day4_binary_class_bag_of_words_imdb.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Neural network\n",
    "Supervised learning with labelled data\n",
    "We have an input vector or tensor. We have layers with weights and biases + activation functions. \n",
    "It is important that the activation functions have to be not linear. We have also the outputs there are the result of the application funtion applied to the result of multiplication between weights and the input vector x. <br>\n",
    "The model give us the prediction vector that compared to the real values, give us loss and we use the optimizer to optimize the weights to decrease the loss.\n",
    "The accuracy is the metric that we are using to evaluate the model also during the training.\n",
    "With the mnist we made classification in ten different classes (loss function = categorical cross entropy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Flow\n",
    "We are in a NLP domain (Natural Language Processing)\n",
    "\n",
    "In this model we will work for binary classification. We have to classify the films' feedbacks based on the sentiment analysis. So we have labels (0 if negative and 1 if positive) as binary output\n",
    "\n",
    "imdb is an array with 25k elements, each one of different length. These arrays are composed by numbers and the maximum lehgth is 2494. \n",
    "\n",
    "To match the numbers with the words, we need to download a dictionary. To use the dictionary (word_index - ~30k elements) we need to reverse its order so we have the numbers as key and the words as values. In this case we can \"translate\" the array into text.\n",
    "\n",
    "But for the model we'll not use the dictionary, because we now that the maximun index in our array is 10k, so we'll use bag of words, where basically we create a new matrix of all zeros with as many rows as the data and 10k columns. Using the fancy index, we'll put 1 when there is correspondence between the index and the word. And this will be the input of our neural network.\n",
    "\n",
    "General:\n",
    "It is important that in the classification problem the training test is balanced. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Let's start importing the dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.datasets import imdb\n",
    "import numpy as np\n",
    "\n",
    "(train_data,train_labels),(test_data,test_labels) = imdb.load_data(num_words=10000)\n",
    "#in this way we are using the most 10000 words used and cutting out the others"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "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>0</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>25000.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean</th>\n",
       "      <td>238.713640</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>std</th>\n",
       "      <td>176.497204</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>min</th>\n",
       "      <td>11.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25%</th>\n",
       "      <td>130.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50%</th>\n",
       "      <td>178.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>75%</th>\n",
       "      <td>291.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>max</th>\n",
       "      <td>2494.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  0\n",
       "count  25000.000000\n",
       "mean     238.713640\n",
       "std      176.497204\n",
       "min       11.000000\n",
       "25%      130.000000\n",
       "50%      178.000000\n",
       "75%      291.000000\n",
       "max     2494.000000"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x=[]\n",
    "for i in train_data:\n",
    "    x.append(len(i))\n",
    "import pandas as pd\n",
    "x = pd.DataFrame(x)\n",
    "x.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Different way to do it\n",
    "import numpy as np\n",
    "lenghts = [len(x) for x in train_data]#list comprehension\n",
    "print('min=',np.min(lenghts))\n",
    "print('mean',np.mean(lenghts))\n",
    "print('std=',np.std(lenghts))\n",
    "print('max',np.max(lenghts))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Cropping (cut away some part) and Padding (adding as many 0 to lists to get them 2494 long)\n",
    "### Normalizing we would loose semantic sense of the numbers\n",
    "### corpus = collections of data of text to analyze vocabulary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Let's import the dictionary to get the words corresponding to the numbers of our array\n",
    "word_index = imdb.get_word_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Let's reverse the dictionary\n",
    "reverse_word_index = dict([(value,key) for (key,value) in word_index.items()])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "the as you with out themselves powerful lets loves their becomes reaching had journalist of lot from anyone to have after out atmosphere never more room and it so heart shows to years of every never going and help moments or of every chest visual movie except her was several of enough more with is now current film as you of mine potentially unfortunately of you than him that with out themselves her get for was camp of you movie sometimes movie that with scary but and to story wonderful that in seeing in character to of 70s musicians with heart had shadows they of here that with her serious to have does when from why what have critics they is you that isn't one will very to as itself with other and in of seen over landed for anyone of and br show's to whether from than out themselves history he name half some br of and odd was two most of mean for 1 any an boat she he should is thought frog but of script you not while history he heart to real at barrel but when from one bit then have two of script their with her nobody most that with wasn't to with armed acting watch an for with heartfelt film want an\n",
      "\n",
      " The label of this text is: 1\n"
     ]
    }
   ],
   "source": [
    "#To get the text from an array we can use the get function that let us index an array - LIST COMPREHENSION\n",
    "decoded_array = ' '.join([reverse_word_index.get(i,'?') for i in train_data[0]])\n",
    "print(decoded_array)\n",
    "print('\\n The label of this text is:',train_labels[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Bag of Words\n",
    "We will use for start the bag of words where we take the unique words in the text loosing also the order of the words and the counting\n",
    "\n",
    "The number of indexes in the array is maximun 10k. So we create a matrix with same rows as our data and 10k columns.\n",
    "We use fancy indexing to put 1 when the word is in the array or 0 if not.\n",
    "\n",
    "### We are using fancy indexing\n",
    "We pass arrays of indices in place of single scalars. This allows us to very quickly access and modify complicated subsets of an array's values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "the maximum number in the array is: 9995 --> for this reason the dimension has to be 10k\n",
      "[0. 1. 1. ... 0. 0. 0.]\n",
      "(10000,)\n",
      "(25000, 10000)\n"
     ]
    }
   ],
   "source": [
    "#Transforming the training data in numpy array\n",
    "\n",
    "def vectorize_sequences(sequences,dimension=10000):  #take sequenses and turn them into vectors\n",
    "    results = np.zeros((len(sequences),dimension))\n",
    "    for i, sequence in enumerate(sequences):\n",
    "        results[i,sequence] = 1.0   #fancy indexing\n",
    "    return results\n",
    "\n",
    "x_train = vectorize_sequences(train_data)\n",
    "x_test = vectorize_sequences(test_data)\n",
    "print('the maximum number in the array is:',max(np.max(train_data)),'--> for this reason the dimension has to be 10k')\n",
    "print(x_train[0])\n",
    "print(x_train[0].shape)\n",
    "print(x_train.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "transforming labels from <class 'numpy.ndarray'> to\n"
     ]
    }
   ],
   "source": [
    "#Transforming the labels\n",
    "print('transforming labels from',type(train_labels),'to')\n",
    "y_train = np.asarray(train_labels).astype('float32')\n",
    "y_test = np.asarray(test_labels).astype('float32')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Defining the Neural Network Architecture"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_1 (Dense)              (None, 16)                160016    \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 1)                 17        \n",
      "=================================================================\n",
      "Total params: 160,033\n",
      "Trainable params: 160,033\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "from keras import models,layers\n",
    "\n",
    "model = models.Sequential()\n",
    "model.add(layers.Dense(16,activation='relu', input_shape=(10000,)))  #the number of columns is the number of features\n",
    "model.add(layers.Dense(1, activation='sigmoid'))  #because we are doing binary classification and we have to\n",
    "#predict\n",
    "\n",
    "model.summary()  #neural network architecture\n",
    "\n",
    "\"\"\"\n",
    "from keras import optimizers, losses, metrics\n",
    "model.compile(optimizer=optimizers.RMSprop(lr=0.001),\n",
    "loss=losses.binary_crossentropy,\n",
    "metrics=[metrics.binary_accuracy]) #square parenthesis because it's not a class\n",
    "\"\"\"\n",
    "\n",
    "model.compile(optimizer='adam',loss='binary_crossentropy',metrics=['accuracy'])\n",
    "#rmsprop because it's always a good start\n",
    "\n",
    "#160033 is number of weights plus biases"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Validation\n",
    "If loss is low, the model is good for optimizing the problem\n",
    "If the loss is low in test set, the model is good in generalizing\n",
    "Information leak\n",
    "Validation data when we split data in 70 train,10 validation, test =20\n",
    "so we validate the model on validation before the test set so we can more fine tuning\n",
    "The first important thing to evaluate in a model is a validation loss, instead of the training loss (also accuracy, always related to the validation)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "#sometimes it's good to shuffle but being sure that labels and data are \n",
    "#connected to each other and there is a method for shuffle\n",
    "x_val = x_train[:10000]\n",
    "partial_x_train = x_train[10000:]  #the rest of 25000\n",
    "y_val = y_train[:10000]\n",
    "partial_y_train=y_train[10000:]\n",
    "#In our case 25k as train (15k trian), 10k as validation and 25k for test"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 15000 samples, validate on 10000 samples\n",
      "Epoch 1/4\n",
      "15000/15000 [==============================] - 6s 367us/step - loss: 0.4490 - acc: 0.8157 - val_loss: 0.3158 - val_acc: 0.8841\n",
      "Epoch 2/4\n",
      "15000/15000 [==============================] - 4s 247us/step - loss: 0.2365 - acc: 0.9207 - val_loss: 0.2796 - val_acc: 0.8918\n",
      "Epoch 3/4\n",
      "15000/15000 [==============================] - 3s 229us/step - loss: 0.1746 - acc: 0.9457 - val_loss: 0.2773 - val_acc: 0.8886\n",
      "Epoch 4/4\n",
      "15000/15000 [==============================] - 3s 201us/step - loss: 0.1394 - acc: 0.9594 - val_loss: 0.2881 - val_acc: 0.8864\n"
     ]
    }
   ],
   "source": [
    "history = model.fit(partial_x_train,partial_y_train,epochs=4,batch_size=256, \n",
    "          validation_data=(x_val,y_val))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX0AAAD8CAYAAACb4nSYAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvIxREBQAAIABJREFUeJzt3XlYVOfZ+PHvzQ6CgAKiMCjuGyoKahbN3ppNs2NM+ybpkrfZtE2aNk3SNs3SpMubZrNN87bp9saIMUlrm61JYxZ/VQMC7kuQRAE3VMAFke3+/TEjImUZcWBmmPtzXXMxZ78fJ7nPeZ7znOeIqmKMMSYwBHk7AGOMMT3Hkr4xxgQQS/rGGBNALOkbY0wAsaRvjDEBxJK+McYEEEv6xhgTQCzpG2NMALGkb4wxASTE2wG0lpCQoEOGDPF2GMYY41fWrFmzX1UTO1vP55L+kCFDyM/P93YYxhjjV0RkhzvrWfOOMcYEEEv6xhgTQCzpG2NMAPG5Nn1jTGCqr6+nrKyM2tpab4fi0yIiIkhNTSU0NLRL27uV9EVkFvAMEAz8TlWfbGe9a4GlQLaq5ovIEGAzsNW1yipV/VaXIjXG9GplZWXExMQwZMgQRMTb4fgkVeXAgQOUlZWRnp7epX10mvRFJBhYCFwClAF5IrJMVTe1Wi8GWACsbrWL7ao6qUvRGWMCRm1trSX8TogI/fv3p6Kiosv7cKdNfypQrKolqloHLAbmtLHeo8DPAKubGWO6xBJ+587038idpJ8ClLaYLnPNaxnEZMChqm+2sX26iBSKyEciMqProXbseEMjT7y9mbLKmu46hDHG+L0z7r0jIkHAU8C9bSzeDaSpaiZwD7BIRPq2sY/bRCRfRPK7Wm3Zd+g4L6/ayd2vFFLf2NSlfRhjAlt0dLS3Q+h27iT9csDRYjrVNe+EGGA88KGIfAFMB5aJSJaqHlfVAwCqugbYDoxsfQBVfVFVs1Q1KzGx06eI2+ToF8WT12ZQuLOKX7y7tfMNjDEmALmT9POAESKSLiJhwFxg2YmFqlqtqgmqOkRVhwCrgNmu3juJrhvBiMhQYARQ4vFSuFwxYRBfmZ7Gix+X8K/Ne7vrMMaYXk5Vue+++xg/fjwZGRnk5uYCsHv3bmbOnMmkSZMYP348n3zyCY2Njdxyyy3N6/7qV7/ycvQd67T3jqo2iMhdwLs4u2y+pKobReQRIF9Vl3Ww+UzgERGpB5qAb6nqQU8E3p6HLh9LwY4q7n11LW/Nn8GguMjuPJwxphv85O8b2bTrkEf3OXZQX3585Ti31n399dcpKipi7dq17N+/n+zsbGbOnMmiRYv48pe/zIMPPkhjYyM1NTUUFRVRXl7Ohg0bAKiqqvJo3J7mVpu+qr6lqiNVdZiqPu6a96O2Er6qnq+q+a7vr6nqOFWdpKqTVfXvng3/P0WEBrPwpsnUNzRZ+74xpktWrFjBjTfeSHBwMAMGDOC8884jLy+P7Oxs/vCHP/Dwww+zfv16YmJiGDp0KCUlJdx9992888479O37H7ctfUqvfCI3PaEPT1w7gfmvFPLLf27lB5eO8XZIxpjT4O4VeU+bOXMmH3/8MW+++Sa33HIL99xzD//1X//F2rVreffdd3nhhRdYsmQJL730krdDbVevHXtn9sRB3Dg1jd9+VMLyLfu8HY4xxo/MmDGD3NxcGhsbqaio4OOPP2bq1Kns2LGDAQMG8M1vfpNvfOMbFBQUsH//fpqamrj22mt57LHHKCgo8Hb4HeqVV/on/PjKsRTurOSeJUW8tWAGA2Otfd8Y07mrr76alStXMnHiRESEn//85yQnJ/OnP/2JX/ziF4SGhhIdHc2f//xnysvLufXWW2lqcjYlP/HEE16OvmOiqt6O4RRZWVnqyZeobK84wpXPrWDcoL688s3phAT32sqNMX5t8+bNjBljTbHuaOvfSkTWqGpWZ9v2+gw4LDGan16dQd4XlTz13jZvh2OMMV7V65M+wFWZKczNdvDrD7fz4VZr3zfGBK6ASPrg7A0wakAM9yxZy55qGxPOGBOYAibpR4Y5++/X1jcyf3EhDdZ/3xgTgAIm6QMMT4rmsavG8+nnB3n6/c+8HY4xxvS4gEr6ANdMTuWGrFQWfljMx9u6/iICY4zxRwGX9AF+Mns8I5Ki+U5uEXsPWfu+MSZwBGTSjwwLZuG8ydTUNbJgcSGNTb71rIIxxvd1NPb+F198wfjx43swGvcFZNIHGDEghkevGs+qkoM88y9r3zfGBIZePQxDZ66bksrK7Qd47oPPmDqkH+eOSPB2SMYYgLfvhz3rPbvP5Ay49Ml2F99///04HA7uvPNOAB5++GFCQkJYvnw5lZWV1NfX89hjjzFnTluvCG9fbW0tt99+O/n5+YSEhPDUU09xwQUXsHHjRm699Vbq6upoamritddeY9CgQdxwww2UlZXR2NjID3/4Q3Jycs6o2K0F7JX+CY9eNY5hidF8O7eIfYetfd+YQJWTk8OSJUuap5csWcLNN9/MG2+8QUFBAcuXL+fee+/ldIeuWbhwISLC+vXreeWVV7j55pupra3lhRdeYMGCBRQVFZGfn09qairvvPMOgwYNYu3atWzYsIFZs2Z5upiBfaUPEBUWwq9vmszs51ew4JUi/u8b0wgOOrO3zRtjzlAHV+TdJTMzk3379rFr1y4qKiqIj48nOTmZ73znO3z88ccEBQVRXl7O3r17SU5Odnu/K1as4O677wZg9OjRDB48mG3btnHWWWfx+OOPU1ZWxjXXXMOIESPIyMjg3nvv5fvf/z5XXHEFM2bM8Hg5A/5KH2DkgBgemTOelSXOph5jTGC6/vrrWbp0Kbm5ueTk5PDyyy9TUVHBmjVrKCoqYsCAAdTWeqZFYN68eSxbtozIyEguu+wyPvjgA0aOHElBQQEZGRk89NBDPPLIIx45VkuW9F2un5LKNZkpPPOvz/h38X5vh2OM8YKcnBwWL17M0qVLuf7666muriYpKYnQ0FCWL1/Ojh07TnufM2bM4OWXXwZg27Zt7Ny5k1GjRlFSUsLQoUOZP38+c+bMYd26dezatYuoqCi+8pWvcN9993XL2PwB37xzgojw6FXjWVtWxYLcIt6aP4PEmHBvh2WM6UHjxo3j8OHDpKSkMHDgQG666SauvPJKMjIyyMrKYvTo0ae9zzvuuIPbb7+djIwMQkJC+OMf/0h4eDhLlizhL3/5C6GhoSQnJ/PAAw+Ql5fHfffdR1BQEKGhofzmN7/xeBl7/Xj6p2vLnkPMef7/kT2kH3/62lRr3zemh9h4+u6z8fQ9aHRyX34yexwrivezcHmxt8MxxhiPsuadNuRkO1hVcoCn39/G1PR+TB/a39shGWN80Pr16/nqV796yrzw8HBWr17tpYg6Z0m/DSLCY1dnsK6smvmvFPLWghkkRFv7vjHdTVUR8Z8m1YyMDIqKinr0mGfaJO9W846IzBKRrSJSLCL3d7DetSKiIpLVYt4PXNttFZEvn1G0PSg6PITn502m6lg938ktosnG5zGmW0VERHDgwIEzTmq9mapy4MABIiIiuryPTq/0RSQYWAhcApQBeSKyTFU3tVovBlgArG4xbywwFxgHDALeF5GRqtrY5Yh70NhBfXn4ynE88MZ6fvPRdu68YLi3QzKm10pNTaWsrIyKChvyvCMRERGkpqZ2eXt3mnemAsWqWgIgIouBOcCmVus9CvwMuK/FvDnAYlU9DnwuIsWu/a3scsQ97MapDlaWHOB//rmVrMHxTLP2fWO6RWhoKOnp6d4Oo9dzp3knBShtMV3mmtdMRCYDDlV983S39XUiwk+vHs/g/n2Yv7iQA0eOezskY4zpsjPusikiQcBTwL1nsI/bRCRfRPJ9sWoXExHK8/Myqayp554la6193xjjt9xJ+uWAo8V0qmveCTHAeOBDEfkCmA4sc93M7WxbAFT1RVXNUtWsxMTE0ytBDxk3KJYfXTGWj7ZV8MLH270djjHGdIk7ST8PGCEi6SIShvPG7LITC1W1WlUTVHWIqg4BVgGzVTXftd5cEQkXkXRgBPCpx0vRQ26alsblEwbyP//cRt4XB70djjHGnLZOk76qNgB3Ae8Cm4ElqrpRRB4RkdmdbLsRWILzpu87wJ3+0nOnLSLCk9dkkBofyd2LCjl4tM7bIRljzGmxsXe6YEN5Ndf8+t+cM7w/v785myAbn8cY42U29k43Gp8Sy0NXjGH51gr+95MSb4djjDFus6TfRV+dPpjLMpL5+btbWbPD2veNMf7Bkn4XiQhPXjuBlDhn+36lte8bY/yAJf0z0NfVf7/iyHG+++paGzPEGOPzLOmfoQmpcTx42Rj+tWUfv/vkc2+HY4wxHbKk7wE3nz2EWeOS+dk7WyjYWentcIwxpl2W9D1ARPjZdRNIjo3g7kWFVNVY+74xxjdZ0veQ2MhQFs6bzL7DtXz31XXWvm+M8UmW9D1ooiOO+y8dw/ub9/L7Fda+b4zxPZb0Pexr5wzhkrED+Nk7WygqrfJ2OMYYcwpL+h4mIvzyuokkxURw16ICqmvqvR2SMcY0s6TfDWKjnP3391TXct9S679vjPEdlvS7SWZaPPdfOpp/btrLH//9hbfDMcYYwJJ+t/r6uelcPCaJn761mbXWvm+M8QGW9LuRiPDL613t+68UUH3M2veNMd5lSb+bxUWF8eyNmeyuquX+16z/vjHGuyzp94Apg+P53qxRvL1hD39eucPb4RhjApgl/R7yjXOHcuHoJB5/czPry6q9HY4xJkBZ0u8hQUHC/1w/kf7RYdy5qIBDtda+b4zpeZb0e1B8nzCen5dJedUxfvDaemvfN8b0OEv6PWzK4H5890ujeHP9bv5v9U5vh2OMCTCW9L3gv2cO5fxRiTz6901sKLf2fWNMz7Gk7wVBQcJTN0yiX58w7lpUwGFr3zfG9BC3kr6IzBKRrSJSLCL3t7H8WyKyXkSKRGSFiIx1zR8iIsdc84tE5AVPF8Bf9esTxnPzMimtPMYPXrf2fWNMz+g06YtIMLAQuBQYC9x4Iqm3sEhVM1R1EvBz4KkWy7ar6iTX51ueCrw3yB7Sj3suGck/1u1m0afWvm+M6X7uXOlPBYpVtURV64DFwJyWK6jqoRaTfQC7bHXT7ecNY+bIRH7y901s3GXt+8aY7uVO0k8BSltMl7nmnUJE7hSR7Tiv9Oe3WJQuIoUi8pGIzGjrACJym4jki0h+RUXFaYTv/5zt+xOJjwrlrkWFHDne4O2QjDG9mMdu5KrqQlUdBnwfeMg1ezeQpqqZwD3AIhHp28a2L6pqlqpmJSYmeiokv5EQHc6zczPZceAoD1j7vjGmG7mT9MsBR4vpVNe89iwGrgJQ1eOqesD1fQ2wHRjZtVB7t2lD+3PPJSNZtnYXi/NKO9/AGGO6wJ2knweMEJF0EQkD5gLLWq4gIiNaTF4OfOaan+i6EYyIDAVGACWeCLw3uuP84cwYkcDDyzayefehzjcwxpjT1GnSV9UG4C7gXWAzsERVN4rIIyIy27XaXSKyUUSKcDbj3OyaPxNY55q/FPiWqh70eCl6iaAg4Vc5k4iNDOXORQUctfZ9Y4yHia+1H2dlZWl+fr63w/CqldsPcNPvVjFnUgpP3TAREfF2SMYYHycia1Q1q7P17IlcH3TWsP4suGgkbxSW82p+mbfDMcb0Ipb0fdRdFw7nnOH9+dGyDWzdc9jb4RhjeglL+j4qOEh4OieT6PBQ7nh5jbXvG2M8wpK+D0uMCefZuZMo2X+UH/5tg7fDMcb0Apb0fdzZwxOYf+EIXi8o59V8679vjDkzlvT9wPyLRnDW0P788G8b2LbX2veNMV1nSd8PBAcJz8ydRHR4CHe+XEBNnbXvG2O6xpK+n0jqG8HTOZkUVxzhx3/b6O1wjDF+ypK+Hzl3RAJ3XzCcV9eU8doa679vjDl9lvT9zIKLRzItvR8P/XUDxfusfd8Yc3os6fuZ4CDh2RsziQoL5s6XCzlW1+jtkIwxfsSSvh8a0DeCX+VMYtu+wzy8zNr3jTHus6Tvp2aOTOSO84eRm1/KG4XWvm+McY8lfT/2nYtHMnVIPx58YwPF+454OxxjjB+wpO/HQoKDePbGTCJCg7lrUQG19da+b4zpWO9J+qqw/AnY+g7UBM57WpJjI3jqhols2XOYn/x9k7fDMcb4uBBvB+Ax1aXwyS+hyfW0asJIcEwFxzRwTIeEEdBLX0Zy/qgkbj9/GL/5cDvTh/ZjzqQUb4dkjPFRvSfpx6XB/aWwqxBKVzs/W96Ewv9zLo+Md50AXCeCQZMhLMq7MXvQvZeMJO/zgzzw+noyUmIZmhjt7ZCMMT6od78uURUOFMPOVSdPBPu3OZcFhUDyBOcJIG2a82/fQZ45rpfsqjrG5c9+QnJsJG/ccTYRocHeDskY00PcfV1i7076bak5CGV5rhPBp1C+BhqOOZfFpjlrAmnTnX+TxkGwf1WGlm/Zx61/zOOmaWk8fnWGt8MxxvQQd5O+f2U0T4jqByO/7PwANNbDnnXOE8DOVbDj/8GGpc5lYdGQMsXVLDQNUrMgMs57sbvhgtFJ/Pd5Q/ntRyVMH9qfKyf6d+3FGONZgXel3xlV503h0k+dzUE7V8HeDaBNgEDSmJMnAcdU6DfU524Q1zc2MffFVWzdc5h/3H0uQxL6eDskY0w3s+YdTzp+BMrzT54ISvPgeLVzWZ/EFieBaTBoEoSEezdeoLzqGJc98wmp8ZG8dru17xvT23k06YvILOAZIBj4nao+2Wr5t4A7gUbgCHCbqm5yLfsB8HXXsvmq+m5Hx/LJpN9aUxNUbDl5c7h0NRwscS4LDoNBma5eQq57A9FJXgnz/U17+caf8/nq9ME8etV4r8RgjOkZHkv6IhIMbAMuAcqAPODGE0ndtU5fVT3k+j4buENVZ4nIWOAVYCowCHgfGKmq7T466hdJvy1H9rWoCax2dh1trHMui08/eXPYMQ0Sx0BQzzwX9/ibm/jfTz7n1zdN5rKMgT1yTGNMz/PkjdypQLGqlrh2vBiYAzQn/RMJ36UPcOJMMgdYrKrHgc9FpNi1v5VulcKfRCfBmCucH4CG47Cr6ORJoPh9WPuKc1l4rPOm8IkTQcoUCI/plrC+N2s0+Tsq+f7SdYwb1JfB/a1935hA5k7STwFKW0yXAdNaryQidwL3AGHAhS22XdVq2/94XFREbgNuA0hLS3Mnbt8XEu7s/5/m+qdShcrPYeeJJqFPYflPAQUJggHjXc8MuE4EsQ6P3CAODQ7iuRszufzZFdy5qIDXbj+b8BBr3zcmUHmsy6aqLgQWisg84CHg5tPY9kXgRXA273gqJp8i4uzp028oTLrROe9YlfMG8YkTQdEiyPtf57KYgSdvDqdNcz5IFhzapUOnxkfxi+smcNtf1vDEW1t4ePY4DxXKGONv3En65YCjxXSqa157FgO/6eK2gSUyDoZf7PwANDbAvo0nnxko/RQ2/dW5LCQSUiaf2l00qp/bh/rSuGS+fm46v1/xOdPS+3Gpte8bE5DcSfp5wAgRSceZsOcC81quICIjVPUz1+TlwInvy4BFIvIUzhu5I4BPPRF4rxQcAgMnOj9Tv+mcd2jXyeag0tXw72fbGFRuuvNE0Mmgct93te9/77V1jBsUS1r/3jP2kDHGPe522bwMeBpnl82XVPVxEXkEyFfVZSLyDHAxUA9UAnep6kbXtg8CXwMagG+r6tsdHctve+/0lLoa16Byq06eCI5VOpedMqjcdGfX0VaDypUerOHyZz8hPaEPr37rbMJCes/o2sYEMns4K1Cowv7PXLUB14mg9aByzd1Fp0Pfgby7cQ///Zc13HrOEH58pbXvG9MbWNIPZDUHT31moHwNNNQ6l7kGlXuzejC/Lu7Pgnlz+FJGqnfjNcacMUv65qSGOti7vkV30dVweDcAR4kg2JFNRPpZzl5CqdkQEevlgI3pJVShvsbZU6+2ytkUe8z1t7bqP7/HD4YrftWlQ9kom+akkDDnA2ApU+CsO5oHldu/+RPee+dvTN1dzNCyXyJtDSqXNs35RLGPDSpnTI9qqDs1aXeawFusd+LJ/LZIEETEOe/HRcYBQ7q9KJb0A5EIxKWRcNZNxEVfyEUvF3DHWUl8L+Poye6iG16DNX9wru+jg8oZc1qaGqG2uu3E3JzAq9pO4PU1He87PBYiY08m8KQxpybzyHjXdMvv8c4n8Xv4gsqSfoC7NGMgN581mF+v3EHmiCwuOe9854LmQeVWnTwRbPmHc5kPDSpnAowqHD/cftJu9wq86uTIuO0JjTo1UccPaZG0405dFtEimYf39auXLVmbvuF4QyPX/ubflB48xpvzzyU1vp3++82DyrlOBO0OKjcdEkf32KByxg/VH2u/OaSzdu/2x2uEoNAWiTmu7e9tXnXH+X3t1W7kmtOy48BRrnh2BcMHRLPkv88iNNiNhF1fC7vXnjrE9NEK57KWg8oljHQOIREUAhIMQcHO70Ehru/BLZadmB908nvz/FZ/JdhOLN7UWN9+c0hnCbzxeAc7lo6TdkcJPDQqYO8/WdI3p+3Ndbu5c1EBt80cygOXjTn9Hag63yvQsrvovs2cHHS1O4gbJ5BWJxoJarXM9be9k85/nKza26b1cdrapoPjd3ri62xfJ2IKcj/xNTU5mz06S9otm0lOrFd3pON9h8W00zTSSQIP72sn8y6w3jvmtF0+YSCrSgbz4sclTEvvx0VjBpzeDkSg/zDnp+WgctWlzpto2uj829Tg+rSY1pbzm05+b57fatsu76vl/k7Mb4L6utM8ftOpx+6oycEbOjuBSBAcP+S8sdnRSTkk4tSr6TgHJGd0ftUdEdvlAQJN97Kkb07x4OVjWLOjkntfXctb82cwKC7yzHZ44kqvt1Nt46TT0YmqgxNIy79ntK/2ToaNzvXCYzq/6g6N8Pa/rPEwS/rmFBGhwSy8aTJXPreCu18pZPFt091r3w90Is4eHH7Ui8MEJvu/2fyH9IQ+PHFNBmt2VPLLf271djjGGA+ypG/adOXEQcyblsZvPyph+ZZ93g7HGOMhlvRNu350xVjGDOzLPUuK2F19zNvhGGM8wJK+aVdEaDAL52VS19DE/FcKaWhs8nZIxpgzZEnfdGhoYjQ/vSaDvC8qeeq9bd4Oxxhzhizpm07NmZTCjVMd/PrD7Xy41dr3jfFnlvSNW3585ThGJ8dwz5K17Kmu9XY4xpgusqRv3BIRGszz8yZTW9/I/MXWvm+Mv7Kkb9w2PCmax68ez6efH+Tp9z/zdjjGmC6wpG9Oy9WZqeRkOVj4YTEfb6vwdjjGmNNkSd+ctodnj2NkUgzfyS1i7yFr3zfGn1jSN6ctMiyYhTdlUlPXaP33jfEzbiV9EZklIltFpFhE7m9j+T0isklE1onIv0RkcItljSJS5Pos82TwxnuGJ8Xw2FXjWf35QZ79l7XvG+MvOk36IhIMLAQuBcYCN4rI2FarFQJZqjoBWAr8vMWyY6o6yfWZ7aG4jQ+4dkoq101J5bnlxaz4bL+3wzHGuMGdK/2pQLGqlqhqHbAYmNNyBVVdrqonXhe/Ckj1bJjGVz0yZxzDE6P5dm4R+w5b+74xvs6dpJ8ClLaYLnPNa8/XgbdbTEeISL6IrBKRq9raQERuc62TX1FhPUL8SVRYCAtvmsyR4/UseKWIxibfev2mMeZUHr2RKyJfAbKAX7SYPdj13sZ5wNMiMqz1dqr6oqpmqWpWYmKiJ0MyPWDkgBgenTOelSUHeO4Da983xpe5k/TLAUeL6VTXvFOIyMXAg8BsVW1+1b2qlrv+lgAfAplnEK/xUddnObhmcgrP/Osz/l1s7fvG+Cp3kn4eMEJE0kUkDJgLnNILR0Qygd/iTPj7WsyPF5Fw1/cE4Bxgk6eCN77l0TnjGZrQhwW5RVQcPt75BsaYHtdp0lfVBuAu4F1gM7BEVTeKyCMicqI3zi+AaODVVl0zxwD5IrIWWA48qaqW9HupPuEh/PqmKRyurec7uda+b4wvElXf+h8zKytL8/PzvR2GOQO5eTv5/mvrueeSkcy/aIS3wzEmIIjIGtf90w7ZE7nG427IcnB1ZgpPv7+NldsPeDscY0wLlvSNx4kIj101niEJfViwuJD9R6x93xhfYUnfdIs+4SEsnDeZ6mP13PzSp7y5bjfHGxq9HZYxAc+Svuk2Ywb25emcSVQerePORQWc9cQHPPqPTWzbe9jboRkTsOxGrul2jU3KiuL95Obt5L1Ne6lvVCanxZGT7eCKCYPoEx7i7RCN8Xvu3si1pG961IEjx3mjsJzFeaUU7ztCn7Bgrpw4iBuyHWQ64hARb4dojF+ypG98mqpSsLOS3LxS/r52N8fqGxk5IJqc7DSuzkyhX58wb4dojF+xpG/8xuHaev6xbje5eaUUlVYRFhzEJeMGMDfbwTnDEggKsqt/YzpjSd/4pS17DpGbV8obheVU1dSTEhfJDVkOrs9KZVBcpLfDM8ZnWdI3fq22vpH3Nu0lN6+UFcX7EYHzRiaSk+XgojEDCAuxjmfGtGRJ3/QapQdreDW/lCX5Zew5VEv/PmFcOyWVG7IcDE+K9nZ4xvgES/qm12lsUj7eVkFuXinvb95LQ5OSNTienGwHl08YSFSYdf00gcuSvunVKg4f5/WCMnLzSinZf5To8BCunDiIudkOJqTGWtdPE3As6ZuAoKrk76hk8aelvLl+F7X1TYxOjiEn2znoW1yUdf00gcGSvgk4h2rr+fvaXeTmlbKurJqwkCBmjUsmJ9vBWUP7W9dP06tZ0jcBbeOuapa4un4eqm3A0S+SnCwH101xkBwb4e3wjPE4S/rG4Oz6+e7GPSz+tJSVJQcIEjh/VBI52Q4uHJ1EaLB1/TS9g7tJ37o7mF4tIjSYOZNSmDMphR0HjrIkv5RX88v4YMs+EqLDuXZKCjlZDoYmWtdPExjsSt8EnIbGJj7aVsHivFI+2LKPxiZl6pB+5GQ7uCxjIJFhwd4O0ZjTZs07xrhh36FalhaUsSSvlC8O1BATHsKczEHkZKUxPqWvdf00fsOSvjGnQVVZ/flBcvNKeWv9bo43NDF2YF/mTnUwZ2IKsVGh3g7RmA48kz/5AAAOE0lEQVRZ0jemi6qP1bOsyDnm/8ZdhwgPCeLS8cnkZKcxfWg/u/o3PsmSvjEesKG8mty8Uv5aVM7h2gYG94/ihiwH101JZUBf6/ppfIdHk76IzAKeAYKB36nqk62W3wN8A2gAKoCvqeoO17KbgYdcqz6mqn/q6FiW9I0vOlbXyNsbnGP+r/78IMFBwgWjEsnJTuOCUYmEWNdP42UeS/oiEgxsAy4ByoA84EZV3dRinQuA1apaIyK3A+erao6I9APygSxAgTXAFFWtbO94lvSNryupOMKS/DKWrilj/5HjJMWEc+2UVHKyHAxJ6OPt8EyAcjfpu3N5MhUoVtUSVa0DFgNzWq6gqstVtcY1uQpIdX3/MvCeqh50Jfr3gFnuFsIYXzQ0MZr7Lx3Nyh9cyItfnUJGSiy//Wg75//yQ+a+uJK/FpZTW9/o7TCNaZM7D2elAKUtpsuAaR2s/3Xg7Q62TWm9gYjcBtwGkJaW5kZIxnhfaHAQXxqXzJfGJbOnupbXXKN+fju3iL5/C+GqzBRysh2MGxTr7VCNaebRJ3JF5Cs4m3LOO53tVPVF4EVwNu94MiZjekJybAR3XjCc288bxqrPD5CbV8rivFL+vHIH41P6kpOdxuyJg4iNtK6fxrvcSfrlgKPFdKpr3ilE5GLgQeA8VT3eYtvzW237YVcCNcYfBAUJZw9L4OxhCfykpo6/Fjq7fv7wrxt4/M1NXDZ+IDnZDqamW9dP4x3u3MgNwXkj9yKcSTwPmKeqG1uskwksBWap6mct5vfDefN2smtWAc4buQfbO57dyDW9jaqyvryaxXmlLCvaxZHjDQxN6MP1WQ6unZJCUox1/TRnztNdNi8DnsbZZfMlVX1cRB4B8lV1mYi8D2QAu12b7FTV2a5tvwY84Jr/uKr+oaNjWdI3vVlNXQNvrd9Dbt5O8r6oJDhIuGh0EnOnOpg5wrp+mq6zh7OM8XHF+47wan4pS9eUceBoHQP6hnP9FAc3ZDlI6x/l7fCMn7Gkb4yfqGto4oMte8nNK+WjbRU0KZw9rD852Q6+PC6ZiFAb9dN0zpK+MX5oV9Uxlq4pY0l+KWWVx4iNDOVqV9fPMQP7ejs848Ms6Rvjx5qalH9vP8DivJ38c+Ne6hqbmJgaS052GldOHEhMhHX9NKeypG9ML1F5tI43CsvJzStl697DRIYGc/mEgczNdjBlcLx1/TSAJX1jeh1Vpai0iiX5zq6fR+saGZrYh7nZDq6ZnEpCdLi3QzReZEnfmF7s6PEG3ly3m8V5OynYWUVIkHDxmAHkuLp+BgfZ1X+gsaRvTID4bO9hcvNKeb2wnINH6xgYG8H1U1K5PsuBo591/QwUlvSNCTB1DU28v3kvi/NK+eSzCgDOHZ5ATraDS8YOIDzEun72Zpb0jQlgZZU1LF1Txqv5ZZRXHSM+KpSrM1PJyXYwKjnG2+GZbmBJ3xhDY5Oyong/S/JK+eemPdQ3KpMcceRkO5gxIoGUuEjr/dNLWNI3xpziwJHjvOEa9bN43xEAkmLCyUyLY3JaPJMHx5OREmtPAPspS/rGmDapKht3HaJgZyUFOyop2FnFzoPOF9+FBAljB/Vlclp888kgNd5qA/7Akr4xxm0Vh49TVFrVfCJYV1bNMdcrHxOiw5mcFsfkwfFkOuKYkBpHZJjVBnyNu0nfo2/OMsb4p8SYcC4ZO4BLxg4AoKGxiS17DlO401kTKNxZyT837QWctYExA/uebBZKi8fRz2oD/sKu9I0xbjlw5DiFO6soLK2kYEcVa8uqqKk7URsIY5IjnsmDnSeCCamxRIXZNWVPsit9Y4xH9Y8O5+KxA7i4RW1g294jziahnZUU7qzi/c3O2kBwkDA6OeaUewOD+0dZbcAH2JW+McZjKo/WNdcECksrKdpZxVFXbaB/nzAy0+LIdJ0IJqbG0Sfcrjs9xa70jTE9Lr5PGBeOHsCFo521gcYmZdvewxTurGquEby/eR8AQQKjkvs6bxK7TgTpCX2sNtDN7ErfGNOjqmrqKCytotDVXbSotIojxxsAiI8KddYEHM7eQhMdcURbbcAtdqVvjPFJcVFhXDAqiQtGJQHO2kDxviOu+wLOE8EHW07WBkYOiCEzLb652+hQqw2cEbvSN8b4nOqaegpLK5ubhYpKqzhc66wNxEaGntJddKIj1t4khl3pG2P8WGxUKOePSuJ8V22gqUnZXuHqKeS6SfzRtgpUQQRGJsUweXAcma5uo0MTogmydwq0ya70jTF+qfpYPWtdTxEXuh4gO+SqDfSNCGHSiSahtHgmpcXRt5fXBjx6pS8is4BngGDgd6r6ZKvlM4GngQnAXFVd2mJZI7DeNblTVWe7VwRjjGlfbGQoM0cmMnNkIuCsDZTsP9L8BHHBjiqe+ddnzbWB4YnRroHlnN1GhycGZm2g0yt9EQkGtgGXAGVAHnCjqm5qsc4QoC/wXWBZq6R/RFWj3Q3IrvSNMZ5yuLaetaXVzTeJC0urqKqpByAmIoRJjpPdRTMd8cRG+W9twJNX+lOBYlUtce14MTAHaE76qvqFa1lTl6I1xphuEBMRyrkjEjh3RALgHGG0ZP9RCnY4TwAFOyp57oPPaHJd+w5Pim7uLjo5LZ7hSdG97n3D7iT9FKC0xXQZMO00jhEhIvlAA/Ckqv619QoichtwG0BaWtpp7NoYY9wnIgxLjGZYYjTXZzkAOHK8gbWlVc3dRd/fvJdX15QBEBMewkRHHJPT4sh0jTIaFxXmzSKcsZ7ovTNYVctFZCjwgYisV9XtLVdQ1ReBF8HZvNMDMRljDADR4SGcMzyBc4afrA18caDG9a4B503i55cXN9cGhib2OWVMoZEDYvyqNuBO0i8HHC2mU13z3KKq5a6/JSLyIZAJbO9wI2OM8RIRIT2hD+kJfbh2SioAR483sLasqrmX0Adb9rHUVRvoExbsqg3EN3cbje/ju7UBd5J+HjBCRNJxJvu5wDx3di4i8UCNqh4XkQTgHODnXQ3WGGO8oU94CGcPS+DsYSdrAzsO1DQPLlews5LffLSdRld1ID2hT3NNIDMtjlEDYggJDvJmEZq51U9fRC7D2SUzGHhJVR8XkUeAfFVdJiLZwBtAPFAL7FHVcSJyNvBboAkIAp5W1d93dCzrvWOM8Uc1dQ2sK6s+5bmB/UfqAIgKC2ZialxzTSAzLY7+0eEePb69LtEYY7xIVSk9eMxVG3DeJN68+xANrtrAkP5RzWMKZabFMzr5zGoDNgyDMcZ4kYiQ1j+KtP5RzJmUAsCxukbWl1c3v4v4k8/280ah8xZpZGgwF41J4vl5k7s1Lkv6xhjTQyLDgpma3o+p6f0AZ22grPJYc5NQVA+8cN6SvjHGeImI4OgXhaPfydpAd/ON28nGGGN6hCV9Y4wJIJb0jTEmgFjSN8aYAGJJ3xhjAoglfWOMCSCW9I0xJoBY0jfGmADic2PviEgFsOMMdpEA7PdQON7UW8oBVhZf1VvK0lvKAWdWlsGqmtjZSj6X9M+UiOS7M+iQr+st5QAri6/qLWXpLeWAnimLNe8YY0wAsaRvjDEBpDcm/Re9HYCH9JZygJXFV/WWsvSWckAPlKXXtekbY4xpX2+80jfGGNMOv0z6IjJLRLaKSLGI3N/G8nARyXUtXy0iQ3o+Sve4UZZbRKRCRIpcn294I87OiMhLIrJPRDa0s1xE5FlXOdeJSPe+HugMuFGW80WkusVv8qOejtEdIuIQkeUisklENorIgjbW8Yvfxc2y+MvvEiEin4rIWldZftLGOt2Xw1TVrz44X86+HRgKhAFrgbGt1rkDeMH1fS6Q6+24z6AstwDPeztWN8oyE5gMbGhn+WXA24AA04HV3o75DMpyPvAPb8fpRjkGApNd32OAbW389+UXv4ubZfGX30WAaNf3UGA1ML3VOt2Ww/zxSn8qUKyqJapaBywG5rRaZw7wJ9f3pcBFIiI9GKO73CmLX1DVj4GDHawyB/izOq0C4kRkYM9Ed3rcKItfUNXdqlrg+n4Y2Ay0fj2TX/wubpbFL7j+rY+4JkNdn9Y3V7sth/lj0k8BSltMl/GfP37zOqraAFQD/XskutPjTlkArnVVvZeKiKNnQvM4d8vqL85yVc/fFpFx3g6mM67mgUycV5Ut+d3v0kFZwE9+FxEJFpEiYB/wnqq2+7t4Oof5Y9IPNH8HhqjqBOA9Tp79jfcU4HzkfSLwHPBXL8fTIRGJBl4Dvq2qh7wdz5nopCx+87uoaqOqTgJSgakiMr6nju2PSb8caHm1m+qa1+Y6IhICxAIHeiS609NpWVT1gKoed03+DpjSQ7F5mju/m19Q1UMnqueq+hYQKiIJXg6rTSISijNJvqyqr7exit/8Lp2VxZ9+lxNUtQpYDsxqtajbcpg/Jv08YISIpItIGM6bHMtarbMMuNn1/TrgA3XdEfExnZalVfvqbJxtmf5oGfBfrt4i04FqVd3t7aC6QkSST7SvishUnP8f+dxFhSvG3wObVfWpdlbzi9/FnbL40e+SKCJxru+RwCXAllardVsOC/HETnqSqjaIyF3Auzh7v7ykqhtF5BEgX1WX4fyP4y8iUozzhtxc70XcPjfLMl9EZgMNOMtyi9cC7oCIvIKz90SCiJQBP8Z5gwpVfQF4C2dPkWKgBrjVO5F2zo2yXAfcLiINwDFgro9eVJwDfBVY72o/BngASAO/+13cKYu//C4DgT+JSDDOE9MSVf1HT+UweyLXGGMCiD827xhjjOkiS/rGGBNALOkbY0wAsaRvjDEBxJK+McYEEEv6xhgTQCzpG2NMALGkb4wxAeT/A3Neew2MaHRMAAAAAElFTkSuQmCC\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",
    "%matplotlib inline\n",
    "\n",
    "plt.plot(history.history['loss'],label='loss')\n",
    "plt.plot(history.history['val_loss'],label='val_loss')  #two graph in same scale \n",
    "plt.legend()\n",
    "plt.show()\n",
    "plt.close()\n",
    "\n",
    "plt.plot(history.history['acc'])\n",
    "plt.plot(history.history['val_acc'],label='val_acc')\n",
    "plt.legend()\n",
    "plt.show()\n",
    "plt.close()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# We can also plot the graphs with subplots\n",
    "plt.figure(figsize=(20, 5))\n",
    "\n",
    "plt.subplot(1, 2, 1)\n",
    "plt.plot(history.history['acc'])\n",
    "plt.plot(history.history['val_acc'],label='val_acc')\n",
    "plt.legend()\n",
    "plt.title(\"Accuracy\")\n",
    "\n",
    "plt.subplot(1, 2, 2)\n",
    "plt.plot(history.history['loss'],label='loss')\n",
    "plt.plot(history.history['val_loss'],label='val_loss')  #two graph in same scale \n",
    "plt.legend()\n",
    "plt.title(\"Loss\")\n",
    "\n",
    "plt.show()\n",
    "plt.close()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}