NN Ising I - dataset from others.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Using NN to study the Ising model phase transition"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Notebook based on the paper:\n",
    "\n",
    "**Machine learning phases of matter.** ([Link](http://arxiv.org/abs/1605.01735%0Ahttp://dx.doi.org/10.1038/nphys4035))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import pickle,os\n",
    "import tensorflow as tf\n",
    "from tensorflow import keras\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense\n",
    "from keras.layers import Dropout\n",
    "from keras import regularizers\n",
    "from keras import optimizers\n",
    "import sys, os\n",
    "from sklearn.model_selection import train_test_split\n",
    "from keras.utils import to_categorical\n",
    "os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "plt.rcParams['figure.figsize'] = [6.32, 5.11]\n",
    "plt.rcParams['figure.constrained_layout.use'] = True\n",
    "\n",
    "plt.rcParams['axes.titlesize'] = 20\n",
    "plt.rcParams['axes.labelsize'] = 20\n",
    "plt.rcParams['axes.linewidth'] = 3\n",
    "plt.rcParams['axes.labelpad'] = 15\n",
    "\n",
    "plt.rcParams['xtick.labelsize'] = 16\n",
    "plt.rcParams['xtick.major.size'] = 8\n",
    "plt.rcParams['xtick.major.width'] = 3\n",
    "\n",
    "plt.rcParams['ytick.labelsize'] = 16\n",
    "plt.rcParams['ytick.major.size'] = 8\n",
    "plt.rcParams['ytick.major.width'] = 3\n",
    "\n",
    "plt.rcParams['legend.fontsize'] = 14\n",
    "plt.rcParams['legend.markerscale'] = 1\n",
    "\n",
    "plt.rcParams['lines.markersize'] = 9\n",
    "plt.rcParams['lines.linewidth'] = 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "path = '../Data/'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "L = 40\n",
    "temp_val = np.arange(0.25, 4.0001, 0.25)\n",
    "temp_samples = 10000\n",
    "total_samples = temp_samples * temp_val.shape[0]\n",
    "test_size = 0.1\n",
    "val_size = 0.2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "split_tc = True"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = pickle.load(open(path + 'Ising2DFM_reSample_L40_T=All.pkl','rb'))\n",
    "data = np.unpackbits(data).reshape(-1, L * L).astype(int)\n",
    "data[np.where(data==0)]=-1\n",
    "\n",
    "labels = pickle.load(open(path + 'Ising2DFM_reSample_L40_T=All_labels.pkl','rb'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "first_tc = np.where(temp_val <= 2.)[0][-1] * temp_samples\n",
    "last_tc = np.where(temp_val >= 3)[0][1] * temp_samples"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "temp_labels = np.hstack([np.repeat(temp, temp_samples) for temp in temp_val])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "if split_tc:\n",
    "    \n",
    "    X_ordered = data[:first_tc, :]\n",
    "    Y_ordered = np.column_stack((labels[:first_tc], temp_labels[:first_tc]))\n",
    "\n",
    "    X_critical = data[first_tc:last_tc, :]\n",
    "    Y_critical = np.column_stack((labels[first_tc:last_tc], temp_labels[first_tc:last_tc]))\n",
    "\n",
    "    X_disordered = data[last_tc:, :]\n",
    "    Y_disordered = np.column_stack((labels[last_tc:], temp_labels[last_tc:]))\n",
    "\n",
    "    X = np.concatenate((X_ordered, X_disordered))\n",
    "    Y = np.concatenate((Y_ordered, Y_disordered))\n",
    "    \n",
    "    del X_ordered, Y_ordered, X_disordered, Y_disordered, data, labels\n",
    "\n",
    "else:\n",
    "    \n",
    "    X = data\n",
    "    Y = np.column_stack((labels, temp_labels))\n",
    "    \n",
    "    del data, labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=test_size)\n",
    "\n",
    "X_train, X_val, Y_train, Y_val = train_test_split(X_train, Y_train, test_size=val_size)\n",
    "\n",
    "del X, Y"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From /home/ronaldo/anaconda3/envs/data/lib/python3.7/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"
     ]
    }
   ],
   "source": [
    "model = Sequential()\n",
    "model.add(Dense(100, input_shape=(L * L, ), activation='relu', kernel_regularizer=regularizers.l2(0.01)))\n",
    "model.add(Dropout(0.15))\n",
    "model.add(Dense(1, activation='sigmoid'))\n",
    "\n",
    "model.compile(optimizer='adam',\n",
    "              loss='binary_crossentropy',\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense_1 (Dense)              (None, 100)               160100    \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 100)               0         \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 1)                 101       \n",
      "=================================================================\n",
      "Total params: 160,201\n",
      "Trainable params: 160,201\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From /home/ronaldo/anaconda3/envs/data/lib/python3.7/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",
      "Train on 79200 samples, validate on 19800 samples\n",
      "Epoch 1/5\n",
      "79200/79200 [==============================] - 3s 39us/step - loss: 0.2259 - accuracy: 0.9891 - val_loss: 0.0156 - val_accuracy: 0.9996\n",
      "Epoch 2/5\n",
      "79200/79200 [==============================] - 3s 36us/step - loss: 0.0171 - accuracy: 0.9994 - val_loss: 0.0145 - val_accuracy: 0.9996\n",
      "Epoch 3/5\n",
      "79200/79200 [==============================] - 3s 36us/step - loss: 0.0183 - accuracy: 0.9994 - val_loss: 0.0091 - val_accuracy: 0.9994\n",
      "Epoch 4/5\n",
      "79200/79200 [==============================] - 3s 36us/step - loss: 0.0284 - accuracy: 0.9988 - val_loss: 0.0293 - val_accuracy: 0.9991\n",
      "Epoch 5/5\n",
      "79200/79200 [==============================] - 3s 36us/step - loss: 0.0258 - accuracy: 0.9989 - val_loss: 0.0121 - val_accuracy: 0.9995\n"
     ]
    }
   ],
   "source": [
    "history = model.fit(X_train, Y_train[:, 0], validation_data=(X_val, Y_val[:, 0]) , batch_size=64, epochs=5)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "score = model.evaluate(X_test, Y_test[:, 0], verbose=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test accuracy: 0.9995 \n",
      "Test loss: 0.0119\n"
     ]
    }
   ],
   "source": [
    "print('Test accuracy: {1:6.4f} \\nTest loss: {0:6.4f}'.format(score[0], score[1]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAc8AAAF4CAYAAAAyt3LmAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8li6FKAAAgAElEQVR4nOzdeZgU1b3/8fd39oVhWIZhEJEdWQTcN1RQwR3USKJR45JozHKjMTG/aOJN4MZEc7PoTUyMiVETTeKSxBg0LgiigoJLIsgOyrDDAMM++8z5/VHVTE/Ts/TQQ3VPf17P009PV52qOtXofOacOnXKnHOIiIhI26UFXQEREZFko/AUERGJkcJTREQkRgpPERGRGCk8RUREYqTwFBERiZHCUySBmdkEM3NmNu0Q93ODv58b4lMzkdSm8BQJ4weMM7MGMxvcQrnXw8recBirKCIJQOEpcrA6wIAvRFtpZkOB8X45EUlBCk+Rg20F3gduNLOMKOtvwgvXFw5rrUQkYSg8RaL7HVACXBK+0MwygeuBt4ElzW1sZkPN7I9mttHMasxsk/95aDPle5vZ781sq5lVmtmHZnZ9SxU0sx5mdq+ZLfO32W1ms8zsvJjP9uB9H2Fm3zOzeWa2Jewc/mxmI1rY7mQze9o/72oz22xmr5rZZ9pTtrVrvmZWamalEcsOXN81swvMbI7/3biwMpeZ2ZNmttLM9pvZPjP7wMxuNbOovxfNLM/Mvm1m75vZXn+bZWb2CzPr7Zd5yj/2Wc3sY6q//pfNfYeSHKL9VS0i8Bfg53itzH+ELZ8C9AbuBIZE29DMTgJeAwqAfwJLgeHANcClZnauc+79sPI98cJ4EDDXf/UBfgO82swx+gNzgAHAW8DLQD5e2L9sZrc4534X+2kfcJZ/jq8DfwP2AUOBqcAUMxvnnFsYUaebgYeAev+8VwHFwInAV4Bn2lP2EEwFLgBewvsuB4Stuw9oABYAG4FC4Bzg/4CTgM9FnFt3vO9iLLACeBSoAQYDnwf+jtdj8WvgSuAW4M0odfqi//7bQzw3CZpzTi+99PJfgAM2+D8/gndd88iw9S8Du4E84B6//A1h6w1Y5i+/JmLfV/rLlwNpYct/6y+/P6L8iUCtv25axLo5eL/8r4pY3g34EKgEeoctvyGyrq18D8VAQZTlY/GC9KWI5SP9upYDo6Jsd2Q7y06Idv5h60uB0ohloXNtAC5oZrvBUZalAX/wtz0lYt2f/eUPhf/b+esKgMKwz4uBKqAootxAv07zgv7vXK9Df6nbVqR5vwPS8VoWodbeJOBPzrmKZrY5Ha+V+Y5z7k/hK5xzT+O1Ko8GzvD3mYnXIt0LTIso/z7QZB/+NmPxBiz9zTn3VMQ2u4DvAznAFW0/1aacc2XOub1Rli8EZgNn+3UP+TJeT9YPnHMHdWc75za0s+yheN4593K0Fc65j6Msa8BreQKcH1puZsV4f/hsBu7wy4Vvt9c5tzts0UNANl73frgv4v1x9XCM5yEJSN22Is1wzi0ws4+Az5vZPXhduGl4odqc4/332c2sn40XnMfhdesNx2vFvhXxCzhkDgf/Ej7Nfy9s5lpgL/+92WuTbWFmFwNfwmsBF3Hw74sivEABONV/f6kNu46l7KF4t7kVflf5t4CL8LrL8yOK9A37+SS8f/c3nXP723DcP+J1C38R+Jl/vEy8FvFO4tMlLQFTeIq07HfAL/Cund0IfOCc+08L5Qv9983NrA8t7xZRfmsz5bdEWdbTf5/kv5rTpYV1LTKzW/FaYTuBmcA6oAKv6/IyvO7b7LBNQuezsQ27j6XsoYj23WFm3YD38LpR38ULu3K8LvpuwG20/9xwzu01syeBL5nZ2c6514FL8QagPeCcq2rHuUiCUXiKtOwJ4Md4XW19gf9ppXyo9VjSzPo+EeVC772bKR9tP6FtbnPO/aKV+sTMvz1nOl74HO+c2xyx/rQom+3y3/viXdNtSSxlQ12kzf2uKqTx+4jkmll+E15wTnfOTQtf4Z/bbRHlw+vbVg/htdpvwRtopIFCnYyueYq0wL+G+FfgSGA/3ijcloRapROaWR9a/m//fTlei+5YMytsoXy4+f77ma3Upb2K8Fpbb0cJzi40dk1Hq9OFbdh/LGV3+u/9IleY2RAaW4WxCI2S/luUdeOjLHsXL8TPMrPI7t2onHOLgHnA5WZ2CjARr9t3WTvqKwlI4SnSuruBy4Hzow2iiTAP71aGM8xsavgK//NZwEq8gUM452rxBgUVEDFgyMxOxBtM1IQ/kOgt4FNm9vlolTCz0f5Al/Yowwv0E/ywDO0zE68rtyjKNg/hdXv+t5mNjFKfI9tZdjmwB+8Wn+KwMrl43entUeq/T4g47nHAXZGFnXPbgKfweg1+GnkfqJl1aeYPn4eALLyQNrzbZaSTULetSCucc+vwrvm1pazzJzeYCTxtZs/jBcDReNcK9wLXRYzY/A5wLvB1PzBD93leCfwL797SSFfjDT76vX99cgFe9+KRwBjgGLyBRWWxna036tTMfoF3n+dH/jlkAWcDPfC6Ic+O2GapmX0FLyD+42+zCu/67In+eZ/djrK1ZvZ/wH/7ZZ/D+701Cdjkv2L1R7zBQg+Y2dn+sYfi3SP7d7zvPdJ/4X2nXwImmNkrePd5DsQbmTsFb3BXuGeB+/G6e7f7+5bOIuh7ZfTSK5FehN3n2YayB93nGbbuaLzrpZvx7mncDDwJHN3MvkrwbrzfhneP5od4ozMn0Mx9jnit1e8AH+Dde1kJrAFexLvGlh9W9obm6tpMfTKAb+BN8FCJd/3zCaA/8Li/rwFRtjsNr6VVhhcum/DujZ3a3rJ4rbY7gY/9cuuA/8UbpVxK8/d5NnuuePea/tM/9n7/O7wJbyIFBzweZZt84LvAIryW+V7/+3kAKG7mOPf7+/tJ0P9t6xXfl/n/wCIiEmdmNgevq/5o59yqgKsjcaRrniIiHcDMTsYbgPSKgrPz0TVPEZE4MrMv413nvBFvlO73g62RdAR124qIxJH/lJcjgU/wrlX/OdgaSUdQeIqIiMRI3bYdKPz5gSIikpyccxa5TAOGREREYqTwFBERiZG6bQ8TXVsWEUkeZgf11DahlqeIiEiMFJ4iIiIxUniKiIjESOEpIiISIw0YEhFpwZ49eygrK6O2tjboqkgcZWZmUlxcTNeuXdu1vcJTRKQZe/bsYevWrfTt25fc3NxWR2BKcnDOUVlZycaNGwHaFaDqthURaUZZWRl9+/YlLy9PwdmJmBl5eXn07duXsrKYnxcPKDxFRJpVW1tLbm5u0NWQDpKbm9vu7niFZwKrrW/g9RVlPPv++qCrIpKy1OLsvA7l31bXPBPYx9v2ceNj79EtL5PLjutLZrr+1hERSQT6bZzAju5dwNDiLuyqqGXu6u1BV0dEhKuuuoqpU6cGXY3AKTwTmJlxyZgjAHhh4eaAayMiycDMWnzdcMMNh7T/hx9+mEceeSQ+lU1i6rZNcJeM7cP9r63k1aVbqK47huyM9KCrJCIJbPPmxj+0X3jhBW6++eYmy5obAFVbW0tmZmar+y8sLDz0SnYCankmuMG9ujCiT1f2VtXx5kp13YpIy0pKSg68unXrdtCywsJCli9fjpnx7LPPMn78eHJycvjDH/7A1q1bufLKKw/cnnPMMcfwpz/9qcn+I7ttTz31VG6//Xa+9a1v0aNHD0pKSvjOd77T6Z8kpfBMApeM6QPAC4s2BVwTEelM7rzzTm6//XaWLVvGRRddRGVlJaeeeiovvvgiixcv5stf/jLXX389c+fObXE/jz76KIWFhSxYsICf/exn/PjHP+Yf//jHYTqLYKjbNglMHnMEP3llBTOXbqWypp7cLHXdigRlwJ0vBnLc0vsujvs+v/GNb3DZZZc1WXb77bcf+PmrX/0qM2fO5KmnnuKMM85odj/HH388d999NwBDhw7lN7/5DbNmzeLyyy+Pe50ThVqeSeConnmMPbKQipp6Xl/RvtkwREQinXjiiU0+19XVMX36dEaPHk2PHj3o0qULL774IuvWrWtxP2PGjGny+Ygjjmj3zD3JQi3PJHHJmCNYuGE3LyzaxEWj+wRdHZGU1REtwKDk5+c3+fzDH/6QX/3qVzzwwAOMGjWK/Px8vvnNb1JdXd3ifiIHGpkZ9fX1ca9vIlHLM0lc7F/3nL28jH3VdQHXRkQ6o7lz53L55Zdz9dVXM3bsWAYNGsTKlSuDrlZCUngmiSO65XJi/+5U1TYwa9nWoKsjIp3QsGHDeOWVV3jnnXdYtmwZt9xyC5s2aaBiNArPJBIadTtDEyaISAeYPn06Y8aMYdKkSUyYMIHi4mLNJtQM6+z34gTJzA58ufH4nsv2VHHKvbPITEvjvbsnUpjb+g3NItJ+y5YtY8SIEUFXQzpQc//G4ZPGO+cOmkFeLc8kUtw1h1MG9qCmvoGZS9V1KyISFIVnkjkw160mTBARCYzCM8lceEwJ6WnG3FXb2bm/JujqiIikJIVnkunZJZvTB/ekrsHx8pItQVdHRCQlKTyT0GR13YqIBErhmYTOH1VCZrrxzsc72La35Zk/REQk/hSeSagwL5OzhvaiwcHLi3XPp4jI4abwTFKXjNWECSIiQVF4JqmJI3qTlZHGe2vL2bK7KujqiIikFIVnkirIyeTso3vhHLz4kVqfIhJfjzzyCN26dWv2czT33XcfQ4YMifuxE5HCM4lpwgQRiTR58mQmTpwYdd2yZcswM2bOnBnzfq+55pq4P2Glrq4OM+Mf//hHhx8r3hSeSezcEcXkZqbzn3W7WF9eEXR1RCQB3HTTTcyePZvS0tKD1v3+97+nf//+nHvuuTHvNzc3l+Li4jjUMLGO1V5JFZ5m1s/M/mpmu81sj5n93cyOasN2J5rZb81suZlVmNk6M/uTmQ2MUjbNzO4ys1IzqzKzhWZ2Rcec0aHJy8rg3BHef2DquhURgIsvvpjevXvz2GOPNVleW1vLE088wec//3nS0tK44447GDZsGLm5uQwcOJA777yzxYdeR+tKvffee+nduzcFBQXccMMNVFQ0/SN+wYIFTJo0iaKiIgoLCznzzDN59913D6wfMGAAAJdffjlmdqDLN9qxfv3rXzN48GCysrIYOnQojz766IF1oRbsI488whVXXEF+fj6DBw/mL3/5S9u/uBglTXiaWR4wGxgOXA98DhgKvG5m+S1tC1wFjAJ+AVwI3AkcD7xvZv0iyv4AmAY86JedDzxrZhfF50ziS123IhIuIyOD66+/nscff5yGhoYDy2fMmMH27du58cYbAejatSuPP/44y5Yt48EHH+TJJ5/kvvvua/Nx/vznPzNt2jTuuecePvjgAwYNGsQDDzzQpMzevXu5/vrreeutt5g/fz6jR4/mwgsvZOfOnQC89957ADz22GNs3ryZ+fPnRz3Ws88+y9e//nW++c1vsnjxYr761a/yxS9+kZdeeqlJuenTp3PFFVewcOFCPvWpT3HDDTewYcOGNp9TLJLmkWRmdhvwc+Bo59xqf9lAYBXw/5xzP29h217OuW0Ry/oDa4B7nHPf85cVA+uB+5xz3w8rOwvo5ZwbE2Od4/pIsmiqaus58Z7X2Fddx5w7JjCgqLW/I0SkraI+rmpaYTCVmba7zUVXrVp14MHW5513HuC1SBsaGg4KnJAHH3yQBx98kOXLlwNe6++OO+5g165dUT+ffPLJnHDCCTz00EMH9jFhwgQ2bNjA6tWrox7DOUdxcTG//OUvueqqq6irqyMzM5PnnnuOyy677EC5yGOdcsopjB07lt/+9rcHylx77bVs2LCBOXPmHNjP3XffzQ9+8AMAampqKCgo4A9/+ANXXXVVs99VKjySbAowPxScAM65NcA84NKWNowMTn/ZWmAb0Dds8flAFvBkRPEngdHRunmDlpOZzqSRvQG1PkXEM3ToUM4666wDXZubNm3ilVde4aabbjpQ5umnn2bcuHGUlJTQpUsX7rjjDtatW9fmYyxbtozTTjutybLIz1u3buWLX/wiw4YNo7CwkIKCAnbs2BHTcULHGjduXJNlZ5xxBkuXLm2ybMyYxvZNVlYWRUVFlJWVxXSstsrokL12jFHA81GWLwE+HevOzGwEUAwsizhGNRD5Z9MS/30kXms1oUwe24fn/rORGQs381/nDA26OiKdWwwtwCDddNNN3HzzzZSXl/P444/To0cPpkyZAsDcuXO55pprmD59Oueddx7dunXjueee4zvf+U5c63Dttdeya9cuHnjgAfr37092djYTJkygpib2J0KFtwSbW5aZmXnQ+vCu63hKppZnD2BnlOXlQPdYdmRmGcBv8Fqev484xi53cB9redj6yH1NMzMX7RVLnQ7FGUN6UZibyYqte1m1de/hOqyIJLCpU6eSk5PDk08+yaOPPsp11113IFzmzZtH//79+e53v8tJJ53E0KFDo47ObcmIESMOukYZ+Xnu3LnceuutXHTRRYwaNYr8/Hy2bGl8GlR6ejrp6enU19e3eqy5c+cetO+RI0fGVOd4SqaWJ0C0QDr4z5HWPQicDlzsnAsPZIvjMQ6brIw0zh/Vm2fe38CMRZv5xqSCoKskIgHLzc3l6quvZtq0aezcuZMvfOELB9YNGzaMdevW8Ze//IWTTz6Zl156iWeeeSam/d9222184Qtf4IQTTuDMM8/kmWee4YMPPmhyi8mwYcN44oknOPHEE9m7dy/f+ta3yM7OPrDezDjqqKOYNWsW48aNIzs7m+7dD24Lfetb3+Lqq6/muOOOY+LEibz44os89dRTzJgxox3fTHwkU8tzJ1Fafnitzmgt0qjM7F7gi8DnnXOvRqwuB7rbwf0D3cPWJ6TwUbfJMghMRDrWTTfdxM6dOzn99NObDIq5/PLLuf3227n11ls59thjmTNnDtOnT49p39dccw133303d911F8cffzzLly/ntttua1Lm8ccfZ9euXRx33HFcffXV3HLLLfTr1/QGh5///OfMnDmTfv36cdJJJ0U91tSpU3nggQf46U9/yqhRo/jVr37Fww8/zIUXXhhTneMpmUbbzgaynHNnRCyfg3ce49uwj+8C9wC3Oud+GWX9dcAfgKHhA5PM7AbgMWCQP0iprXXu8NG2IXX1DZz8o1mU76/hX7eeycgjunbo8URSQXMjMaXzSIXRtv8ETjWzQaEFZjYAGOeva5GZ3YoXnN+NFpy+l4Ea4JqI5dcCi2MJzsMtIz2NC48pAWCGRt2KiHSoZArP3wGlwPNmdqmZTcEbfbseeDhUyMz6m1mdmX0vbNlVwAN44TjbzE4Nex244uycKwPuB+4ys2+Y2QQzewg4B4jvMLQOoK5bEZHDI2kGDDnn9pvZOXjh9gTeIJ5ZwNedc/vCihqQTtM/DC7wl1/gv8K9AUwI+/xdYB9wG1ACrAA+45wL7sp0G508sAe9CrJZX17Jog27GdsvsZ9KICKSrJImPAGcc+uAFueZdc6VEjE61jl3A3BDG49Rj9e9e0976hik9DTj4tF9ePztUl5YtEnhKSLSQZKp21baYPLYPgC8sGgzDQ3quhU5VLoE0nkdyr+twrOTOa5fd44ozGHz7ir+va7Nd/CISBSZmZlUVlYGXQ3pIJWVlQfNStRWCs9OJi3NuHhMY+tTRNqvuLiYjRs3UlFRoRZoJ+Kco6Kigo0bN7b7uaFJc59nMjqc93mGW7h+F5f+ah69CrKZf9e5pKcl9ARJIgltz549lJWVUVtbG3RVJI4yMzMpLi6ma9fo98S3dp9nUg0YkrYZc2QhR/XIY115Be+uKee0wT2DrpJI0uratWuzv2AldanbthMyMy7xu241YYKISPwpPDup0IQJLy/eQl19xzySR0QkVSk8O6kRfQoY1Cuf8v01vP3xjqCrIyLSqSg8Oymv67Zxuj4REYkfhWcnNtm/7vny4i3U1KnrVkQkXhSendjQ3gUMLylgT1Udb63aFnR1REQ6DYVnJ3eJJkwQEYk7hWcnF7ruOXPpVqpq6wOujYhI56Dw7OQGFOVzTN+u7KuuY84Kdd2KiMSDwjMFTPZbn5owQUQkPhSeKSA0UfzsZWVU1NQFXBsRkeSn8EwBR3bP47ijulFZW8+sZWVBV0dEJOkpPFOEJkwQEYkfhWeKuHh0H8zg9RXb2FulRyuJiBwKhWeKKCnM4aQBPaipa2Dm0q1BV0dEJKkpPFPIZE2YICISFwrPFHLBMX1IM3hr1TZ2V6jrVkSkvRSeKaRXQTanDe5Jbb3jlSVbgq6OiEjSUnimGE2YICJy6BSeKeaCY0rISDPe/ngHO/ZVB10dEZGkpPBMMd3ysjhjaBH1DY6XFqvrVkSkPRSeKUgTJoiIHBqFZwo6b1RvstLTWLCmnLI9VUFXR0Qk6Sg8U1DXnEzGH90L5+DFj3TPp4hIrBSeKeoSTZggItJuCs8UNXFEb3Iy0/hg7U427aoMujoiIklF4Zmi8rMzOGd4MQAvqvUpIhIThWcKm6xRtyIi7aLwTGFnDy8mPyudhRt2s3bH/qCrIyKSNBSeKSwnM52JI3sDGjgkIhILhWeKa5wwQeEpItJWCs8Ud9awIgpyMli2eQ8fb9sXdHVERJKCwjPFZWekc/6oEgBeWKjWp4hIWyg85cCECTMWbcI5F3BtREQSn8JTGDekiO55mawu28eKrXuDro6ISMJTeAqZ6WlccIy6bkVE2krhKUDTCRPUdSsi0jKFpwBwyqCeFHXJpnRHBYs37gm6OiIiCU3hKQCkpxkXjfa7bjVdn4hIixSeckD4hAnquhURaZ7CUw44sX93SrrmsHFXJf9Zvyvo6oiIJKykCk8z62dmfzWz3Wa2x8z+bmZHtXHbH5nZq2a2w8ycmd3QTLlSf33k67K4nkwCSkszLg49JFujbkVEmpU04WlmecBsYDhwPfA5YCjwupnlt2EXXwNygRfaUPYV4LSI1xvtqHbSCU2Y8OJHm2hoUNetiEg0GUFXIAY3A4OAo51zqwHMbBGwCrgF+Hkr2xc65xrMbAhwXStltzvn5h9qhZPRsf26cWT3XDbsrOS90nJOGdQz6CqJiCScpGl5AlOA+aHgBHDOrQHmAZe2trFzrqED69ZpmIV13epJKyIiUSVTeI4CFkdZvgQYGedjTTazCjOrNrP5qXC9M1xowoSXFm+mrl5/c4iIREqm8OwB7IyyvBzoHsfjzMC7Pno+cA1QBTxnZtdGK2xm05oZYJS0FwxHHdGVgUX5bN9Xw4I15UFXR0Qk4SRTeAJECySL6wGc+5pz7o/Oubecc38FzgXeB+6N53ESmZk1PmlloSZMEBGJlEzhuROv9RmpO9FbpHHhnKsHngWONLM+HXWcRBOaMOHlJVuoVdetiEgTyRSeS/Cue0YaCSzt4GOHWrcHtXydc9Occxbt1cF16lBHlxQwtLgLuypqmbt6e9DVERFJKMkUnv8ETjWzQaEFZjYAGOev6xBmlgF8GljnnNvSUcdJRJPH+tP1acIEEZEmkik8fweUAs+b2aVmNgV4HlgPPBwqZGb9zazOzL4XvrGZjTezqcAF/qITzWyqvyxU5rNm9pSZXWdmZ5vZVcDrwAnAtzv07BJQ6Lrnq0u2UFVbH3BtREQSR9JMkuCc229m5wD3A0/gdaXOAr7unNsXVtSAdA7+w2A6MD7s81f9V2gbgDVAMfATvOurFcB7wAXOuVfidzbJYVCvLozs05Wlm/fw5sptnDeqJOgqiYgkBNPTMzpO+O0qyfo9/3rOav735RVMGXsEv/jscUFXR0TksDBrHLYSbQxLMnXbSgAuGe1d93xt2VYqa9R1KyICCk9pxVE98xjbrxsVNfW8vqIs6OqIiCQEhae0arImTBARaULhKa26aLQXnrOXl7Gvui7g2oiIBE/hKa06olsuJ/bvTnVdA7OWbQ26OiIigVN4SpuEJkyYoQkTREQUntI2F44uIc3gzZXb2F1ZG3R1REQCpfCUNikuyOGUgT2pqW/g1SUpNUuhiMhBFJ7SZpeM9QYOvbBIXbciktoUntJmFx7Th/Q0Y97q7ezcXxN0dUREAqPwlDbrkZ/FuCFF1DU4XlbXrYikMIWnxOQSTZggIhLf8DSz7maWH899SmI5f2QJmenG/E92sG1vddDVEREJRMzhaWbnmtn/mln3sGXFZvYGsB0oN7Ofx7OSkjgK8zI5a2gvGhy8tFgDh0QkNbWn5fk14FPOuZ1hy34KnAmsBnYAt5nZZ+JQP0lAB0bdasIEEUlR7QnPscDc0AczywWmAjOdc0cDRwPrgS/FpYaScCaO6E12RhrvrS1n8+7KoKsjInLYtSc8i4Hw0SKnADnA4wDOub3AC3ghKp1QQU4mZx9djHPwou75FJEU1J7wrAZywz6fCTjgzbBle4Aeh1AvSXCaMEFEUll7wnMNcE7Y5yuAVc65jWHL+uENHpJO6pzhxeRmpvPh+l2sL68IujoiIodVe8LzD8BoM1tgZm8Bo4E/R5Q5HlhxqJWTxJWXlcHEkb0BePEjtT5FJLW0JzwfAp4CTgTG4V3f/HFopZmdDIwA5sShfpLAQhMmvLBIEyaISGrJiHUD51wtcLWZfcn76PZGFPkEOA4oPfTqSSIbP6wXBdkZLN64hzXb9zOwSPNjiEhqaPcMQ865PVGCE+fcdufcQufc7kOrmiS6nMx0Jvldty9ouj4RSSHtmWGou5mNNLPsiOU3mtnzZvZnv+tWUoBG3YpIKmpPy/NHwILwbc3sa8AjwGTgKmCOmY2MSw0loZ0xpBeFuZms2LqXlVsP6ogQEemU2hOe44BZzrnwqWXuADYCZwGhafm+cYh1kySQlZHGBaNKAHXdikjqaE949sW71xMAv4XZD/ilc26uc+6vwAy8IJUUEN5165wLuDYiIh2vPeGZC1SFfR6HN8PQa2HLPsYLWUkBpw3qSc/8LD7Zvp+lm/cEXR0RkQ7XnvDcCAwP+3w+3nR8C8OWdQc0Y3iKyEhP48LRftetBg6JSApoT3i+DlxkZv9lZjcBU4CXnXMNYWWG4D1ZRVLEJWOOALwJE9R1KyKdXXvC815gH/B/wG/xunCnhVaaWTEwHng7DvWTJHHSgB4UF2SzvryShRt0i6+IdCg3sTwAACAASURBVG4xh6dzbg0wCrgNuBU4xjkXPo9tf+BX+I8ok9SQnmZcNDr0kGyNuhWRzs3UxdZxzOzAl5sK3/MHa8u54qF36FOYw7xvn0NamgVdJRGRdjFr/P3lnDvol1m7p+fzd55pZqPN7EwzG2NmmYeyP0lux/XrTt9uuWzeXcW/1+0MujoiIh2mXeFpZl3N7DfALuBDvCeo/AfYZWa/MbNu8auiJIu0NOPiMZquT0Q6v/bMbdsVmAd8EagD3gKe8d9r/eVz/XKSYkKPKXvxo83UN3T+rmoRSU3taXnehTdg6CGgv3NugnPus865CTQOFhrpl5MUM7pvIUf1yGPb3moWrNkRdHVERDpEe8LzU8B859xXnXO7wlc453Y7574GvANcEY8KSnIxs7CHZKvrVkQ6p/aE51F41zhb8gbefLeSgiaP9SZMeHnxFmrrG1opLSKSfNoTnhVAcStlevnlJAUNLylgcK98yvfX8PbH6roVkc6nPeH5HvBpMxsabaWZDcZ7LNl7h1IxSV5e160/XZ8mTBCRTqg94fkToAvwnpn9wMzOMbMRZna2mU3HC80uwE/jWVFJLpP9x5S9smQLNXXquhWRzqU90/PNAr4C5ADfAWYCi/EeSfbfQD7wX86515rdiXR6Q4oLGF5SwJ6qOt5atS3o6oiIxFVGezZyzj1sZi8BnwOOAwqB3XgTJTzpnFsbvypKspo89giWb1nBC4s2c+6I3kFXR0QkbjpkblszywGynHMp/WTkVJvbNtLaHfsZ/5M55Gel88F/TyInMz3oKomItEmHzm3bgoeA8g7atySJ/j3zGd23kP019cxZURZ0dURE4qajwhMg7o/UMLN+ZvZXM9ttZnvM7O9mdlQbt/2Rmb1qZjvMzJnZDc2USzOzu8ys1MyqzGyhmWnCh3YKTZgwQxMmiEgn0pHhGVdmlgfMBoYD1+Ndbx0KvG5m+W3YxdeAXOCFVsr9AO/h3g8CFwLzgWfN7KL21Ty1hSaKn72sjIqauoBrIyISH0kTnsDNwCDgMufcP5xzzwNT8ObTvaUN2xc6587EC8eozKwYuAO4zzn3U+fc6865W4DXgfsO+QxS0JHd8zj+qG5U1tYza5m6bkWkc0im8JyCN6fu6tAC59wavCe8XNraxs65ttxseD6QBTwZsfxJYLSZDWx7dSUkNGHCDE2YICKdRDKF5yi8+0kjLcF7iku8jlENrI5YvsR/j9dxUsrFY/pgBnNWbmNvVW3Q1REROWTJFJ49gJ1RlpcD3eN4jF3u4PtKysPWN2Fm0/wBSAe94lSnpNe7aw4nD+hBTV0DM5duDbo6IiKHrE3haWb1sbyA6zqovtECKZ6jeu0wHCMlXeI/aUWPKRORzqCtLU9rxyvedhKl5YfX6ozWIm2PcqC7hd8d23iM0HpphwuPKSHN4M2V29hVURN0dUREDkmbwtM5l9aOV7ynk1mCd00y0khgaRyPkQ0MjnIMoh3HOTfNOWfRXnGqU6dQ1CWb0wcXUdfgeGXJlqCrIyJySJLpmuc/gVPNbFBogZkNAMb56+LhZaAGuCZi+bXAYn90r7RTaMIEdd2KSLJLpvD8HVAKPG9ml5rZFOB5YD3wcKiQmfU3szoz+174xmY23symAhf4i040s6n+MgCcc2XA/cBdZvYNM5tgZg8B5+A9QUYOwQXHlJCRZrz98Q527KsOujoiIu3WrqeqBME5t9/MzsELtyfwrqvOAr7unNsXVtSAdA7+w2A6MD7s81f9V2ibkO8C+4DbgBJgBfAZ59yMOJ1KyuqWl8WZQ4t4fcU2Xlq8hWtP7R90lURE2qVDnqoinlR/qko0f/tgA998diGnDOzB07ecFnR1RESiCuqpKiJRTRrVm6z0NN4tLWfrnqqgqyMi0i4KTzmsuuZkMv7oXjgH//pIA4dEJDkpPOWwm6wJE0QkySk85bA7d3gxOZlpfLB2Jxt3VQZdHRGRmCk85bDLz87g3OG9AXhxkZ60IiLJR+EpgdCECSKSzBSeEoizhxeTn5XOog27Wbtjf9DVERGJicJTApGTmc6kkV7XrVqfIpJsFJ4SmEvGeKNuZyzUdU8RSS4KTwnMmcOKKMjJYPmWvawu29f6BiIiCULhKYHJzkjn/FElALygUbcikkQUnhKo8AkTNP+viCQLhacE6vTBPemel8nqsn2s2Lo36OqIiLSJwlMClZmexgXHePd8auCQiCQLhacEbnLYhAnquhWRZKDwlMCdMqgnRV2yWbujgsUb9wRdHRGRVik8JXDpacbFozXqVkSSh8JTEsIlGnUrIklE4SkJ4YSjulPSNYeNuyr597pdQVdHRKRFCk9JCGlpxsUHBg6p61ZEEpvCUxJGaMKEf320mYYGdd2KSOJSeErCGHtkIf165LJ1TzXvlZYHXR0RkWYpPCVhmBkXj/aftKKuWxFJYApPSSiX+Nc9X/poC3X1DQHXRkQkOoWnJJRRR3RlYFE+O/bXMP8Tdd2KSGJSeEpCMbOw6frUdSsiiUnhKQknNGHCy0u2UFOnrlsRSTwKT0k4w3oXMKx3F3ZV1DJv9fagqyMichCFpySkS8Zo1K2IJC6FpySk0KjbmUu2UlVbH3BtRESaUnhKQhrUqwujjujK3uo63ly5LejqiIg0ofCUhBXqun1h0eaAayIi0pTCUxJWqOv2tWVbqaxR162IJA6FpySsfj3yGNuvGxU19cxeXhZ0dUREDlB4SkLThAkikogUnpLQQs/4nL28jH3VdQHXRkTEo/CUhNanMJeTBnSnuq6B15ZuDbo6IiKAwlOSQOOoW3XdikhiUHhKwrtwdAlpBm+s3MbuytqgqyMiovCUxFdckMOpg3pSW+94dcmWoKsjIqLwlOSgCRNEJJEoPCUpXHBMCelpxtzV2ynfXxN0dUQkxSk8JSn0yM9i3JAi6hscLy9W162IBEvhKUlDEyaISKJQeErSOG9UCVnpacz/ZAfb9lYHXR0RSWEKT0kahbmZnDWsiAYHLy3WwCERCU5ShaeZ9TOzv5rZbjPbY2Z/N7Oj2rhtjpn9xMw2m1mlmb1jZmdFKVdqZi7K67L4n5HEKjTqdsZCdd2KSHCSJjzNLA+YDQwHrgc+BwwFXjez/Dbs4vfAzcD3gEuAzcArZnZslLKvAKdFvN441HOQQzdxZG+yM9J4r3Qnm3dXBl0dEUlRSROeeME3CLjMOfcP59zzwBSgP3BLSxua2VjgauB259zvnHOzgM8A64D/ibLJdufc/IjXzriejbRLl+wMzhleDMCLuudTRAKSTOE5BZjvnFsdWuCcWwPMAy5tw7a1wNNh29YBTwHnm1l2/KsrHUUTJohI0JIpPEcBi6MsXwKMbMO2a5xzFVG2zQKGRCyfbGYVZlZtZvNbut5pZtOauUbqWqmTtNM5w4vJy0rnw/W7WF8e+U8qItLxkik8ewDRuk7Lge6HsG1ofcgM4GvA+cA1QBXwnJldG1NtpcPkZqVz7ojegFqfIhKMZApPgGitOWvDdtbWbZ1zX3PO/dE595Zz7q/AucD7wL0x1VQ6lCZMEJEgJVN47qRpCzGkO9FbleHKW9g2tD4q51w98CxwpJn1ibJ+mnPOor1aqZMcgvFH96IgO4Mlm/awZvv+oKsjIikmmcJzCd61y0gjgaVt2Hagf7tL5LY1wOqDN2kiFIS6jpkgsjPSmTTK77rVPZ8icpglU3j+EzjVzAaFFpjZAGCcv661bTOBT4dtmwFcCbzqnGt2rje/3KeBdc45zUieQCaHJkxQ162IHGbJFJ6/A0qB583sUjObAjwPrAceDhUys/5mVmdm3wstc859iHebygNmdpOZnYt3m8pA4Pth237WzJ4ys+vM7Gwzuwp4HTgB+HbHn6LEYtyQIgpzM1m5dR8rt+4NujoikkKSJjydc/uBc4CVwBPAn4A1wDnOuX1hRQ1I5+BzuxF4DLgHeBHoB1zgnPt3WJk1QDHwE+BVvFCu9ss9Fe9zkkOTlZHGhceUAOq6FZHDy5zTZbyOEn6vp77njjF31Xau/f0CBhXlM+ub4zHTOC0ROXThv0uiDQBNmpanSDSnDupBz/wsPtm+n6Wb9wRdHRFJEQpPSWoZ6WlcONrrup2xUBMmiMjhofCUpDf5wFy3m9Q9LiKHhcJTkt5JA3rQu2s2G3ZWsnDD7qCrIyIpQOEpSS8tzbhotD9dn0bdishhoPCUTiH8MWUNDeq6FZGOpfCUTuH4o7rRt1suW/ZU8cE6PbdcRDqWwlM6BTPjkjHquhWRw0PhKZ1GqOv2X4u3UK+uWxHpQApP6TSO6duV/j3z2La3mgVrdgRdHRHpxBSe0mmEd91qwgQR6UgKT+lUQl23Ly/eTG19Q8C1EZHOSuEpncrwkgKGFHdhZ0Utb3+srlsR6RgKT+lUNOpWRA4Hhad0OqGu21eWbKG6rj7g2ohIZ6TwlE5nSHEXhpcUsKeqjrdWbg+6OiKta6iHTf+BbStBDzdIChlBV0CkI0weewTLt6zghUWbmDiyd9DVETnYvjJYPQtWz8R9PBur9GbGqsvpSU3fU2nofzppA8aR1ecYMjIzA66sRDI9wqnjmNmBL1ff8+G1bkcFZ/3kdfKz0vngvyeRk5kedJUkxe2rrKZs2VzqV8ykcOMcivcta7J+XUMvcqyWYtvVZPlul8cHbjj/SRvJR+mjWJs1lMysLHIz08nOTCc3M52czDT/vfF1YHlWOjkZ6eRkpZOT4X1uWrbptulpdji/loRl1vg9OOcO+lIUnh1I4RmsKQ/OZdGG3fzm2uO54Jg+QVdHUsDeqlrW7qigdMd+1u6oYNuW9RRteZNhexZwcsOHdLP9B8pWu0zmN4xgTsNY5nEcdd0HkZeVTo/qDYyoXszo+sUc27CUI21bk2NUuGw+aBjKuw3DebdhBB+6wVSTFbdzyEpPIyczzQvgKMGbk+Ev98uEgjoU1k3DO53crDSyD2wTVi4jnbQEDmqFZ4AUnsH67Zsf86N/Lefi0X341TXHB10d6ST2VNWydnsFa3bsZ+32/ZQeCMv97NxXybG2mgnpC5mQ9iGj00qbbLvRSlhecCplxWfCwDPoV1xE/555HNEtt9kWn9u1jvo1b1NfOo/09e+QUb6qyfqGtEx2dR/Dth4nsKnweNZ3Gc1el0N1bT2VtfVU1Tb476FX4+fK2nqqIz4fzl9VWRlpzbScGz/nZobC2wviUHg33+pu/FyYm0l+dvuuTio8A6TwDNbGXZWMu282OZlpfHD3pHb/TySpZ3dFLaU79nuv7RWsDf28o4Ly/TVNyvZiJ+PTFzEhbSFnpn1EYVjrsi4tm70lp2LDJlFwzEWkFw0+9MrtK4N178Dat6F0HmxdDIT9frF06DMG+o+D/qfDUadBXo827do5R019A1U1TQO1uaCtqm04EMqVNfVU1dVTWdNAVV09VQc+R5QLC/CO9pUJg/l/Fwxv17YKzwApPIN3xUNv88Hanfzis8cxZewRQVdHEsiuihqv1bg9FJJeOK7dsZ+dFbXNbpef6bigcD0TMxZyXM0HlFSsbFqgx2AYOgmGTIIB4yAzt2NPpHInrFsAa+d5gbrpP+Aigql4lBekoVdBScfWqQ2cc1TXNTQNVD9wq8KDOCzA29JyDg/qz48byM1nDWpX/RSeAVJ4Bu+xeWuYPmMp543szW+vOzHo6shh5Jxjp9+CXLtjP2sOtCC9wNxd2XxA5mam079nHgOL8unfM5/h+fsYXfkefbfPI3vdG1j1nsbCGbkw8EwvLIdOhB7t+2UdN9X7YMN7XpCufdv7ub66aZkeg/0g9Vun3Y4CS9zrj0FQeAZI4Rm8rXuqOPXeWWSmp/H+3RPpmqMh/52Jc47y/TVNulfX+K3H0u372VNV1+y2eVnpDOiZz4CiPPr3zGdgz3z698xjQFE+xXlp2Ib3YPVMWPUabP2o6cY9hzSGZf/D0Lo8FHXVsPHfjS3T9QugZl/TMl2PDGuZjoOioSkfpgrPACk8E8OVD7/DgjXl/OzTY7nihCODro7EyDnH9n01futx/4HRrKU79rN2ewV7q5sPyPysdAYU5Xuvnn5IFnkh2atLdpNfkOzZBKtfg1Uz4ZM5cFDr8iy/O3Yi9BjYcSfc0errYMuixjBd+zZUNb09hryipi3T3qMgLbVu91J4BkjhmRienL+Wu/+xmLOP7sVjN54cdHUkCucc2/ZVU7q9ceRq488V7GshIAuyMxjgB6LXkmwMyqIuWU0DMlx9rdcKWzXTC82ti5uu7zm0MSz7j4PMnDiecQJpaIBty/wg9QN139amZXIKvYFHoUDtMxbSO3cvjsIzQArPxLBjXzUn/2gWBrx/90S65cXvnjhpO+ccZXurGwfo+N2roWuRFTXNj74syMlgYFG+F45+MIZCskd+CwEZafdGvyt2JnzyBtTsbVyXmQcDx3tdsUMmQvcBh3bCyco5KP8krGU6D3ata1omMw/6ndzYMu17QmJ3XbeDwjNACs/E8bnfL+CtVdv58RWjufKko4KuTqfV0ODYurcq7Pqj17UaakG2dHtCYW5mRPeqH5I98+mel9n2gAxXVwPr5ze2LsuWNl1fdHRY6/J0yMiO/RipYNd6//aYed7tMTua3mtKepYXoKHrpv1OgeyCYOoaJwrPACk8E8fT763j23/7iDOHFvHEF04JujqBcc5R3+Co81/19Y7ahgbq6h11B96b/lzf0EBtvYtaZk9lXeM9kNsrWFu+n6ra5h9C3j0vs8l1x/Bu1rj1COze0BiWB7Uu82HQeC8sh0yE7v3jc8xUs6+s8Xrp2rfjeq9polB4BkjhmTh2VdRw4j2v0eAc7353IkVdorcwIsOlrt4LjvoGR219g7+uwV8XpUxDA/X1TcuEtqttcNTXNzTZ94EyB7aLCKj6Bn+7yONGD7rwfTZd11imo/XIz2KAH4z9/dGsA/wWZGFeB1wnq6vxWkWhkbHbms4ZS6/hXlAOneT9ElfrMv7adK/pyKYjehPgXtOWKDwDpPBMLJ9//D1mLy+jqEsWGWlpUQPucIRL0NIMMtLSyEg30tOMzPQ07z3NyEhPIyPN/HVpZIbKRCufbuRmZnjdrEXerR5H9cyjMPcwDCTZtb4xLNe80fTWi6wuMGhCY+uyW7+Or4801QnuNVV4BuiQw7Oh3usOScsIe6W38jkjof4DTCQzl27l5j++32q58HDJCA+UiHBpUiYiXJqUiRZGEfsMhVGoTNPtmgZW47rw7cL32bTukdsl8mTczaqr9n4Jr37Ne21b3nR98cjG1mW/UyFDg8ISSpN7Ted5rdTa/U3LJNi9pgrPAB1yeFbuhB8PaMeBWwvYVj6nZxza9ge9oq1vYR/pmbFv08Y/HMr2VFFV29A0GMMDMFnDpTPauTasdflm01+2WQXetcvQYJ9C3b+bVOrrYMvCptdNE+xeU4VngOISnn+Y7LVAG+rCXhGf68N+jrzOkGosDdIyW/7DIKebN3ghryfk+u95PRqXHVjeQ9fHDqe6aq9Vsuo1LzS3R8wZWzzKv41kkjeaU63LziPyXtPSebC/rGmZ7EI46tTGQD3i2A6911ThGaBArnk613rYHrSstg3bNLOP+toYtom2j1jqGfa5PkqdO+IPh6wuzQSt/54bFrihz531ZvqOsLO0cWTsmjehtqJxXXZXf2RsqHXZN7BqymEWea9p6TzYfXjvNVV4BkgDhg6ztvzhUF8Dlbugshwqdvgv/+fK8safQ+/tCeTM/INbsy22cHumTuDWVnm/EEPT4EXeL9h7dOMkBf1O6fSz2EgMdq2Dte80BmoH32uq8AyQwjPJOQdVu8NCNSxwK8NDtrzpsobmp5JrVmZe863ZvJ6Q2/3gZckyo0v5J7B6lheWpW9FtC4LYfCExtZl1z6BVVOSTKv3mqbB6bfCpOnt2r3CM0AKzxTknDeheChU29rCbWj+8VjNyshtuTUbreWblRf/c45UW+l1s4WmwSv/uOn6ktH+E0kmwZEnqXUp8RHtXtMLfwwn39yu3Sk8A6TwlDZxzrtP8UDI7mymhbujaSDX18R+rIycVlq44UEcauHmtX7LwI6PG7tiS+dCXWXjuuxCGHx248jYBL85XjqJ6n2Aa3fXrcIzQApP6TDOQc3+iJAtj9LCjehajrxRvS0ycsKCtnvTFm7Vbq+FWf5J021KxvhhGWpdZsTnvEUOE4VngBSeklCc8643NulGbqZrOXx5XVXr+87pBoPP8QJz8LlQ0Lvjz0ekAyk8A6TwlE6hpiJKC9cP1rR07zFefU9Q61I6FYVngBSeIiLJqbXwTDustREREekEFJ4iIiIxUniKiIjEKKnC08z6mdlfzWy3me0xs7+b2VFt3DbHzH5iZpvNrNLM3jGzs6KUSzOzu8ys1MyqzGyhmV0R/7MREZFklTThaWZ5wGxgOHA98DlgKPC6meW3YRe/B24GvgdcAmwGXjGzYyPK/QCYBjwIXAjMB541s4vicBoiItIJJM1oWzO7Dfg5cLRzbrW/bCCwCvh/zrmft7DtWOBD4PPOucf8ZRnAEmCFc26Kv6wYWA/c55z7ftj2s4BezrkxMdZZo21FRJJQZxptOwWYHwpOAOfcGmAecGkbtq0Fng7btg54CjjfzEIPbTwfyAKejNj+SWC0H9YiIpLikik8RwGLoyxfAoxsw7ZrnHMVEcuX4IXlkLBy1cDqKOWIdhwzm2ZmLtqrlTqJiEiSSqbw7AHsjLK8HOh+CNuG1ofed7mD+1gjy4mISApLtvm0orXmWnncw4Eybdm2reViZq09lUJERJJGMrU8dxK95ded6K3KcOUtbBtaH3rvbgcnXWS5A5xz05xzFu3VSp1ERCRJJVN4LsG7JhlpJLC0DdsO9G93idy2hsZrnEuAbGBwlHK04TgiIpICkqnb9p/AT81skHPuEwAzGwCMA+5sw7bTgU8Df/C3zQCuBF51zoUecvgyXphe45cPuRZY7I/ubbN4tD4jbndRa/YQ6fuMH32X8aXvM746+vtMpvD8HfBfwPNmdjfetckf4N2X+XCokJn1Bz4G/sc59z8AzrkPzexp4AEzywTWAF8GBuIFJX65MjO7H7jLzPYC/8YL2HNo/XYYERFJEUkTns65/WZ2DnA/8ATeIJ5ZwNedc/vCihqQzsFd0jcCPwTuAboBC4ELnHP/jij3XWAfcBtQAqwAPuOcmxHfMxIRkWSVNDMMpSp15cSXvs/40XcZX/o+46ujv89kGjAkIiKSEBSeIiIiMUqaa54pbHrrRSQG+j7jR99lfOn7jK8O/T51zVNERCRG6rYVERGJkcJTREQkRgrPBGVm/czsr2a228z2mNnfzeyooOuVjMzsSDP7pZm9Y2YV/iPjBgRdr2RkZlPN7G9mttbMKs1shZnda2YFQdctGZnZ+WY228y2mFm1mW0ws2fMrLXHLEobmNnL/v/v98R73wrPBOTPwTsbGA5cD3wOGAq8bmb5QdYtSQ0BPoP3AIG3Aq5LsrsDqAe+A1wAPIQ3W9dMM9Pvk9j1AD7Amz3tPOAuvDm85/uzpUk7mdlngbEdtX+Ntk1MNwODgKOdc6sBzGwRsAq4Bfh5gHVLRm8653oDmNlNeL+kpH0mO+e2hX1+w8zK8eaMnoD3R5+0kXPuL8BfwpeZ2bvAcmAq8LMg6pXszKwb3mx0twN/7ohj6C/FxDQFmB8KTgB/Uvp5aI7dmDnnGoKuQ2cREZwh7/nvfQ9nXTqxHf57baC1SG7/Cyzx/zjpEArPxDQKWBxl+RIaH48mkijG++/LAq1FEjOzdDPLMrOheA+62AI8FXC1kpKZnQFcB3ylI4+jbtvE1IPoD/gup/HB3CKBM7O+wP8Arznn3g+6PklsAXCC//Nq4BznXFmA9UlK/lOzHgZ+6pxb0ZHHUsszcUWbvUKTRUvCMLMuwPNAHd5Ti6T9PgecClwN7MEbgDUgyAolqW8DuXhP0OpQankmpp14rc9I3YneIhU5rMwsB+8h84OA8c65DQFXKak550Jd3gvM7CWgFLgT+FJglUoy/q183wVuArLNLDtsdbY/iGivc64+HsdTyzMxLcG77hlpJLD0MNdFpAm/a+xvwMnARc65jwKuUqfinNuF13U7JOi6JJlBQA7wJF4jI/QC7xarncDoeB1M4ZmY/gmcamaDQgv8Lpxx/jqRQPj3cv4JOBe41Dk3P+AqdTpm1hvvHu+Pg65LkvkQODvKC7xAPRvvj5K40MTwCcifCGEhUAncjXf98wdAATDGObcvwOolJTOb6v94Ll5X2FeAbcA259wbgVUsyZjZQ3jf3w+BFyJWb1D3bWzM7Dng38AivGudw/DuTSwBTnbOrQywep2C/1DsHzrn7o7rfhWeicnvv78fmIQ3UGgW8HXnXGmQ9UpW4U+Vj/CGc27C4axLMjOzUqC5mW+mO+emHb7aJD8z+zbe7FeDgSxgPTAHuFf/r8eHwlNERCRB6JqniIhIjBSeIiIiMVJ4ioiIxEjhKSIiEiOFp4iISIwUniIiIjFSeIqIiMRI4SkiIhIjhaeIiEiMFJ4iIiIxUniKiIjESOEpIiISI4WniIhIjBSeIiIiMVJ4ioiIxEjhKSIiEiOFp4iISIwUniKSMMxsmpk5M5sQdF1EWqLwFOlE/OBp7TUh6HqKJLuMoCsgIh1iegvrSg9XJUQ6K4WnSCfknJsWdB1EOjN124qksPBrjGZ2vZn9x8wqzazMzB41s5JmthtqZn80s41mVmNmm/zPQ5spn25mXzKzeWa22z/GajN7pIVtpprZu2ZWYWblZvaUmfWN5/mLtJdaniICcDtwHvA08DJwBnAjMMHMTnHObQsVNLOTgNeAAuCfwFJgOHANcKmZneucez+sfBbwIjARWA/8GdgDDAAuB+YCqyLq8xVgir//N4BTgCuBsWZ2rHOuOp4nLxIrhadIJ2Rm05pZVeWcuy/K8guBU5xz/wnbx/3AG/QOCQAAAolJREFU14H7gC/4ywz4I9AVuNY596ew8lcCTwFPmtlI51yDv2oaXnDOAD4dHnxmlu3vK9IFwEnOuY/Cyv4Z+CxwKfBMsycvchiYcy7oOohInJhZa/9D73bOdQsrPw34PvCoc+4LEfsqBNYC2UA351y1mY3Daym+45w7Pcrx38JrtY53zr1pZunADiALGOKc29RK/UP1+aFz7u6IdWcDs4GfOefuaOU8RTqUrnmKdELOOWvm1a2ZTd6Iso/dwIdADjDCX3y8/z67mf2Elh/nvw8HCoFFrQVnhPejLFvvv3ePYT8iHULhKSIAW5tZvsV/L4x439xM+dDybhHvG2Osz64oy+r89/QY9yUSdwpPEQHo3czy0Gjb3RHvUUfhAn0iyoVCUKNkpVNReIoIwPjIBf41z2OBKmCZvzg0oGhCM/sJLf+3/74cL0DHmNkR8aioSCJQeIoIwOfM7LiIZdPwumn/EjZCdh6wAjjDzKaGF/Y/nwWsxBtUhHOuHvg1kAv8xh9dG75Nlpn1ivO5iHQ43aoi0gm1cKsKwD+ccx9GLHsJmGdmz+BdtzzDf5UCd4YKOeecmV0PzASeNrPn8VqXRwOXAXuB68JuUwFvqsBTgMnASjN7wS/XD+/e0m8Bj7frREUCovAU6Zy+38K6UrxRtOHuB57Du6/zSmAfXqB9xzlXFl7QObfAnyjhbrz7NycD24G/AD9wzq2IKF9j/799O7aBEAaCKDou7TKq4NIrk4BaaMMkBKQToRPvNWA7+rJWO8aS5JdkTfJNMpIc15l7/zx4lj1PeLHbXuVnzrk9exv4H2aeAFASTwAoiScAlMw8AaDk5wkAJfEEgJJ4AkBJPAGgJJ4AUBJPACidAu5DQAa+5VEAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 455.04x367.92 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot training & validation accuracy values\n",
    "plt.plot(history.history['loss'])\n",
    "plt.plot(history.history['val_loss'])\n",
    "plt.title('Model accuracy')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Validation'], loc='upper right')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAc8AAAF4CAYAAAAyt3LmAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8li6FKAAAgAElEQVR4nOzdd3hc1bX4/e9StZrlIksyxr3gSjMYgymmmBYCIZDAa7iB0BJucklIID8g3MSGJJBCuQmhpDiACQFSKAEXHDcwCTUJuMgNcNxky01Wsfqs9499Rj4ajcpIY49mZn2eZ54ZnbP3OfvIMEu7i6pijDHGmM5LiXUBjDHGmHhjwdMYY4yJkAVPY4wxJkIWPI0xxpgIWfA0xhhjImTB0xhjjImQBU9j4oyITBcRFZFZ3bzOtd51ro1OyYxJHhY8jemAF2BURAIiMrKddEt9aa89jEU0xhxmFjyN6ZxGQIDrw50UkdHAGV46Y0yCs+BpTOfsBN4HviwiaWHO34ALrq8e1lIZY2LCgqcxnfdroBi4yH9QRNKBa4C/A6vbyiwio0XkaRHZJiL1IrLd+3l0G+mLROS3IrJTRGpE5N8ick17BRSRfiJyn4iUeHn2i8hiETk34qdtfe0jROR7IvKWiOzwPcOzIjKunXxTROR577nrRKRURF4XkS92JW1Hfb4isklENoUca+7fFZHzRWSZ97tRX5rPicgzIrJeRKpFpEpEPhCRW0Qk7HeliGSLyP8TkfdFpNLLUyIiPxeRIi/Nc969T2/jGpd753/R1u/Q9Dzh/oI2xoT3B+BBXC3zJd/xi4Ei4A5gVLiMInIi8DcgD3gFWAOMBa4CLhGRs1X1fV/6/rhgPAJY4b0GAo8Dr7dxj6HAMmAY8CawAMjBBfsFIvIVVf115I/d7HTvGZcCfwaqgNHA5cDFIjJNVT8MKdONwGNAk/fcG4BC4ATgv4EXupK2Gy4Hzgfm436Xw3zn7gcCwDvANiAfOAv4P+BE4L9Cnq0v7ndxDLAOmAPUAyOB64C/4FosHgWuAL4CvBGmTDd577/q5rOZw0lV7WUve7XzAhTY6n3+Da5f80jf+QXAfiAb+IGX/lrfeQFKvONXhVz7Cu/4WiDFd/xX3vGHQtKfADR452aFnFuG+/K/MuR4H+DfQA1Q5Dt+bWhZO/g9FAJ5YY4fgwuk80OOj/fKuheYECbfkV1MOz3c8/vObwI2hRwLPmsAOL+NfCPDHEsBnvLynhRy7lnv+GP+fzvvXB6Q7/t5FVALFISkG+6V6a1Y/3dur8he1mxrTGR+DaTiahbB2t4M4PeqeqCNPKfgapn/UNXf+0+o6vO4WuVRwKneNdNxNdJKYFZI+veBFtfw8hyDG7D0Z1V9LiRPOfB9oBdwWecftSVVLVPVyjDHPwSWAGd6ZQ+6Gde6da+qtmrOVtWtXUzbHS+r6oJwJ1T14zDHAriaJ8B5weMiUoj7w6cUuM1L589Xqar7fYceAzJxzft+N+H+uHoiwucwMWbNtsZEQFXfEZGVwHUi8gNcE24KLqi25XjvfUkb55fgAudxuGa9sbha7JshX8BBy2j9JXyy957fRl/gAO+9zb7JzhCRzwBfxdWAC2j9HVKACygAU733+Z24dCRpu+Pdtk54TeW3AxfimstzQpIM8n0+Effv/oaqVnfivk/jmoVvAh7w7peOqxHvIzpN0uYwsuBpTOR+Dfwc13f2ZeADVf1XO+nzvffSNs4Hj/cJSb+zjfQ7whzr773P8F5tyW3nXLtE5BZcLWwfsAjYDBzANV1+Dtd8m+nLEnyebZ24fCRpuyPc7w4R6QO8h2tGfRcX7Pbimuj7AN+g68+GqlaKyDPAV0XkTFVdClyCG4D2sKrWduFZTAxZ8DQmcnOBH+Oa2gYB93SQPlh7LG7j/MCQdMH3ojbSh7tOMM83VPXnHZQnYt70nNm44HO8qpaGnD85TLZy730Qrk+3PZGkDTaRtvX9lc/B30cobeP4DbjAOVtVZ/lPeM/2jZD0/vJ21mO4WvtXcAONbKBQHLM+T2Mi5PUh/gk4EqjGjcJtT7BWOr2N88Hj//Te1+JqdMeKSH476f3e9t5P66AsXVWAq239PUzgzOVg03S4Ml3QietHknaf9z449ISIjOJgrTASwVHSfw5z7owwx97FBfHTRSS0eTcsVf0IeAu4VEROAs7BNfuWdKG8JsYseBrTNXcDlwLnhRtEE+It3FSGU0Xkcv8J7+fTgfW4gUOoagNuUFAeIQOGROQE3GCiFryBRG8CnxeR68IVQkQmeQNduqIMF9Ane8EyeM10XFNuQZg8j+GaPf9XRMaHKc+RXUy7FqjATfEp9KXJwjWnd8Um7316yH2PA+4MTayqu4DncK0GPwudByoiuW384fMYkIEL0oKbLmPikDXbGtMFqroZ1+fXmbTqLW6wCHheRF7GBYCjcH2FlcCXQkZs3gWcDXzTC5jBeZ5XAPNwc0tDzcQNPvqt1z/5Dq558UjgaGAibmBRWWRP60adisjPcfM8V3rPkAGcCfTDNUOeGZJnjYj8Ny5A/MvLswHXP3uC99xndiFtg4j8H/C/XtoXcd9lM4Dt3itST+MGCz0sImd69x6NmyP7F9zvPdTXcb/TrwLTRWQhbp7ncNzI3Itxg7v8/gg8hGvu3e1d28SjWM+VsZe9evoL3zzPTqRtNc/Td+4oXH9pKW5OYynwDHBUG9cqxk2834Wbo/lv3OjM6bQxzxFXW70L+AA397IG+BR4DdfHluNLe21bZW2jPGnAt3ALPNTg+j/nAkOBJ71rDQuT72RcTasMF1y24+bGXt7VtLha2x3Ax166zcBPcKOUN9H2PM82nxU31/QV797V3u/wBtxCCgo8GSZPDvBd4CNczbzS+/08DBS2cZ+HvOv9NNb/bdur6y/x/jGNMcYcBiKyDNdUf5SqbohxcUwXWZ+nMcYcJiIyBTcAaaEFzvhmfZ7GGHOIicjNuH7OL+NG6X4/tiUy3RVXNU8ROVJEfiEi/xCRA95OBMM6mbeXiPzU26WhxrtGq10ORCRFRO70dmaoFZEPRSTskmYicqOIrPV2f1gnIl/t3hMaYxLU/8P10VYD/6Wqba50ZOJDXAVP3FysL+Lmeb0ZYd7fAjcC38ONoCsFForIsSHp7sVND3gEN+fsbeCPInKhP5G3A8QTuMEN5+NG0T3q/YVpjDHNVHWYqqap6hhVfTbW5TFREOsRS5G8aLnrxA20MbovTL5jvLRf9h1Lw829e8V3rBCow60y4s+/GPgoJG8Z8FRIujm44efpET6X2ste9rKXvXrmK9z3dlzVPDVk54IIXIybGvC871qNuEnO54lIcM3K83Bz154Jyf8MMElEhns/n4xbaDs03VzcvLRTu1hOY4wxcSCugmc3TAA+1dZbRq3GBctRvnR1wMYw6cDNAwumA7dHX3vpjDHGJKBkGW3bj4PrYfrt9Z0Pvpdr68mv4dIR5pqh6Zp520R9v6OC2rxbY4yJPRFp93yy1DwF13Yd7nhX09FGWmOMMQkuWYLnXsLUBoG+vvPB977S+k+OcOkIc81+IeeNMcYkoGQJnquB4SKSHXJ8PG5dzI2+dJnAyDDpwK1ZGUwHB/s+20rXTFVnqaqEe0X4LMYYY2IsWYLnK0A68IXgAW9z3yuA11W1zju8ABdMQ7d8uhpYpaqfej//AzclJVy6vbgtqIwxxiSouBsw5NsPcbL3foGI7AJ2qepyERmK22nhHlW9B0BV/y0iz+O2G0rH7TJxM27roOYAqKplIvIQcKeIVOI2J74COAu4xJeuQUT+F7cowjbgb16a64D/UdX6Q/X8xhhjYi/ugiduJR+/R7335bitmgRIpXWt+svAD3FbRvUBPgTOV9V/hqT7Lm4rp2/gtoRaB3xRVf/qT6Sqj4uIAt/G7QO4Gfi6qj6KMcaYhGZbkvUAXhAGbKqKMcb0BP5xo+HGpiRLn6cxxhgTNfHYbGtC7d4AW2yThqhIzYCMbMjIgfQc9+5/pfWCDiZPG2MSnwXPRPDpG/Dat2JdiuQgKb6g6gXZjFxIz24daFuky/WOBT+H5E3PsqBsTLQEAlBdBtW7oXjiIbmFBc9EUDAajpkZ61IkhqZ6qK+G+ipoOOB99r2a6qC+0r2iSloG1Ra13uyQAN1WujCB24KySUSN9VCxDfZvgfItvvfN7r1im/t/uVc+3LH5kBTBBgz1ADZgKI40NUJDMJgeCAmyVd6x6tZp6qu9dL40/nSNtYeowOILqKG13ghrx/7AnZ5tQdkcOvXVvqC4uXWQrCylw9VRs/tD/mC4biGk94q4CB0NGLLg2QNY8DQEmnwBNTTItlELbidAq5dXGmsOSXEVoSGlF3UpWdSlZFEvvaiVLGolk7rUHFLzisjpN5C+hYPoWzgIyS2G3AGQU+iCr0leqnBg78FaYovA6AXKmnD7ePhICuQdAX0GuwCZf6T3eYj3fqT7Y68bLHjGAQuePZ+q0hhQGpuU+qYADcFXo9IQOPg5eK6xSWloCrRM6x1raHSf633p/GkP3kO9tIEwadVLe/BazdcOuHOqkEKALOrIppZsqSOH2ubP2aHHxfuZOrKllhxqyaKOHPHe/fmkruNfWhua0nKQvEJScosOBtTcQsgZALlFvs+F3f4CNDEQaHI1w/1bWzal+oNkQ+jukCFSM30BcTD0GeK9ez/3PgJS0w/pY1jwjAMWPGMvEFBufeHf/HtLOQ2NLjg1BloGungjAhmpKWSkppCelkJaipCemkJGWgrpqe5zWmoKGd7ng6+DP2ekeelSUkhPE3et1BTSUyCLOrKoJUtr6UUNWVpLRqCWQE05+3dtp7a8FK0qI7dxHwOknAKpYAD7yZSGzj9Eeo4Lov6AGgy2zZ+9AJyZe+h+meagxjovMIZpTt2/GSq2Q6Cx/Wtk5vsCo7/26AXJnAGQEtuZlBY844AFz9h7f9NeLn/8H+2mSU2RFoEl+DkYUNL8P/uCTjAAhQ1OviCWnipeYGudLj3VBa40/329tGkp/nwH86SmxL5PUlXZVVVHSWkla0srWFtawX9Kd1Cxazv9dB8Fsv/gi/0MTK3gyPQqCmQ/eU37SAtEUMNNzwlTky1sGXCba7S51mfbltr9XiDcGr7PsWpnx9fIKQwJjkNa/twr/9A/RzdZ8IwDFjxj795X1/DbFZ8y86Qh/Pf0kQdrWL4aW08IRomivjHAx7uqKCmtYO2OSkpKKygprWR3lT9YKrnUMED2MzG/lon5dYzOqWFIZjVFKfvJbdiLVO9yUxKqyiIbdJWeHVKT9ZqMw9VuEynQqkL1rrabU8u3QN3+9q+RkuaaTUMDYr6vBtmFATo9jQXPOGDBM7ZUlVN/vJRt5TX8+eZTmDy0b8eZzCGxq7KOtTsqWFvqBdQdlWwsq6ShqfX/FzkZqRxVnMfYgb0ZV5zHhP7CmNwachv2HQyoVWXe52CQ3ek+RzKQKi2rZY02XE02GHgz82IbaJsavSkcW1tP39jv1SY7+iMjLat1UPT3OeYNhJTUw/M8MWTBMw5Y8IytD7eUc8kv36Kodyb/uONsUqyG2aM0NLlaqj+gri2toKwyfJPu4H5ZjCvu3RxUxw3szZB+2Qf/XVXd6OSqMlcLq9rp+xzmWEeDW/zSerVTkw0e8z5n9o480NYf8AJjG7XGyu2gHfTPZ/UNPwgn+J7dP3Fq2t1gwTMOWPCMrfvml/DE8k+49pRhzLo4dH9z01Ptqapr0eS7dkcFG3ZWhR3clR2spRb3ZtxAF1CPKs6jd69OjNisq2qnJhtyrKG68w+Qmhl+4FNuoQtg1btb9zke2N3BRQXyisMERV8Tqw2s6hQLnnHAgmfsqCrTf7aM/+w5wHM3TWXqiP6xLpLphoamAJ/urm4RUEtKK9hZEb6WOqhPFuMGHgyoY4vzGNo/p+v9282B1hdc26rd1ldFfv2U9DBzGv1TOAZBWkbXym5asOAZByx4xs7q7fv5zM9XUJCbwTt3nWODghLU3up6L5C6Jt+SHRWs31lFfWPrWmpWeipjivOam3zHev2q+VlRnldYXx0SUH012QN7XO0ztM8xtyjmUziSRUfB09a2NUltwaodAJw7odgCZwLrl5PBKSMLOGVkQfOxxqYAm/ZUsyYYUL2Rv6X7a/lwSzkfbilvcY1BfbIYGwyoA10T8PCCbtRSM3Kg33D3MnHHap49gNU8Y+fsB5bx8a5qnrn+JE4dXdBxBpPwyg/UNzf5ri2tpGRHBet2VFIXppaamZbCUcV53gClgzXVPtnWdBrvrNk2DljwjI0NOyuZ8dAb9MlO573vnkN6qjWHmfCaAsqnu6tbTKNZu6OSbeXhp7wMzO/VHEiDfarD+ueQZv+NxQ1rtjWmDfNWek2244sscJp2paYIowpzGVWYy0VHHzy+/0CDC6g7Dk6jWbejgtL9tZTur2XJ2rLmtJlpKYwpymvR9DuuuDd9c5KvlqqqNHhLYAbXjG5scusyB9drbvLWaG70HWsMuDWem495741Nbo1pl8c7FnAVka+dOeqQPIMFT5O05q8qBeCCiQNjXBITr/Kz0zlpRH9O8o3Sbgoo/9lT3WoazdZ9Nazctp+V21qu4FPcu1eLJt9xA3szouBgLdW/KUFDMHg0BVod8weacOkbAkpTwAtCwbWbQ67V6ljztf2fvSAVErj85xqDgc93TX8wbAocnha2jLQUC57GRJNrgqskr1cap4yy6SkmelJThBEDchkxIJcLJx38w6yitoF1IQF13Y5KdlTUsqOilmXrdjWnTUsRUlPksAaawy0tRdx60CkppKaK23wg9eCxtFQhNXgsRZrXdU5LSWnOm5aaQrrvXGqK/zpunelDVv5DdmVjerBgrfOccUVkpiX+UmMm9nr3SufEYf04cVi/5mOBgLJ574EWKyeV7Khgy96a5mZHOBho0rygEhpoUlsEF38wCQaXloHGf71037nUlOAGBy3PBTdFaD7mK4c/fXBzhNSUgwHQHwwP5pMWfYrxyIKnSUrzvf7OCyYWx7gkJpmlpAjDCnIYVpDDBb5aam1DE6okTKBJRBY8TdLZsvcAK7ftJzsjldPHDIh1cYxppVe6tYb0dDbE0CSd4MIIZ40ttC8pY0yXWPA0ScdG2RpjusuCp0kqpftr+OfmcnqlpzD9KGuyNcZ0jQVPk1QWek22Z4wZQE6mdfkbY7rGgqdJKvO84Omff2eMMZGy4GmSxq7KOt7btJeM1BTOGlsY6+IYY+KYBU+TNBau3oEqnDa6gLxeUd6b0RiTVCx4mqQRnKJyvi2MYIzpJgueJinsq67nH5/sIS1FmDG+KNbFMcbEOQueJiksWrOTpoBy8sj+tlGxMabbLHiapDDPWxjBRtkaY6LBgqdJePtrGnhr425SxG18bYwx3RVXwVNEBovIn0Rkv4hUiMhfRGRIJ/MO9/KWi0i1iCwVkRPCpCsQkTkisktEakTkHRE5L0y6bBGZLSLrvXRbRORpERnW/Sc10bS4ZCcNTcpJw/vTPzcz1sUxxiSAuAmeIpINLAHGAtcA/wWMBpaKSE4HefsDK4CJwFeAK71TS0VknC9dpneP84HvAJ8HtgCvisj0kMv+Brgd+DVwIXA3cDqwWERyu/ygJurme6NsL5hko2yNMdERT+uT3QiMAI5S1Y0AIvIRsAEXEB9sJ+/NQBFwhi/vEuATYDbwRS/dF4BJwJmqusxLtwD4EPgJMMU7luXl+Ymq/jR4ExHZCcwHpgELu/3Eptuq6hpZvn4XInDeBAuexpjoiJuaJ3Ax8HYw+AGo6qfAW8AlHeSdCmwIyVsNvAlcJCJpvnQ1wHJfOgVeB04UkUHe4TQgFagIuU+59x5Pv9eEtnRtGfWNASYP6UtR716xLo4xJkHE05f8BGBVmOOrgfEd5G0C6sMcrwOygJG+dA1ewAxNB67ZF1WtBOYCt4jImSKSKyITgJ/iaqmLQ28kIrNERMO9Oii76Ybm7cdslK0xJoriKXj2A/aFOb4X6NtB3nXAaK/vEwARScFrhvWuHUzX298P6jk5JB3Al4EXcX2klbjAng7MUNVwgdocZjX1TSxduwuwVYWMMdEVT8ETIFwtTTqR73Hcsz4tIiNFZCDwc2C4dz7gvT8L7AKeEpFJ3sjbu3ADgfzpAH4AXA3cBpyBG8DUH5jf0QAmc3gsX7+LmoYmjhnch0F9smJdHGNMAomn4LmPljW/oL6Er5E2U9VPgKuAycBGYDuuNvmQl6TUS1cOXAYUAB/hAul1wCx/Oq+J9g7gW6r6gKq+oarP4EbdTgZuCFOGWaoq4V6dfH4ToeYmW6t1GmOiLJ6C52pcv2eo8cCajjKr6p+BQV76Uao6GcgFtqjqZl+6N3F9oGOAcd57A24g0T+9ZJO89/dC7rEBN2gotNnXHGZ1jU0sLikDLHgaY6IvnoLnK8BUERkRPOAtSDDNO9chVW1S1RJV/VhEjgCuAB4Lk05VdYOqrgWycdNk5qpqlZdkh/c+xZ9PRMYAfYBtkTyYib4VG3ZTVdfI+IG9GdrfWtGNMdEVT/M8fw18HXhZRO7G9X/ei1vE4IlgIhEZCnwM3KOq93jH0nHzNJfjppdMAO7E1WYf8N9ERO4DPgB2A6NwCyE0eOmD3sSNqn1ARPoC7wNDcAsl7AeeiuJzmy4ILoxwoS2MYIw5BOImeKpqtYicheunnIsbKLQY+KavRoh3PJWWtWrFrUY0E1cz3ArMAX4UZmRsEfAwUAiU4UbUfl9V9/rK0iQiZwN3ATcB9+CC7d+B7/mbgc3h19AUYNGanQCcP9GmqBhjok9aT2k0h5t/rqf9e3TfG+t38aU57zKmKJfXbz0j1sUxxsQhkYNjOcMN7IynPk9jOiU4ytZqncaYQ8WCp0kojU0BXl/tmmytv9MYc6hY8DQJ5d1Ne9lTXc/wghyOKsqLdXGMMQnKgqdJKAuC249NLG7RZ2GMMdFkwdMkjEBAfcHT+juNMYeOBU+TMP65eR9llXUc2TeLiYN6x7o4xpgEZsHTJIx5K63J1hhzeFjwNAlBVVlge3caYw4TC54mIXy4dT/b99dS3LsXxx7ZJ9bFMcYkOAueJiEcXBihmJQUa7I1xhxaFjxN3FNV5vv6O40x5lCz4Gni3prSCjbvPUBBbiYnDAu3X7oxxkSXBU8T94K1zvMmFJFqTbbGmMPAgqeJa6rKPK+/80IbZWuMOUwseJq4tqGsik92VdM3O52ThluTrTHm8LDgaeJasMn23PHFpKXaf87GmMPDvm1MXGueomLbjxljDiMLniZufbKrirU7Ksnrlca0kQWxLo4xJolY8DRxa763g8qMcUVkpNl/ysaYw8e+cUzcat5+zEbZGmMOMwueJi5t2XuAldv2k5ORymmjrcnWGHN4WfA0cSlY6zxrXBG90lNjXBpjTLKx4GniUnBhBFvL1hgTCxY8Tdwp3V/DvzaX0ys9helHDYh1cYwxSciCp4k7wSbb6WMKyc5Ii3FpjDHJyIKniTvzm0fZWpOtMSY2LHiauFJWWct7m/aSkZrCWWMLY10cY0ySsuBp4srrq3eiCqePKSCvV3qsi2OMSVIWPE1caV7LdqItjGCMiR0LniZu7K2u5+1P9pKWIswYVxTr4hhjkpgFTxM3Fq3ZQVNAOWVUAfnZ1mRrjIkdC54mbgRH2V5oCyMYY2LMgqeJC/trGnhr425SBGaMtyZbY0xsWfA0cWFxyU4ampSpI/rTPzcz1sUxxiQ5C54mLsxb6S2MYE22xpgeIK6Cp4gMFpE/ich+EakQkb+IyJBO5h3u5S0XkWoRWSoiJ4RJVyAic0Rkl4jUiMg7InJeG9fsKyIPi8hmEakTka0i8mQ3H9OEqKpr5I0NuxCB8yZY8DTGxF7cLAwqItnAEqAOuAZQ4AfAUhE5WlWr28nbH1gBVAJfAQ4A3/LyTlHVEi9dpnePAuA7wA7geuBVEZmhqst81+zrXVOBu4FNwBHAtOg9tQFYsraM+sYAJw7rS2HvXrEujjHGdD54ishxqvqvQ1mYDtwIjACOUtWNXpk+AjbgAuKD7eS9GSgCzvDlXQJ8AswGvuil+wIwCTgzGChFZAHwIfATYIrvmvcBucAkVa3wHX+u649owlnQvP2YLYxgjOkZImm2/cBrwrzOqwUebhcDbweDH4Cqfgq8BVzSQd6pwIaQvNXAm8BFIpLmS1cDLPelU+B14EQRGQQgIjnAl4DfhAROE2U19U0sXbsLgPOtv9MY00NEEjznAccDvwa2i8gvRGTSoSlWWBOAVWGOrwbGd5C3CagPc7wOyAJG+tI1eAEzNB3ARO99spdvp9ePWiMiVSLykogM76AsJgLL15dR09DEsYP7cESfrFgXxxhjgAiCp6peBAwD7gUqgK8B/xaRt0TkSyJyqDuj+gH7whzfC/TtIO86YLTX9wmAiKRwsBm2ny9dbxEZF5L/5JB0R3jvP8MF3IuBm4DjgGUikhdaABGZJSIa7tVB2ZOajbI1xvREEY22VdVtqjoLF0QvAebjAtDvgG0i8lCYwBNN4QKNdCLf47hnfVpERorIQODnQLCWGPDenwV2AU+JyCRv5O1dwOkh6YK/t0+BK1V1kao+i+s7HQJcHclDmfBqG5pYsrYMsP5OY0zP0qWpKqoaUNW/+mqj9+CaRW8BVonIMhG5PHrFBFyts1+Y430JXyP1l/cT4Cpcc+tGYDuuNvmQl6TUS1cOXIYbbfsRLpBeB8zypwP2eO9/8zfxquo7uFr5cZ1/LNOWFRt2U1XXyIQjejOkfyy62Y0xJrxozPOcABwN9MfVAvcApwHPi8gHIjIsCvcA17c5Iczx8cCajjKr6p+BQV76Uao6GTdadouqbvalexPXBzoGGOe9N+AGEv3TVxYIXxOGgzVU//1nqaqEe3VU9mTVvJbtJKt1GmN6li4FTxEpFJE7RORjXNPt54BlwOeBYmAU8ARwLPBodIrKK8BUERnhK8cw3LzKVzpzAVVtUtUSVf1YRI4ArgAeC5NOVXWDqq4FsnHTZOaqapV3fivwPnCuiDQHPxE5GegNvNe1RzRB9Y0BFq1xwdNG2RpjehppPbC0ncQiZ+PmVF4CpOOaS58EHvNPA/Gl/y3wRVVtNcnPgqsAACAASURBVIAm4oK66SEf4mqAd+NqffcCecDRwcAmIkOBj4F7VPUe71g6bp7mclyz6gTgTi/d2apa77vPfcAHwG7cHwG342qS01R1b8jvYiHwMvAbYADwQ6AKOF5VayJ4Nn/Tb6d/J4ls+fpdXDPnXY4qymPhrad3nMEYY6LIVy8iXAthJIskbMAtUiC4WtejwHOqWttOtg1ATmfv0R5VrRaRs3D9lHO9ciwGvhkMnMGiAqm0rFUrMBqYCfQBtgJzgB/5A6enCHgYKATKgBeB7/sDp1eexSLyWVx/74tANfAacHskgdOEN3+l6162WqcxpifqdM1TRA4AfwAeVdUPOplnMDBCVZd3mDiJWc2zpcamAFN+tJi91fUs/ObpHFXc7YYLY4yJSNRqnsAR3mjUTlPVLcCWSPIY8+6ne9lbXc+IghzGFOXGujjGGNNKJIskRBQ4jemq4CjbCyYVt/jrzxhjeopOB08R+aqIBEephjs/yDt/ffSKZ5JNIKAsWB1cVcimqBhjeqZIpqrMBEpVdXu4k6q6DTcQx1bXMV32weZ97KqsY3C/LCYc0TvWxTHGmLAiCZ5H4aaKtOcjYGzXi2OS3byVB7cfsyZbY0xPFUnwzAc66vesoONF2o0JKxBQFqyyheCNMT1fJMGzFLcMX3uOxq0Ha0zEPtxaTun+Wgbm9+KYI/vEujjGGNOmSILnUuB8ETk13EkROQ24ALdwgTERC9Y6z59YTEqKNdkaY3quSILnj3E7p/xNRB4UkXNFZIL3/hCwCLdp9I8PRUFNYlNV5q062N9pjDE9WacXSVDVdSLyRdyel98EvuE7Lbj+zpmqWhLdIppksHp7BVv21jAgL5PJQ63b3BjTs0WywhCq+pq3q8m1wEm4dWLLgbeBp1R1TzvZjWnTfK/Wed6EIlKtydYY08NFFDwBvAD5wCEoi0lSqsr8ld7endZka4yJA9HYDNuYblm/s4pPdlfTNzudKcP7xbo4xhjToYhrngAiciQwCMgMd15V3+hOoUxyOdhkW0xaqv09Z4zp+SIKniJyLm4/zY5WEUrtcolM0gk22drencaYeBHJwvAnAa/iBgk9ghth+wbwa2Ct9/NfcZtDG9MpH++qYt3OSnr3SuOUkQWxLo4xxnRKJG1kdwG1wImqGpymslRVvwpMBO4FzgH+FN0imkQWXBjhnPFFZKRZk60xJj5E8m11MvBKyK4qKQDqfB8oAWZHsXwmwQX7O22UrTEmnkS6MPxm38/1QE5ImreA07tbKJMcNu85wKptFeRkpHLqaGuyNcbEj0iCZxktd0wpA0aGpEkHsrpbKJMcFqx2tc6zxhXRK93GmBlj4kckwXM9LYPl28AMERkDICLFwGXAhugVzySyec0LI9goW2NMfIkkeC4AzhCR4Cz2/8PVMv8lIu/hRtwOAB6ObhFNItpeXsO/t5TTKz2FM44aEOviGGNMRCIJnk/g+jMbAFT1LeALwKe40balwM2q+nS0C2kST3CU7ZlHFZKd0aW1OowxJmYi2VWlAngn5NiLwIvRLpRJfP69O40xJt5EskjCHBG59VAWxiSHsspa3vvPXjLSUjhrbGGsi2OMMRGLpNl2JmDfdKbbFq7eiSqcPrqAvF7psS6OMcZELJLguQkLniYK5q90U1QusIURjDFxKpLg+SxwgYj07TClMW3YU1XHO5/uJS1FOGdcUayLY4wxXRJJ8LwPeB9YKiIXiYh985mILVqzk6aAMm1UAfnZ1mRrjIlPkcwRqPXeBXgZQETCpVNVtbkHJqz53ijbC2yUrTEmjkUS5N4E9FAVxCS+/QcaeGvjblJThHMnWPA0xsSvSOZ5Tj+E5TBJ4G8lO2kMKKeM7E+/nIxYF8cYY7rMNlA0h01w+7ELJtkoW2NMfLPgaQ6LytoG3tiwGxE4b4KNNTPGxLdON9uKyPc6mVRV9d4ulsckqCVry6hvDDBlWD8K83rFujjGGNMtkQwYmtXOueBAIvE+W/A0LdhatsaYRBJJs+2Zbbwuxc0BrQaeB86KchmbichgEfmTiOwXkQoR+YuIDOlk3uFe3nIRqRaRpSJyQph0Bd46vrtEpEZE3hGR8zq49ikiEhARFRGbphPiQH0jS9eVARY8jTGJIZLRtsvbOf2yiDwPvAs81+1ShSEi2cASoA64BlfD/QFu0YajVbW6nbz9gRVAJfAV4ADwLS/vFFUt8dJlevcoAL4D7ACuB14VkRmquizMtdNx27XtBCwyhLF83S5qGwIcO7gPR/TJinVxjDGm26I2YEhVV+IWT7grWtcMcSMwAvicqr6kqi8DFwNDcQGxPTcDRcBFqvq8qv4VuAgXRGf70n0BmATMVNXfqep871gJ8JM2rn07rrl6TtceK/HN85psL5xkf1sYYxJDtEfbbsZtjH0oXAy8raobgwdU9VPgLeCSDvJOBTaE5K3GLfxwka+pdSpQAyz3pVPgdeBEERnkv6iIjAS+C/w33ibhpqXahiaWlOwEbCF4Y0ziiHbwPAkXfA6FCcCqMMdXA+M7yNsE1Ic5XgdkASN96Rq8gBmaDlr/YfAY8CdVfaOD+yetNzfsprq+iYmDejO4X3asi2OMMVERyVSVtgbmpAGDcc2qpwIvRKFc4fQD9oU5vhfoaKeXdcAMEemvqnsARCQFmOK7djBdbxEZF+wH9Zwckg4RuRo4ARjbmcKLyCzg+51Jm0iaF0awWqcxJoFEMjJ0E+2vbSvABuC27hSoA+HuH3Z1+hCPA7cAT4vILbi+zu8Cw73zAe/9WdyUnKdE5HqgFLgJON2fTkT6AQ8Ad6lqWeSPkRzqGwMsWhNssrX+TmNM4ogkeD5N+OAVwNUI3wVeVtW6MGmiYR++mp9PX8LXSJup6icichXwSyDY7/lP4CFcsC/10pWLyGXAU8BHXrqPcQH13mA63CjfncALItLHOxac+Z8vIrXtjf5NFn//eDeVtY0cVZTHiAG5sS6OMcZETSRTVa49hOXojNW4fs9Q44E1HWVW1T+LyEvAGKBeVT8WkceALaq62ZfuTW8g0CggFViPG1Fbgwu4wXtOAvaEudVu3Kjjz4XcfxZtLDQhIgm5W838ld72YzbK1hiTYOJpQv8rwM9EZISqfgIgIsOAacAdnbmAqjbhpp0gIkcAVwA/DZNOcU3QiEgurj93rqpWeUm+CfQJyXYtbv7pObhaaVJrbArw+prg3p3W32mMSSyRDBgaiQtUrwUH3YScLwAuBFYEg1uU/Rr4Om5Bhrs5uAzgFtwiBcFyDMU1td6jqvd4x9Jx8zSXAxW4GuyduNrsAyHPcR/wAa4GOQpX62zw0gOgqv8OLZyITPc+LlfVxm4/bZx759O97DvQwIgBOYwpsiZbY0xiiWSqyh24QFPRxvn9wM9wwSbqvD7Es3DNqHOB3wOfAmf5aoTgBhCl0vLZFBiNC7LzcTXHOcB5qho6haUIeBg3t3OW9z5NVfdG+ZES2sFRtsWIdGZMlzHGxA9pPaWxjYQiG4D3VfX/ayfN74ETVPWoKJUvKfj7PDv779GTNQWUk360mN1Vdbz6P6cycVB+rItkjDER8f/Rr6qtagCR1DwH4aartGczcEQE1zQJ6IP/7GN3VR2D+2Ux4YjesS6OMcZEXSTBsx7o6Jswj/bngpokMG+la7K9cOJAa7I1xiSkSILnKuAz3uCbVkQkA7fYeofTRkziCgSUhatt705jTGKLJHg+AwzBLQzQ4lvR+/kF3DJ9T0eveCbe/HtrOaX7azkivxfHDg6dzWOMMYkhknmevwIuw+1gMkNEPgK24fpCjwaygb/hlsIzSWqBt/3YeTbK1hiTwDpd81TVAG4e5/24eY9TccF0Kq4/9EfAZ7x0Jgmp6sH+zkm2MIIxJnFFtMKQqjYAd3mLFIzFrbJTDqy1oGlWb69g674aBuRlMnlIRxvdGGNM/OrS8nxeoLSBQaaFYK3z/AnFpKRYk60xJnF1utlWREaKyJdEpH8b5wu88yOiVzwTL1SV+auCa9naKFtjTGKLm+X5TM+2bmcln+6upl9OBlOGh9s5zhhjEkckwXM68Dev37MV7/gi3PqzJskEtx87d3wRaamR/GdljDHxx5bnM1HRvBC8jbI1xiQBW57PdNvGsirW76yid680Th4RtkvcGGMSii3PZ7ptgVfrnDG+mIw0a7I1xiQ+W57PdJuNsjXGJBtbns90y+Y9B1i9vYLczDROHV0Q6+IYY8xhYcvzmW4JDhQ6a2whvdJTY1waY4w5PCLqoFLVBlW9C+gPTARO9d4LVPVuoElELol+MU1PNc9rsr1wkjXZGmOSR1SW5xORoSJyA/BlYCBgVZAksK28hg+3lJOVnsoZYwpjXRxjjDlsuhQ8AUQkFdf/eRNwDq4Wq7h+T5MEgtuPnTl2AFkZ9veSMSZ5RBw8vbVrbwCuBYq8w7uBJ4Dfqup/olY606MFp6icP9EWRjDGJJdOBU8RSQMuxdUyz8TVMuuBv+AGDb2sqt87VIU0PU9ZRS3v/2cfGWkpnDXWmmyNMcml3eApIqOBG4FrgAJAgH8CTwLPqupeEbHRtUlo4eodqMLpoweQm9nl1n9jjIlLHX3rrcP1Y5YBDwG/U9XVh7xUpsebt9JG2RpjkldnpqooMA/4kwVOA7Cnqo53Pt1Deqpw9riijjMYY0yC6Sh4/i/wH9wUlLdEZI2IfEdEbIRIEnt9zU4CCtNGFZCfFXapY2OMSWjtBk9V/aGqjgQuAF4ERuJWGNosIq+JyBcPQxlND2Nr2Rpjkl2nVhhS1YWqejlu4fe7cLXRC4A/4Jp1jxWRyYeslKbH2H+ggb9v3E1qijBjvAVPY0xyinR5vjJVvV9VRwEzgD/h1rk9AXhXRP4lIl87BOU0PcSikp00BpSpI/rRLycj1sUxxpiY6PLmi6q6WFWvAI4EvgOsB44Bfh6lspkeaP5KtzDCBbYwgjEmiXV752JV3a2qP1PVccBZuKZck4Aqaxt4c8NuRODcCTbK1hiTvKI6u11VlwHLonlN03MsWVtGfVOAKcP7UZjXK9bFMcaYmOl2zdMkj/krbZStMcaABU/TSQfqG1m2vgyA8y14GmOSnAVP0ynL1u2itiHAcUP6MDA/K9bFMcaYmLLgaTplnjfK9kIbZWuMMfEVPEVksIj8SUT2i0iFiPxFRIZ0Mu9wL2+5iFSLyFIROSFMugIRmSMiu0SkRkTeEZHzQtIMFJH7ROR9ryy7RGSxiJwerWftSWobmli61ppsjTEmKG6Cp4hkA0uAsbgt0v4LGA0sFZGcDvL2B1YAE4GvAFd6p5aKyDhfukzvHufj5q5+HtgCvCoi032XnAxcAbwMXI7bGLwWWCYiF3XnOXuiN9bvorq+iUmD8hncLzvWxTHGmJiLp40YbwRGAEep6kYAEfkI2IALiA+2k/dmoAg4w5d3CfAJMBsIrtH7BWAScKY37QYRWQB8CPwEmOKlWwGMUdXG4A1EZCGwGhd0X+3ms/YoC7y1bK3WaYwxTtzUPIGLgbeDwQ9AVT8F3gIu6SDvVGBDSN5q4E3gIhFJ86WrAZb70inwOnCiiAzyjpX7A6d3rBH4NzCoa4/XM9U3BlhUshOwKSrGGBMUT8FzArAqzPHVwPgO8jYB9WGO1wFZuN1igukavIAZmg5cs29YIpIBnAyUtHF+lohouFcHZY+ptz7eTWVtI2OL8xgxIDfWxTHGmB4hnoJnP2BfmON7gb4d5F0HjPb6PgEQkRQONsP286Xr7e8H9Zwcki6cWbh1fn/cQVniiq1la4wxrcVT8AS3/Vko6US+x3HP+rSIjPQ28/45MNw7H/DenwV2AU+JyCRv5O1dwOkh6VoWQGQmcAdwr6q+2blH6fkamgK8vsZrsp1kTbbGGBMUT8FzH+Frfn0JXyNtpqqfAFfhRsluBLbjapMPeUlKvXTlwGVAAfARLpBeh6tVNqfzE5HPAk8Cv1XV77dThlmqKuFe7ZU9lt75ZC/lBxoYOSCH0YXWZGuMMUHxFDxX4/o9Q40H1nSUWVX/jBvMMx4YpaqTgVxgi6pu9qV7E9cHOgYY57034AYS/dN/TRE5G/gj8CJuxG9Cmb/qYJOtSI+N8cYYc9jFU/B8BZgqIiOCB0RkGDDNO9chVW1S1RJV/VhEjsDN1XwsTDpV1Q2quhbIxk2TmauqVb57n4yb57kYuFpVwzbpxqumgLJwtbcQvDXZGmNMC9J6YGnP5C2E8CGuBng3rv/zXiAPODoY2ERkKPAxcI+q3uMdS8fN01wOVOBqsHd66c5W1Xrffe4DPgB2A6OA23F9ndNUda+XZizwd+9a1+IWSGimqm9H+GzN/wg95d/jnU/2cMWv3mZIv2yW3z7dap7GmKTi/84L170WN4skqGq1iJyF66ecixsotBj4pr9G6B1PpWWtWnGrEc0E+gBbgTnAj/yB01MEPAwUAmW4JtnvBwOnZyqur7UvsDRMceM+0sxfdbDWaYHTGGNaipuaZyLraTXPQEA55f4l7Kio5aWvTePYwX1iXSRjjDmsOqp5xlOfpzlM/rWlnB0VtRyR34tjjsyPdXGMMabHseBpWlngjbI930bZGmNMWBY8TQuqyryVrr/zQhtla4wxYVnwNC2s2lbBtvIaCvMyOX5IR6seGmNMcrLgaVqY19xkW0xKijXZGmNMOBY8TTNVbV4I3vbuNMaYtlnwNM3W7qhk054D9M/JYMqw9jaQMcaY5GbB0zQLLoxw7oQi0lLtPw1jjGmLfUOaZrZ3pzHGdI4FTwPAxrJKNpRVkZ+Vzskj+3ecwRhjkpgFTwPAfG9u54zxRaRbk60xxrTLviUN4FsI3kbZGmNMhyx4Gv6zp5o1pRXkZqZx6uiCWBfHGGN6PAueprnWefa4QjLTUmNcGmOM6fkseBobZWuMMRGy4Jnktu47wIdb95OVnsoZYwbEujjGGBMXLHgmuQVek+1ZYwvJyrAmW2OM6QwLnkkuGDxtLVtjjOk8C55JbGdFLe//Zx+ZaSmcObYw1sUxxpi4YcEziS1c7Wqdp48ZQG5mWoxLY4wx8cOCZxKb542yvXCSNdkaY0wkLHgmqd1Vdbz76V7SU4WzxhbFujjGGBNXrK0uSb2+eicBhTNGFZCflR7r4hjTY1VUVFBWVkZDQ0Osi2KiKD09ncLCQnr37t2l/BY8k9T8VbYwgjEdqaioYOfOnQwaNIisrCxEJNZFMlGgqtTU1LBt2zaALgVQa7ZNQuUH6vnHx3tITRFmjLcmW2PaUlZWxqBBg8jOzrbAmUBEhOzsbAYNGkRZWVmXrmHBMwktWrOTxoBy8oj+9M3JiHVxjOmxGhoayMrKinUxzCGSlZXV5eZ4C55JqHn7MRtla0yHrMaZuLrzb2vBM8lU1DawYsNuRODc8RY8jTGmKyx4JpklJWXUNwWYMqwfA/IyY10cY0ycufLKK7n88stjXYyYs9G2SebgKFurdRqTiDpqirzmmmt48sknu3z9J554AlXtcv5EYcEziVTXNbJs3S4AzrcpKsYkpNLS0ubPr776KjfeeGOLY20NgGpoaCA9veM53/n5+d0vZAKwZtsksmzdLuoaAxw/pA/F+b1iXRxjzCFQXFzc/OrTp0+rY/n5+axduxYR4Y9//CNnnHEGvXr14qmnnmLnzp1cccUVzdNzJk6cyO9///sW1w9ttp06dSq33nort99+O/369aO4uJi77ror4WunFjyTyLxVwbVsrdZpjIE77riDW2+9lZKSEi688EJqamqYOnUqr732GqtWreLmm2/mmmuuYcWKFe1eZ86cOeTn5/POO+/wwAMP8OMf/5iXXnrpMD1FbFizbZKobWhi6Vo3Gfi8CdbfaUxXDbvjtZjcd9P9n4n6Nb/1rW/xuc99rsWxW2+9tfnz1772NRYtWsRzzz3Hqaee2uZ1jj/+eO6++24ARo8ezeOPP87ixYu59NJLo17mnsJqnkli+fpdHKhv4ugj8xncLzvWxTHG9AAnnHBCi58bGxuZPXs2kyZNol+/fuTm5vLaa6+xefPmdq9z9NFHt/j5iCOO6PLKPfEirmqeIjIYeAiYAQjwN+Cbqtr+v6zLOxz4KXAOkA68C9yuqu+HpCsAfgJ8FsgFPgK+p6oLw1zzRuDbwHBgE/CQqj7e1ec7lBZ4CyOcb6NsjemWQ1EDjJWcnJwWP//whz/kl7/8JQ8//DATJkwgJyeHb3/729TV1bV7ndCBRiJCU1NT1Mvbk8RN8BSRbGAJUAdcAyjwA2CpiBytqtXt5O0PrAAqga8AB4BveXmnqGqJly7Tu0cB8B1gB3A98KqIzFDVZb5r3gg8AdyHC+JnA4+KiKjqY9F89u6qa2zib2t2ArYQvDGmbStWrODSSy9l5syZAAQCAdavX8/QoUNjXLKeJ26CJ3AjMAI4SlU3AojIR8AGXEB8sJ28NwNFwBm+vEuAT4DZwBe9dF8AJgFnBgOliCwAPsTVRqd4x9KAHwJzVfW7Xt6lInIEcK+I/EZVe8z+RX/fuIfKukbGFucxvCCn4wzGmKQ0ZswYXnvtNf7xj3/Qp08fHnzwQbZv327BM4x46vO8GHg7GPwAVPVT4C3gkg7yTgU2hOStBt4ELvKCYTBdDbDcl06B14ETRWSQd/hkYADwTMh95gL9gbZ71mNg3kobZWuM6djs2bM5+uijmTFjBtOnT6ewsNBWE2pDPNU8JwAvhzm+GldjbE8TUB/meB2QBYwE1nnpGrT1BKVgg/9EYJtXFoBVYcoCMB5Y2kGZDouGpgCLSoJNttbfaUwyufzyy8POtxw7dmzY4wUFBbzyyivtXvO5555r8fPbb7/dYZpEFE81z37AvjDH9wJ9O8i7Dhjt9X0CICIpeM2w3rWD6XqLyLiQ/CeHpAu+h5Znb8j5ZiIyS0Q03KuDsnfL25/sofxAA6MKcxldlHcob2WMMUkjnoInuEFCoTqzp8zjuGd9WkRGishA4Oe4UbIAAe/9WWAX8JSITBKRAhG5Czg9JF3wnj1+CY3m7ces1mmMMVETT8FzH2FqdLhaZ7gaaTNV/QS4CpgMbAS242qTD3lJSr105cBluNG2H+EC6XXALH862q5h9gs5H1NNAeX11cHgaf2dxhgTLfEUPFdzsK/RbzywpqPMqvpnYJCXfpSqTsbN49zinyeqqm/i+kDHAOO89wbcQKJ/+spCmPKM995blUdVZ6mqhHt1VPauem/TXnZX1TO0fzbjBlqTrTHGREs8Bc9XgKkiMiJ4QESGAdO8cx1S1SZVLVHVj71pJVcAreZkqrNBVdcC2bhpMnNVtcpL8g9gN64263c1rtb5ViQPdqjMXxncfmxgt3ZMN8YY01I8jbb9NfB14GURuRvX33gvsAW3WAEAIjIU+Bi4R1Xv8Y6l4+ZpLgcqcDXGO3E1yAf8NxGR+4APcMFxFHA7ruZ5ZzCNqjaIyP/iFkXYhlsk4SxcE+//qGq4kb2HVSCgLFht/Z3GGHMoxE3wVNVqETkL1085FzdoZzFueb4qX1IBUmlZq1ZgNDAT6ANsBeYAPwoT6IqAh4FCoAx4Efi+qrbox1TVx72Rst/GBdjNwNdV9dEoPG63/WvLPnZW1DGoTxZHH2n77xljTDTFTfAE8PomL+sgzSZCRuCqaiNwUSfvcV0E5XkCX623J5m/8uBattZka4wx0RVPfZ6mk1S1eYrKhZOsydYYY6LNgmcCWrltP9vKayjqnclxgztaP8IYY1r7zW9+Q58+fdr8OZz777+fUaNGRf3ePZEFzwQ0L9hkO6GYlBRrsjUmmXz2s5/lnHPOCXuupKQEEWHRokURX/eqq65i/fr13S1eC42NjYgIL7300iG/V7RZ8EwwqsqCVW6Kyvm2MIIxSeeGG25gyZIlbNq0qdW53/72twwdOpSzzz474utmZWVRWFgYhRL2rHt1lQXPBFNSWsmmPQfon5PBlOHhFmQyxiSyz3zmMxQVFfG73/2uxfGGhgbmzp3LddddR0pKCrfddhtjxowhKyuL4cOHc8cdd7S76XW4ptT77ruPoqIi8vLyuPbaazlw4ECL8++88w4zZsygoKCA/Px8TjvtNN59993m88OGDQPg0ksvRUSam3zD3evRRx9l5MiRZGRkMHr0aObMmdN8LliD/c1vfsNll11GTk4OI0eO5A9/+EPnf3ERsuCZYIK1znMnFJNqTbbGJJ20tDSuueYannzySQKBQPPxv/71r+zevZsvf/nLAPTu3Zsnn3ySkpISHnnkEZ555hnuv//+Tt/n2WefZdasWfzgBz/ggw8+YMSIETz88MMt0lRWVnLNNdfw5ptv8vbbbzNp0iQuuOAC9u1zK6q+9957APzud7+jtLQ07A4tAH/84x/55je/ybe//W1WrVrF1772NW666Sbmz5/fIt3s2bO57LLL+PDDD/n85z/Ptddey9atWzv9TJGQcNvSmMPLv7NKd/89znlwORvLqph7/RROGz2g22UzJpmVlJQwblzIJkuzYjRvetb+TifdsGEDY8aMYeHChZx77rmAq5EGAoFWASfokUce4ZFHHmHt2rWAq/3ddtttlJeXh/15ypQpTJ48mcceO7hI2/Tp09m6dSsbN25sfQPc91thYSG/+MUvuPLKK2lsbCQ9PZ0XX3yRz33uc83pQu910kknccwxx/CrX/2qOc3VV1/N1q1bWbZsWfN17r77bu69914A6uvrycvL46mnnuLKK69s83cV9t8YWkzxC7eMqtU8E8iGnZVsLKsiPyudqSP6d5zBGJOQRo8ezemnn97ctLl9+3YWLlzIDTfc0Jzm+eefZ9q0aRQXF5Obm8ttt93G5s2b27pkKyUlJZx88sktjoX+vHPnTm666SbGjBlDfn4+eXl57NmzJ6L7BO81bdq0FsdOPfVU1qxpuYz40Ucf3fw5IyODgoICysrKIrpXZ8XVIgmmfcG5neeOLyI91f4uMuaQiKAGGEs33HADJS4IfAAACxdJREFUN954I3v37uXJJ5+kX79+XHzxxQCsWLGCq666itmzZ3PuuefSp08fXnzxRe66666oluHqq6+mvLychx9+mKFDh5KZmcn06dOpr498BdNwi72EHktPT2913t90HU32DZtAmvfutIURjEl6l19+Ob169eKZZ55hzpw5fOlLX2oOLm+99RZDhw7lu9/9LieeeCKjR48OOzq3PePGjWvVRxn684oVK7jlllu48MILmTBhAjk5OezYsaP5fGpqKqmpqTQ1NXV4rxUrVrS69vjx49vIcehZzTNBbNpdTUlpBXmZaUwbVRDr4hhjYiwrK4uZM2cya9Ys9u3bx/XXX998bsyYMWzevJk//OEPTJkyhfnz5/PCCy9EdP1vfOMbXH/99UyePJnTTjuNF154gQ8++KDFFJMxY8Ywd+5cTjjhBCorK7n99tvJzMxsPi8iDBkyhMWLFzNt2jQyMzPp27f1wi633347M2fO5LjjjuOcc87htdde47nnnuOvf/1rF34z0WE1zwQRrHWePa6QzLTUGJfGGNMT3HDDDezbt49TTjmlxaCYSy+9lFtvvZVbbrmFY489lmXLljF79uyIrn3VVVdx9913c+edd3L88cezdu1avvGNb7RI8+STT1JeXs5xxx3HzJkz+cpXvsLgwYNbpHnwwQdZtGgRgwcP5sQTTwx7r8svv5yHH36Yn/3sZ0yYMIFf/vKXPPHEE1xwwQURlTmabLRtDxCN0bYXP7KCj7bu54n/msx5E6zZ1phoaGskpkkcNto2iW3dd4D/v717jbGjrOM4/v1ZbNGKLYqigorITQxeo/VCbCte0ATRpEq8QGPUiMQXmmi8kVhUojEqL0y8JIao4SbxihpJVLQiERQVUVQQYw0UERBcNCK19O+LmU2Oxz0tw57d2dl+P8nJ7D4zZ+Z/ZnP2d+aZeeZcfeMMD1y5gvVHODxFkhaa4bkMXNx22W486uHse3+7bCVpoXnB0DKw7nEPZfOzH8uGI5f2vSAlabkwPJeBYw5ewzEH93TXE0naC9ltK0lSR4anJO3GQt2hRv2bz9/W8JSkCVavXs327dvZsWPHvL+0QUtHVbFjxw62b9/O6tWr79M6HOe5BEzzW1UkTc+uXbu47bbbmJmZYefOnX2XoynaZ599WLNmDQcccAD3u9//H0fuaZyn4bkEGJ6StLR4kwRJkqbM8JQkqSPDU5KkjgxPSZI68g5DS8xc35YuSVpaPPKUJKkjw1OSpI4c57lMjI0Vte/3PnAfzp/7cDrcj/O30PvQI09JkjoyPCVJ6sjwlCSpI8NTkqSODE9JkjryJgnLxxl9F7AMuA/nz304He7H+VvQfehQFUmSOrLbVpKkjgxPSZI6MjwHLsmjk3w5yUySO5N8Nclj+q5rSJIcnOSTSX6S5F9JKskhfdc1FEk2JflKkj8nuSvJtUk+nGS/vmsbkiQvTnJJkpuT3J3kxiQXJjm679qGLMnF7Xv6Q9Ncr+E5YEkeCFwCHAVsBk4GDgd+kGR1n7UNzGHAq4A7gEt7rmWI3gHcA7wXOB74NPAW4LtJ/B9z7z0E+DnwVuBFwHuAJwKXJ3lsn4UNVZJXA09eiHV7te2wvQk4FDiyqq4HSHI18AfgzcAneqxtSH5UVQcCJHkjzT8u3XsnVNWtI79vTXI78AVgA80HPO1BVZ0PnD/aluSnwO+BTcDH+6hrqJKsBc4C3g6cN+31+6lw2F4GXD4bnABV9SfgMuDE3qoamKra1XcNQzYWnLN+1k4PWsxalqG/tdP/9FrFMH0UuKb9UDJ1huewPRH4zRzt1wCeJ1Gf1rfT3/VaxQAlWZFkZZLDgc8CNwMX9FzWoCQ5FjgFOG2htmG37bA9hOY83bjbgf0XuRYJgCQHAR8AvldVV/ZdzwBdATy9/fl64PlVdUuP9QxKkvvTfOj4WFVdu1Db8chz+Oa6y4Xf/6deJHkQ8A1gJ/D6nssZqpOBZwGvAe6kufDqkD4LGph3AQ8AzlzIjXjkOWx30Bx9jtufuY9IpQWTZF/gIpqL2NZX1Y09lzRIVTXb1X1Fku8A24B3A6f2VtRAtMP03ge8EViVZNXI7FXtRUT/qKp75rstjzyH7Rqa857jjgZ+u8i1aC/WdpV9BXgm8NKq+nXPJS0LVfV3mq7bw/quZSAOBfYFzqE5gJh9QDOk6g7gmGlsyPActouAZyU5dLah7d55bjtPWnDtWM5zgeOAE6vq8p5LWjaSHEgzjvuPfdcyEFcBG+d4QBOoG2k+jMybN4YfsPZGCL8C7gJOpzn/+UFgP+BJVfXPHssblCSb2h+Po+keOw24Fbi1qrb2VtgAJPk0zT47E/jW2Owb7b69d5J8DfgFcDXNuc4jaMYoPgJ4ZlVd12N5g5akgDOr6vSprdPwHLa2j/8s4IU0Fwp9H3hbVW3rs66had9cc9laVRsWs5ahSbINmHQHnDOqasviVTNcSd5Fc6erxwMrgRuAHwIf9v08P4anJElLgOc8JUnqyPCUJKkjw1OSpI4MT0mSOjI8JUnqyPCUJKkjw1OSpI4MT0mSOjI8JUnqyPCUJKkjw1OSpI4MT0mSOjI8JUnqyPCUJKkjw1OSpI4MT0mSOjI8JUnqyPCUtKQl2ZKkkmzouxZpluEpLXNt8OzpsaHvOqUh2afvAiQtmjN2M2/bYhUhLQeGp7SXqKotfdcgLRd220r6H6PnGJNsTvLLJHcluSXJ2UkeMeF5hyf5YpLtSXYkuan9/fAJy69IcmqSy5LMtNu4PsnndvOcTUl+muRfSW5PckGSg6b5+qV7wyNPSZO8HXgR8CXgYuBY4PXAhiTrqurW2QWTPAP4HrAfcBHwW+Ao4LXAiUmOq6orR5ZfCXwbeAFwA3AecCdwCPAK4MfAH8bqOQ14Wbv+rcA64CTgyUmeUlV3T/PFS7tjeEp7iSRbJsz6d1V9ZI72lwDrquqXI+s4C3gb8BHgDW1bgC8CDwZeV1Xnjix/EnABcE6So6tqVztrC01wfhN45WjwJVnVrmvc8cAzqurXI8ueB7waOBG4cOKLl6YsVdV3DZIWUJI9vclnqmrtyPJbgPcDZ1fVG8bWtQb4M7AKWFtVdyd5Ls2R4k+q6jlzbP9SmqPW9VX1oyQrgL8BK4HDquqmPdQ/W8+ZVXX62LyNwCXAx6vqHXt4ndLUeM5T2ktUVSY81k54ytY51jEDXAXsCzyhbX5aO71kwnpm25/aTo8C1gBX7yk4x1w5R9sN7XT/DuuR5s3wlDTJXye039xO14xN/zJh+dn2tWPT7R3r+fscbTvb6YqO65LmxfCUNMmBE9pnr7adGZvOeRUu8Mix5WZD0KtkNViGp6RJ1o83tOc8nwL8G/hd2zx7QdGGCeuZbf9FO/09TYA+KcmjplGotNgMT0mTnJzkqWNtW2i6ac8fuUL2MuBa4Ngkm0YXbn9/HnAdzUVFVNU9wKeABwCfaa+uHX3OyiQPm/JrkabKoSrSXmI3Q1UAvl5VV421fQe4LMmFNOctj20f24B3zy5UVZVkM/Bd4EtJvkFzdHkk8HLgH8ApI8NUoLlV4DrgBOC6JN9ql3s0zdjSdwKfv08vVFoEhqe093j/buZto7mKdtRZwNdoxnWeBPyTJtDeW1W3jC5YVVe0N0o4nWb85gnAbcD5wAer6tqx5XckOR44FTgF2AwEuKnd5o+7vzxp8TjOU9L/GBlXubGqfthvNdLS5DlPSZI6MjwlSerI8JQkqSPPeUqS1JFHnpIkdWR4SpLUkeEpSVJHhqckSR0ZnpIkdWR4SpLU0X8BhBnHtZCuvLoAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 455.04x367.92 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot training & validation accuracy values\n",
    "plt.plot(history.history['accuracy'])\n",
    "plt.plot(history.history['val_accuracy'])\n",
    "plt.title('Model accuracy')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Validation'], loc='lower right')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "res_test = model.predict(X_test)\n",
    "res_tc = model.predict(X_critical)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "f_tc = int(first_tc / temp_samples)\n",
    "l_tc = int(last_tc / temp_samples)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 455.04x367.92 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "acc = np.zeros(temp_val.shape[0])\n",
    "acc_std = np.zeros(temp_val.shape[0])\n",
    "\n",
    "for idx, temp in enumerate(temp_val):\n",
    "    \n",
    "    if temp < temp_val[f_tc] or temp >= temp_val[l_tc]:\n",
    "        \n",
    "        pos = np.argwhere(Y_test[:, 1] == temp)[:, 0]\n",
    "    \n",
    "        aux = res_test[pos, 0] - Y_test[pos, 0]\n",
    "        acc[idx] = np.mean(np.abs(aux))\n",
    "        acc_std[idx] = np.std(np.abs(aux))\n",
    "        \n",
    "    else:\n",
    "       \n",
    "        pos = np.argwhere(Y_critical[:, 1] == temp)[:, 0]\n",
    "    \n",
    "        aux = res_tc[pos, 0] - Y_critical[pos, 0]\n",
    "        acc[idx] = np.mean(np.abs(aux))\n",
    "        acc_std[idx] = np.std(np.abs(aux))\n",
    "    \n",
    "    \n",
    "plt.errorbar(x=temp_val, y=acc,\n",
    "             yerr=acc_std, fmt='o', capsize=5)\n",
    "\n",
    "plt.xlabel('Temperature')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 455.04x367.92 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "#%matplotlib notebook\n",
    "output_mean = np.zeros(temp_val.shape[0])\n",
    "output_std = np.zeros(temp_val.shape[0])\n",
    "\n",
    "for idx, temp in enumerate(temp_val):\n",
    "    \n",
    "    if temp < temp_val[f_tc] or temp >= temp_val[l_tc]:\n",
    "        \n",
    "        pos = np.argwhere(Y_test[:, 1] == temp)[:, 0]\n",
    "    \n",
    "        output_mean[idx] = np.mean(res_test[pos])\n",
    "        output_std[idx] = np.std(res_test[pos])\n",
    "    \n",
    "    else:\n",
    "    \n",
    "        pos = np.argwhere(Y_critical[:, 1] == temp)[:, 0]\n",
    "        \n",
    "        output_mean[idx] = np.mean(res_tc[pos])\n",
    "        output_std[idx] = np.std(res_tc[pos])\n",
    "    \n",
    "    \n",
    "plt.errorbar(x=temp_val, y=output_mean, yerr=output_std,\n",
    "             fmt='o', capsize=5, linestyle='--')\n",
    "\n",
    "plt.errorbar(x=temp_val, y=np.abs(output_mean - 1),\n",
    "             yerr=output_std, fmt='o', capsize=5, linestyle='--')\n",
    "\n",
    "plt.ylabel('Output')\n",
    "plt.xlabel('Temperature')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}