mriClassification-checkpoint.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['train', 'val']\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "WARNING: Logging before flag parsing goes to stderr.\n",
      "W1013 00:08:21.382179  9380 deprecation_wrapper.py:119] From C:\\Users\\SplendidUnivers\\Miniconda3\\envs\\test\\lib\\site-packages\\keras\\backend\\tensorflow_backend.py:4070: The name tf.nn.max_pool is deprecated. Please use tf.nn.max_pool2d instead.\n",
      "\n",
      "W1013 00:08:21.975985  9380 deprecation.py:323] From C:\\Users\\SplendidUnivers\\Miniconda3\\envs\\test\\lib\\site-packages\\tensorflow\\python\\ops\\nn_impl.py:180: add_dispatch_support.<locals>.wrapper (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.where in 2.0, which has the same broadcast rule as np.where\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 243 images belonging to 2 classes.\n",
      "Found 243 images belonging to 2 classes.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "W1013 00:08:24.589663  9380 deprecation_wrapper.py:119] From C:\\Users\\SplendidUnivers\\Miniconda3\\envs\\test\\lib\\site-packages\\keras\\backend\\tensorflow_backend.py:422: The name tf.global_variables is deprecated. Please use tf.compat.v1.global_variables instead.\n",
      "\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/30\n",
      "100/100 [==============================] - 30s 303ms/step - loss: 0.5363 - accuracy: 0.7453 - val_loss: 0.4045 - val_accuracy: 0.8160\n",
      "Epoch 2/30\n",
      "100/100 [==============================] - 22s 219ms/step - loss: 0.3482 - accuracy: 0.8610 - val_loss: 0.3605 - val_accuracy: 0.8134\n",
      "Epoch 3/30\n",
      "100/100 [==============================] - 22s 216ms/step - loss: 0.2766 - accuracy: 0.8830 - val_loss: 0.3173 - val_accuracy: 0.8647\n",
      "Epoch 4/30\n",
      "100/100 [==============================] - 22s 218ms/step - loss: 0.2462 - accuracy: 0.8990 - val_loss: 0.1846 - val_accuracy: 0.9192\n",
      "Epoch 5/30\n",
      "100/100 [==============================] - 23s 229ms/step - loss: 0.2260 - accuracy: 0.8994 - val_loss: 0.0548 - val_accuracy: 0.9185\n",
      "Epoch 6/30\n",
      "100/100 [==============================] - 23s 227ms/step - loss: 0.2075 - accuracy: 0.9110 - val_loss: 0.0729 - val_accuracy: 0.9310\n",
      "Epoch 7/30\n",
      "100/100 [==============================] - 23s 230ms/step - loss: 0.1995 - accuracy: 0.9142 - val_loss: 0.0489 - val_accuracy: 0.9192\n",
      "Epoch 8/30\n",
      "100/100 [==============================] - 21s 214ms/step - loss: 0.1868 - accuracy: 0.9195 - val_loss: 0.0854 - val_accuracy: 0.9026\n",
      "Epoch 9/30\n",
      "100/100 [==============================] - 22s 217ms/step - loss: 0.1889 - accuracy: 0.9142 - val_loss: 0.3024 - val_accuracy: 0.9330\n",
      "Epoch 10/30\n",
      "100/100 [==============================] - 22s 219ms/step - loss: 0.1740 - accuracy: 0.9212 - val_loss: 0.1157 - val_accuracy: 0.9336\n",
      "Epoch 11/30\n",
      "100/100 [==============================] - 22s 222ms/step - loss: 0.1761 - accuracy: 0.9198 - val_loss: 0.1950 - val_accuracy: 0.9363\n",
      "Epoch 12/30\n",
      "100/100 [==============================] - 22s 225ms/step - loss: 0.1775 - accuracy: 0.9205 - val_loss: 0.0609 - val_accuracy: 0.9311\n",
      "Epoch 13/30\n",
      "100/100 [==============================] - 23s 228ms/step - loss: 0.1722 - accuracy: 0.9215 - val_loss: 0.0888 - val_accuracy: 0.9304\n",
      "Epoch 14/30\n",
      "100/100 [==============================] - 23s 233ms/step - loss: 0.1575 - accuracy: 0.9225 - val_loss: 0.1497 - val_accuracy: 0.9330\n",
      "Epoch 15/30\n",
      "100/100 [==============================] - 23s 227ms/step - loss: 0.1726 - accuracy: 0.9175 - val_loss: 0.2005 - val_accuracy: 0.9284\n",
      "Epoch 16/30\n",
      "100/100 [==============================] - 23s 235ms/step - loss: 0.1604 - accuracy: 0.9241 - val_loss: 0.0200 - val_accuracy: 0.9357\n",
      "Epoch 17/30\n",
      "100/100 [==============================] - 22s 223ms/step - loss: 0.1612 - accuracy: 0.9274 - val_loss: 0.1445 - val_accuracy: 0.9304\n",
      "Epoch 18/30\n",
      "100/100 [==============================] - 24s 243ms/step - loss: 0.1528 - accuracy: 0.9244 - val_loss: 0.1028 - val_accuracy: 0.9396\n",
      "Epoch 19/30\n",
      "100/100 [==============================] - 22s 215ms/step - loss: 0.1561 - accuracy: 0.9231 - val_loss: 0.1175 - val_accuracy: 0.9382\n",
      "Epoch 20/30\n",
      "100/100 [==============================] - 22s 220ms/step - loss: 0.1582 - accuracy: 0.9297 - val_loss: 0.1839 - val_accuracy: 0.9284\n",
      "Epoch 21/30\n",
      "100/100 [==============================] - 22s 218ms/step - loss: 0.1595 - accuracy: 0.9251 - val_loss: 0.0831 - val_accuracy: 0.9310\n",
      "Epoch 22/30\n",
      "100/100 [==============================] - 22s 219ms/step - loss: 0.1507 - accuracy: 0.9248 - val_loss: 0.1535 - val_accuracy: 0.9238\n",
      "Epoch 23/30\n",
      "100/100 [==============================] - 22s 219ms/step - loss: 0.1507 - accuracy: 0.9284 - val_loss: 0.1384 - val_accuracy: 0.9363\n",
      "Epoch 24/30\n",
      "100/100 [==============================] - 22s 219ms/step - loss: 0.1529 - accuracy: 0.9297 - val_loss: 0.1877 - val_accuracy: 0.9423\n",
      "Epoch 25/30\n",
      "100/100 [==============================] - 22s 217ms/step - loss: 0.1419 - accuracy: 0.9254 - val_loss: 0.1104 - val_accuracy: 0.9310\n",
      "Epoch 26/30\n",
      "100/100 [==============================] - 22s 221ms/step - loss: 0.1431 - accuracy: 0.9281 - val_loss: 0.2732 - val_accuracy: 0.9310\n",
      "Epoch 27/30\n",
      "100/100 [==============================] - 22s 218ms/step - loss: 0.1377 - accuracy: 0.9268 - val_loss: 0.1327 - val_accuracy: 0.9415\n",
      "Epoch 28/30\n",
      "100/100 [==============================] - 22s 221ms/step - loss: 0.1381 - accuracy: 0.9284 - val_loss: 0.0861 - val_accuracy: 0.9384\n",
      "Epoch 29/30\n",
      "100/100 [==============================] - 22s 217ms/step - loss: 0.1385 - accuracy: 0.9297 - val_loss: 0.2105 - val_accuracy: 0.9323\n",
      "Epoch 30/30\n",
      "100/100 [==============================] - 23s 228ms/step - loss: 0.1384 - accuracy: 0.9323 - val_loss: 0.0400 - val_accuracy: 0.9304\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>epochs</th>\n",
       "      <th>val_loss</th>\n",
       "      <th>val_accuracy</th>\n",
       "      <th>loss</th>\n",
       "      <th>accuracy</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.404463</td>\n",
       "      <td>0.816032</td>\n",
       "      <td>0.536477</td>\n",
       "      <td>0.745299</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0.360501</td>\n",
       "      <td>0.813403</td>\n",
       "      <td>0.347772</td>\n",
       "      <td>0.861038</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>0.317332</td>\n",
       "      <td>0.864652</td>\n",
       "      <td>0.275489</td>\n",
       "      <td>0.883049</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>3</td>\n",
       "      <td>3</td>\n",
       "      <td>0.184610</td>\n",
       "      <td>0.919152</td>\n",
       "      <td>0.244956</td>\n",
       "      <td>0.899043</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>0.054820</td>\n",
       "      <td>0.918528</td>\n",
       "      <td>0.227569</td>\n",
       "      <td>0.899373</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   epochs  val_loss  val_accuracy      loss  accuracy\n",
       "0       0  0.404463      0.816032  0.536477  0.745299\n",
       "1       1  0.360501      0.813403  0.347772  0.861038\n",
       "2       2  0.317332      0.864652  0.275489  0.883049\n",
       "3       3  0.184610      0.919152  0.244956  0.899043\n",
       "4       4  0.054820      0.918528  0.227569  0.899373"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Conv2D\n",
    "from keras.layers import MaxPool2D\n",
    "from keras.layers import Flatten\n",
    "from keras.layers import Dense\n",
    "from keras.layers import Dropout\n",
    "from keras.layers import BatchNormalization\n",
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "from sklearn.metrics import accuracy_score, classification_report, confusion_matrix, label_ranking_average_precision_score, label_ranking_loss, coverage_error \n",
    "\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "import os\n",
    "print(os.listdir(\"C:/Users/SplendidUnivers/Desktop/data2\"))\n",
    "\n",
    "sns.set_style({'xtick.bottom':False,\n",
    "               'ytick.left':False,\n",
    "               'axes.spines.bottom': False,\n",
    "               'axes.spines.left': False,\n",
    "               'axes.spines.right': False,\n",
    "               'axes.spines.top': False})\n",
    "#read data\n",
    "#im_list = [1+i for i in range(81)]\n",
    "#fig, axes = plt.subplots(nrows=21, ncols=3, figsize=(28,28))\n",
    "#for i in enumerate(im_list):\n",
    "    #img = plt.imread(\"C:/Users/SplendidUnivers/Desktop/data2/train/\"+str(i[1])+\".jpg\")\n",
    "    #ax=axes[i[0]//3,i[0]%3]\n",
    "    #ax.imshow(img) \n",
    "\n",
    "#Classification\n",
    "classifier = Sequential()\n",
    "\n",
    "#Adding 1st Convolution and Pooling Layer\n",
    "classifier.add(Conv2D(32,kernel_size=(3,3),input_shape = (64, 64, 3), activation = 'elu'))\n",
    "classifier.add(BatchNormalization())\n",
    "classifier.add(MaxPool2D(pool_size=(2,2)))\n",
    "\n",
    "#Adding 2nd Convolution and Pooling Layer\n",
    "classifier.add(Conv2D(32,kernel_size=(3,3),activation = 'elu'))\n",
    "classifier.add(BatchNormalization())\n",
    "classifier.add(MaxPool2D(pool_size=(2,2)))\n",
    "\n",
    "#Adding 3rd Convolution and Pooling Layer\n",
    "classifier.add(Conv2D(32,kernel_size=(3,3), activation = 'elu'))\n",
    "classifier.add(BatchNormalization())\n",
    "classifier.add(MaxPool2D(pool_size=(2,2)))\n",
    "\n",
    "#Adding 4th Convolution and Pooling Layer\n",
    "classifier.add(Conv2D(32,kernel_size=(3,3), activation = 'elu'))\n",
    "classifier.add(BatchNormalization())\n",
    "classifier.add(MaxPool2D(pool_size=(2,2)))\n",
    "\n",
    "\n",
    "#Flattening\n",
    "classifier.add(Flatten())\n",
    "\n",
    "classifier.add(Dense(units=128,activation='elu'))\n",
    "classifier.add(BatchNormalization())\n",
    "classifier.add(Dense(units = 1, activation = 'sigmoid'))\n",
    "\n",
    "classifier.compile(optimizer = 'adam', loss = 'binary_crossentropy', metrics = ['accuracy'])\n",
    "\n",
    "\n",
    "#split data train and validation\n",
    "train_datagen = ImageDataGenerator(rescale=1./255,shear_range=0.2,zoom_range=0.2,horizontal_flip=True)\n",
    "val_datagen = ImageDataGenerator(rescale=1./255)\n",
    "train_data = train_datagen.flow_from_directory('C:/Users/SplendidUnivers/Desktop/data2/',\n",
    "                                                     target_size=(64,64),\n",
    "                                                     batch_size=32,\n",
    "                                                     class_mode = 'binary')\n",
    "val_data = val_datagen.flow_from_directory('C:/Users/SplendidUnivers/Desktop/data2/',\n",
    "                                                        target_size=(64,64),\n",
    "                                                        batch_size=32,\n",
    "                                                        class_mode = 'binary')\n",
    "\n",
    "\n",
    "#accuracies calculation\n",
    "accuracies = classifier.fit_generator(train_data,\n",
    "                         steps_per_epoch = 100,\n",
    "                         epochs = 30,\n",
    "                         validation_data = val_data,\n",
    "                         validation_steps = 50)\n",
    "\n",
    "#show accuracy\n",
    "acc = pd.DataFrame.from_dict(accuracies.history)\n",
    "acc = pd.concat([pd.Series(range(0,30),name='epochs'),acc],axis=1)\n",
    "acc.head()\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEICAYAAABS0fM3AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nO3deXhU9fX48fcx7CD7IhIQVFxwA7WodUPckLpg1RZcKlbrioqKiv5spVhF+61VFJeKValVEa0Lai1uIGpdiAZZRSJSCEtE9kUCSc7vj3OH3ExmJpNkkjCZ83qeeTJzt/ncmck997OLquKccy5z7VLXCXDOOVe3PBA451yG80DgnHMZzgOBc85lOA8EzjmX4TwQOOdchvNA4HYQkSwR2SQi3VK5bV0Skb1FJOVtpEXkJBFZHHq9QESOTWbbKrzXkyJye1X3d64iDeo6Aa7qRGRT6GUzoBAoDl5foarPVeZ4qloMtEj1tplAVfdNxXFE5DLgQlXtFzr2Zak4tnPxeCBIY6q640Ic3HFepqrvxdteRBqoalFtpM25ivjvcefhRUP1mIj8SUReFJEXRGQjcKGIHCUin4nIOhFZISIPiUjDYPsGIqIi0j14/c9g/dsislFEPhWRHpXdNlh/moh8KyLrReRhEflERIbGSXcyabxCRPJEZK2IPBTaN0tEHhCR1SLyHTAgwedzh4hMjFr2iIj8NXh+mYjMD87nu+BuPd6x8kWkX/C8mYg8G6RtLnBYjPddFBx3roicGSw/CBgHHBsUu/0Y+mxHhfa/Mjj31SLymoh0TuazqcznHEmPiLwnImtEZKWI3BJ6n98Hn8kGEckRkd1jFcOJyMeR7zn4PKcH77MGuENEeorI1OBcfgw+t1ah/fcIznFVsH6siDQJ0rx/aLvOIrJFRNrFO1+XgKr6ox48gMXASVHL/gRsA87Agn5T4GfAEVhucE/gW2BYsH0DQIHuwet/Aj8ChwMNgReBf1Zh247ARuCsYN2NwHZgaJxzSSaNrwOtgO7Amsi5A8OAuUA20A6Ybj/zmO+zJ7AJaB469g/A4cHrM4JtBOgP/AQcHKw7CVgcOlY+0C94/hdgGtAG2AOYF7Xtr4DOwXdyfpCGTsG6y4BpUen8JzAqeH5KkMbeQBPgUeCDZD6bSn7OrYAC4HqgMdAS6Busuw34GugZnENvoC2wd/RnDXwc+Z6DcysCrgKysN/jPsCJQKPgd/IJ8JfQ+cwJPs/mwfZHB+ueAO4Ovc9NwKt1/X+Yro86T4A/UvRFxg8EH1Sw3wjgpeB5rIv746FtzwTmVGHb3wIfhdYJsII4gSDJNB4ZWv8KMCJ4Ph0rIousGxh9cYo69mfA+cHz04BvE2z7JnBN8DxRIFgS/i6Aq8PbxjjuHOAXwfOKAsEE4J7QupZYvVB2RZ9NJT/ni4CcONt9F0lv1PJkAsGiCtJwLjAjeH4ssBLIirHd0cD3gASvZwK/TPX/VaY8vGio/lsafiEi+4nIW0FWfwMwGmifYP+VoedbSFxBHG/b3cPpUPvPzY93kCTTmNR7Af9LkF6A54EhwfPzgR0V7CJyuoh8HhSNrMPuxhN9VhGdE6VBRIaKyNdB8cY6YL8kjwt2fjuOp6obgLVAl9A2SX1nFXzOXYG8OGnoigWDqoj+Pe4mIpNEZFmQhmei0rBYrWFCGar6CZa7OEZEDgS6AW9VMU0ZzwNB/RfddPJv2B3o3qraEvgDdodek1Zgd6wAiIhQ9sIVrTppXIFdQCIqat76InCSiGRjRVfPB2lsCrwMjMGKbVoD7ySZjpXx0iAiewKPYcUj7YLjfhM6bkVNXZdjxU2R4+2KFUEtSyJd0RJ9zkuBveLsF2/d5iBNzULLdovaJvr87sNaux0UpGFoVBr2EJGsOOn4B3AhlnuZpKqFcbZzFfBAkHl2BdYDm4PKtitq4T3fBA4VkTNEpAFW7tyhhtI4CRguIl2CisNbE22sqgVY8cXTwAJVXRisaoyVW68CikXkdKwsO9k03C4ircX6WQwLrWuBXQxXYTHxMixHEFEAZIcrbaO8AFwqIgeLSGMsUH2kqnFzWAkk+pwnA91EZJiINBKRliLSN1j3JPAnEdlLTG8RaYsFwJVYo4QsEbmcUNBKkIbNwHoR6YoVT0V8CqwG7hGrgG8qIkeH1j+LFSWdjwUFV0UeCDLPTcDFWOXt37A74hoVXGx/DfwV+8feC8jF7gRTncbHgPeB2cAM7K6+Is9jZf7Ph9K8DrgBeBWrcD0XC2jJuBPLmSwG3iZ0kVLVWcBDwBfBNvsBn4f2fRdYCBSISLiIJ7L/f7AinFeD/bsBFySZrmhxP2dVXQ+cDJyDVU5/CxwfrP4/4DXsc96AVdw2CYr8fgfcjjUc2Dvq3GK5E+iLBaTJwL9CaSgCTgf2x3IHS7DvIbJ+MfY9b1PV/1by3F1IpKLFuVoTZPWXA+eq6kd1nR6XvkTkH1gF9Ki6Tks68w5lrlaIyAAsq78Va35YhN0VO1clQX3LWcBBdZ2WdOdFQ662HAMswooMBgCDvHLPVZWIjMH6MtyjqkvqOj3pzouGnHMuwyWVIxCRAWKjK+aJyMgY6/cQkfdFZJaITAua4kXWFYvIzOAxObS8R9BGe6HYMAiNUnNKzjnnKqPCHEFQsfct1oIgH2uJMURV54W2eQl4U1UniEh/4BJVvShYt0lDg6OF9pkEvKKqE0XkceBrVX2sgvR69sU55yovYf+XZHIEfYE8VV2kqtuAiVgFTVgvrCkZwNQY68umyDoU9ae0ad8EYFASaXHOOZdiyQSCLpTtFp5P+V6hX2PtjQHOBnYNjQLYJBid8DMRiVzs2wHrtHQI2ljHdM45VwuSaT4aK0sRXUQzAhgXDDc7HevuHrnId1PV5UFTrw9EZDbWCaWiYzrnnKsFyQSCfMqOm5KNdQbaQVWXA78EEJEWwDlBz8TIOlR1kYhMA/pgvQdbS+nEFOWO6ZxzrnYkUzQ0A+gZtPJpBAzGuoLvICLtRSRyrNuAp4LlbYLxUBCR9tjQsfOCruhTKe0ufjE2hrpzzrlaVmEgCO7YhwFTgPnYKH9zRWS0BDMrAf2ABSLyLdAJuDtYvj+QIyJfYxf+e0OtjW4FbhSRPKzO4O8pOifnnHOVkG4dytIqsc45t5OodvNR55xz9ZgHAufcTu9//4NHH4WtW+s6JfWTBwLn3E5ryxb4wx9gv/3gmmvg/POhqKji/VJJFcaOhVtvhenTa//9a4PXETjndjqq8OKLcPPNkJ8Pv/419OoFd94Jv/0tPPkkSE1PsBqk4/rr4eGHISsLiouhbVsYOBDOOANOPRVatar5dKRAwk/L5yNwLk3MmQP33msXnu7dSx89ekC7drVzYawNublw3XXw8cfQuzc8/zwce6ytKy6G0aOhfXu4776aTYcqDB9uQeCGG2DUKHj3XZg8Gd56C/75T2jQAI4/3oLCGWfAnnvWbJpqiucInEsDH3wAZ59tz7OyYO3asuubNy8fHH7xCytSSRerVsH/+392t9+uHdxzj939Z4WmrleFYcOsvuC+++CWW2omLap28R871oLBX/9aNtAWF8Nnn8Ebb1hgmD/flh9wAAwaZPu0b18zaauixLcJqppOD+cyzrPPqjZsqHrAAar/+58tW7dOdeZM1ddeU33wQdXhw1UHDVLt3Vu1VStVUG3RQvWDD+o27cnYtk31gQcs3Q0a2LmsXRt/++Ji1cGD7RzHj099ekpKVG+4wY5//fX2uiJ5eXYO/furZmWptm5t38u2balPXxUlvLbW9YXdA4HT1atVX39d9ZNP6jolO5eSEtV77rH/0n79El8coy1ebIGjcWPVyZNrLo3VNW2a6n772TmeeqrqvHnJ7VdYqDpggOouu6j+61+pS09JieqNN1p6rrsuuSAQbc4c1ZNPtmPst5/qf/6TuvRVgwcCt3NZs8Yu/MOH2x2siP0SQfW441Tff79q/4A7kxdeUP3Zz1THjlX96afK7799u+oVV9hncsEFqlu3Vv4YP/5oacjKUn3uucrvX9M++8wC1V57WbCq7He+aZPqUUepNmqk+t571U9PSYnqTTfZZ37ttdX7DZaU2G98r73seKefrrpgQfXTWA0eCFzdWrvW/tFvvFG1T5/SC3+TJpaVHj1a9cMP7aLZubOtO+YY1XffTb+AsHat6vnn2zl07Gh/s7NVH3/c7mKTsXGj6i9+YfvedpsVhVTVhg2WmxBRffTRqh8n1ZYsUe3USbVHD9VVq6p+nNWrVQ880IrBvvii6scpKVEdMcI+82HDUve727pV9c9/Vt11VyveGzHCivXqgAcCl1qFhXZ38/nnqlOmqL74ourf/qZ633124brqKtUhQ1RPO031kEMs+w5293fCCap//KNd+GPdKf/0k+rDD6t26WL7/Pzn9h41GRBKSlTXr6/+cT74QLVrV7sDHz3a7urfe8/uWkG1e3fVp56y5fGsXKl62GH2mT32WPXTpKq6ZYvqGWdYGsaMSc0xq2PjRvtdtGypOndu9Y+3bJkFlHbtVOfPr/z+JSWqN99sn8/VV9fMb23FCtXf/tYCcseOqk8+qVpUlPr3ScADgUut007THUU50Y8GDVQ7dFDt2dOKJQYMUB01ysqCK1NEsnWr6iOP2N00qB55pOrbb6fmn7SkRHXhQqtovOCC0qDTv7/qG29U/g5861a70xOx8/788/Lv99Zbqoceau/Ts6cV1URfCL75xi5ozZqlvlx/27bSnMqtt9ZdTqu42Cq1d9nFvs9UWbjQchjZ2aUV6skoKVG95Rb7XK66quY/lxkz7OYG7PcwYYLq7NkJbg62bLE7jDvvTD5LGZsHgvpo06a6ed/ly+2C95vf2MXtk0/srm7ZMtXNm1P/j7R1q90Zd+tmv9YjjrCWMnl5ViyQzF1VSYlt/+STqhdeWBpcwC4egwdbTiayvGdP1XHj7M61IrNnqx58sO135ZWJv5eSEtVXX7WiDFDt1Uv15Zft4vjxx6pt21oQrU4RRyLFxXaxi6S1sgFvw4bEuZlkjBxp7z92bPWOE8vMmdbyaN99VWfNsu+8okckCFTl86iqkhLV558v+zts3Fj18MNVLxu6Xcdd+41+cvHfdOPPT7EKELDI+dVX1XlbDwT1zYQJVvwwblztv/fYsfarqUoWvDoKC1WfeMKKV8I5EBHVNm1U99zT/pFOOcUu7FddpXr77aoXXWTFNZHtO3ZU/fWvLbjMn182cG3bZpW8ffvatq1bW5FBrDvM4mJrLti4sR3zzTeTP5fiYtWJE0tbyxx4oB2nZ0+7ONWkkhILemA5hETNG1evtsA1fHhp3U6PHqrTp1ftvZ95xt73iitq7s77o4+s7ilejjXW44orai8IhG3frjrni836z1tn6Ygjp+uJrWZoO1aV/rYp1p5tftDzjlmm9/zhJ125slpvl/Da6h3K0szHH0P//tCokY3D8q9/lXY0qg1HHWUDf+Xm1t57hm3fbp2rfvgB1qwpfaxdG/t1u3bQr589TjjBOlgl0wP300/hgQfs8xWBc86xDkZHHmlDHgwdCu+/D2eeCePHQ8eOlT+X4mLrNTt6NHTuDK+8UnudkO67D0aOhNMPWcKkXS+j6Zkns+rsy5n+dSs+/BA+/BBmz7ZLUpMm9r3//Ofwwgvw/fc29MPo0dC4cXLvF/ndHnss/Oc/0LBhnA1VoaAAFi+2kebCf9evty/i4ovti41j7lz46qvk0tWmjQ0XsUtNjrq2ZYulf/Fi+/Aij0WLrLt4UZH1mjvsMPT4fiw78FRmNjmS3AXNmDkTZs60Tb//3joLVpF3KKsvFi1Sbd9edZ99VPPzrdy8SZPaa3+/aJHdqdx7b+28X3WVlFT/znPxYssVRDpp/exnlgNp1sxyKKmqs6j1MvtvvtHHe/5FhWI9pOEcPYDZO+5EmzUt1pNOUr3rLrvDDjdd3bhR9fLLbbuDDrLimIpEfrc9e1ouY4eCAmvKdPnllpXbZ5/Yt/Nt21qBeqSSpXFjK5v873+r98Ft22aVC8mUAVZk3TprDzthgmVFhwyxf9BOncqfT5Mmlh0cMMAqbN5+28rdKjh8NX8j1c8RiMgAYCyQBTypqvdGrd8Dm56yA7AGuFBV80WkN/AY0BIoBu5W1ReDfZ4BjgfWB4cZqqozK0hKxuYINmywO7Lly61r+z77wI8/2rLVq+G//4V9963ZNETuIqt5Z5KWNm2CCRNsaIMOHSwX0LNnCg6cn2/jFIjAgAE1/8EWFtoXeffd0KwZL5z3Cne834+eu23k+J+m0G/WQxxW/AWNzhxgWaDjj4+ZhXrrLbj0Ust13XUXjBhRdiiIiMjvdtky+Pxz2KfTenj1VctavP++ZYs6dCgdG2OPPco+32MP2HXX0gPOmgV/+xs8+yxs3AgHHwxXXgkXXAAtWyY+9+3bLaswbZo9Pv7YvliwHEZ4jI7odLRsCSUlsGQJfPONPRYsKH2+cmXp+2RlQbduNs5H5BEZ96NHD+jUqYazIDFVL0eAXfy/A/YEGgFfA72itnkJuDh43h94Nni+D9AzeL47sAJoHbx+Bji3ovePemSkoiLVgQOtRc7775ddl5dnZdTdu1sTtZrUu7fd5LhqmjdP9e67rVIj+m5xv/2sUH7KlKr1REtk+vTSiokhQzRmofPy5ap33GG38GCVAxMmxGyxsmqV6rnn2mZHH6363Xdl10d+t1lZJfre76epnn223c2DVTbcfrvVtlfFxo3WZrl3b90xnsYVV6jm5pZus3273aXfe6/dfbdoUfo59+plbUWffNLWX3mlbbPffqpNm5b/Xtq0KZ9badPG2gZfcom1nX79dWtXvRONKxGS+Dpf4QZwFDAl9Po24LaobeYC2cFzATbEOdbXocDggSBJkXFPHn889vovvrCiisMOS00uN5b58y0NDz5YM8ev14qLrU3pyJHWpCVyIenb1xr2z59vbUcfeMDGWYhcLJs1s55l48aVv8pWxpo1qr/7ne7ozJBMu80tW6zsq1cv22+33aysKCfHmol9953qsmVa8uNqffbvhdqqVYk2bx4qLtu2TW8YtMh+t42vLT3G9dfbxTlVZWElJfbZXnJJ6YX6iCPiX/gnTYodAKOPWVBgx33xResRdvXV9o/4xBMWUH/4Id16O1Y7EJyLFQdFXl8EjIva5nng+uD5L7EinHZR2/QF5gO7aGkgWADMAh4AGleUlhr/qHZCTzxh39L11yfe7s03rSXRgAE1c0MyapS1Glm2LPXHrpdKSiz7dvXVpR0VGjRQPekku7AvXRp/382brW3usGGlYxRE2rVec431uHvjDbubThT5S0qseVKnTvbjGDGi8u2OS0pssJxTTy1/lxx6LCFb+8sHCqqnN5qi9zX+vYLqdY0fV73sMvssaroH1Zo1dqdy4IGq++9vTceSufBnhoTX1grrCETkPOBUVb0seH0R0FdVrw1tszswDugBTAfOAQ5Q1fXB+s7AtKD46LPQspVBcdMTwHeqOjphYjKsjmDqVDjlFDj5ZBvqtkEFs0eMHw+XX576iTtUbVKQ3XazNLkKfPutlVtPnQpNm1rZ/9lnw+mnWzOVysrLg7fftseHH1orlLD27cuWb3fvDtnZ8MQT8O9/w2GH2Y+jT5/qndeCBXZuP/1kTce2bi3zvOSnQh7+9HBGfnI6W4sbcWqfH3jz49Y0aNaoeu/rUiHh1SCZQHAUMEpVTw1e3wagqmPibN8C+EZVs4PXLbEgMEZVX4qzTz9ghKqenjAxGRQIFi6EI46wZoX//W/ysyD94Q9WeTdqlM3mlApff20ThDz+OFxxRWqOWWeKiqzt6YoVVsG3cqU937IFhgyBgw6q+rG3bYM//xn+9CdrczlmjDV1bNYsdelXtfRHmiOGH99/b38LC23b5s2tUnjYsNg1uTVk/nx47jlrYpoms3dlgmoHggbAt8CJwDJgBnC+qs4NbdMeWKOqJSJyN1Csqn8QkUbA28Abqvpg1HE7q+oKERGsaGirqo6s4GQyIhCsXWvt1dessZYWlZn1SNVyBM88Y7mCSy+tfnpGjoT777frZaXauataI+jWre0ONW7j8WpShXXrSi/s8R4rVlhTq1i/+V12sVYhZ5wBt91mDecr45NPLDs2bx786lfw4IMWxWtbSUlpoIi0UHGuuoEAQEQGAg9iLYieUtW7RWQ0kKOqk0XkXGAMdqGeDlyjqoUiciHwNFaZHDFUVWeKyAdYc1MBZgJXquqmCpJS7wPB9u1w2mk2Sfb775dO0VfZY5xxBrz3nrVMPO20qqdH1a4nvXpZKUPSvv8eLrvMen+BXWi7di07hVb4b5cudte6fbtd1Neujf133TqLkAUFZS/y27aVT0PDhnYh7NzZHrvtZo/w88hj82YYNw4eesiOf/zxcPvtVi6XqIxt3ToLHI8/bk0GH33UpgZzbudS/UCwE0mrxEbMmmU3aq1bW1a5ZcvYOXVVuOYaeOwxePpp671aVRs32rXs22+tyfThh1eww8yZ1rV10CA49NAdiz/7zG6OJ0yA3/wmiTcuKbGL4i232AV09Gg76UixReTv8uVl78wbNCjtLp1Iw4ZWzh59IY88OnUqfd6mTeUrSjZtsvL0+++3xu+HHmoX+rPPLj9n4ssv2+S6P/xgM5yPHg0tWlTu/ZyrHR4I6tKMGdC3b/nlu+5q18dWrUoDBNhd9y23pGZi7hUrSks45s61IuNyZs+2CoVXXildNmiQLTvkEIYPt+t6QUES5b2LFllZ1LRpdic9frx1xomlsNA654QDRGGhXbzbtLEPJdbfJk1qZ5b2wkKbnfy++6zCZt994dZbrePSypUWsd980ypgx4+3Clnndl4eCOrSww/bTePTT9vN8vr1pY9168q/PvFEyxGkquPhxx9b8dKIEfB//xdaMXcu/PGP8NJLlkW54QarXJgwwe6G16+n+Jfn0fWj5zjymIZl4kQ5JSXwyCNWmZCVZTN9X3pp7Vywa1pxsQ04NGaM5Zqys62YStVq5a+7ruLmXM7VPR9rqC4NHWo9f+uy78nvfmfNyL/6Sq1X6+DB1ilg112tF+maNWV3WLtW9c47dWqzgQqqLx711/gziCxcaPNLgnViWLKkxs+nTpSUWEesU06x7rTff1/XKXKuMqrXj2Ank1aJBWt2udtuNuJiXVm7Fvbfp4iuJUv4bM0+ZDVvYneyN92UcBTHKy/Zyj+fEwoadaP5llUweLC1T91vP8sFPPywlZ83amRDdQ4dWj9yAc7VPwn/MWt95KNMUlhoJTDV7cdTZaowezZthl/M2NUXkbNmT8b1f8XK4++5J2EQ2L4dXn6jCWee25jm/5tnxT6TJ8MBB8CFF1pN9PDhNrbznDlwySUeBJxLUx4IatDcudZ/qVYDQXExfPSRVQr07GmjM06axK9u6MJpJxby/z4/kyU/dajwMO+9Z6OaDh6MBYx77rEAMmKEjR45Z47VJ7z5ppWbO+fSlhcN1aC//92a0i9cCHvvXYNvtGULvPsuvP66dRz48Ucrrunf31oAnX02dOzI4sV2Q3/iibZpohv4iy+2bQoKYkw+sn69VQp7U0nn0kXC7Lo3d6hBubnWTLQyPYOT9uOPdjf+2mvwzjs25kurVtaZadAgOPXUcuOzd+9uTd1HjLDWouecE/vQW7faTf+558aZgcrHDXCuXvEcQQ06+mi7cZ4+PcUHfu01+PWvrTdtdrZd+M86y8rtKxjGoajI+jWsXGljwsS6pr/6KvzylxZfTj45xWl3ztUFryyuC8XF1uw85fUDH31kBfd9+sCXX1qnrIcfhpNOSmosnwYNbFDKggJr8BPLxIk2adQJJ6Q47c65nZIHghqycKEV3ac0EMyZY7Old+9uxUKHHlqlljqHH26tRx9/3EY2Ddu0yaoZzjvP+0k5lyk8ENSQ3Fz7m7JAsGSJjWvftClMmVLJYUDLu+suK1W6/PKy47W98YZVNwweXM30OufShgeCGpKbaw13evVKwcFWr7bK302brGdavPF7KqFFCxsVYu5c+MtfSpdPnGgDgR59dLXfwjmXJjwQ1JDcXJvjpNpD8G/ZYmNKf/+9tec8+OCUpA/ssOeeay2JFi60Hshvv2310Kka68g5t/Pzf/caoGqBoNrFQkVFdlX+7DOb8un441OSvrCxY62J6JVXWpPS7du9WMi5TOPVgTVg6VIrzalWIFC1eSHffNMmO4nX6L+adt8d7r0Xrr7apqTcc88k5i5wztUrniOoASmpKP797+Gpp+zvVVelJF3xXHEF/PznpUNK+JBBzmWWpAKBiAwQkQUikici5eYVFpE9ROR9EZklItNEJDu07mIRWRg8Lg4tP0xEZgfHfCiYu7heyM21i2mVi/PHjbNJxy+7zOYMqGG77GJzq/Tta1MSOOcySzKT12dhk9efDORjk9cPUdV5oW1eAt5U1Qki0h+4RFUvEpG2QA5wONYr+EvgMFVdKyJfANcDnwH/Bh5S1bcrSG9a9Cw+6yybInL+/Crs/NJLVi9wxhk2IYo35nfOVV+1xxrqC+Sp6iIAEZkInAXMC23TC7gheD4VeC14firwrqquCfZ9FxggItOAlqr6abD8H8AgoKJAkBZyc+GYY0ILNm+2gdq2bk38KCiw4Z6POgpeeMGDgHOuViRzpekCLA29zgeOiNrma+AcYCxwNrCriLSLs2+X4JEfY3naW73aKot3zP8+bRoMHGi9tJJx0EHWq6tZs5pKonPOlZFMIIiVpYguohkBjBORocB0YBlQlGDfZI6ZlspUFG/ebHP37r473HyzTbwefjRtWn5Z584p6HzgnHPJSyYQ5ANdQ6+zgeXhDVR1OfBLABFpAZyjqutFJB/oF7XvtOCY2VHLyxwzXX31lf3t0we44w5YtAg+/BCOO65O0+Wcc/Ek02poBtBTRHqISCNgMDA5vIGItBeRyLFuA54Knk8BThGRNiLSBjgFmKKqK4CNInJk0FroN8DrKTifOpebC926QdsFn1pvrauv9iDgnNupVRgIVLUIGIZd1OcDk1R1roiMFpEzg836AQtE5FugE3B3sO8a4C4smMwARkcqjoGrgCeBPOA76lFFcZ+Di60dZteu1qR487cAABY2SURBVFvLOed2Yj4xTQpt2mSTgt15zPvc+dFJNkroKafUdbKcc84npqkts2bZyBB9PnkYLrnEg4BzLi14Q/UUys0pBrI4tN0SuP/puk6Oc84lxXMEKZT7j9m0ZxVdnrgT2rSp6+Q451xSvI4gVebN49ADC2nfMYt3VqZuzgDnnEsBryOoccXFbBt6OXP0APqct3ddp8Y55yrF6whSYexY5s3YxHYa0efoRnWdGuecqxTPEVRXXh7ccQe5h1wCpHCyeuecqyUeCKqjpMTmDGjUiNxDL6V5c+jZs64T5ZxzleOBoDqeeMLGEbr/fnIXtqB3b5/03TmXfvyyVVVLlsAtt8BJJ1Ey9LfMnOnFQs659OSBoKqGD7eiofHjyftO2LTJA4FzLj15IKiKOXPg1VdhxAjo3j01k9U751wd8UBQFffeC82bw3XXATbiaMOGcMABdZwu55yrAg8ElbVokc0nfNVV0LYtYIHggAOgkXchcM6lIQ8ElfXnP9uk8jfcANhoo7m5XizknEtfHggqY8UKePppG2J6990BWL4cVq0KTVbvnHNpJqlAICIDRGSBiOSJyMgY67uJyFQRyRWRWSIyMFh+gYjMDD1KRKR3sG5acMzIuo6pPbUa8Ne/QlGRNRsNlJmj2Dnn0lCFYw2JSBbwCHAyNun8DBGZrKrzQpvdgU1h+ZiI9AL+DXRX1eeA54LjHAS8rqozQ/tdoKo5KTqXmrVmDTz2GAweDHvuuWNxbi6IwCGH1GHanHOuGpLJEfQF8lR1kapuAyYCZ0Vto0DL4HkrYHmM4wwBXqhqQuvcww/D5s0wsmyGKDfXhpVo0aKO0uWcc9WUTCDoAiwNvc4PloWNAi4UkXwsN3BtjOP8mvKB4OmgWOj3IpJwvOw6tWkTPPQQnHkmHHRQmVVeUeycS3fJBIJYF+joCWKGAM+oajYwEHhWRHYcW0SOALao6pzQPheo6kHAscHjokqlvDY98YQVDd12W5nFa9bA//7ngcA5l96SCQT5QNfQ62zKF/1cCkwCUNVPgSZA+9D6wUTlBlR1WfB3I/A8VgS18ykshPvvhxNOgCOPLLNqZlDb4YHAOZfOkgkEM4CeItJDRBphF/XJUdssAU4EEJH9sUCwKni9C3AeVrdAsKyBiLQPnjcETgfmsDP6xz+sjejtt5db5S2GnHP1QYWthlS1SESGAVOALOApVZ0rIqOBHFWdDNwEjBeRG7Bio6FaOhnycUC+qi4KHbYxMCUIAlnAe8D4lJ1VqhQVwX33weGHw4knlludmwvZ2dChQx2kzTnnUsQnr0/khRfg/PPhlVfg7LPLre7VC/beGyZH54+cc27n4pPXV4kqjBkD++8PZ0W3loUtW2DBAi8Wcs6lP5+8Pp633oLZs2HChJjTjs2aZdMReCBwzqU7zxHEogr33AN77AFDhsTcxOcgcM7VF54jiGX6dPj0Uxg3ziYaiCE3F9q0gW7dajltzjmXYl5ZHMupp1ongcWLoWnTmJscdpgFgvfeq5UUOedcdXhlcaV8+SW88w7ceGPcILB1q1Uf/OxntZw255yrAR4Ioo0ZA61a2QxkccyeDdu3W/cC55xLdx4IwhYvtj4Dw4ZBy5ZxN8sJBs72QOCcqw88EIQtWGAthk47LeFmM2ZA+/ZeUeycqx88EIStXGl/d9st4WY5OVY/sBMPnO2cc0nzQBBWUGB/O3WKu8mWLTB3rhcLOefqDw8EYQUF0KxZwunGZs60HsUeCJxz9YUHgrCCgoS5AbD6AfBA4JyrPzwQhCURCHJyYPfd7eGcc/WBB4KwJAOB5wacc/WJB4KwCgLBhg3WwtQDgXOuPkkqEIjIABFZICJ5IjIyxvpuIjJVRHJFZJaIDAyWdxeRn0RkZvB4PLTPYSIyOzjmQyJ13BizqAhWrUoYCL76yroZeCBwztUnFQYCEckCHgFOA3oBQ0SkV9RmdwCTVLUPNqfxo6F136lq7+BxZWj5Y8DlQM/gMaDqp5ECP/5oV/kEfQi8R7Fzrj5KJkfQF8hT1UWqug2bhD56yi4FImMytAKWJzqgiHQGWqrqp8Hcxv8ABlUq5amWRB+CnBybosDnKHbO1SfJBIIuwNLQ6/xgWdgo4EIRyQf+DVwbWtcjKDL6UESODR0zv4Jj1q4kAsGMGZ4bcM7VP8kEglhl99HzAgwBnlHVbGAg8KyI7AKsALoFRUY3As+LSMskj1m7KggEa9bAokUeCJxz9U8yM5TlA11Dr7MpX/RzKUEZv6p+KiJNgPaq+gNQGCz/UkS+A/YJjpldwTFrVwWB4Msv7a/PQeCcq2+SyRHMAHqKSA8RaYRVBk+O2mYJcCKAiOwPNAFWiUiHoLIZEdkTqxRepKorgI0icmTQWug3wOspOaOqKiiAJk1g111jro5UFB96aC2myTnnakGFOQJVLRKRYcAUIAt4SlXnishoIEdVJwM3AeNF5AasiGeoqqqIHAeMFpEioBi4UlXXBIe+CngGaAq8HTzqzsqVlhuI04p1xgzYe2+bntI55+oTn7M44pRTrMfYZ5/FXN2tGxxzDDz/fI2lwDnnaorPWZyUBL2KCwpg6VKvKHbO1U8eCCISBIJIRbEHAudcfeSBAKC4OOHwEjNmWNVBnz61nC7nnKsFHggAVq+22WbiBIKcHNh//7gNipxzLq15IICEfQhUfehp51z95oEASietjxEIli+31R4InHP1lQcCKM0RxBh51KemdM7Vdx4IIGHRUE4OZGVB7961nCbnnKslHgjAAkHjxtCyZblVOTlw4IHQtGkdpMs552qBBwIo7UMQNbyEVxQ75zKBBwKI25ls8WJrWeqBwDlXn3kggLiBIDLiqA897ZyrzzwQQOnIo1FycqBRI6sjcM65+soDQUmJDS8Ro+loTg4cfLDVIzvnXH3lgWD1ahtrKCpHUFLiFcXOuczggSBOH4K8PJuewOsHnHP1XVKBQEQGiMgCEckTkZEx1ncTkakikisis0RkYLD8ZBH5UkRmB3/7h/aZFhxzZvDomLrTqoQ4gSBSUew5AudcfVfhVJXBnMOPACdjk87PEJHJqjovtNkdwCRVfUxEegH/BroDPwJnqOpyETkQm+6yS2i/C1Q1JzWnUkUJAkGTJtCrVx2kyTnnalEyOYK+QJ6qLlLVbcBE4KyobRSIdMttBSwHUNVcVV0eLJ8LNBGRnavqNU4gmDHD5h9oUGGodM659JZMIOgCLA29zqfsXT3AKOBCEcnHcgPXxjjOOUCuqhaGlj0dFAv9XiTOrPE1raDA2oi2br1jUXExfPWV1w845zJDMoEg1gU6ehL5IcAzqpoNDASeFZEdxxaRA4D7gCtC+1ygqgcBxwaPiyqT8JSJ9CEIxaFvvoEtW7x+wDmXGZIJBPlA19DrbIKin5BLgUkAqvop0ARoDyAi2cCrwG9U9bvIDqq6LPi7EXgeK4KqfTF6FXtFsXMukyQTCGYAPUWkh4g0AgYDk6O2WQKcCCAi+2OBYJWItAbeAm5T1U8iG4tIAxGJBIqGwOnAnOqeTJXECAQzZkCLFrDPPnWSIuecq1UVBgJVLQKGYS1+5mOtg+aKyGgROTPY7CbgdyLyNfACMFRVNdhvb+D3Uc1EGwNTRGQWMBNYBoxP9cklJU6O4LDDbB4C55yr78Su12kjtYktKbHxI26+Ge65B4Dt222S+mHD4C9/Sem7OedcXUnYGCezexavXQtFRWVyBHPnQmGh1w845zJHZgeCGH0IfI5i51ymyYjuUsOH29hB/frZo0+foPx/5UrbIDTyaE6OdSnYa6+6SKlzztW+jAgErVvDwoXw1lv2umVLOO44OKFVa/rRh0PadyJSLxwZcbSOurc551yty6jK4uXL4cMPYepUmDbNggNAq5YlHHf8LvTrByNHwk03wZgx1U+sc87tJBLe2mZUIIi2bNgYpj3+DdOGPsO0D4W8PFv+2mtwVvRoSs45l748EMT129/CO+9Afj4AS5fC/Plw0kmwS2ZXozvn6peEgSAj6gjiiupM1rWrPZxzLpNk9n1vjF7FzjmXaTI7EKxcGXPSeuecyySZGwhU4YcfPEfgnMt4mRsI1q61gYU8EDjnMlzmBoI4U1Q651ym8UDggcA5l+E8EHggcM5lOA8EHgiccxkuqUAgIgNEZIGI5InIyBjru4nIVBHJFZFZIjIwtO62YL8FInJqssescStXQoMG0LZtrb+1c87tTCoMBCKSBTwCnAb0AoaISK+oze7AprDsg81p/Giwb6/g9QHAAOBREclK8pg1q6AAOnb0sSSccxkvmatgXyBPVRep6jZgIhA9JJsCLYPnrYDlwfOzgImqWqiq3wN5wfGSOWbN8l7FzjkHJBcIugBLQ6/zg2Vho4ALRSQf+DdwbQX7JnPMmuWBwDnngOQCQaxR66JHAR0CPKOq2cBA4FkR2SXBvskcs2Z5IHDOOSC50UfzgfCYnNmUFv1EXIrVAaCqn4pIE6B9BftWdMyao+qBwDnnAsnkCGYAPUWkh4g0wip/J0dtswQ4EUBE9geaAKuC7QaLSGMR6QH0BL5I8pg1Z/162LbNA4FzzpFEjkBVi0RkGDAFyAKeUtW5IjIayFHVycBNwHgRuQEr4hmqNuPNXBGZBMwDioBrVLUYINYxa+D8Yosxab1zzmWqzJyh7MMPoV8/eO89OPHElBzSOed2YglnKMvMRvTeq9g553bwQOCccxkucwNBVha0a1fXKXHOuTqXuYGgQwcfXsI558jkQODFQs45B2RqIPBJ651zbofMDASeI3DOuR0yLxD48BLOOVdG5gWCDRugsNADgXPOBTIvEHgfAuecK8MDgXPOZTgPBM45l+EyLxD4yKPOOVdG5gWCggLrUezDSzjnHJCpgaBDBxtryDnnXIYGAq8fcM65HZIKBCIyQEQWiEieiIyMsf4BEZkZPL4VkXXB8hNCy2eKyFYRGRSse0ZEvg+t653aU4vDA4FzzpVR4VSVIpIFPAKcjE1GP0NEJqvqvMg2qnpDaPtrgT7B8qlA72B5WyAPeCd0+JtV9eUUnEfyCgqgZ89afUvnnNuZJZMj6AvkqeoiVd0GTATOSrD9EOCFGMvPBd5W1S2VT2aK+PASzjlXTjKBoAuwNPQ6P1hWjojsAfQAPoixejDlA8TdIjIrKFpqnERaqmfTJvjpJ2866pxzIckEgliTHsebRH4w8LKqFpc5gEhn4CBgSmjxbcB+wM+AtsCtSaSleiJ9CDxH4JxzOyQTCPKBrqHX2cDyONvGuusH+BXwqqpujyxQ1RVqCoGnsSKomuW9ip1zrpxkAsEMoKeI9BCRRtjFfnL0RiKyL9AG+DTGMcrVGwS5BEREgEHAnMolvQo8EDjnXDkVthpS1SIRGYYV62QBT6nqXBEZDeSoaiQoDAEmqmqZYiMR6Y7lKD6MOvRzItIBK3qaCVxZnRNJigcC55wrR6Ku2zu76iX2zjvhrrtg2zZoUGEMdM65+iJWXe8OmdWzuKAA2rf3IOCccyGZFwi8WMg558rIrECwcqX3IXDOuSiZFQg8R+Ccc+V4IHDOuQyXOYFg0ybYssUDgXPORcmcQOB9CJxzLiYPBM45l+E8EDjnXIbLnEAQGXnUm48651wZmRMICgpAxCaud845t0NmBYJ27Xx4Ceeci5JZgcDrB5xzrhwPBM45l+E8EDjnXIbzQOCccxkuqUAgIgNEZIGI5InIyBjrHxCRmcHjWxFZF1pXHFo3ObS8h4h8LiILReTFYBrMmrF5sw0x4U1HnXOunAoDgYhkAY8ApwG9gCEi0iu8jareoKq9VbU38DDwSmj1T5F1qnpmaPl9wAOq2hNYC1xazXOJzzuTOedcXMnkCPoCeaq6SFW3AROBsxJsX26i+mjBhPX9gZeDRROwCexrhgcC55yLK5lA0AVYGnqdHywrR0T2AHoAH4QWNxGRHBH5TEQiF/t2wDpVLaromCnhgcA55+JKpndVrEmP400iPxh4WVWLQ8u6qepyEdkT+EBEZgMbKnHM6vNA4JxzcSWTI8gHuoZeZwPL42w7mKhiIVVdHvxdBEwD+gA/Aq1FJBKIEh2z+iKBoGPHGnsL55xLV8kEghlAz6CVTyPsYj85eiMR2RdoA3waWtZGRBoHz9sDRwPzVFWBqcC5waYXA69X50QSKiiAtm2hYcMaewvnnEtXFQaCoBx/GDAFmA9MUtW5IjJaRMKtgIYAE4OLfMT+QI6IfI1d+O9V1XnBuluBG0UkD6sz+Hv1TycOn7TeOefikrLX7Z1e1RI7Zgxs2GB/nXMu88Sq6y1dmRGBwDnnMlvCQJA5Q0w455yLyQOBc85lOA8EzjmX4TwQOOdchvNA4JxzGc4DgXPOZTgPBM45l+E8EDjnXIZLZvTRnUnCThHOOecqz3MEzjmX4TwQOOdchvNA4JxzGc4DgXPOZTgPBM45l+E8EDjnXIbzQOCccxkuIwKBiAwQkQUikiciI+s6PdUlIotFZLaIzBSRnLpOT1WIyFMi8oOIzAktaysi74rIwuBvm7pMY2XEOZ9RIrIs+J5misjAukxjZYhIVxGZKiLzRWSuiFwfLE/n7yjeOaXl9yQiTUTkCxH5OjifPwbLe4jI58F39GIw13ziY6XZDGWVJiJZwLfAyUA+MAMYEpo7Oe2IyGLgcFX9sa7TUlUichywCfiHqh4YLPszsEZV7w0CdhtVvbUu05msOOczCtikqn+py7RVhYh0Bjqr6lcisivwJTAIGEr6fkfxzulXpOH3JCICNFfVTSLSEPgYuB64EXhFVSeKyOPA16r6WKJjZUKOoC+Qp6qLVHUbMBE4q47TlPFUdTqwJmrxWcCE4PkE7J80LcQ5n7SlqitU9avg+UZgPtCF9P6O4p1TWlKzKXjZMHgo0B94OVie1HeUCYGgC7A09DqfNP7yAwq8IyJfisjldZ2YFOqkqivA/mmBjnWcnlQYJiKzgqKjtClGCROR7kAf4HPqyXcUdU6Qpt+TiGSJyEzgB+Bd4DtgnaoWBZskdb3LhEAQa3yidC8PO1pVDwVOA64JiiXczucxYC+gN7ACuL9uk1N5ItIC+BcwXFU31HV6UiHGOaXt96SqxaraG8jGSj/2j7VZRcfJhECQD3QNvc4GltdRWlJCVZcHf38AXsV+APVBQVCOGynP/aGO01MtqloQ/KOWAONJs+8pKHf+F/Ccqr4SLE7r7yjWOaX79wSgquuAacCRQGsRiQwomtT1LhMCwQygZ1CT3ggYDEyu4zRVmYg0Dyq6EJHmwCnAnMR7pY3JwMXB84uB1+swLdUWuWAGziaNvqegIvLvwHxV/WtoVdp+R/HOKV2/JxHpICKtg+dNgZOweo+pwLnBZkl9R/W+1RBA0BzsQSALeEpV767jJFWZiOyJ5QLAhhF/Ph3PR0ReAPoB7YEC4E7gNWAS0A1YApynqmlRARvnfPphxQ0KLAauiJSv7+xE5BjgI2A2UBIsvh0rU0/X7yjeOQ0hDb8nETkYqwzOwm7qJ6nq6OAaMRFoC+QCF6pqYcJjZUIgcM45F18mFA0555xLwAOBc85lOA8EzjmX4TwQOOdchvNA4JxzGc4DgXPOZTgPBM45l+H+P3ohBI43bX28AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "accuracy = accuracies.history['accuracy']\n",
    "val_accuracy = accuracies.history['val_accuracy']\n",
    "loss = accuracies.history['loss']\n",
    "val_loss = accuracies.history['val_loss']\n",
    "\n",
    "epochs = range(len(acc))\n",
    "\n",
    "plt.plot(epochs, accuracy, 'r', label='Training accuracy')\n",
    "plt.plot(epochs, val_accuracy, 'b', label='Validation accuracy')\n",
    "plt.title('Training and validation accuracy')\n",
    "\n",
    "plt.figure()\n",
    "\n",
    "plt.plot(epochs, loss, 'r', label='Training Loss')\n",
    "plt.plot(epochs, val_loss, 'b', label='Validation Loss')\n",
    "plt.title('Training and validation loss')\n",
    "plt.legend()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "test",
   "language": "python",
   "name": "test"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}