6QEAAAAAgPSlFUbZtq37779fl19+uc4880xJ0syZMzVz\n5sxBKQ655UBDmyZaLYqaBSoqKMl2OQAAAAAAYBhKa0wvEAjoV7/6lVpbWwerHuSwlgN75DMcdRXW\nyDCMbJcDAAAAAACGobR3Rs2aNUsbN27MWAGPPPKIFi1apJkzZ2rJkiVav379Mc9dtmyZzjjjjD7/\n+dKXvpSxenBs9qHdkiSjbFyWKwEAAAAAAMNV2jujbr/9dn3zm9+Uz+fTwoULVVlZ2adLpqCgYEDv\ntWLFCt1999266667NHv2bD300EP6whe+oGeffVYVFRV9zv/Zz36mWCyW+r25uVmf+MQndMUVV6T7\nNXASrBbvSXqh6glZrgQAAAAAAAxXaYdRS5YskST94Ac/0H/8x3/0e877778/oPd68MEHdf311+u6\n666TJN11111atWqVli9frs9//vN9zi8rK+v1+zPPPKNQKEQYNUTC0YOSJZWMOz3bpQAAAAAAgGEq\n7TDqP//zPzOyL8i2bW3YsEG33npr6phpmrr44ou1bt26Ab3HE088oauuukqFhYUnXYdpDu/dR8n6\nB/t7dEXjGq1GSVJw1HgZw/zfDSPTUN0vwHDHvQIMDPcKMHDcL8DAjJR7Je0w6tprr83IBzc1NSmR\nSGjUqFG9jldWVmrXrl0nvH79+vX64IMPjtmdNRA+n6nKyqKTvj6XlJeHB/X99254V6OsdrWapZo0\npmpQPwsYbIN9vwD5gnsFGBjuFWDguF+Agcn3eyXtMCpXPP7445o2bZpmzZp10u8Rjztqbe3KYFVD\nzzQNlZeH1dTUIcdxB+Uz3FhUnX+5Rz5Jm4ou0PjD7YPyOcBgG4r7BcgH3CvAwHCvAAPH/QIMTD7c\nKyUlBfL7reOek3YYdeGFF55wTG/16tUnfJ/y8nJZlqVDhw71On748GFVVR2/86azs1PPPPOMbrvt\nthMXfALD9Y97NMdxB+27RF57VL7OBv3drlXDqHPz5t8MI9dg3i9APuFeAQaGewUYOO4XYGDy/V5J\nO4y68cYb+4RRLS0tWrNmjdrb21PLyE8kEAhoxowZeu2117Ro0SJJkuM4Wr16tW666abjXvvss8/K\ntm1dffXV6ZaPNDmRNsXef1Exq1C/67hIVxQFs10SAAAAAAAYxtIOo7761a/2e9x1XX3ta1+Tzzfw\nt7z55pv17W9/WzNmzNCsWbP00EMPKRKJ6JprrpEkfetb31J1dbW+8Y1v9Lru8ccf1+LFi1VeXp5u\n+UiTc3iPJOlQ+HS1NRSoJBzIckUAAAAAAGA4y9jOKMMw9KlPfUr/9m//pltuuWVA11x55ZVqbGzU\nT3/6UzU0NGj69Om67777VFFRIUk6cOCATNPsdc327dv11ltv6YEHHshU6TgOp2mfJOmQKiVJpUWE\nUQAAAAAA4ORldIH5nj17FIvF0rpm6dKlWrp0ab+vPfzww32OTZo0SZs3bz6p+pA+p9HrjDqQ8LrQ\nSsOM6QEAAAAAgJOXdhj1yCOP9DkWi8W0fft2PfXUU7riiisyUhhyQ6JxryRpl10iSYzpAQAAAACA\nU5J2GPX973+/z7FAIKCamhrdcMMN+pd/+ZeMFIbsc11HTtN+yR/Svs6gDMNWcYE/22UBAAAAAIBh\nLO0watOmTYNRB3KQ23ZYikVkjp6s1oaYSgoDMk3jxBcCAAAAAAAcg3niUzBSOd0jek7pOCUcV6WM\n6AEAAAAAgFOUdhj14x//WN/97nf7fe273/2ufvKTn5xyUcgNiSYvjOoqGC1JKuFJegAAAAAA4BSl\nHUY9/fTTmjdvXr+vnXvuuXr66adPuSjkhmRnVFvAC6NKCwmjAAAAAADAqUk7jKqvr1d1dXW/r40e\nPVr19fWnXBRyQzKMOmxWSqIzCgAAAAAAnLq0w6iqqipt3Lix39c2btyoioqKUy4K2ecm4nKa62QU\nlKrJ9p6gVxoOZrkqAAAAAAAw3KUdRl1xxRX6xS9+oVWrVvU6/tJLL+mXv/ylrrzyykzVhixyOxol\nNyGzrEYtHbYkscAcAAAAAACcMl+6F3zta1/Tpk2bdMstt6isrExVVVVqaGhQS0uLLrnkEn39618f\njDoxxJyOJkmSES5XSzthFAAAAAAAyIy0w6hgMKgHHnhAr7zyil5//XU1NzerrKxMF110kS655JLB\nqBFZ4HY2S5KMwjK11kclSaXsjAIAAAAAAKco7TAqaf78+Zo/f34ma0EOSYZRZmFZakyvhM4oAAAA\nAABwitLeGfXMM8/ovvvu6/e1+++/XytWrDjlopB9TscRnVEdtnyWocLgSWeXAAAAAAAAkk4ijLr3\n3nsVDPb/VLVQKKR77733lItC9iU7o9yCUrV1xlQaDsgwjCxXBQAAAAAAhru0w6hdu3Zp6tSp/b42\nefJk7dq165SLQvYlw6guIyxXUnEhI3oAAAAAAODUpR1GhUIh1dXV9ftaXV2dAgFCi3zgdrZIktpU\nKIkwCgAAAAAAZEbaYdTFF1+se+65R4cPH+51vLGxUffccw9P1MsTTmeT5A+pPWZJkooL/VmuCAAA\nAAAA5IO0N1J/85vf1JIlS7R48WLNnz9fo0ePVn19vV599VWVlJTo9ttvH4w6MYTceFSyu2SU1qit\nKyZJKiogjAIAAAAAAKcu7c6osWPH6k9/+pOWLl2quro6vfzyy6qrq9OyZcv05JNPasyYMYNRJ4ZQ\nckTPLCxVe6ctic4oAAAAAACQGWl3RklSRUWFvvGNb2S6FuQIp6NJkmQUlqut0+uMYmcUAAAAAADI\nhJMKo1asWKHHHntMO3fuVDQa7fP66tWrT7kwZE+yM8ooLFVbO2N6AAAAAAAgc9Ie03vqqaf07W9/\nWxMmTFBdXZ0WLVqkyy67TI7jqKioSDfeeONg1Ikh5HZ6nVFmuEztXcnOKMIoAAAAAABw6tIOo+6/\n/3595Stf0R133CFJ+vSnP627775bL7zwgsrLy1VQUJDxIjG0ejqjytTWvTOKzigAAAAAAJAJaYdR\nu3bt0jnnnCPLsmRZltrb2yVJRUVF+uIXv6hHHnkk40ViaPXsjCpTOzujAAAAAABABqUdRoXDYdm2\n1y1TXV2tbdu2pV5zXVdNTU2Zqw4Z47quEvXb5CZiJz63K/k0vTK1dcVkGFJh6KTWiwEAAAAAAPSS\ndsIwc+ZMbd68WfPnz9eiRYv0y1/+Uj6fT36/X7/4xS80Z86cwagTp8ip36bOP/5A1oTZKviHr8sw\njGOe63Y0ez8UlKqtM6aiAr/M45wPAAAAAAAwUGmHUV/+8pe1f/9+SdJtt92mffv26c4775TjOJo5\nc6a+973vZbxInDqn5aAkKbH7XcXeX6XAWZcf+9zOZskXlG0EFE84Ki5kDxgAAAAAAMiMtMOoOXPm\npLqfSkpKdM8998i2bdm2raKioowXiMxw7c7Uz9E1j8o3drrMspq+58Wjkt0po7Rabd37olheDgAA\nAAAAMiXtnVH9CQQCBFE5zo14i+bN8rFS3FZ07ZP9nhff/a53XlGl2ruSy8sJowAAAAAAQGZkJIxC\n7nOjHZKkwNyrJX9I8V3r5Npdvc5xOpsVfeV/vPNmXaG2Tm9RfTGdUQAAAAAAIEMIo0aIZBhlhMvl\nO+0cKWErvusdua6j2JbXFH3rD+r668/lRtvlP+tD8o2f1TOmR2cUAAAAAADIkLR3RmF4SoVRwbD8\nky9UfMtrim1dI7erTdE1j6bOM0trFLxwiST1jOkVBIa+YAAAAAAAkJcIo0YIN+rtjDKCYZllNTJC\nxUrs3aDE/o2S6VPwon+S4S+Qb+IcGb6gJNEZBQAAAAAAMo4waoRwo97T9IxgWIbpk2/SeYptXCkl\npOCF/6TAjMV9rmnv6t4ZRRgFAAAAAAAyhJ1RI0W0Q7ICMnzeyJ1/2iWSJGvsdPlnfqTfS5KdUYzp\nAQAAAACATKEzagRwXUdutF1GQWnqmDV6sgqv+57M0hoZRv+ZZFv3zqginqYHAAAAAAAyhDBqJIhF\nJNeVESzqddiqnHDcy9gZBQAAAAAAMo0xvRHAjXQ/SS8UTuu69k5bAb+poN8ajLIAAAAAAMAIRBg1\nArjR7jAqOPAwKuE46ozE2RcFAAAAAAAyijBqBDiZMKq9Ky5XjOgBAAAAAIDMIowaAZJhlNIIoxqa\nuiRJo0pDg1ESAAAAAAAYoQijRgA32i5JfRaYH8/Bpk5JUk1F4aDUBAAAAAAARibCqBGgZ0xv4MFS\nXaMXRo0uLxiUmgAAAAAAwMhEGDUC9IRR6XRGeWN6dEYBAAAAAIBMIowaAdxI+gvMD3Z3RlWXE0YB\nAAAAAIDMIYwaCZI7o0IDC6Nc19XBpk4VBH0q5ml6AAAAAAAggwijRgDX9rqcBtoZ1dxuy445qi4v\nkGEYg1kaAAAAAAAYYQijRoCeMb2B7YxKLi9nXxQAAAAAAMg0wqgRwI22S4Yp+UMDOv9gU/e+KMIo\nAAAAAACQYYRRI4Ab7ZARDA945K5neXnBYJYFAAAAAABGIMKoPOfGbSkRk9J6kl6XJDqjAAAAAABA\n5hFG5Tk3mtwXlUYYlRzTKyeMAgAAAAAAmUUYlefSDaMcx1V9U5dKCv0qDPkGszQAAAAAADACEUbl\nuXTDqD317Uo4Lk/SAwAAAAAAg4IwKs+50XZJAw+jXl6/X5J0zhmjB60mAAAAAAAwchFG5buot/9p\nIGFU1E5ozYY6+SxTF59dM9iVAQAAAACAEYilQHnOTYVR/Y/dtXfF9NjKrZo0rkSWYagrmtCFM6pV\nVOAfyjIBAAAAAMAIQRiV51y7O4wK9A2jInZcP/nf72r7/la9+t4B+SyvUW7h7LFDWiMAAAAAABg5\nGNPLc8kF5jqqMyoWd/TzJ9/T9v2tOn1MiaorChVPOKqpKNS08WVZqBQAAAAAAIwEdEbluZ7OqJ6d\nUQnH0b1/2qCNO5tUW1Wkf71+tnymqRff2afpE8tlGEa2ygUAAAAAAHmOMCrPHb0zynVdPfTsZr31\nQYNGlxXoG9fPVjjk7Ye64oIJWasTAAAAAACMDIzp5bujdkZt29+qV9cfUFlRQN/4pzkqLQpmszoA\nAAAAADDCEEbluaM7ow42er9fMnOMqsoKslYXAAAAAAAYmQij8pxrd0qGIflDkqTm9qgkqYyOKAAA\nAAAAkAWEUXnOtTulQKEMw/tTN7fZkgijAAAAAABAdhBG5THXSUixSGpflNTTGVVeTBgFAAAAAACG\nHmFUHnOPWl4uSU2pMb1AVmoCAAAAAAAjG2FUPjtqebnkdUYZkkrChFEAAAAAAGDoEUblsaM7oxzX\nVUu7rZJwQD6LPz0AAAAAABh6JBJ5zD2qM6qtM6aE47K8HAAAAAAAZA1hVB5z7Q7vh+7OqOY2lpcD\nAAAAAIDsIozKYz2dUWFJLC8HAAAAAADZRxiVz47aGdWcDKPojAIAAAAAAFlCGJXHjt4ZlRzTY2cU\nAAAAAADIFsKoPHb00/SSnVHsjAIAAAAAANlCGJXHkp1RSu6MarMl0RkFAAAAAACyhzAqj9EZBQAA\nAAAAcg1hVB5LhVHdO6Oa2qLyWabCIV82ywIAAAAAACMYYVQ+i3ZI8jqjYnFH7V0xlRUFZBhGlgsD\nAAAAAAAjFWFUHnOjnZJpSb6AWjq6n6THiB4AAAAAAMgiwqg85tqdMgKFMgxDzd3Ly8tZXg4AAAAA\nALKIMCpPuXFbSsSk7n1RrZ1eGFVSGMhmWQAAAAAAYIQjjMpTPU/SC0uSInZckhQKWlmrCQAAAAAA\ngDAqTx39JL2onZAkhQKEUQAAAAAAIHsIo/JVNNkZ5YVRkVQY5ctaSQAAAAAAAIRReapnTM8Lo7ro\njAIAAAAAADmAMCpPudHeY3qpnVGEUQAAAAAAIIsIo/KUG4t4P/hDko7cGcWYHgAAAAAAyB7CqHwV\n65IkGYECSUfujKIzCgAAAAAAZA9hVJ5yba8zyujujEqGUUHCKAAAAAAAkEWEUXnq6DE9dkYBAAAA\nAIBcQBiVr2JRSZLhD0piZxQAAAAAAMgNhFF5qqczip1RAAAAAAAgdxBG5Sk3ucD8iDE9n2XIZ/En\nBwAAAAAA2UMyka9ifReYM6IHAAAAAACyjTAqT7ndO6PkD8pxXNlxhxE9AAAAAACQdYRReco9ojOK\nfVEAAAAAACBXEEblq1hEMizJ8itixyVJQcIoAAAAAACQZYRRecqNRaRASIZhKBpLdkaxMwoAAAAA\nAGQXYVQecl1HikVl+IKSxJgeAAAAAADIGYRR+ShuS3JlBLqfpBf1xvQIowAAAAAAQLYRRuWh5PJy\n+bvDqGRnlJ8xPQAAAAAAkF2EUfnITj5Jr0CSFEnujArSGQUAAAAAALKLMCoPufFkGMXOKAAAAAAA\nkFsIo/KQax89ppfcGcWYHgAAAAAAyC7CqHwUS3ZGJReYe51RQT+dUQAAAAAAILsIo/KQe1QYFY0x\npgcAAAAAAHIDYVQe6vs0ve4xPRaYAwAAAACALMt6GPXII49o0aJFmjlzppYsWaL169cf9/yWlhbd\ncccduvjiizVz5kx99KMf1RtvvDFE1Q4TR4/pJReY+9kZBQAAAAAAsiur6cSKFSt0991366677tLs\n2bP10EMP6Qtf+IKeffZZVVRU9Dnftm3dfPPNqqqq0s9//nONHj1ae/bsUWVlZRaqz11Hj+nxND0A\nAAAAAJArshpGPfjgg7r++ut13XXXSZLuuusurVq1SsuXL9fnP//5Puc/8cQTam1t1e9//3v5/X5J\nUm1t7ZDWPBykxvQChFEAAAAAACC3ZC2Msm1bGzZs0K233po6ZpqmLr74Yq1bt67fa1auXKk5c+bo\nzjvv1IsvvqjKykpdd911uummm2QYxknVYZond12uSNZ/5Pcw4lHvWCAk0zQU7Q6jCkP+Yf99gVPR\n3/0CoC/uFWBguFeAgeN+AQZmpNwrWQujmpqalEgkNGrUqF7HKysrtWvXrn6v2bNnj1avXq1rrrlG\nv/71r7V161Z973vfk2EYuummm9KuweczVVlZdFL155ry8nDq53ojLltS2agKhSqLZCccSdLYMSXy\n++iOAo68XwAcG/cKMDDcK8DAcb8AA5Pv98qw2mjtuq6qqqp05513yrIszZgxQ3v27NHvfve7kwqj\n4nFHra1dg1Dp0DFNQ+XlYTU1dchxXElSpL1NktTa6arjcLs6umKyTEOtLcP7uwKnqr/7BUBf3CvA\nwHCvAAPH/QIMTD7cKyUlBfL7j98Ik7Uwqry8XJZl6dChQ72OHz58WFVVVf1eM2rUKPn9fllWz5ea\nPHmyDhw4cNJ1DNc/7tEcx019F9f2dka5vqAcx1XUTigUsPLmuwKn6sj7BcCxca8AA8O9Agwc9wsw\nMPl+r5jZ+uBAIKAZM2botddeSx1zHEerV6/WnDlz+r1m7ty52r17txzHSR3buXOnxowZM+j1Didu\n984o+UNyXFfRWEKhwLBqggMAAAAAAHkqa2GUJN188836/e9/r+XLl2vbtm268847FYlEdM0110iS\nvvWtb+lHP/pR6vwbbrhBTU1N+uEPf6gdO3bo+eef14MPPqhPf/rT2foKucn2xvEMfyi1vJwn6QEA\nAAAAgFyQ1XaZK6+8Uo2NjfrpT3+qhoYGTZ8+Xffdd58qKiokSQcOHJBp9uRl48aN03333ae7775b\njz76qMaMGaNbbrlFN954Y7a+Qk5yYxHJMCXLr0iXLYkwCgAAAAAA5Iasz24tXbpUS5cu7fe1hx9+\nuM+xefPm6fHHHx/ssoY1NxaR/CEZhqGIHZdEGAUAAAAAAHJDVsf0kHmu60qxqAx/SJIUSY3pZT13\nBAAAAAAAIIzKO3FbktsnjArSGQUAAAAAAHIAYVSecWPe8nJ1h1EsMAcAAAAAALmEMCrfxCKSJMMf\nlCR1RmOSpIIgY3oAAAAAACD7CKPyjBuLSlJqTK+90wujigv8WasJAAAAAAAgiTAqz7jdnVHJMb32\niBdGhQmjAAAAAABADiCMyjfdO6OMQIEkqb0rLkkqIowCAAAAAAA5gDAqz7h2d2eUz9sZ1d5pS5KK\nCgmjAAAAAABA9hFG5ZnkmJ4R6B7T6/LG9OiMAgAAAAAAuYAwKt/YyTG9QkmM6QEAAAAAgNxCGJVn\nXLtT0pFhlC3TMFQQ9GWzLAAAAAAAAEmEUXnHjXZIkoxgWK7rqr0rrnCBT6ZhZLkyAAAAAAAAwqi8\nk+yMUrBQdsxRPOEwogcAAAAAAHIGYVSecaM9Y3ptXd1P0iOMAgAAAAAAOYIwKs+kdkYFw+pgeTkA\nAAAAAMgxhFH55ojOqPaumCQpTBgFAAAAAAByBGFUnnHtTsmwJF+AMT0AAAAAAJBzCKPyjBvtlBEs\nlGEYqTG9YsIoAAAAAACQIwij8oibiEkJWwoWShJjegAAAAAAIOcQRuWRI5+kJ0ntnV4YxZgeAAAA\nAADIFYRR+cQ+KoyKEEYBAAAAAIDcQhiVR9xkGHXUmB5hFAAAAAAAyBWEUXmkZ0wvLIkxPQAAAAAA\nkHsIo/KIG+2Q1LczKlzgy1pNAAAAAAAARyKMyiPJMT0dsTOqMOiTZfJnBgAAAAAAuYGUIo8cuTMq\nFncUtROM6AEAAAAAgJxCGJVPoj1P00stLy8kjAIAAAAAALmDMCqPpBaYBwvVwZP0AAAAAABADiKM\nyiOpMb1AodqSy8tDhFEAAAAAACB3EEblkdQC8yM6o4oZ0wMAAAAAADmEMCqPpMb0AmG1R5KdUb5s\nlgQAAAAAANALYVQece0OSZIRKEh1RoXZGQUAAAAAAHIIYVQ+iXZKll+GL6COSFwSO6MAAAAAAEBu\nIYzKE67ryrU7ZQQKJemIzijG9AAAAAAAQO4gjMoXcVtyEjKCYUmiMwoAAAAAAOQkwqg84Ua9fVEK\nep1RnRF2RgEAAAAAgNxDGJUnXDv5JD0vjGrv8jqjiniaHgAAAAAAyCGEUXnCjfYOozoiMRmSQkHC\nKAAAAAAAkDsIo/JEqjMq2BNGFYZ8Mg0jm2UBAAAAAAD0QhiVJ5I7o4xAoWLxhOyYw74oAAAAAACQ\ncwij8kSyM0qBQp6kBwAAAAAAchZhVL6I2ZIkwx9QR1fySXrsiwIAAAAAALmFMCpPuInuMMoKpDqj\niuiMAgAAAAAAOYYwKl/EvW4o+QLqiHR3RhFGAQAAAACAHEMYlSfcuNcZJZ9fHV1eZ1RhiDE9AAAA\nAACQWwij8kTvMb3kzig6owAAAAAAQG4hjMoX/Y7p0RkFAAAAAAByC2FUnujpjOoZ06MzCgAAAAAA\n5BrCqHzRT2cUT9MDAAAAAAC5hjAqTyQXmBu+gDoiyc4oxvQAAAAAAEBuIYzKF4nuzijLr44u7+dC\nOqMAAAAAAECOIYzKE707o1hgDgAAAAAAchNhVL7o1RkVVzBgyWfx5wUAAAAAALmFtCJPJDujHMOn\nzmhcRXRFAQAAAACAHEQYlS/itmT51WUnJElh9kUBAAAAAIAcRBiVJ9xETPIFUsvLwwWEUQAAAAAA\nIPcQRuWLuC3D8qsjEpfE8nIAAAAAAJCbCKPygJuIS67jdUZ1P0mvkDE9AAAAAACQgwij8kByeblh\nHTmmR2cUAAAAAADIPYRRecCJeWGUfH51Rb0xvcIgYRQAAAAAAMg9hFF5INUZ5QsoGnMkSQGflc2S\nAAAAAAAA+kUYlQeSYZQsv2LxhCQp4OdPCwAAAAAAcg+JRR5wYz2dUXaczigAAAAAAJC7CKPyQE9n\nVEB2ckyPzigAAAAAAJCDSCzygBOPSpIMn19295ien84oAAAAAACQgwij8kByTE+WP9UZFaQzCgAA\nAAAA5CASizzgJmLeD74AnVEAAAAAACCnEUblgdQCc8uvWJydUQDw/7d378FRlfcfxz97DcglVxKl\nOiAIG8idYhwgSIUiNE5niqY4iFoKarWAHS6iVDEXfhjajkIDttLBUiSJijY6VhTpVGDGRgFtYxSN\nCLTWKikhIYYYyG52z++PkCMrIFHh7CXv18wO7DnPWZ7D4csyH77nOQAAAADCF4lFFDAXMHe65fV1\ndka5nVxaAAAAAAAQfkgsokBXGGVzuOU1O6O4TQ8AAAAAAIQfwqgoEOhawNzpojMKAAAAAACENRKL\nKGB0tEuSbE46owAAAAAAQHgjjIoC5ppRjs41o2w2yWG3hXZSAAAAAAAAZ0AYhUJVMAAAFR5JREFU\nFQXMp+k5O5+m53Y5ZLMRRgEAAAAAgPBDGBUFTu2MavcFWC8KAAAAAACELVKLKBA4GUYZDpc6/AG5\nnawXBQAAAAAAwhNhVBTouk2vQ05JktvFZQUAAAAAAOGJ1CIKdN2m16HOjig6owAAAAAAQLgijIoC\nXWGUN3AyjKIzCgAAAAAAhClSiyjQdZue7+TlZAFzAAAAAAAQrkgtokDA7IzqWjOK2/QAAAAAAEB4\nIoyKAuZtekZnCOWiMwoAAAAAAIQpUosoYHR4JZtN7f7O93RGAQAAAACAcEUYFQWMDq/kcMvXYUhi\nzSgAAAAAABC+SC2igOHzyuZ0y9vR2RpFZxQAAAAAAAhXhFERzjCMzs4op0vtvpNhFJ1RAAAAAAAg\nTJFaRDq/T5Jkc7jl6whIojMKAAAAAACEL8KoSHcyjJLTLa+vM4ziaXoAAAAAACBckVpEOKPDK0my\nOV3mmlExdEYBAAAAAIAwRRgV4brCKDnojAIAAAAAAOGP1CLSda0ZderT9Jx0RgEAAAAAgPBEGBXh\nzM4op8vsjHK7uKwAAAAAACA8kVpEuq41oxwu+czOKC4rAAAAAAAIT6QWEc7oepreKWtGuVnAHAAA\nAAAAhCnCqEhnPk3PLW/HyTCKzigAAAAAABCmSC0iXPCaUZ236bnojAIAAAAAAGGKMCrSdT1Nz/FF\nZ1QMnVEAAAAAACBMkVpEOCPoNr2TnVFOOqMAAAAAAEB4IoyKdF0LmDtd8pkLmHNZAQAAAABAeCK1\niHDO74xUr8tGyHlputo7/LLbbHI6uKwAAAAAACA8kVpEOEfSIA289f/kiB8ony9AVxQAAAAAAAhr\nJBdRxNvhl5vFywEAAAAAQBgjuYgSgYChDr8ht4vFywEAAAAAQPgijIoSXzxJj0sKAAAAAADCF8lF\nlPCaT9KjMwoAAAAAAIQvwqgo0dUZxZpRAAAAAAAgnJFcRAk6owAAAAAAQCQgjIoSdEYBAAAAAIBI\nQHIRJXx0RgEAAAAAgAgQ8jCqoqJCEydOVEZGhqZPn67a2tqzjq2qqpLH4wl6ZWRkWDjb8NXO0/QA\nAAAAAEAEcIbyF3/ppZdUWlqq4uJiZWVlaePGjbrtttu0detWJSQknPGYuLg4vfjii+Z7m81m1XTD\nWldnVIyTzigAAAAAABC+QhpGbdiwQTfeeKNuuOEGSVJxcbF27Nih5557TnPmzDnrcQMGDDhvc7Db\nIzvM6pq/z38yjHI7Iv6cgAulqzaoEeCrUStA91ArQPdRL0D39JRaCVkY5fV6tXfvXt11113mNrvd\nrrFjx6qmpuasx7W2tup73/ueDMNQWlqaFi5cqCuuuOIbzcHptCsxse83OjbcuNydlzK2f6+oOSfg\nQomP7xPqKQARgVoBuodaAbqPegG6J9prJWRh1NGjR+X3+5WUlBS0PTExUR999NEZj7n88su1YsUK\neTweHTt2TH/84x81Y8YMvfjii0pJSfnac+joCKil5fg3mn+4sNttio/vo8ajbZKkDl+HGhtbQzwr\nIDx11cvRo58rEDBCPR0gbFErQPdQK0D3US9A90RDrfTv31uuczxcLaS36X1dOTk5ysnJCXqfn5+v\nZ555RvPmzftGnxmpF/fL3v/oqCRpYGKfqDkn4EIJBAzqBOgGagXoHmoF6D7qBeieaK+VkD16LT4+\nXg6HQ0eOHAna3tjY2O01oVwul0aMGHHWTqqewuvz692DTYpxOTRycHyopwMAAAAAAHBWIQuj3G63\n0tLSVF1dbW4LBAJ6/fXXlZ2d3a3P8Pv92rdv33ld0DwSvf1hg9p9fqUPSZCLp+kBAAAAAIAwFtLb\n9H7605/q3nvvVVpamjIzM7Vx40adOHFC06ZNkyQtWbJEKSkpWrRokSRp7dq1ys7O1qBBg9TS0qLH\nH39chw4dUkFBQShPI+TeeLdekjRqWM8O5QAAAAAAQPgLaRiVn5+vpqYmlZWVqaGhQSNGjND69euV\nkJAgSTp06JDs9i+at1paWrRs2TI1NDQoNjZW6enpevrppzVkyJBQnULIBQxDu9+rl91mU8bQxFBP\nBwAAAAAA4CvZDMOI3hWxzsHn86u5uS3U0/hWDn7aov974k2NGBSve2bknPsAoAez221KTOyrxsbW\nqF4MEPi2qBWge6gVoPuoF6B7oqFW4uIuOufT9EK2ZhTOj/qmzjDtytTkEM8EAAAAAADg3EJ6mx6+\nvTHpKRr0nVh9J753qKcCAAAAAABwTnRGRTiH3a7s4cmy222hngoAAAAAAMA5EUYBAAAAAADAMoRR\nAAAAAAAAsAxhFAAAAAAAACxDGAUAAAAAAADLEEYBAAAAAADAMoRRAAAAAAAAsAxhFAAAAAAAACxD\nGAUAAAAAAADLEEYBAAAAAADAMoRRAAAAAAAAsAxhFAAAAAAAACxDGAUAAAAAAADLEEYBAAAAAADA\nMoRRAAAAAAAAsAxhFAAAAAAAACxDGAUAAAAAAADLEEYBAAAAAADAMoRRAAAAAAAAsAxhFAAAAAAA\nACxDGAUAAAAAAADLEEYBAAAAAADAMjbDMIxQTyJUAgFDfn8g1NP41lwuh3w+f6inAUQE6gXoHmoF\n6B5qBeg+6gXonkivFYfDLrvd9pVjenQYBQAAAAAAAGtxmx4AAAAAAAAsQxgFAAAAAAAAyxBGAQAA\nAAAAwDKEUQAAAAAAALAMYRQAAAAAAAAsQxgFAAAAAAAAyxBGAQAAAAAAwDKEUQAAAAAAALAMYRQA\nAAAAAAAsQxgFAAAAAAAAyxBGAQAAAAAAwDKEUQAAAAAAALAMYRQAAAAAAAAsQxgV4SoqKjRx4kRl\nZGRo+vTpqq2tDfWUAEvt2bNHd955p/Ly8uTxeLR9+/ag/e3t7SouLtZVV12lnJwczZ8/X42NjUFj\nPv30U91xxx3KysrSmDFj9Otf/1p+v9/K0wAuuHXr1umGG25QTk6OxowZo3nz5unf//530BjqBZCe\nfvpp/fCHP9SoUaM0atQo3Xjjjdq5c6e5nzoBzqywsFAej0fl5eXmtubmZi1atEijRo3SlVdeqfvv\nv19tbW1Bx9XV1emmm25SRkaGJkyYoPXr11s9deCCW7NmjTweT9Br6tSp5v6e+N1CGBXBXnrpJZWW\nlmru3Ll67rnn5PF4dNttt6mpqSnUUwMs09bWJo/Ho8LCwjPuf+ihh7R9+3atXr1amzZt0uHDh3X3\n3Xeb+/1+v372s5/J5/Ppqaee0sqVK1VVVaW1a9dadQqAJXbv3q2ZM2dq8+bN2rBhg7xer2bPnq0T\nJ06YY6gXQEpOTtaiRYtUVVWlP//5zxozZozmzp2rAwcOSKJOgDPZvn27ampqlJycHLR98eLF2r9/\nvzZs2KDf//732rNnj4qKisz9ra2tmjNnjgYOHKiqqiotWbJEa9as0bPPPmvxGQAXXmpqql577TXz\nVVlZae7rkd8tBiJWQUGBUVJSYr73+/1GXl6esX79+hDOCgid4cOHG6+++qr5vqWlxUhLSzO2bt1q\nbtu/f78xfPhwo7a21jAMw9ixY4cxYsQIo6GhwRxTWVlpjB492vB6vdZNHrBYY2OjMXz4cOOtt94y\nDIN6Ab7KlVdeaVRVVVEnwBk0NDQYV199tVFXV2dcc801xqZNmwzD+KI23nnnHXPszp07jdTUVLM+\nKioqjNzc3KDa+M1vfmNcd9111p4EcIGVlZUZ06ZNO+O+nvrdQmdUhPJ6vdq7d6/GjRtnbrPb7Ro7\ndqxqampCODMgfLz77rvy+XxBdTJ06FANHDjQrJOamhqlpqYqKSnJHJOXl6eWlhYdPHjQ8jkDVjl2\n7JgkKTY2VhL1ApyJ3+/Xli1bdPz4cWVlZVEnwBksXbpUt9xyizweT9D2f/7zn4qLi1N6erq5bezY\nsbLZbObSIjU1NcrNzZXL5TLH5OXl6cMPP1Rra6s1JwBY5ODBg8rLy9OkSZN0zz33qL6+XlLP/TcY\nYVSEOnr0qPx+f9AfRklKTExUQ0NDiGYFhJcjR46oV69e6tu3b9D2xMREHTlyxByTmJgYtL+rrrrG\nANHGMAyVlpYqNzdXQ4cOlUS9AKf64IMPlJOTo4yMDBUWFurRRx/VkCFDqBPgS8rLy3X8+HHNnj37\ntH1nqgWn06nY2FjqBT1OZmamSktLtX79ehUVFenjjz/WzJkz1dbW1mO/W5yhngAAALBWSUmJ9u3b\npyeffDLUUwHC0uWXX67nn39ex44d0yuvvKJ7771XFRUVoZ4WEFYOHDig3/3ud9q8ebPsdnocgK8y\nYcIE8+epqanKysrSNddco1deeUVOZ8+MZfhbI0LFx8fL4XCcloI2NjZqwIABIZoVEF6SkpJ04sSJ\n09q8Gxsbzf9JSEpKOu1JFV119eXOQyAaLF++XK+++qo2btyolJQUczv1AnzB7XZr0KBBSk9P16JF\ni+TxeLRp0ybqBDjF22+/raamJl177bUaOXKkRo4cqU8++UQrVqzQ1KlTz1gLHR0d+uyzz6gX9Hj9\n+/fX4MGD9dFHH/XY7xbCqAjldruVlpam6upqc1sgENDrr7+u7OzsEM4MCB/p6elyuVxBdXLw4EF9\n+umnZp1kZ2errq4u6CmU1dXV6t+/v4YMGWL5nIELxTAMlZSUaNu2bdq4caMuu+yyoP3UC3B2hmHI\n6/VSJ8Apvv/97+uFF17Q888/b76Sk5N1xx136LHHHlNOTo6am5u1d+9e85g33nhDhmEoMzNTUme9\n7N69Wz6fzxxTXV2tYcOGnXbLEhBNPv/8c3388ccaMGBAj/1ucRSd+mxNRJS+fftq9erVuuSSS+R2\nu/Xb3/5WdXV1WrFihXr37h3q6QGW+Pzzz3XgwAEdOXJETz31lLKzs+V2uyV1dhD+73//U0VFhVJT\nU9Xc3KzCwkJdeumluuuuuyRJl112mbZt26bq6mp5PB69//77Wr58uWbMmBG0iCAQ6YqLi/WXv/xF\nZWVlSk5OVltbm9ra2uRwOOR0OhUTE0O9AJJWrVolh8MhwzBUX1+vJ554Qi+88ILuueceDR06lDoB\nToqJiVFiYmLQq7y8XHl5ebr66quVkJCgt99+W1u2bNHIkSP13//+V4WFhRo/frx+9KMfSZIGDx6s\nyspKffjhhxo8eLB27dqlRx55RPPnz1daWlqIzxA4f371q1/J7XbLMAzt379fRUVFampqUlFRkWJj\nY3vkd4vNMAwj1JPAN1deXq7HH39cDQ0NGjFihJYtW2b+TwPQE+zatUu33nrradvnzZun+fPnq729\nXStXrtSWLVvk9Xo1fvx4FRYWBrWzfvLJJyoqKtLu3bvVu3dvTZs2TYsXL5bD4bDyVIAL6stPOepS\nWlqq66+/XpKoF0DSgw8+qNdee02HDx9Wv3795PF4dPvtt5v/2KdOgLObOHGiZs+erZtvvlmS1Nzc\nbN4ebrfbNWXKFD3wwAO66KKLzGPq6upUUlKid955R/Hx8brlllt0++23h+oUgAtiwYIF2rNnj5qb\nm5WQkKDRo0drwYIFZqd6T/xuIYwCAAAAAACAZVgzCgAAAAAAAJYhjAIAAAAAAIBlCKMAAAAAAABg\nGcIoAAAAAAAAWIYwCgAAAAAAAJYhjAIAAAAAAIBlCKMAAAAAAABgGcIoAACAKLZr1y55PB7t27cv\n1FMBAACQRBgFAAAAAAAACxFGAQAAAAAAwDKEUQAAABfAm2++qZtvvllZWVm66qqr9MADD6i1tVWS\nVFVVJY/Ho9raWt10003KzMzUlClT9Ne//vW0zykvL9e1116r9PR0TZ48WX/6059OG1NXV6c777xT\no0ePVk5OjgoKCvT3v/89aMzRo0d19913KycnR5MmTVJFRcUFOW8AAIBzIYwCAAA4z9566y3NmjVL\nSUlJKisr09KlS7Vz50798pe/DBq3YMECTZo0SWvWrNHw4cP1i1/8QnV1deb+zZs3a/ny5Zo4caIe\ne+wxTZ06VStXrtQf/vAHc8yBAwc0Y8YMHT58WMXFxVq7dq0mT56sQ4cOBf1ay5YtU2pqqtauXavc\n3FyVlJSotrb2wv5GAAAAnIEz1BMAAACINg8//LBycnK0evVqc1tKSopmzZoVtJD4j3/8Y82ZM0eS\nNH78eOXn52vdunVatWqVAoGA1qxZo+uvv1733XefJCkvL0/Hjh3TunXr9JOf/EQxMTF69NFH1a9f\nP1VWVqpXr16SpHHjxp02p+uuu04///nPJUm5ubnavn27tm3bpszMzAv2+wAAAHAmdEYBAACcR8eP\nH1dNTY1+8IMfqKOjw3x997vflcvl0t69e82xkydPNn9ut9s1adIks1upvr5ehw8f1tSpU4M+Pz8/\nX62trfrggw8kSW+88Yby8/PNIOpsTg2oXC6XBg8erPr6+m99vgAAAF8XnVEAAADnUUtLi/x+v4qL\ni1VcXHza/kOHDuniiy+WJCUkJATtS0xMVENDgySZPyYmJp42RpI+++wzSVJzc7MGDBhwznn1798/\n6L3L5ZLX6+3OKQEAAJxXhFEAAADnUb9+/WSz2TRv3jxNmDDhtP3Jycnm4uJNTU2Kj4839zU2NprB\nUtePjY2NQcd3vY+NjZUkxcXFmcEVAABAJOA2PQAAgPPooosuUnZ2tv71r38pIyPjtFdKSoo59tSn\n5wUCAf3tb38z13C6+OKLlZycrK1btwZ9/ssvv6y+ffvK4/FIksaMGaOXX35Z7e3tFpwdAADAt0dn\nFAAAwHm2ePFizZo1S3a7XVOmTFGfPn106NAh7dixQwsWLDDHPfPMM3K5XBo2bJieffZZ/ec//9Ej\njzwiqXMNqfnz5+vBBx9UXFycxo0bpz179ujJJ5/UwoULFRMTI0maO3euCgoKNHPmTM2ePVtxcXF6\n7733FBcXp4KCgpCcPwAAwFchjAIAADjPRo8erYqKCpWVlWnJkiUKBAIaOHCgxo8fr6SkJHPcqlWr\n9NBDD2n16tW65JJLtGrVKo0cOdLcP336dLW3t+uJJ57Qpk2blJKSovvuu0+zZs0yxwwZMkSVlZV6\n+OGHdf/990uSrrjiCi1cuNCy8wUAAPg6bIZhGKGeBAAAQE9SVVWlpUuX6h//+If69OkT6ukAAABY\nijWjAAAAAAAAYBnCKAAAAAAAAFiG2/QAAAAAAABgGTqjAAAAAAAAYBnCKAAAAAAAAFiGMAoAAAAA\nAACWIYwCAAAAAACAZQijAAAAAAAAYBnCKAAAAAAAAFiGMAoAAAAAAACW+X+h4uMrNrNyjAAAAABJ\nRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 1440x720 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6v-c6zlSagrD",
        "colab_type": "text"
      },
      "source": [
        "Test Accuracy"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "i2uMl4cGaJqP",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "c4e78574-9703-45de-f09a-e1c1bd9807d3"
      },
      "source": [
        "y_predict = np.round(model.predict(X_test_norm))\n",
        "y_predict = [i[0] for i in y_predict.tolist()]\n",
        "sum(y_predict == y_test)/len(y_test)"
      ],
      "execution_count": 31,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.849609375"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 31
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "yVpBte8HaYUl",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 285
        },
        "outputId": "50759e15-584c-4014-a6e6-631272c17e01"
      },
      "source": [
        "from sklearn.metrics import confusion_matrix\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "target_names = ['negative', 'positive']\n",
        "C = confusion_matrix(y_test,y_predict) \n",
        "C = C / C.astype(np.float).sum(axis=1)*100\n",
        "sns.heatmap(C, annot=True, fmt=\".2f\",cmap=\"GnBu\",xticklabels=target_names, yticklabels=target_names)\n",
        "plt.ylabel(\"True Label\")\n",
        "plt.xlabel(\"Predicted Label\")\n",
        "plt.show()"
      ],
      "execution_count": 32,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXEAAAEMCAYAAAAyO4a7AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nJIxBnIhIwhjEiYgkjEGciEjC/h+OD3PLqVZDMAAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mBRz9hT-cT68",
        "colab_type": "text"
      },
      "source": [
        "# Remark for multiple classification"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6ncjtg-S2gHM",
        "colab_type": "text"
      },
      "source": [
        "Change label to tensor flow acceptable format"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Metp7rYvcXR9",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "ed59b708-a2c0-483b-971f-8539e2ef1984"
      },
      "source": [
        "num_label = len(np.unique(y_train))\n",
        "y_train_multi = (np.arange(num_label) == y_train[:,None]).astype(np.float32)\n",
        "y_train_multi\n",
        "\n",
        "y_val_multi = (np.arange(num_label) == y_val[:,None]).astype(np.float32)\n",
        "y_val_multi"
      ],
      "execution_count": 46,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [1., 0.],\n",
              "       [0., 1.],\n",
              "       [1., 0.],\n",
              "       [0., 1.]], dtype=float32)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 46
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "u35E5rUU3gOn",
        "colab_type": "text"
      },
      "source": [
        "Change output layers to softmax with #node = number of class"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KiGhYRQy3e9C",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 329
        },
        "outputId": "a7bfbd60-70c7-4be1-ac2f-4e8494df3bc7"
      },
      "source": [
        "tf.reset_default_graph() # Clear Model\n",
        "np.random.seed(1) # Fix Randomness\n",
        "tf.random.set_random_seed(1) # Fix Randomness\n",
        "model = tf.keras.models.Sequential([\n",
        "  tf.keras.layers.Dense(16, activation='relu', input_shape=(14,)),\n",
        "  tf.keras.layers.Dense(8, activation='relu'),\n",
        "  tf.keras.layers.Dense(4, activation='relu'),\n",
        "  tf.keras.layers.Dense(2, activation='relu'),\n",
        "  tf.keras.layers.Dense(num_label, activation='softmax')    \n",
        "])\n",
        "model.summary()"
      ],
      "execution_count": 43,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "dense (Dense)                (None, 16)                240       \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 8)                 136       \n",
            "_________________________________________________________________\n",
            "dense_2 (Dense)              (None, 4)                 36        \n",
            "_________________________________________________________________\n",
            "dense_3 (Dense)              (None, 2)                 10        \n",
            "_________________________________________________________________\n",
            "dense_4 (Dense)              (None, 2)                 6         \n",
            "=================================================================\n",
            "Total params: 428\n",
            "Trainable params: 428\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IH1Y0DJi34CW",
        "colab_type": "text"
      },
      "source": [
        "Change loss to categorical_crossentropy (Multi Classification)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-m7QMqDS34ix",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VgZNdm1b4YML",
        "colab_type": "text"
      },
      "source": [
        "Train Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vGi5t2Z34aAW",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "575fc0a8-859d-4b92-af08-043491e2e47b"
      },
      "source": [
        "train_acc = list()\n",
        "val_acc = list()\n",
        "for i in range(0,500):\n",
        "  history = model.fit(X_train_norm, y_train_multi, epochs= 1, batch_size = 200, validation_data= (X_val_norm, y_val_multi))\n",
        "  tmp_avg = np.mean(history.history['acc'])\n",
        "  tmp_avg_val = np.mean(history.history['val_acc'])\n",
        "  train_acc.append(tmp_avg)\n",
        "  val_acc.append(tmp_avg_val)"
      ],
      "execution_count": 47,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 91us/sample - loss: 0.7622 - acc: 0.4930 - val_loss: 0.7086 - val_acc: 0.5110\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.7178 - acc: 0.5021 - val_loss: 0.6867 - val_acc: 0.5648\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.6979 - acc: 0.5474 - val_loss: 0.6768 - val_acc: 0.5917\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 18us/sample - loss: 0.6862 - acc: 0.5768 - val_loss: 0.6728 - val_acc: 0.6284\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.6798 - acc: 0.6116 - val_loss: 0.6696 - val_acc: 0.6577\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 18us/sample - loss: 0.6742 - acc: 0.6508 - val_loss: 0.6658 - val_acc: 0.6650\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.6686 - acc: 0.6642 - val_loss: 0.6613 - val_acc: 0.6870\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.6628 - acc: 0.6661 - val_loss: 0.6557 - val_acc: 0.6822\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.6569 - acc: 0.6648 - val_loss: 0.6504 - val_acc: 0.6797\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.6510 - acc: 0.6661 - val_loss: 0.6451 - val_acc: 0.6846\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.6452 - acc: 0.6673 - val_loss: 0.6401 - val_acc: 0.6993\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.6391 - acc: 0.6789 - val_loss: 0.6352 - val_acc: 0.6968\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.6329 - acc: 0.6936 - val_loss: 0.6302 - val_acc: 0.6993\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.6268 - acc: 0.7095 - val_loss: 0.6253 - val_acc: 0.7017\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.6209 - acc: 0.7174 - val_loss: 0.6213 - val_acc: 0.7188\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.6150 - acc: 0.7205 - val_loss: 0.6154 - val_acc: 0.7262\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.6094 - acc: 0.7174 - val_loss: 0.6101 - val_acc: 0.7262\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.6041 - acc: 0.7205 - val_loss: 0.6058 - val_acc: 0.7311\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.5995 - acc: 0.7260 - val_loss: 0.6017 - val_acc: 0.7311\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.5948 - acc: 0.7303 - val_loss: 0.5984 - val_acc: 0.7335\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.5900 - acc: 0.7352 - val_loss: 0.5945 - val_acc: 0.7335\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.5854 - acc: 0.7382 - val_loss: 0.5908 - val_acc: 0.7359\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.5808 - acc: 0.7413 - val_loss: 0.5874 - val_acc: 0.7384\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.5764 - acc: 0.7437 - val_loss: 0.5839 - val_acc: 0.7384\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.5720 - acc: 0.7541 - val_loss: 0.5809 - val_acc: 0.7433\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.5680 - acc: 0.7609 - val_loss: 0.5771 - val_acc: 0.7408\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.5637 - acc: 0.7602 - val_loss: 0.5735 - val_acc: 0.7384\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.5596 - acc: 0.7682 - val_loss: 0.5708 - val_acc: 0.7433\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.5556 - acc: 0.7676 - val_loss: 0.5684 - val_acc: 0.7531\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 19us/sample - loss: 0.5526 - acc: 0.7786 - val_loss: 0.5662 - val_acc: 0.7555\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.5481 - acc: 0.7786 - val_loss: 0.5642 - val_acc: 0.7457\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.5455 - acc: 0.7743 - val_loss: 0.5625 - val_acc: 0.7482\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.5416 - acc: 0.7841 - val_loss: 0.5606 - val_acc: 0.7579\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.5382 - acc: 0.7896 - val_loss: 0.5581 - val_acc: 0.7506\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.5346 - acc: 0.7920 - val_loss: 0.5563 - val_acc: 0.7506\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.5310 - acc: 0.7914 - val_loss: 0.5560 - val_acc: 0.7457\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.5283 - acc: 0.7969 - val_loss: 0.5540 - val_acc: 0.7604\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.5253 - acc: 0.7933 - val_loss: 0.5531 - val_acc: 0.7579\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.5225 - acc: 0.7963 - val_loss: 0.5519 - val_acc: 0.7482\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.5208 - acc: 0.8024 - val_loss: 0.5511 - val_acc: 0.7457\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.5177 - acc: 0.8018 - val_loss: 0.5498 - val_acc: 0.7579\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.5140 - acc: 0.8012 - val_loss: 0.5472 - val_acc: 0.7531\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.5119 - acc: 0.8055 - val_loss: 0.5453 - val_acc: 0.7604\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.5090 - acc: 0.8006 - val_loss: 0.5438 - val_acc: 0.7677\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.5071 - acc: 0.8110 - val_loss: 0.5425 - val_acc: 0.7653\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.5049 - acc: 0.8135 - val_loss: 0.5420 - val_acc: 0.7702\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.5014 - acc: 0.8128 - val_loss: 0.5401 - val_acc: 0.7628\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.4995 - acc: 0.8122 - val_loss: 0.5403 - val_acc: 0.7677\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.4965 - acc: 0.8141 - val_loss: 0.5395 - val_acc: 0.7726\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.4937 - acc: 0.8147 - val_loss: 0.5390 - val_acc: 0.7653\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.4910 - acc: 0.8135 - val_loss: 0.5389 - val_acc: 0.7677\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 18us/sample - loss: 0.4891 - acc: 0.8159 - val_loss: 0.5383 - val_acc: 0.7702\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4860 - acc: 0.8165 - val_loss: 0.5366 - val_acc: 0.7702\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.4838 - acc: 0.8165 - val_loss: 0.5343 - val_acc: 0.7628\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.4818 - acc: 0.8190 - val_loss: 0.5350 - val_acc: 0.7751\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4791 - acc: 0.8190 - val_loss: 0.5331 - val_acc: 0.7751\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4791 - acc: 0.8196 - val_loss: 0.5318 - val_acc: 0.7751\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.4752 - acc: 0.8190 - val_loss: 0.5333 - val_acc: 0.7726\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.4735 - acc: 0.8226 - val_loss: 0.5307 - val_acc: 0.7726\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4708 - acc: 0.8226 - val_loss: 0.5299 - val_acc: 0.7751\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4694 - acc: 0.8220 - val_loss: 0.5281 - val_acc: 0.7775\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4668 - acc: 0.8300 - val_loss: 0.5258 - val_acc: 0.7800\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.4650 - acc: 0.8287 - val_loss: 0.5294 - val_acc: 0.7751\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 18us/sample - loss: 0.4633 - acc: 0.8257 - val_loss: 0.5265 - val_acc: 0.7824\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.4628 - acc: 0.8226 - val_loss: 0.5248 - val_acc: 0.7848\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.4585 - acc: 0.8269 - val_loss: 0.5232 - val_acc: 0.7848\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.4565 - acc: 0.8281 - val_loss: 0.5220 - val_acc: 0.7800\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.4552 - acc: 0.8281 - val_loss: 0.5236 - val_acc: 0.7848\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.4532 - acc: 0.8269 - val_loss: 0.5222 - val_acc: 0.7775\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.4515 - acc: 0.8306 - val_loss: 0.5201 - val_acc: 0.7873\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.4492 - acc: 0.8355 - val_loss: 0.5192 - val_acc: 0.7873\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4469 - acc: 0.8318 - val_loss: 0.5172 - val_acc: 0.7824\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4459 - acc: 0.8324 - val_loss: 0.5178 - val_acc: 0.7800\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.4437 - acc: 0.8330 - val_loss: 0.5173 - val_acc: 0.7775\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.4416 - acc: 0.8355 - val_loss: 0.5151 - val_acc: 0.7824\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.4403 - acc: 0.8361 - val_loss: 0.5165 - val_acc: 0.7751\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 19us/sample - loss: 0.4381 - acc: 0.8367 - val_loss: 0.5138 - val_acc: 0.7726\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4366 - acc: 0.8398 - val_loss: 0.5121 - val_acc: 0.7726\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4343 - acc: 0.8398 - val_loss: 0.5132 - val_acc: 0.7775\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4328 - acc: 0.8416 - val_loss: 0.5121 - val_acc: 0.7824\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4311 - acc: 0.8398 - val_loss: 0.5127 - val_acc: 0.7824\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.4292 - acc: 0.8440 - val_loss: 0.5089 - val_acc: 0.7824\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.4290 - acc: 0.8404 - val_loss: 0.5079 - val_acc: 0.7873\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.4260 - acc: 0.8422 - val_loss: 0.5095 - val_acc: 0.7775\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4248 - acc: 0.8422 - val_loss: 0.5081 - val_acc: 0.7897\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.4246 - acc: 0.8398 - val_loss: 0.5073 - val_acc: 0.7873\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.4272 - acc: 0.8410 - val_loss: 0.5101 - val_acc: 0.7775\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.4237 - acc: 0.8398 - val_loss: 0.5033 - val_acc: 0.7800\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4201 - acc: 0.8434 - val_loss: 0.5078 - val_acc: 0.7800\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4180 - acc: 0.8428 - val_loss: 0.5032 - val_acc: 0.7775\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4195 - acc: 0.8391 - val_loss: 0.5001 - val_acc: 0.7824\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4175 - acc: 0.8440 - val_loss: 0.5080 - val_acc: 0.7775\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.4131 - acc: 0.8465 - val_loss: 0.5008 - val_acc: 0.7873\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.4118 - acc: 0.8453 - val_loss: 0.5039 - val_acc: 0.7824\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.4118 - acc: 0.8434 - val_loss: 0.4999 - val_acc: 0.7922\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.4084 - acc: 0.8489 - val_loss: 0.5008 - val_acc: 0.7873\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.4074 - acc: 0.8477 - val_loss: 0.4990 - val_acc: 0.7946\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.4058 - acc: 0.8483 - val_loss: 0.4988 - val_acc: 0.7873\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.4043 - acc: 0.8471 - val_loss: 0.4979 - val_acc: 0.7873\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.4025 - acc: 0.8526 - val_loss: 0.4986 - val_acc: 0.7897\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.4013 - acc: 0.8532 - val_loss: 0.4964 - val_acc: 0.7971\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3999 - acc: 0.8477 - val_loss: 0.4974 - val_acc: 0.7824\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.4002 - acc: 0.8459 - val_loss: 0.4955 - val_acc: 0.7824\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3965 - acc: 0.8489 - val_loss: 0.4975 - val_acc: 0.7848\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3969 - acc: 0.8532 - val_loss: 0.4909 - val_acc: 0.7946\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3951 - acc: 0.8514 - val_loss: 0.4897 - val_acc: 0.7922\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3925 - acc: 0.8538 - val_loss: 0.4932 - val_acc: 0.7946\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3928 - acc: 0.8508 - val_loss: 0.4895 - val_acc: 0.7897\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3941 - acc: 0.8514 - val_loss: 0.4900 - val_acc: 0.8020\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3945 - acc: 0.8520 - val_loss: 0.4906 - val_acc: 0.8044\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3902 - acc: 0.8550 - val_loss: 0.4867 - val_acc: 0.7946\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3880 - acc: 0.8508 - val_loss: 0.4893 - val_acc: 0.7946\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3856 - acc: 0.8514 - val_loss: 0.4872 - val_acc: 0.7897\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3856 - acc: 0.8557 - val_loss: 0.4893 - val_acc: 0.7971\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3854 - acc: 0.8557 - val_loss: 0.4888 - val_acc: 0.7971\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3826 - acc: 0.8569 - val_loss: 0.4856 - val_acc: 0.7946\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3816 - acc: 0.8581 - val_loss: 0.4871 - val_acc: 0.7995\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 18us/sample - loss: 0.3803 - acc: 0.8593 - val_loss: 0.4877 - val_acc: 0.7897\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3785 - acc: 0.8563 - val_loss: 0.4870 - val_acc: 0.7922\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3791 - acc: 0.8526 - val_loss: 0.4900 - val_acc: 0.7873\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3776 - acc: 0.8575 - val_loss: 0.4873 - val_acc: 0.7897\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3764 - acc: 0.8544 - val_loss: 0.4866 - val_acc: 0.7971\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3749 - acc: 0.8544 - val_loss: 0.4850 - val_acc: 0.7897\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3754 - acc: 0.8532 - val_loss: 0.4871 - val_acc: 0.7946\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3731 - acc: 0.8569 - val_loss: 0.4840 - val_acc: 0.7995\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3715 - acc: 0.8532 - val_loss: 0.4832 - val_acc: 0.7995\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3718 - acc: 0.8563 - val_loss: 0.4853 - val_acc: 0.7946\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3714 - acc: 0.8575 - val_loss: 0.4832 - val_acc: 0.7946\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3707 - acc: 0.8587 - val_loss: 0.4835 - val_acc: 0.7995\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3677 - acc: 0.8544 - val_loss: 0.4834 - val_acc: 0.8020\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3666 - acc: 0.8557 - val_loss: 0.4851 - val_acc: 0.7946\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 19us/sample - loss: 0.3678 - acc: 0.8593 - val_loss: 0.4846 - val_acc: 0.7922\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3660 - acc: 0.8563 - val_loss: 0.4826 - val_acc: 0.8044\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3648 - acc: 0.8569 - val_loss: 0.4835 - val_acc: 0.8044\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3640 - acc: 0.8599 - val_loss: 0.4827 - val_acc: 0.8020\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3637 - acc: 0.8544 - val_loss: 0.4851 - val_acc: 0.7946\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3624 - acc: 0.8599 - val_loss: 0.4841 - val_acc: 0.7971\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3620 - acc: 0.8575 - val_loss: 0.4834 - val_acc: 0.7971\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3611 - acc: 0.8563 - val_loss: 0.4803 - val_acc: 0.8020\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3597 - acc: 0.8581 - val_loss: 0.4831 - val_acc: 0.7971\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 18us/sample - loss: 0.3597 - acc: 0.8550 - val_loss: 0.4836 - val_acc: 0.7897\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3602 - acc: 0.8544 - val_loss: 0.4799 - val_acc: 0.7971\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3598 - acc: 0.8563 - val_loss: 0.4799 - val_acc: 0.7995\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3592 - acc: 0.8624 - val_loss: 0.4788 - val_acc: 0.8020\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3584 - acc: 0.8557 - val_loss: 0.4796 - val_acc: 0.7995\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3555 - acc: 0.8618 - val_loss: 0.4804 - val_acc: 0.7897\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3549 - acc: 0.8606 - val_loss: 0.4783 - val_acc: 0.8044\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 20us/sample - loss: 0.3543 - acc: 0.8599 - val_loss: 0.4763 - val_acc: 0.7995\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.3555 - acc: 0.8563 - val_loss: 0.4787 - val_acc: 0.8044\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3536 - acc: 0.8618 - val_loss: 0.4786 - val_acc: 0.8044\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3521 - acc: 0.8642 - val_loss: 0.4778 - val_acc: 0.8117\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3506 - acc: 0.8642 - val_loss: 0.4769 - val_acc: 0.8044\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.3503 - acc: 0.8612 - val_loss: 0.4770 - val_acc: 0.7897\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.3494 - acc: 0.8593 - val_loss: 0.4759 - val_acc: 0.8020\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3484 - acc: 0.8642 - val_loss: 0.4753 - val_acc: 0.8020\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3516 - acc: 0.8587 - val_loss: 0.4802 - val_acc: 0.8068\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3503 - acc: 0.8661 - val_loss: 0.4801 - val_acc: 0.8117\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3473 - acc: 0.8642 - val_loss: 0.4768 - val_acc: 0.8068\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3481 - acc: 0.8661 - val_loss: 0.4735 - val_acc: 0.8068\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3471 - acc: 0.8606 - val_loss: 0.4756 - val_acc: 0.7995\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3481 - acc: 0.8667 - val_loss: 0.4749 - val_acc: 0.8044\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3457 - acc: 0.8575 - val_loss: 0.4762 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3443 - acc: 0.8661 - val_loss: 0.4744 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3440 - acc: 0.8642 - val_loss: 0.4750 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3452 - acc: 0.8661 - val_loss: 0.4709 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3451 - acc: 0.8569 - val_loss: 0.4736 - val_acc: 0.7995\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3426 - acc: 0.8630 - val_loss: 0.4737 - val_acc: 0.7971\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3406 - acc: 0.8661 - val_loss: 0.4734 - val_acc: 0.8020\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3396 - acc: 0.8691 - val_loss: 0.4714 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3414 - acc: 0.8636 - val_loss: 0.4741 - val_acc: 0.8068\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3430 - acc: 0.8709 - val_loss: 0.4744 - val_acc: 0.8068\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.3463 - acc: 0.8636 - val_loss: 0.4756 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3429 - acc: 0.8697 - val_loss: 0.4738 - val_acc: 0.7995\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 18us/sample - loss: 0.3385 - acc: 0.8624 - val_loss: 0.4737 - val_acc: 0.8020\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3362 - acc: 0.8642 - val_loss: 0.4754 - val_acc: 0.7971\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3376 - acc: 0.8697 - val_loss: 0.4743 - val_acc: 0.8044\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3370 - acc: 0.8679 - val_loss: 0.4726 - val_acc: 0.8068\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3363 - acc: 0.8673 - val_loss: 0.4724 - val_acc: 0.8044\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3370 - acc: 0.8642 - val_loss: 0.4720 - val_acc: 0.7995\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 11us/sample - loss: 0.3353 - acc: 0.8661 - val_loss: 0.4702 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3344 - acc: 0.8667 - val_loss: 0.4732 - val_acc: 0.7971\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 11us/sample - loss: 0.3336 - acc: 0.8691 - val_loss: 0.4715 - val_acc: 0.8068\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3334 - acc: 0.8679 - val_loss: 0.4695 - val_acc: 0.8117\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3332 - acc: 0.8679 - val_loss: 0.4655 - val_acc: 0.8117\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3335 - acc: 0.8716 - val_loss: 0.4680 - val_acc: 0.8044\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3327 - acc: 0.8691 - val_loss: 0.4704 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.3331 - acc: 0.8771 - val_loss: 0.4694 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3302 - acc: 0.8667 - val_loss: 0.4660 - val_acc: 0.8117\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.3312 - acc: 0.8709 - val_loss: 0.4640 - val_acc: 0.8068\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3309 - acc: 0.8636 - val_loss: 0.4669 - val_acc: 0.8020\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3309 - acc: 0.8685 - val_loss: 0.4654 - val_acc: 0.8044\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3297 - acc: 0.8642 - val_loss: 0.4640 - val_acc: 0.8044\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3281 - acc: 0.8667 - val_loss: 0.4702 - val_acc: 0.7995\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3283 - acc: 0.8648 - val_loss: 0.4674 - val_acc: 0.8068\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3274 - acc: 0.8709 - val_loss: 0.4685 - val_acc: 0.8020\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3261 - acc: 0.8673 - val_loss: 0.4656 - val_acc: 0.8166\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3284 - acc: 0.8667 - val_loss: 0.4640 - val_acc: 0.8142\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3314 - acc: 0.8618 - val_loss: 0.4683 - val_acc: 0.7995\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3290 - acc: 0.8722 - val_loss: 0.4619 - val_acc: 0.8044\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3257 - acc: 0.8636 - val_loss: 0.4641 - val_acc: 0.8020\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.3232 - acc: 0.8667 - val_loss: 0.4643 - val_acc: 0.8117\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3232 - acc: 0.8685 - val_loss: 0.4631 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3252 - acc: 0.8624 - val_loss: 0.4669 - val_acc: 0.8044\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3244 - acc: 0.8734 - val_loss: 0.4632 - val_acc: 0.7995\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3280 - acc: 0.8642 - val_loss: 0.4667 - val_acc: 0.8068\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3267 - acc: 0.8716 - val_loss: 0.4595 - val_acc: 0.8068\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3216 - acc: 0.8667 - val_loss: 0.4612 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3219 - acc: 0.8703 - val_loss: 0.4600 - val_acc: 0.8166\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.3203 - acc: 0.8709 - val_loss: 0.4617 - val_acc: 0.8117\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3196 - acc: 0.8728 - val_loss: 0.4612 - val_acc: 0.8068\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.3180 - acc: 0.8697 - val_loss: 0.4634 - val_acc: 0.8068\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.3189 - acc: 0.8716 - val_loss: 0.4607 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3186 - acc: 0.8777 - val_loss: 0.4661 - val_acc: 0.8117\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3204 - acc: 0.8716 - val_loss: 0.4634 - val_acc: 0.8117\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3215 - acc: 0.8728 - val_loss: 0.4644 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3234 - acc: 0.8703 - val_loss: 0.4569 - val_acc: 0.8142\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3203 - acc: 0.8722 - val_loss: 0.4572 - val_acc: 0.8117\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3165 - acc: 0.8758 - val_loss: 0.4568 - val_acc: 0.8068\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3156 - acc: 0.8728 - val_loss: 0.4556 - val_acc: 0.8117\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 18us/sample - loss: 0.3154 - acc: 0.8728 - val_loss: 0.4565 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 26us/sample - loss: 0.3159 - acc: 0.8722 - val_loss: 0.4645 - val_acc: 0.8166\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3138 - acc: 0.8771 - val_loss: 0.4595 - val_acc: 0.8093\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3168 - acc: 0.8746 - val_loss: 0.4600 - val_acc: 0.8117\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3126 - acc: 0.8758 - val_loss: 0.4578 - val_acc: 0.8142\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3124 - acc: 0.8740 - val_loss: 0.4569 - val_acc: 0.8142\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3125 - acc: 0.8746 - val_loss: 0.4551 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.3136 - acc: 0.8746 - val_loss: 0.4566 - val_acc: 0.8166\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3118 - acc: 0.8771 - val_loss: 0.4560 - val_acc: 0.8142\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3129 - acc: 0.8752 - val_loss: 0.4536 - val_acc: 0.8142\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3137 - acc: 0.8765 - val_loss: 0.4608 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3101 - acc: 0.8734 - val_loss: 0.4573 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3104 - acc: 0.8789 - val_loss: 0.4539 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3138 - acc: 0.8740 - val_loss: 0.4585 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3106 - acc: 0.8752 - val_loss: 0.4559 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3133 - acc: 0.8734 - val_loss: 0.4618 - val_acc: 0.8166\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3112 - acc: 0.8746 - val_loss: 0.4593 - val_acc: 0.8166\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3098 - acc: 0.8746 - val_loss: 0.4537 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3099 - acc: 0.8758 - val_loss: 0.4532 - val_acc: 0.8117\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.3069 - acc: 0.8783 - val_loss: 0.4574 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3077 - acc: 0.8771 - val_loss: 0.4536 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.3067 - acc: 0.8728 - val_loss: 0.4551 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3057 - acc: 0.8716 - val_loss: 0.4542 - val_acc: 0.8289\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3060 - acc: 0.8765 - val_loss: 0.4526 - val_acc: 0.8289\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3051 - acc: 0.8777 - val_loss: 0.4529 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.3048 - acc: 0.8807 - val_loss: 0.4556 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3054 - acc: 0.8820 - val_loss: 0.4522 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3054 - acc: 0.8758 - val_loss: 0.4582 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3029 - acc: 0.8789 - val_loss: 0.4518 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3053 - acc: 0.8807 - val_loss: 0.4620 - val_acc: 0.8166\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.3069 - acc: 0.8728 - val_loss: 0.4510 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.3041 - acc: 0.8728 - val_loss: 0.4587 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3028 - acc: 0.8740 - val_loss: 0.4527 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3012 - acc: 0.8832 - val_loss: 0.4591 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.3022 - acc: 0.8783 - val_loss: 0.4563 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2999 - acc: 0.8813 - val_loss: 0.4555 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3013 - acc: 0.8813 - val_loss: 0.4510 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3022 - acc: 0.8795 - val_loss: 0.4556 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2996 - acc: 0.8789 - val_loss: 0.4513 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2991 - acc: 0.8801 - val_loss: 0.4514 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 19us/sample - loss: 0.2988 - acc: 0.8832 - val_loss: 0.4492 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2990 - acc: 0.8746 - val_loss: 0.4500 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2994 - acc: 0.8801 - val_loss: 0.4571 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.3022 - acc: 0.8801 - val_loss: 0.4488 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2968 - acc: 0.8826 - val_loss: 0.4500 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2966 - acc: 0.8807 - val_loss: 0.4488 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2972 - acc: 0.8838 - val_loss: 0.4511 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2966 - acc: 0.8850 - val_loss: 0.4471 - val_acc: 0.8142\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2961 - acc: 0.8820 - val_loss: 0.4444 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2965 - acc: 0.8789 - val_loss: 0.4493 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2983 - acc: 0.8813 - val_loss: 0.4479 - val_acc: 0.8289\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2952 - acc: 0.8813 - val_loss: 0.4515 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2956 - acc: 0.8832 - val_loss: 0.4459 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2962 - acc: 0.8795 - val_loss: 0.4538 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2925 - acc: 0.8844 - val_loss: 0.4475 - val_acc: 0.8289\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.2941 - acc: 0.8856 - val_loss: 0.4476 - val_acc: 0.8313\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2928 - acc: 0.8826 - val_loss: 0.4489 - val_acc: 0.8191\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2923 - acc: 0.8826 - val_loss: 0.4506 - val_acc: 0.8337\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2923 - acc: 0.8826 - val_loss: 0.4494 - val_acc: 0.8289\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2916 - acc: 0.8850 - val_loss: 0.4517 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2926 - acc: 0.8813 - val_loss: 0.4478 - val_acc: 0.8313\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2925 - acc: 0.8813 - val_loss: 0.4522 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2922 - acc: 0.8862 - val_loss: 0.4593 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2957 - acc: 0.8795 - val_loss: 0.4464 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2930 - acc: 0.8832 - val_loss: 0.4502 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2886 - acc: 0.8869 - val_loss: 0.4481 - val_acc: 0.8289\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2882 - acc: 0.8862 - val_loss: 0.4535 - val_acc: 0.8215\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2884 - acc: 0.8850 - val_loss: 0.4450 - val_acc: 0.8289\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2885 - acc: 0.8875 - val_loss: 0.4595 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 11us/sample - loss: 0.2874 - acc: 0.8844 - val_loss: 0.4444 - val_acc: 0.8313\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 11us/sample - loss: 0.2882 - acc: 0.8850 - val_loss: 0.4524 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2869 - acc: 0.8844 - val_loss: 0.4445 - val_acc: 0.8289\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2863 - acc: 0.8862 - val_loss: 0.4578 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2907 - acc: 0.8838 - val_loss: 0.4430 - val_acc: 0.8411\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2857 - acc: 0.8905 - val_loss: 0.4503 - val_acc: 0.8362\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2851 - acc: 0.8887 - val_loss: 0.4442 - val_acc: 0.8386\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2888 - acc: 0.8869 - val_loss: 0.4520 - val_acc: 0.8289\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2843 - acc: 0.8869 - val_loss: 0.4449 - val_acc: 0.8337\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2838 - acc: 0.8881 - val_loss: 0.4502 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2856 - acc: 0.8869 - val_loss: 0.4437 - val_acc: 0.8313\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2865 - acc: 0.8869 - val_loss: 0.4571 - val_acc: 0.8240\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2825 - acc: 0.8930 - val_loss: 0.4425 - val_acc: 0.8386\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2838 - acc: 0.8930 - val_loss: 0.4443 - val_acc: 0.8337\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2820 - acc: 0.8905 - val_loss: 0.4408 - val_acc: 0.8386\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2874 - acc: 0.8820 - val_loss: 0.4426 - val_acc: 0.8337\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2840 - acc: 0.8942 - val_loss: 0.4500 - val_acc: 0.8386\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2813 - acc: 0.8881 - val_loss: 0.4502 - val_acc: 0.8337\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2812 - acc: 0.8905 - val_loss: 0.4451 - val_acc: 0.8337\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 18us/sample - loss: 0.2803 - acc: 0.8917 - val_loss: 0.4594 - val_acc: 0.8362\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 18us/sample - loss: 0.2811 - acc: 0.8869 - val_loss: 0.4441 - val_acc: 0.8337\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2797 - acc: 0.8905 - val_loss: 0.4561 - val_acc: 0.8313\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2796 - acc: 0.8887 - val_loss: 0.4435 - val_acc: 0.8411\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2798 - acc: 0.8905 - val_loss: 0.4464 - val_acc: 0.8337\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2789 - acc: 0.8905 - val_loss: 0.4468 - val_acc: 0.8337\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2792 - acc: 0.8948 - val_loss: 0.4513 - val_acc: 0.8484\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2799 - acc: 0.8844 - val_loss: 0.4442 - val_acc: 0.8386\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2783 - acc: 0.8893 - val_loss: 0.4474 - val_acc: 0.8484\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2775 - acc: 0.8930 - val_loss: 0.4549 - val_acc: 0.8460\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2781 - acc: 0.8911 - val_loss: 0.4486 - val_acc: 0.8362\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2767 - acc: 0.8899 - val_loss: 0.4494 - val_acc: 0.8509\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2765 - acc: 0.8899 - val_loss: 0.4518 - val_acc: 0.8460\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2761 - acc: 0.8899 - val_loss: 0.4540 - val_acc: 0.8460\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2766 - acc: 0.8936 - val_loss: 0.4429 - val_acc: 0.8386\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2750 - acc: 0.8942 - val_loss: 0.4643 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2795 - acc: 0.8875 - val_loss: 0.4452 - val_acc: 0.8386\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 18us/sample - loss: 0.2770 - acc: 0.8911 - val_loss: 0.4614 - val_acc: 0.8337\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2761 - acc: 0.8924 - val_loss: 0.4440 - val_acc: 0.8411\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2767 - acc: 0.8869 - val_loss: 0.4392 - val_acc: 0.8362\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2783 - acc: 0.8948 - val_loss: 0.4610 - val_acc: 0.8435\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2754 - acc: 0.8924 - val_loss: 0.4413 - val_acc: 0.8337\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2760 - acc: 0.8948 - val_loss: 0.4728 - val_acc: 0.8264\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2769 - acc: 0.8924 - val_loss: 0.4449 - val_acc: 0.8362\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2734 - acc: 0.8917 - val_loss: 0.4565 - val_acc: 0.8362\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2737 - acc: 0.8917 - val_loss: 0.4447 - val_acc: 0.8386\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2716 - acc: 0.8893 - val_loss: 0.4616 - val_acc: 0.8484\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2724 - acc: 0.8917 - val_loss: 0.4477 - val_acc: 0.8411\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2737 - acc: 0.8911 - val_loss: 0.4403 - val_acc: 0.8484\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2734 - acc: 0.8979 - val_loss: 0.4467 - val_acc: 0.8509\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2719 - acc: 0.8985 - val_loss: 0.4497 - val_acc: 0.8435\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2693 - acc: 0.8954 - val_loss: 0.4428 - val_acc: 0.8411\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2704 - acc: 0.8954 - val_loss: 0.4515 - val_acc: 0.8435\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2726 - acc: 0.8856 - val_loss: 0.4409 - val_acc: 0.8362\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2742 - acc: 0.8905 - val_loss: 0.4679 - val_acc: 0.8289\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2725 - acc: 0.8936 - val_loss: 0.4413 - val_acc: 0.8435\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2705 - acc: 0.8966 - val_loss: 0.4562 - val_acc: 0.8386\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2695 - acc: 0.8948 - val_loss: 0.4474 - val_acc: 0.8435\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2687 - acc: 0.8911 - val_loss: 0.4429 - val_acc: 0.8484\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2722 - acc: 0.8979 - val_loss: 0.4593 - val_acc: 0.8362\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2714 - acc: 0.8893 - val_loss: 0.4397 - val_acc: 0.8460\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2689 - acc: 0.8991 - val_loss: 0.4528 - val_acc: 0.8460\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2666 - acc: 0.8930 - val_loss: 0.4469 - val_acc: 0.8509\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2654 - acc: 0.8966 - val_loss: 0.4570 - val_acc: 0.8460\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2688 - acc: 0.8948 - val_loss: 0.4395 - val_acc: 0.8460\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2661 - acc: 0.8960 - val_loss: 0.4603 - val_acc: 0.8337\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2688 - acc: 0.8966 - val_loss: 0.4395 - val_acc: 0.8509\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2660 - acc: 0.8997 - val_loss: 0.4558 - val_acc: 0.8386\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 20us/sample - loss: 0.2679 - acc: 0.8936 - val_loss: 0.4467 - val_acc: 0.8557\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2652 - acc: 0.8960 - val_loss: 0.4534 - val_acc: 0.8582\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2651 - acc: 0.8948 - val_loss: 0.4462 - val_acc: 0.8557\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2662 - acc: 0.8905 - val_loss: 0.4535 - val_acc: 0.8460\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2690 - acc: 0.8917 - val_loss: 0.4532 - val_acc: 0.8435\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2654 - acc: 0.8942 - val_loss: 0.4452 - val_acc: 0.8411\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2676 - acc: 0.8960 - val_loss: 0.4563 - val_acc: 0.8386\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2641 - acc: 0.8948 - val_loss: 0.4393 - val_acc: 0.8533\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2690 - acc: 0.8911 - val_loss: 0.4652 - val_acc: 0.8460\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2655 - acc: 0.8942 - val_loss: 0.4434 - val_acc: 0.8509\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2658 - acc: 0.8936 - val_loss: 0.4477 - val_acc: 0.8533\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2636 - acc: 0.8997 - val_loss: 0.4598 - val_acc: 0.8386\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.2632 - acc: 0.8942 - val_loss: 0.4450 - val_acc: 0.8533\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 18us/sample - loss: 0.2619 - acc: 0.8985 - val_loss: 0.4549 - val_acc: 0.8484\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2640 - acc: 0.8985 - val_loss: 0.4467 - val_acc: 0.8557\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2625 - acc: 0.8948 - val_loss: 0.4433 - val_acc: 0.8533\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2657 - acc: 0.8985 - val_loss: 0.4641 - val_acc: 0.8533\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2674 - acc: 0.8930 - val_loss: 0.4400 - val_acc: 0.8484\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2638 - acc: 0.8948 - val_loss: 0.4435 - val_acc: 0.8557\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2613 - acc: 0.8979 - val_loss: 0.4483 - val_acc: 0.8557\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2611 - acc: 0.8966 - val_loss: 0.4626 - val_acc: 0.8435\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2672 - acc: 0.8911 - val_loss: 0.4468 - val_acc: 0.8460\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2617 - acc: 0.8979 - val_loss: 0.4625 - val_acc: 0.8484\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2604 - acc: 0.8936 - val_loss: 0.4439 - val_acc: 0.8460\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.2635 - acc: 0.8954 - val_loss: 0.4481 - val_acc: 0.8533\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.2611 - acc: 0.9003 - val_loss: 0.4440 - val_acc: 0.8557\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2645 - acc: 0.8942 - val_loss: 0.4600 - val_acc: 0.8460\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2628 - acc: 0.8985 - val_loss: 0.4422 - val_acc: 0.8362\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2630 - acc: 0.8911 - val_loss: 0.4598 - val_acc: 0.8460\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2588 - acc: 0.8966 - val_loss: 0.4452 - val_acc: 0.8435\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2627 - acc: 0.8979 - val_loss: 0.4759 - val_acc: 0.8337\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2625 - acc: 0.8972 - val_loss: 0.4409 - val_acc: 0.8533\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2592 - acc: 0.8997 - val_loss: 0.4622 - val_acc: 0.8435\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 22us/sample - loss: 0.2589 - acc: 0.8979 - val_loss: 0.4473 - val_acc: 0.8533\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.2591 - acc: 0.8979 - val_loss: 0.4451 - val_acc: 0.8509\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2639 - acc: 0.8960 - val_loss: 0.4661 - val_acc: 0.8337\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.2614 - acc: 0.8985 - val_loss: 0.4457 - val_acc: 0.8411\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.2616 - acc: 0.8954 - val_loss: 0.4580 - val_acc: 0.8533\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2590 - acc: 0.8960 - val_loss: 0.4429 - val_acc: 0.8533\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2588 - acc: 0.8936 - val_loss: 0.4473 - val_acc: 0.8533\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2578 - acc: 0.8985 - val_loss: 0.4557 - val_acc: 0.8509\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2544 - acc: 0.8985 - val_loss: 0.4483 - val_acc: 0.8533\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2564 - acc: 0.9021 - val_loss: 0.4615 - val_acc: 0.8533\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2560 - acc: 0.8972 - val_loss: 0.4550 - val_acc: 0.8509\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2563 - acc: 0.8979 - val_loss: 0.4544 - val_acc: 0.8509\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2557 - acc: 0.8966 - val_loss: 0.4422 - val_acc: 0.8509\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.2599 - acc: 0.8954 - val_loss: 0.4604 - val_acc: 0.8557\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2554 - acc: 0.9021 - val_loss: 0.4504 - val_acc: 0.8557\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2549 - acc: 0.8997 - val_loss: 0.4440 - val_acc: 0.8557\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2590 - acc: 0.8997 - val_loss: 0.4594 - val_acc: 0.8509\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2546 - acc: 0.8966 - val_loss: 0.4485 - val_acc: 0.8533\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2567 - acc: 0.9021 - val_loss: 0.4657 - val_acc: 0.8435\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2555 - acc: 0.8985 - val_loss: 0.4483 - val_acc: 0.8484\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2595 - acc: 0.8948 - val_loss: 0.4534 - val_acc: 0.8484\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2568 - acc: 0.9003 - val_loss: 0.4576 - val_acc: 0.8460\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2554 - acc: 0.8972 - val_loss: 0.4549 - val_acc: 0.8509\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2525 - acc: 0.8991 - val_loss: 0.4529 - val_acc: 0.8533\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2525 - acc: 0.9034 - val_loss: 0.4533 - val_acc: 0.8582\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2522 - acc: 0.9034 - val_loss: 0.4503 - val_acc: 0.8509\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2517 - acc: 0.9046 - val_loss: 0.4502 - val_acc: 0.8509\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2520 - acc: 0.9046 - val_loss: 0.4540 - val_acc: 0.8557\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2529 - acc: 0.9058 - val_loss: 0.4534 - val_acc: 0.8606\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2548 - acc: 0.8942 - val_loss: 0.4555 - val_acc: 0.8411\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2542 - acc: 0.8997 - val_loss: 0.4653 - val_acc: 0.8509\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2529 - acc: 0.9015 - val_loss: 0.4488 - val_acc: 0.8484\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2541 - acc: 0.9021 - val_loss: 0.4460 - val_acc: 0.8557\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2508 - acc: 0.9046 - val_loss: 0.4630 - val_acc: 0.8509\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2535 - acc: 0.8972 - val_loss: 0.4534 - val_acc: 0.8533\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2515 - acc: 0.9028 - val_loss: 0.4540 - val_acc: 0.8533\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2514 - acc: 0.9058 - val_loss: 0.4556 - val_acc: 0.8484\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2522 - acc: 0.9021 - val_loss: 0.4659 - val_acc: 0.8435\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2520 - acc: 0.9003 - val_loss: 0.4479 - val_acc: 0.8533\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2510 - acc: 0.9028 - val_loss: 0.4643 - val_acc: 0.8460\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2491 - acc: 0.9021 - val_loss: 0.4505 - val_acc: 0.8509\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2496 - acc: 0.9034 - val_loss: 0.4509 - val_acc: 0.8582\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2486 - acc: 0.9052 - val_loss: 0.4663 - val_acc: 0.8557\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2494 - acc: 0.9021 - val_loss: 0.4564 - val_acc: 0.8484\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2497 - acc: 0.9015 - val_loss: 0.4498 - val_acc: 0.8557\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2475 - acc: 0.8985 - val_loss: 0.4603 - val_acc: 0.8484\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2489 - acc: 0.9003 - val_loss: 0.4474 - val_acc: 0.8411\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2507 - acc: 0.8991 - val_loss: 0.4817 - val_acc: 0.8313\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2519 - acc: 0.9015 - val_loss: 0.4448 - val_acc: 0.8435\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2499 - acc: 0.8985 - val_loss: 0.4697 - val_acc: 0.8484\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2491 - acc: 0.9040 - val_loss: 0.4445 - val_acc: 0.8411\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.2474 - acc: 0.9009 - val_loss: 0.4621 - val_acc: 0.8460\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.2470 - acc: 0.9015 - val_loss: 0.4557 - val_acc: 0.8533\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2461 - acc: 0.9046 - val_loss: 0.4561 - val_acc: 0.8509\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2537 - acc: 0.9009 - val_loss: 0.4757 - val_acc: 0.8411\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2501 - acc: 0.9028 - val_loss: 0.4443 - val_acc: 0.8533\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2476 - acc: 0.9003 - val_loss: 0.4638 - val_acc: 0.8484\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2471 - acc: 0.9052 - val_loss: 0.4544 - val_acc: 0.8557\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2481 - acc: 0.9040 - val_loss: 0.4594 - val_acc: 0.8484\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2489 - acc: 0.8979 - val_loss: 0.4497 - val_acc: 0.8509\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2498 - acc: 0.8966 - val_loss: 0.4582 - val_acc: 0.8582\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2471 - acc: 0.9076 - val_loss: 0.4581 - val_acc: 0.8533\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2466 - acc: 0.9034 - val_loss: 0.4585 - val_acc: 0.8435\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2486 - acc: 0.9040 - val_loss: 0.4633 - val_acc: 0.8509\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2450 - acc: 0.9046 - val_loss: 0.4598 - val_acc: 0.8631\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2462 - acc: 0.9015 - val_loss: 0.4613 - val_acc: 0.8460\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2442 - acc: 0.9064 - val_loss: 0.4568 - val_acc: 0.8533\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.2442 - acc: 0.9015 - val_loss: 0.4572 - val_acc: 0.8606\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2451 - acc: 0.9034 - val_loss: 0.4642 - val_acc: 0.8484\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2470 - acc: 0.9028 - val_loss: 0.4512 - val_acc: 0.8533\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2439 - acc: 0.9034 - val_loss: 0.4548 - val_acc: 0.8509\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2441 - acc: 0.9028 - val_loss: 0.4629 - val_acc: 0.8435\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2435 - acc: 0.9040 - val_loss: 0.4537 - val_acc: 0.8509\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2441 - acc: 0.9058 - val_loss: 0.4739 - val_acc: 0.8460\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2473 - acc: 0.9040 - val_loss: 0.4486 - val_acc: 0.8411\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2437 - acc: 0.9064 - val_loss: 0.4772 - val_acc: 0.8435\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2490 - acc: 0.9028 - val_loss: 0.4460 - val_acc: 0.8411\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2446 - acc: 0.9021 - val_loss: 0.4666 - val_acc: 0.8362\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2433 - acc: 0.9064 - val_loss: 0.4534 - val_acc: 0.8509\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2422 - acc: 0.9052 - val_loss: 0.4670 - val_acc: 0.8557\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2425 - acc: 0.9052 - val_loss: 0.4571 - val_acc: 0.8582\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2419 - acc: 0.9046 - val_loss: 0.4584 - val_acc: 0.8509\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2421 - acc: 0.9076 - val_loss: 0.4498 - val_acc: 0.8509\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2413 - acc: 0.9064 - val_loss: 0.4577 - val_acc: 0.8509\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2457 - acc: 0.8991 - val_loss: 0.4688 - val_acc: 0.8509\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2431 - acc: 0.9021 - val_loss: 0.4551 - val_acc: 0.8533\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2419 - acc: 0.9021 - val_loss: 0.4581 - val_acc: 0.8386\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 17us/sample - loss: 0.2401 - acc: 0.9064 - val_loss: 0.4722 - val_acc: 0.8509\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2422 - acc: 0.9095 - val_loss: 0.4559 - val_acc: 0.8484\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2413 - acc: 0.9058 - val_loss: 0.4518 - val_acc: 0.8557\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2421 - acc: 0.9046 - val_loss: 0.4606 - val_acc: 0.8484\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2407 - acc: 0.9083 - val_loss: 0.4551 - val_acc: 0.8533\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2401 - acc: 0.9046 - val_loss: 0.4477 - val_acc: 0.8582\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 16us/sample - loss: 0.2423 - acc: 0.9046 - val_loss: 0.4530 - val_acc: 0.8557\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2406 - acc: 0.9083 - val_loss: 0.4634 - val_acc: 0.8533\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2396 - acc: 0.9064 - val_loss: 0.4548 - val_acc: 0.8484\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2409 - acc: 0.9009 - val_loss: 0.4577 - val_acc: 0.8533\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2413 - acc: 0.9040 - val_loss: 0.4529 - val_acc: 0.8557\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2405 - acc: 0.9040 - val_loss: 0.4511 - val_acc: 0.8509\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2435 - acc: 0.9046 - val_loss: 0.4543 - val_acc: 0.8509\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2396 - acc: 0.9089 - val_loss: 0.4603 - val_acc: 0.8460\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2411 - acc: 0.9076 - val_loss: 0.4561 - val_acc: 0.8484\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2396 - acc: 0.9064 - val_loss: 0.4527 - val_acc: 0.8533\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2400 - acc: 0.9076 - val_loss: 0.4436 - val_acc: 0.8460\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2415 - acc: 0.9064 - val_loss: 0.4652 - val_acc: 0.8460\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2410 - acc: 0.9052 - val_loss: 0.4609 - val_acc: 0.8533\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2404 - acc: 0.9003 - val_loss: 0.4549 - val_acc: 0.8509\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 12us/sample - loss: 0.2386 - acc: 0.9064 - val_loss: 0.4650 - val_acc: 0.8460\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2381 - acc: 0.9052 - val_loss: 0.4522 - val_acc: 0.8435\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 14us/sample - loss: 0.2398 - acc: 0.9034 - val_loss: 0.4802 - val_acc: 0.8386\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 15us/sample - loss: 0.2425 - acc: 0.9034 - val_loss: 0.4561 - val_acc: 0.8460\n",
            "Train on 1635 samples, validate on 409 samples\n",
            "1635/1635 [==============================] - 0s 13us/sample - loss: 0.2367 - acc: 0.9064 - val_loss: 0.4603 - val_acc: 0.8533\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1Q7_ahVz5yPb",
        "colab_type": "text"
      },
      "source": [
        "Check Overfitting and Stopped Iteration"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "E6OJofS-5yi6",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 663
        },
        "outputId": "1c481498-62ab-40fc-f65c-6426596e4a1a"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "sns.set()\n",
        "plt.figure(num=None, figsize=(16, 8), dpi=90, facecolor='w', edgecolor='k')\n",
        "plt.plot()\n",
        "plt.plot(train_acc)\n",
        "plt.plot(val_acc)\n",
        "plt.title('model accuracy')\n",
        "plt.ylabel('accuracy')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['Train', 'Test'], loc='upper left')\n",
        "plt.show()"
      ],
      "execution_count": 50,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAABKMAAAJyCAYAAAAYSAOhAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAoGz+Pwg2IbKPc/q/AAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 1440x720 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "baX39XFu2q0o",
        "colab_type": "text"
      },
      "source": [
        "Change test accuracy calculation"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ITMM1GFW2A5D",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "4af4730e-75da-4a86-8ec2-f8f2b6c8124b"
      },
      "source": [
        "y_predict = model.predict(X_test_norm)\n",
        "y_predict = np.argmax(y_predict, 1)\n",
        "sum(y_predict == y_test)/len(y_test)"
      ],
      "execution_count": 49,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.8359375"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 49
        }
      ]
    }
  ]
}