tensorflow-nn-zt.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd  \n",
    "import numpy as np\n",
    "\n",
    "df = pd.read_csv(r'dataSet/ONLINE EDUCATION SYSTEM REVIEW.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "# preprocessing data\n",
    "# drop the grade column\n",
    "df = df.drop(columns=\"Average marks scored before pandemic in traditional classroom\")\n",
    "# split data to inputs and target\n",
    "inputs = df.drop(columns=[\"Performance in online\",\"Your level of satisfaction in Online Education\"])\n",
    "inputs = pd.DataFrame(inputs)\n",
    "inputs.head(5)\n",
    "target = df[\"Your level of satisfaction in Online Education\"]\n",
    "target = pd.DataFrame(target)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "\n",
    "def print_confusion_matrix(confusion_matrix, axes, class_label, class_names, fontsize=14):\n",
    "\n",
    "    df_cm = pd.DataFrame(\n",
    "        confusion_matrix, index=class_names, columns=class_names,\n",
    "    )\n",
    "\n",
    "    try:\n",
    "        heatmap = sns.heatmap(df_cm, annot=True, fmt=\"d\", cbar=False, ax=axes)\n",
    "    except ValueError:\n",
    "        raise ValueError(\"Confusion matrix values must be integers.\")\n",
    "    heatmap.yaxis.set_ticklabels(heatmap.yaxis.get_ticklabels(), rotation=0, ha='right', fontsize=fontsize)\n",
    "    heatmap.xaxis.set_ticklabels(heatmap.xaxis.get_ticklabels(), rotation=45, ha='right', fontsize=fontsize)\n",
    "    axes.set_ylabel('True label')\n",
    "    axes.set_xlabel('Predicted label')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [],
   "source": [
    "#preprocssing inputs\n",
    "#declare catergory columns\n",
    "category = [\"Gender\", \"Home Location\", \"Level of Education\",\"Device type used to attend classes\",\"Economic status\",\"Are you involved in any sports?\",\"Do elderly people monitor you?\",\"Interested in Gaming?\",\"Have separate room for studying?\",\"Engaged in group studies?\",\"Interested in?\"]\n",
    "\n",
    "# print(\"before encoded\")\n",
    "# for i in category:\n",
    "#     unique = np.unique(inputs[i])\n",
    "#     print(i)\n",
    "#     print(unique)\n",
    "\n",
    "for i in category:\n",
    "    inputs[i]=inputs[i].astype(\"category\")\n",
    "    inputs[i]=inputs[i].replace(np.nan, 0)\n",
    "    # number encoded the category item\n",
    "    unique = np.unique(inputs[i])\n",
    "    for j,z in zip(unique,range(len(unique))):\n",
    "        inputs[i] = inputs[i].replace(j,z)\n",
    "\n",
    "# print(\"after encoded\")\n",
    "# for i in category:\n",
    "#     unique = np.unique(inputs[i])\n",
    "#     print(i)\n",
    "#     print(unique)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [],
   "source": [
    "#convert inputs into one-hot-encoded\n",
    "from sklearn.preprocessing import OneHotEncoder\n",
    "import os\n",
    "base =os.curdir\n",
    "\n",
    "def one_hot_encode(column, df):\n",
    "    encoder = OneHotEncoder(sparse = False)\n",
    "    seleted_col = df[[column]]\n",
    "    encoder.fit(seleted_col)\n",
    "    encoded = encoder.transform(seleted_col)\n",
    "    encoded = pd.DataFrame(data=encoded,columns=encoder.categories_)\n",
    "    return encoded\n",
    "\n",
    "def one_hot_encoding(inputs):\n",
    "    for x in inputs:\n",
    "        encoded_col = one_hot_encode(x,inputs)\n",
    "        inputs.drop(x,axis=1,inplace =True)\n",
    "        for y in encoded_col.columns:\n",
    "            path = str(x)+str(y)\n",
    "            path = str(path)\n",
    "            path = path.replace(\"'\",\"\")\n",
    "            path = path.replace(\",\",\"\")\n",
    "            inputs.insert(len(inputs.columns),path,encoded_col[y])\n",
    "    return inputs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1033, 20)\n",
      "(1033, 20)\n"
     ]
    },
    {
     "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>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "      <th>6</th>\n",
       "      <th>7</th>\n",
       "      <th>8</th>\n",
       "      <th>9</th>\n",
       "      <th>10</th>\n",
       "      <th>11</th>\n",
       "      <th>12</th>\n",
       "      <th>13</th>\n",
       "      <th>14</th>\n",
       "      <th>15</th>\n",
       "      <th>16</th>\n",
       "      <th>17</th>\n",
       "      <th>18</th>\n",
       "      <th>19</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.290323</td>\n",
       "      <td>0.526316</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.250</td>\n",
       "      <td>1.00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.222222</td>\n",
       "      <td>0.555556</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.322581</td>\n",
       "      <td>0.315789</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.250</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>0.444444</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.290323</td>\n",
       "      <td>0.210526</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.375</td>\n",
       "      <td>0.25</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.555556</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.290323</td>\n",
       "      <td>0.210526</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.250</td>\n",
       "      <td>0.75</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.222222</td>\n",
       "      <td>0.555556</td>\n",
       "      <td>0.111111</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.25</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.290323</td>\n",
       "      <td>0.210526</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.250</td>\n",
       "      <td>0.50</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.777778</td>\n",
       "      <td>0.666667</td>\n",
       "      <td>0.111111</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.5</td>\n",
       "      <td>0.50</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    0    1    2         3         4    5    6      7     8    9    10  \\\n",
       "0  1.0  1.0  1.0  0.290323  0.526316  0.5  0.0  0.250  1.00  0.0  1.0   \n",
       "1  1.0  1.0  1.0  0.322581  0.315789  0.5  0.0  0.250  0.00  1.0  1.0   \n",
       "2  1.0  0.0  1.0  0.290323  0.210526  0.5  0.0  0.375  0.25  0.0  1.0   \n",
       "3  1.0  1.0  1.0  0.290323  0.210526  0.5  0.0  0.250  0.75  1.0  1.0   \n",
       "4  1.0  0.0  1.0  0.290323  0.210526  0.5  0.0  0.250  0.50  0.0  0.0   \n",
       "\n",
       "         11        12        13   14   15   16   17    18   19  \n",
       "0  0.222222  0.555556  0.000000  0.0  0.0  0.0  0.0  0.00  0.5  \n",
       "1  0.666667  0.444444  0.000000  1.0  1.0  0.0  0.0  0.00  1.0  \n",
       "2  0.555556  0.666667  0.000000  0.0  1.0  0.0  0.0  0.00  0.0  \n",
       "3  0.222222  0.555556  0.111111  0.0  0.0  1.0  0.0  0.25  1.0  \n",
       "4  0.777778  0.666667  0.111111  1.0  1.0  1.0  0.5  0.50  0.0  "
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# inputs = one_hot_encoding(inputs)\n",
    "# inputs.head(5)\n",
    "from sklearn.preprocessing import MinMaxScaler\n",
    "# scale inputs\n",
    "scaler = MinMaxScaler(feature_range=(0, 1))\n",
    "scaled = scaler.fit_transform(inputs)\n",
    "print(inputs.shape)\n",
    "print(scaled.shape)\n",
    "inputs = pd.DataFrame(scaled)\n",
    "inputs.head(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "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>Your level of satisfaction in Online Education(Average)</th>\n",
       "      <th>Your level of satisfaction in Online Education(Bad)</th>\n",
       "      <th>Your level of satisfaction in Online Education(Good)</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\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": [
       "   Your level of satisfaction in Online Education(Average)  \\\n",
       "0                                                1.0         \n",
       "1                                                0.0         \n",
       "2                                                0.0         \n",
       "3                                                0.0         \n",
       "4                                                1.0         \n",
       "\n",
       "   Your level of satisfaction in Online Education(Bad)  \\\n",
       "0                                                0.0     \n",
       "1                                                1.0     \n",
       "2                                                1.0     \n",
       "3                                                1.0     \n",
       "4                                                0.0     \n",
       "\n",
       "   Your level of satisfaction in Online Education(Good)  \n",
       "0                                                0.0     \n",
       "1                                                0.0     \n",
       "2                                                0.0     \n",
       "3                                                0.0     \n",
       "4                                                0.0     "
      ]
     },
     "execution_count": 55,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#preprocessing target\n",
    "unique = np.unique(target)\n",
    "# number encoded target\n",
    "# for i, j in zip(unique,range(len(unique))):\n",
    "#     target = target.replace(i,j)\n",
    "\n",
    "target = one_hot_encoding(target)\n",
    "    \n",
    "target.head(5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1033, 20)\n",
      "(1033, 3)\n"
     ]
    }
   ],
   "source": [
    "inputs= inputs.values\n",
    "target = target.values\n",
    "print(inputs.shape)\n",
    "print(target.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "from keras.wrappers.scikit_learn import KerasClassifier\n",
    "from sklearn.model_selection import cross_val_score\n",
    "from sklearn.model_selection import KFold\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "import tensorflow as tf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras import layers\n",
    "model = Sequential([\n",
    "    layers.Dense(15, input_dim = 20, activation = 'relu'), # Rectified Linear Unit Activation Function\n",
    "    layers.Dense(10, activation = 'relu'),\n",
    "    layers.Dense(5, activation = 'relu'),\n",
    "    layers.Dropout(0.2),\n",
    "    layers.Dense(3, activation = 'softmax') # Softmax for multi-class classification\n",
    "])\n",
    "    # Compile model here\n",
    "model.compile(optimizer='adam',\n",
    "              loss = 'categorical_crossentropy',\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split \n",
    "X_train, X_test, y_train, y_test = train_test_split(inputs, target, test_size=0.2, random_state=0)   "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [],
   "source": [
    "# seed = 10\n",
    "# np.random.seed(seed)\n",
    "# estimator = KerasClassifier(build_fn = model, epochs = 100, batch_size = 10, verbose = 0)\n",
    "\n",
    "# kfold = KFold(n_splits = 10, shuffle = True, random_state = seed)\n",
    "\n",
    "# results = cross_val_score(estimator, inputs, target, cv = kfold)\n",
    "# print(\"Result: %.2f%% (%.2f%%)\" % (results.mean()*100, results.std()*100))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/250\n",
      "9/9 [==============================] - 0s 16ms/step - loss: 1.1024 - accuracy: 0.4056 - val_loss: 1.0992 - val_accuracy: 0.4251\n",
      "Epoch 2/250\n",
      "9/9 [==============================] - 0s 3ms/step - loss: 1.0930 - accuracy: 0.4976 - val_loss: 1.0951 - val_accuracy: 0.4203\n",
      "Epoch 3/250\n",
      "9/9 [==============================] - 0s 6ms/step - loss: 1.0878 - accuracy: 0.4976 - val_loss: 1.0926 - val_accuracy: 0.4493\n",
      "Epoch 4/250\n",
      "9/9 [==============================] - 0s 6ms/step - loss: 1.0839 - accuracy: 0.5133 - val_loss: 1.0895 - val_accuracy: 0.4589\n",
      "Epoch 5/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 1.0784 - accuracy: 0.5266 - val_loss: 1.0872 - val_accuracy: 0.4493\n",
      "Epoch 6/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 1.0741 - accuracy: 0.5194 - val_loss: 1.0848 - val_accuracy: 0.4493\n",
      "Epoch 7/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 1.0717 - accuracy: 0.5230 - val_loss: 1.0814 - val_accuracy: 0.4589\n",
      "Epoch 8/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 1.0656 - accuracy: 0.5375 - val_loss: 1.0787 - val_accuracy: 0.4686\n",
      "Epoch 9/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 1.0596 - accuracy: 0.5363 - val_loss: 1.0778 - val_accuracy: 0.4493\n",
      "Epoch 10/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 1.0597 - accuracy: 0.5085 - val_loss: 1.0737 - val_accuracy: 0.4638\n",
      "Epoch 11/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 1.0539 - accuracy: 0.5303 - val_loss: 1.0703 - val_accuracy: 0.4686\n",
      "Epoch 12/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 1.0463 - accuracy: 0.5339 - val_loss: 1.0676 - val_accuracy: 0.4686\n",
      "Epoch 13/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 1.0407 - accuracy: 0.5242 - val_loss: 1.0636 - val_accuracy: 0.4783\n",
      "Epoch 14/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 1.0333 - accuracy: 0.5315 - val_loss: 1.0567 - val_accuracy: 0.4928\n",
      "Epoch 15/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 1.0238 - accuracy: 0.5363 - val_loss: 1.0466 - val_accuracy: 0.4831\n",
      "Epoch 16/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 1.0162 - accuracy: 0.5508 - val_loss: 1.0350 - val_accuracy: 0.5121\n",
      "Epoch 17/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 1.0073 - accuracy: 0.5605 - val_loss: 1.0245 - val_accuracy: 0.4976\n",
      "Epoch 18/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.9936 - accuracy: 0.5654 - val_loss: 1.0153 - val_accuracy: 0.4831\n",
      "Epoch 19/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.9768 - accuracy: 0.5642 - val_loss: 1.0004 - val_accuracy: 0.5362\n",
      "Epoch 20/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.9653 - accuracy: 0.5799 - val_loss: 0.9894 - val_accuracy: 0.5459\n",
      "Epoch 21/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.9516 - accuracy: 0.5944 - val_loss: 0.9818 - val_accuracy: 0.5411\n",
      "Epoch 22/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.9469 - accuracy: 0.5787 - val_loss: 0.9712 - val_accuracy: 0.5459\n",
      "Epoch 23/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.9307 - accuracy: 0.5920 - val_loss: 0.9607 - val_accuracy: 0.5652\n",
      "Epoch 24/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.9202 - accuracy: 0.6029 - val_loss: 0.9550 - val_accuracy: 0.5266\n",
      "Epoch 25/250\n",
      "9/9 [==============================] - 0s 3ms/step - loss: 0.9146 - accuracy: 0.5763 - val_loss: 0.9472 - val_accuracy: 0.5362\n",
      "Epoch 26/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.9110 - accuracy: 0.5738 - val_loss: 0.9481 - val_accuracy: 0.5169\n",
      "Epoch 27/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.8997 - accuracy: 0.5593 - val_loss: 0.9419 - val_accuracy: 0.5266\n",
      "Epoch 28/250\n",
      "9/9 [==============================] - 0s 3ms/step - loss: 0.8936 - accuracy: 0.5763 - val_loss: 0.9357 - val_accuracy: 0.5411\n",
      "Epoch 29/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.8919 - accuracy: 0.5690 - val_loss: 0.9364 - val_accuracy: 0.5411\n",
      "Epoch 30/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.8809 - accuracy: 0.5642 - val_loss: 0.9327 - val_accuracy: 0.5362\n",
      "Epoch 31/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.8765 - accuracy: 0.5811 - val_loss: 0.9319 - val_accuracy: 0.5362\n",
      "Epoch 32/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.8873 - accuracy: 0.5690 - val_loss: 0.9287 - val_accuracy: 0.5314\n",
      "Epoch 33/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.8780 - accuracy: 0.5690 - val_loss: 0.9314 - val_accuracy: 0.5314\n",
      "Epoch 34/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.8583 - accuracy: 0.5932 - val_loss: 0.9281 - val_accuracy: 0.5411\n",
      "Epoch 35/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.8612 - accuracy: 0.5884 - val_loss: 0.9249 - val_accuracy: 0.5362\n",
      "Epoch 36/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.8696 - accuracy: 0.5751 - val_loss: 0.9240 - val_accuracy: 0.5362\n",
      "Epoch 37/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.8659 - accuracy: 0.5811 - val_loss: 0.9290 - val_accuracy: 0.5411\n",
      "Epoch 38/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.8506 - accuracy: 0.5714 - val_loss: 0.9389 - val_accuracy: 0.5362\n",
      "Epoch 39/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.8638 - accuracy: 0.5787 - val_loss: 0.9251 - val_accuracy: 0.5362\n",
      "Epoch 40/250\n",
      "9/9 [==============================] - 0s 5ms/step - loss: 0.8563 - accuracy: 0.5751 - val_loss: 0.9253 - val_accuracy: 0.5459\n",
      "Epoch 41/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.8526 - accuracy: 0.6138 - val_loss: 0.9221 - val_accuracy: 0.5507\n",
      "Epoch 42/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.8465 - accuracy: 0.6138 - val_loss: 0.9233 - val_accuracy: 0.5507\n",
      "Epoch 43/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.8419 - accuracy: 0.6029 - val_loss: 0.9257 - val_accuracy: 0.5604\n",
      "Epoch 44/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.8371 - accuracy: 0.6186 - val_loss: 0.9210 - val_accuracy: 0.5749\n",
      "Epoch 45/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.8364 - accuracy: 0.6077 - val_loss: 0.9206 - val_accuracy: 0.5700\n",
      "Epoch 46/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.8436 - accuracy: 0.6138 - val_loss: 0.9152 - val_accuracy: 0.5894\n",
      "Epoch 47/250\n",
      "9/9 [==============================] - 0s 3ms/step - loss: 0.8323 - accuracy: 0.6053 - val_loss: 0.9302 - val_accuracy: 0.5942\n",
      "Epoch 48/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.8316 - accuracy: 0.6005 - val_loss: 0.9149 - val_accuracy: 0.5749\n",
      "Epoch 49/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.8261 - accuracy: 0.6090 - val_loss: 0.9134 - val_accuracy: 0.5942\n",
      "Epoch 50/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.8287 - accuracy: 0.6199 - val_loss: 0.9197 - val_accuracy: 0.6184\n",
      "Epoch 51/250\n",
      "9/9 [==============================] - 0s 5ms/step - loss: 0.8367 - accuracy: 0.5872 - val_loss: 0.9076 - val_accuracy: 0.6087\n",
      "Epoch 52/250\n",
      "9/9 [==============================] - 0s 5ms/step - loss: 0.8266 - accuracy: 0.6090 - val_loss: 0.9148 - val_accuracy: 0.6039\n",
      "Epoch 53/250\n",
      "9/9 [==============================] - 0s 5ms/step - loss: 0.8457 - accuracy: 0.6138 - val_loss: 0.9255 - val_accuracy: 0.6184\n",
      "Epoch 54/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.8157 - accuracy: 0.6271 - val_loss: 0.9105 - val_accuracy: 0.6039\n",
      "Epoch 55/250\n",
      "9/9 [==============================] - 0s 3ms/step - loss: 0.8060 - accuracy: 0.6247 - val_loss: 0.9224 - val_accuracy: 0.6135\n",
      "Epoch 56/250\n",
      "9/9 [==============================] - 0s 5ms/step - loss: 0.8160 - accuracy: 0.6126 - val_loss: 0.9238 - val_accuracy: 0.6087\n",
      "Epoch 57/250\n",
      "9/9 [==============================] - 0s 6ms/step - loss: 0.8055 - accuracy: 0.6283 - val_loss: 0.9121 - val_accuracy: 0.6087\n",
      "Epoch 58/250\n",
      "9/9 [==============================] - 0s 5ms/step - loss: 0.8073 - accuracy: 0.6174 - val_loss: 0.9201 - val_accuracy: 0.5942\n",
      "Epoch 59/250\n",
      "9/9 [==============================] - 0s 5ms/step - loss: 0.8202 - accuracy: 0.6174 - val_loss: 0.9167 - val_accuracy: 0.5942\n",
      "Epoch 60/250\n",
      "9/9 [==============================] - 0s 5ms/step - loss: 0.8186 - accuracy: 0.6174 - val_loss: 0.9123 - val_accuracy: 0.6039\n",
      "Epoch 61/250\n",
      "9/9 [==============================] - 0s 7ms/step - loss: 0.8059 - accuracy: 0.6150 - val_loss: 0.9198 - val_accuracy: 0.6039\n",
      "Epoch 62/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.8058 - accuracy: 0.6259 - val_loss: 0.9192 - val_accuracy: 0.6039\n",
      "Epoch 63/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.8046 - accuracy: 0.6186 - val_loss: 0.9171 - val_accuracy: 0.5942\n",
      "Epoch 64/250\n",
      "9/9 [==============================] - 0s 5ms/step - loss: 0.8074 - accuracy: 0.6114 - val_loss: 0.9172 - val_accuracy: 0.5990\n",
      "Epoch 65/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.8071 - accuracy: 0.6077 - val_loss: 0.9189 - val_accuracy: 0.6039\n",
      "Epoch 66/250\n",
      "9/9 [==============================] - 0s 5ms/step - loss: 0.8043 - accuracy: 0.6295 - val_loss: 0.9159 - val_accuracy: 0.5942\n",
      "Epoch 67/250\n",
      "9/9 [==============================] - 0s 5ms/step - loss: 0.8050 - accuracy: 0.6259 - val_loss: 0.9257 - val_accuracy: 0.5990\n",
      "Epoch 68/250\n",
      "9/9 [==============================] - 0s 5ms/step - loss: 0.7999 - accuracy: 0.6077 - val_loss: 0.9186 - val_accuracy: 0.6087\n",
      "Epoch 69/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.8121 - accuracy: 0.6053 - val_loss: 0.9200 - val_accuracy: 0.6039\n",
      "Epoch 70/250\n",
      "9/9 [==============================] - 0s 6ms/step - loss: 0.7893 - accuracy: 0.6308 - val_loss: 0.9277 - val_accuracy: 0.5942\n",
      "Epoch 71/250\n",
      "9/9 [==============================] - 0s 6ms/step - loss: 0.8079 - accuracy: 0.6077 - val_loss: 0.9248 - val_accuracy: 0.6039\n",
      "Epoch 72/250\n",
      "9/9 [==============================] - 0s 8ms/step - loss: 0.8044 - accuracy: 0.6271 - val_loss: 0.9203 - val_accuracy: 0.6087\n",
      "Epoch 73/250\n",
      "9/9 [==============================] - 0s 5ms/step - loss: 0.7997 - accuracy: 0.6271 - val_loss: 0.9239 - val_accuracy: 0.6039\n",
      "Epoch 74/250\n",
      "9/9 [==============================] - 0s 8ms/step - loss: 0.7988 - accuracy: 0.6223 - val_loss: 0.9213 - val_accuracy: 0.6087\n",
      "Epoch 75/250\n",
      "9/9 [==============================] - 0s 7ms/step - loss: 0.8060 - accuracy: 0.5993 - val_loss: 0.9220 - val_accuracy: 0.6087\n",
      "Epoch 76/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.8050 - accuracy: 0.6102 - val_loss: 0.9244 - val_accuracy: 0.6087\n",
      "Epoch 77/250\n",
      "9/9 [==============================] - 0s 5ms/step - loss: 0.8002 - accuracy: 0.6223 - val_loss: 0.9216 - val_accuracy: 0.5990\n",
      "Epoch 78/250\n",
      "9/9 [==============================] - 0s 5ms/step - loss: 0.7883 - accuracy: 0.6271 - val_loss: 0.9331 - val_accuracy: 0.5894\n",
      "Epoch 79/250\n",
      "9/9 [==============================] - 0s 5ms/step - loss: 0.7989 - accuracy: 0.6308 - val_loss: 0.9257 - val_accuracy: 0.5990\n",
      "Epoch 80/250\n",
      "9/9 [==============================] - 0s 6ms/step - loss: 0.7910 - accuracy: 0.6114 - val_loss: 0.9266 - val_accuracy: 0.6039\n",
      "Epoch 81/250\n",
      "9/9 [==============================] - 0s 7ms/step - loss: 0.8082 - accuracy: 0.6199 - val_loss: 0.9271 - val_accuracy: 0.5990\n",
      "Epoch 82/250\n",
      "9/9 [==============================] - 0s 5ms/step - loss: 0.7898 - accuracy: 0.6271 - val_loss: 0.9287 - val_accuracy: 0.5990\n",
      "Epoch 83/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7928 - accuracy: 0.6199 - val_loss: 0.9286 - val_accuracy: 0.5990\n",
      "Epoch 84/250\n",
      "9/9 [==============================] - 0s 7ms/step - loss: 0.7920 - accuracy: 0.6295 - val_loss: 0.9335 - val_accuracy: 0.5942\n",
      "Epoch 85/250\n",
      "9/9 [==============================] - 0s 7ms/step - loss: 0.7962 - accuracy: 0.6150 - val_loss: 0.9269 - val_accuracy: 0.6039\n",
      "Epoch 86/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.8001 - accuracy: 0.6077 - val_loss: 0.9301 - val_accuracy: 0.6039\n",
      "Epoch 87/250\n",
      "9/9 [==============================] - 0s 5ms/step - loss: 0.7906 - accuracy: 0.6138 - val_loss: 0.9366 - val_accuracy: 0.5894\n",
      "Epoch 88/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7962 - accuracy: 0.6271 - val_loss: 0.9335 - val_accuracy: 0.5990\n",
      "Epoch 89/250\n",
      "9/9 [==============================] - 0s 5ms/step - loss: 0.7962 - accuracy: 0.6162 - val_loss: 0.9291 - val_accuracy: 0.6039\n",
      "Epoch 90/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7933 - accuracy: 0.6162 - val_loss: 0.9364 - val_accuracy: 0.5990\n",
      "Epoch 91/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7812 - accuracy: 0.6199 - val_loss: 0.9319 - val_accuracy: 0.5990\n",
      "Epoch 92/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7931 - accuracy: 0.6211 - val_loss: 0.9314 - val_accuracy: 0.5990\n",
      "Epoch 93/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7874 - accuracy: 0.6295 - val_loss: 0.9342 - val_accuracy: 0.5990\n",
      "Epoch 94/250\n",
      "9/9 [==============================] - 0s 3ms/step - loss: 0.7891 - accuracy: 0.6223 - val_loss: 0.9335 - val_accuracy: 0.5990\n",
      "Epoch 95/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7829 - accuracy: 0.6199 - val_loss: 0.9310 - val_accuracy: 0.6087\n",
      "Epoch 96/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7851 - accuracy: 0.6320 - val_loss: 0.9288 - val_accuracy: 0.5990\n",
      "Epoch 97/250\n",
      "9/9 [==============================] - 0s 5ms/step - loss: 0.7916 - accuracy: 0.6344 - val_loss: 0.9314 - val_accuracy: 0.6087\n",
      "Epoch 98/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7615 - accuracy: 0.6489 - val_loss: 0.9335 - val_accuracy: 0.6087\n",
      "Epoch 99/250\n",
      "9/9 [==============================] - 0s 7ms/step - loss: 0.7626 - accuracy: 0.6416 - val_loss: 0.9345 - val_accuracy: 0.5990\n",
      "Epoch 100/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7882 - accuracy: 0.6380 - val_loss: 0.9452 - val_accuracy: 0.6039\n",
      "Epoch 101/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7928 - accuracy: 0.5969 - val_loss: 0.9404 - val_accuracy: 0.5990\n",
      "Epoch 102/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7589 - accuracy: 0.6295 - val_loss: 0.9383 - val_accuracy: 0.6039\n",
      "Epoch 103/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7790 - accuracy: 0.6247 - val_loss: 0.9444 - val_accuracy: 0.5894\n",
      "Epoch 104/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7749 - accuracy: 0.6392 - val_loss: 0.9429 - val_accuracy: 0.5894\n",
      "Epoch 105/250\n",
      "9/9 [==============================] - 0s 5ms/step - loss: 0.7838 - accuracy: 0.6271 - val_loss: 0.9390 - val_accuracy: 0.5990\n",
      "Epoch 106/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7741 - accuracy: 0.6489 - val_loss: 0.9350 - val_accuracy: 0.5990\n",
      "Epoch 107/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7796 - accuracy: 0.6138 - val_loss: 0.9369 - val_accuracy: 0.5990\n",
      "Epoch 108/250\n",
      "9/9 [==============================] - 0s 3ms/step - loss: 0.7748 - accuracy: 0.6247 - val_loss: 0.9400 - val_accuracy: 0.6087\n",
      "Epoch 109/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7719 - accuracy: 0.6186 - val_loss: 0.9476 - val_accuracy: 0.5942\n",
      "Epoch 110/250\n",
      "9/9 [==============================] - 0s 3ms/step - loss: 0.7810 - accuracy: 0.6223 - val_loss: 0.9437 - val_accuracy: 0.5990\n",
      "Epoch 111/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7672 - accuracy: 0.6453 - val_loss: 0.9402 - val_accuracy: 0.6135\n",
      "Epoch 112/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7692 - accuracy: 0.6404 - val_loss: 0.9423 - val_accuracy: 0.6039\n",
      "Epoch 113/250\n",
      "9/9 [==============================] - 0s 3ms/step - loss: 0.7773 - accuracy: 0.6283 - val_loss: 0.9433 - val_accuracy: 0.6039\n",
      "Epoch 114/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7740 - accuracy: 0.6114 - val_loss: 0.9433 - val_accuracy: 0.5942\n",
      "Epoch 115/250\n",
      "9/9 [==============================] - 0s 3ms/step - loss: 0.7732 - accuracy: 0.6344 - val_loss: 0.9440 - val_accuracy: 0.6039\n",
      "Epoch 116/250\n",
      "9/9 [==============================] - 0s 3ms/step - loss: 0.7733 - accuracy: 0.6174 - val_loss: 0.9429 - val_accuracy: 0.6039\n",
      "Epoch 117/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7749 - accuracy: 0.6404 - val_loss: 0.9424 - val_accuracy: 0.6087\n",
      "Epoch 118/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7740 - accuracy: 0.6308 - val_loss: 0.9554 - val_accuracy: 0.5942\n",
      "Epoch 119/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7572 - accuracy: 0.6247 - val_loss: 0.9371 - val_accuracy: 0.6039\n",
      "Epoch 120/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7555 - accuracy: 0.6259 - val_loss: 0.9469 - val_accuracy: 0.6039\n",
      "Epoch 121/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7660 - accuracy: 0.6489 - val_loss: 0.9606 - val_accuracy: 0.5894\n",
      "Epoch 122/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7690 - accuracy: 0.6404 - val_loss: 0.9473 - val_accuracy: 0.6135\n",
      "Epoch 123/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7660 - accuracy: 0.6199 - val_loss: 0.9513 - val_accuracy: 0.5990\n",
      "Epoch 124/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7516 - accuracy: 0.6477 - val_loss: 0.9492 - val_accuracy: 0.6135\n",
      "Epoch 125/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7654 - accuracy: 0.6283 - val_loss: 0.9476 - val_accuracy: 0.6135\n",
      "Epoch 126/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7626 - accuracy: 0.6441 - val_loss: 0.9523 - val_accuracy: 0.5942\n",
      "Epoch 127/250\n",
      "9/9 [==============================] - 0s 3ms/step - loss: 0.7613 - accuracy: 0.6429 - val_loss: 0.9514 - val_accuracy: 0.5845\n",
      "Epoch 128/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7577 - accuracy: 0.6392 - val_loss: 0.9534 - val_accuracy: 0.5990\n",
      "Epoch 129/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7700 - accuracy: 0.6295 - val_loss: 0.9526 - val_accuracy: 0.5990\n",
      "Epoch 130/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7410 - accuracy: 0.6368 - val_loss: 0.9439 - val_accuracy: 0.5990\n",
      "Epoch 131/250\n",
      "9/9 [==============================] - 0s 5ms/step - loss: 0.7642 - accuracy: 0.6332 - val_loss: 0.9621 - val_accuracy: 0.6039\n",
      "Epoch 132/250\n",
      "9/9 [==============================] - 0s 3ms/step - loss: 0.7624 - accuracy: 0.6332 - val_loss: 0.9527 - val_accuracy: 0.6087\n",
      "Epoch 133/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7690 - accuracy: 0.6259 - val_loss: 0.9564 - val_accuracy: 0.6039\n",
      "Epoch 134/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7528 - accuracy: 0.6453 - val_loss: 0.9482 - val_accuracy: 0.6087\n",
      "Epoch 135/250\n",
      "9/9 [==============================] - 0s 3ms/step - loss: 0.7594 - accuracy: 0.6392 - val_loss: 0.9505 - val_accuracy: 0.5990\n",
      "Epoch 136/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7557 - accuracy: 0.6525 - val_loss: 0.9620 - val_accuracy: 0.5942\n",
      "Epoch 137/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7586 - accuracy: 0.6162 - val_loss: 0.9475 - val_accuracy: 0.5894\n",
      "Epoch 138/250\n",
      "9/9 [==============================] - 0s 10ms/step - loss: 0.7529 - accuracy: 0.6392 - val_loss: 0.9593 - val_accuracy: 0.5990\n",
      "Epoch 139/250\n",
      "9/9 [==============================] - 0s 7ms/step - loss: 0.7483 - accuracy: 0.6574 - val_loss: 0.9606 - val_accuracy: 0.5990\n",
      "Epoch 140/250\n",
      "9/9 [==============================] - 0s 5ms/step - loss: 0.7528 - accuracy: 0.6368 - val_loss: 0.9527 - val_accuracy: 0.6087\n",
      "Epoch 141/250\n",
      "9/9 [==============================] - 0s 3ms/step - loss: 0.7527 - accuracy: 0.6404 - val_loss: 0.9581 - val_accuracy: 0.5990\n",
      "Epoch 142/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7654 - accuracy: 0.6174 - val_loss: 0.9612 - val_accuracy: 0.6087\n",
      "Epoch 143/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7372 - accuracy: 0.6501 - val_loss: 0.9609 - val_accuracy: 0.6087\n",
      "Epoch 144/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7512 - accuracy: 0.6465 - val_loss: 0.9592 - val_accuracy: 0.5990\n",
      "Epoch 145/250\n",
      "9/9 [==============================] - 0s 3ms/step - loss: 0.7715 - accuracy: 0.6320 - val_loss: 0.9615 - val_accuracy: 0.6135\n",
      "Epoch 146/250\n",
      "9/9 [==============================] - 0s 3ms/step - loss: 0.7430 - accuracy: 0.6404 - val_loss: 0.9612 - val_accuracy: 0.6135\n",
      "Epoch 147/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7605 - accuracy: 0.6429 - val_loss: 0.9558 - val_accuracy: 0.5990\n",
      "Epoch 148/250\n",
      "9/9 [==============================] - 0s 3ms/step - loss: 0.7517 - accuracy: 0.6344 - val_loss: 0.9603 - val_accuracy: 0.6039\n",
      "Epoch 149/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7417 - accuracy: 0.6610 - val_loss: 0.9612 - val_accuracy: 0.6087\n",
      "Epoch 150/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7525 - accuracy: 0.6429 - val_loss: 0.9537 - val_accuracy: 0.5990\n",
      "Epoch 151/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7536 - accuracy: 0.6429 - val_loss: 0.9606 - val_accuracy: 0.6087\n",
      "Epoch 152/250\n",
      "9/9 [==============================] - 0s 3ms/step - loss: 0.7413 - accuracy: 0.6429 - val_loss: 0.9582 - val_accuracy: 0.5942\n",
      "Epoch 153/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7515 - accuracy: 0.6295 - val_loss: 0.9603 - val_accuracy: 0.5942\n",
      "Epoch 154/250\n",
      "9/9 [==============================] - 0s 3ms/step - loss: 0.7453 - accuracy: 0.6477 - val_loss: 0.9733 - val_accuracy: 0.6039\n",
      "Epoch 155/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7468 - accuracy: 0.6332 - val_loss: 0.9606 - val_accuracy: 0.6039\n",
      "Epoch 156/250\n",
      "9/9 [==============================] - 0s 3ms/step - loss: 0.7452 - accuracy: 0.6501 - val_loss: 0.9584 - val_accuracy: 0.5894\n",
      "Epoch 157/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7441 - accuracy: 0.6368 - val_loss: 0.9647 - val_accuracy: 0.5990\n",
      "Epoch 158/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7458 - accuracy: 0.6404 - val_loss: 0.9710 - val_accuracy: 0.5942\n",
      "Epoch 159/250\n",
      "9/9 [==============================] - 0s 3ms/step - loss: 0.7393 - accuracy: 0.6525 - val_loss: 0.9673 - val_accuracy: 0.5990\n",
      "Epoch 160/250\n",
      "9/9 [==============================] - 0s 3ms/step - loss: 0.7482 - accuracy: 0.6513 - val_loss: 0.9667 - val_accuracy: 0.6039\n",
      "Epoch 161/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7402 - accuracy: 0.6489 - val_loss: 0.9673 - val_accuracy: 0.5990\n",
      "Epoch 162/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7529 - accuracy: 0.6283 - val_loss: 0.9646 - val_accuracy: 0.6039\n",
      "Epoch 163/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7587 - accuracy: 0.6477 - val_loss: 0.9618 - val_accuracy: 0.6039\n",
      "Epoch 164/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7295 - accuracy: 0.6429 - val_loss: 0.9751 - val_accuracy: 0.5942\n",
      "Epoch 165/250\n",
      "9/9 [==============================] - 0s 3ms/step - loss: 0.7486 - accuracy: 0.6416 - val_loss: 0.9701 - val_accuracy: 0.5894\n",
      "Epoch 166/250\n",
      "9/9 [==============================] - 0s 3ms/step - loss: 0.7460 - accuracy: 0.6586 - val_loss: 0.9654 - val_accuracy: 0.5990\n",
      "Epoch 167/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7383 - accuracy: 0.6416 - val_loss: 0.9737 - val_accuracy: 0.6087\n",
      "Epoch 168/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7500 - accuracy: 0.6429 - val_loss: 0.9655 - val_accuracy: 0.6087\n",
      "Epoch 169/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7258 - accuracy: 0.6489 - val_loss: 0.9670 - val_accuracy: 0.5942\n",
      "Epoch 170/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7409 - accuracy: 0.6610 - val_loss: 0.9757 - val_accuracy: 0.6039\n",
      "Epoch 171/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7413 - accuracy: 0.6320 - val_loss: 0.9735 - val_accuracy: 0.6039\n",
      "Epoch 172/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7484 - accuracy: 0.6453 - val_loss: 0.9787 - val_accuracy: 0.6087\n",
      "Epoch 173/250\n",
      "9/9 [==============================] - 0s 5ms/step - loss: 0.7368 - accuracy: 0.6538 - val_loss: 0.9711 - val_accuracy: 0.5990\n",
      "Epoch 174/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7350 - accuracy: 0.6453 - val_loss: 0.9772 - val_accuracy: 0.5990\n",
      "Epoch 175/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7212 - accuracy: 0.6634 - val_loss: 0.9784 - val_accuracy: 0.6039\n",
      "Epoch 176/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7436 - accuracy: 0.6416 - val_loss: 0.9812 - val_accuracy: 0.6039\n",
      "Epoch 177/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7366 - accuracy: 0.6368 - val_loss: 0.9820 - val_accuracy: 0.6087\n",
      "Epoch 178/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7271 - accuracy: 0.6501 - val_loss: 0.9804 - val_accuracy: 0.6087\n",
      "Epoch 179/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7355 - accuracy: 0.6501 - val_loss: 0.9791 - val_accuracy: 0.5990\n",
      "Epoch 180/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7393 - accuracy: 0.6501 - val_loss: 0.9832 - val_accuracy: 0.6039\n",
      "Epoch 181/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7513 - accuracy: 0.6453 - val_loss: 0.9789 - val_accuracy: 0.5990\n",
      "Epoch 182/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7401 - accuracy: 0.6344 - val_loss: 0.9913 - val_accuracy: 0.5845\n",
      "Epoch 183/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7252 - accuracy: 0.6538 - val_loss: 0.9773 - val_accuracy: 0.6039\n",
      "Epoch 184/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7297 - accuracy: 0.6574 - val_loss: 0.9751 - val_accuracy: 0.5942\n",
      "Epoch 185/250\n",
      "9/9 [==============================] - 0s 5ms/step - loss: 0.7250 - accuracy: 0.6646 - val_loss: 0.9943 - val_accuracy: 0.5942\n",
      "Epoch 186/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7335 - accuracy: 0.6634 - val_loss: 0.9859 - val_accuracy: 0.5942\n",
      "Epoch 187/250\n",
      "9/9 [==============================] - 0s 3ms/step - loss: 0.7279 - accuracy: 0.6513 - val_loss: 0.9856 - val_accuracy: 0.6039\n",
      "Epoch 188/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7214 - accuracy: 0.6562 - val_loss: 0.9891 - val_accuracy: 0.5990\n",
      "Epoch 189/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7329 - accuracy: 0.6550 - val_loss: 0.9864 - val_accuracy: 0.6087\n",
      "Epoch 190/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7291 - accuracy: 0.6562 - val_loss: 0.9875 - val_accuracy: 0.5942\n",
      "Epoch 191/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7310 - accuracy: 0.6525 - val_loss: 0.9885 - val_accuracy: 0.5942\n",
      "Epoch 192/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7286 - accuracy: 0.6465 - val_loss: 0.9875 - val_accuracy: 0.6087\n",
      "Epoch 193/250\n",
      "9/9 [==============================] - 0s 5ms/step - loss: 0.7255 - accuracy: 0.6441 - val_loss: 0.9907 - val_accuracy: 0.5990\n",
      "Epoch 194/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7308 - accuracy: 0.6513 - val_loss: 0.9929 - val_accuracy: 0.5990\n",
      "Epoch 195/250\n",
      "9/9 [==============================] - 0s 6ms/step - loss: 0.7270 - accuracy: 0.6380 - val_loss: 0.9999 - val_accuracy: 0.5942\n",
      "Epoch 196/250\n",
      "9/9 [==============================] - 0s 6ms/step - loss: 0.7292 - accuracy: 0.6501 - val_loss: 1.0003 - val_accuracy: 0.5894\n",
      "Epoch 197/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7171 - accuracy: 0.6441 - val_loss: 1.0005 - val_accuracy: 0.5894\n",
      "Epoch 198/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7239 - accuracy: 0.6574 - val_loss: 1.0035 - val_accuracy: 0.5845\n",
      "Epoch 199/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7081 - accuracy: 0.6598 - val_loss: 0.9983 - val_accuracy: 0.6184\n",
      "Epoch 200/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7285 - accuracy: 0.6707 - val_loss: 1.0065 - val_accuracy: 0.5942\n",
      "Epoch 201/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7297 - accuracy: 0.6501 - val_loss: 1.0053 - val_accuracy: 0.5942\n",
      "Epoch 202/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7167 - accuracy: 0.6671 - val_loss: 1.0042 - val_accuracy: 0.5942\n",
      "Epoch 203/250\n",
      "9/9 [==============================] - 0s 5ms/step - loss: 0.7446 - accuracy: 0.6441 - val_loss: 0.9995 - val_accuracy: 0.6087\n",
      "Epoch 204/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7139 - accuracy: 0.6562 - val_loss: 1.0077 - val_accuracy: 0.5894\n",
      "Epoch 205/250\n",
      "9/9 [==============================] - 0s 3ms/step - loss: 0.7231 - accuracy: 0.6501 - val_loss: 1.0010 - val_accuracy: 0.6039\n",
      "Epoch 206/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7325 - accuracy: 0.6477 - val_loss: 1.0028 - val_accuracy: 0.6184\n",
      "Epoch 207/250\n",
      "9/9 [==============================] - 0s 3ms/step - loss: 0.7304 - accuracy: 0.6562 - val_loss: 1.0076 - val_accuracy: 0.5942\n",
      "Epoch 208/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7148 - accuracy: 0.6453 - val_loss: 1.0025 - val_accuracy: 0.6039\n",
      "Epoch 209/250\n",
      "9/9 [==============================] - 0s 3ms/step - loss: 0.7204 - accuracy: 0.6562 - val_loss: 1.0057 - val_accuracy: 0.6184\n",
      "Epoch 210/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7092 - accuracy: 0.6646 - val_loss: 1.0090 - val_accuracy: 0.6184\n",
      "Epoch 211/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7282 - accuracy: 0.6525 - val_loss: 1.0103 - val_accuracy: 0.5990\n",
      "Epoch 212/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7173 - accuracy: 0.6550 - val_loss: 1.0090 - val_accuracy: 0.6135\n",
      "Epoch 213/250\n",
      "9/9 [==============================] - 0s 7ms/step - loss: 0.7281 - accuracy: 0.6659 - val_loss: 1.0125 - val_accuracy: 0.6039\n",
      "Epoch 214/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7173 - accuracy: 0.6671 - val_loss: 1.0115 - val_accuracy: 0.5942\n",
      "Epoch 215/250\n",
      "9/9 [==============================] - 0s 7ms/step - loss: 0.7240 - accuracy: 0.6453 - val_loss: 1.0077 - val_accuracy: 0.6135\n",
      "Epoch 216/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7166 - accuracy: 0.6719 - val_loss: 1.0154 - val_accuracy: 0.5845\n",
      "Epoch 217/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7202 - accuracy: 0.6646 - val_loss: 1.0101 - val_accuracy: 0.6087\n",
      "Epoch 218/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7103 - accuracy: 0.6634 - val_loss: 1.0103 - val_accuracy: 0.6184\n",
      "Epoch 219/250\n",
      "9/9 [==============================] - 0s 3ms/step - loss: 0.7251 - accuracy: 0.6550 - val_loss: 1.0227 - val_accuracy: 0.5797\n",
      "Epoch 220/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7238 - accuracy: 0.6610 - val_loss: 1.0214 - val_accuracy: 0.5990\n",
      "Epoch 221/250\n",
      "9/9 [==============================] - 0s 5ms/step - loss: 0.7180 - accuracy: 0.6659 - val_loss: 1.0190 - val_accuracy: 0.6135\n",
      "Epoch 222/250\n",
      "9/9 [==============================] - 0s 15ms/step - loss: 0.7241 - accuracy: 0.6525 - val_loss: 1.0288 - val_accuracy: 0.5652\n",
      "Epoch 223/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7085 - accuracy: 0.6719 - val_loss: 1.0161 - val_accuracy: 0.6184\n",
      "Epoch 224/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7241 - accuracy: 0.6610 - val_loss: 1.0310 - val_accuracy: 0.5604\n",
      "Epoch 225/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7192 - accuracy: 0.6610 - val_loss: 1.0233 - val_accuracy: 0.5990\n",
      "Epoch 226/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7202 - accuracy: 0.6513 - val_loss: 1.0269 - val_accuracy: 0.5990\n",
      "Epoch 227/250\n",
      "9/9 [==============================] - 0s 5ms/step - loss: 0.7210 - accuracy: 0.6586 - val_loss: 1.0334 - val_accuracy: 0.5845\n",
      "Epoch 228/250\n",
      "9/9 [==============================] - 0s 6ms/step - loss: 0.7162 - accuracy: 0.6562 - val_loss: 1.0256 - val_accuracy: 0.5942\n",
      "Epoch 229/250\n",
      "9/9 [==============================] - 0s 7ms/step - loss: 0.7125 - accuracy: 0.6671 - val_loss: 1.0269 - val_accuracy: 0.6039\n",
      "Epoch 230/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7200 - accuracy: 0.6574 - val_loss: 1.0309 - val_accuracy: 0.5894\n",
      "Epoch 231/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7041 - accuracy: 0.6719 - val_loss: 1.0294 - val_accuracy: 0.5942\n",
      "Epoch 232/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.6975 - accuracy: 0.6525 - val_loss: 1.0342 - val_accuracy: 0.6039\n",
      "Epoch 233/250\n",
      "9/9 [==============================] - 0s 5ms/step - loss: 0.6972 - accuracy: 0.6768 - val_loss: 1.0404 - val_accuracy: 0.5942\n",
      "Epoch 234/250\n",
      "9/9 [==============================] - 0s 6ms/step - loss: 0.7149 - accuracy: 0.6743 - val_loss: 1.0368 - val_accuracy: 0.5845\n",
      "Epoch 235/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7081 - accuracy: 0.6538 - val_loss: 1.0286 - val_accuracy: 0.5990\n",
      "Epoch 236/250\n",
      "9/9 [==============================] - 0s 7ms/step - loss: 0.7151 - accuracy: 0.6550 - val_loss: 1.0422 - val_accuracy: 0.5845\n",
      "Epoch 237/250\n",
      "9/9 [==============================] - 0s 3ms/step - loss: 0.7180 - accuracy: 0.6538 - val_loss: 1.0368 - val_accuracy: 0.5990\n",
      "Epoch 238/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7082 - accuracy: 0.6646 - val_loss: 1.0363 - val_accuracy: 0.6039\n",
      "Epoch 239/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7129 - accuracy: 0.6719 - val_loss: 1.0340 - val_accuracy: 0.6039\n",
      "Epoch 240/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7101 - accuracy: 0.6562 - val_loss: 1.0404 - val_accuracy: 0.6039\n",
      "Epoch 241/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7015 - accuracy: 0.6731 - val_loss: 1.0391 - val_accuracy: 0.6087\n",
      "Epoch 242/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7111 - accuracy: 0.6513 - val_loss: 1.0519 - val_accuracy: 0.5700\n",
      "Epoch 243/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7097 - accuracy: 0.6538 - val_loss: 1.0313 - val_accuracy: 0.6039\n",
      "Epoch 244/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7084 - accuracy: 0.6634 - val_loss: 1.0392 - val_accuracy: 0.5845\n",
      "Epoch 245/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7034 - accuracy: 0.6792 - val_loss: 1.0492 - val_accuracy: 0.5990\n",
      "Epoch 246/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7147 - accuracy: 0.6598 - val_loss: 1.0473 - val_accuracy: 0.6039\n",
      "Epoch 247/250\n",
      "9/9 [==============================] - 0s 7ms/step - loss: 0.7191 - accuracy: 0.6538 - val_loss: 1.0532 - val_accuracy: 0.5845\n",
      "Epoch 248/250\n",
      "9/9 [==============================] - 0s 6ms/step - loss: 0.6969 - accuracy: 0.6586 - val_loss: 1.0488 - val_accuracy: 0.5990\n",
      "Epoch 249/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7215 - accuracy: 0.6538 - val_loss: 1.0554 - val_accuracy: 0.5845\n",
      "Epoch 250/250\n",
      "9/9 [==============================] - 0s 4ms/step - loss: 0.7238 - accuracy: 0.6586 - val_loss: 1.0513 - val_accuracy: 0.5894\n",
      "0.6183574795722961\n",
      "0.9197027087211609\n",
      "49\n"
     ]
    }
   ],
   "source": [
    "epochs=250\n",
    "history = model.fit(\n",
    "  X_train,\n",
    "  y_train,\n",
    "  validation_data = (X_test,y_test),\n",
    "  batch_size = 100,\n",
    "  epochs=epochs\n",
    ")\n",
    "\n",
    "\n",
    "acc = history.history['accuracy']\n",
    "val_acc = history.history['val_accuracy']\n",
    "\n",
    "loss = history.history['loss']\n",
    "val_loss = history.history['val_loss']\n",
    "\n",
    "epochs_range = range(epochs)\n",
    "\n",
    "print(np.max(val_acc))\n",
    "print(val_loss[np.argmax(val_acc)])\n",
    "print(np.argmax(val_acc))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 576x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8, 8))\n",
    "plt.subplot(1, 2, 1)\n",
    "plt.plot(epochs_range, acc, label='Training Accuracy')\n",
    "plt.plot(epochs_range, val_acc, label='Validation Accuracy')\n",
    "plt.legend(loc='lower right')\n",
    "plt.title('Training and Validation Accuracy')\n",
    "\n",
    "plt.subplot(1, 2, 2)\n",
    "plt.plot(epochs_range, loss, label='Training Loss')\n",
    "plt.plot(epochs_range, val_loss, label='Validation Loss')\n",
    "plt.legend(loc='upper right')\n",
    "plt.title('Training and Validation Loss')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy :  0.5893719806763285\n",
      "Mean Square Error :  1.0483091787439613\n",
      "[0 0 0 0 0]\n",
      "Confusion Matrix for each label : \n",
      "[[[ 46  56]\n",
      "  [ 20  85]]\n",
      "\n",
      " [[148  20]\n",
      "  [ 21  18]]\n",
      "\n",
      " [[135   9]\n",
      "  [ 44  19]]]\n",
      "Classification Report : \n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.60      0.81      0.69       105\n",
      "           1       0.47      0.46      0.47        39\n",
      "           2       0.68      0.30      0.42        63\n",
      "\n",
      "    accuracy                           0.59       207\n",
      "   macro avg       0.59      0.52      0.53       207\n",
      "weighted avg       0.60      0.59      0.57       207\n",
      "\n"
     ]
    }
   ],
   "source": [
    "from sklearn.metrics import classification_report, confusion_matrix, multilabel_confusion_matrix\n",
    "from sklearn.metrics import mean_squared_error, accuracy_score, precision_score, recall_score\n",
    "prediction = model.predict(X_test)\n",
    "prediction  = np.argmax(prediction, axis=1)\n",
    "y_val=np.argmax(y_test, axis=1)\n",
    "print(\"Accuracy : \", accuracy_score(y_val, prediction))\n",
    "print(\"Mean Square Error : \", mean_squared_error(y_val, prediction))\n",
    "\n",
    "print(prediction[:5])\n",
    "print(\"Confusion Matrix for each label : \")\n",
    "conf_matrix = multilabel_confusion_matrix(y_val, prediction)\n",
    "print(conf_matrix)\n",
    "\n",
    "print(\"Classification Report : \")\n",
    "print(classification_report(y_val, prediction))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 864x504 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "vis_arr = np.array(conf_matrix)\n",
    "labels = [\"\".join(\"c\" + str(i)) for i in range(0, 3)]\n",
    "fig, ax = plt.subplots(1, 3, figsize=(12, 7))\n",
    "for axes, cfs_matrix, label in zip(ax.flatten(), vis_arr, labels):\n",
    "    print_confusion_matrix(cfs_matrix, axes, label, [\"N\", \"Y\"])\n",
    "\n",
    "fig.tight_layout()\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "30/30 [==============================] - 0s 1ms/step - loss: 0.7890 - accuracy: 0.6534\n",
      "Accuracy for batch  1  :  0.6442307692307693\n",
      "Mean Square Error for batch  1  :  1.1057692307692308\n",
      "33/33 [==============================] - 0s 1ms/step - loss: 0.7827 - accuracy: 0.6341\n",
      "Accuracy for batch  2  :  0.6730769230769231\n",
      "Mean Square Error for batch  2  :  0.6730769230769231\n",
      "33/33 [==============================] - 0s 1ms/step - loss: 0.7736 - accuracy: 0.6467\n",
      "Accuracy for batch  3  :  0.6442307692307693\n",
      "Mean Square Error for batch  3  :  0.6153846153846154\n",
      "33/33 [==============================] - 0s 955us/step - loss: 0.7780 - accuracy: 0.6428\n",
      "Accuracy for batch  4  :  0.6116504854368932\n",
      "Mean Square Error for batch  4  :  0.7378640776699029\n",
      "33/33 [==============================] - 0s 981us/step - loss: 0.7793 - accuracy: 0.6283\n",
      "Accuracy for batch  5  :  0.6796116504854369\n",
      "Mean Square Error for batch  5  :  0.8737864077669902\n",
      "33/33 [==============================] - 0s 1ms/step - loss: 0.7721 - accuracy: 0.6399\n",
      "Accuracy for batch  6  :  0.6601941747572816\n",
      "Mean Square Error for batch  6  :  1.0097087378640777\n",
      "33/33 [==============================] - 0s 970us/step - loss: 0.7680 - accuracy: 0.6476\n",
      "Accuracy for batch  7  :  0.6893203883495146\n",
      "Mean Square Error for batch  7  :  0.7766990291262136\n",
      "33/33 [==============================] - 0s 1ms/step - loss: 0.7680 - accuracy: 0.6486\n",
      "Accuracy for batch  8  :  0.6990291262135923\n",
      "Mean Square Error for batch  8  :  0.941747572815534\n",
      "33/33 [==============================] - 0s 1ms/step - loss: 0.7677 - accuracy: 0.6457\n",
      "Accuracy for batch  9  :  0.6796116504854369\n",
      "Mean Square Error for batch  9  :  0.9611650485436893\n",
      "30/30 [==============================] - 0s 1ms/step - loss: 0.7613 - accuracy: 0.6484\n",
      "Accuracy for batch  10  :  0.6893203883495146\n",
      "Mean Square Error for batch  10  :  0.7766990291262136\n",
      "Average Accuracy =  0.6670276325616131\n",
      "Average MSE =  0.8471900672143391\n"
     ]
    }
   ],
   "source": [
    "X = inputs\n",
    "y = target\n",
    "\n",
    "from sklearn.model_selection import KFold\n",
    "\n",
    "n_splits=10\n",
    "kf = KFold(n_splits=n_splits)\n",
    "\n",
    "acc = 0\n",
    "mse = 0\n",
    "\n",
    "i = 0\n",
    "for train_indices, test_indices in kf.split(X):\n",
    "    #print(train_indices)\n",
    "    start_train, stop_train = train_indices[0], train_indices[-1]+1\n",
    "    start_test, stop_test = test_indices[0], test_indices[-1]+1\n",
    "    \n",
    "    mlp = model\n",
    "    mlp.fit(X[start_train:stop_train], y[start_train:stop_train])\n",
    "    pred = mlp.predict(X[start_test:stop_test])\n",
    "    #pred\n",
    "    i+=1\n",
    "\n",
    "    pred  = np.argmax(pred, axis=1)\n",
    "    y=np.argmax(y, axis=1)\n",
    "    acc += accuracy_score(y[start_test:stop_test], pred)\n",
    "    mse += mean_squared_error(y[start_test:stop_test], pred)\n",
    "\n",
    "\n",
    "    \n",
    "    print(\"Accuracy for batch \", i, \" : \", accuracy_score(y[start_test:stop_test], pred))\n",
    "    print(\"Mean Square Error for batch \", i, \" : \", mean_squared_error(y[start_test:stop_test], pred))\n",
    "    X = inputs\n",
    "    y = target\n",
    "\n",
    "print('Average Accuracy = ', acc / n_splits)\n",
    "print('Average MSE = ', mse / n_splits)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [],
   "source": [
    "import keras_tuner as kt\n",
    "from tensorflow import keras\n",
    "def model_builder(hp):\n",
    "    model = keras.Sequential()\n",
    "    model.add(keras.layers.Dense(15, input_dim = 20, activation = 'relu'))\n",
    "\n",
    "\n",
    "    # Tune the number of units in the first Dense layer\n",
    "    # Choose an optimal value between 32-512\n",
    "    hp_units = hp.Int('units', min_value=3, max_value=20, step=20)\n",
    "    model.add(keras.layers.Dense(units=hp_units, activation='relu'))\n",
    "    model.add(keras.layers.Dense(10, activation='relu'))\n",
    "    model.add(keras.layers.Dropout(0.2))\n",
    "    model.add(keras.layers.Dense(3, activation = 'softmax'))\n",
    "\n",
    "    # Tune the learning rate for the optimizer\n",
    "    # Choose an optimal value from 0.01, 0.001, or 0.0001\n",
    "    hp_learning_rate = hp.Choice('learning_rate', values=[1e-2, 1e-3, 1e-4])\n",
    "\n",
    "    model.compile(optimizer=keras.optimizers.Adam(learning_rate=hp_learning_rate),\n",
    "                loss = 'categorical_crossentropy',\n",
    "                metrics=['accuracy'])\n",
    "\n",
    "    return model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Reloading Oracle from existing project nn_kt_dir/intro_to_kt/oracle.json\n",
      "INFO:tensorflow:Reloading Tuner from nn_kt_dir/intro_to_kt/tuner0.json\n",
      "INFO:tensorflow:Oracle triggered exit\n",
      "\n",
      "The hyperparameter search is complete. The optimal number of units in the first densely-connected\n",
      "layer is 3 and the optimal learning rate for the optimizer\n",
      "is 0.001.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "\n",
    "\n",
    "tuner = kt.Hyperband(model_builder,\n",
    "                     objective='val_accuracy',\n",
    "                     max_epochs=10,\n",
    "                     factor=3,\n",
    "                     directory='nn_kt_dir',\n",
    "                     project_name='intro_to_kt')\n",
    "stop_early = tf.keras.callbacks.EarlyStopping(monitor='val_loss', patience=5)\n",
    "tuner.search(X_train, y_train, epochs=200, validation_split=0.2, callbacks=[stop_early])\n",
    "\n",
    "# Get the optimal hyperparameters\n",
    "best_hps=tuner.get_best_hyperparameters(num_trials=1)[0]\n",
    "\n",
    "print(f\"\"\"\n",
    "The hyperparameter search is complete. The optimal number of units in the first densely-connected\n",
    "layer is {best_hps.get('units')} and the optimal learning rate for the optimizer\n",
    "is {best_hps.get('learning_rate')}.\n",
    "\"\"\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      "21/21 [==============================] - 1s 12ms/step - loss: 1.0737 - accuracy: 0.4924 - val_loss: 1.0599 - val_accuracy: 0.5422\n",
      "Epoch 2/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 1.0647 - accuracy: 0.4924 - val_loss: 1.0522 - val_accuracy: 0.5422\n",
      "Epoch 3/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 1.0536 - accuracy: 0.5045 - val_loss: 1.0397 - val_accuracy: 0.5422\n",
      "Epoch 4/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 1.0483 - accuracy: 0.5106 - val_loss: 1.0358 - val_accuracy: 0.5422\n",
      "Epoch 5/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 1.0367 - accuracy: 0.5182 - val_loss: 1.0302 - val_accuracy: 0.5422\n",
      "Epoch 6/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 1.0354 - accuracy: 0.5167 - val_loss: 1.0259 - val_accuracy: 0.5422\n",
      "Epoch 7/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 1.0273 - accuracy: 0.5136 - val_loss: 1.0238 - val_accuracy: 0.5422\n",
      "Epoch 8/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 1.0312 - accuracy: 0.5121 - val_loss: 1.0222 - val_accuracy: 0.5422\n",
      "Epoch 9/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 1.0301 - accuracy: 0.5167 - val_loss: 1.0221 - val_accuracy: 0.5422\n",
      "Epoch 10/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 1.0280 - accuracy: 0.5197 - val_loss: 1.0197 - val_accuracy: 0.5422\n",
      "Epoch 11/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 1.0248 - accuracy: 0.5273 - val_loss: 1.0178 - val_accuracy: 0.5422\n",
      "Epoch 12/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 1.0250 - accuracy: 0.5152 - val_loss: 1.0157 - val_accuracy: 0.5422\n",
      "Epoch 13/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 1.0172 - accuracy: 0.5212 - val_loss: 1.0147 - val_accuracy: 0.5422\n",
      "Epoch 14/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 1.0207 - accuracy: 0.5212 - val_loss: 1.0123 - val_accuracy: 0.5422\n",
      "Epoch 15/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 1.0169 - accuracy: 0.5227 - val_loss: 1.0098 - val_accuracy: 0.5422\n",
      "Epoch 16/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 1.0200 - accuracy: 0.5212 - val_loss: 1.0073 - val_accuracy: 0.5422\n",
      "Epoch 17/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 1.0147 - accuracy: 0.5227 - val_loss: 1.0039 - val_accuracy: 0.5422\n",
      "Epoch 18/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.9988 - accuracy: 0.5242 - val_loss: 1.0013 - val_accuracy: 0.5422\n",
      "Epoch 19/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 1.0097 - accuracy: 0.5242 - val_loss: 0.9990 - val_accuracy: 0.5422\n",
      "Epoch 20/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 1.0065 - accuracy: 0.5212 - val_loss: 0.9948 - val_accuracy: 0.5422\n",
      "Epoch 21/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.9965 - accuracy: 0.5242 - val_loss: 0.9901 - val_accuracy: 0.5422\n",
      "Epoch 22/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.9891 - accuracy: 0.5242 - val_loss: 0.9847 - val_accuracy: 0.5422\n",
      "Epoch 23/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.9787 - accuracy: 0.5242 - val_loss: 0.9786 - val_accuracy: 0.5422\n",
      "Epoch 24/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.9796 - accuracy: 0.5258 - val_loss: 0.9683 - val_accuracy: 0.5422\n",
      "Epoch 25/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.9635 - accuracy: 0.5379 - val_loss: 0.9602 - val_accuracy: 0.5542\n",
      "Epoch 26/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.9500 - accuracy: 0.5591 - val_loss: 0.9507 - val_accuracy: 0.5361\n",
      "Epoch 27/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.9426 - accuracy: 0.5758 - val_loss: 0.9414 - val_accuracy: 0.5422\n",
      "Epoch 28/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.9265 - accuracy: 0.5788 - val_loss: 0.9336 - val_accuracy: 0.5602\n",
      "Epoch 29/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.9155 - accuracy: 0.5742 - val_loss: 0.9244 - val_accuracy: 0.5542\n",
      "Epoch 30/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.9029 - accuracy: 0.5894 - val_loss: 0.9160 - val_accuracy: 0.5602\n",
      "Epoch 31/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8831 - accuracy: 0.5864 - val_loss: 0.9074 - val_accuracy: 0.5663\n",
      "Epoch 32/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8772 - accuracy: 0.5955 - val_loss: 0.9034 - val_accuracy: 0.5542\n",
      "Epoch 33/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8647 - accuracy: 0.5955 - val_loss: 0.8973 - val_accuracy: 0.5482\n",
      "Epoch 34/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8679 - accuracy: 0.5803 - val_loss: 0.8916 - val_accuracy: 0.5542\n",
      "Epoch 35/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8321 - accuracy: 0.6106 - val_loss: 0.8849 - val_accuracy: 0.5482\n",
      "Epoch 36/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8422 - accuracy: 0.5909 - val_loss: 0.8846 - val_accuracy: 0.5602\n",
      "Epoch 37/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8319 - accuracy: 0.5985 - val_loss: 0.8784 - val_accuracy: 0.5542\n",
      "Epoch 38/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8423 - accuracy: 0.6045 - val_loss: 0.8788 - val_accuracy: 0.5723\n",
      "Epoch 39/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8205 - accuracy: 0.6258 - val_loss: 0.8774 - val_accuracy: 0.5723\n",
      "Epoch 40/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8143 - accuracy: 0.6227 - val_loss: 0.8756 - val_accuracy: 0.5602\n",
      "Epoch 41/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8192 - accuracy: 0.5985 - val_loss: 0.8739 - val_accuracy: 0.5843\n",
      "Epoch 42/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8041 - accuracy: 0.6379 - val_loss: 0.8727 - val_accuracy: 0.5723\n",
      "Epoch 43/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8188 - accuracy: 0.6106 - val_loss: 0.8735 - val_accuracy: 0.5783\n",
      "Epoch 44/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7961 - accuracy: 0.6152 - val_loss: 0.8761 - val_accuracy: 0.5843\n",
      "Epoch 45/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7958 - accuracy: 0.6318 - val_loss: 0.8751 - val_accuracy: 0.5843\n",
      "Epoch 46/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8001 - accuracy: 0.6288 - val_loss: 0.8755 - val_accuracy: 0.5783\n",
      "Epoch 47/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7884 - accuracy: 0.6364 - val_loss: 0.8747 - val_accuracy: 0.5904\n",
      "Epoch 48/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7900 - accuracy: 0.6348 - val_loss: 0.8748 - val_accuracy: 0.5843\n",
      "Epoch 49/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7895 - accuracy: 0.6197 - val_loss: 0.8764 - val_accuracy: 0.5904\n",
      "Epoch 50/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7956 - accuracy: 0.6258 - val_loss: 0.8770 - val_accuracy: 0.5843\n",
      "Epoch 51/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7710 - accuracy: 0.6470 - val_loss: 0.8749 - val_accuracy: 0.5663\n",
      "Epoch 52/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7779 - accuracy: 0.6364 - val_loss: 0.8757 - val_accuracy: 0.5964\n",
      "Epoch 53/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7849 - accuracy: 0.6348 - val_loss: 0.8768 - val_accuracy: 0.5663\n",
      "Epoch 54/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7656 - accuracy: 0.6409 - val_loss: 0.8766 - val_accuracy: 0.5843\n",
      "Epoch 55/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7782 - accuracy: 0.6500 - val_loss: 0.8760 - val_accuracy: 0.5783\n",
      "Epoch 56/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7701 - accuracy: 0.6424 - val_loss: 0.8770 - val_accuracy: 0.5904\n",
      "Epoch 57/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7815 - accuracy: 0.6394 - val_loss: 0.8796 - val_accuracy: 0.5783\n",
      "Epoch 58/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7554 - accuracy: 0.6470 - val_loss: 0.8793 - val_accuracy: 0.5723\n",
      "Epoch 59/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7674 - accuracy: 0.6364 - val_loss: 0.8830 - val_accuracy: 0.5904\n",
      "Epoch 60/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7490 - accuracy: 0.6636 - val_loss: 0.8844 - val_accuracy: 0.5723\n",
      "Epoch 61/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7537 - accuracy: 0.6576 - val_loss: 0.8822 - val_accuracy: 0.5904\n",
      "Epoch 62/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7477 - accuracy: 0.6455 - val_loss: 0.8816 - val_accuracy: 0.5964\n",
      "Epoch 63/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7625 - accuracy: 0.6682 - val_loss: 0.8832 - val_accuracy: 0.6024\n",
      "Epoch 64/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7630 - accuracy: 0.6470 - val_loss: 0.8853 - val_accuracy: 0.5723\n",
      "Epoch 65/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7608 - accuracy: 0.6379 - val_loss: 0.8848 - val_accuracy: 0.5843\n",
      "Epoch 66/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.7565 - accuracy: 0.6576 - val_loss: 0.8890 - val_accuracy: 0.5843\n",
      "Epoch 67/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.7530 - accuracy: 0.6682 - val_loss: 0.8873 - val_accuracy: 0.5783\n",
      "Epoch 68/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7552 - accuracy: 0.6409 - val_loss: 0.8906 - val_accuracy: 0.5663\n",
      "Epoch 69/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7509 - accuracy: 0.6515 - val_loss: 0.8907 - val_accuracy: 0.5663\n",
      "Epoch 70/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7406 - accuracy: 0.6470 - val_loss: 0.8889 - val_accuracy: 0.5723\n",
      "Epoch 71/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7354 - accuracy: 0.6667 - val_loss: 0.8903 - val_accuracy: 0.5964\n",
      "Epoch 72/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7417 - accuracy: 0.6576 - val_loss: 0.8927 - val_accuracy: 0.5723\n",
      "Epoch 73/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7473 - accuracy: 0.6591 - val_loss: 0.8910 - val_accuracy: 0.5964\n",
      "Epoch 74/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7321 - accuracy: 0.6697 - val_loss: 0.8946 - val_accuracy: 0.5602\n",
      "Epoch 75/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7297 - accuracy: 0.6636 - val_loss: 0.8960 - val_accuracy: 0.5783\n",
      "Epoch 76/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7320 - accuracy: 0.6939 - val_loss: 0.8962 - val_accuracy: 0.5783\n",
      "Epoch 77/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7524 - accuracy: 0.6667 - val_loss: 0.8976 - val_accuracy: 0.5723\n",
      "Epoch 78/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7404 - accuracy: 0.6591 - val_loss: 0.8960 - val_accuracy: 0.5783\n",
      "Epoch 79/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7347 - accuracy: 0.6621 - val_loss: 0.8987 - val_accuracy: 0.5723\n",
      "Epoch 80/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7223 - accuracy: 0.6621 - val_loss: 0.8980 - val_accuracy: 0.5723\n",
      "Epoch 81/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7285 - accuracy: 0.6773 - val_loss: 0.9018 - val_accuracy: 0.5783\n",
      "Epoch 82/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7415 - accuracy: 0.6636 - val_loss: 0.8981 - val_accuracy: 0.5723\n",
      "Epoch 83/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7268 - accuracy: 0.6591 - val_loss: 0.9038 - val_accuracy: 0.5723\n",
      "Epoch 84/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7257 - accuracy: 0.6864 - val_loss: 0.9014 - val_accuracy: 0.5783\n",
      "Epoch 85/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7143 - accuracy: 0.6591 - val_loss: 0.9026 - val_accuracy: 0.5723\n",
      "Epoch 86/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7300 - accuracy: 0.6652 - val_loss: 0.9055 - val_accuracy: 0.5723\n",
      "Epoch 87/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7048 - accuracy: 0.6682 - val_loss: 0.9053 - val_accuracy: 0.5663\n",
      "Epoch 88/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.7244 - accuracy: 0.6727 - val_loss: 0.9086 - val_accuracy: 0.5663\n",
      "Epoch 89/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.7189 - accuracy: 0.6758 - val_loss: 0.9114 - val_accuracy: 0.5542\n",
      "Epoch 90/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.7071 - accuracy: 0.6818 - val_loss: 0.9115 - val_accuracy: 0.5663\n",
      "Epoch 91/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.7086 - accuracy: 0.6697 - val_loss: 0.9119 - val_accuracy: 0.5663\n",
      "Epoch 92/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.7154 - accuracy: 0.6712 - val_loss: 0.9129 - val_accuracy: 0.5663\n",
      "Epoch 93/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6912 - accuracy: 0.6909 - val_loss: 0.9154 - val_accuracy: 0.5602\n",
      "Epoch 94/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.7147 - accuracy: 0.6712 - val_loss: 0.9199 - val_accuracy: 0.5542\n",
      "Epoch 95/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.7128 - accuracy: 0.6894 - val_loss: 0.9205 - val_accuracy: 0.5542\n",
      "Epoch 96/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.7013 - accuracy: 0.6803 - val_loss: 0.9217 - val_accuracy: 0.5663\n",
      "Epoch 97/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6969 - accuracy: 0.6864 - val_loss: 0.9238 - val_accuracy: 0.5602\n",
      "Epoch 98/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7066 - accuracy: 0.6788 - val_loss: 0.9288 - val_accuracy: 0.5482\n",
      "Epoch 99/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.6867 - accuracy: 0.7000 - val_loss: 0.9238 - val_accuracy: 0.5542\n",
      "Epoch 100/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.7128 - accuracy: 0.6803 - val_loss: 0.9247 - val_accuracy: 0.5602\n",
      "Epoch 101/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.7029 - accuracy: 0.6773 - val_loss: 0.9266 - val_accuracy: 0.5542\n",
      "Epoch 102/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.6993 - accuracy: 0.6682 - val_loss: 0.9316 - val_accuracy: 0.5542\n",
      "Epoch 103/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.7077 - accuracy: 0.6818 - val_loss: 0.9305 - val_accuracy: 0.5542\n",
      "Epoch 104/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.7086 - accuracy: 0.6742 - val_loss: 0.9341 - val_accuracy: 0.5422\n",
      "Epoch 105/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6970 - accuracy: 0.6924 - val_loss: 0.9299 - val_accuracy: 0.5361\n",
      "Epoch 106/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6941 - accuracy: 0.6924 - val_loss: 0.9397 - val_accuracy: 0.5602\n",
      "Epoch 107/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7120 - accuracy: 0.6667 - val_loss: 0.9306 - val_accuracy: 0.5361\n",
      "Epoch 108/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6954 - accuracy: 0.6879 - val_loss: 0.9376 - val_accuracy: 0.5422\n",
      "Epoch 109/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6776 - accuracy: 0.6924 - val_loss: 0.9363 - val_accuracy: 0.5663\n",
      "Epoch 110/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6936 - accuracy: 0.6788 - val_loss: 0.9355 - val_accuracy: 0.5482\n",
      "Epoch 111/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6922 - accuracy: 0.6985 - val_loss: 0.9388 - val_accuracy: 0.5542\n",
      "Epoch 112/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6879 - accuracy: 0.6848 - val_loss: 0.9382 - val_accuracy: 0.5542\n",
      "Epoch 113/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6905 - accuracy: 0.6712 - val_loss: 0.9412 - val_accuracy: 0.5301\n",
      "Epoch 114/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6935 - accuracy: 0.6848 - val_loss: 0.9390 - val_accuracy: 0.5542\n",
      "Epoch 115/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6806 - accuracy: 0.6909 - val_loss: 0.9463 - val_accuracy: 0.5482\n",
      "Epoch 116/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6816 - accuracy: 0.6833 - val_loss: 0.9447 - val_accuracy: 0.5663\n",
      "Epoch 117/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6876 - accuracy: 0.6848 - val_loss: 0.9490 - val_accuracy: 0.5542\n",
      "Epoch 118/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6946 - accuracy: 0.6848 - val_loss: 0.9496 - val_accuracy: 0.5602\n",
      "Epoch 119/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6842 - accuracy: 0.7030 - val_loss: 0.9485 - val_accuracy: 0.5422\n",
      "Epoch 120/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6796 - accuracy: 0.7076 - val_loss: 0.9530 - val_accuracy: 0.5542\n",
      "Epoch 121/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6865 - accuracy: 0.6939 - val_loss: 0.9536 - val_accuracy: 0.5422\n",
      "Epoch 122/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6838 - accuracy: 0.6879 - val_loss: 0.9534 - val_accuracy: 0.5301\n",
      "Epoch 123/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6794 - accuracy: 0.6970 - val_loss: 0.9563 - val_accuracy: 0.5723\n",
      "Epoch 124/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6775 - accuracy: 0.7045 - val_loss: 0.9542 - val_accuracy: 0.5542\n",
      "Epoch 125/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6951 - accuracy: 0.6818 - val_loss: 0.9522 - val_accuracy: 0.5482\n",
      "Epoch 126/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6701 - accuracy: 0.7061 - val_loss: 0.9567 - val_accuracy: 0.5361\n",
      "Epoch 127/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6743 - accuracy: 0.7076 - val_loss: 0.9607 - val_accuracy: 0.5482\n",
      "Epoch 128/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6810 - accuracy: 0.6818 - val_loss: 0.9602 - val_accuracy: 0.5542\n",
      "Epoch 129/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6771 - accuracy: 0.6939 - val_loss: 0.9611 - val_accuracy: 0.5542\n",
      "Epoch 130/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6744 - accuracy: 0.6848 - val_loss: 0.9648 - val_accuracy: 0.5542\n",
      "Epoch 131/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6708 - accuracy: 0.6985 - val_loss: 0.9696 - val_accuracy: 0.5422\n",
      "Epoch 132/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6870 - accuracy: 0.6939 - val_loss: 0.9652 - val_accuracy: 0.5542\n",
      "Epoch 133/200\n",
      "21/21 [==============================] - 0s 4ms/step - loss: 0.6776 - accuracy: 0.6864 - val_loss: 0.9644 - val_accuracy: 0.5241\n",
      "Epoch 134/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.6646 - accuracy: 0.6924 - val_loss: 0.9640 - val_accuracy: 0.5602\n",
      "Epoch 135/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.6638 - accuracy: 0.7000 - val_loss: 0.9737 - val_accuracy: 0.5723\n",
      "Epoch 136/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.6759 - accuracy: 0.6924 - val_loss: 0.9684 - val_accuracy: 0.5301\n",
      "Epoch 137/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6664 - accuracy: 0.6955 - val_loss: 0.9668 - val_accuracy: 0.5663\n",
      "Epoch 138/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6679 - accuracy: 0.7136 - val_loss: 0.9732 - val_accuracy: 0.5542\n",
      "Epoch 139/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.6586 - accuracy: 0.6939 - val_loss: 0.9716 - val_accuracy: 0.5482\n",
      "Epoch 140/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.6562 - accuracy: 0.7091 - val_loss: 0.9764 - val_accuracy: 0.5602\n",
      "Epoch 141/200\n",
      "21/21 [==============================] - 0s 5ms/step - loss: 0.6880 - accuracy: 0.6803 - val_loss: 0.9769 - val_accuracy: 0.5663\n",
      "Epoch 142/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6705 - accuracy: 0.7106 - val_loss: 0.9733 - val_accuracy: 0.5482\n",
      "Epoch 143/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.6508 - accuracy: 0.7121 - val_loss: 0.9857 - val_accuracy: 0.5663\n",
      "Epoch 144/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6655 - accuracy: 0.6955 - val_loss: 0.9789 - val_accuracy: 0.5602\n",
      "Epoch 145/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6586 - accuracy: 0.7227 - val_loss: 0.9834 - val_accuracy: 0.5542\n",
      "Epoch 146/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6487 - accuracy: 0.7212 - val_loss: 0.9814 - val_accuracy: 0.5482\n",
      "Epoch 147/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6643 - accuracy: 0.7167 - val_loss: 0.9871 - val_accuracy: 0.5663\n",
      "Epoch 148/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6626 - accuracy: 0.6924 - val_loss: 0.9903 - val_accuracy: 0.5482\n",
      "Epoch 149/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6510 - accuracy: 0.7242 - val_loss: 0.9932 - val_accuracy: 0.5482\n",
      "Epoch 150/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6675 - accuracy: 0.6985 - val_loss: 0.9899 - val_accuracy: 0.5542\n",
      "Epoch 151/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6598 - accuracy: 0.7182 - val_loss: 0.9874 - val_accuracy: 0.5663\n",
      "Epoch 152/200\n",
      "21/21 [==============================] - 0s 5ms/step - loss: 0.6586 - accuracy: 0.7076 - val_loss: 0.9897 - val_accuracy: 0.5422\n",
      "Epoch 153/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6564 - accuracy: 0.7045 - val_loss: 0.9922 - val_accuracy: 0.5542\n",
      "Epoch 154/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6677 - accuracy: 0.6939 - val_loss: 0.9984 - val_accuracy: 0.5663\n",
      "Epoch 155/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6686 - accuracy: 0.7076 - val_loss: 1.0001 - val_accuracy: 0.5542\n",
      "Epoch 156/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6553 - accuracy: 0.6970 - val_loss: 0.9989 - val_accuracy: 0.5723\n",
      "Epoch 157/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6608 - accuracy: 0.7015 - val_loss: 1.0100 - val_accuracy: 0.5542\n",
      "Epoch 158/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.6623 - accuracy: 0.7167 - val_loss: 0.9945 - val_accuracy: 0.5482\n",
      "Epoch 159/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.6485 - accuracy: 0.7303 - val_loss: 1.0006 - val_accuracy: 0.5542\n",
      "Epoch 160/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.6410 - accuracy: 0.7076 - val_loss: 1.0016 - val_accuracy: 0.5542\n",
      "Epoch 161/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.6577 - accuracy: 0.6985 - val_loss: 0.9980 - val_accuracy: 0.5361\n",
      "Epoch 162/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.6472 - accuracy: 0.7152 - val_loss: 1.0057 - val_accuracy: 0.5542\n",
      "Epoch 163/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.6456 - accuracy: 0.6955 - val_loss: 1.0019 - val_accuracy: 0.5542\n",
      "Epoch 164/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.6432 - accuracy: 0.7197 - val_loss: 1.0018 - val_accuracy: 0.5301\n",
      "Epoch 165/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6526 - accuracy: 0.7197 - val_loss: 1.0044 - val_accuracy: 0.5361\n",
      "Epoch 166/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6449 - accuracy: 0.7227 - val_loss: 1.0082 - val_accuracy: 0.5663\n",
      "Epoch 167/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6437 - accuracy: 0.7182 - val_loss: 1.0150 - val_accuracy: 0.5602\n",
      "Epoch 168/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6300 - accuracy: 0.7318 - val_loss: 1.0191 - val_accuracy: 0.5542\n",
      "Epoch 169/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6425 - accuracy: 0.7182 - val_loss: 1.0155 - val_accuracy: 0.5542\n",
      "Epoch 170/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6328 - accuracy: 0.7197 - val_loss: 1.0275 - val_accuracy: 0.5542\n",
      "Epoch 171/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6566 - accuracy: 0.7015 - val_loss: 1.0278 - val_accuracy: 0.5542\n",
      "Epoch 172/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6413 - accuracy: 0.7152 - val_loss: 1.0305 - val_accuracy: 0.5542\n",
      "Epoch 173/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6267 - accuracy: 0.7273 - val_loss: 1.0168 - val_accuracy: 0.5361\n",
      "Epoch 174/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6481 - accuracy: 0.7061 - val_loss: 1.0188 - val_accuracy: 0.5542\n",
      "Epoch 175/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6313 - accuracy: 0.7076 - val_loss: 1.0358 - val_accuracy: 0.5361\n",
      "Epoch 176/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6501 - accuracy: 0.7182 - val_loss: 1.0193 - val_accuracy: 0.5542\n",
      "Epoch 177/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6267 - accuracy: 0.7455 - val_loss: 1.0259 - val_accuracy: 0.5602\n",
      "Epoch 178/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6453 - accuracy: 0.7091 - val_loss: 1.0243 - val_accuracy: 0.5482\n",
      "Epoch 179/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6322 - accuracy: 0.7364 - val_loss: 1.0265 - val_accuracy: 0.5602\n",
      "Epoch 180/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6304 - accuracy: 0.7379 - val_loss: 1.0302 - val_accuracy: 0.5482\n",
      "Epoch 181/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6244 - accuracy: 0.7227 - val_loss: 1.0373 - val_accuracy: 0.5422\n",
      "Epoch 182/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6369 - accuracy: 0.7000 - val_loss: 1.0428 - val_accuracy: 0.5422\n",
      "Epoch 183/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6197 - accuracy: 0.7167 - val_loss: 1.0357 - val_accuracy: 0.5422\n",
      "Epoch 184/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6436 - accuracy: 0.7167 - val_loss: 1.0327 - val_accuracy: 0.5542\n",
      "Epoch 185/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6440 - accuracy: 0.7152 - val_loss: 1.0477 - val_accuracy: 0.5361\n",
      "Epoch 186/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6391 - accuracy: 0.7212 - val_loss: 1.0323 - val_accuracy: 0.5301\n",
      "Epoch 187/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6451 - accuracy: 0.7121 - val_loss: 1.0419 - val_accuracy: 0.5422\n",
      "Epoch 188/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6384 - accuracy: 0.7212 - val_loss: 1.0397 - val_accuracy: 0.5482\n",
      "Epoch 189/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6274 - accuracy: 0.7197 - val_loss: 1.0427 - val_accuracy: 0.5482\n",
      "Epoch 190/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6271 - accuracy: 0.7273 - val_loss: 1.0497 - val_accuracy: 0.5422\n",
      "Epoch 191/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6188 - accuracy: 0.7303 - val_loss: 1.0507 - val_accuracy: 0.5241\n",
      "Epoch 192/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6164 - accuracy: 0.7288 - val_loss: 1.0626 - val_accuracy: 0.5361\n",
      "Epoch 193/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6236 - accuracy: 0.7439 - val_loss: 1.0614 - val_accuracy: 0.5482\n",
      "Epoch 194/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.6356 - accuracy: 0.7333 - val_loss: 1.0590 - val_accuracy: 0.5422\n",
      "Epoch 195/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.6374 - accuracy: 0.7045 - val_loss: 1.0510 - val_accuracy: 0.5241\n",
      "Epoch 196/200\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.6303 - accuracy: 0.7288 - val_loss: 1.0461 - val_accuracy: 0.5482\n",
      "Epoch 197/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6389 - accuracy: 0.7242 - val_loss: 1.0468 - val_accuracy: 0.5301\n",
      "Epoch 198/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6298 - accuracy: 0.7152 - val_loss: 1.0511 - val_accuracy: 0.5301\n",
      "Epoch 199/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6370 - accuracy: 0.7152 - val_loss: 1.0613 - val_accuracy: 0.5422\n",
      "Epoch 200/200\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.6381 - accuracy: 0.7242 - val_loss: 1.0650 - val_accuracy: 0.5422\n",
      "Best epoch: 63\n",
      "Epoch 1/63\n",
      "21/21 [==============================] - 1s 9ms/step - loss: 1.3366 - accuracy: 0.2500 - val_loss: 1.1429 - val_accuracy: 0.2289\n",
      "Epoch 2/63\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 1.1620 - accuracy: 0.2848 - val_loss: 1.0629 - val_accuracy: 0.3133\n",
      "Epoch 3/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 1.0918 - accuracy: 0.3909 - val_loss: 1.0325 - val_accuracy: 0.5060\n",
      "Epoch 4/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 1.0830 - accuracy: 0.4182 - val_loss: 1.0109 - val_accuracy: 0.5602\n",
      "Epoch 5/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 1.0468 - accuracy: 0.4470 - val_loss: 0.9981 - val_accuracy: 0.5482\n",
      "Epoch 6/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 1.0121 - accuracy: 0.4894 - val_loss: 0.9878 - val_accuracy: 0.5482\n",
      "Epoch 7/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 1.0196 - accuracy: 0.4879 - val_loss: 0.9789 - val_accuracy: 0.5482\n",
      "Epoch 8/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 1.0049 - accuracy: 0.4909 - val_loss: 0.9735 - val_accuracy: 0.5422\n",
      "Epoch 9/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 1.0020 - accuracy: 0.4955 - val_loss: 0.9660 - val_accuracy: 0.5422\n",
      "Epoch 10/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.9758 - accuracy: 0.5227 - val_loss: 0.9567 - val_accuracy: 0.5422\n",
      "Epoch 11/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.9616 - accuracy: 0.5561 - val_loss: 0.9461 - val_accuracy: 0.5482\n",
      "Epoch 12/63\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.9576 - accuracy: 0.5424 - val_loss: 0.9363 - val_accuracy: 0.5482\n",
      "Epoch 13/63\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.9478 - accuracy: 0.5576 - val_loss: 0.9280 - val_accuracy: 0.5482\n",
      "Epoch 14/63\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.9508 - accuracy: 0.5470 - val_loss: 0.9203 - val_accuracy: 0.5422\n",
      "Epoch 15/63\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.9294 - accuracy: 0.5561 - val_loss: 0.9117 - val_accuracy: 0.5422\n",
      "Epoch 16/63\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.9315 - accuracy: 0.5485 - val_loss: 0.9043 - val_accuracy: 0.5663\n",
      "Epoch 17/63\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.9109 - accuracy: 0.5818 - val_loss: 0.8975 - val_accuracy: 0.5783\n",
      "Epoch 18/63\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.9111 - accuracy: 0.5515 - val_loss: 0.8905 - val_accuracy: 0.5663\n",
      "Epoch 19/63\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.9019 - accuracy: 0.5652 - val_loss: 0.8823 - val_accuracy: 0.5723\n",
      "Epoch 20/63\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.8878 - accuracy: 0.5924 - val_loss: 0.8758 - val_accuracy: 0.5602\n",
      "Epoch 21/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8812 - accuracy: 0.5682 - val_loss: 0.8712 - val_accuracy: 0.5602\n",
      "Epoch 22/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8710 - accuracy: 0.5682 - val_loss: 0.8653 - val_accuracy: 0.5602\n",
      "Epoch 23/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8600 - accuracy: 0.5833 - val_loss: 0.8616 - val_accuracy: 0.5602\n",
      "Epoch 24/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8624 - accuracy: 0.5818 - val_loss: 0.8593 - val_accuracy: 0.5663\n",
      "Epoch 25/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8476 - accuracy: 0.5788 - val_loss: 0.8580 - val_accuracy: 0.5542\n",
      "Epoch 26/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8410 - accuracy: 0.5894 - val_loss: 0.8535 - val_accuracy: 0.5723\n",
      "Epoch 27/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8527 - accuracy: 0.5939 - val_loss: 0.8515 - val_accuracy: 0.5663\n",
      "Epoch 28/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8520 - accuracy: 0.5758 - val_loss: 0.8510 - val_accuracy: 0.5663\n",
      "Epoch 29/63\n",
      "21/21 [==============================] - 0s 5ms/step - loss: 0.8365 - accuracy: 0.6167 - val_loss: 0.8518 - val_accuracy: 0.5723\n",
      "Epoch 30/63\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.8267 - accuracy: 0.5939 - val_loss: 0.8506 - val_accuracy: 0.5602\n",
      "Epoch 31/63\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.8237 - accuracy: 0.5864 - val_loss: 0.8528 - val_accuracy: 0.5783\n",
      "Epoch 32/63\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.8254 - accuracy: 0.5894 - val_loss: 0.8467 - val_accuracy: 0.5904\n",
      "Epoch 33/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8074 - accuracy: 0.6273 - val_loss: 0.8499 - val_accuracy: 0.5783\n",
      "Epoch 34/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8265 - accuracy: 0.6121 - val_loss: 0.8479 - val_accuracy: 0.6084\n",
      "Epoch 35/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8413 - accuracy: 0.6030 - val_loss: 0.8476 - val_accuracy: 0.5904\n",
      "Epoch 36/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8146 - accuracy: 0.6152 - val_loss: 0.8473 - val_accuracy: 0.6024\n",
      "Epoch 37/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8255 - accuracy: 0.5985 - val_loss: 0.8470 - val_accuracy: 0.6084\n",
      "Epoch 38/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8260 - accuracy: 0.6136 - val_loss: 0.8489 - val_accuracy: 0.6205\n",
      "Epoch 39/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8006 - accuracy: 0.6136 - val_loss: 0.8480 - val_accuracy: 0.6145\n",
      "Epoch 40/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8319 - accuracy: 0.5879 - val_loss: 0.8480 - val_accuracy: 0.6024\n",
      "Epoch 41/63\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.8097 - accuracy: 0.6212 - val_loss: 0.8497 - val_accuracy: 0.6145\n",
      "Epoch 42/63\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.8061 - accuracy: 0.6152 - val_loss: 0.8474 - val_accuracy: 0.5964\n",
      "Epoch 43/63\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.8010 - accuracy: 0.6424 - val_loss: 0.8464 - val_accuracy: 0.6024\n",
      "Epoch 44/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8089 - accuracy: 0.6212 - val_loss: 0.8534 - val_accuracy: 0.6205\n",
      "Epoch 45/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8046 - accuracy: 0.6030 - val_loss: 0.8470 - val_accuracy: 0.6145\n",
      "Epoch 46/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8208 - accuracy: 0.6106 - val_loss: 0.8471 - val_accuracy: 0.6145\n",
      "Epoch 47/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7934 - accuracy: 0.6167 - val_loss: 0.8453 - val_accuracy: 0.6024\n",
      "Epoch 48/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7962 - accuracy: 0.5970 - val_loss: 0.8479 - val_accuracy: 0.6205\n",
      "Epoch 49/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7767 - accuracy: 0.6364 - val_loss: 0.8468 - val_accuracy: 0.6145\n",
      "Epoch 50/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8153 - accuracy: 0.6015 - val_loss: 0.8477 - val_accuracy: 0.6205\n",
      "Epoch 51/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7892 - accuracy: 0.6227 - val_loss: 0.8455 - val_accuracy: 0.6024\n",
      "Epoch 52/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8064 - accuracy: 0.6242 - val_loss: 0.8460 - val_accuracy: 0.6024\n",
      "Epoch 53/63\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.7996 - accuracy: 0.6318 - val_loss: 0.8471 - val_accuracy: 0.5964\n",
      "Epoch 54/63\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.7998 - accuracy: 0.6242 - val_loss: 0.8480 - val_accuracy: 0.6265\n",
      "Epoch 55/63\n",
      "21/21 [==============================] - 0s 6ms/step - loss: 0.7868 - accuracy: 0.6091 - val_loss: 0.8482 - val_accuracy: 0.5904\n",
      "Epoch 56/63\n",
      "21/21 [==============================] - 0s 3ms/step - loss: 0.7976 - accuracy: 0.6318 - val_loss: 0.8506 - val_accuracy: 0.6265\n",
      "Epoch 57/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7959 - accuracy: 0.6182 - val_loss: 0.8493 - val_accuracy: 0.5843\n",
      "Epoch 58/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7979 - accuracy: 0.6061 - val_loss: 0.8502 - val_accuracy: 0.6084\n",
      "Epoch 59/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7981 - accuracy: 0.6227 - val_loss: 0.8495 - val_accuracy: 0.6024\n",
      "Epoch 60/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7889 - accuracy: 0.6485 - val_loss: 0.8494 - val_accuracy: 0.5843\n",
      "Epoch 61/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.8081 - accuracy: 0.6333 - val_loss: 0.8542 - val_accuracy: 0.6145\n",
      "Epoch 62/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7888 - accuracy: 0.6333 - val_loss: 0.8491 - val_accuracy: 0.5904\n",
      "Epoch 63/63\n",
      "21/21 [==============================] - 0s 2ms/step - loss: 0.7831 - accuracy: 0.6273 - val_loss: 0.8491 - val_accuracy: 0.6024\n"
     ]
    }
   ],
   "source": [
    "model = tuner.hypermodel.build(best_hps)\n",
    "history = model.fit(X_train, y_train, epochs=200, validation_split=0.2)\n",
    "\n",
    "val_acc_per_epoch = history.history['val_accuracy']\n",
    "best_epoch = val_acc_per_epoch.index(max(val_acc_per_epoch)) + 1\n",
    "print('Best epoch: %d' % (best_epoch,))\n",
    "\n",
    "hypermodel = tuner.hypermodel.build(best_hps)\n",
    "\n",
    "# Retrain the model\n",
    "history = hypermodel.fit(X_train, y_train, epochs=best_epoch, validation_split=0.2)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "7/7 [==============================] - 0s 1ms/step - loss: 0.9652 - accuracy: 0.5845\n",
      "[test loss, test accuracy]: [0.9652019143104553, 0.5845410823822021]\n"
     ]
    }
   ],
   "source": [
    "eval_result = hypermodel.evaluate(X_test, y_test)\n",
    "print(\"[test loss, test accuracy]:\", eval_result)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeMAAAHiCAYAAADbK6SdAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAAsTAAALEwEAmpwYAACMFUlEQVR4nOzdd3yb1bnA8d/R8t4jw3ZiO3s6w0lIAhlAKauEWUhpIaWFQgeFTmgpcFu4pb3cDm6hLbRAaSmjFGjYbYAMSAhkk73sJI4T7z1kSzr3j1eSZVuekS3Jfr6fTz6x3/fVqyMn1qPnjOcorTVCCCGECB5TsBsghBBCDHcSjIUQQoggk2AshBBCBJkEYyGEECLIJBgLIYQQQSbBWAghhAiyIReMlVJvKaVuDPS1waSUKlRKnT8A912rlPqq++vrlVL/7s21/XieMUqpeqWUub9tFaIv5H2gT/eV94EQEBLB2P0P5PnjUko1+Xx/fV/upbW+SGv9l0BfG4qUUncppdb7OZ6qlGpRSk3v7b201s9qrS8IULvavWlorY9rrWO11s5A3N/P8yml1FGl1N6BuL8YHPI+0D/yPgBKKa2UGh/o+w6mkAjG7n+gWK11LHAc+JzPsWc91ymlLMFrZUj6G7BIKZXT4fh1wKda691BaFMwLAHSgVyl1LzBfGL5Pxk48j7Qb/I+MASERDDuilJqmVKqSCn1Q6XUaeAppVSSUup1pVSZUqrK/XWmz2N8u1xWKaU+UEo97L62QCl1UT+vzVFKrVdK1Sml1iilHlVK/a2LdvemjT9TSn3ovt+/lVKpPue/pJQ6ppSqUEr9uKufj9a6CHgP+FKHUzcAz/TUjg5tXqWU+sDn+88opfYrpWqUUr8DlM+5cUqp99ztK1dKPauUSnSf+yswBnjNndH8QCmV7f7kanFfM1optVopVamUOqyUutnn3vcrpV5USj3j/tnsUUrld/UzcLsR+Bfwpvtr39c1TSn1H/dzlSilfuQ+blZK/UgpdcT9PFuVUlkd2+q+tuP/kw+VUr9WSlUA93f383A/Jksp9bL736FCKfU7pZTN3aYZPtelK6UalVJpPbzeYUXeB+R9oJfvA/5eT4L7HmXun+U9SimT+9x4pdQ692srV0q94D6u3L/fpUqpWqXUp6oPvQv9FdLB2G0kkAyMBW7BaPNT7u/HAE3A77p5/ALgAJAK/BL4s1JK9ePavwMfAynA/XT+j++rN238AvBljIzOBnwPQCk1Ffi9+/6j3c/n9xfH7S++bVFKTQJmudvb15+V5x6pwMvAPRg/iyPAYt9LgJ+72zcFyML4maC1/hLts5pf+nmK54Ei9+OvBv5bKXWuz/nL3NckAqu7a7NSKtp9j2fdf65TStnc5+KANcDb7ucaD7zrfuh3gJXAxUA8cBPQ2N3PxccC4CgwAniQbn4eyhgfex04BmQDGcDzWusW92v8os99VwLvaq3LetmO4UTeB+R9oMc2+/F/QAKQCyzF+IDyZfe5nwH/BpIwfrb/5z5+AUZv20T3Yz8PVPTjuftGax1Sf4BC4Hz318uAFiCym+tnAVU+368Fvur+ehVw2OdcNKCBkX25FuM/sAOI9jn/N+BvvXxN/tp4j8/3Xwfedn99L8abtedcjPtncH4X944GaoFF7u8fBP7Vz5/VB+6vbwA+8rlOYfzSfLWL+14ObPf3b+j+Ptv9s7Rg/MI6gTif8z8HnnZ/fT+wxufcVKCpm5/tF4Ey970jgRrgCve5lb7t6vC4A8AKP8e9be3m53S8h39v788DWOhpn5/rFmC8YSn391uAzw/071g4/EHeB+R9oG/vAxoY3+GY2f0zm+pz7GvAWvfXzwCPA5kdHncucBA4CzAN1v/5cMiMy7TWzZ5vlFLRSqk/urscaoH1QKLqeobeac8XWmtP5hPbx2tHA5U+xwBOdNXgXrbxtM/XjT5tGu17b611A918KnO36R/ADe5P79dj/Cfrz8/Ko2MbtO/3SqkRSqnnlVIn3ff9G8Yn597w/CzrfI4dw8gYPTr+bCJV1+OENwIvaq0d7v8n/6StqzoL49O8P92d60m7f/sefh5ZwDGttaPjTbTWmzFe3zKl1GSMzH11P9s01Mn7gLwPdPc+4E8qYHXf199z/ADjA8bH7m7wmwC01u9hZOGPAqVKqceVUvF9eN5+CYdg3HFbqe8Ck4AFWut4jO4E8BnLGACngGR3l6hHVjfXn0kbT/ne2/2cKT085i8YXSmfAeKA186wHR3boGj/ev8b499lhvu+X+xwz+62AivG+FnG+RwbA5zsoU2dKGPc61zgi0qp08oYT7wauNjdxXYCo3vKnxPAOD/HG9x/+/5bj+xwTcfX193P4wQwpps3kb+4r/8S8JJvwBHtyPuAvA/0VTnQitE93+k5tNantdY3a61HY2TMjyn3jGyt9SNa67kYGflE4PsBbJdf4RCMO4rDGPOoVkolA/cN9BNqrY9hdCHer4yJNwuBzw1QG18CLlVKne0e+/wpPf87bQCqMbpcPOORZ9KON4BpSqkr3UHkdtoHpDigHqhRSmXQ+T9qCV0EQa31CWAj8HOlVKRSaibwFYxP1X31JYzuJM/42CyMX5wijC7q14FRSqk7lFIRSqk4pdQC92P/BPxMKTXBPWFjplIqRRvjtScxArzZ/WnZX9D21d3P42OMN7WHlFIx7tfsO+72N+AKjDeyZ/rxMxiu5H2gs+H6PuBhc98rUikV6T72IvCg+3d/LMZckb8BKKWuUW0T2aowPjy4lFLzlFILlFJWjA/nzYDrDNrVK+EYjH8DRGF86vkIY3LOYLgeY/yvAngAeAGwd3Htb+hnG7XWe4BvYEy8OIXxn6Soh8dojDfysbR/Q+9XO7TW5cA1wEMYr3cC8KHPJf8FzMEYn30DY5KHr58D9yilqpVS3/PzFCsxxo+KgVeA+7TWa3rTtg5uBB5zf8L1/gH+ANzo7gL7DMYb5mngELDc/dhfYfyi/htjrO3PGD8rgJsx3lgqgGkYbxrd6fLnoY01lZ/D6II+jvFvea3P+RPANow3gg19/xEMW79B3gc6Pma4vg947MH40OH582XgWxgB9SjwAcbP80n39fOAzUqpeozhoW9rrY9iTOh8AuNnfgzjtf/PGbSrVzwTR0QfKWMa/H6t9YB/IhdDm1LqSaBYa31PsNsi+kbeB0SghGNmHBTurotxSimTUupCYAXwapCbJcKcUiobuBIjMxchTt4HxECRSja9NxKjGyYFo7voNq319uA2SYQzpdTPgDuBn2utC4LdHtEr8j4gBoR0UwshhBBBJt3UQgghRJBJMBZCCCGCLGhjxqmpqTo7OztYTy9E2Ni6dWu51jqkN4+Q32chetbd73LQgnF2djZbtmwJ1tMLETaUUsd6viq45PdZiJ5197ss3dRCCCFEkEkwFkIIIYJMgrEQQggRZFL0QwghQlhraytFRUU0N8uGXuEiMjKSzMxMrFZrrx8jwVgIIUJYUVERcXFxZGdnY+xiKEKZ1pqKigqKiorIycnp9eOkm1oIIUJYc3MzKSkpEojDhFKKlJSUPvdkSDAWQogQJ4E4vPTn30uCsRBCiC5VVFQwa9YsZs2axciRI8nIyPB+39LS0u1jt2zZwu23397jcyxatCggbV27di2XXnppQO412GTMWAghRJdSUlLYsWMHAPfffz+xsbF873vf8553OBxYLP5DSX5+Pvn5+T0+x8aNGwPS1nAmmbEQQog+WbVqFbfeeisLFizgBz/4AR9//DELFy5k9uzZLFq0iAMHDgDtM9X777+fm266iWXLlpGbm8sjjzzivV9sbKz3+mXLlnH11VczefJkrr/+ejw7C7755ptMnjyZuXPncvvtt/cpA37uueeYMWMG06dP54c//CEATqeTVatWMX36dGbMmMGvf/1rAB555BGmTp3KzJkzue666878h9VLkhkLIUSY+K/X9rC3uDag95w6Op77Pjetz48rKipi48aNmM1mamtr2bBhAxaLhTVr1vCjH/2If/7zn50es3//ft5//33q6uqYNGkSt912W6flP9u3b2fPnj2MHj2axYsX8+GHH5Kfn8/XvvY11q9fT05ODitXrux1O4uLi/nhD3/I1q1bSUpK4oILLuDVV18lKyuLkydPsnv3bgCqq6sBeOihhygoKCAiIsJ7bDBIZiyEEKLPrrnmGsxmMwA1NTVcc801TJ8+nTvvvJM9e/b4fcwll1xCREQEqamppKenU1JS0uma+fPnk5mZiclkYtasWRQWFrJ//35yc3O9S4X6Eow/+eQTli1bRlpaGhaLheuvv57169eTm5vL0aNH+da3vsXbb79NfHw8ADNnzuT666/nb3/7W5fd7wNBMmMhhAgT/clgB0pMTIz365/85CcsX76cV155hcLCQpYtW+b3MREREd6vzWYzDoejX9cEQlJSEjt37uSdd97hD3/4Ay+++CJPPvkkb7zxBuvXr+e1117jwQcf5NNPPx2UoCyZsRBCiDNSU1NDRkYGAE8//XTA7z9p0iSOHj1KYWEhAC+88EKvHzt//nzWrVtHeXk5TqeT5557jqVLl1JeXo7L5eKqq67igQceYNu2bbhcLk6cOMHy5cv5xS9+QU1NDfX19QF/Pf5IZiyEEOKM/OAHP+DGG2/kgQce4JJLLgn4/aOionjssce48MILiYmJYd68eV1e++6775KZmen9/h//+AcPPfQQy5cvR2vNJZdcwooVK9i5cydf/vKXcblcAPz85z/H6XTyxS9+kZqaGrTW3H777SQmJgb89fijPDPVBlt+fr6W/U+F6JlSaqvWuuf1IUEkv88DZ9++fUyZMiXYzQi6+vp6YmNj0VrzjW98gwkTJnDnnXcGu1ld8vfv1t3vsnRTCyEGXG1zK40tAzP2J4aHJ554glmzZjFt2jRqamr42te+FuwmBZQEYyEGwGW/+4BH3z8c7GaEjPwH1vDIu/LzEP135513smPHDvbu3cuzzz5LdHR0sJsUUBKMhQiw5lYnu4pqeG9/abCbEjJsZhOtTlewmyFEyJJgLESAFVc3AbD7ZI0EIDerWdHikJ+FEF2RYCxEgJ10B2O7w8XBkrogtyY0WCUzFqJbEoyFCLCTVU3er3cV1QSxJaHDZjHRIsFYiC5JMBYho97uYE9x+AevoqomzCZFfKSFnSeqg92ckGCMGQdnGaU4M8uXL+edd95pd+w3v/kNt912W5ePWbZsGZ6lbhdffLHfGs/3338/Dz/8cLfP/eqrr7J3717v9/feey9r1qzpQ+v9C8WtFiUYi5Dx1AcFXPHoRppanMFuyhk5Wd3EyPhI8rIS2SmZMeDuppYx47C0cuVKnn/++XbHnn/++V7Xh37zzTf7XTijYzD+6U9/yvnnn9+ve4U6CcYiZBRUNNDidFFY0RDsppyRk1VNZCRFkZeZyMGSurD/cBEIVouSMeMwdfXVV/PGG2/Q0tICQGFhIcXFxZxzzjncdttt5OfnM23aNO677z6/j8/Ozqa8vByABx98kIkTJ3L22Wd7t1kEYw3xvHnzyMvL46qrrqKxsZGNGzeyevVqvv/97zNr1iyOHDnCqlWreOmllwCj0tbs2bOZMWMGN910E3a73ft89913H3PmzGHGjBns37+/1681mFstSjlMETI8s5ALyhuYMio+yK3pv6KqRs7KTSEvKxGnS7OnuIb87ORgNyuorGYZMw6It+6C058G9p4jZ8BFD3V5Ojk5mfnz5/PWW2+xYsUKnn/+eT7/+c+jlOLBBx8kOTkZp9PJeeedx65du5g5c6bf+2zdupXnn3+eHTt24HA4mDNnDnPnzgXgyiuv5Oabbwbgnnvu4c9//jPf+ta3uOyyy7j00ku5+uqr292rubmZVatW8e677zJx4kRuuOEGfv/733PHHXcAkJqayrZt23jsscd4+OGH+dOf/tTjjyHYWy1KZixCxqmaZsAIxuGq1enidG2zOzNOAJCuatzBWLqpw5ZvV7VvF/WLL77InDlzmD17Nnv27GnXpdzRhg0buOKKK4iOjiY+Pp7LLrvMe2737t2cc845zJgxg2effbbLLRg9Dhw4QE5ODhMnTgTgxhtvZP369d7zV155JQBz5871bi7Rk2BvtSiZsQgJLpfmVLURjI+WhW8wPl3TjEtDZlIU6fGRjEqIlElcQITFRINdymGesW4y2IG0YsUK7rzzTrZt20ZjYyNz586loKCAhx9+mE8++YSkpCRWrVpFc3Nzv+6/atUqXn31VfLy8nj66adZu3btGbXXsw1jILZgHKytFiUzFiGhoqHF241ZUH5mW5Z9XFDJK9uLAtGsPityL2vKSDRK9c3MTGBXUXVQ2hJKrDKbOqzFxsayfPlybrrpJm9WXFtbS0xMDAkJCZSUlPDWW291e48lS5bw6quv0tTURF1dHa+99pr3XF1dHaNGjaK1tZVnn33WezwuLo66us5r9SdNmkRhYSGHDxslVv/617+ydOnSM3qNwd5qUTJjERI848XpcRFn3E395w+Osv14NVfMzuz54gDzFPzITIoCYGZmIu/sKaG6sYXEaNugtydUWM0ygSvcrVy5kiuuuMLbXZ2Xl8fs2bOZPHkyWVlZLF68uNvHz5kzh2uvvZa8vDzS09PbbYP4s5/9jAULFpCWlsaCBQu8Afi6667j5ptv5pFHHvFO3AKIjIzkqaee4pprrsHhcDBv3jxuvfXWPr2eUNtqUbZQFCHhrU9Pcduz27hidgavbD/J9p98hqSY/gWva/6wkW3Hqzn0wEWYTCrALe3eb9Yc5DdrDnHggQuJsJj58HA51/9pM8/cNJ8lE9P6dc+hsIXiN/++jb3Ftbz3vWWD16ghQrZQDE993UJRMmMREordk7cWj0/lle0nKaho6HcwrmxowenSVDe1ktzPe/TXyaom0uMiiLCYAZieYUzi+tErn5ISG4ECvr5sHBdMGzmo7Qo2m8ymFqJbMmYsQkJxdRNRVjOzxyQCUHAGk7gqG4z1kOX19l5df7yikWM9rG0uKG/gdE3Pk1OKqpq8XdQACVFWvr5sHOPSYkmMslJS28z9q/dgdwyvtcc2i9SmFqI7khmLkFBc3cSoxEjGJEdjNimO9nMSlycjBiivszNxRFy317tcmlVPf4zLpXnvu8u67Na+7W9bGREfyV9umt/t/U5WN5GXldju2A8unOz92tNt/ffNx/ny4pxevKKhQSZwCdE9yYxFSCiuaSYjMQqr2cSY5Oh+T+KqbmzBMw2irBeZ8YdHyjla1kBhRSPrD5X5vcbp0hwta2BLYSWObrI7p0tzqqZ9ZtzR4vGpLMxN4dH3D9PYMnyW+kg5zDMTrLk9on/68+8lwViEhOLqJkYnGEEsNzWm32uNqxpbvF+X1fUcjJ/ZdIyUGBupsRH8ddMxv9ecrm2mxemiocXJ/tNdb4lYWtdMq1OTkdh1MAb43mcnUV7fwlMfFvbYvqHCalHYpZu6XyIjI6moqJCAHCa01lRUVBAZGdmnx0k3tfBLa41SgzMT2e5wUlZnZ1Si8Z83JzWGD4+U43LpPs+GrqhvC8blPl/7c7K6iXf3lXDr0nGYTYrfvX+YE5WNZCVHt7vumE+WvvVYlXdSVqf7edYYd5MZA8wdm8R5k9P547ojfPGssSREWbu9fiiwufczHsz/V0NFZmYmRUVFlJX577kRoScyMrLdsqnekGAsOjlaVs8lj3zAK99YxOSRA18juqTGyGBHuzPKnLQYmluNspKje8gyO/LNjHuawPX3zUYm/IUFYzCbFI+tPcJr6zfz9eUTIKHtF+lYZSNjVAlYItlyrIobF2X7vZ9njXFWD8EY4DsXTOSSRz7g8fVH+P5nJ/d4fbizmk1obXTlW8wSjPvCarWSkzN85hcMV9JNLTo5WFJHU6uTDw6VD8rzeYKYp3s3JzUG6F+N6gr3TOq0uIhug7Hd4eSFT05w7uQRZCZFMyohivOnpLNgx904/3V7u2sLKxp41Pp//DrhebYUVnZ5T0/1rd58gJg2OoH7PzeVK+cMfmGSYLBZjLcamcQlhH8SjEUnnoC2a5A2OPBU3xqVYHRT56bGAnC0H8G4yt32iSNiux0zfnv3acrrW7hh4VjvsRsWZjNal9BQWtju2uMVjWSZy8k0V3Oqptn74aGjoqomkmNsRNt61+G0anEO49Jie3VtuLOajbcaWWsshH8SjEUnle6x1p2DVFP5VE37jHJEfARRVnO/1hpXNLQQG2FhdEKU38y4uLqJ+1fv4Qcv7SI3NYazx6d6zy3KSSRdVUNj+x6B4+V1xOs6EjGWW3WVHRdVNXY7kzoUKKWeVEqVKqV2d3F+hVJql1Jqh1Jqi1Lq7EA8r83dNS1rjYXwT4Kx6MSTGR+raKS6sftJUIFwsrqZ5BgbkVajapVSipzUmH5tGFHV0EJyjI3UuAgq6ltwudq6RZ/7+DhL/+d9/vbRMS7LG81fbprfboKYaizHjItYZy0uh7FWWWtNbeVpTGhsrdXE2MxsPVbVxeto6nEmdQh4Griwm/PvAnla61nATUDPG8H2gjczluVNQvglwVh04gnGMDBd1Q12R7uAVlzdxOjE9ssActJi+j1mnORequRwaWrcBUAAXvjkBLmpsaz9/jL+55q8TrOmqS0GwKQ0x4qOA8aM7JhWo62qqYo5WQl8Utg5GDe3OjlW0egd7w5VWuv1QJcD31rret22hiYGCMggrycYS2YshH8SjEUnFfV2Jo4wxjL7vBdv9XF4+25o9T+uCvD8Jye46vcb2XbcCGqnatrWGHssijzGFTXP0NzbwhhH18KmR6lqbHGvGzZqUjd9+i/Y+jRaa46W1TMvJ4nMpGj/96g75f3ycEEhAMcqGkhR7g8k2sXCDAsHTtdS19za7qGHSupxujTTRvtf9hROlFJXKKX2A29gZMdnrG0ClwRjIfyRYCw6qWxoYWxKDOPSYtjZ18x421/ho8fgk657Nz0Ttv733wfQWnOyqqnTDORzal7j25aXKdn6r94976ZH4Z0fMaJ2D0nRNtLiIoingbT3vgvv/pTKeju1zQ5yUruZMOXOjAGKi4xlT8cqGkml7WcwbwS4NGw/Xt3uoXtPGddMGz3wS8EGmtb6Fa31ZOBy4GddXaeUusU9rrylpzWwbd3UMptaCH8kGItOKhqM7DIvM5GdRdV9q/xTsM74e8OvoLnW7yWl7lnOHx6u4J09JTS0ODt1U6c1HQUgcdND4OpFNlW6D4CbWv5GSqyNtNgIvmp5A2tLDTRWcMIdXHO760auO+39sqLUCMzHKhpIM7W9jmlJDkyq8ySuPcW1xEZYGNOx6zuMubu0c5VSqV2cf1xrna+1zk9L6357SJtFJnAJ0R0JxsNEbwOqy6WpdE+CmpmZQFmdndO1Pe9WBBjBt2gLjD8fmiqNDNmPsrpmZmYmMDI+kvtWG5N622XGLheRVQc5odNJqD0Ie17usq1Ol4bmGqg5gStxLIvVp0yz7yRN1fAV81vURmUAUF24E6D7Md26U2AzNpawV5+mxeHiWGUj2ZGN3kuiHbVMHR3P5oL2wXhvcS1TRsUN+v7JgaaUGq/cJbKUUnOACKDiTO8rS5uE6J4E42Ggot7O9PveYdORnt9Ta5tbcbo0KbER3t2Hdp7oZVf1sY2gnbDoWzDlMtj4O2jo/JxldXYyk6K4/bwJlNQaWfIo3zHj6mOo1gZeib6a49ZceP9BcLZ2us9PX9/Lpf/3gTcrrlp0D6d0MouPPUbClkeIoJXVOfcC4Dy9B6tZdb/0qLYY0ibiNFlJ1NXsP11LYUUjWbYGwB1kGytZNC6V7ceraWoxtkF0uTT7TtUydVTod1ErpZ4DNgGTlFJFSqmvKKVuVUrd6r7kKmC3UmoH8ChwrQ5AUWTvBC6ZTS2EXxKMh4HjlY00tDi9E6a645lJnRJjY8qoeCwmxS4/640r6u38/K197fflLVgHlkjIOguW/xhaG+CDX3V6bFmdnbTYCK7Jz2RsitGt225JkDu4OtKn8X9cB5VHYfvf2t2joLyBv350jH2namk+uct4WNxUfuu4ktTqnaiPH+dN83J2qikQk0Zk5X7GJEdjMXfzX77uFMSNQkenkUItO4tqOF7RwAhzHSSOMa5prGDRuBRanC4+cXdVH3P/fMNh8pbWeqXWepTW2qq1ztRa/1lr/Qet9R/c53+htZ6mtZ6ltV6otf4gEM8rmbEQ3ZNgPAx46jUfq+h5qZBnowXPut/Jo+L8Fv/457Yi/rjuKB/7dtceXYfOWsCukma2NqVTPu4KXB8/AY1t1zS3OjmrZRN37P8CVlcL/3XZNC6YOoL0uIi2+5TuASAuawb/qJuGI2MevH4H3J9g/PnjEn7znwNGFzVQd/xTiIinRKXxknMJzXHZYLbySvz1xjaK6VNJbTzS/eQtgNpTED8ac1waoy11bDhYRlVjK8nUQMo4UGZoqmR+TjJWs2Kju6dhT7HRczC1q8lbT18KW57s/rmHuAgphylEtyQYDwOVDUYXb2FFYw9XQmWD0W2c4l4alJeZyK6imnbFM8CYfAXGxCUA6sugdA97I2dz2e8+5Krfb+KWPTMwOe3U7HvP+7iyOjsrzB+S1HQMyvaxbFI6j9+Q336stWQvJI4lN2MkoNi/4CFYdjcsvQumroBTO9n96VYunDay7fr0KVQ2tuLAQsUlT8B1z6ETsiivt6PTp5LlOE5uajdd1C0NYK+BuFGomHQybPWsPWDMEI5zVkHsCIhKgsZKom0WZmclsfGIUalrb3EtFpNiwgg/wd5hh8INcHRdjz/7oUzWGQvRPQnGw0B1XzJjbze1kanmZSZS1+xoVye6xeHyZsR7PcHYPYt6vWMqcREWnrlpPhd85kLqdST2g77BuImFpr3GNyV7/TeidC+kT2XSSGMy1c6mNFh2Fyy/G86/H4Dl1n08eMV0bGZFbM1BSJ9KpbvtsWNnw4TzSY2NoLyuhaq4CUQrOzOiu+mm98ykjhsFMWmkqBp3l6omwl4BMakQnWxMTAMWjkvh05M11DS2sqe4lgkj4oiwmP3c1712uaqw6+ceBqzucphSgUsI/yQYDwOeIFVSa/dOOuqKp5s6KcbYY3fhuBQA/rO3xHvNzqJqmlqdxNjM3i5aCtZBRAJvV4xkZlYCSyamcf70TDa7phBd1Dbs2HxiJ8nKXeay1E8wdtih4jCMmEpGYhSxERYOnq7znt5Rn0SRTuW61KOkxEYwJ7mZKGetNxhbTIr4SGOjhtTYCCoa7BwzGZtBTFTHu37hnjXG8aMgJtVddUsTQzMmpx1i0iAq2dvlvnh8KlrDpqMV7O1u8latOxhXH+v6uYcBGTMWonsSjIcB3z1+j1d231Vd2dBCXKTFm+VlJUczKyuR1TvbCmJ8eLgcpeCa/CyOljfQ2OKAo+twjl3MntMNzMxMBIzlShtd04htOAY1RQBEnNgAgDN2tP9gXH4IXA5In4pSiokjYjlQ0haMH1t7hC1qBuMatoHLyaLYUuPECCMYJ8XYvJvXp8VF0OrUbKpPx6UVGS2FXb9wTwYbNxpi0zG7WoiliUmx7kpiMekQnQJNRnY9KyuRKKuZ1TtPUlZn77rYR53759ZUZSzBGqakApcQ3ZNgPAxUNrTgGZLtqavaU/DDS2tuyq5k36laDpcaQXHj4Qqmj05g4bgUtIajh/ZC9TGKk+fjcGnyMo1ZxdE2C7sjZhn3cY+ZJpds4pArA5W7xH83tSdAp08FYNLIOA6crjMqdVU3sWZfCabxy1HNNXBqJ3m2kwDYUyYb66Oj29ruKYm58XgTRaQTU32g6xfeLjM2ClhMT7AzNd79QSYmDaKToNEYK7dZTMzLSWb/nu0kUdv15K3athKbVA3f7NgmS5uE6JYE42GgqrGViSOM8ddjPUziqqi3kxLrM7O58AMu++SLzDMdYPXOUzS2ONh+oopF41K82aBjxwsAfGKaBeBdnwzQnDSJWlOCUTva0UJGzXa2mWdiGjEN6k+3m2kNGMHYZIXUCQBMGhFHVWMrZfV2/r7ZCGb5y1YY1xasI8d1jNM6icKGCG+xEo809+vYdryKk7ZclL9M3MNT8CMizhgfBu5eksqqWe4iITGpbd3U7mW3i8el8KTllzxh+xVTR8V1fV+PYTxubJXZ1EJ0S4LxMFDV0EJOagwJUVaOVXafGXcMaFQaZSkvTSvltZ3FfFJYRatTs2h8KhmJUWRFNjPpyNMw6RI+qEoiLS6CkfFtpS1HJ8WwxTTTGFMu+gSbbmZ/1GwYYWS+nbqqS/ZC6kQwG2PWE92TuHafrOGFT05w7uQRjM7MNjLno+tIazrCQVcmh0vrqWxs3/ZU93KpxhYnNfHjoeIItHZRTazulJEVg9ElDeQltzI+2v3hJTbdmMDltEOrcezsnDjGqFLyTQeJP/F+1/eNNsbdh/O4sXcCl3RTC+GXBONhoKrRGEvNTonuOTPu2E3tzuzOji+loLyBx9cfwWpWzMtOQinF92LfJsLVAOf+mJ1F1eRlJnjHbAEyk6J41z4F6kvg48dxYuJU8jxvN3SnrupSY5mSxyR3Rv/bNYcor2/hSwuNyVjkLIXjm4isPswBPcYIxh0+SKT6ZPjO1ClGdbDyg/5feK1R8APwdlNTXwoNxvIlot2ZMXiz+SlRNZiUO9N792f+a2jXnoK0yRCZMLwzY5PsZyxEdyQYD3Faa6oaW0mKtjImJYbCbsaMPXWpPWuMAe9Y6ljnMaxmxYeHK5g9JolomwXqTnNx47943bWY6rgJHClrIM89ecsjIzGKdQ534N37KgdULjHxKUbgi0z0FvgAjNrWNSfasmYgJTaC1NgIdhbVkJ0SzTnj3XsW5C4FRzPK0UxJVC4HS+qoaWolyScYJ0ZZsbgHyyMzZxgHu+qqrjMKfgDebmoayqGhzGinxWZkxuBd3mSqMTLd+hk3QMmnsPcVP/ctNl5rUvawHjM2mRQWk5IJXEJ0QYLxEFfb7MDp0iRFG5nxyaqmLrMTT13q5BifMWP3+ltL+X6WTTC6WxePcwer9Q9j1k4ebr2Kf+0wgvZMn/FigIykaIp0Ova4LOMhjqmkxUWAUjBimrf0JdD2dfq0dveYNNIopvHFs8a2FQcZu9ioiAW0pkxhy7FKtKZdVm8yKe8Hi7SxU8Fsg5I9dOJyuUthuouImK1GgY+GUiM79mTKHTJjT6Ybe/4PjTa/9yA4ffZf1tpd1csTjAs7P/cwYrOYJBgL0YVeBWOl1IVKqQNKqcNKqbu6uObzSqm9Sqk9Sqm/B7aZor+qGtrKW45Jjsal4aR7P+GOfOtSe3mW5rTUs3KS8d9lycRUI8vb+jQ1U67juB7Bcx8ba3g9M6k9PBszlKScBcAG5zQjGIPRVV26zzshihJjByffzBhgZmYiMTYz18zNajsYGQ8Zc0CZiBo9xbvhhG9mDG1d1dnpCZA6CXa/DC99xfjjqXfdWG4sp4ob3fbAmDQjK24obwvGnrHfJk8wPgbmCONx594DlUdgp89//aYqY4w5bjQkjoXq4+27sv95M+zxk00PUVazSSZwCdGFHoOxUsqMsXvLRcBUYKVSamqHayYAdwOLtdbTgDsC31TRH5WNniIeNrLd2wd2tbzJU/CjfTf1KW+mujypjDXfWcLsMUmw91/gaiX2vO9js5jYf7qOsSnRJEa3D4YZ7mC8Lfli6secyxbXJJ9gPAXs7q5prWHbM8aGDAlZ7e7xrXPH886dS0iItrZv8PxbYO4qske2bbeb4icYp8ZGEB9phVkrwRIBxduNCWWv32kESN9lTR4xaW3d1LGeYOwnM04cAyYTTLoIknPhwNs+Pzuf+yZlG4G53l3pq/o4fPoi1LUVUxnqrGaTTOASogu9yYznA4e11ke11i3A88CKDtfcDDyqta4C0FqXBraZor88mXFStI2x7o3vu5rE5alL7Z0E5bAbWeO45QCo0r2MT3cv4SlYB6mTsKZkeydZzewwXgwQH2klLtLCNj2RnUuewI6tbVOIEe7u6JK9sG81nNph1KD2mQAGxnrlzKTozg2e+Xm49NeMT2+rCZ3U4cPAqkXZfPeCicY3C78Bt28z/tzsnv287hc+pTA7ZMb1pUZXtbebOsn42xOMq48ZQRaMNo/Kaz8G7i0kMgqS3BPPPF3VnlrVuUs7v64hymZWMoFLiC70JhhnACd8vi9yH/M1EZiolPpQKfWRUupCfzdSSt2ilNqilNpSVlbWvxaLPqlqNDaJSI62kRYXQZTV3GUwLvdkxp4xY0+QSpsECWPaJj85Woy9i92BxFMKsmMXtUdmkjFWXVZnBPt2mTEYk5/ee8DoRp55bZ9fo28wbpfVA8snp7Ny/pjOD0rMgnlfhR1/h4L1xrGOmXHdKaOr2ROMzVaIiPfppi5sC7JgdLtXFYLdXe6zXTDOcT/GPYmrYJ2x+UTa5D6/3nBllTFjIboUqAlcFmACsAxYCTyhlErseJHW+nGtdb7WOj8tLS1ATy26482MY6wopRibEt1lN3Wlz/gy0L5EZPqUtglWJ7cYa21zjGA8LcMdjDtM3vLISIyiyF8wjkyA+EzY9Jix5Ojce8DkZ7OFHiRG27zVtjpmxt06+ztgiYLNfwBl8q4vBowA3FLf9rWHe+cmb3lLT2YMbcu1yvYbf9f6BOOETEAZwVprIzPOWdKpF2AoM8aMJRgL4U9vgvFJwHcQL9N9zFcRsFpr3aq1LgAOYgRnEWSVjS1YzYrYCGPzhLEp0Rzroj61py61p46wNxjHjzImVZUfNLLio+uM4JV9NgCXz87g/s9NZe6YJL/3zUyK4mR1E6V1zURYTMS52wIY922qhFGzYMrn+v06x6XFEhfh0/beiE2DhV831h/HjgCzpf05D99g7Nm5yZPhJvpkxh0LmdQVG+uTLTZjrDo+w+jaLt1ndH/nDJ8uajBKYrY4ZAKXEP705p3rE2CCUipHKWUDrgNWd7jmVYysGKVUKka39dHANVP0V1VDC0nRbZsnZKfEcLyiEaer85tieb29/QQo38wufZox47jikFHactQsiEoEjHHhVYtz2u9J7CMzKYp6u4PDpfWkx0e0KwriHTc+794zyhI/M3UEZ09I7fnCjhZ+01hHHD+6/XHfABzrkzF7SmJ6qmn5ZsaJ2WCNbitk4lnW5JE01siMC4bfeDFIN7UQ3bH0dIHW2qGU+ibwDmAGntRa71FK/RTYorVe7T53gVJqL+AEvq+1rhjIhoveqWpsadd1OyYlmhani9O1zWQkRrW71ij44bvGuBgskUbXrCfrO/Gx0U296Fu9boPneXacqCbHPaPba97NkDIBxp3btxfWwVfPyeWr5/TjgVGJcN3fO38QiOkqM04xtnj0TMTyHTM2mYwxYM8krrri9pPCkrLhyPtGz0JSjjETexixmaXohxBd6TEYA2it3wTe7HDsXp+vNfAd9x8RJM2tTn71n4Ncv2AMY1OMoFfV0OrdmxiMzBjgWHmD32Cclewza7nWXQhDKSNgmizwyZ+NDLkPXayemdBVja3Mj4tofzIhA2Zf35eXGXjZizsfaxeMfTLu6GRjvLjqmPEhJbLDpLURU9uWN9WegtFz2s4ljjUCtL0WZlwduPaHCavZJLOpheiCVOAKorI6O82tzoDd7+F3DvD4+qOs3tG293DHzROmjorHalb8e2/n9a3l9X7qUnsyO4vNCMglnxqFLsac1et2edYag8/krVDnnUEdYcyg9ohKNoJpxaH248Ue6dOM5WA1Rcbf8R0yYzAmhg2z8WKQCVxCdEeCcRBd/uiH/HpNFxsX9NHGw+X86YMCAA6U1HmPV3fopk6KsXHxjFH8c2sRjS1tpRtdLk1Vo5+61L5jnp6u6qz5YG2fVXcnKdpKtM2YJZ0eF9nD1SEiIs4IxDFp7buwPYU/ine2Hy/28PyMjrjXMcd1GDP2yFkS0OaGA5vFRItU4BLCLwnGQdJgd3CyuokDp+t6vrgHNU2tfO8fO8lNjeHs8aneexoBtrXTcp8vnTWWOruDV7e3ZdC1za1M1IUke67V2lhn7BtMPEt3+jjxSCnl7RIPm8xYKWPiVkyHSWGewh/2mvbB1cPzMzq8xvjbX2Y8Ykbn+w4DNsmMheiSBOMgOVVj1Ic+0cUyo764f/UeSurs/OraWeRlJXC0vAG7w0mdZ5OIDiUi545NYsqoeJ7ZVIh214WuP7ietyLuZmr9JuOi5mpwNLUPJlnzAQUTLuhzGz1d1WmxYRKMAVLGQ2qHFXqezBj8Z8ax6cZypqNrje99P8zEjjAy7Ul+a+IMeVaZwCVElyQYB8nJamOT+xNVTbj8LDPqrYLyBl7ZfpJbl+YyKyuRSSPjcbo0R8savHWpk30mcIGRqd6wcCz7T9ex9VgVLQ4Xu9f9E4AJtR8ZF3mXNY1se2DOEvjOPqPsYx95NowIm8wY4Nq/wud+2/5YVA/BGIyu6uZq42vfDzNKwdc3w5IfBLKVYUMmcAnRNQnGQXLKvXNSi8NFWb293/d5d58xEeu6ecYyGU+d6AOn66jybBLhpyrVilmjiYu08KcNBXztr1sYWbEZgNQydzD27NYU12H9re8Ych94ZlSnx4dRMI6IA1uHpVi+mbG/CVzQtgWkOaKtW9sjJsWYDDcMyTpjIbrWq6VNIvCKfbYxPFHZyIj4/k1sendfKZNGxHmXJOWkxmA1Kw6U1BEXafzz+gvG0TYLV8/N5KkPC4lXDcyMKDC6UMsPGhO3an2qbwXANXMzSY62MSqh9xO/QpJnG0Vl6rS7lJdnEpdnWZgAPBW4JBgL4Y9kxkFysroZT8Gq4/0cN65pauWTwkrOndJWIcpmMZGbGsuB03Wda0138OVFOYxLi+Hxc5ox4YJzvmecKFjffpODAEiJjeDz87oIXuHEGm1kvPEZXWe4nklcHat6DXM2i+xnLERXJBgHyamaJqaONtavnqhs6uFq/zYcKsPh0pw3Ob3d8Ukj49p3U3cRjMekRPPud5dxFruNDRPmrjIyv6PrjGAcnWLUVBZtlDK6qrsaL4a2nZgC9EFmqJAJXEJ0TYJxkBRXN5GdEsPI+EhOVHWdGf/uvUM89/Fx7I7OxUHe21dKUrSV2R02aJg0Mo6T1U2cqGzCZjYRY+thJ6SCdTB2IVgjIfsc4/vaU53Hi4Vh+lUw7fKuz0fEwqwvwqSLB61J4cBqNuFw6TOasCjEUCVjxgOkot5OhNXs3S3Jl9aa4ppmPjttJCXJzV12U5+uaebhfxtFQX675hA3L8nl+gVjiLSacbo07x8oZfmkdMwdNmjwTOLaXFDh3TqxS7WnjC3/8lYa3+cuhb2vGlskZuT3/YUPB599sOdrLn904NsRZqxm47N/i9NFZD+2yhRiKJPMeIDc+NTHPPD6Xr/nKhpaaHG4GJUQSVZSNEVdBOOdRdUA3HXRZLJTo/nZ63u54cmPqbc72H68iqrG1nbjxR6TRhrB+GBJfc/7+xasN/72FPLwlGlsqgrY5C0hwJjABUhXtRB+SDAeIAVlDRwta/B7zjOTenRiFFnJ0ZyqbfbbDb2rqBqLSbFqUTbP37KQ3143i23Hqrj+T5t5eftJLCbFkolpnR6XkRjl7ZruORivM7YQHDnT+D45t22WsHRTiwDy7DUtk7iE6EyC8QBosDtoaHFyurbZ7/lid8EPTzDWuu2Yr50napg0Mo5IqxFYV8zK4LHr57CvuJa/bz7O/Jxk4iOtnR5nMikmuLuqu5pJDRglL4+ug5xzwNNtqFRbdiyZsQggq2TGQnRJgvEAKK+382vro3yu7gVvuUlf7TJjd2WqjmUxtdbsKqpmZmZiu+MXTBvJEzfmE2Mzc/nsjC7bMNndVZ0U0zlYe1UVQG1R5x2EPF3WkhmLALKajbkLstZYiM5kAtcAqCwvYYVpI8f0YWqaWkns0FVcXN1EpNVEUrSVMSlGsY6Ok7gKKxqpbXaQl9lhv1xg6cQ0dtx3gTfT8GeiOzPutpv61E7j78x57Y9PuQw+W9bnDSGE6I6nm7pFMmMhOpHMeAC4CjZgUpocUwnlJ490On+qppnRCVEopRgRF4nNbOq0vGnniWoA8rIS/T5Hd4EYfDLj7oJxyV6jklTapA43j4SF35A1xiKgpJtaiK5JMB4AMSc/xKmNLjnH4fc7nT9Z3cRo95aCJpMiIymKog6FP3YWVRNpNTEhPbZfbZiemcDkkXHMGpPY9UWle40JW33Ym1iI/vLOpnbIBC4hOpJgPADSyzaxwTWTMp1A1IkPOp0/VdPE6MS2WtRZydGduql3FdUwIyMBSw8ZcFfiI628fccS5nQoCNJO6d620o1CDDCrdFML0SUJxoFWc5Lk5uNss+Sx0TXN2AXJZxJXS8VxGuqq222YkJUU1a6butXpYvfJmk6Tt7pUfgicju6vKTsILp83wZYGqCyAEdN69xxCnCHPBC7pphaiMwnGgVawzvgrPp8dljxiWiuMClcAjZVYHj+be8x/JSPRJxgnR1Pd2EptcysAB0vqsDtczPQzeauThgp47Cx4/4Fu2rQBHp0Hn77YdqzsAKAlMxaDxtNNLbOphehMgnGgHV1HjYqnIWESR+PmeY8B8OFvMdlrWGbeyeiEtm7qMe7tDz3Lm3YV1QAwq4vJW+2UfAouB3z0B6g73fm81vDufxlfH17TdrzUXR1MMmMxSGQClxBdk2AcSFpDwTo+VtNJjY9CJWVRbBplZMt1p2HzH7HbEhmpqhjDSe/DspI8wdiYxLXzRDWJ0VZvkO5WiTuoOltg/cOdzx98G4o+gahko/Slp8u8ZK+xU1N3uw8JEUASjIXomgTjQCo/BHWnWNsylfS4SEbGR/KRng6FH8Dah8DVyjuTjO7kkeWbvQ/zBN03Pj3Fcx8fZ+ORCmZkJHS/wYNH6R6IToW5N8LWp6GqsO2cywXv/syYMX3evVBf0tZlXrrHWNIkBfvFIGlbZyyzqYXoSIJxILnHizc4p5EWF8GI+Ejea5kC9lrY+hTMuYHNKo9i0rAdX+99WEK0lazkKF7bWczdL3/K8cpGFo9P7d1zluyFEVNhyfeNwLr2obZze142gu7yH8P484xjR9caf5fuky5qMajaljZJZixER1KBK5COrqUlNpPjzemkxUVgNZv4q3MqWAFLJCz5PsX/LGK3bRajCzeAy+nNTP99x1JqmowJXCYFaXEdCm5oDR/9HiZ+FlLGGcdcLiPTnXMjxI+G+TfDxt+B2Wbc9+A7MGI6TLsSTCZIyjHGr2d83siSZfKWGERWi7scpnRTC9GJBONAOrWL6pQ5UK5Ii4sg0mqikngqx11Ocs5siB/NqZrDHEucB6X/McpRZswBIMpmJsrWTZfxwbfhnbuh/CB87jfGsepCY9/h9CnG94vvhMPvwoE3je9NVmPvXZO7AyR3Kex+GU67y2B6HifEIJAxYyG6JsE4UBwtUFtEefrFAKTHRRDl3m3p49kPceH0UbQ4XBSUN1A/dxGUYnRru4Nxtzxjv+DtCgfaJm95uptjUuDrm7q+T85SY1x5x9/bP06IQWCVpU1CdEnGjAOl5gRoF6dNIwFIjY1gpHv50ukaY3vE3cU12B0upkwYD2lT2pY89cQz9jtmIVQeheoTxnHP8qS0yb27T84S9/1eMWZXx47o3eOECIAI2c9YiC5JMA6U6mMAnNBpxNjMxERYSI62YTUrSursAGwtrAJg7thko8v4+CZo9b/nsZezFd5/0Bj7vfh/jGOe7Lhkj7E0KaKX9atjUmHEDGNd8ohpxt7FQgwS6aYWomsSjAPFvaToUGuqd/KVyaRIj4ukxJ0Zf1JYydiUaON87jJwNEPRx93fd8ezRjZ87k+MgByT1pZRl+7r+yQsz7aIMnlLDDKzSWFSEoyF8Ce8x4z3robtf4UvvBj8LK/qGJhtHG6KJT2ubSLWiPgITtc2o7Vm67Eqlk1KN06MXQzKDM+tNGZad8VeC5nzjVnUShldzQXrjIy64jBM+Vzf2pm7DDb9TiZviaCwmk0yZiyEH+EdjI99CIf+7V4zG+RMr6oQErIobXAwZWRb5ayRCZEcOF1HQXkDFQ0t5Ge7d1GKjIdL/hdOf9r9fZUJ8m9q+7CRsxR2/xP2vw7a2ffXnbsMzr8fpl/Zt8cJEQA2s0mWNgnhR3gH4+Za4++CdQMejI+U1QMwLq2L8dnqY5CUTdkRO0smtK0RTo+LZP3BcrYcM8aL88f6bGmY/+W+N8TTzfzRY+4n6OOMaLMVzr6z788rRABYLSbpphbCj/AeM7a7g7GnqtQA+s4LO/j637Z1fUFVIY6EMdQ1O9oV7BiZEEm93cG6g2UkRlu7Dua9lZQNiWPh5FajuIenAIgQYcBmNtHqkNnUQnQU3sG42djdiMIPe97Pt5f+998H+P4/drZ/mlYne4prOVBS591ZyePbz2/noVc3Q1MV9VGZQPvqWSPjjfHg9/aVMndMEiZTAMa2Pdlx6kQj0xUiTFgtSjJjIfwI72BsrzWqTLXUQXE3WWsffHi4nNU7i9tNMtl7qhaHy/g0/97+Uu/x0zXNrN5ZzMZPtgJQYRsFtA/G6fHG102tTuZm+3RRn4kcmREtwpPVbMIuwViITsI7GDfXQs45xte9LaDRg7J6O3aHi93FNd5ju05UA5ASY+Ndn2D8+q5itIZR2jhWYjKKaKTFds6MAfLHJgekjeQsNbqoM+YG5n5CDBKjm1qCsRAdhXcwttcaY6gjZ7QvE9lPWmtKa40CHVsKK73HdxXVkB4XwRWzM/joSAX1dqNLfPXOYqZnxLMkrQGAAmca0JYNA4xwB2Ob2cTMzIQzbiMAsWnwzU+MWdZChBGrWSZwCeFP+AZjrY3MOCLeyBRPbIaWxp4f1406uwO7+1P7Fne1LIAdRdXMzEzkvCkjaHG6+OBQOQXlDewqquGyvNGck1pPjY7mrztqMClIiWkLxjERFuIiLEzPiCfSGsC9g5OywWIL3P2EGAQ2i0nKYQrhR/gGY0czuFqN9bq5y8DZAic+OqNblrnLVkbbzGw9VoXWmtrmVo6WNZCXmUB+dhJxkRbe21/CazuLAbh05mgyVSmnTSPYf7qO5JgIzB0mad2waCw3Lso+o7YJMRRYzUrWGQvhR/gGY88a44h4YwMFk+WMx409wfjcyelUNLRQWNHI7iJj7DgvKxGr2cTSiWm8t7+M1TuLmZ+dzOjEKExVxzCn5AB+9iEGvv/ZyayYlXFGbRMiEJRSTyqlSpVSu7s4f71SapdS6lOl1EalVF4gn1+6qYXwL3yLfnjWGEcmGBslZM4zKlM57J2vHTkDZl/f4y1L6+wsMu3mu663mWMppfWNN4lpcXKvpYoFB9YCn+H8KdN4fdcpyuvt/Ozy6cb2htXHGT37AqynlN9gLEQIeRr4HfBMF+cLgKVa6yql1EXA48CCQD25TcphCuFX+AZj38wYIG8l/OcnbXv1erhaobXJ2De4h3rM1VWVPGL9HSkFdq62mLEeM6HR5Fo0ETtdsPt5lt28BZMCpRQXTx8J9afBaSd6RC4PXD7dO2FLiFCktV6vlMru5vxGn28/AjID+fySGQvhX/gGY29m7A7Gc280/nTUWAm/zYP3HoDrnu32ltmHnyFV1aK//C53rNEcq2igqcVJfnYyj5wbAb9fROL2x1g68RKibRZSYiPgmLF1IknZXDt+TABfoBBB9xXgrUDeUCZwCeFf+I4Z2ztkxl2JToZF3zI2Vji5tevrGiuZV/w31pnmozLzmTs2iSNlDRTXNBtLkkZMhZmfh81/5M9XZvF/K2cbj3NvnUhi9pm+IiFChlJqOUYw/mE319yilNqilNpSVlbWq/vKrk1C+Be+wbi5Q2bcnbNug+gUePenXV/z4W+IcDXyUryRXftu6DArK9H4Ytld4GrF9MHDbWUtq48BChKz+v4ahAhBSqmZwJ+AFVrriq6u01o/rrXO11rnp6Wl9ereNimHKYRf4RuMe5sZA0TEwTnfNTaU8DfjuvYUbH6c961LaUqaDHhmTyvMJsW00e5iHcm5MOcG2Po0VBYYx6oKIX40WGTilgh/SqkxwMvAl7TWBwN9f6tsoSiEX+E7ZuydwBXXu+vzvwKbHoXnvwCRie3PtTaAq5VH1DVMdc+GjrSayctMpKnVSZTNp1jHkh8Yk8QeXwa2WGgok7KUImwopZ4DlgGpSqki4D7ACqC1/gNwL5ACPKaMPbQdWuv8QD2/VcphCuFX+AZje60RDE29rGpljYTLfw+7XvR72pl1FrteSmKpz9KkX187y7tBhFf8KFjxKBx5v+3YtCv62nohgkJrvbKH818FvjpQz2/MppYJXEJ0FL7B2FMKsy9yl7ZtP9hBeW0zWr/bbp1wVnK0//vMuNr4I4ToE5vF6KbWWuPOvIUQhPWYcU3vJm/1kqf6VroU7RBiwNjMRgDu1OMkxDAXvsG4P5lxNzzBWCpoCTFwrGbjLUeWNwnRXvgGY3vtgGTGvnsRCyECyxOMZXmTEO2FbzAOcGZcWtcMSGYsxECyWtyZsQRjIdoJ32Bsrwt4ZhwfaQnsnsNCiHYivJmxjBkL4SuMg3GAx4zr7ZIVCzHArBZjApesNRaivfAMxo4WcDQHNDMurZVgLMRAkzFjIfwLz2DsLYWZELBbltXbSY+T7Q+FGEieYGyXzFiIdsIzGDfXGH/3thRmL5TVSWYsxECzSWYshF/hGYw77mV8hhrsDhpbnBKMhRhgVpnAJYRf4RmMm/uwY1MvlMoaYyEGhc0imbEQ/oRnMA5wZuwthRkvwViIgWR1l8OUdcZCtBeewTjAmbGUwhRicEg5TCH8C89g7M2MAzObusxTfUu6qYUYUNJNLYR/4RmMvZlxYGZTl9bZsZgUSdG2gNxPCOGfrDMWwr/wDMb2WrBGg9kakNuV1dlJjY3AZJL9VYUYSN7M2CGzqYXwFb7BOIClMEvq7DJ5S4hBIBO4hPAvPINxc2C3TyyubmJ0QlTA7ieE8M8mE7iE8Cs8g3EAM2OttRGMEyUYCzHQZMxYCP/CMxg31wZs8lZtk1F9a3Si1KUWYqBJMBbCv/AMxvbAdVOfrG4CkMxYiEHgHTOWbmoh2gnPYNwcuG7qYgnGQgwapRRRVjONLc5gN0WIkBKewdheG7CCH6dq3ME4QbqphRgM8VEW6podwW6GECEl/IKxsxVaGwOWGZ+sbsZqVqRK9S0hBkVcpJXa5tZgN0OIkBJ+wdheZ/wdoDHj4uomRiVEScEPIQZJXKRkxkJ0FH7BuLnG+DtAmfGpmiZGSRe1EIMmPtJKnWTGQrQTfsE4wNsnFlc3kyGTt4QYNHGRFmolMxainTAMxu5u6gBkxk6X5nRtM6NkjbEQgyZOMmMhOgm/YNwcuMy4tK4Zp0vLsiYhBlF8lGTGQnQUfsHY000dgMxY1hgLMfjiI620OFw0t8paYyE8wi8YNwcyGDcDyCYRQgyiuEgLgMyoFsJH+AVju3s2dQC6qdsyYxkzFmKwxEca+5DLuLEQbcIvGDvsoExgOfMiHcXVTcRFWohzvzkIIQaeJzOWcWMh2oRfMHY5wGQJyK2Ka5qli1qIwbD/DTi9G4D4KMmMheioV8FYKXWhUuqAUuqwUuouP+dXKaXKlFI73H++GvimujlbAxeMq5uki1qIwfDSV2DX84CMGQvhT49RTSllBh4FPgMUAZ8opVZrrfd2uPQFrfU3B6CN7bmcAQ3Gs7ISA3IvIUQ3ImK9NQI8w0K1TZIZC+HRm8x4PnBYa31Ua90CPA+sGNhmdcPlAJP5jG/T1OKkqrFVljUJMRgi4sBeD0C8ZMZCdNKbYJwBnPD5vsh9rKOrlFK7lFIvKaWyAtI6f1wOMJ35hKviGplJLcSgsbVlxjE2C0rJmLEQvgI1ges1IFtrPRP4D/AXfxcppW5RSm1RSm0pKyvr3zMFaAKXd1mTTOASYuBFxEOLkRmbTIrYCKnCJYSv3gTjk4BvppvpPualta7QWtvd3/4JmOvvRlrrx7XW+Vrr/LS0tP60N2DB+JSn4Id0Uwsx8CJi26rnYaw1lj2NhWjTm2D8CTBBKZWjlLIB1wGrfS9QSo3y+fYyYF/gmthBgMaMT1Y3oRSMiJduaiEGnM+YMciexkJ01GOKqbV2KKW+CbwDmIEntdZ7lFI/BbZorVcDtyulLgMcQCWwasBaHKDM+HBZPWmxEdgs4bfUWoiw4zNmDO7MWGZTC+HVq6imtX4TeLPDsXt9vr4buDuwTeuCywHmM5vAtfFwOW/sOsWqRdmBaZMQonsRsd4xYzB2bvLUhhdChGUFLucZdVPXNLXyvX/sJDc1hh9cOCmADRNCdCkiHhzNRtEe3Hsa2yUzFsIjDIPxmXVT3796DyV1dn517SyibYEpHiKE6IEt1vjbW/jDQm2TjBkL4RF+wfgMymH+a8dJXtl+km8uHy+Vt4QYTBFxxt8tnsIfVurtDrTWQWyUEKEj/IJxLzJjh9NFg739p+7XdxXz3Rd3MndsEt88d/xAtlAI0VFE58zY6dI0tjiD2CghQkf49dP2ojb1o+8f4XfvH+KqOZnctmwcWwqr+P5LRiB+ctU8rObw+wwiRFjzZMbu5U3e+tTNrcREhN/bkBCBFn6/BS5Hj3sZHymrx6QUL28/yYtbTqCBReNSeOKGfBknFiIYbJ5gbGTG8VFt9alHJQSrUUKEjvCLTL3opq5qbGHyqHie+NJc/vxhAbVNDu773FQirWdeLEQI0Q/eMWPZuUkIf8IwGPc8gau6sZWUWBvp8ZHcfdGUQWqYEKJLfsaMQXZuEsIj/AZPe7HOuKqxhaRo2yA1SAjRow5jxvE+Y8ZCiLAMxj1X4KpubJVgLEQo6Thm7M6MZecmIQzhGYy76aZucbiotztIij7zPY+FEAFitoAlqtOYsexpLIRhyAXj6sYWABJjJDMWIqREtG0WEWk1YTUrGTMWwi38grGz+y0UqxqNT9qSGQsRYmyx3jFjpRRxsnOTEF7hF4x7yIyr3JmxjBkLEWIi4trv3CR7GgvhFabBuOus19tNLZmxEKElIq7dnsZxkVYZMxbCLQyDcfflMNu6qSUzFiKkdArGFplNLYRbGAbjnsaMpZtaiJBki20XjOMlMxbCKwyDcfcVuKobW4mwmIiySelLIUJKhzHjOBkzFsIrDINxDxO4GqT6lhBdUUo9qZQqVUrt7uL8ZKXUJqWUXSn1vYA+eURspzFjmU0thCG8grHLBdrVbQWuqsZWmbwlRNeeBi7s5nwlcDvwcMCfOSIeHM3gNAJwfJSFhhYnTpcO+FMJEW7CKxhr90bk3YwZV0tdaiG6pLVejxFwuzpfqrX+BAh8ymrruFmE8aG5XrqqhQizYOxy/9L2sM44KUYyYyFCjncbRWPcOM5bn1q6qoUIr2Ds7t7qaQJXomTGQgw4pdQtSqktSqktZWVlPT+gwzaKsnOTEG3CKxj3kBlrraluapVSmEIMAq3141rrfK11flpaWs8P6LSNouxpLIRHmAVjz5ix/2Bc2+zA6dIyZixEKOqwjaJnzFhmVAsBXff3hqIeMuO2UpgSjIXwRyn1HLAMSFVKFQH3AVYArfUflFIjgS1APOBSSt0BTNVa157xk3vHjN3d1FGSGQvhMaSCsezYJET3tNYrezh/GsgckCf3jhl7JnDJnsZCeIRZN3X3E7i8pTBlL2MhQk+npU0WzCZFYUVjEBslRGgIs2Dc/ZhxtdSlFiJ0dVjaZDWbuHjGKP6x5YT3d1eI4SrMgrG7m9rcRWbcIN3UQoQssxUskWBvG37+xvJxNLQ4eXpjYfDaJUQICM9g3E1mbFJt6xeFECEmIs47ZgwweWQ8508ZwVMfFlJvl4lcYvgaUsG4qrGVhCgrJpMaxEYJIXqtwzaKAN88dzw1Ta08+9GxIDVKiOALr2Ds7CkYS11qIUJah20UAWZlJXL2+FSe2FBAc6szSA0TIrjCKxh7M2P/G0VUy45NQoS2iLhOmTHAN5aPp7zezuodxUFolBDBF6bB2H9mXCl7GQsR2roIxmflJhNlNXOgpPM5IYaDMA3G/rPf6sYWqb4lRCjzM2YMoJRidGIkxdVNQWiUEMEXZsG4+3XGVY2ySYQQIc3PmLHH6MQoimuaB7lBQoSGMAvGXY8ZN7c6aWp1SvUtIUJZhP/MGGB0QpRkxmLYCrNg3HU5zGp3XWqZwCVECIuIB0dz297kPkYnRlFWZ8fukBnVYvgJs2Dc9QSuKimFKUTo61Cf2tfoxEgATktXtRiGwiwYuz8xmztnv1Xe7RMlMxYiZHl2bvIzbpyRGAXASemqFsNQmAXjrseMq73bJ0pmLETI8mwWYe8cjEe7g/GpasmMxfATpsFYuqmFCEs2TzDu3E09MsHoppZJXGI4Cq9g7JQJXEKENe82ip2DcaTVTGqsjeIaCcZi+AmvYNxNZlzZ0EK0zUyk1X+pTCFECIjoegIXGF3VJ6WbWgxDYRaMuy76cbq2mZHxkYPcICFEn3QzZgzGWuNT0k0thqEwC8ZdZ8ana5q9Y05CiBDVzdImcFfhqm5Caz2IjRIi+CQYCyEGT0TXE7jAWGvc0OKktskxiI0SIvjCLBj7n8DldGlKapsZJcFYiNBmtkJUEtSd8nt6tKw1FsNUmAVj/2PGFfV2HC7NyISoIDRKCNEnSTlQVeD3lCcYy/ImMdyEWTB2gDKBqX2zT7nL542SCVxChL7kXKg86veUpyTmKVneJIaZ8AvGfsaLPcFYxoyFCAPJOVBTBI6WTqdSYyKwmU2yvEkMO0MiGJ92f4qWMWMhwkByLmgXVB/vdMpkUoxMiJRuajHshFcwdnadGdvMJpJlL2MhQl9yrvF3l+PGEozF8BNewdjl8LtJxCn3sialVBAaJYTok6Qc4+8ux42jJBiLYScMg3Hn2tOyxliIMBKbDtYYqPSfGWckRlFSZ8fhdA1yw4QInjAMxn66qWubGC3BWIjwoJQxiaubzNjp0pTW2Qe5YUIET5gFY2enYOxyaUpq7LLGWIhwktz1WuNRspWiGIbCLBi3dhozrmxsocXpkpnUQoSTpByoKmwr5OMjQ6pwiWEozIJx527q07LGWIjwk5wLzhaoLe50qq0Kl6w1FsNH+AVjc/sJXN7qWxKMhQgfyV3PqI6JsJAYbeVkdeMgN0qI4AmzYOzs1E3tKfghmbEQYaSHtcYZiVGcrJJuajF8hFkw7txNXVzTjMWkSI2JCFKjhBB9Fp9hLFPsYkZ1RmKUjBmLYSW8grGz1e+Y8Yj4SEwmKfghRNgwmSEpu+u1xklRFFU1obUe3HYJESThFYz9ZManappkvFiIcJSc023hj8YWJ9WNrYPcKCGCI8yCced1xlJ9S4gwlZxrjBn7yX4zk6IBWd4kho8wC8btM2OtNadqmiUzFiIcJeVASz00lHU6lZlkLG8qkklcYpgI62Bc3diK3eGS6ltChCPPjGo/XdVS+EMMN2EWjNtP4JI1xkKEsW7WGidGW4m2mWV5kxg2wiwYt19nfLrW+EWVYCxEGEocC8oMFYc6nVJKuZc3SeEPMTyEWTBuX4GrLTOWbmohwo7FBmmT4dQuv6czkmStsRg+wi8Y+3ZTVzdjUpAaawtio4QQ/TYqD07t8DujWqpwieEkrINxcU0TI+MjsZjD62UIIdxG5RmzqetOdzqVmRRNVWMrDXZHEBomxOAKryjmdLQbMy6ubmJUonRRCxG2RuUZf5/a2elURpLMqBbDR3gF447d1DXN3u3WhBBhaOQMQPkPxp7lTdJVLYaBsA3GLpfmVHUzoxNlJrUQYSsiFlIn+A3G3sIfkhmLYSDMgrHT2OkFKG+w0+J0MVpmUgsR3kbl+Q3GabER2MwmyYzFsBBmwbhtzPhUtbGsSbqphQhzo/KgtggaytsdNpkUoxIj240Zn6ppwuWSnZzE0BOGwdjopi52/4JKN7UQYa67SVyJUZysMgp/7DxRzdm/eJ+XthYNZuuEGBRhFozbymEWuwt+SDe1EGFu5Ezj766CcbWxr/HPXt+L06X5997Oy6CECHfhE4xdLtCudplxlNVMYrS1hwcKIUJaVCIkZXcxiSuaklo7r+44yZZjVWQkRvHh4QrsDuegN1OIgRQ+wVi7f/nMbcF4dGIkSqkgNkoIERCjZnW71vi+f+1h8sg47r9sGk2tTj4uqBzkBgoxsMInGLvcVXh8uqll8pYQQ8SoPKgqgKbqdoc9a41rmx385NKpnD0+FZvFxNoDnfdAFiKchW8wrm6S8WIhhgrPJK7T7TeN8Kw1Pn9KOovHpxJlM3NWbgrvHygd7BYKMaDCJxg7W42/TRbsDidldXbJjIUYKkbNMv4u2tLucGZSFD+7fDoPXjHDe2z5pDSOljVwvEK2VxRDR6+CsVLqQqXUAaXUYaXUXd1cd5VSSiul8gPXRDeXe8zYZKGkxg7AKFnWJMTQEJMCo2fD/tfbHVZK8aWzxjIivu13fdmkdADWHpTsWAwdPQZjpZQZeBS4CJgKrFRKTfVzXRzwbWBzoBsJtOum9hQByJDMWIihY+oKOLkVqo93e1lOagzZKdG8v1+CsRg6epMZzwcOa62Paq1bgOeBFX6u+xnwC6A5gO1r4xOMT9UYwXhUgmTGQgwZUy4z/t73Wo+XLpuUzqajFTS3yhInMTT0JhhnACd8vi9yH/NSSs0BsrTWbwSwbe35BOO26luSGQsxZKSMM3Zx2vuvHi9dNimN5lYXT31YiNZSHlOEvzOewKWUMgG/Ar7bi2tvUUptUUptKSvr49KEdt3UzaTE2Ii0mrt/jBCiHaXUk0qpUqXU7i7OK6XUI+75IbvcH7QHz9QVcGIz1Jzs9rJF41JZOjGNX7y9n5ue/oTS2oHpkBNisPQmGJ8Esny+z3Qf84gDpgNrlVKFwFnAan+TuLTWj2ut87XW+WlpaX1rqTcYmzlV0ySTt4Ton6eBC7s5fxEwwf3nFuD3g9CmNlMvN/7uoavaZjHx1Kp5/Ndl09h0tILP/mY9BeUNA98+IQZIb4LxJ8AEpVSOUsoGXAes9pzUWtdorVO11tla62zgI+AyrfUW/7frJ08wNltljbEQ/aS1Xg90V75qBfCMNnwEJCqlRg1O6zD2Nk6f2quuapNJceOibF75+mKqGlt5TyZ0iTDWYzDWWjuAbwLvAPuAF7XWe5RSP1VKXTbQDfTyncBVLdW3hBggPc4R8TijYafuTL0cjm+Cut5tCDFlVDxJ0VYOl9YHrg1CDLJejRlrrd/UWk/UWo/TWj/oPnav1nq1n2uXBTwrBu8640YH1NkdsnWiEEF2RsNO3Zm6AtCw59VeP2R8eiyHS+sC1wYhBlnYVeAqbzSCsmTGQgyInuaIDLz0yUYBkK1PQy9nShvBWDJjEb7CJxi7u6krGl2ABGMhBshq4Ab3rOqzgBqt9alBb0X+TVC2D45/1KvLx6XFUtXYSkW9fYAbJsTACLtgXOrJjGUClxB9ppR6DtgETFJKFSmlvqKUulUpdav7kjeBo8Bh4Ang60Fp6PSrICIetjzZq8vHp8cCSHYswpYl2A3oNfeYcWWTkRmnxUUEszVChCWt9coezmvgG4PUnK7ZYiDvOqOr+sKHjNrV3fAG47J6FuR2f60QoSjsMuNmhyLCYsJsUkFukBBiQOXfBM4W2PFsj5eOTogi2maWzFiErTAKxsYELrvLCMZCiCEufQqMWQRbnwKXq9tLTSbFuDSZxCXCV/hENU9m7DIRIWUwhRge8m+CyqNw9P0eL5UZ1SKchVEwNsaMm50mbObwabYQ4gxMvQxiR8DGR3q8dHx6LKdqmqm3OwahYUIEVvhENU9m7IQIa/g0WwhxBiwRsPCbcHQtFG3t9tJxacYkriOSHYswFD5RzR2MmyQzFmJ4yf8yRCbCB7/q9jJZ3iTCWfhENXcFLrtTyZixEMNJRBwsuBX2vw6l+7u8bGxKNBaT4nCZBGMRfsInGHszY2Q2tRDDzYKvgTUGPvh1l5dYzSZyUmMkMxZhKXyimmejCKdJgrEQw010stFd/ek/oLKgy8vGp8fKmLEIS+ET1byZsawzFmJYWvhNsETCK7eCo8XvJePTYzlW2Yjd4RzkxglxZsInqnmCsUNhk2AsxPATPwpW/B+c+Aj+8xO/l4xPj8Xp0hSWNw5y44Q4M+ET1dwVuJocEGGRCVxCDEvTr4KzvgGb/wC7/tHp9MQRcQDsPlkz2C0T4oyEUTB2jxk7ZAKXEMPaZ/4LxiyE126HsgPtTk0aEUdaXATvHygNUuOE6J/wiWouBygTzU6km1qI4cxshWueBrMN3vohaO09ZTIpzp2UzrqDZbQ6u69nLUQoCZ+o5nKAyUKLwyWZsRDDXdxIWPpDo2b14TXtTp07JZ26ZgdbCquC1Dgh+i58opo7GNsdLhkzFkLAvK9Cci6882NwttWjPnt8Kjaziff2lwSxcUL0TfgEY6cDbTLjdGnpphZCgMUGn/kZlB+AbU97D8dEWDhrXArv7pdxYxE+wiequTNjkAlcQgi3yZdA9jnw/n9DU7X38HmT0zla1kBBeUPw2iZEH4RPVHM50MoIxpIZCyEAUAoueAAaK+DjJ7yHz52cDsC7+6SrWoSH8IlqLgfamxnLmLEQwm30LBh/Pnz8OLQ2A5CVHM3EEbG8J13VIkyEUTB2opURhKWbWgjRzqJvQUOpUbva7dzJI/i4oJLa5tYgNkyI3gmfqOZy4HIHY+mmFkK0k7MURsyATb/zrju+YNoIHC7NPa/sljXHIuSFT1RztXrHjCUzFkK0oxQs+iaU7feuO54zJokfXjiZ1TuL+fqz22hulc0jROgKn6gmmbEQojvTroS40bDxEe+h25aN46crpvGfvSXc/MwWCcgiZIVPVHM5cSqZwCWE6ILFBgu+BgXr4dQu7+EbFmbzy6tmsuFQOS98ciKIDRSia2EUjNsy4whr+DRbCDGI5t4IlijY8mS7w5+fl0VeViJ/2VSIy6W7eLAQwRM+Uc3lwIW7m9ocPs0WQgyiqCSYdoUxq9pe3+7UqkVjOVrWwIbD5UFqnBBdC5+o5mzF6c6MIyUzFkJ0Ze4qaKmH3f9sd/jiGaNIjY3gLxsLg9IsIboTPlHN5cTpbq6MGQshupQ1H9KmwNan2x2OsJi5fsEY3j9QSqG7TOb+07X8/K19NLXIxC4RXGEUjB04kXKYQogeKGVkx8Xb4NTOdqeuXzAGs1L8ZVMhz2wq5LLffcgf1x1lw6Gy4LRVCLfwiWouh09mHD7NFkIEwczPgyUStv6l3eH0+EgumTmKpz4s5N5/7WFhbgomBbtP1gSpoUIYwiequRw4kHXGQoheiE6GqZcbE7la2u/cdMuSXEYnRHLvpVN5+svzmJAex+7i2uC0Uwi38IlqvsFYZlMLIXoydxXYa9vVqwaYNjqBjXefx01n56CUYlpGvGTGIujCJ6q5HDgwYTEpLBKMhRA9GXOWUa968+PeetX+TB+dQGmdndLa5kFsnBDthU9UczlwaLN0UQshekcpoyJX6R4o/KDLy6ZnJACwu1iyYxE84RPZXA5aMcnkLSFE7824GqKSYfMfurxk6uh4lILdJ2XcWARP+EQ2lxOHSzJjIUQfWKOMseMDb0LVMb+XxEZYyEmNkXFjEVThE9mcre7MWAp+CCH6YN5XAAWf/KnLS6aPTmCPzKgWQRQ+wdjloFVLN7UQoo8SMmHK52DbXzotc/KYnhHPyeomKhtaBrlxQhjCJ7K5nLRIN7UQoj/Oug2aazoVAfGYPto9iUu6qkWQhE9kczlokcxYCNEfY86C7HPgg19DS2On09NkRrUIsvCJbN5uahkzFkL0w/IfQUNpp72OARKirIxJjmaPzKgWQRJGwbiVFpeSbmohRP+MXQQ5S93Zceex4+kZ8ewurqGwvIHvvLiDxQ+9R2mdFAIRgyM8IpvLBdpFi0u6qYUQZ2D5j6Cx3O/M6mmjEzhW0ch5v1rH6h3FnKxu4uOCyiA0UgxH4RHZtLHXqF2bJDMWQvTfmLNg3Lnw4W/BXt/u1NKJaaTG2rhxYTbrfrAcm9nEp0UyhiwGR3hENpcDALtLxoyFEGdo+Y+hsQI2Pdru8PSMBLbc8xnu/dxUMhKjmDIqjl0SjMUgCb9gbA2PJgshQlRmPky5zMiO60u7vGxGZgK7T9bgcnW9yYQQgRIekc3ZCoDdpWT7RCHEmTv/fnDaYe3Pu7xkZmYidXYHBRX+C4UIEUjhEdlc7jFjp5LMWAhx5lLGQf5NRhGQsoN+L5mZaaw9lnFjMRjCI7K5u6mbXSYiJDMWQgTCkh+ANRrW3O/39Pi0WCKtJhk3FoMiPCKbOxg7MBNhlQlcQogAiE2Ds78NB96A45s7nbaYTUwbncCuourBb5sYdsIqGDulHKYQIpDO+jpEp8D6X/o9PTPT2M3J4XQNcsPEcBMekc0nM5Z1xkKIgLHFwMJvwOE1cHJrp9MzMxNoanVypEwmcYmBFR6RzZMZY5bMWAgRWPNuhshEWP9wp1MzMhIB2Cld1WKAhUdk8x0zlqIfQohAiow3uqsPvAmnP213Kjc1htgIi8yoFgMuzIKxlMMUQgyABV+DiHhY/z/tDptMiukZ8eySfY7FAAuPyOZeZyzd1EKIARGVCPNvgb3/6pQdz8xMZF9xLQ+/c4D7/rWbn7+1j1aZ0CUCLDwim0zgEkIMtIXfgKgkePtu0G0lMM+ZkEqry8Vjaw/z8raT/HHdUV7bWRzEhoqhKDwim7scphPZKEIIMUCik2HZj6BwA+x/w3v4nAlp7P/ZhRz574vZed8FTEiP5fH1R9FaalaLwAmPYOzOjFu1dFMLIQZQ/k2QNhn+fQ847N7DERYzSilMJsUtS3LZf7qO9YfKg9hQMdSER2TzGTOWbmohxIAxW+Cz/w1VBbD5D34vWTErgxHxETy+/sggN04MZeER2dotbQqPJgsRipRSFyqlDiilDiul7vJzfqxS6l2l1C6l1FqlVGYw2hlU48+DiRfCuv+BupJOp20WE19enMOHhyvYLbOsRYCER2TzFv2QpU1C9JdSygw8ClwETAVWKqWmdrjsYeAZrfVM4KdA13sMDmWf/W9ji8V3fuT39BcWjCE2wsLj648OcsPEUBUekc1lTOCSoh9CnJH5wGGt9VGtdQvwPLCiwzVTgffcX7/v5/zwkDIOzvku7H4JDr/b6XR8pJUvLBjDG5+eorSuOQgNFENNmARjz5ixSfYzFqL/MoATPt8XuY/52glc6f76CiBOKZUyCG0LPYvvgORx8MZ3obWp0+krZmfgdGnW7i8b/LaJISc8IptnNjUWbLKfsRAD6XvAUqXUdmApcBJw+rtQKXWLUmqLUmpLWdkQDEjWSLj018Zkrg3/2+n05JFxjE6IZM2+zuPKQvRVeEQ22UJRiEA4CWT5fJ/pPualtS7WWl+ptZ4N/Nh9rNrfzbTWj2ut87XW+WlpaQPU5CDLXQozr4UPfgPlh9qdUkpx7pR0PjhcTnOr388rQvRaeEQ2dzA2WawopYLcGCHC1ifABKVUjlLKBlwHrPa9QCmVqpTyvC/cDTw5yG0MPRc8CNZoePP77SpzAZw3eQSNLU42F1QGqXFiqAiPYOyuwGU2W4LcECHCl9baAXwTeAfYB7yotd6jlPqpUuoy92XLgANKqYPACODBoDQ2lMSmwbn3wNH3jdrVPhaOSyHSauJd6aoWZyg8grF7ApfZYg1yQ4QIb1rrN7XWE7XW47TWD7qP3au1Xu3++iWt9QT3NV/VWtu7v+MwkX8TjJxhLHWy13sPR1rNnD0+jXf3lUp5THFGwiQYG93UyizBWAgRBGYLXPy/UHuy0zaL501J52R1EwdL6rt4sBA9C6tgbLHagtwQIcSwNWYBzLoeNv0OKtuKfSyflA4gs6rFGQmTYGyMGVssMmYshAii8+4FkwXW/dJ7aGRCJNMz4nlvf2kQGybCXXgE49ZmXJgwWSQzFkIEUdxImH8z7HoByg54D583eQTbjldxqKQuiI0T4Sw8grGjGbuKIMIqpTCFEEG2+A6wRMHatrLdNywcS2KUlR/+cxdOl0zkEn0XHsG4tZEWbLJJhBAi+GJS4azbYM8rcHo3ACmxEfzk0qlsO17N3z46FuQGinAUHtGttYlmFSGbRAghQsOib0JEArz/395DV8zO4JwJqfzy7f2crO5cy1qI7oRNMLZjk1KYQojQEJUEi74FB96Ao+sAozzmf18xA5eGe175VNYdiz4Jj+jW2kSTdFMLIULJom9Cci68dju0NAKQlRzNHedP4P0DZewsqglyA0U4CY/o5miiSUdIZiyECB3WKPjcI1BV2G4y1xcWjCHaZpaxY9En4RHdWpto0lYZMxZChJacc2DOjUYhkOLtAMRFWlkxK4PXdhZT3dgS5AaKcNGrYKyUulApdUApdVgpdZef87cqpT5VSu1QSn2glJoa0Fa2NtGopZtaCBGCPvNTiEmHf30LWo2JW188awx2h4uXthYFuXEiXPQY3ZRSZuBR4CJgKrDST7D9u9Z6htZ6FvBL4FcBbWVrE40uq3RTCyFCT1QiXPYIlOyGl24Cp4NpoxOYPSaRv28+LhO5RK/0JrrNBw5rrY9qrVuA54EVvhdorWt9vo0BAvq/T7c20agjJDMWQoSmiZ+FSx6GA28aE7q05ktnjeVoeQMbj1QEu3UiDPQmumUAJ3y+L3Ifa0cp9Q2l1BGMzPj2wDTPzdFEMzJmLIQIYfO+Csvuhh3Pwpr7uHjGKJKirTKRS/RKwFJNrfWjWutxwA+Be/xdo5S6RSm1RSm1paysrPc3b22iCZlNLYQIcUt/CPlfgQ9/S+TB17gmP4t/7y2htK452C0TIa430e0kkOXzfab7WFeeBy73d0Jr/bjWOl9rnZ+Wlta7FrpcKEczzbLOWAgR6pSCi34BGfnw2u18YZIJp0vzyrbu3jKF6F0w/gSYoJTKUUrZgOuA1b4XKKUm+Hx7CXAoYC10GJ8om7VU4BJChAGzFa76E7hcZK/7NvPHxPPClhMykUt0q8foprV2AN8E3gH2AS9qrfcopX6qlLrMfdk3lVJ7lFI7gO8ANwashe6lAk3IBC4hRJhIzoFLfwXHN3FfwhscLWtg67GqYLdKhDBLby7SWr8JvNnh2L0+X387wO1q4zCCcTM2mcAlhAgfMz8Ph99l2q7f83VbPS9uySQ/OznYrRIhKvRTTU9mrG1EWEO/uUII4fW538LUFfzA9FemffoQ9c1GRa7mVidldfYgN06EktCPbq0+mbE59JsrhBBe1ki4+mlKpn6ZG9WbVD91LW98sJVl/7OW83+1juZWZ7BbKEJE6Ec332AsmbEQItyYTKRf/Ssei7iJtNPrOfc/F3Kr8zkcTbVsLqgMdutEiAj96NZqbE1m7NokY8ZCiPCjTCbSL/guX459jPLM81jl/AcfRXyT5LduhU9fgmbZbnG469UErqDyLG2SdcZCiDB29dxMrp57LXAtnNzKtr//kpnVm+CfayA6FVa9AemTg91MESShH908mTGyzlgIMURkzOXIwoeY2/QopVe9DCYz/PVyY29kMSyFfnRr9WTGss5YCDF0LJ2UhgsT79SPgy+9YsyPeeZyqDsd7KaJrjRWwqH/wAAUcAn96ObOjI0KXDJmLIQYGnJTY8hKjmLtgTIYMQ2++E+oL4U/XwDb/0ZZdR3v7S/heEUjLlcv3vwLNsDxzQPf8L5qbYJ/fBkOrQl2S86MvQ6eWQHPXg0fPxHw24f+mLG3Apd0Uwshhg6lFMsmpvPS1iLsDicRmfnwpZfhze/Bv76ByfxffGo/m9dcIyg3p3PO+BRumdgAp3cZNxh3How/z+jaXnM/FKwDazTc+gGkjAvmS2tvw69gz8tG+77xMcSkBrtFfedshX+sgpI9MHoOvH0XpI6HcecG7ClCPxj7TOCSYCyEGEqWTUrjrx8d45OCKs6ekApjzoKvbaBuz9scefF+vm15ue3iAvef2JHgaoWdz4EygXZBVDKc+xP48BF49Tb48lvGODRAyV7jfTRjzsC8iJqTEJ0M1qjO5yqOwIe/gexz4MRm44PGNU8PTDt6w2GHzX+A3S9D5jyYdKHxd9lBKN4GVccgJRfSp0LyOLBEgMkC79wNh9fA5x6B6VfCnz8LL66Cm9+F1Ak9Pm1vhH4wbm3EoSzER0dikaIfQoghZOG4FGxmE2sPlBrBGEApXq2bwk9a7uWtr+czJbqWgwf38sDre/jS5Z/jM/NngMsFxdvh0L/BYoN5N0NkPCRkwSu3wMZHYPEd8PHj8M6PjeA9drFxbMJnjN2lelJxBLY+bQT7jLnGn5hUcDnB2QIH3oLtfzWCbOwIWHYXzP6SsVEGGOOqb34PLJHGxhnb/wbv/QymXQFTVwzQT7QDZ6vRvWyvNX5ea+43ehJGzjDa80mH7mZzBDi7qIy25Psw173twsrn4Ilz4U/nQdwoMFnBbIGvvgem/sWpMAjGzbSqCFJjI4LdEiGECKhom4X5Ocm8t7+UH108BZPJCJIvbz/J5JFxTBkzAhhBbtI4tv87ktEnNJ+Zj/GGnznX+ONr5udh/+vw/n/DsU1w6B2YeBFknw0fPQZ/vwasMRCbBjHpEJUEEbFgi4XIBIhOMf4+9G8j2JosRvbt/J3/F5A6EZbfY2SNr98JG38Hs79oZPg1RXDkPbjwIYgbCYu/DftWwxvfNTJUZytoJ+QsgaRs435OB3z8R9jwv5CQCbnLYcxCqD4OJ7dA+UEjmM+/pXMmrjWUH4LC9VC01ch0yw4APuPtaVPgiy8b3futTVCwHk7tMpaUZcw1AmvtSSjdZwRtZyu4HMaHjZmfb7tP0lhjjP+jx4z7uBzGn34GYgiLYNxIMzbS4iQYCyGGnqvnZnLHCzt49uPjfOmssRSWN7D9eDV3XdS25thiNrFoXAobDpWjtUZ1ldkqBZf+Gh7bZATIz/wUFt1uHJ9/C+x5BU7tMCaKNZRC3SmoaICWemiqbssKo1Ngyfdg3leNLvDSPXBym5FlmsxGgM7Ih6z5xr2XfA8Ovg1rfw7v/ldbe0bMMLJ2MDLmFY/Bn86Hl2/2bTRMugimXg4fPQqndkLOUiO4bfqd0c0NRvd8/Cj4z72w+Y9wzneMMfKak1BxGAo3GIEUjHXbGXNhyueMryPiICbNGOM1u8OeNQomftb44ysh0/jTk9Gz4MrHe76ul0I/GDuaadQ2yYyFEEPSilmj+cfWE/zirf2cPyWdV7afRCnjuK9zJqTxzp4SCsobyE2L7fqGManGmLHDDiOntx232CDvWuOPP1pDSwM0VRpZszWy7dzo2cafrih3QJ10ETRUGF3Xxdtg+lVtwQ+M9ty521giZLEZmeeuF2DLU3DgTSMDveZpIzArZQT/UzuNzDk+wzhWsAHW3Gdk2B6xI4wMOvf7kLsUknJ61xUfQkI/GLc20uiySjAWQgxJSil+fsVMLvjNOn7y6m4OldazMDeFUQntu2GXTEgDYMOh8u6DMfRvUpFSRpd1RA/37klMCky+2Pjj93xq+xnV595jjMce/whG5UFUYtu5iDiji91Xzjnw1XeNIB0RZwRp3w8OYSrkZ0Q57Y00aBupcbZgN0UIIQbEmJRovvuZSazZV8qxikaumJ3h95qxKdFsOFQWkOc8Wd0UOts4WiKMjNY3EHdHKaObOGXckAjEEAbBuNXeiB3pphZCDG1fXpzNjIwEIq0mLpw+0u8150xIZdORClocrjN+vhv+vJm7X/70jO8jAiPkg7HT3kiTjpAJXEKIIc1iNvHnG/N5/paFxEVa/V5zzoQ0GlqcbD9edUbPVVjewJGyBvadqj2j+4jACflg7GppogkbaZIZCyGGuPT4SGZlJXZ5fuG4FMwmxYZD5Wf0PGsPlAJGV3VTi/OM7iUCI+SDsWdpk3RTCyGGu/hIK7OzEnnfHUz7a+3BtnHno+X1Z9osEQAhH4xNjiaatI2UWJnAJYQQl84cxZ7iWnacqO7X45tbnWw6UsGicSkAHC6VYBwKQj8YO+1gicQqpTCFEIKr5mYSG2HhLxsL+/X4j45WYHe4WLUoG5OCI2UNgW2g6JeQj3BWVzPKFh3sZgghREiIi7Ry9dxMXt9VTGldc58fv/ZAGREWE0smppGVHM2Rsr5nxk9/WMDWY2c2iUy0F9rB2OnAggNLhARjIYTwuGHhWFqdmuc2n+jzY9cdLGPhuBQirWbGpcVypI/d1A6niwfe2MeTHxT0+blF10I7GDuMvYwtkTFBbogQQoSO3LRYlk5M42+bj/VpzfGxigYKyhtYNtGo5jU+PZaj5Q04XbqHR7Y5VdOMw6XZK8uiAiq0g3GrEYxtEoyFEKKdVYuzKauz89buU71+zNoDxizqZZPSARiXFkOLw8XJqqZe3+N4ZSMABeUN1NsdfWix6E5IB+PmRmNiQVT0GdZKFUKIIWbphDRyUmP4338fZG9x+yzV4XThcHbOmNceKCU7JZrsVCPBGZ9uvLceLqvzXtPY4mDrsUr+teMkj609zP7T7e99rKLR+/V+yY4DJqSDcVVNNQBRMRKMhRDCl8mk+OXVM2lqdXL5ox/ypw1HOVpWz4Nv7CX/wTXc+ret7a53uTRbjlWx0L2kCSA31XhvPVLa4L3mqt9v4qrfb+Lbz+/gl28f4Df/OdTuPscrG70bIu0plmAcKCEdjKvrjE9rMTFxQW6JEEKEnnnZybz97XNYMjGNB97Yx7n/u46nPiwkxmZhw6Fy7I626lpHy+upa3Ywe0yS91hSjI2UGJt3RvVHRyvYd6qW73xmImu+s4RzJ6dzsLSu3XMer2wgOyWG5Bhbp4xc9F9Ib6FYW2v8Q8fGSjAWQgh/UmIjeOKGuby64ySna+xcOSeD7cerufVvW9l9spa5Y43gu+14NQBzxiS2e/y49Fhv4Y8XtpwgLtLCLUtyibSamToqnnUHy7A7nERYzICRGY9JjsalNXtO1Qza6xzqQjozrq83PpHFx8cHuSVCCBG6lFJcMTuT25aNY0R8pDcAbz1W6b1m+/Fq4iMt3q5pj3FpsRwpq6emsZW3dp/m8lkZRFqNwDthRCxOl6awvG2c+HiFEYynjorn4Ol6Wv2MTfdGc6uTG5/8WNYru4V0MG5sMD6tJcRJMBZCiN5Ki4sgJzWGTwrbAt3241XMGpOEyaTaXTsuLYaqxlae2lhAi8PFtfOyvOcmpBu9kgdLjMSourGF2mYHY1OimTo6nhanq9/lNLceq2LdwTLW7Cvp1+OHmpAOxs2Nxn8AqyxtEkKIPskfm8TWY1Voram3OzhQUsdsPztCeWZUP77+KFNHxTM9I8F7LjctBpOCQ+6A61nWlJUczbTRRpLU33HjDw8bO08d7UcFsKEotINxk7tmqiUyuA0RQogwk5+dRGVDC0fLG9h1ohqtYXaH8WIwuqkBGluc7bJigEirmTHJ0Rx2T+LyLGsakxxNTmoskVZTv2dUbzxSAcBRqY0NhHgwbmlyf2KySjlMIYToi/zsZAC2FFay3b3D0+yspE7XZSRGEWk1YbOYuHxWRqfzE0bEcbCkfWY8Jjkas0kxeWQ8e/sxiau2uZVdRdXYLCYKKxr8ronuqLzeztL/eZ/dJ4fmpLGQDsatze5JA9ao4DZECCHCTG6qsfzok8Iqth2rYlxaDAnR1k7XmUyKReNS+Xx+pt/zE9JjKSxvoMXh4nhFI6mxNmIijIU4U0fHs7e4Fq17X04TYPPRSlza2A6y1akp6kUFsAOn6zhW0cjGI+V9eq5wEdLB2NXiDsbSTS2EEH2ilGLu2CRvZuy7vrijJ1fN42crpvs9N3FEHA6X5lhFg3dZk8e00fHUNjt6FUx9bTxSToTFxNVzMwFjDXRPTtUYO1QN1f2XQzYYN7c6MTmbcZgiwBSyzRRCiJA1LzuJwopGKhtamNNNMAYjePvjmeB1sKS+UzCeOsqYxNXXceONhyuYl53MlJHG4z0VwLpzusYI+BKMB1lZnZ1IWnCaI4LdFCGECEtzxyZ7v/Y3eas3xqXFohTsPVVDcU0TY1LaVrdMHhmPScEr24u63T1q3cEy797LZXV2DpTUsWh8CkkxNpJjbH3KjA+V1ve5WzwchGwwLq+3E0UL2iLjxUII0R/TM+KJsJiIsZmZOKJ/lQyjbMaM6rUHytCadplxlM3MnedP5J09Jdz45MdUN7Z0evwnhZXc+OTHXPOHTZyuaWbTUWMW9aJxqYAxtn2kFzOqT7uDcV2zg7I6e79ey5naeLicL/15c5+2reytkA3GZXV2IlULSiZvCSFEv0RYzCwen8qi8amYTf67oXtjQnqstyt6bEr71S3fOm8Cv/p8HluPVXHFYxspLG8LrFprfvHWflJibFTUt7DyiY94bWcxcZEWprvXKeemxfRqrfGpmmZibEZlMN+u6ga7g/9+cx8V9QMfoFfvLGbDoXL2DcBuVSEbjMvrW4jCjskmy5qEEKK/Hrt+Dv+3cvYZ3WN8eltW7ZsZe1w5J5O/37yA6sYWvvz0J9Q0tgLw/oFSthyr4s7PTOTpL8+jpLaZ/+wtYUFOChazEX7GpcVSXt9CTVNrt204XdvMWbnGjlOHfYL3mn0lPL7+KP/12t4zeo29sd1d33v78cCX8AzZYOxwuYg1t2KOkOpbQgSKUupCpdQBpdRhpdRdfs6PUUq9r5TarpTapZS6OBjtFIETaTV7a03318QRxiSuCIuJ9Dj/83jys5N54oZ8iqoa+eZz22hxuPjl2wcYmxLNtfOyyM9O5qlV84iPtPC5vFHex+W6i450lx03tzqpbGhhVlYisRGWdpnxxwVG/e3VO4tZf7DsjF5nd2qbW707WHnWbQdSyAbjGxZms2hMDCarLGsSIhCUUmbgUeAiYCqwUik1tcNl9wAvaq1nA9cBjw1uK0Uo8tSoHpMc3eWsazAC8oOXz2DDoXKu+cNG9p+u47sXTMLqzoIX5Kaw/d4LWOFTXCQ3zUi4uhs3Lqk1xotHJUa122UKYHNBJYvGpZCbGsNP/rWb5lZnV7c5I7tO1KA1JMfYvBlyIIX0Foq0NkLsyGC3IuBaW1spKiqiubk52E0RISQyMpLMzEys1s6FFwJkPnBYa30UQCn1PLAC8O3f04BnZ5YEoHigGiPCx7h0I2B2HC/25/Pzsth3upanPixk6qh4Lp0xqt35jmPXY5KjsZhUt5mxZyb1qIRIxqfFsv6QkQGX19s5XFrPVXMyyctM4At/2syj7x/muxdMAowx6+4+PPTF9uNVKAVfmD+G371/mIp6OymxgVvtE+LBuBmGYGZcVFREXFwc2dnZAfuPIsKb1pqKigqKiorIyckZqKfJAE74fF8ELOhwzf3Av5VS3wJigPMHqjEifETbLHx22gjOmZDWq+t/fPEUEqNsXDBtRKddojqymk2MSYnutka1JzMemRDJ+PRY/rmtiJqmVj5xd1EvyE1mzpgkrpidwWNrj/Dcxyeot7eiULz2rcXtxrz7a/uJasanxbJkYhq/e/8wO05Uc96UEWd8X48QD8ZNQ7IudXNzswRi0Y5SipSUFMrKBm7Mq5dWAk9rrf9XKbUQ+KtSarrWutNaDqXULcAtAGPGjBnkZorB9scv5ff6WovZxLfPn9Dr63NTjT2Vu+LJjEfGRzLBXYTkcGk9mwsqibKameHeaereS6eSEGXF7nARYzPz5IcF/GtHsTdTBmP8+e3dp1k0PoX0OCPZc7o0/9lbwjt7TlPb1Eq93UFcpIVHVs4m2mZBa83241V8ZuoIZmQkYDYpth8fTsHY0TRk61JLIBYdDcL/iZOA77Y8me5jvr4CXAigtd6klIoEUoHSjjfTWj8OPA6Qn58/9KowiEEzLi2G9QfLcLo0DpeLd/eVcu7kdO/Es9M1zcRFWoiJsHgrgh1xB+O5Y5O8Y9JJMTbuv2ya9757imt5a/fpdsH4LxsL+flb+zGbFOdOTmfu2CRe+OQEBeUNpMZGMCI+ggiLiTX7SvnHliJuXJTNsYpGqhpbmT0miSibmSmj4th+IrAzqkN2AhdgZMZS9CPgKioqmDVrFrNmzWLkyJFkZGR4v29p6bxo39eWLVu4/fbbe3yORYsWBaq5ANxxxx1kZGTgcgV+sf0w8gkwQSmVo5SyYUzQWt3hmuPAeQBKqSlAJBD0dF0MbePSYmlxunh9VzGf+78P+Pqz23j+4+Pe86dqmhiVYGSxWcnR2Cwmth6rYv/pWhbkJHd1Wy6aMZLDpfXeLSC11rzwyQmmZ8Tz1XNy2H68mofe2k98pIVHvzCHzT86jzduP4eXv76YuWOTeGLDURxOlzfweqqYzc5KYueJGpyuwH0GDd1grLUxgWuIZsbBlJKSwo4dO9ixYwe33nord955p/d7m82Gw+Ho8rH5+fk88sgjPT7Hxo0bA9Zel8vFK6+8QlZWFuvWrQvYfTvq7nUPBVprB/BN4B1gH8as6T1KqZ8qpS5zX/Zd4Gal1E7gOWCVHoq1B0VI8cyo/vbzO6hrdpAYbWWbz4zl0zXNjEwwYoHZpMhNjeG1XcVoDfO7CcafnWZMAH7r09OAsQzqaHkDNy7M5u6LprDp7nNZ9/1lvPqNxVwyc1S7yWW3LMmlqKqJt3afZvvxamJsZu+s8llZidTbHQGtkx26wdjZCto1JCdwhaJVq1Zx6623smDBAn7wgx/w8ccfs3DhQmbPns2iRYs4cOAAAGvXruXSSy8F4P777+emm25i2bJl5ObmtgvSsbGx3uuXLVvG1VdfzeTJk7n++uu9dWXffPNNJk+ezNy5c7n99tu99+1o7dq1TJs2jdtuu43nnnvOe7ykpIQrrriCvLw88vLyvB8AnnnmGWbOnEleXh5f+tKXvK/vpZde8tu+c845h8suu4ypU41VPpdffjlz585l2rRpPP74497HvP3228yZM4e8vDzOO+88XC4XEyZM8I7zulwuxo8fHwrjvl3SWr+ptZ6otR6ntX7QfexerfVq99d7tdaLtdZ5WutZWut/B7fFYjiYNDKO3NQYvrBgDP++cwkLc1PY4bOW91RNM6Pi22LBuPRYGluc2Cwm8rISu7zviPhI5o5N4q3dRjB+/pMTxEVYuGSmMcPbajYxNiXG7xDRZ6aMIDc1hj+uP8K241XkZSV6g7UnQw5k8Y/QHTNu9exlPPQmcPn6r9f2sLePO570ZOroeO773LSeL+ygqKiIjRs3Yjabqa2tZcOGDVgsFtasWcOPfvQj/vnPf3Z6zP79+3n//fepq6tj0qRJ3HbbbZ2W5mzfvp09e/YwevRoFi9ezIcffkh+fj5f+9rXWL9+PTk5OaxcubLLdj333HOsXLmSFStW8KMf/YjW1lasViu33347S5cu5ZVXXsHpdFJfX8+ePXt44IEH2LhxI6mpqVRWVvb4urdt28bu3bu9s5iffPJJkpOTaWpqYt68eVx11VW4XC5uvvlmb3srKysxmUx88Ytf5Nlnn+WOO+5gzZo15OXlkZbWuxmnQghDXKSV9763zPv9rKxE3tp9mop6O/FRVsrq7YxMaAvGnklcs7MSeyxoctH0kTzwxj4+LarhzU9PcU1+JtG2nkOfyaS4ZUkud738KQDfXD7eey4nNYaEKCvbj1dz3fzATF4M3czY4V6DK3sZD5prrrkGs9n4j11TU8M111zD9OnTufPOO9mzZ4/fx1xyySVERESQmppKeno6JSUlna6ZP38+mZmZmEwmZs2aRWFhIfv37yc3N9cbALsKxi0tLbz55ptcfvnlxMfHs2DBAt555x0A3nvvPW677TYAzGYzCQkJvPfee1xzzTWkphpF6JOTu+7C8m2f73KiRx55hLy8PM466yxOnDjBoUOH+Oijj1iyZIn3Os99b7rpJp555hnACOJf/vKXe3w+IUT3PHsv7zhRTWmdHa3xjhlD27aO3Y0Xe1w43eiqvuOF7dgdLq6b1/vgefnsDFLda4l9d71SSjF7TGK77P1MSWYcZP3JYAdKTExb6dGf/OQnLF++nFdeeYXCwkKWLVvm9zEREW2L3s1ms99x195c05V33nmH6upqZsyYAUBjYyNRUVFddml3xWKxeCd/uVyudhPVfF/32rVrWbNmDZs2bSI6Opply5Z1W5wlKyuLESNG8N577/Hxxx/z7LPP9qldQojOfJcPJUYbPW2+mfGcMUmMiI/ggmk9F4XKTIpmZmYCu4pqmJ4Rz3T3MqjeiLSauWVJDv/774PeDwge+WOT+N+DB9l6rLLdVpX9FbqZcauxkbRM4AqOmpoaMjKMknVPP/10wO8/adIkjh49SmFhIQAvvPCC3+uee+45/vSnP1FYWEhhYSEFBQX85z//obGxkfPOO4/f//73ADidTmpqajj33HP5xz/+QUWFsU2bp5s6OzubrVu3ArB69WpaW/0Xpa+pqSEpKYno6Gj279/PRx99BMBZZ53F+vXrKSgoaHdfgK9+9at88YtfbNezIITovyibmckjjeVDbdW32mLB6MQoNv/o/F4HVk92fG0fsmKPm8/J5cO7ziU5xtbu+A2LsslMiuL253Z4N8Y4EyEcjN3ZiATjoPjBD37A3XffzezZswdklnFUVBSPPfYYF154IXPnziUuLo6EhPa/WI2Njbz99ttccskl3mMxMTGcffbZvPbaa/z2t7/l/fffZ8aMGcydO5e9e/cybdo0fvzjH7N06VLy8vL4zne+A8DNN9/MunXryMvLY9OmTe2yYV8XXnghDoeDKVOmcNddd3HWWWcBkJaWxuOPP86VV15JXl4e1157rfcxl112GfX19dJFLUQAzR6TyM4TNRRXG4mZb2bcVyvnjeFrS3O5ak5Gzxd3oJTydlX7io+08n8r51BS28xdL+/iTBcdqGCtWsjPz9dbtmzp+oKCDfCXS+HG1yBnyeA1bBDs27ePKVOmBLsZQVdfX09sbCxaa77xjW8wYcIE7rzzzmA3q8+2bNnCnXfeyYYNG874Xv7+byiltmqte1/+KAh6/H0Woo9e2lrE9/6xkyUT09hSWMme//psSBZL+uO6I/z8rf08cPl0vnjW2G6v7e53OXQzY+8ELsmMh6onnniCWbNmMW3aNGpqavja174W7Cb12UMPPcRVV13Fz3/+82A3RYghxTNhauPhckYmRIZkIAajG3vJxDR+9vpeSmv7v/lPGEzgkmA8VN15551hmQn7uuuuu7jrrk7bAgshzlBOSgzxkRZqmx3tZlKHGpNJ8avP57HjeDXp8f1vZ+hmxjKBSwghhi2TSTHLPYN5xBkEucGQGhvB+VPPbNMICcZCCCFC0ix3da1QzowDRYKxEEKIkOQZNx6ZMPTjQOgG4+gUGD1nyBf9EEII4d+CnGQ+lzeapROGfonZ0A3GedfCLe+DpfP6LnFmli9f7i0p6fGb3/zGW1rSn2XLluFZunLxxRdTXV3d6Zr777+fhx9+uNvnfvXVV9m7d6/3+3vvvZc1a9b0ofXdk60WhRg6om0W/m/lbMakDP2kLHSDsRgwK1eu5Pnnn2937Pnnn+92swZfb775JomJif167o7B+Kc//Snnn39+v+7VkWy1KIQIVxKMh6Grr76aN954w1ufubCwkOLiYs455xxuu+028vPzmTZtGvfdd5/fx2dnZ1NeXg7Agw8+yMSJEzn77LO92yyCsYZ43rx55OXlcdVVV9HY2MjGjRtZvXo13//+95k1axZHjhxpt7Xhu+++y+zZs5kxYwY33XQTdrvd+3z33Xcfc+bMYcaMGezfv99vu2SrRSFEuArddcbDxVt3welPA3vPkTPgooe6PJ2cnMz8+fN56623WLFiBc8//zyf//znUUrx4IMPkpycjNPp5LzzzmPXrl3MnDnT7322bt3K888/z44dO3A4HMyZM4e5c+cCcOWVV3LzzTcDcM899/DnP/+Zb33rW1x22WVceumlXH311e3u1dzczKpVq3j33XeZOHEiN9xwA7///e+54447AEhNTWXbtm089thjPPzww/zpT3/q1B7ZalEIEa4kMx6mfLuqfbuoX3zxRebMmcPs2bPZs2dPuy7ljjZs2MAVV1xBdHQ08fHxXHbZZd5zu3fv5pxzzmHGjBk8++yzXW7B6HHgwAFycnKYOHEiADfeeCPr16/3nr/yyisBmDt3rndzCV+y1aIQIpxJZhxs3WSwA2nFihXceeedbNu2jcbGRubOnUtBQQEPP/wwn3zyCUlJSaxatarb7QO7s2rVKl599VXy8vJ4+umnWbt27Rm117MNY1dbMMpWi0KIcCaZ8TAVGxvL8uXLuemmm7xZcW1tLTExMSQkJFBSUsJbb73V7T2WLFnCq6++SlNTE3V1dbz22mvec3V1dYwaNYrW1tZ2gScuLo66urpO95o0aRKFhYUcPnwYgL/+9a8sXbq0169HtloUQoQzCcbD2MqVK9m5c6c3GOfl5TF79mwmT57MF77wBRYvXtzt4+fMmcO1115LXl4eF110EfPmzfOe+9nPfsaCBQtYvHgxkydP9h6/7rrr+J//+R9mz57NkSNHvMcjIyN56qmnuOaaa5gxYwYmk4lbb721V69DtloUQoS70N1CcQiTLRSHp95stShbKAoxdHX3uyxjxkIMgoceeojf//73MlYshPBLuqmFGAR33XUXx44d4+yzzw52U4QQIUiCsRBCCBFkEoyDJFhj9SJ0yf8JIYYvCcZBEBkZSUVFhbz5Ci+tNRUVFURGDv19W4UQnckEriDIzMykqKhIahOLdiIjI8nMzAx2M4QQQSDBOAisVmu7sopCCCGGN+mmFkIIIYJMgrEQQggRZBKMhRBCiCALWjlMpVQZcKyHy1KB8kFoTqgYTq93OL1WOLPXO1ZrHdKbH8vvcyfD6bXC8Hq9A/K7HLRg3BtKqS2hXpM3kIbT6x1OrxWG3+v1Zzj9DIbTa4Xh9XoH6rVKN7UQQggRZBKMhRBCiCAL9WD8eLAbMMiG0+sdTq8Vht/r9Wc4/QyG02uF4fV6B+S1hvSYsRBCCDEchHpmLIQQQgx5IRuMlVIXKqUOKKUOK6XuCnZ7AkkplaWUel8ptVcptUcp9W338WSl1H+UUofcfycFu62BopQyK6W2K6Ved3+fo5Ta7P73fUEpZQt2GwNFKZWolHpJKbVfKbVPKbVwKP/b9mQo/y6D/D67v5ff5zMUksFYKWUGHgUuAqYCK5VSU4PbqoByAN/VWk8FzgK+4X59dwHvaq0nAO+6vx8qvg3s8/n+F8CvtdbjgSrgK0Fp1cD4LfC21noykIfxuofyv22XhsHvMsjvM8jv85n/22qtQ+4PsBB4x+f7u4G7g92uAXy9/wI+AxwARrmPjQIOBLttAXp9me7/sOcCrwMKY9G8xd+/dzj/ARKAAtzzMXyOD8l/2178PIbV77L7Ncrvs/w+9/lPSGbGQAZwwuf7IvexIUcplQ3MBjYDI7TWp9ynTgMjgtWuAPsN8APA5f4+BajWWjvc3w+lf98coAx4yt2N9yelVAxD99+2J8Pmdxnk99n9/VD6Nx603+dQDcbDglIqFvgncIfWutb3nDY+coX9VHel1KVAqdZ6a7DbMkgswBzg91rr2UADHbqwhsq/rWhPfp+HpEH7fQ7VYHwSyPL5PtN9bMhQSlkxfnGf1Vq/7D5copQa5T4/CigNVvsCaDFwmVKqEHgeo2vrt0CiUsqzn/ZQ+vctAoq01pv/v737VYkgCsMw/nwIewFmERHE6gUYBJsYRQSFxXuwaLd6BeoFiGkvQIPRYBC0WSwmu+kznAFNsuDIGc8+vzT7J5zh4+Ud5iw73etrSphbnO00ms8ymGfM869nO9QyvgdWul/ojYA9YFJ5Tb2JiAAugOfMPPv20QQYd8djyt7Tv5aZx5m5kJlLlDneZOY+cAvsdF9r4lwBMvMNeI2I1e6tTeCJBmc7paazDObZPPdzvoP904+I2KLsTcwBl5l5WndF/YmIdeAOeORr3+WEss90BSxSnoCzm5nvVRb5ByJiAzjKzO2IWKZcWc8DD8BBZn5UXF5vImINOAdGwAtwSLnwbXa2P2k5y2CezXM/sx1sGUuSNCuGeptakqSZYRlLklSZZSxJUmWWsSRJlVnGkiRVZhlLklSZZSxJUmWWsSRJlX0ClNzT116y9DcAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 576x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "acc = history.history['accuracy']\n",
    "val_acc = history.history['val_accuracy']\n",
    "\n",
    "loss = history.history['loss']\n",
    "val_loss = history.history['val_loss']\n",
    "\n",
    "epochs_range = range(best_epoch)\n",
    "\n",
    "plt.figure(figsize=(8, 8))\n",
    "plt.subplot(1, 2, 1)\n",
    "plt.plot(epochs_range, acc, label='Training Accuracy')\n",
    "plt.plot(epochs_range, val_acc, label='Validation Accuracy')\n",
    "plt.legend(loc='lower right')\n",
    "plt.title('Training and Validation Accuracy')\n",
    "\n",
    "plt.subplot(1, 2, 2)\n",
    "plt.plot(epochs_range, loss, label='Training Loss')\n",
    "plt.plot(epochs_range, val_loss, label='Validation Loss')\n",
    "plt.legend(loc='upper right')\n",
    "plt.title('Training and Validation Loss')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy :  0.5845410628019324\n",
      "Mean Square Error :  1.1111111111111112\n",
      "[0 2 0 2 0]\n",
      "Confusion Matrix for each label : \n",
      "[[[ 46  56]\n",
      "  [ 23  82]]\n",
      "\n",
      " [[149  19]\n",
      "  [ 19  20]]\n",
      "\n",
      " [[133  11]\n",
      "  [ 44  19]]]\n",
      "Classification Report : \n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.59      0.78      0.67       105\n",
      "           1       0.51      0.51      0.51        39\n",
      "           2       0.63      0.30      0.41        63\n",
      "\n",
      "    accuracy                           0.58       207\n",
      "   macro avg       0.58      0.53      0.53       207\n",
      "weighted avg       0.59      0.58      0.56       207\n",
      "\n"
     ]
    }
   ],
   "source": [
    "prediction = hypermodel.predict(X_test)\n",
    "prediction  = np.argmax(prediction, axis=1)\n",
    "y_val=np.argmax(y_test, axis=1)\n",
    "\n",
    "print(\"Accuracy : \", accuracy_score(y_val, prediction))\n",
    "print(\"Mean Square Error : \", mean_squared_error(y_val, prediction))\n",
    "\n",
    "print(prediction[:5])\n",
    "print(\"Confusion Matrix for each label : \")\n",
    "conf_matrix = multilabel_confusion_matrix(y_val, prediction)\n",
    "print(conf_matrix)\n",
    "\n",
    "print(\"Classification Report : \")\n",
    "print(classification_report(y_val, prediction))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 864x504 with 3 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "vis_arr = np.array(conf_matrix)\n",
    "labels = [\"\".join(\"c\" + str(i)) for i in range(0, 3)]\n",
    "fig, ax = plt.subplots(1, 3, figsize=(12, 7))\n",
    "for axes, cfs_matrix, label in zip(ax.flatten(), vis_arr, labels):\n",
    "    print_confusion_matrix(cfs_matrix, axes, label, [\"N\", \"Y\"])\n",
    "\n",
    "fig.tight_layout()\n",
    "plt.show()\n"
   ]
  }
 ],
 "metadata": {
  "interpreter": {
   "hash": "aee8b7b246df8f9039afb4144a1f6fd8d2ca17a180786b69acc140d282b71a49"
  },
  "kernelspec": {
   "display_name": "Python 3.10.3 64-bit",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.3"
  },
  "orig_nbformat": 4
 },
 "nbformat": 4,
 "nbformat_minor": 2
}