notebooks/0c - Regression with Tensorflow and Keras API.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Regression with Tensorflow 2.0 and Keras API\n",
    "Builds a non-linear regression with TensorFlow [code](../examples/regression/non_linear_regression.py)\n",
    "\n",
    "### Code Overview:\n",
    "1. Import libraries\n",
    "2. Generate data for training and testing \n",
    "3. Define Model\n",
    "4. Define Loss function\n",
    "5. Define Training procedure\n",
    "6. Run Model code \n",
    "    a. Train model\n",
    "    b. Test model\n",
    "7. Plot accuracy of results"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Regression Example"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Import libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'\n",
    "\n",
    "import math\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras.layers import Dense"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Let's create a basic dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x7f47be6331d0>"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAD4CAYAAAD8Zh1EAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nO3df5Ac9Xnn8fezsyNphG1GGJzAwFqySxEBy0h4C7hQdQZCELYLoWCMREydnZCo7AtJGRPVSWUVyJyvUKLyQVLHnaO4OCeBAAKTvSXIJzsBV65UFoeolSxLtoyAM2hEgmK0uMwOaLT7vT+mZ+nt7Z7pmemenR+fV5VKOzM9PQ+9y6PvPv18v19zziEiIt1vYK4DEBGRZCihi4j0CCV0EZEeoYQuItIjlNBFRHrE4Fx98JlnnukWL148Vx8vItKVnn/++X9zzp0V9tqcJfTFixezd+/eufp4EZGuZGY/i3pNJRcRkR6hhC4i0iOU0EVEeoQSuohIj1BCFxHpEUroIiI9QgldRKRHKKGLiPSIOZtYJCLS60bGimzbdZhj4yXOyefYsGoZa1YWUvs8JXQRkRSMjBXZ9MQBSuVJAIrjJTY9cQAgtaSukouISAq27To8ncyrSuVJtu06nNpnKqGLiKTg2HipoeeToJKLiEhC/DXzATMmQ/ZsPiefS+3zldBFRBIQrJmHJfNcNsOGVctSi0ElFxGRBITVzIMWZNNNuUroIiIJiFMbPzFRZtMTBxgZK6YSgxK6iEgC4tbGS+VJtoweTCUGJXQRkQRsWLWMXDYT69jxUjmVUXrdhG5mD5jZ62b2o4jXzcz+wsyOmNkPzezixKMUEelAI2NFLt/6NEs2PsW2XYe5eOh0LOZ7v/pk8qP0OCP0bwHX1nj9E8BS78964H+0HpaISGerdrUUx0s4KjNBd7/4BrN7W8KdmCgnHlPdhO6c+2fgjRqHXA/8javYA+TN7OykAhQR6URxulraLYkaegF41ff4qPfcLGa23sz2mtne48ePJ/DRIiJzo9UZn/lcNqFI3pVEQg8rGYX+1uGc2+6cG3bODZ911lkJfLSIyNxoZcZndsDYsvrCBKOpSCKhHwXO8z0+FziWwHlFRDpWI10tfoV8jm2fuSiVFReTmPo/CtxmZo8AlwJvOudeS+C8IiIdbf7gQEN1dAN2b7wqtXjqJnQzexi4AjjTzI4CdwFZAOfcN4CdwCeBI8AE8LtpBSsiMtdGxopsGT3IeKnxLpU0F+aCGAndOXdzndcd8IeJRSQi0qGCC3A1Iu2FuUCrLYqI1BRnSdwoGe/4Qhu2nwMldBGRSHGWxI1yy2VDfG3N8rRCC6W1XEREIrQyeejbzxdTW1UxihK6iEiEViYPpb1/aBgldBGRCK12paS5f2gYJXQRkRCbRw7ETsgWscRi2m2KQbopKiLiqXa0FBsYWRvw2UuH+PbzxRn19na0KQZphC4iwszlcBvhgK+tWc49NyynkM9hVKb333PD8tTbFIM0QhcRofmOloJXVlmzstD2BB6kEbqICM3dwJyLskotSugiIkB+YWPrk2fM5qSsUosSuoj0vZGxIr98+1RD75l0rqOSOSihi0ifGxkrcseO/ZSn4k/rh0p3S7tngtajhC4ifava2dLIGi1VDto+E7QedbmISN/wr5x4Tj7HxMlTLW303O6ZoPUooYtIXwiunBi33zxjxvtyg5yYmL2hRbtngtajkouI9IVm+sxz2Qxfv+ki7rruwln7h3ZayyJohC4ifaKZ8kiwLdFfrmnHhhWNUkIXkb5wTj4XWmYxg7B7ooV8bkbC7oSZoPWo5CIifWHDqmWzyiYAucEBspmZyyV2YjklDiV0EekLa1YWuOeG5eRzM2eETpSnwMGihdk5XVgrCSq5iEhP8y+Jm4nY5Lk85Vg4b5CxO6+ZgwiToxG6iPSs4JK4tSYQFcdLXL716Y6b/dkIJXQR6VmNtioWx0tseuJA1yZ1JXQR6VmNblYBc7O5c1KU0EWkJ42MFYnY6rOuTpvSH5duiopIT2lmX9CgTpvSH5cSuoj0jOB6Lc3IZqwre9BBJRcR6SFfffJg3WRerwxz2rzBruxBByV0EekRI2PF0BURg+qtfP5mqf45OpUSuoj0hKQ6U7q1fg5K6CLSI5LoTOnWNVyqlNBFpCe0OrJetDDbtWu4VKnLRUR6woZVy5rqcCl06NrmzYiV0M3sWuDPgQzwTefc1sDrQ8BfA3nvmI3OuZ0JxyoiEqmakLeMHmQ8xo3NXDbT9SPyoLolFzPLAPcDnwAuAG42swsCh20GdjjnVgLrgP+edKAiIvWsWVlg313XcN/aFRTyuenlcO9bu2LWc72WzCHeCP0S4Ihz7iUAM3sEuB445DvGAe/zvj4dOJZkkCIi1RmgUVvA1Xsd6LkEHhQnoReAV32PjwKXBo7ZAnzXzP4IOA24OuxEZrYeWA8wNDTUaKwi0qeCM0CrqyJWbXrih5TKU9OP/a/3ehL3i5PQwyZWBXvzbwa+5Zz7upn9O+BvzewjzrmpGW9ybjuwHWB4eLhef7+I9LmRsWJkTbxUnmTL6EF+8XaZqZBsUipPcseO/UD/JPU4Cf0ocJ7v8bnMLqncClwL4Jz7gZktAM4EXk8iSBHpPyNjRTY8tp9yWLb21Lv5OelcX43U4/ShPwcsNbMlZjaPyk3P0cAxrwC/CWBmvw4sAI4nGaiI9Jdtuw7XTOZxdfP65o2qO0J3zp0ys9uAXVRaEh9wzh00s7uBvc65UeAO4K/M7HYq5ZjPO1djrycRkRD+G5tJJpBuXd+8UbH60L2e8p2B5+70fX0IuDzZ0ESknySx9G2Ubl6fpRGa+i8iHaHR/T/j6vb1WRqhhC4iHSHJskivTyCKorVcRGTOjYwVGTBjMoFbb4V8jt0br0ogqu6jEbqIzKlq7TyJZN7N28clQQldROZUnNp5vW3jps9140V9U14Jo4QuInMqTu3cAflctuYxhXyur5M5qIYuIm0StXjWOfkcxRhJvTw5VfP1fi61VGmELiKpq9bJi96EoeriWSNjRTasWkYum6l7jrdORpdl8rls34/OQSN0EWmDsDp5qTzJ7Y/uI5cdmLFSYqMM2LL6whYj7A0aoYtI6qLq5A6YaDGZf/ayIY3OPRqhi0jq4tbJ68nnspw2f7DmJhb9TAldRFLX7AbOfrlshi2rL1QCr0EJXURSVe1uaSWZL1qY5a7rlMzrUUIXkdSMjBXZ8Ph+ypPNzwK9b+0KJfKYlNBFJBUjY0Vu37GPVmb0a7JQY9TlIiKJq/adt5LM+31dlmZohC4iLQvOAp04earltc1Pmzeo0XmDlNBFpCXBnYaSaE8EeLPOBtAym0ouItKStHYa6pdt45KkhC4iLUljA+Z+2jYuSSq5iEhTqnXzOPc9DchlB2pO88+YMeWcZoC2QAldRBoWrJvXU2/Nllw201d7f6ZFCV1EGpZE3Vwj8uQpoYtITWEbUyRRN59yjpe3fiqBCKVKCV1EIoW1JN7+6L5YdfN61MWSPCV0EZllZKzIltGDjIf0gieRzNXFkg4ldBGZYfPIAR7c80qqn6EboOlQH7qITBsZK/JQyslcC26lRwldRKbF7Stvlkot6VJCFxGgMjpPah2WKCq1pEs1dJE+5W9HPD2X5a2Tp1L9PJVa0qeELtKHgu2IYd0sSVKppT1UchHpQ2mtkBgmY6ZSS5sooYv0oTRWSITKLkN+uWyGr990kZJ5m8RK6GZ2rZkdNrMjZrYx4pibzOyQmR00s79LNkwRSVJaszS33XgRhXwOo1Iz18i8verW0M0sA9wP/BZwFHjOzEadc4d8xywFNgGXO+dOmNkH0gpYRFp35flnJT55qHrTUwl87sS5KXoJcMQ59xKAmT0CXA8c8h3zB8D9zrkTAM6515MOVESa10pHSz6X5bT5g9Pv/cXbZaYCzera0LkzxEnoBeBV3+OjwKWBY34NwMx2Axlgi3PufwdPZGbrgfUAQ0NDzcQrIjFUE3hYX3lUR4sBgxmjPDkzW5sxY3nb4DovixZmueu6CzUy7wBxErqFPBecTDYILAWuAM4F/o+ZfcQ5Nz7jTc5tB7YDDA8PpzkhTaRvNbr5RJWDWckc4MREmU1PHACYLqkoeXemOAn9KHCe7/G5wLGQY/Y458rAy2Z2mEqCfy6RKEWkrlqj8laVypNs23VYibzDxelyeQ5YamZLzGwesA4YDRwzAlwJYGZnUinBvJRkoCISrToqT3PqflqtjpKcugndOXcKuA3YBfwY2OGcO2hmd5vZau+wXcDPzewQ8AywwTn387SCFpGZ2jFRSBtSdL5YU/+dczuBnYHn7vR97YAve39EpM3SHj1r6n530FouIl2sWjdPs8OgoE2cu4YSukgXCesnD+tMSUohn2P3xqtSO78kSwldpAuE7fHZ6AqJuWyGBdkBTkzEf59uhHYXLc4l0uGqHSyNJPBCPsctlw2Rsco0kowZn/5Ygbuuu5BcNhP7PLoR2l00QhfpcM10sBTHSzy055Xp2vqkczy455Xp9VsGjFnT94N0I7T7KKGLdLhmyx618nW9ZK4bod1JCV2kw52Tz6W+16efboR2L9XQRTrchlXLGqp7t0Jllu6mEbpIhwq2KDbaoRKHf2ncc1Rm6XpK6CIdaGSsyIbH9lP2it3jpTIDwMLsABPlqUQ+I5sxtqzWsre9RCUXkQ60ZfTgdDKvmoKmkvn/2/op7lu7gnwuO/3cooVZtt2ovT57jUboInPMX1qplj0anTQUpeD1kWsN8/6ghC4yB/xrlxvvthgWx0vTm0k0KjNgTPpG9brB2X+U0EXaLLijULAlvNllcN87f1A3OPucErpIyoIllYmTp1JZu/zNUpl9d12T+Hmleyihi6QoOBpPc4KQ1l0RJXSRFKS5v2cY1csFlNBFEhcclachlx3gjNPmq14uMyihiySsHft7vl2e0norMosmFokkrB2bQqheLmE0QhdpQnAHoUULs9x1XWUafdqrI6peLlE0Qhdp0MhYkS/v2DdjNueJiTIbHt9fWYOlwdURc9kMt1w2VPOYRQuzGJWZn/fcsFz1cgmlEbpIg7765MHQDSLKk45tuw5P17bv2LGfSTf7QP/MUKhMJKruJBRl7E71l0t9GqGLxDQyVuTyrU/XXMK2Wj9fs7LAVEgyh9o7CYWp7gsqUo9G6CIBYYtlAbFaEf03K5Oqpd986Xktn0P6gxK6iE/YzM5NTxxgQXYgVivileefxeVbn+bYeIkF2dZ/Ab7lsiG+tmZ5y+eR/qCELkLtmZ2l8mSsZL70A6fx0J5XpksqpRY3oijkc0rm0hAldOl7rc7szJgx6RwvvP5WYjGpNVGaoYQufa/VmZ1hnSzNGDCYcpWRuabySzOU0KXvtWNmZ5hsxrQNnCRKCV36QljnSjWRpj2zM4x/ZqlIUpTQpedFda5ApV98w6plfOnRfanHcd/aFUrgkipNLJKeF1YjL5Un2bbrcNtiKORzSuaSulgJ3cyuNbPDZnbEzDbWOO5GM3NmNpxciCKtiaqRHxsvMTJW5I4d+1OPQR0r0g51E7qZZYD7gU8AFwA3m9kFIce9F/hj4NmkgxRpRdRSs6fnsmx64kBiXSpR8rmsRufSFnFq6JcAR5xzLwGY2SPA9cChwHH/Gfgz4E8SjVAkhqjp+tXJQsEFsaCyqXKrqTyfy/LWyVOUJ8PPlMtm2LL6whY/RSSeOAm9ALzqe3wUuNR/gJmtBM5zzv2DmUUmdDNbD6wHGBqqvVyoSC3+BH56IKkWx0tseGw/GNPPhaXbVpJ5LpuZXsY2GIsZjE+UtTWctF2chB621Nv0/wtmNgDcC3y+3omcc9uB7QDDw8Pp/p4rPSvYteJfl7yqHLa+bUKCE3/WrCwoaUtHiJPQjwL+5d7OBY75Hr8X+Ajwfass8/mrwKiZrXbO7U0qUJGqduzZGUYTgaTTxelyeQ5YamZLzGwesA4Yrb7onHvTOXemc26xc24xsAdQMpfUpLu92wAF7yaq/1fTRQuzSubS8eqO0J1zp8zsNmAXkAEecM4dNLO7gb3OudHaZxBJzshYMfQGZxIMuOeGjyppS9eKNVPUObcT2Bl47s6IY69oPSzpd5tHDvDws69OtxTGSeLVxa2a5WB6spGSunQjTf2XOVFrbZXNIwdm7bEZJ08ncR80uCyASDfR1H9pu2qXSnG8hOPdJDoyVgTg4WdfrX2ClLV7WQCRpCihS9tFra2yZfQgkNz64q2YqyV1RVqhkoukJqqsEpUsx0tlRsaK0zsAxZHWDdKo5QJEOplG6JKKsLLK7Y/uY/HGpxiwsLlqFdt2HY61y/3lHz6DXDaTSjLX9m/SrTRCl1R89cmDs8oq1eRba/RdHC/x0J5XyA5ArT2Wd7/4RgJRzqaNJ6SbKaFL4kbGipyYmD0dPy5HJZlnB4z3LBhkfKLM6bls6BT/pCiRSy9QQpfEVGvmSc3kLE85Fs4bZOzOa1h593cTOWeQGdx7k3YSkt6ghC6JCC6YlZTieInNIwdaGvFH8a+YKNILlNAlVK2JP2HSXDDrocAko2bltbSt9DgldJml3qbKYdLs226lk8WAz142xNfWLE8qHJGOpYQus9TbVDlsZ6DUGsKbkDFjyjmNwqXvKKHLLFGj7eJ4iS89um/G4w2P7WfSOdo9uTM7UJl8FFy/RWuWSz/TxCKZpZFZkuWp2Um1HbZ95iL+600ryOey089pzXLpdxqhyywbVi1LpWMlKYV8bsb2byJSoYQus1STZLVW3iGlcUDT8kVqUUKXUP6Njz+8aWcqKyDmc1neeudU3Q2ddZNTJB4l9B4R1jcOsztSGkmGI2NFtoweTG052zdLZe5du2I6xtNzWd46eYry5Lufp8k/IvGZm6O1p4eHh93evdpHOglhszSjukCgUoOOSu7VJJ7muin+OHZvvGrW57fyj5BIrzOz551zw2GvaYTeA8L6xmuVMfwTharvr46Qf/F2uS1dK1G1cH+pR0Qao4TeIYIj0yvPP4t/2P/ajJFy1IqAzczSrO4Q9M6pqel/DNoxKodKTVxlFJHkqQ+9A4RtBvHgnldmJdgTE2W+9Og+Vt793en9N6H53XXGS+XUWxODW1nkshm+fpN6xUXSoBH6HPGPyAca2HINKondv7ZKp/aN57IZPv2xAs/85Lhq4iJtoJuibeRfLzyppU+qZRiYm1p4lFo3XkWkebVuiqrk0iabRw5w+6P7pjd/SCrXnpgos+Hx/UBlhmd1Z5+5SubZjHHf2hXs3niVkrlIm6nkkoB6rXYjY0Ue2vNKajMuy5NuxqJZc0XbuInMLZVcGhTWjfLt54uz6tenzcvw2xdX6sdxt2Tzl2EWZiu/PE3U2im5g4T1lItI8tSHnpCwjR8ejNhN562Tk5GvhckOGBjTsyQnylPkshkWLcymsv1as8Jq/1pfRaQzKKHX0Uo3SlyLFlaWgA0m7lJ5kvmDA+SymY7oYKlOw4fWlhQQkXQooTOz+yTjJe3q3/4RaVprmiycNxg5OWi8VCafy85JQs9lBzjjtPmhiVsJXKTz9G1Cj2ohrCbt6t/tuMNQTZhRtfZ2zeD0M+CeGz6qxC3SRfouoYctPtWu28JRvefV0W+nTA6qbqysZC7SXfoioacxoaeeRQuzfOqjZ8+YJRnWEVO9oVhNnnfs2J9KaaeQzzFx8lTkDdbqddGEIJHu1fMJPdiZknYy99edn/nJca48/6zppP7MT47z6Y8VZiy6tSD77tyuNSsLqfST+7tQwn4LUP+4SG+I1YduZtcCfw5kgG8657YGXv8y8PvAKeA48HvOuZ/VOmcrfeiNrJl9+danY/eBx1UdzQZH+8HWwyiZAWMyMJVzwOB3Lh3i7559JfYsz+yA8Z4Fg5yYKE/fxC14vwlErZ+i9cZFulutPvS6Cd3MMsBPgd8CjgLPATc75w75jrkSeNY5N2FmXwSucM6trXXeZhN62GYOtXa1WbzxqYY/oxZ/ScKfHNuxfko+l8UMxifKSsYifarViUWXAEeccy95J3sEuB6YTujOuWd8x+8Bbmk+3NrCNnMolSfZtutwaHLLJNg7bjCrda+a2Dc9cSDVZK6ZmCJST5yEXgBe9T0+Clxa4/hbge+EvWBm64H1AENDQzFDnCmqXzvq+SRvMDqY8Q+H/2Zr2jQTU0TqiZPQg3sUQMS9RTO7BRgGPh72unNuO7AdKiWXmDHOENWvHbXJQ9JT54vjJZZsfCp0Q+O05HNZlVZEpK44Cf0ocJ7v8bnAseBBZnY18BXg4865d5IJb7awfu3sgDFx8hSLNz416+bgL98+lXgMjvZN9sllM2xZfWFbPktEuluchP4csNTMlgBFYB3wO/4DzGwl8JfAtc651xOP0qc6Uq3ejBwcqGyIXB2FV0ssxfFSqkvWtoP23hSRRtTd4MI5dwq4DdgF/BjY4Zw7aGZ3m9lq77BtwHuAx8xsn5mNphYxlaS+e+NV/MaHz6DW6rLdkMyjvgHZjGnvTRFpSKyJRc65ncDOwHN3+r6+OuG46to8coDdL77R7o+tKe6qiBkzppybbj0EZixHoIk+ItKMrpwpunnkQENrjadp0cLsjL7wel0vUT3zSt4i0qquS+jV7dya4Z8UlNQM0rE7r5n1XPCmrdZJEZF26LqEvm3X4aZr44vfn5tOpo2sbmgGYe3shZBWyeBNW83oFJF26bqEHjWBKI7dL77ByFhxeoYnMGPqfrCv3L9DT9hyA1GTffznFxFpl65L6LU2gojDP9MzmHjrLVylUbeIdLJYqy2mIcnFuRphwMtbP9XUe0VE5lqtxbnq9qF3mjUrC9xzw3IK+RxGpY5939oV3HLZUOgaBUFRSwSIiHS7riu5QHiNes3KAsMfPCMwg3Tm+2rVvUVEul1XJvQojdbERUR6SU8l9CB1m4hIP+m6GrqIiIRTQhcR6RFK6CIiPUIJXUSkRyihi4j0CCV0EZEeoYQuItIjlNBFRHrEnC3OZWbHgZ+1eJozgX9LIJwkdWJMoLga0YkxgeJqVCfGlURMH3TOnRX2wpwl9CSY2d6oVcfmSifGBIqrEZ0YEyiuRnViXGnHpJKLiEiPUEIXEekR3Z7Qt891ACE6MSZQXI3oxJhAcTWqE+NKNaaurqGLiMi7un2ELiIiHiV0EZEe0fEJ3cw+Y2YHzWzKzCLbfczsWjM7bGZHzGyj7/klZvasmb1gZo+a2bwEYjrDzL7nnfN7ZrYo5JgrzWyf78/bZrbGe+1bZvay77UVrcYUNy7vuEnfZ4/6nk/8WsWNy8xWmNkPvO/1D81sre+1xK5X1M+J7/X53n/7Ee9aLPa9tsl7/rCZrWo2hibj+rKZHfKuzT+Z2Qd9r4V+P9sQ0+fN7Ljvs3/f99rnvO/3C2b2uaRiihnXvb6Yfmpm477X0rpWD5jZ62b2o4jXzcz+wov5h2Z2se+15K6Vc66j/wC/DiwDvg8MRxyTAV4EPgTMA/YDF3iv7QDWeV9/A/hiAjH9GbDR+3oj8Kd1jj8DeANY6D3+FnBjCtcqVlzALyOeT/xaxY0L+DVgqff1OcBrQD7J61Xr58R3zH8EvuF9vQ541Pv6Au/4+cAS7zyZhK5PnLiu9P38fLEaV63vZxti+jzw3yJ+3l/y/l7kfb2oXXEFjv8j4IE0r5V33n8PXAz8KOL1TwLfAQy4DHg2jWvV8SN059yPnXOH6xx2CXDEOfeSc+4k8AhwvZkZcBXwuHfcXwNrEgjreu9ccc95I/Ad59xEAp9dS6NxTUvxWsWKyzn3U+fcC97Xx4DXgdDZcC0I/TmpEevjwG961+Z64BHn3DvOuZeBI9752hKXc+4Z38/PHuDchD676ZhqWAV8zzn3hnPuBPA94No5iutm4OGEPjuSc+6fqQzaolwP/I2r2APkzexsEr5WHZ/QYyoAr/oeH/Weez8w7pw7FXi+Vb/inHsNwPv7A3WOX8fsH6r/4v3qda+ZzU8gpkbiWmBme81sT7UMRHrXqpG4ADCzS6iMvl70PZ3E9Yr6OQk9xrsWb1K5NnHe26xGz30rldFeVdj3s10xfdr7vjxuZuc1+N4048IrSy0BnvY9nca1iiMq7kSvVUdsEm1m/wj8ashLX3HO/a84pwh5ztV4vqWY4rzfd56zgeXALt/Tm4B/oZK0tgP/Cbi7jXENOeeOmdmHgKfN7ADwi5DjYve0Jny9/hb4nHNuynu66esVPH3Ic8H/xsR/lmKIfW4zuwUYBj7ue3rW99M592LY+xOO6UngYefcO2b2BSq/2VwV871pxlW1DnjcOTfpey6NaxVHW36uOiKhO+eubvEUR4HzfI/PBY5RWQQnb2aD3mir+nxLMZnZv5rZ2c6517wE9HqNU90E/L1zruw792vel++Y2f8E/iROTEnF5ZU0cM69ZGbfB1YC36bJa5VUXGb2PuApYLP3a2n13E1fr4Con5OwY46a2SBwOpVfpeO8t1mxzm1mV1P5B/Ljzrl3qs9HfD9bTVJ1Y3LO/dz38K+AP/W994rAe7/fYjyx4/JZB/yh/4mUrlUcUXEneq16peTyHLDUKl0a86h8I0dd5a7DM1Rq2ACfA+KM+OsZ9c4V55yzanheUqvWrdcAoXfG04jLzBZVSxZmdiZwOXAoxWsVN655wN9TqTM+FngtqesV+nNSI9Ybgae9azMKrLNKF8wSYCnwf5uMo+G4zGwl8JfAaufc677nQ7+fbYrpbN/D1cCPva93Add4sS0CrmHmb6ipxuXFtozKTcYf+J5L61rFMQr8B6/b5TLgTW+gkuy1SuOOb5J/gN+m8q/YO8C/Aru8588BdvqO+yTwUyr/2n7F9/yHqPyPdwR4DJifQEzvB/4JeMH7+wzv+WHgm77jFgNFYCDw/qeBA1QS04PAexK6VnXjAn7D++z93t+3pnmtGojrFqAM7PP9WZH09Qr7OaFSvlntfb3A+28/4l2LD/ne+xXvfYeBTyT8c14vrn/0fv6r12a03vezDTHdAxz0PvsZ4Hzfe3/Pu4ZHgN9t57XyHm8Btgbel+a1ephKZ1aZSr66FfgC8AXvdQPu92I+gK9jL8lrpan/IiI9oldKLiIifU8JXUSkRyihi/JptEIAAAAfSURBVIj0CCV0EZEeoYQuItIjlNBFRHqEErqISI/4/wl0E7Fg2QRpAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Create dataset\n",
    "N = 400\n",
    "data = lambda: None\n",
    "data.x = np.linspace(-1, 1, N)\n",
    "data.y = (10 * np.exp(data.x) + 2 * np.random.rand(N)) / (10 * math.exp(1))\n",
    "plt.scatter(data.x, data.y)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Set training parameters"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Training Parameters\n",
    "num_epochs = 100"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Define model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define Network\n",
    "model = tf.keras.models.Sequential([\n",
    "    Dense(256, activation=tf.nn.relu, input_shape=[1]),\n",
    "    Dense(256, activation=tf.nn.relu),\n",
    "    Dense(1)\n",
    "])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Compile model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "dense (Dense)                (None, 256)               512       \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 256)               65792     \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 1)                 257       \n",
      "=================================================================\n",
      "Total params: 66,561\n",
      "Trainable params: 66,561\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.compile(optimizer='adam',\n",
    "              loss='mean_squared_error')\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Train model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 320 samples, validate on 80 samples\n",
      "Epoch 1/100\n",
      "320/320 [==============================] - 1s 2ms/sample - loss: 0.0772 - val_loss: 0.0041\n",
      "Epoch 2/100\n",
      "320/320 [==============================] - 0s 114us/sample - loss: 0.0166 - val_loss: 0.1440\n",
      "Epoch 3/100\n",
      "320/320 [==============================] - 0s 153us/sample - loss: 0.0060 - val_loss: 0.0019\n",
      "Epoch 4/100\n",
      "320/320 [==============================] - 0s 157us/sample - loss: 0.0017 - val_loss: 5.2227e-04\n",
      "Epoch 5/100\n",
      "320/320 [==============================] - 0s 155us/sample - loss: 6.6088e-04 - val_loss: 6.1439e-04\n",
      "Epoch 6/100\n",
      "320/320 [==============================] - 0s 163us/sample - loss: 6.2182e-04 - val_loss: 0.0027\n",
      "Epoch 7/100\n",
      "320/320 [==============================] - 0s 162us/sample - loss: 5.5471e-04 - val_loss: 6.8344e-04\n",
      "Epoch 8/100\n",
      "320/320 [==============================] - 0s 154us/sample - loss: 4.7960e-04 - val_loss: 8.8918e-04\n",
      "Epoch 9/100\n",
      "320/320 [==============================] - 0s 154us/sample - loss: 4.9501e-04 - val_loss: 8.5234e-04\n",
      "Epoch 10/100\n",
      "320/320 [==============================] - 0s 157us/sample - loss: 5.2429e-04 - val_loss: 0.0011\n",
      "Epoch 11/100\n",
      "320/320 [==============================] - 0s 158us/sample - loss: 4.6224e-04 - val_loss: 9.5142e-04\n",
      "Epoch 12/100\n",
      "320/320 [==============================] - 0s 152us/sample - loss: 4.8985e-04 - val_loss: 8.6305e-04\n",
      "Epoch 13/100\n",
      "320/320 [==============================] - 0s 151us/sample - loss: 4.9336e-04 - val_loss: 0.0011\n",
      "Epoch 14/100\n",
      "320/320 [==============================] - 0s 151us/sample - loss: 4.8130e-04 - val_loss: 8.9196e-04\n",
      "Epoch 15/100\n",
      "320/320 [==============================] - 0s 152us/sample - loss: 5.1359e-04 - val_loss: 9.7877e-04\n",
      "Epoch 16/100\n",
      "320/320 [==============================] - 0s 149us/sample - loss: 4.8045e-04 - val_loss: 0.0010\n",
      "Epoch 17/100\n",
      "320/320 [==============================] - 0s 155us/sample - loss: 4.6877e-04 - val_loss: 9.8348e-04\n",
      "Epoch 18/100\n",
      "320/320 [==============================] - 0s 164us/sample - loss: 4.7853e-04 - val_loss: 0.0010\n",
      "Epoch 19/100\n",
      "320/320 [==============================] - 0s 156us/sample - loss: 5.0106e-04 - val_loss: 9.4133e-04\n",
      "Epoch 20/100\n",
      "320/320 [==============================] - 0s 157us/sample - loss: 5.3613e-04 - val_loss: 9.6152e-04\n",
      "Epoch 21/100\n",
      "320/320 [==============================] - 0s 160us/sample - loss: 4.9365e-04 - val_loss: 0.0011\n",
      "Epoch 22/100\n",
      "320/320 [==============================] - 0s 156us/sample - loss: 4.7442e-04 - val_loss: 0.0011\n",
      "Epoch 23/100\n",
      "320/320 [==============================] - 0s 153us/sample - loss: 4.8114e-04 - val_loss: 0.0011\n",
      "Epoch 24/100\n",
      "320/320 [==============================] - 0s 156us/sample - loss: 5.4165e-04 - val_loss: 0.0012\n",
      "Epoch 25/100\n",
      "320/320 [==============================] - 0s 154us/sample - loss: 5.0626e-04 - val_loss: 0.0013\n",
      "Epoch 26/100\n",
      "320/320 [==============================] - 0s 157us/sample - loss: 5.0575e-04 - val_loss: 0.0011\n",
      "Epoch 27/100\n",
      "320/320 [==============================] - 0s 169us/sample - loss: 4.7043e-04 - val_loss: 9.1901e-04\n",
      "Epoch 28/100\n",
      "320/320 [==============================] - 0s 163us/sample - loss: 4.7122e-04 - val_loss: 9.3780e-04\n",
      "Epoch 29/100\n",
      "320/320 [==============================] - 0s 167us/sample - loss: 4.7646e-04 - val_loss: 0.0013\n",
      "Epoch 30/100\n",
      "320/320 [==============================] - 0s 159us/sample - loss: 4.7491e-04 - val_loss: 8.0009e-04\n",
      "Epoch 31/100\n",
      "320/320 [==============================] - 0s 159us/sample - loss: 4.8701e-04 - val_loss: 0.0014\n",
      "Epoch 32/100\n",
      "320/320 [==============================] - 0s 170us/sample - loss: 4.9433e-04 - val_loss: 7.8409e-04\n",
      "Epoch 33/100\n",
      "320/320 [==============================] - 0s 169us/sample - loss: 5.0442e-04 - val_loss: 8.9959e-04\n",
      "Epoch 34/100\n",
      "320/320 [==============================] - 0s 156us/sample - loss: 5.0160e-04 - val_loss: 0.0012\n",
      "Epoch 35/100\n",
      "320/320 [==============================] - 0s 163us/sample - loss: 5.0126e-04 - val_loss: 9.2366e-04\n",
      "Epoch 36/100\n",
      "320/320 [==============================] - 0s 165us/sample - loss: 4.6383e-04 - val_loss: 0.0010\n",
      "Epoch 37/100\n",
      "320/320 [==============================] - 0s 158us/sample - loss: 4.5527e-04 - val_loss: 0.0012\n",
      "Epoch 38/100\n",
      "320/320 [==============================] - 0s 163us/sample - loss: 4.7160e-04 - val_loss: 9.4593e-04\n",
      "Epoch 39/100\n",
      "320/320 [==============================] - 0s 159us/sample - loss: 4.8793e-04 - val_loss: 0.0012\n",
      "Epoch 40/100\n",
      "320/320 [==============================] - 0s 155us/sample - loss: 5.1955e-04 - val_loss: 9.7925e-04\n",
      "Epoch 41/100\n",
      "320/320 [==============================] - 0s 158us/sample - loss: 4.5714e-04 - val_loss: 0.0015\n",
      "Epoch 42/100\n",
      "320/320 [==============================] - 0s 170us/sample - loss: 5.1185e-04 - val_loss: 9.1493e-04\n",
      "Epoch 43/100\n",
      "320/320 [==============================] - 0s 160us/sample - loss: 4.7244e-04 - val_loss: 0.0012\n",
      "Epoch 44/100\n",
      "320/320 [==============================] - 0s 152us/sample - loss: 4.5766e-04 - val_loss: 8.2507e-04\n",
      "Epoch 45/100\n",
      "320/320 [==============================] - 0s 159us/sample - loss: 4.7688e-04 - val_loss: 0.0011\n",
      "Epoch 46/100\n",
      "320/320 [==============================] - 0s 157us/sample - loss: 4.6006e-04 - val_loss: 0.0017\n",
      "Epoch 47/100\n",
      "320/320 [==============================] - 0s 156us/sample - loss: 4.8253e-04 - val_loss: 6.9520e-04\n",
      "Epoch 48/100\n",
      "320/320 [==============================] - 0s 158us/sample - loss: 4.4889e-04 - val_loss: 0.0015\n",
      "Epoch 49/100\n",
      "320/320 [==============================] - 0s 159us/sample - loss: 4.5367e-04 - val_loss: 0.0010\n",
      "Epoch 50/100\n",
      "320/320 [==============================] - 0s 143us/sample - loss: 4.8770e-04 - val_loss: 0.0010\n",
      "Epoch 51/100\n",
      "320/320 [==============================] - 0s 151us/sample - loss: 4.5532e-04 - val_loss: 0.0011\n",
      "Epoch 52/100\n",
      "320/320 [==============================] - 0s 146us/sample - loss: 4.5371e-04 - val_loss: 0.0012\n",
      "Epoch 53/100\n",
      "320/320 [==============================] - 0s 147us/sample - loss: 4.8186e-04 - val_loss: 0.0011\n",
      "Epoch 54/100\n",
      "320/320 [==============================] - 0s 147us/sample - loss: 4.8831e-04 - val_loss: 0.0012\n",
      "Epoch 55/100\n",
      "320/320 [==============================] - 0s 157us/sample - loss: 4.8514e-04 - val_loss: 9.5140e-04\n",
      "Epoch 56/100\n",
      "320/320 [==============================] - 0s 151us/sample - loss: 4.7597e-04 - val_loss: 0.0019\n",
      "Epoch 57/100\n",
      "320/320 [==============================] - 0s 158us/sample - loss: 4.6937e-04 - val_loss: 8.1256e-04\n",
      "Epoch 58/100\n",
      "320/320 [==============================] - 0s 155us/sample - loss: 4.6046e-04 - val_loss: 0.0010\n",
      "Epoch 59/100\n",
      "320/320 [==============================] - 0s 150us/sample - loss: 4.7725e-04 - val_loss: 7.8477e-04\n",
      "Epoch 60/100\n",
      "320/320 [==============================] - 0s 151us/sample - loss: 4.9120e-04 - val_loss: 0.0017\n",
      "Epoch 61/100\n",
      "320/320 [==============================] - 0s 153us/sample - loss: 4.9521e-04 - val_loss: 0.0019\n",
      "Epoch 62/100\n",
      "320/320 [==============================] - 0s 158us/sample - loss: 4.9941e-04 - val_loss: 7.2252e-04\n",
      "Epoch 63/100\n",
      "320/320 [==============================] - 0s 161us/sample - loss: 5.3363e-04 - val_loss: 9.0120e-04\n",
      "Epoch 64/100\n",
      "320/320 [==============================] - 0s 156us/sample - loss: 5.2596e-04 - val_loss: 0.0024\n",
      "Epoch 65/100\n",
      "320/320 [==============================] - 0s 161us/sample - loss: 5.7217e-04 - val_loss: 0.0012\n",
      "Epoch 66/100\n",
      "320/320 [==============================] - 0s 157us/sample - loss: 5.5174e-04 - val_loss: 0.0013\n",
      "Epoch 67/100\n",
      "320/320 [==============================] - 0s 159us/sample - loss: 5.2418e-04 - val_loss: 9.6051e-04\n",
      "Epoch 68/100\n",
      "320/320 [==============================] - 0s 154us/sample - loss: 5.7900e-04 - val_loss: 0.0011\n",
      "Epoch 69/100\n",
      "320/320 [==============================] - 0s 157us/sample - loss: 4.9362e-04 - val_loss: 0.0012\n",
      "Epoch 70/100\n",
      "320/320 [==============================] - 0s 161us/sample - loss: 4.8833e-04 - val_loss: 0.0016\n",
      "Epoch 71/100\n",
      "320/320 [==============================] - 0s 156us/sample - loss: 4.4545e-04 - val_loss: 9.8938e-04\n",
      "Epoch 72/100\n",
      "320/320 [==============================] - 0s 177us/sample - loss: 4.5675e-04 - val_loss: 0.0016\n",
      "Epoch 73/100\n",
      "320/320 [==============================] - 0s 165us/sample - loss: 4.4567e-04 - val_loss: 6.9098e-04\n",
      "Epoch 74/100\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "320/320 [==============================] - 0s 156us/sample - loss: 5.2568e-04 - val_loss: 0.0013\n",
      "Epoch 75/100\n",
      "320/320 [==============================] - 0s 158us/sample - loss: 4.9513e-04 - val_loss: 0.0015\n",
      "Epoch 76/100\n",
      "320/320 [==============================] - 0s 154us/sample - loss: 5.1221e-04 - val_loss: 0.0013\n",
      "Epoch 77/100\n",
      "320/320 [==============================] - 0s 148us/sample - loss: 5.0396e-04 - val_loss: 9.3571e-04\n",
      "Epoch 78/100\n",
      "320/320 [==============================] - 0s 159us/sample - loss: 4.9962e-04 - val_loss: 8.4033e-04\n",
      "Epoch 79/100\n",
      "320/320 [==============================] - 0s 151us/sample - loss: 5.1969e-04 - val_loss: 0.0015\n",
      "Epoch 80/100\n",
      "320/320 [==============================] - 0s 157us/sample - loss: 5.6723e-04 - val_loss: 0.0017\n",
      "Epoch 81/100\n",
      "320/320 [==============================] - 0s 158us/sample - loss: 5.1441e-04 - val_loss: 0.0013\n",
      "Epoch 82/100\n",
      "320/320 [==============================] - 0s 160us/sample - loss: 4.9459e-04 - val_loss: 0.0010\n",
      "Epoch 83/100\n",
      "320/320 [==============================] - 0s 157us/sample - loss: 4.9200e-04 - val_loss: 0.0024\n",
      "Epoch 84/100\n",
      "320/320 [==============================] - 0s 155us/sample - loss: 4.9710e-04 - val_loss: 0.0014\n",
      "Epoch 85/100\n",
      "320/320 [==============================] - 0s 158us/sample - loss: 4.6709e-04 - val_loss: 0.0011\n",
      "Epoch 86/100\n",
      "320/320 [==============================] - 0s 157us/sample - loss: 4.6342e-04 - val_loss: 0.0013\n",
      "Epoch 87/100\n",
      "320/320 [==============================] - 0s 155us/sample - loss: 4.6996e-04 - val_loss: 0.0022\n",
      "Epoch 88/100\n",
      "320/320 [==============================] - 0s 152us/sample - loss: 4.9746e-04 - val_loss: 6.8635e-04\n",
      "Epoch 89/100\n",
      "320/320 [==============================] - 0s 154us/sample - loss: 5.7267e-04 - val_loss: 8.8120e-04\n",
      "Epoch 90/100\n",
      "320/320 [==============================] - 0s 155us/sample - loss: 4.9368e-04 - val_loss: 0.0020\n",
      "Epoch 91/100\n",
      "320/320 [==============================] - 0s 156us/sample - loss: 4.4431e-04 - val_loss: 7.7545e-04\n",
      "Epoch 92/100\n",
      "320/320 [==============================] - 0s 160us/sample - loss: 5.2439e-04 - val_loss: 0.0012\n",
      "Epoch 93/100\n",
      "320/320 [==============================] - 0s 159us/sample - loss: 5.3396e-04 - val_loss: 0.0022\n",
      "Epoch 94/100\n",
      "320/320 [==============================] - 0s 168us/sample - loss: 5.9803e-04 - val_loss: 7.2407e-04\n",
      "Epoch 95/100\n",
      "320/320 [==============================] - 0s 147us/sample - loss: 6.0550e-04 - val_loss: 0.0015\n",
      "Epoch 96/100\n",
      "320/320 [==============================] - 0s 156us/sample - loss: 5.7814e-04 - val_loss: 0.0016\n",
      "Epoch 97/100\n",
      "320/320 [==============================] - 0s 158us/sample - loss: 4.7321e-04 - val_loss: 8.6771e-04\n",
      "Epoch 98/100\n",
      "320/320 [==============================] - 0s 158us/sample - loss: 4.5634e-04 - val_loss: 0.0027\n",
      "Epoch 99/100\n",
      "320/320 [==============================] - 0s 156us/sample - loss: 4.8495e-04 - val_loss: 0.0018\n",
      "Epoch 100/100\n",
      "320/320 [==============================] - 0s 157us/sample - loss: 5.2397e-04 - val_loss: 9.2338e-04\n"
     ]
    }
   ],
   "source": [
    "history = model.fit(data.x, data.y, epochs=num_epochs, validation_split=0.2, shuffle=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Test model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictions = model.predict(data.x)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Plot results"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAD4CAYAAAD8Zh1EAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nO3deXhU5fn/8feTyQADAgFBTYIBVBZFNkXFpShahapFRJRFXFqVVqtWa1FQVLQqKH5/aivVUpdW6wIqhqiUaAE3FAUMEUGCbAIBJAoBJQEmyfP7Y2biMDmTmUlmQjL5vK6Li8w5Z865ORnuPLnPsxhrLSIi0vClHOwAREQkPpTQRUSShBK6iEiSUEIXEUkSSugiIkki9WBduF27drZTp04H6/IiIg3S0qVLv7fWtnfad9ASeqdOnViyZMnBuryISINkjPk23D6VXEREkoQSuohIklBCFxFJEkroIiJJQgldRCRJKKGLiCQJJXQRkSShhC4ikiQO2sAiEZFkl51XyNTcArYUl5KR5mHcoG4M7ZuZsOspoYuIJEB2XiETZi2n1FsOQGFxKRNmLQdIWFJXyUVEJAGm5hZUJvOAUm85U3MLEnZNJXQRkQTYUlwa0/Z4UMlFRCROgmvmKcZQ7rBmc0aaJ2HXV0IXEYmD0Jq5UzL3uF2MG9QtYTGo5CIiEgdONfNQzdyJTblK6CIicRBNbXxniZcJs5aTnVeYkBiU0EVE4iDa2nipt5xJOSsSEoMSuohIHIwb1A2P2xXVscWl3oS00iMmdGPMc8aY7caYr8LsN8aYvxpj1hhjvjTGnBD3KEVE6qHsvEJOnzKfzuPfYWpuASdktcZE+d773op/Kz2aFvq/gMHV7P8V0MX/ZyzwVO3DEhGp3wK9WgqLS7H4RoIuXLuDqn1bnO0s8cY9pogJ3Vr7IbCjmkMuAl6wPouANGNMerwCFBGpj6Lp1VLX4lFDzwQ2Bb3e7N9WhTFmrDFmiTFmSVFRURwuLSJycNR2xGeaxx2nSH4Wj4TuVDJy/K3DWjvdWtvPWtuvffv2cbi0iMjBUZsRn+4Uw6QhPeIYjU88Evpm4Mig1x2ALXE4r4hIvRVLr5aAVnt/IjPNw9RLeydkxsV4DP3PAW40xrwKnALsstZujcN5RUTqtaapKVHX0a/7bBbXLsnm8BV5kJWY6XMjJnRjzCvAWUA7Y8xm4F7ADWCtfRqYA5wPrAFKgN8kJFIRkXogO6+QSTkrKC6NvpfKHz6ZwbiPXmRez7M4PD1xfUYiJnRr7agI+y3wh7hFJCJST4VOwBWRtdz68cv88ZNXyDl+IPbZ58Ed/4ehAZptUUSkGtFMievIWm7/6EVu+HQm75w4iIqnpzO0X1ZCY1VCFxEJI5opcR1Zy38KXuOMT2fC2LFc8NRTkJL4mVaU0EVEwqjR4CFruXfedM5Y+hbrRlzNUU8/DSbaCQFqR5NziYiEEevgIWMrePDdafxm6Vs80+8irug9ps6SOaiFLiISVkaah8Iok3pKRTlT5v6Ny5b/j7/3H84jA67C7Nqb4AgPpIQuIuJgYvbyqFvoqbacR+Y8zrAVC3jitFE8dsZoMCah64c6xlGnVxMRqccCPVqibZUDuMvLeGPR0/RasYBHfzGGJ08bCSR+/VAnSugiItSgjzngLvfy15yp9Fr9CV/dfCdvpp+DKS4lI83DuEHdEjK8vzpK6CIixN6jpUmZl2mzJ3Pums/hscc4/pZbWJjA+KKhXi4iIsTWo6Wpdx/TZz3AuWs+J3/8g3DLLQmMLHpK6CIiQFrz6IbkN/Pu5Zk3/sKA9V+Qd/cj9J58Z4Iji54Suog0etl5hfy0tyzicc33l/L86/dx2sYvGXf+LfS9f1wdRBc9JXQRadSy8wq5bWY+3orqh/Ufsq+Ef8+8l5M2reDWC//ErJ7nkJ1XWEdRRkcJXUQarUDPlkhztLTa+xMvzribPlsLuHnI7eQcdxYW34PU+kS9XESk0QieOTEjzUPJ/rKIPVtal/7IizPvpvv2DdwwdALvdelfua+264rGmxK6iDQKof3Moxk81LZkF/+ZMZGjd2xm7LC7eP/okw7YX9cjQSNRQheRRiHWfubtf9rJyzMnctTubSx64l98trUtBL3/YIwEjUQ1dBFpFGIpj2Ts3s6Ml++g809FuObM4fQbRjN5WE8y0zwYIDPNw+RhPet8JGgkaqGLSKMQbuZEYyD4mWjHnVt46dW7aL2/hNR578FppwEwtG9mvUvgodRCF5FGYdygbnjcrirbPakpuF2+OcuP+X4jM18eTwvvPpY8+3plMm8olNBFpFEY2jeTycN6kuY5cERoibcCLJxavIGZL4/HZQxfvPAmA0cPPkiR1pxKLiKS1IKnxHWFWeS558aVTH9jEi0POxTmzeOcY445CJHWnlroIpK0Al0VA7Vzp2R+6rf5vDjzboo8rRk2agrZP9avroixUAtdRJJWpK6KA9cu5uk3H2J9mwyuGPEARbTi61nLAer9A1AnaqGLSNKqbvDQ+as+ZvqsByho35GRoydTdEgbAEq95fVuSH+0lNBFJCll5xViwuy7ZPk8/pbzCHkZ3bh85IMUe1odsL++DemPlkouIpJUIq0LOuaLd3jgvaf4qGMfxg6bSGmTZlWOqW9D+qOlhC4iSSPSuqBjP3uDO99/nveOOYUbL7qDfalNqhzjdpl6N6Q/WkroIpI07ntrhXMyt5ZbP36ZP37yCm91/wW3XngbZS7n9NeiSWqDfCAKSugikiSy8wrZWeKtusNa7lrwLNctzmZGz3OZMPhGKlKqjhgN2FXqcI4GQgldRJKCU88UYyt44N2/c/myuTx/4q+5/5zrsKb6viANtX4OSugikiRCe6a4KsqZOudxhq1YwLT+lzJ1wJW+mbiqUR+nxI2Fui2KSFIIblk3KfPy5OyHGbZiAY8MuJKpZ14VMZm3ae6ul1PixkItdBFJCuMGdWPCrOXYkj08/eZkzlq/lPvOuY7n+11U7fsy0zyMG9StQSfygKgSujFmMPAE4AKesdZOCdmfBfwbSPMfM95aOyfOsYqIhDW0byapP+4m/YrL6LtxBXcMvokZvQeFPd7jdjX4FnmoiAndGOMCpgHnApuBxcaYHGvtyqDDJgIzrbVPGWOOA+YAnRIQr4iIs+3bufCPo2FrAYsnT+NjumH8i0EH6uLBC0QnS6s8WDQt9JOBNdbadQDGmFeBi4DghG6BwNjZ1sCWeAYpIhIYAeqYkL/9lh/PPBv3lkJ+N3Qia+jmmLCTLYGHiiahZwKbgl5vBk4JOWYS8K4x5iagBfBLpxMZY8YCYwGysrJijVVEGqnQEaCFxaVM8M+KeMj6NfS4ejgt9pdy+WV/YWmH4yBof7In8WDR9HJxejQcOqnwKOBf1toOwPnAi8ZU7exprZ1ure1nre3Xvn372KMVkUYlO6+QPve9yy0zllUZAVrqLWfG09n0HTOE1IoyRoye7EvmQftvm5lPdl5hXYd90ETTQt8MHBn0ugNVSyrXAIMBrLWfGmOaAe2A7fEIUkQan+y8Qsa9lo+3ouqiFAD9N37J9Df+wq5mLRkz4i9saFu1JV5ubaNqqUfTQl8MdDHGdDbGNAFGAjkhx2wEzgEwxhwLNAOK4hmoiDQuU3MLwibzc9Z8xr9n3su2lu24ZMwjjsk8oCHPbx6riC10a22ZMeZGIBdfl8TnrLUrjDH3A0ustTnAbcA/jTG34ivHXG2tw1pPIiLVCH7wGS6BDF2xgEffeYyvjjia3wyfxM7mrSOet6HObx6rqPqh+/uUzwnZdk/Q1yuB0+Mbmog0JpGmvgW4culb3P+/f/BJVi+uGzaRPU2bR3Xuhjw/Syw0UlRE6oVq1/+0lps+eZXbPn6Jd7v056YhtzvOZe6koc/PEgsldBGpF8KVRYytYOL8Z7lmyWzeOP5sbv/VHymvZvpb8A3nT+YBROEooYvIQZedV0iKMZSHPHpLLS/j4f8+wSUrFkQ9/W1mmoeF489OZLj1lhK6iBxUgdp5aDL37N/L32dPZuC6pTz6izE8eeqIiDMmNuTl4+JBCV1EDiqn2nla6W6ef+0+em37hvGDbmRGn8HRnWt470ZTXnGihC4iB1Vo7Txj93ZemHEPR+76jhuGjie362kApHncFFezPFxmmqdRJ3NQQheROhJucq2MNA+F/qTepehbXph5Dy28e7lixF/4/MjjK9/vLa+o9vyNudQSoBWLRCThAnXyQv+AocDkWtl5hYwb1A2P28UJm7/mtZfvwGUruGz0lAOSOcCe/eH7p6d53I2+dQ5qoYtIHXCqk5d6y7l1xjI87hRO/XoR02Y/zNaWh3LlZfezOe2IqM9tgElDesQ54oZJLXQRSbhwfcwtcP4X7zJ91gOsbpfF8DFTY07ml/fPUuvcTy10EUm44Dp5JWv5/WdvMP6Df/Fhp75cP3RCxKH8aR43LZqmNspBQ9FQQheRhAss4BwouxhbwV3zn+XaJbPJOXYAt11wK16Xu9pzeNwuJg3poQReDSV0EUmoQO+WQDJ3l3uZOudxhq78IOrRn22au7n310rmkSihi0jCZOcVMu71fLzlvlGgzfeX8lT2ZM5c/wWPDLiSv/e/NOLoz8dH9FEij5ISuogkRHZeIbfOXEZgRH+bkl08//p99Ny2htsH38zM3udFPIcGC8VGCV1E4i7Q7zyQzDN3beeFmfeQuXs7v7v4Lv7XJXSd+aoa+7wsNaGELiK1FjoKtGR/WWXNvGvRBl6YeQ8e7z6uuOx+FocMGAqnRZNUtc5jpIQuIrUSutJQcPfEX6z/gmmzH6bE3ZTLLn+Ygvadoj7vrmrmbRFnSugiUiuOKw1Zy9VL3+Lu+c+wul0W115yD4WtD4vpvI1l2bh4UkIXkVoJHQXqLvdy33tPMzo/l3e79OeWC2+jpElsybkxLRsXT0roIlIjgbp58LIUbUp28VT2ZPpv+opp/S/l0QFXYE0KBvC4Uyjxhp8x0WUMFdZqBGgtKKGLSMxC6+YAPbd+w1PZD9F+TzF/vPA2ZvcYWLnPQrXJ3ON2MXlYTyXxWlJCF5GYhdbNR+Tncv97T1HUog3DL3+E5eldIp5DLfL4U0IXkWo5LUwRqJs39e7j/veeZsTy9/iwU1/++Os/s7N566jOW2Et66dckMjQGx0ldBEJy6lL4q0zlmGBDru+46k3H6Lnd2v566kjePyM0VSkuKI+t3qxxJ8SuohUkZ1XyKScFY5reFrgzHVLefytR3HZCq655G7mHRN55Gcw9WJJDCV0ETnAxOzl/GfRRsd9xlZw4yczuPXjlylo35HfX3wn37bJiPkaegCaGEroIlIpO6+Ql8Ik81Z7f+Kxt/+Pc9YuZlaPgdw56A/sdTeL+RqacCtxlNBFpFJov/KA475bx1PZD5G++3vuPvf3vNj3gojT3jpRqSWxlNBFBPC1zqssEwcM+2oeD+VOo7jZIYwcPZkvMo+t8TVUakksJXSRRiq4O2Jrj5s9+8sO2N/Uu4975v+Ty5fN5dOsntw05Ha+b9GmxtdTqSXxlNBFGqHQ7oihvVk67yhk2uwpHLd9PU+fcglTB1xJeQxdEkOp1FI3lNBFGiHHGRL9hqz8gIdyn8Sbkspvht/LgqNPqtW1XMao1FJHlNBFGqHQGRLBV2K5d94/GZ0/lyWZx3LTkNvZ2qp9TOd1u0zl+qGgOVrqWvVLbfsZYwYbYwqMMWuMMePDHHOZMWalMWaFMebl+IYpIvEUOkrzqB82k/3ibYzOn8tTpwxn5KjJMSdzgKnDe5OZ5sHgq5krmdetiC10Y4wLmAacC2wGFhtjcqy1K4OO6QJMAE631u40xsQ2k72I1KmB3dtXDh4KlFj2u9xcPfxe3q9hiSXw0FMJ/OCJpuRyMrDGWrsOwBjzKnARsDLomOuAadbanQDW2u3xDlREas6pR0twiWVx5nHcPGScY6s8zeOmRdPUyvfu3uulIqSzuhZ0rh+iSeiZwKag15uB0IkbugIYYxYCLmCStXZu6ImMMWOBsQBZWVk1iVdEohBI4E79yotLvRz1w2amzZ7CsUUb+Hv/4fy/M8ZQ7kqtUgMH3/ih4OltQ+d5adPczb2/7qGWeT0QTUJ3Gg4WOpgsFegCnAV0AD4yxhxvrS0+4E3WTgemA/Tr189pQJqI1JLT4hPBhqx8n8lzn2RfahOuHj6J94/uV7kvNJkD7CzxMmHWcoDKkoqSd/0UTULfDBwZ9LoDsMXhmEXWWi+w3hhTgC/BL45LlCISUXWtcgj0YpnO6PxcPu9wHDf/+na2tWoX1blLveVMzS1QIq/noknoi4EuxpjOQCEwEhgdckw2MAr4lzGmHb4SzLp4Bioi4UVqlR/9wyaenP0wxxZtYFr/S/l/vxgT80Ahp66OUr9ETOjW2jJjzI1ALr76+HPW2hXGmPuBJdbaHP++84wxK4FyYJy19odEBi4iP6tuoNCFX3/Iw//9K3tTm3DVpffxwVEn1ugaWpCi/otqYJG1dg4wJ2TbPUFfW+BP/j8iUsecWs/uci93LniO3yx9iyWZx3LjkDuiLrGE0tD9hkEjRUUasEDdPPRRZvruIv6ePYW+Wwt4pt9FTDnrN5S5avbfPVOLODcYSugiDYhTf/LQnikD1i3l8bf/D3e5l+svGs9/u59R4+tlpnlYOP7s2oYtdUQJXaQBcFrjM3SGxJSKcm7+5FVuXvgqBe07csPQCaxv+3Or2uN20cydws6SquuEhqMHoQ1LVHO5iMjBE+jB4rRgc0Cbkl3867VJ3LLwFd48fiB/uHEap59/Ki7/qkIuY7jkxEzu/XUPPO7oe7foQWjDoha6SD1XXQ8WgD5bCpiWPYV2JcWMH3Qjr/YeBCWW9Ys2VtbWy63lP4s2Vs7fkmKoMnw/lB6ENjxK6CL1XNiyh7Vc9cXb3DX/Wb5reSiXjJnKV0cc8/Puas4ZKZnrQWjDpIQuUs9lpHmqjP5svr+UKXP/xpCvP+R/R5/Eny68jd3NDonL9fQgtOFSDV2knhs3qNsBde9jvt9Izr9v5YJVH/PIgCu57pK745bMVWZp2NRCF6mnQrsoNnOn8Isl/2Py3L9R4m7GmBEP8GnHXrW6RvDUuBkqszR4Sugi9VB2XiHjXsvH6y92l/xYwt3zn+GKvHf4vMNx3DjkDra3PLRW13C7DJOGaNrbZKKELlIPTcpZUZnMM3dtZ9rsKfTZupp/nDyMqQOujGnU54YpF2gO80ZCCV3kIAsurQTKHoHEe+a6pTz+1qO4Ksr53cV3ktv1tJjOnenvR645zBsHJXSRgyB47nLDz10MC4tLmTBrOSkV5fxx4Svc9MkMCtp35PqhE9jQtvqE7EoxlAf1R9QDzsZHCV2kjoXOXR7aJdyzawf/eOtRBmzIY2bPX3L3udezz9004nlbNk3VA85GTgldJMFCSyol+8vCjvw8ofBrnpz9MIeW7OL2wTczs/d5UV9nV6mXZfdGf7wkHyV0kQQKbY2HWx4upaKc33/2Bn/66D8Utj6MYVc8yorDj47pWpp3RZTQRRIg0vqewY7+YRMP5k6j/6avyDl2ABPPuyHmgUKqlwsooYvEXaT1PQNa7f2Jmxe+wlVfvE2puxm3nX8rbxx/NvhnSKyOx51C2xZNVS+XAyihi8RZpNkRUyrKuezL9/jzRy/StmQ3r/YexKMDrmBH89ZRX2Ovt0LzrUgVSugicVbdohAnbfqKe+f9k+O/W8tnHXpw1WVjY66Vg+rl4kwJXaQGqht56TQ7YvruIia8/zxDvv6Qwpbt+cOQO3in+xlRlVdCqV4u4Sihi8QoO6+QP81cdsCc4jtLvIx7PR/wzY4YqKE3LdvPtZ+/yR8WzSTFWp44bRRP9b+Eve5mle/1uF1ccmJm5eITTto0d1Nc4lW9XKqlhC4So/veWuG4QIS33DI1t8BX27aWuQ/+gzvn/ZOsXd8xp+tpPHT2NWxufTihbfJSb3m1yRwg7x71L5fIlNBFohToiljdIstbikthyRKG3nYbQz/8kNWHZjF6xAN80qlP5TERFguqwlWDsow0TkroIiGcJssCInZFzNi9nXs/fQkengft2zN16C083WUg5SnRL8rsZNQpR9bq/dJ4KKGLBHEa2Tlh1nKauVPCJvND9pVw/aLXuGbJbFwphhfOGsXUPhdTdsghlHsrahXPmP5ZPDC0Z63OIY2HEroI1Y/sLPWWOyZzV0U5o/JzueXjl2hXsov/nfBL7j1lNIWtDvMdUMtknpnmUTKXmCihS6MX7cjOStZy1rol3LXgObr8sInPO/Tgt8Pv5cv0rnGLSV0TpSaU0KXRizSyM1jXog1MnP8sAzbksa5NBmMvvot3u/SvUX/yUCkGKqyvZa6uiVITSujS6FU3sjOgbcku/vTRfxiVn8tPTTzcf/Z1vHjC+Xhd7hpf1+0yTB3eW4lb4kYJXRoFp54rgUTqNLIzoEmZl6uX5nDjJzNo7t3LCydcwBOnj6LY06pW8WhNT0kEJXRJeuF6roBvrc1xg7pxy4xlB77JWgat/pQ733+OjsXbmH9UPx4ceA1r29W8C+HjI/oogUtCKaFL0nOqkZd6y5maW+CYYDvvKOSh3Cc5deNyCtplccVl9/NR5xNqFUNmmkfJXBIuqoRujBkMPAG4gGestVPCHDcceA04yVq7JG5RitRCuBr5luJSsvMKuW2mbw4WYyu4eulb3P7BC+xLdTPxvBt4pfegWg8MAtRjRepExIRujHEB04Bzgc3AYmNMjrV2ZchxLYGbgc8SEahITYWrkbf2uJkwaznl1pK1cytT5zzOKZtXMO/ok5gw6Ea2tzw0LtdP87jVOpc6EU0L/WRgjbV2HYAx5lXgImBlyHF/AR4B/hzXCEWiEG64fmCwkKHqHCq7Sr1gK7gibw4T3n+eMuPiz+ffwuvHnxN1N8Q0j5s9+8vwljvP0OJxu5g0pEct/mUi0YsmoWcCm4JebwZOCT7AGNMXONJa+7YxJmxCN8aMBcYCZGVlxR6tiF9wAm8dklQLi0sZ91o+GCq3OaXbzF3f8cicJzht45d80PkE7hh8M9tatYvq+h63i8nDejK0b2aVWIxBU93KQRFNQndqqlT+/zDGpACPAVdHOpG1djowHaBfv36xTjonAlTttRJYZCKY12l+2wBrGZWfy10LnsUCdwy+iRm9zou6VR468Gdo30wlbakXoknom4HgvlodgC1Br1sCxwPvG99/iCOAHGPMED0YlUSIZWRnqKydW3ko90nO+DafhR17cfuvbqGw9WFRvVcDgaS+iyahLwa6GGM6A4XASGB0YKe1dhdQ+XuqMeZ94M9K5pIo4QYBVcdVUc61i9/k1o9fZn9KKneddwMv9xmMNSkHHOdxp9C2RdMqdXcNBJKGIGJCt9aWGWNuBHLxdVt8zlq7whhzP7DEWpuT6CBFArLzCh0fcFanx7Y1PDz3bxz/3Vpyu/TnnnN/z3ctq9bKDTB5WC8lbWmwouqHbq2dA8wJ2XZPmGPPqn1Y0thNzF7OK59totz6Unc0STwwuVVAM+9ebv34Za5dnM0PzVvzu6F3ktvttLDvt/jKOYCSujRIGikqB0V1c6tMzF5eZY3NaFrkgWRubAXnr1rI7R/+m47F23i592CmnHU1u5sdEvEcodMCiDQkSuhS5yLNrfLKZ5uqe7sza+mxfR3nr/qY8ws+pvPOrRS0y2LEqMl8lhXbIhHVTQsgUp8poUudCze3yqScFQztm1lZZqmOq6KczjsK6fL9RnptW8OvChbSqXgrZSaFT7N68dgZl/N2919QUcNh+9FMqStS3yihS8KEK6uES5bFpV6y8wpxGXNAUje2gqN+KKTP1tX02raa3lu/oXvRBpqV7QegzKTwScfePNV/OO926c/O5q1rHXtGmqfW5xCpa8ZG0RpKhH79+tklS9SzMVk5LesWeLAZmrCDZaZ5GNi9PZ++vZCLVy7gxMKvOX7bGlru9/0Q+KmJh+VHHMPObj34wNOBr9p0YF3bDpQ2aRa32INHgYrUN8aYpdbafk771EKXhLjvrRVVyiqBFB4umTf17uPkj+dz0bS5PLB5Jd4UFysOP4o3e5xNfnpX8tO7sK5tZo3LKNFQf3NpyJTQJe6y8wrZWVJ1OH44XYs2MCo/l2Ffzaf1vj2sa5PBg2f9lpxe57D/0HYUl3hp7XE7DvGPFyVySQZK6BI3gZp5NCM5m3n3cuGqjxm1bC4nblnFPlcqc7uezit9BrHoyJ6V86pkNkkl757z6Hv/uwmJ2Rh47DKtJCTJQQld4sKpZu6ka9EGxuT9l6ErFtBqfwlr2nbgL2dfy6weAx0fZhYWlzIxe3lMLf5oqVYuyUYJXRxVN/DHSbUTZlnLad/mM/bzNzlr/VL2udy83f0MXu09iMUdekSc5fClkEFGNZWmqW0lySmhSxWRBv44ceqKmFpexvkFHzP28zc5/ru1FLVI45EBV/Jyn8EUe1pFHU9t+mEZ4PL+WTwwNLbBRSINkRK6VFHdosqB/aErAwVPtnLIvhJG5Ofy2yU5ZP5YxDeHHsm4X93M7OMGsj/VnfD4XcZQYa1a4dLoqB+6VNF5/DtRt4rdKb4+5RUWDv/xe36zJIfRy+bSan8Jn2b1ZPrJw3j/qBOrTFNbW8HXPWC75iyXJKd+6BKTcIsqO/FWWLpvX891i99kyMoPSLGWOd1OZ/rJw1ie3iVhMU69tDcAk3JWVHZnVNdDaeyU0KWKcYO6Re6xYi2nf5vP2M9nceb6L9jjbsaLfS/guX5D2Jx2RELjy0zzHLD8m4j4KKFLFYEkGaiVB1c1UsvLuGDVR4z9/E16bF/H9hZteGTAlbzU51fs8rRMeGwet+vnur2IHEAJXRwFL3x89IQ5ePbuYWT+XH67JIeMH7+Py4PONI+bPfvKql/QGT3kFImWEnqScOo3DlV7pMSSDLPzCvn7i+8zbuGsygedn2T14s5Bf+CDODzo3FXq5bERfSpjbO1xs2d/Gd7ynxO8Bv+IRE+9XJKA0yjNcL1AwFeDDpfcs/MKmZSzgvRvC7j28zcZ8vWHGGuZ0/0M/nnSxXF90JmZ5mHh+LOrXL82P4REkp16uSQ5p37j1ZUxggcKBSBw48oAAA8mSURBVN6/pbiU1s1S6bVqMX/9bBYDNuSxx92MF064kOf7DWFz68PjGnO4WnhwqUdEYqOEXk+EtkwHdm/P2/lbD5hhMFy3vJqsrhNYIWhfWQVmz09c9vWHXPXF2xy3fT3bW7Th4TOv4qU+v4pqHc5YuYxRGUUkAVRyqQeindgqIDSxnz5lftT9xrGWDru307dwFX23FHDCllUc9906mlSUsapdR549aSizjzsrbiM6gwaQAqqJi9SWSi71UHCLPKWaFXyc7CzxHjC3SnX9xj3799Jr2zeVybvvllW031MMQGlqU75M78Jz/YbwXpf+LM08NuJEWbHwuF1ccmImC1YVqSYuUgfUQq9DwfOFh7ZcayrQWgeYOncVTdev5bSib+j+7Ur6Fq6iW9EGUm0FAOvaZJCX0Y28jO7kZXanoF1HylyJ+Zle3YNXEak5tdDrgYnZy3lp0cbKJB6PZN5y3x56rF/NhnkvMrK8kAVf5dFkl6/1/WMTD8vSu/H3/pfyRWZ38tO7xmXx5Eg0l4rIwaOEHgeRutpl5xUekMxrIqWinGN+2ETfLQW+8knhKo75YRMpWCowfNPuSBZ0PNnX+s7oxtpDOyR07U0nmktF5OBSySVGTr1R3lhaWKV+3aKJi4tP8NWPo31gGVyG6VS6g55bVnPs5lX03rqaXlu/qVz5fmezluRldOOLzO7kZXTny/Qu/Ni0RRz/lbFz6lMuIvGnkkucOC388J8wq+ns2V8edl+olvv20Pe7NfTaupqehQX03rqaI37aAYDXlcrK9p3J7jGQL/yt7w1tMuL68DIWTrV/za8iUj8ooUdQm94oTpqUeeletJ7eW1fTZ+tqem9ZzTE7NlfuX9s2k0869iY/vSv56V3ZktWVH01q1F0aEynQ5RBqN6WAiCSGEjoH9j5x+ZN24O/gFmmsydzYCjrv2ELvrasrE/ix29fRtLwMgKIWaSxL78a8fufxcVpn8tO7VB3IUwZpnpSDktA97hTatmjqmLiVwEXqn0ab0MN1IQwk7cDfsaTw9j/t8LW6/S3v3tu+odW+PQD81MTD8iOO4fl+F7HM3/re2rIdGIPBt6jE7jC19uDRonXFAJOH9VLiFmlAGl1CD0w+FZwka1JEOWRfCT23ralsfffeupqMH78HwJviYlX7TuQcO4D89K4sS+/K2kM7YFNcjtcKtH5jGS2aSIGFlZXMRRqWRpHQazugx13upVvRt5U1795bV1d2GQRY3yadz4/sUVn3XnHYUTRvfQgX9Ernk6BRkk49YgIPFAPJ87aZ+bWu0zvJTPNQsr+MnSXOrf3AfdGAIJGGK+kTemjPlIip0lo67dzy80PLravp8d06mpb7EuH3zVuTn96Vt4/9BcvSu/JleheKPa0q3+5xp9DOX3desKqIgd3bVw59X7CqiEtOzDxg0q1m7p/nFB/aN5NbZiyL67/fF9PPvVCcfgtQ/3GR5BBVP3RjzGDgCcAFPGOtnRKy/0/AtUAZUAT81lr7bXXnrE0/9FjmzI40cVW7PTsra959tq6m17ZvSNv7EwAl7qYsP/yYypZ3fkZXNrc6DGMMlqpd+NwpBgwHLNDgxJViKA+Z3jbFwOhTsnj5s42Oc5g7cacYDmmWys4Sb+VD3Ez/bwLh5k/RfOMiDVt1/dAjJnRjjAtYDZwLbAYWA6OstSuDjhkIfGatLTHGXA+cZa0dUd15a5rQnWYmrG4Gv07j36n8uvn+0ip17w67iwAoMymsbt+x8oFlfnpXvmmXRXnIaMvgkkRwcmztcbN7rzfqZFwTaR43xkBxiVfJWKSRqu3AopOBNdbadf6TvQpcBFQmdGvtgqDjFwFjah5u9ZwWcyj1ljM1t+DA5Ob1wldfMWbZXHpu8Q3W6fLDJlz+iao2tj6cvIzuPH/iEPIzurLisKMpbdKs2msbqNJ1L5DYJ8xantBkrpGYIhJJNAk9E9gU9HozcEo1x18D/NdphzFmLDAWICsrK8oQD+S4mIO1pK5fB698B59/7vvzxRewdy8PADs8rchP78Lcbqf5695d2VGDiaosHPCDI/hha6JpJKaIRBJNQncaY+7YFjXGjAH6AWc67bfWTgemg6/kEmWMB8hI8/BD0U5O3bicPlt+Lp202fuj78zNmsGJJ8L118PJJ3Phon181fTQuA2VLywupfP4dxwXNE6UNI9bpRURiSiahL4ZODLodQdgS+hBxphfAncBZ1pr98UnvKrGDerGE8/P5/nX76PcpLC6XRbvdTuVgqxj+fTQo1nbviP7Uly+h4PN2lPQfBPxroVY6m6wj8ftYtKQHnVyLRFp2KJJ6IuBLsaYzkAhMBIYHXyAMaYv8A9gsLV2e9yjDDK0bybYgVy//3E+bJ7J/mYevBVVjyssLq31lLUHm9beFJFYpEQ6wFpbBtwI5AJfAzOttSuMMfcbY4b4D5sKHAK8ZoxZZozJSVjEwNATOvDU03+kz7GZjsm8MvZEBhEn4b4Bbpfh/y7TQhEiEr2oBhZZa+cAc0K23RP09S/jHFdEE7OXs3Dtjrq+bLU8bldUQ/ddxlBhbWXXQ+CA6Qg00EdEaqJBjhSdmL086rnGE61Nc/cB/cIj9XoJ12deyVtEaqvBJfTAcm41ETwoKNII0mjl3XNelW2hA580T4qI1IUGl9Cn5hbUuDbe6VBPZTKNZXZDY8BpQG1mmqfKtsD5NbxeROpag0vojgOLorRw7Q6y8worR3gCBwzdD+1XHrxCj9N0A+EG+wSfX0SkrjS4hJ6R5qlVqSR4pGdo4o00cZVa3SJSn0U122IixHNyrlgYYP2UC2r0XhGRg626ybki9kOvb4b2zWTysJ5kpnkw+OrYj4/ow5j+WY5zFITKcKh7i4gkgwZXcgHnGvXQvpn069i2siySmkKVQUfV1b1FRBq6BpnQw4m1Ji4ikkySKqGHUm8TEWlMGlwNXUREnCmhi4gkCSV0EZEkoYQuIpIklNBFRJKEErqISJJQQhcRSRJK6CIiSeKgTc5ljCkCvq3ladoB38chnHiqjzGB4opFfYwJFFes6mNc8Yipo7W2vdOOg5bQ48EYsyTcrGMHS32MCRRXLOpjTKC4YlUf40p0TCq5iIgkCSV0EZEk0dAT+vSDHYCD+hgTKK5Y1MeYQHHFqj7GldCYGnQNXUREftbQW+giIuKnhC4ikiTqfUI3xlxqjFlhjKkwxoTt7mOMGWyMKTDGrDHGjA/a3tkY85kx5htjzAxjTJM4xNTWGPOe/5zvGWPaOBwz0BizLOjPXmPMUP++fxlj1gft61PbmKKNy39cedC1c4K2x/1eRRuXMaaPMeZT//f6S2PMiKB9cbtf4T4nQfub+v/ta/z3olPQvgn+7QXGmEE1jaGGcf3JGLPSf2/mGWM6Bu1z/H7WQUxXG2OKgq59bdC+q/zf72+MMVfFK6Yo43osKKbVxpjioH2JulfPGWO2G2O+CrPfGGP+6o/5S2PMCUH74nevrLX1+g9wLNANeB/oF+YYF7AWOApoAuQDx/n3zQRG+r9+Grg+DjE9Aoz3fz0eeDjC8W2BHUBz/+t/AcMTcK+iigv4Kcz2uN+raOMCugJd/F9nAFuBtHjer+o+J0HH3AA87f96JDDD//Vx/uObAp3953HF6f5EE9fAoM/P9YG4qvt+1kFMVwNPhvm8r/P/3cb/dZu6iivk+JuA5xJ5r/znHQCcAHwVZv/5wH8BA/QHPkvEvar3LXRr7dfW2oIIh50MrLHWrrPW7gdeBS4yxhjgbOB1/3H/BobGIayL/OeK9pzDgf9aa0vicO3qxBpXpQTeq6jistauttZ+4/96C7AdcBwNVwuOn5NqYn0dOMd/by4CXrXW7rPWrgfW+M9XJ3FZaxcEfX4WAR3idO0ax1SNQcB71tod1tqdwHvA4IMU1yjglThdOyxr7Yf4Gm3hXAS8YH0WAWnGmHTifK/qfUKPUiawKej1Zv+2Q4Fia21ZyPbaOtxauxXA//dhEY4fSdUP1YP+X70eM8Y0jUNMscTVzBizxBizKFAGInH3Kpa4ADDGnIyv9bU2aHM87le4z4njMf57sQvfvYnmvTUV67mvwdfaC3D6ftZVTJf4vy+vG2OOjPG9iYwLf1mqMzA/aHMi7lU0wsUd13tVLxaJNsb8DzjCYddd1trZ0ZzCYZutZnutYorm/UHnSQd6ArlBmycA2/AlrenAHcD9dRhXlrV2izHmKGC+MWY5sNvhuKj7tMb5fr0IXGWtrfBvrvH9Cj29w7bQf2PcP0tRiPrcxpgxQD/gzKDNVb6f1tq1Tu+Pc0xvAa9Ya/cZY36P7zebs6N8byLjChgJvG6tLQ/aloh7FY06+VzVi4Rurf1lLU+xGTgy6HUHYAu+SXDSjDGp/tZWYHutYjLGfGeMSbfWbvUnoO3VnOoy4E1rrTfo3Fv9X+4zxjwP/DmamOIVl7+kgbV2nTHmfaAv8AY1vFfxissY0wp4B5jo/7U0cO4a368Q4T4nTsdsNsakAq3x/SodzXtrKqpzG2N+ie8H5JnW2n2B7WG+n7VNUhFjstb+EPTyn8DDQe89K+S979cynqjjCjIS+EPwhgTdq2iEizuu9ypZSi6LgS7G10ujCb5vZI71PXVYgK+GDXAVEE2LP5Ic/7miOWeVGp4/qQXq1kMBxyfjiYjLGNMmULIwxrQDTgdWJvBeRRtXE+BNfHXG10L2xet+OX5Oqol1ODDff29ygJHG1wumM9AF+LyGccQclzGmL/APYIi1dnvQdsfvZx3FlB70cgjwtf/rXOA8f2xtgPM48DfUhMblj60bvoeMnwZtS9S9ikYOcKW/t0t/YJe/oRLfe5WIJ77x/ANcjO+n2D7gOyDXvz0DmBN03PnAanw/be8K2n4Uvv94a4DXgKZxiOlQYB7wjf/vtv7t/YBngo7rBBQCKSHvnw8sx5eY/gMcEqd7FTEu4DT/tfP9f1+TyHsVQ1xjAC+wLOhPn3jfL6fPCb7yzRD/1838//Y1/ntxVNB77/K/rwD4VZw/55Hi+p//8x+4NzmRvp91ENNkYIX/2guA7kHv/a3/Hq4BflOX98r/ehIwJeR9ibxXr+DrmeXFl6+uAX4P/N6/3wDT/DEvJ6jHXjzvlYb+i4gkiWQpuYiINHpK6CIiSUIJXUQkSSihi4gkCSV0EZEkoYQuIpIklNBFRJLE/wc14O/aFyEToAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot Results\n",
    "plt.scatter(data.x, data.y)\n",
    "plt.plot(data.x, predictions, 'r')\n",
    "plt.show()"
   ]
  }
 ],
 "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": "3.6.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}