Visualize Architecture.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "from matplotlib import pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from os import listdir\n",
    "from os.path import join, exists\n",
    "import matplotlib"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "params = {'backend': 'ps',\n",
    "#           'text.latex.preamble': [r'\\usepackage{gensymb}'],\n",
    "          'axes.labelsize': 8, # fontsize for x and y labels (was 10)\n",
    "          'axes.titlesize': 8,\n",
    "          'font.size': 8, # was 10\n",
    "          'legend.fontsize': 8, # was 10\n",
    "          'xtick.labelsize': 8,\n",
    "          'ytick.labelsize': 8,\n",
    "          'font.family': 'serif',\n",
    "}\n",
    "\n",
    "matplotlib.rcParams.update(params)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Unnamed: 0</th>\n",
       "      <th>eval-saturation_classifier-2</th>\n",
       "      <th>eval-saturation_classifier-6</th>\n",
       "      <th>eval-saturation_features-0</th>\n",
       "      <th>eval-saturation_features-11</th>\n",
       "      <th>eval-saturation_features-15</th>\n",
       "      <th>eval-saturation_features-18</th>\n",
       "      <th>eval-saturation_features-22</th>\n",
       "      <th>eval-saturation_features-25</th>\n",
       "      <th>eval-saturation_features-4</th>\n",
       "      <th>...</th>\n",
       "      <th>train-saturation_features-0</th>\n",
       "      <th>train-saturation_features-11</th>\n",
       "      <th>train-saturation_features-15</th>\n",
       "      <th>train-saturation_features-18</th>\n",
       "      <th>train-saturation_features-22</th>\n",
       "      <th>train-saturation_features-25</th>\n",
       "      <th>train-saturation_features-4</th>\n",
       "      <th>train-saturation_features-8</th>\n",
       "      <th>training_accuracy</th>\n",
       "      <th>training_loss</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>76.56</td>\n",
       "      <td>50.0</td>\n",
       "      <td>3.12</td>\n",
       "      <td>16.80</td>\n",
       "      <td>13.87</td>\n",
       "      <td>35.35</td>\n",
       "      <td>52.93</td>\n",
       "      <td>61.33</td>\n",
       "      <td>3.91</td>\n",
       "      <td>...</td>\n",
       "      <td>3.12</td>\n",
       "      <td>22.66</td>\n",
       "      <td>17.97</td>\n",
       "      <td>44.14</td>\n",
       "      <td>66.02</td>\n",
       "      <td>73.44</td>\n",
       "      <td>5.47</td>\n",
       "      <td>7.03</td>\n",
       "      <td>0.5575</td>\n",
       "      <td>0.005392</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>78.91</td>\n",
       "      <td>50.0</td>\n",
       "      <td>4.69</td>\n",
       "      <td>19.14</td>\n",
       "      <td>16.02</td>\n",
       "      <td>40.43</td>\n",
       "      <td>60.94</td>\n",
       "      <td>68.95</td>\n",
       "      <td>4.69</td>\n",
       "      <td>...</td>\n",
       "      <td>4.69</td>\n",
       "      <td>25.00</td>\n",
       "      <td>19.14</td>\n",
       "      <td>45.70</td>\n",
       "      <td>67.77</td>\n",
       "      <td>75.00</td>\n",
       "      <td>6.25</td>\n",
       "      <td>8.20</td>\n",
       "      <td>0.6132</td>\n",
       "      <td>0.005160</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>79.69</td>\n",
       "      <td>50.0</td>\n",
       "      <td>4.69</td>\n",
       "      <td>20.31</td>\n",
       "      <td>16.80</td>\n",
       "      <td>41.60</td>\n",
       "      <td>63.09</td>\n",
       "      <td>71.09</td>\n",
       "      <td>5.47</td>\n",
       "      <td>...</td>\n",
       "      <td>4.69</td>\n",
       "      <td>26.17</td>\n",
       "      <td>20.12</td>\n",
       "      <td>46.68</td>\n",
       "      <td>68.36</td>\n",
       "      <td>75.39</td>\n",
       "      <td>7.03</td>\n",
       "      <td>9.38</td>\n",
       "      <td>0.6481</td>\n",
       "      <td>0.004980</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>79.69</td>\n",
       "      <td>50.0</td>\n",
       "      <td>6.25</td>\n",
       "      <td>21.48</td>\n",
       "      <td>17.38</td>\n",
       "      <td>42.38</td>\n",
       "      <td>64.06</td>\n",
       "      <td>71.88</td>\n",
       "      <td>7.03</td>\n",
       "      <td>...</td>\n",
       "      <td>6.25</td>\n",
       "      <td>27.34</td>\n",
       "      <td>20.90</td>\n",
       "      <td>47.07</td>\n",
       "      <td>68.75</td>\n",
       "      <td>75.59</td>\n",
       "      <td>8.59</td>\n",
       "      <td>10.16</td>\n",
       "      <td>0.6718</td>\n",
       "      <td>0.004738</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>79.69</td>\n",
       "      <td>50.0</td>\n",
       "      <td>6.25</td>\n",
       "      <td>22.27</td>\n",
       "      <td>17.58</td>\n",
       "      <td>42.77</td>\n",
       "      <td>64.84</td>\n",
       "      <td>72.27</td>\n",
       "      <td>7.81</td>\n",
       "      <td>...</td>\n",
       "      <td>7.81</td>\n",
       "      <td>28.12</td>\n",
       "      <td>21.29</td>\n",
       "      <td>47.27</td>\n",
       "      <td>68.75</td>\n",
       "      <td>75.78</td>\n",
       "      <td>8.59</td>\n",
       "      <td>10.94</td>\n",
       "      <td>0.6832</td>\n",
       "      <td>0.004636</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>5</td>\n",
       "      <td>79.69</td>\n",
       "      <td>50.0</td>\n",
       "      <td>6.25</td>\n",
       "      <td>23.05</td>\n",
       "      <td>17.97</td>\n",
       "      <td>43.36</td>\n",
       "      <td>65.23</td>\n",
       "      <td>72.66</td>\n",
       "      <td>8.59</td>\n",
       "      <td>...</td>\n",
       "      <td>7.81</td>\n",
       "      <td>28.91</td>\n",
       "      <td>21.68</td>\n",
       "      <td>47.66</td>\n",
       "      <td>68.95</td>\n",
       "      <td>75.78</td>\n",
       "      <td>9.38</td>\n",
       "      <td>11.33</td>\n",
       "      <td>0.7002</td>\n",
       "      <td>0.004503</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>6</td>\n",
       "      <td>79.69</td>\n",
       "      <td>50.0</td>\n",
       "      <td>6.25</td>\n",
       "      <td>23.83</td>\n",
       "      <td>18.16</td>\n",
       "      <td>43.55</td>\n",
       "      <td>65.43</td>\n",
       "      <td>72.85</td>\n",
       "      <td>8.59</td>\n",
       "      <td>...</td>\n",
       "      <td>9.38</td>\n",
       "      <td>29.30</td>\n",
       "      <td>21.88</td>\n",
       "      <td>47.66</td>\n",
       "      <td>68.95</td>\n",
       "      <td>75.78</td>\n",
       "      <td>10.16</td>\n",
       "      <td>12.11</td>\n",
       "      <td>0.7104</td>\n",
       "      <td>0.004389</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>7</td>\n",
       "      <td>79.30</td>\n",
       "      <td>50.0</td>\n",
       "      <td>6.25</td>\n",
       "      <td>24.22</td>\n",
       "      <td>18.36</td>\n",
       "      <td>43.75</td>\n",
       "      <td>65.62</td>\n",
       "      <td>73.05</td>\n",
       "      <td>9.38</td>\n",
       "      <td>...</td>\n",
       "      <td>9.38</td>\n",
       "      <td>29.30</td>\n",
       "      <td>22.07</td>\n",
       "      <td>47.85</td>\n",
       "      <td>68.95</td>\n",
       "      <td>75.78</td>\n",
       "      <td>10.16</td>\n",
       "      <td>12.11</td>\n",
       "      <td>0.7269</td>\n",
       "      <td>0.004234</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>8</td>\n",
       "      <td>79.30</td>\n",
       "      <td>50.0</td>\n",
       "      <td>7.81</td>\n",
       "      <td>25.00</td>\n",
       "      <td>18.55</td>\n",
       "      <td>43.95</td>\n",
       "      <td>65.82</td>\n",
       "      <td>73.24</td>\n",
       "      <td>10.16</td>\n",
       "      <td>...</td>\n",
       "      <td>9.38</td>\n",
       "      <td>29.69</td>\n",
       "      <td>22.27</td>\n",
       "      <td>48.05</td>\n",
       "      <td>69.14</td>\n",
       "      <td>75.59</td>\n",
       "      <td>10.94</td>\n",
       "      <td>12.50</td>\n",
       "      <td>0.7239</td>\n",
       "      <td>0.004229</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>9</td>\n",
       "      <td>79.30</td>\n",
       "      <td>50.0</td>\n",
       "      <td>7.81</td>\n",
       "      <td>25.39</td>\n",
       "      <td>18.75</td>\n",
       "      <td>44.14</td>\n",
       "      <td>66.02</td>\n",
       "      <td>73.24</td>\n",
       "      <td>10.16</td>\n",
       "      <td>...</td>\n",
       "      <td>10.94</td>\n",
       "      <td>30.08</td>\n",
       "      <td>22.46</td>\n",
       "      <td>48.05</td>\n",
       "      <td>69.14</td>\n",
       "      <td>75.78</td>\n",
       "      <td>10.94</td>\n",
       "      <td>12.89</td>\n",
       "      <td>0.7346</td>\n",
       "      <td>0.004144</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>10</td>\n",
       "      <td>78.91</td>\n",
       "      <td>0.0</td>\n",
       "      <td>7.81</td>\n",
       "      <td>25.78</td>\n",
       "      <td>18.95</td>\n",
       "      <td>44.34</td>\n",
       "      <td>66.02</td>\n",
       "      <td>73.24</td>\n",
       "      <td>10.94</td>\n",
       "      <td>...</td>\n",
       "      <td>10.94</td>\n",
       "      <td>30.47</td>\n",
       "      <td>22.46</td>\n",
       "      <td>48.24</td>\n",
       "      <td>69.14</td>\n",
       "      <td>75.59</td>\n",
       "      <td>11.72</td>\n",
       "      <td>13.28</td>\n",
       "      <td>0.7445</td>\n",
       "      <td>0.004050</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>11</td>\n",
       "      <td>78.52</td>\n",
       "      <td>0.0</td>\n",
       "      <td>9.38</td>\n",
       "      <td>26.17</td>\n",
       "      <td>19.14</td>\n",
       "      <td>44.34</td>\n",
       "      <td>66.02</td>\n",
       "      <td>73.24</td>\n",
       "      <td>10.94</td>\n",
       "      <td>...</td>\n",
       "      <td>10.94</td>\n",
       "      <td>30.47</td>\n",
       "      <td>22.66</td>\n",
       "      <td>48.24</td>\n",
       "      <td>69.14</td>\n",
       "      <td>75.59</td>\n",
       "      <td>11.72</td>\n",
       "      <td>13.28</td>\n",
       "      <td>0.7551</td>\n",
       "      <td>0.003944</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>12</td>\n",
       "      <td>78.52</td>\n",
       "      <td>0.0</td>\n",
       "      <td>9.38</td>\n",
       "      <td>26.17</td>\n",
       "      <td>19.34</td>\n",
       "      <td>44.34</td>\n",
       "      <td>66.21</td>\n",
       "      <td>73.24</td>\n",
       "      <td>11.72</td>\n",
       "      <td>...</td>\n",
       "      <td>12.50</td>\n",
       "      <td>30.47</td>\n",
       "      <td>22.66</td>\n",
       "      <td>48.24</td>\n",
       "      <td>69.14</td>\n",
       "      <td>75.59</td>\n",
       "      <td>11.72</td>\n",
       "      <td>13.67</td>\n",
       "      <td>0.7566</td>\n",
       "      <td>0.003890</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>13</td>\n",
       "      <td>78.12</td>\n",
       "      <td>0.0</td>\n",
       "      <td>9.38</td>\n",
       "      <td>26.56</td>\n",
       "      <td>19.34</td>\n",
       "      <td>44.53</td>\n",
       "      <td>66.02</td>\n",
       "      <td>73.24</td>\n",
       "      <td>11.72</td>\n",
       "      <td>...</td>\n",
       "      <td>12.50</td>\n",
       "      <td>30.86</td>\n",
       "      <td>22.66</td>\n",
       "      <td>48.24</td>\n",
       "      <td>69.14</td>\n",
       "      <td>75.59</td>\n",
       "      <td>12.50</td>\n",
       "      <td>13.67</td>\n",
       "      <td>0.7668</td>\n",
       "      <td>0.003797</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>14</td>\n",
       "      <td>77.73</td>\n",
       "      <td>0.0</td>\n",
       "      <td>9.38</td>\n",
       "      <td>26.95</td>\n",
       "      <td>19.53</td>\n",
       "      <td>44.53</td>\n",
       "      <td>66.21</td>\n",
       "      <td>73.24</td>\n",
       "      <td>12.50</td>\n",
       "      <td>...</td>\n",
       "      <td>12.50</td>\n",
       "      <td>30.86</td>\n",
       "      <td>22.85</td>\n",
       "      <td>48.24</td>\n",
       "      <td>68.95</td>\n",
       "      <td>75.39</td>\n",
       "      <td>12.50</td>\n",
       "      <td>14.06</td>\n",
       "      <td>0.7662</td>\n",
       "      <td>0.003753</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>15</td>\n",
       "      <td>77.73</td>\n",
       "      <td>0.0</td>\n",
       "      <td>10.94</td>\n",
       "      <td>26.95</td>\n",
       "      <td>19.53</td>\n",
       "      <td>44.53</td>\n",
       "      <td>66.21</td>\n",
       "      <td>73.24</td>\n",
       "      <td>12.50</td>\n",
       "      <td>...</td>\n",
       "      <td>14.06</td>\n",
       "      <td>31.25</td>\n",
       "      <td>22.85</td>\n",
       "      <td>48.24</td>\n",
       "      <td>68.95</td>\n",
       "      <td>75.39</td>\n",
       "      <td>12.50</td>\n",
       "      <td>14.06</td>\n",
       "      <td>0.7757</td>\n",
       "      <td>0.003642</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>16</td>\n",
       "      <td>77.34</td>\n",
       "      <td>0.0</td>\n",
       "      <td>10.94</td>\n",
       "      <td>27.34</td>\n",
       "      <td>19.73</td>\n",
       "      <td>44.53</td>\n",
       "      <td>66.21</td>\n",
       "      <td>73.24</td>\n",
       "      <td>12.50</td>\n",
       "      <td>...</td>\n",
       "      <td>14.06</td>\n",
       "      <td>31.25</td>\n",
       "      <td>23.05</td>\n",
       "      <td>48.24</td>\n",
       "      <td>68.95</td>\n",
       "      <td>75.39</td>\n",
       "      <td>13.28</td>\n",
       "      <td>14.45</td>\n",
       "      <td>0.7787</td>\n",
       "      <td>0.003594</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>17</td>\n",
       "      <td>76.95</td>\n",
       "      <td>0.0</td>\n",
       "      <td>10.94</td>\n",
       "      <td>27.34</td>\n",
       "      <td>19.73</td>\n",
       "      <td>44.53</td>\n",
       "      <td>66.21</td>\n",
       "      <td>73.24</td>\n",
       "      <td>13.28</td>\n",
       "      <td>...</td>\n",
       "      <td>14.06</td>\n",
       "      <td>31.25</td>\n",
       "      <td>23.05</td>\n",
       "      <td>48.24</td>\n",
       "      <td>68.95</td>\n",
       "      <td>75.39</td>\n",
       "      <td>13.28</td>\n",
       "      <td>14.45</td>\n",
       "      <td>0.7799</td>\n",
       "      <td>0.003565</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>18</td>\n",
       "      <td>76.95</td>\n",
       "      <td>0.0</td>\n",
       "      <td>10.94</td>\n",
       "      <td>27.73</td>\n",
       "      <td>19.92</td>\n",
       "      <td>44.53</td>\n",
       "      <td>66.21</td>\n",
       "      <td>73.24</td>\n",
       "      <td>13.28</td>\n",
       "      <td>...</td>\n",
       "      <td>14.06</td>\n",
       "      <td>31.64</td>\n",
       "      <td>23.05</td>\n",
       "      <td>48.24</td>\n",
       "      <td>68.95</td>\n",
       "      <td>75.39</td>\n",
       "      <td>13.28</td>\n",
       "      <td>14.84</td>\n",
       "      <td>0.7798</td>\n",
       "      <td>0.003537</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>19</td>\n",
       "      <td>76.56</td>\n",
       "      <td>0.0</td>\n",
       "      <td>12.50</td>\n",
       "      <td>27.73</td>\n",
       "      <td>19.92</td>\n",
       "      <td>44.73</td>\n",
       "      <td>66.21</td>\n",
       "      <td>73.24</td>\n",
       "      <td>13.28</td>\n",
       "      <td>...</td>\n",
       "      <td>15.62</td>\n",
       "      <td>31.64</td>\n",
       "      <td>23.24</td>\n",
       "      <td>48.24</td>\n",
       "      <td>68.95</td>\n",
       "      <td>75.39</td>\n",
       "      <td>13.28</td>\n",
       "      <td>14.84</td>\n",
       "      <td>0.7941</td>\n",
       "      <td>0.003442</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>20 rows × 25 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    Unnamed: 0  eval-saturation_classifier-2  eval-saturation_classifier-6  \\\n",
       "0            0                         76.56                          50.0   \n",
       "1            1                         78.91                          50.0   \n",
       "2            2                         79.69                          50.0   \n",
       "3            3                         79.69                          50.0   \n",
       "4            4                         79.69                          50.0   \n",
       "5            5                         79.69                          50.0   \n",
       "6            6                         79.69                          50.0   \n",
       "7            7                         79.30                          50.0   \n",
       "8            8                         79.30                          50.0   \n",
       "9            9                         79.30                          50.0   \n",
       "10          10                         78.91                           0.0   \n",
       "11          11                         78.52                           0.0   \n",
       "12          12                         78.52                           0.0   \n",
       "13          13                         78.12                           0.0   \n",
       "14          14                         77.73                           0.0   \n",
       "15          15                         77.73                           0.0   \n",
       "16          16                         77.34                           0.0   \n",
       "17          17                         76.95                           0.0   \n",
       "18          18                         76.95                           0.0   \n",
       "19          19                         76.56                           0.0   \n",
       "\n",
       "    eval-saturation_features-0  eval-saturation_features-11  \\\n",
       "0                         3.12                        16.80   \n",
       "1                         4.69                        19.14   \n",
       "2                         4.69                        20.31   \n",
       "3                         6.25                        21.48   \n",
       "4                         6.25                        22.27   \n",
       "5                         6.25                        23.05   \n",
       "6                         6.25                        23.83   \n",
       "7                         6.25                        24.22   \n",
       "8                         7.81                        25.00   \n",
       "9                         7.81                        25.39   \n",
       "10                        7.81                        25.78   \n",
       "11                        9.38                        26.17   \n",
       "12                        9.38                        26.17   \n",
       "13                        9.38                        26.56   \n",
       "14                        9.38                        26.95   \n",
       "15                       10.94                        26.95   \n",
       "16                       10.94                        27.34   \n",
       "17                       10.94                        27.34   \n",
       "18                       10.94                        27.73   \n",
       "19                       12.50                        27.73   \n",
       "\n",
       "    eval-saturation_features-15  eval-saturation_features-18  \\\n",
       "0                         13.87                        35.35   \n",
       "1                         16.02                        40.43   \n",
       "2                         16.80                        41.60   \n",
       "3                         17.38                        42.38   \n",
       "4                         17.58                        42.77   \n",
       "5                         17.97                        43.36   \n",
       "6                         18.16                        43.55   \n",
       "7                         18.36                        43.75   \n",
       "8                         18.55                        43.95   \n",
       "9                         18.75                        44.14   \n",
       "10                        18.95                        44.34   \n",
       "11                        19.14                        44.34   \n",
       "12                        19.34                        44.34   \n",
       "13                        19.34                        44.53   \n",
       "14                        19.53                        44.53   \n",
       "15                        19.53                        44.53   \n",
       "16                        19.73                        44.53   \n",
       "17                        19.73                        44.53   \n",
       "18                        19.92                        44.53   \n",
       "19                        19.92                        44.73   \n",
       "\n",
       "    eval-saturation_features-22  eval-saturation_features-25  \\\n",
       "0                         52.93                        61.33   \n",
       "1                         60.94                        68.95   \n",
       "2                         63.09                        71.09   \n",
       "3                         64.06                        71.88   \n",
       "4                         64.84                        72.27   \n",
       "5                         65.23                        72.66   \n",
       "6                         65.43                        72.85   \n",
       "7                         65.62                        73.05   \n",
       "8                         65.82                        73.24   \n",
       "9                         66.02                        73.24   \n",
       "10                        66.02                        73.24   \n",
       "11                        66.02                        73.24   \n",
       "12                        66.21                        73.24   \n",
       "13                        66.02                        73.24   \n",
       "14                        66.21                        73.24   \n",
       "15                        66.21                        73.24   \n",
       "16                        66.21                        73.24   \n",
       "17                        66.21                        73.24   \n",
       "18                        66.21                        73.24   \n",
       "19                        66.21                        73.24   \n",
       "\n",
       "    eval-saturation_features-4      ...        train-saturation_features-0  \\\n",
       "0                         3.91      ...                               3.12   \n",
       "1                         4.69      ...                               4.69   \n",
       "2                         5.47      ...                               4.69   \n",
       "3                         7.03      ...                               6.25   \n",
       "4                         7.81      ...                               7.81   \n",
       "5                         8.59      ...                               7.81   \n",
       "6                         8.59      ...                               9.38   \n",
       "7                         9.38      ...                               9.38   \n",
       "8                        10.16      ...                               9.38   \n",
       "9                        10.16      ...                              10.94   \n",
       "10                       10.94      ...                              10.94   \n",
       "11                       10.94      ...                              10.94   \n",
       "12                       11.72      ...                              12.50   \n",
       "13                       11.72      ...                              12.50   \n",
       "14                       12.50      ...                              12.50   \n",
       "15                       12.50      ...                              14.06   \n",
       "16                       12.50      ...                              14.06   \n",
       "17                       13.28      ...                              14.06   \n",
       "18                       13.28      ...                              14.06   \n",
       "19                       13.28      ...                              15.62   \n",
       "\n",
       "    train-saturation_features-11  train-saturation_features-15  \\\n",
       "0                          22.66                         17.97   \n",
       "1                          25.00                         19.14   \n",
       "2                          26.17                         20.12   \n",
       "3                          27.34                         20.90   \n",
       "4                          28.12                         21.29   \n",
       "5                          28.91                         21.68   \n",
       "6                          29.30                         21.88   \n",
       "7                          29.30                         22.07   \n",
       "8                          29.69                         22.27   \n",
       "9                          30.08                         22.46   \n",
       "10                         30.47                         22.46   \n",
       "11                         30.47                         22.66   \n",
       "12                         30.47                         22.66   \n",
       "13                         30.86                         22.66   \n",
       "14                         30.86                         22.85   \n",
       "15                         31.25                         22.85   \n",
       "16                         31.25                         23.05   \n",
       "17                         31.25                         23.05   \n",
       "18                         31.64                         23.05   \n",
       "19                         31.64                         23.24   \n",
       "\n",
       "    train-saturation_features-18  train-saturation_features-22  \\\n",
       "0                          44.14                         66.02   \n",
       "1                          45.70                         67.77   \n",
       "2                          46.68                         68.36   \n",
       "3                          47.07                         68.75   \n",
       "4                          47.27                         68.75   \n",
       "5                          47.66                         68.95   \n",
       "6                          47.66                         68.95   \n",
       "7                          47.85                         68.95   \n",
       "8                          48.05                         69.14   \n",
       "9                          48.05                         69.14   \n",
       "10                         48.24                         69.14   \n",
       "11                         48.24                         69.14   \n",
       "12                         48.24                         69.14   \n",
       "13                         48.24                         69.14   \n",
       "14                         48.24                         68.95   \n",
       "15                         48.24                         68.95   \n",
       "16                         48.24                         68.95   \n",
       "17                         48.24                         68.95   \n",
       "18                         48.24                         68.95   \n",
       "19                         48.24                         68.95   \n",
       "\n",
       "    train-saturation_features-25  train-saturation_features-4  \\\n",
       "0                          73.44                         5.47   \n",
       "1                          75.00                         6.25   \n",
       "2                          75.39                         7.03   \n",
       "3                          75.59                         8.59   \n",
       "4                          75.78                         8.59   \n",
       "5                          75.78                         9.38   \n",
       "6                          75.78                        10.16   \n",
       "7                          75.78                        10.16   \n",
       "8                          75.59                        10.94   \n",
       "9                          75.78                        10.94   \n",
       "10                         75.59                        11.72   \n",
       "11                         75.59                        11.72   \n",
       "12                         75.59                        11.72   \n",
       "13                         75.59                        12.50   \n",
       "14                         75.39                        12.50   \n",
       "15                         75.39                        12.50   \n",
       "16                         75.39                        13.28   \n",
       "17                         75.39                        13.28   \n",
       "18                         75.39                        13.28   \n",
       "19                         75.39                        13.28   \n",
       "\n",
       "    train-saturation_features-8  training_accuracy  training_loss  \n",
       "0                          7.03             0.5575       0.005392  \n",
       "1                          8.20             0.6132       0.005160  \n",
       "2                          9.38             0.6481       0.004980  \n",
       "3                         10.16             0.6718       0.004738  \n",
       "4                         10.94             0.6832       0.004636  \n",
       "5                         11.33             0.7002       0.004503  \n",
       "6                         12.11             0.7104       0.004389  \n",
       "7                         12.11             0.7269       0.004234  \n",
       "8                         12.50             0.7239       0.004229  \n",
       "9                         12.89             0.7346       0.004144  \n",
       "10                        13.28             0.7445       0.004050  \n",
       "11                        13.28             0.7551       0.003944  \n",
       "12                        13.67             0.7566       0.003890  \n",
       "13                        13.67             0.7668       0.003797  \n",
       "14                        14.06             0.7662       0.003753  \n",
       "15                        14.06             0.7757       0.003642  \n",
       "16                        14.45             0.7787       0.003594  \n",
       "17                        14.45             0.7799       0.003565  \n",
       "18                        14.84             0.7798       0.003537  \n",
       "19                        14.84             0.7941       0.003442  \n",
       "\n",
       "[20 rows x 25 columns]"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "shallow = pd.read_csv(r'.\\logs\\VGG11\\CatVsDog\\VGG11_bs128_e20_id0.csv', sep=';')\n",
    "shallow.head(20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "deep = pd.read_csv('cd_128_VGG11VGG19VGG19_XXXS_A1.csv', sep=';')\n",
    "deep.head(20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "medium = pd.read_csv('cd_64_VGG11VGG19VGG19_XXXS_A1.csv', sep=';')\n",
    "medium.head(20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "def extract_layer_saturation(df, excluded = 'classifier6', epoch=20):\n",
    "    cols = list(df.columns)\n",
    "    print(cols)\n",
    "    train_cols = [col for col in cols if 'train' in col and not excluded in col and not 'accuracy' in col and not 'loss' in col]\n",
    "    epoch_df = df[df.index.values == epoch]\n",
    "    epoch_df = epoch_df[train_cols]\n",
    "    return epoch_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['Unnamed: 0', 'eval-saturation_classifier-2', 'eval-saturation_classifier-6', 'eval-saturation_features-0', 'eval-saturation_features-11', 'eval-saturation_features-15', 'eval-saturation_features-18', 'eval-saturation_features-22', 'eval-saturation_features-25', 'eval-saturation_features-4', 'eval-saturation_features-8', 'test_accuracy', 'test_loss', 'train-saturation_classifier-2', 'train-saturation_classifier-6', 'train-saturation_features-0', 'train-saturation_features-11', 'train-saturation_features-15', 'train-saturation_features-18', 'train-saturation_features-22', 'train-saturation_features-25', 'train-saturation_features-4', 'train-saturation_features-8', 'training_accuracy', 'training_loss']\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>train-saturation_classifier-2</th>\n",
       "      <th>train-saturation_classifier-6</th>\n",
       "      <th>train-saturation_features-0</th>\n",
       "      <th>train-saturation_features-11</th>\n",
       "      <th>train-saturation_features-15</th>\n",
       "      <th>train-saturation_features-18</th>\n",
       "      <th>train-saturation_features-22</th>\n",
       "      <th>train-saturation_features-25</th>\n",
       "      <th>train-saturation_features-4</th>\n",
       "      <th>train-saturation_features-8</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "Empty DataFrame\n",
       "Columns: [train-saturation_classifier-2, train-saturation_classifier-6, train-saturation_features-0, train-saturation_features-11, train-saturation_features-15, train-saturation_features-18, train-saturation_features-22, train-saturation_features-25, train-saturation_features-4, train-saturation_features-8]\n",
       "Index: []"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "shallow_sats = extract_layer_saturation(shallow)\n",
    "shallow_sats.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'deep' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-15-f3106342efe8>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[1;32m----> 1\u001b[1;33m \u001b[0mdeep_sats\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mextract_layer_saturation\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mdeep\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m      2\u001b[0m \u001b[0mdeep_sats\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mhead\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mNameError\u001b[0m: name 'deep' is not defined"
     ]
    }
   ],
   "source": [
    "deep_sats = extract_layer_saturation(deep)\n",
    "deep_sats.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'medium' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-8-ee876a132afd>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[1;32m----> 1\u001b[1;33m \u001b[0mmedium_sats\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mextract_layer_saturation\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mmedium\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m      2\u001b[0m \u001b[0mmedium_sats\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mhead\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mNameError\u001b[0m: name 'medium' is not defined"
     ]
    }
   ],
   "source": [
    "medium_sats = extract_layer_saturation(medium)\n",
    "medium_sats.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'deep_sats' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-9-1ebfa71a6d30>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[1;32m----> 1\u001b[1;33m \u001b[0mdeep_sats\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mvalues\u001b[0m\u001b[1;33m[\u001b[0m\u001b[1;36m0\u001b[0m\u001b[1;33m]\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m: name 'deep_sats' is not defined"
     ]
    }
   ],
   "source": [
    "deep_sats.values[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_saturation_level(df, title=''):\n",
    "    cols = list(df.columns)\n",
    "    col_names = ['Layer {}'.format(i+1) for i in range(len(df.columns))]\n",
    "    plt.figure(figsize=(7,5))\n",
    "    plt.grid()\n",
    "    plt.bar(list(range(len(cols))), df.values[0])\n",
    "    plt.xticks(list(range(len(cols))), col_names, rotation='vertical', fontsize=16)\n",
    "    plt.ylim((0,100))\n",
    "    plt.yticks(fontsize=16)\n",
    "    plt.xlabel('Layers', fontsize=16)\n",
    "    plt.title(title, fontsize=16)\n",
    "    plt.ylabel('Saturation in %', rotation='vertical', fontsize=16)\n",
    "    plt.show()\n",
    "    \n",
    "def plot_saturation_level_ax(df, ax):\n",
    "    cols = list(df.columns)\n",
    "    col_names = [i+1 for i in range(len(df.columns))]\n",
    "    ax.grid()\n",
    "    ax.bar(list(range(len(cols))), df.values[0])\n",
    "    ax.set_xticks([])\n",
    "    ax.set_ylim((0,100))\n",
    "    #plt.yticks(fontsize=16)\n",
    "    #plt.xlabel('Layers', fontsize=16)\n",
    "    #plt.title(title, fontsize=16)\n",
    "    #plt.ylabel('Saturation in %', rotation='vertical', fontsize=16)\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'shallow_sats' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-11-d67fcaf66ed9>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[1;32m----> 1\u001b[1;33m \u001b[0mplot_saturation_level\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mshallow_sats\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;34m'60.57% Test Accuracy'\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m: name 'shallow_sats' is not defined"
     ]
    }
   ],
   "source": [
    "plot_saturation_level(shallow_sats, '60.57% Test Accuracy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'deep_sats' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-12-871fd7a7e4d1>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[1;32m----> 1\u001b[1;33m \u001b[0mplot_saturation_level\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mdeep_sats\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;34m'87.52% Test Accuracy'\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m: name 'deep_sats' is not defined"
     ]
    }
   ],
   "source": [
    "plot_saturation_level(deep_sats, '87.52% Test Accuracy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'medium_sats' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-13-e1839e9e4fff>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[1;32m----> 1\u001b[1;33m \u001b[0mplot_saturation_level\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mmedium_sats\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;34m'88.74% Test Accuracy'\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m: name 'medium_sats' is not defined"
     ]
    }
   ],
   "source": [
    "plot_saturation_level(medium_sats, '88.74% Test Accuracy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Unnamed: 0</th>\n",
       "      <th>eval_features0</th>\n",
       "      <th>eval_features3</th>\n",
       "      <th>eval_features7</th>\n",
       "      <th>eval_features10</th>\n",
       "      <th>eval_features14</th>\n",
       "      <th>eval_features17</th>\n",
       "      <th>eval_features21</th>\n",
       "      <th>eval_features24</th>\n",
       "      <th>eval_features28</th>\n",
       "      <th>...</th>\n",
       "      <th>train_features28</th>\n",
       "      <th>train_features31</th>\n",
       "      <th>train_classifier2</th>\n",
       "      <th>train_classifier6</th>\n",
       "      <th>train_accuracy</th>\n",
       "      <th>test_accuracy</th>\n",
       "      <th>train_loss</th>\n",
       "      <th>test_loss</th>\n",
       "      <th>epoch</th>\n",
       "      <th>time_per_step</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>1.56</td>\n",
       "      <td>1.56</td>\n",
       "      <td>0.78</td>\n",
       "      <td>0.78</td>\n",
       "      <td>0.39</td>\n",
       "      <td>0.39</td>\n",
       "      <td>0.20</td>\n",
       "      <td>0.20</td>\n",
       "      <td>0.20</td>\n",
       "      <td>...</td>\n",
       "      <td>1.95</td>\n",
       "      <td>1.76</td>\n",
       "      <td>3.12</td>\n",
       "      <td>50.0</td>\n",
       "      <td>0.51446</td>\n",
       "      <td>0.5768</td>\n",
       "      <td>0.010431</td>\n",
       "      <td>0.009561</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000833</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>3.12</td>\n",
       "      <td>4.69</td>\n",
       "      <td>3.12</td>\n",
       "      <td>3.12</td>\n",
       "      <td>1.56</td>\n",
       "      <td>1.56</td>\n",
       "      <td>0.78</td>\n",
       "      <td>0.78</td>\n",
       "      <td>0.78</td>\n",
       "      <td>...</td>\n",
       "      <td>3.71</td>\n",
       "      <td>3.12</td>\n",
       "      <td>5.86</td>\n",
       "      <td>60.0</td>\n",
       "      <td>0.68556</td>\n",
       "      <td>0.7027</td>\n",
       "      <td>0.006931</td>\n",
       "      <td>0.006698</td>\n",
       "      <td>1</td>\n",
       "      <td>0.000937</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>3.12</td>\n",
       "      <td>6.25</td>\n",
       "      <td>4.69</td>\n",
       "      <td>4.69</td>\n",
       "      <td>2.34</td>\n",
       "      <td>2.34</td>\n",
       "      <td>1.17</td>\n",
       "      <td>0.98</td>\n",
       "      <td>0.98</td>\n",
       "      <td>...</td>\n",
       "      <td>4.49</td>\n",
       "      <td>4.10</td>\n",
       "      <td>7.03</td>\n",
       "      <td>70.0</td>\n",
       "      <td>0.75382</td>\n",
       "      <td>0.7651</td>\n",
       "      <td>0.005552</td>\n",
       "      <td>0.005424</td>\n",
       "      <td>2</td>\n",
       "      <td>0.001055</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>6.25</td>\n",
       "      <td>10.94</td>\n",
       "      <td>6.25</td>\n",
       "      <td>6.25</td>\n",
       "      <td>3.12</td>\n",
       "      <td>3.12</td>\n",
       "      <td>1.56</td>\n",
       "      <td>1.56</td>\n",
       "      <td>1.37</td>\n",
       "      <td>...</td>\n",
       "      <td>5.08</td>\n",
       "      <td>4.69</td>\n",
       "      <td>8.59</td>\n",
       "      <td>70.0</td>\n",
       "      <td>0.79044</td>\n",
       "      <td>0.7809</td>\n",
       "      <td>0.004745</td>\n",
       "      <td>0.005298</td>\n",
       "      <td>3</td>\n",
       "      <td>0.001152</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>6.25</td>\n",
       "      <td>14.06</td>\n",
       "      <td>8.59</td>\n",
       "      <td>8.59</td>\n",
       "      <td>4.69</td>\n",
       "      <td>4.69</td>\n",
       "      <td>2.34</td>\n",
       "      <td>2.34</td>\n",
       "      <td>1.95</td>\n",
       "      <td>...</td>\n",
       "      <td>5.47</td>\n",
       "      <td>5.08</td>\n",
       "      <td>8.98</td>\n",
       "      <td>70.0</td>\n",
       "      <td>0.81698</td>\n",
       "      <td>0.7989</td>\n",
       "      <td>0.004189</td>\n",
       "      <td>0.004644</td>\n",
       "      <td>4</td>\n",
       "      <td>0.001151</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>5</td>\n",
       "      <td>7.81</td>\n",
       "      <td>15.62</td>\n",
       "      <td>9.38</td>\n",
       "      <td>9.38</td>\n",
       "      <td>5.47</td>\n",
       "      <td>5.47</td>\n",
       "      <td>2.73</td>\n",
       "      <td>2.73</td>\n",
       "      <td>2.15</td>\n",
       "      <td>...</td>\n",
       "      <td>6.05</td>\n",
       "      <td>5.66</td>\n",
       "      <td>10.16</td>\n",
       "      <td>70.0</td>\n",
       "      <td>0.83842</td>\n",
       "      <td>0.8111</td>\n",
       "      <td>0.003738</td>\n",
       "      <td>0.004345</td>\n",
       "      <td>5</td>\n",
       "      <td>0.001223</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>6</td>\n",
       "      <td>9.38</td>\n",
       "      <td>15.62</td>\n",
       "      <td>10.94</td>\n",
       "      <td>10.94</td>\n",
       "      <td>5.86</td>\n",
       "      <td>6.25</td>\n",
       "      <td>3.12</td>\n",
       "      <td>2.93</td>\n",
       "      <td>2.34</td>\n",
       "      <td>...</td>\n",
       "      <td>6.45</td>\n",
       "      <td>5.86</td>\n",
       "      <td>10.94</td>\n",
       "      <td>70.0</td>\n",
       "      <td>0.85158</td>\n",
       "      <td>0.8355</td>\n",
       "      <td>0.003414</td>\n",
       "      <td>0.003851</td>\n",
       "      <td>6</td>\n",
       "      <td>0.001290</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>7</td>\n",
       "      <td>10.94</td>\n",
       "      <td>17.19</td>\n",
       "      <td>11.72</td>\n",
       "      <td>12.50</td>\n",
       "      <td>6.64</td>\n",
       "      <td>7.03</td>\n",
       "      <td>3.52</td>\n",
       "      <td>3.12</td>\n",
       "      <td>2.54</td>\n",
       "      <td>...</td>\n",
       "      <td>6.64</td>\n",
       "      <td>6.05</td>\n",
       "      <td>11.33</td>\n",
       "      <td>70.0</td>\n",
       "      <td>0.86240</td>\n",
       "      <td>0.8508</td>\n",
       "      <td>0.003145</td>\n",
       "      <td>0.003407</td>\n",
       "      <td>7</td>\n",
       "      <td>0.001462</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>8</td>\n",
       "      <td>12.50</td>\n",
       "      <td>18.75</td>\n",
       "      <td>12.50</td>\n",
       "      <td>13.28</td>\n",
       "      <td>7.03</td>\n",
       "      <td>7.81</td>\n",
       "      <td>3.71</td>\n",
       "      <td>3.52</td>\n",
       "      <td>2.54</td>\n",
       "      <td>...</td>\n",
       "      <td>7.23</td>\n",
       "      <td>6.64</td>\n",
       "      <td>12.11</td>\n",
       "      <td>70.0</td>\n",
       "      <td>0.87564</td>\n",
       "      <td>0.8627</td>\n",
       "      <td>0.002862</td>\n",
       "      <td>0.003252</td>\n",
       "      <td>8</td>\n",
       "      <td>0.001589</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>9</td>\n",
       "      <td>12.50</td>\n",
       "      <td>20.31</td>\n",
       "      <td>14.06</td>\n",
       "      <td>14.84</td>\n",
       "      <td>7.81</td>\n",
       "      <td>8.59</td>\n",
       "      <td>4.10</td>\n",
       "      <td>3.71</td>\n",
       "      <td>2.54</td>\n",
       "      <td>...</td>\n",
       "      <td>7.23</td>\n",
       "      <td>6.84</td>\n",
       "      <td>12.11</td>\n",
       "      <td>70.0</td>\n",
       "      <td>0.88294</td>\n",
       "      <td>0.8526</td>\n",
       "      <td>0.002663</td>\n",
       "      <td>0.003563</td>\n",
       "      <td>9</td>\n",
       "      <td>0.001704</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>10</td>\n",
       "      <td>15.62</td>\n",
       "      <td>21.88</td>\n",
       "      <td>15.62</td>\n",
       "      <td>16.41</td>\n",
       "      <td>9.38</td>\n",
       "      <td>10.16</td>\n",
       "      <td>4.88</td>\n",
       "      <td>4.49</td>\n",
       "      <td>2.93</td>\n",
       "      <td>...</td>\n",
       "      <td>7.03</td>\n",
       "      <td>7.03</td>\n",
       "      <td>12.50</td>\n",
       "      <td>70.0</td>\n",
       "      <td>0.89416</td>\n",
       "      <td>0.8614</td>\n",
       "      <td>0.002443</td>\n",
       "      <td>0.003274</td>\n",
       "      <td>10</td>\n",
       "      <td>0.001695</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>11</td>\n",
       "      <td>15.62</td>\n",
       "      <td>23.44</td>\n",
       "      <td>16.41</td>\n",
       "      <td>17.97</td>\n",
       "      <td>9.77</td>\n",
       "      <td>10.94</td>\n",
       "      <td>5.27</td>\n",
       "      <td>4.88</td>\n",
       "      <td>3.12</td>\n",
       "      <td>...</td>\n",
       "      <td>7.42</td>\n",
       "      <td>7.23</td>\n",
       "      <td>12.89</td>\n",
       "      <td>70.0</td>\n",
       "      <td>0.90134</td>\n",
       "      <td>0.8810</td>\n",
       "      <td>0.002270</td>\n",
       "      <td>0.002866</td>\n",
       "      <td>11</td>\n",
       "      <td>0.001915</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>12</td>\n",
       "      <td>17.19</td>\n",
       "      <td>26.56</td>\n",
       "      <td>17.97</td>\n",
       "      <td>19.53</td>\n",
       "      <td>10.94</td>\n",
       "      <td>11.72</td>\n",
       "      <td>5.66</td>\n",
       "      <td>5.27</td>\n",
       "      <td>3.32</td>\n",
       "      <td>...</td>\n",
       "      <td>7.81</td>\n",
       "      <td>7.42</td>\n",
       "      <td>13.28</td>\n",
       "      <td>70.0</td>\n",
       "      <td>0.90806</td>\n",
       "      <td>0.8613</td>\n",
       "      <td>0.002103</td>\n",
       "      <td>0.003415</td>\n",
       "      <td>12</td>\n",
       "      <td>0.003310</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>13</td>\n",
       "      <td>17.19</td>\n",
       "      <td>26.56</td>\n",
       "      <td>18.75</td>\n",
       "      <td>19.53</td>\n",
       "      <td>11.33</td>\n",
       "      <td>12.11</td>\n",
       "      <td>5.86</td>\n",
       "      <td>5.47</td>\n",
       "      <td>3.32</td>\n",
       "      <td>...</td>\n",
       "      <td>8.01</td>\n",
       "      <td>7.62</td>\n",
       "      <td>13.67</td>\n",
       "      <td>70.0</td>\n",
       "      <td>0.91404</td>\n",
       "      <td>0.8831</td>\n",
       "      <td>0.001947</td>\n",
       "      <td>0.002803</td>\n",
       "      <td>13</td>\n",
       "      <td>0.004378</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>14</td>\n",
       "      <td>18.75</td>\n",
       "      <td>28.12</td>\n",
       "      <td>19.53</td>\n",
       "      <td>21.09</td>\n",
       "      <td>12.11</td>\n",
       "      <td>13.28</td>\n",
       "      <td>6.25</td>\n",
       "      <td>5.86</td>\n",
       "      <td>3.52</td>\n",
       "      <td>...</td>\n",
       "      <td>8.01</td>\n",
       "      <td>7.62</td>\n",
       "      <td>13.67</td>\n",
       "      <td>80.0</td>\n",
       "      <td>0.92048</td>\n",
       "      <td>0.8721</td>\n",
       "      <td>0.001809</td>\n",
       "      <td>0.003117</td>\n",
       "      <td>14</td>\n",
       "      <td>0.009161</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>15</td>\n",
       "      <td>18.75</td>\n",
       "      <td>28.12</td>\n",
       "      <td>20.31</td>\n",
       "      <td>21.88</td>\n",
       "      <td>12.50</td>\n",
       "      <td>13.67</td>\n",
       "      <td>6.45</td>\n",
       "      <td>5.86</td>\n",
       "      <td>3.52</td>\n",
       "      <td>...</td>\n",
       "      <td>8.40</td>\n",
       "      <td>8.40</td>\n",
       "      <td>13.67</td>\n",
       "      <td>80.0</td>\n",
       "      <td>0.92538</td>\n",
       "      <td>0.8771</td>\n",
       "      <td>0.001696</td>\n",
       "      <td>0.003078</td>\n",
       "      <td>15</td>\n",
       "      <td>0.008208</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>16</td>\n",
       "      <td>20.31</td>\n",
       "      <td>29.69</td>\n",
       "      <td>21.09</td>\n",
       "      <td>23.44</td>\n",
       "      <td>13.67</td>\n",
       "      <td>14.84</td>\n",
       "      <td>7.23</td>\n",
       "      <td>6.45</td>\n",
       "      <td>3.91</td>\n",
       "      <td>...</td>\n",
       "      <td>8.20</td>\n",
       "      <td>7.81</td>\n",
       "      <td>12.50</td>\n",
       "      <td>70.0</td>\n",
       "      <td>0.92950</td>\n",
       "      <td>0.8765</td>\n",
       "      <td>0.001602</td>\n",
       "      <td>0.003074</td>\n",
       "      <td>16</td>\n",
       "      <td>0.007203</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>17</td>\n",
       "      <td>20.31</td>\n",
       "      <td>31.25</td>\n",
       "      <td>21.88</td>\n",
       "      <td>25.00</td>\n",
       "      <td>14.06</td>\n",
       "      <td>15.62</td>\n",
       "      <td>7.62</td>\n",
       "      <td>6.84</td>\n",
       "      <td>3.91</td>\n",
       "      <td>...</td>\n",
       "      <td>8.59</td>\n",
       "      <td>8.40</td>\n",
       "      <td>12.11</td>\n",
       "      <td>70.0</td>\n",
       "      <td>0.93474</td>\n",
       "      <td>0.8830</td>\n",
       "      <td>0.001494</td>\n",
       "      <td>0.002812</td>\n",
       "      <td>17</td>\n",
       "      <td>0.005743</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>18</td>\n",
       "      <td>20.31</td>\n",
       "      <td>32.81</td>\n",
       "      <td>23.44</td>\n",
       "      <td>25.78</td>\n",
       "      <td>14.84</td>\n",
       "      <td>16.02</td>\n",
       "      <td>7.81</td>\n",
       "      <td>7.23</td>\n",
       "      <td>4.10</td>\n",
       "      <td>...</td>\n",
       "      <td>8.98</td>\n",
       "      <td>8.40</td>\n",
       "      <td>12.50</td>\n",
       "      <td>70.0</td>\n",
       "      <td>0.93888</td>\n",
       "      <td>0.8815</td>\n",
       "      <td>0.001390</td>\n",
       "      <td>0.002985</td>\n",
       "      <td>18</td>\n",
       "      <td>0.007388</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>19</td>\n",
       "      <td>20.31</td>\n",
       "      <td>32.81</td>\n",
       "      <td>24.22</td>\n",
       "      <td>25.78</td>\n",
       "      <td>15.62</td>\n",
       "      <td>16.80</td>\n",
       "      <td>8.20</td>\n",
       "      <td>7.42</td>\n",
       "      <td>4.30</td>\n",
       "      <td>...</td>\n",
       "      <td>8.98</td>\n",
       "      <td>8.59</td>\n",
       "      <td>11.72</td>\n",
       "      <td>80.0</td>\n",
       "      <td>0.94232</td>\n",
       "      <td>0.8914</td>\n",
       "      <td>0.001313</td>\n",
       "      <td>0.002831</td>\n",
       "      <td>19</td>\n",
       "      <td>0.008512</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>20 rows × 31 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    Unnamed: 0  eval_features0  eval_features3  eval_features7  \\\n",
       "0            0            1.56            1.56            0.78   \n",
       "1            1            3.12            4.69            3.12   \n",
       "2            2            3.12            6.25            4.69   \n",
       "3            3            6.25           10.94            6.25   \n",
       "4            4            6.25           14.06            8.59   \n",
       "5            5            7.81           15.62            9.38   \n",
       "6            6            9.38           15.62           10.94   \n",
       "7            7           10.94           17.19           11.72   \n",
       "8            8           12.50           18.75           12.50   \n",
       "9            9           12.50           20.31           14.06   \n",
       "10          10           15.62           21.88           15.62   \n",
       "11          11           15.62           23.44           16.41   \n",
       "12          12           17.19           26.56           17.97   \n",
       "13          13           17.19           26.56           18.75   \n",
       "14          14           18.75           28.12           19.53   \n",
       "15          15           18.75           28.12           20.31   \n",
       "16          16           20.31           29.69           21.09   \n",
       "17          17           20.31           31.25           21.88   \n",
       "18          18           20.31           32.81           23.44   \n",
       "19          19           20.31           32.81           24.22   \n",
       "\n",
       "    eval_features10  eval_features14  eval_features17  eval_features21  \\\n",
       "0              0.78             0.39             0.39             0.20   \n",
       "1              3.12             1.56             1.56             0.78   \n",
       "2              4.69             2.34             2.34             1.17   \n",
       "3              6.25             3.12             3.12             1.56   \n",
       "4              8.59             4.69             4.69             2.34   \n",
       "5              9.38             5.47             5.47             2.73   \n",
       "6             10.94             5.86             6.25             3.12   \n",
       "7             12.50             6.64             7.03             3.52   \n",
       "8             13.28             7.03             7.81             3.71   \n",
       "9             14.84             7.81             8.59             4.10   \n",
       "10            16.41             9.38            10.16             4.88   \n",
       "11            17.97             9.77            10.94             5.27   \n",
       "12            19.53            10.94            11.72             5.66   \n",
       "13            19.53            11.33            12.11             5.86   \n",
       "14            21.09            12.11            13.28             6.25   \n",
       "15            21.88            12.50            13.67             6.45   \n",
       "16            23.44            13.67            14.84             7.23   \n",
       "17            25.00            14.06            15.62             7.62   \n",
       "18            25.78            14.84            16.02             7.81   \n",
       "19            25.78            15.62            16.80             8.20   \n",
       "\n",
       "    eval_features24  eval_features28      ...        train_features28  \\\n",
       "0              0.20             0.20      ...                    1.95   \n",
       "1              0.78             0.78      ...                    3.71   \n",
       "2              0.98             0.98      ...                    4.49   \n",
       "3              1.56             1.37      ...                    5.08   \n",
       "4              2.34             1.95      ...                    5.47   \n",
       "5              2.73             2.15      ...                    6.05   \n",
       "6              2.93             2.34      ...                    6.45   \n",
       "7              3.12             2.54      ...                    6.64   \n",
       "8              3.52             2.54      ...                    7.23   \n",
       "9              3.71             2.54      ...                    7.23   \n",
       "10             4.49             2.93      ...                    7.03   \n",
       "11             4.88             3.12      ...                    7.42   \n",
       "12             5.27             3.32      ...                    7.81   \n",
       "13             5.47             3.32      ...                    8.01   \n",
       "14             5.86             3.52      ...                    8.01   \n",
       "15             5.86             3.52      ...                    8.40   \n",
       "16             6.45             3.91      ...                    8.20   \n",
       "17             6.84             3.91      ...                    8.59   \n",
       "18             7.23             4.10      ...                    8.98   \n",
       "19             7.42             4.30      ...                    8.98   \n",
       "\n",
       "    train_features31  train_classifier2  train_classifier6  train_accuracy  \\\n",
       "0               1.76               3.12               50.0         0.51446   \n",
       "1               3.12               5.86               60.0         0.68556   \n",
       "2               4.10               7.03               70.0         0.75382   \n",
       "3               4.69               8.59               70.0         0.79044   \n",
       "4               5.08               8.98               70.0         0.81698   \n",
       "5               5.66              10.16               70.0         0.83842   \n",
       "6               5.86              10.94               70.0         0.85158   \n",
       "7               6.05              11.33               70.0         0.86240   \n",
       "8               6.64              12.11               70.0         0.87564   \n",
       "9               6.84              12.11               70.0         0.88294   \n",
       "10              7.03              12.50               70.0         0.89416   \n",
       "11              7.23              12.89               70.0         0.90134   \n",
       "12              7.42              13.28               70.0         0.90806   \n",
       "13              7.62              13.67               70.0         0.91404   \n",
       "14              7.62              13.67               80.0         0.92048   \n",
       "15              8.40              13.67               80.0         0.92538   \n",
       "16              7.81              12.50               70.0         0.92950   \n",
       "17              8.40              12.11               70.0         0.93474   \n",
       "18              8.40              12.50               70.0         0.93888   \n",
       "19              8.59              11.72               80.0         0.94232   \n",
       "\n",
       "    test_accuracy  train_loss  test_loss  epoch  time_per_step  \n",
       "0          0.5768    0.010431   0.009561      0       0.000833  \n",
       "1          0.7027    0.006931   0.006698      1       0.000937  \n",
       "2          0.7651    0.005552   0.005424      2       0.001055  \n",
       "3          0.7809    0.004745   0.005298      3       0.001152  \n",
       "4          0.7989    0.004189   0.004644      4       0.001151  \n",
       "5          0.8111    0.003738   0.004345      5       0.001223  \n",
       "6          0.8355    0.003414   0.003851      6       0.001290  \n",
       "7          0.8508    0.003145   0.003407      7       0.001462  \n",
       "8          0.8627    0.002862   0.003252      8       0.001589  \n",
       "9          0.8526    0.002663   0.003563      9       0.001704  \n",
       "10         0.8614    0.002443   0.003274     10       0.001695  \n",
       "11         0.8810    0.002270   0.002866     11       0.001915  \n",
       "12         0.8613    0.002103   0.003415     12       0.003310  \n",
       "13         0.8831    0.001947   0.002803     13       0.004378  \n",
       "14         0.8721    0.001809   0.003117     14       0.009161  \n",
       "15         0.8771    0.001696   0.003078     15       0.008208  \n",
       "16         0.8765    0.001602   0.003074     16       0.007203  \n",
       "17         0.8830    0.001494   0.002812     17       0.005743  \n",
       "18         0.8815    0.001390   0.002985     18       0.007388  \n",
       "19         0.8914    0.001313   0.002831     19       0.008512  \n",
       "\n",
       "[20 rows x 31 columns]"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "vgg13 = pd.read_csv('10_VGG13_A0.csv', sep=';')\n",
    "vgg13.head(20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 504x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_saturation_level(extract_layer_saturation(vgg13), '89.14% Test Accuracy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "cd_vgg13 = pd.read_csv('catdog_VGG13_A0.csv', sep=';')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Unnamed: 0</th>\n",
       "      <th>train_features0</th>\n",
       "      <th>train_features3</th>\n",
       "      <th>train_features7</th>\n",
       "      <th>train_features10</th>\n",
       "      <th>train_features14</th>\n",
       "      <th>train_features17</th>\n",
       "      <th>train_features21</th>\n",
       "      <th>train_features24</th>\n",
       "      <th>train_features28</th>\n",
       "      <th>train_features31</th>\n",
       "      <th>train_classifier2</th>\n",
       "      <th>train_classifier6</th>\n",
       "      <th>train_accuracy</th>\n",
       "      <th>test_accuracy</th>\n",
       "      <th>train_loss</th>\n",
       "      <th>test_loss</th>\n",
       "      <th>epoch</th>\n",
       "      <th>time_per_step</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>1.56</td>\n",
       "      <td>1.56</td>\n",
       "      <td>0.78</td>\n",
       "      <td>0.78</td>\n",
       "      <td>0.39</td>\n",
       "      <td>0.39</td>\n",
       "      <td>0.20</td>\n",
       "      <td>0.20</td>\n",
       "      <td>0.20</td>\n",
       "      <td>0.20</td>\n",
       "      <td>0.39</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.6225</td>\n",
       "      <td>0.6995</td>\n",
       "      <td>0.005065</td>\n",
       "      <td>0.004839</td>\n",
       "      <td>0</td>\n",
       "      <td>0.001485</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>3.12</td>\n",
       "      <td>6.25</td>\n",
       "      <td>3.12</td>\n",
       "      <td>3.12</td>\n",
       "      <td>1.56</td>\n",
       "      <td>1.17</td>\n",
       "      <td>0.59</td>\n",
       "      <td>0.59</td>\n",
       "      <td>0.59</td>\n",
       "      <td>0.39</td>\n",
       "      <td>0.78</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.7218</td>\n",
       "      <td>0.7340</td>\n",
       "      <td>0.004275</td>\n",
       "      <td>0.004171</td>\n",
       "      <td>1</td>\n",
       "      <td>0.001489</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>3.12</td>\n",
       "      <td>7.81</td>\n",
       "      <td>3.91</td>\n",
       "      <td>3.91</td>\n",
       "      <td>1.95</td>\n",
       "      <td>1.95</td>\n",
       "      <td>0.98</td>\n",
       "      <td>0.78</td>\n",
       "      <td>0.78</td>\n",
       "      <td>0.59</td>\n",
       "      <td>1.17</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.7654</td>\n",
       "      <td>0.7625</td>\n",
       "      <td>0.003784</td>\n",
       "      <td>0.003929</td>\n",
       "      <td>2</td>\n",
       "      <td>0.001514</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>6.25</td>\n",
       "      <td>12.50</td>\n",
       "      <td>7.03</td>\n",
       "      <td>7.03</td>\n",
       "      <td>3.52</td>\n",
       "      <td>3.52</td>\n",
       "      <td>1.37</td>\n",
       "      <td>1.17</td>\n",
       "      <td>1.17</td>\n",
       "      <td>0.59</td>\n",
       "      <td>1.17</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.7871</td>\n",
       "      <td>0.7995</td>\n",
       "      <td>0.003504</td>\n",
       "      <td>0.003280</td>\n",
       "      <td>3</td>\n",
       "      <td>0.001593</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>7.81</td>\n",
       "      <td>15.62</td>\n",
       "      <td>7.81</td>\n",
       "      <td>8.59</td>\n",
       "      <td>4.30</td>\n",
       "      <td>4.30</td>\n",
       "      <td>1.76</td>\n",
       "      <td>1.56</td>\n",
       "      <td>1.56</td>\n",
       "      <td>0.78</td>\n",
       "      <td>1.56</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.7976</td>\n",
       "      <td>0.8040</td>\n",
       "      <td>0.003336</td>\n",
       "      <td>0.003363</td>\n",
       "      <td>4</td>\n",
       "      <td>0.001533</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>5</td>\n",
       "      <td>10.94</td>\n",
       "      <td>18.75</td>\n",
       "      <td>10.16</td>\n",
       "      <td>10.94</td>\n",
       "      <td>5.47</td>\n",
       "      <td>5.86</td>\n",
       "      <td>2.34</td>\n",
       "      <td>2.15</td>\n",
       "      <td>1.95</td>\n",
       "      <td>1.17</td>\n",
       "      <td>1.56</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.8258</td>\n",
       "      <td>0.8160</td>\n",
       "      <td>0.003005</td>\n",
       "      <td>0.003159</td>\n",
       "      <td>5</td>\n",
       "      <td>0.001652</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>6</td>\n",
       "      <td>10.94</td>\n",
       "      <td>20.31</td>\n",
       "      <td>11.72</td>\n",
       "      <td>11.72</td>\n",
       "      <td>6.25</td>\n",
       "      <td>6.64</td>\n",
       "      <td>2.54</td>\n",
       "      <td>2.34</td>\n",
       "      <td>1.95</td>\n",
       "      <td>1.37</td>\n",
       "      <td>1.56</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.8341</td>\n",
       "      <td>0.8480</td>\n",
       "      <td>0.002919</td>\n",
       "      <td>0.002829</td>\n",
       "      <td>6</td>\n",
       "      <td>0.001558</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>7</td>\n",
       "      <td>12.50</td>\n",
       "      <td>21.88</td>\n",
       "      <td>12.50</td>\n",
       "      <td>13.28</td>\n",
       "      <td>6.64</td>\n",
       "      <td>7.03</td>\n",
       "      <td>2.93</td>\n",
       "      <td>2.54</td>\n",
       "      <td>2.15</td>\n",
       "      <td>1.37</td>\n",
       "      <td>1.95</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.8427</td>\n",
       "      <td>0.8010</td>\n",
       "      <td>0.002770</td>\n",
       "      <td>0.003457</td>\n",
       "      <td>7</td>\n",
       "      <td>0.001567</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>8</td>\n",
       "      <td>14.06</td>\n",
       "      <td>23.44</td>\n",
       "      <td>13.28</td>\n",
       "      <td>14.06</td>\n",
       "      <td>7.42</td>\n",
       "      <td>7.81</td>\n",
       "      <td>3.12</td>\n",
       "      <td>2.73</td>\n",
       "      <td>2.34</td>\n",
       "      <td>1.56</td>\n",
       "      <td>1.95</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.8532</td>\n",
       "      <td>0.8195</td>\n",
       "      <td>0.002597</td>\n",
       "      <td>0.003333</td>\n",
       "      <td>8</td>\n",
       "      <td>0.001579</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>9</td>\n",
       "      <td>15.62</td>\n",
       "      <td>25.00</td>\n",
       "      <td>14.84</td>\n",
       "      <td>14.84</td>\n",
       "      <td>7.81</td>\n",
       "      <td>8.20</td>\n",
       "      <td>3.32</td>\n",
       "      <td>2.93</td>\n",
       "      <td>2.34</td>\n",
       "      <td>1.56</td>\n",
       "      <td>1.95</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.8591</td>\n",
       "      <td>0.8070</td>\n",
       "      <td>0.002520</td>\n",
       "      <td>0.003280</td>\n",
       "      <td>9</td>\n",
       "      <td>0.001593</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>10</td>\n",
       "      <td>17.19</td>\n",
       "      <td>28.12</td>\n",
       "      <td>16.41</td>\n",
       "      <td>17.19</td>\n",
       "      <td>8.98</td>\n",
       "      <td>9.38</td>\n",
       "      <td>3.71</td>\n",
       "      <td>3.32</td>\n",
       "      <td>2.54</td>\n",
       "      <td>1.76</td>\n",
       "      <td>1.95</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.8664</td>\n",
       "      <td>0.8525</td>\n",
       "      <td>0.002399</td>\n",
       "      <td>0.002721</td>\n",
       "      <td>10</td>\n",
       "      <td>0.001766</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>11</td>\n",
       "      <td>18.75</td>\n",
       "      <td>29.69</td>\n",
       "      <td>17.19</td>\n",
       "      <td>17.97</td>\n",
       "      <td>9.38</td>\n",
       "      <td>10.16</td>\n",
       "      <td>4.10</td>\n",
       "      <td>3.52</td>\n",
       "      <td>2.73</td>\n",
       "      <td>1.76</td>\n",
       "      <td>1.95</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.8726</td>\n",
       "      <td>0.8575</td>\n",
       "      <td>0.002312</td>\n",
       "      <td>0.002743</td>\n",
       "      <td>11</td>\n",
       "      <td>0.001614</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>12</td>\n",
       "      <td>20.31</td>\n",
       "      <td>29.69</td>\n",
       "      <td>17.97</td>\n",
       "      <td>18.75</td>\n",
       "      <td>10.16</td>\n",
       "      <td>10.55</td>\n",
       "      <td>4.30</td>\n",
       "      <td>3.71</td>\n",
       "      <td>2.73</td>\n",
       "      <td>1.95</td>\n",
       "      <td>1.95</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.8805</td>\n",
       "      <td>0.7730</td>\n",
       "      <td>0.002207</td>\n",
       "      <td>0.004624</td>\n",
       "      <td>12</td>\n",
       "      <td>0.001627</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>13</td>\n",
       "      <td>21.88</td>\n",
       "      <td>31.25</td>\n",
       "      <td>19.53</td>\n",
       "      <td>19.53</td>\n",
       "      <td>10.55</td>\n",
       "      <td>11.33</td>\n",
       "      <td>4.49</td>\n",
       "      <td>3.91</td>\n",
       "      <td>2.73</td>\n",
       "      <td>2.15</td>\n",
       "      <td>1.95</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.8690</td>\n",
       "      <td>0.8655</td>\n",
       "      <td>0.002363</td>\n",
       "      <td>0.002565</td>\n",
       "      <td>13</td>\n",
       "      <td>0.001626</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>14</td>\n",
       "      <td>23.44</td>\n",
       "      <td>32.81</td>\n",
       "      <td>20.31</td>\n",
       "      <td>20.31</td>\n",
       "      <td>11.33</td>\n",
       "      <td>11.72</td>\n",
       "      <td>4.88</td>\n",
       "      <td>4.10</td>\n",
       "      <td>2.73</td>\n",
       "      <td>2.34</td>\n",
       "      <td>1.95</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.8872</td>\n",
       "      <td>0.8200</td>\n",
       "      <td>0.002061</td>\n",
       "      <td>0.003342</td>\n",
       "      <td>14</td>\n",
       "      <td>0.001639</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>15</td>\n",
       "      <td>25.00</td>\n",
       "      <td>35.94</td>\n",
       "      <td>21.09</td>\n",
       "      <td>21.88</td>\n",
       "      <td>12.11</td>\n",
       "      <td>12.50</td>\n",
       "      <td>5.27</td>\n",
       "      <td>4.49</td>\n",
       "      <td>2.93</td>\n",
       "      <td>2.54</td>\n",
       "      <td>1.95</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.8945</td>\n",
       "      <td>0.8760</td>\n",
       "      <td>0.001971</td>\n",
       "      <td>0.002495</td>\n",
       "      <td>15</td>\n",
       "      <td>0.001877</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>16</td>\n",
       "      <td>25.00</td>\n",
       "      <td>37.50</td>\n",
       "      <td>22.66</td>\n",
       "      <td>22.66</td>\n",
       "      <td>12.50</td>\n",
       "      <td>12.89</td>\n",
       "      <td>5.47</td>\n",
       "      <td>4.69</td>\n",
       "      <td>2.93</td>\n",
       "      <td>2.54</td>\n",
       "      <td>1.95</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.9040</td>\n",
       "      <td>0.8530</td>\n",
       "      <td>0.001858</td>\n",
       "      <td>0.002760</td>\n",
       "      <td>16</td>\n",
       "      <td>0.001664</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>17</td>\n",
       "      <td>26.56</td>\n",
       "      <td>40.62</td>\n",
       "      <td>23.44</td>\n",
       "      <td>24.22</td>\n",
       "      <td>13.67</td>\n",
       "      <td>14.06</td>\n",
       "      <td>5.66</td>\n",
       "      <td>4.88</td>\n",
       "      <td>2.93</td>\n",
       "      <td>2.73</td>\n",
       "      <td>1.56</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.8985</td>\n",
       "      <td>0.8395</td>\n",
       "      <td>0.001824</td>\n",
       "      <td>0.003089</td>\n",
       "      <td>17</td>\n",
       "      <td>0.001935</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>18</td>\n",
       "      <td>28.12</td>\n",
       "      <td>40.62</td>\n",
       "      <td>25.00</td>\n",
       "      <td>25.00</td>\n",
       "      <td>14.06</td>\n",
       "      <td>14.45</td>\n",
       "      <td>5.66</td>\n",
       "      <td>4.88</td>\n",
       "      <td>2.73</td>\n",
       "      <td>2.54</td>\n",
       "      <td>1.17</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.9095</td>\n",
       "      <td>0.8380</td>\n",
       "      <td>0.001633</td>\n",
       "      <td>0.003209</td>\n",
       "      <td>18</td>\n",
       "      <td>0.001696</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>19</td>\n",
       "      <td>28.12</td>\n",
       "      <td>42.19</td>\n",
       "      <td>25.78</td>\n",
       "      <td>25.00</td>\n",
       "      <td>14.45</td>\n",
       "      <td>14.84</td>\n",
       "      <td>5.86</td>\n",
       "      <td>5.08</td>\n",
       "      <td>2.73</td>\n",
       "      <td>2.73</td>\n",
       "      <td>1.17</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.9078</td>\n",
       "      <td>0.8575</td>\n",
       "      <td>0.001693</td>\n",
       "      <td>0.002992</td>\n",
       "      <td>19</td>\n",
       "      <td>0.001711</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    Unnamed: 0  train_features0  train_features3  train_features7  \\\n",
       "0            0             1.56             1.56             0.78   \n",
       "1            1             3.12             6.25             3.12   \n",
       "2            2             3.12             7.81             3.91   \n",
       "3            3             6.25            12.50             7.03   \n",
       "4            4             7.81            15.62             7.81   \n",
       "5            5            10.94            18.75            10.16   \n",
       "6            6            10.94            20.31            11.72   \n",
       "7            7            12.50            21.88            12.50   \n",
       "8            8            14.06            23.44            13.28   \n",
       "9            9            15.62            25.00            14.84   \n",
       "10          10            17.19            28.12            16.41   \n",
       "11          11            18.75            29.69            17.19   \n",
       "12          12            20.31            29.69            17.97   \n",
       "13          13            21.88            31.25            19.53   \n",
       "14          14            23.44            32.81            20.31   \n",
       "15          15            25.00            35.94            21.09   \n",
       "16          16            25.00            37.50            22.66   \n",
       "17          17            26.56            40.62            23.44   \n",
       "18          18            28.12            40.62            25.00   \n",
       "19          19            28.12            42.19            25.78   \n",
       "\n",
       "    train_features10  train_features14  train_features17  train_features21  \\\n",
       "0               0.78              0.39              0.39              0.20   \n",
       "1               3.12              1.56              1.17              0.59   \n",
       "2               3.91              1.95              1.95              0.98   \n",
       "3               7.03              3.52              3.52              1.37   \n",
       "4               8.59              4.30              4.30              1.76   \n",
       "5              10.94              5.47              5.86              2.34   \n",
       "6              11.72              6.25              6.64              2.54   \n",
       "7              13.28              6.64              7.03              2.93   \n",
       "8              14.06              7.42              7.81              3.12   \n",
       "9              14.84              7.81              8.20              3.32   \n",
       "10             17.19              8.98              9.38              3.71   \n",
       "11             17.97              9.38             10.16              4.10   \n",
       "12             18.75             10.16             10.55              4.30   \n",
       "13             19.53             10.55             11.33              4.49   \n",
       "14             20.31             11.33             11.72              4.88   \n",
       "15             21.88             12.11             12.50              5.27   \n",
       "16             22.66             12.50             12.89              5.47   \n",
       "17             24.22             13.67             14.06              5.66   \n",
       "18             25.00             14.06             14.45              5.66   \n",
       "19             25.00             14.45             14.84              5.86   \n",
       "\n",
       "    train_features24  train_features28  train_features31  train_classifier2  \\\n",
       "0               0.20              0.20              0.20               0.39   \n",
       "1               0.59              0.59              0.39               0.78   \n",
       "2               0.78              0.78              0.59               1.17   \n",
       "3               1.17              1.17              0.59               1.17   \n",
       "4               1.56              1.56              0.78               1.56   \n",
       "5               2.15              1.95              1.17               1.56   \n",
       "6               2.34              1.95              1.37               1.56   \n",
       "7               2.54              2.15              1.37               1.95   \n",
       "8               2.73              2.34              1.56               1.95   \n",
       "9               2.93              2.34              1.56               1.95   \n",
       "10              3.32              2.54              1.76               1.95   \n",
       "11              3.52              2.73              1.76               1.95   \n",
       "12              3.71              2.73              1.95               1.95   \n",
       "13              3.91              2.73              2.15               1.95   \n",
       "14              4.10              2.73              2.34               1.95   \n",
       "15              4.49              2.93              2.54               1.95   \n",
       "16              4.69              2.93              2.54               1.95   \n",
       "17              4.88              2.93              2.73               1.56   \n",
       "18              4.88              2.73              2.54               1.17   \n",
       "19              5.08              2.73              2.73               1.17   \n",
       "\n",
       "    train_classifier6  train_accuracy  test_accuracy  train_loss  test_loss  \\\n",
       "0                 0.0          0.6225         0.6995    0.005065   0.004839   \n",
       "1                 0.0          0.7218         0.7340    0.004275   0.004171   \n",
       "2                 0.0          0.7654         0.7625    0.003784   0.003929   \n",
       "3                 0.0          0.7871         0.7995    0.003504   0.003280   \n",
       "4                 0.0          0.7976         0.8040    0.003336   0.003363   \n",
       "5                 0.0          0.8258         0.8160    0.003005   0.003159   \n",
       "6                 0.0          0.8341         0.8480    0.002919   0.002829   \n",
       "7                 0.0          0.8427         0.8010    0.002770   0.003457   \n",
       "8                 0.0          0.8532         0.8195    0.002597   0.003333   \n",
       "9                 0.0          0.8591         0.8070    0.002520   0.003280   \n",
       "10                0.0          0.8664         0.8525    0.002399   0.002721   \n",
       "11                0.0          0.8726         0.8575    0.002312   0.002743   \n",
       "12                0.0          0.8805         0.7730    0.002207   0.004624   \n",
       "13                0.0          0.8690         0.8655    0.002363   0.002565   \n",
       "14                0.0          0.8872         0.8200    0.002061   0.003342   \n",
       "15                0.0          0.8945         0.8760    0.001971   0.002495   \n",
       "16                0.0          0.9040         0.8530    0.001858   0.002760   \n",
       "17                0.0          0.8985         0.8395    0.001824   0.003089   \n",
       "18                0.0          0.9095         0.8380    0.001633   0.003209   \n",
       "19                0.0          0.9078         0.8575    0.001693   0.002992   \n",
       "\n",
       "    epoch  time_per_step  \n",
       "0       0       0.001485  \n",
       "1       1       0.001489  \n",
       "2       2       0.001514  \n",
       "3       3       0.001593  \n",
       "4       4       0.001533  \n",
       "5       5       0.001652  \n",
       "6       6       0.001558  \n",
       "7       7       0.001567  \n",
       "8       8       0.001579  \n",
       "9       9       0.001593  \n",
       "10     10       0.001766  \n",
       "11     11       0.001614  \n",
       "12     12       0.001627  \n",
       "13     13       0.001626  \n",
       "14     14       0.001639  \n",
       "15     15       0.001877  \n",
       "16     16       0.001664  \n",
       "17     17       0.001935  \n",
       "18     18       0.001696  \n",
       "19     19       0.001711  "
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "cd_vgg13 = pd.read_csv('catdog_VGG13_A0.csv', sep=';')\n",
    "cd_vgg13.head(20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 504x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_saturation_level(extract_layer_saturation(cd_vgg13), '85.57% Test Accuracy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'deep' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-19-ba8b43d7102e>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[0;32m      1\u001b[0m \u001b[1;32mfor\u001b[0m \u001b[0mi\u001b[0m \u001b[1;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;36m20\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m----> 2\u001b[1;33m     \u001b[0mdeep_sats\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mextract_layer_saturation\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mdeep\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mepoch\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mi\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m      3\u001b[0m     \u001b[0mplot_saturation_level\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mdeep_sats\u001b[0m\u001b[1;33m,\u001b[0m \u001b[1;34m'85.75% Test Accuracy'\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mNameError\u001b[0m: name 'deep' is not defined"
     ]
    }
   ],
   "source": [
    "for i in range(20):\n",
    "    deep_sats = extract_layer_saturation(deep, epoch=i)\n",
    "    plot_saturation_level(deep_sats, '85.75% Test Accuracy')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'shallow' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-20-1543c948c6f1>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[0;32m      1\u001b[0m \u001b[1;32mfor\u001b[0m \u001b[0mi\u001b[0m \u001b[1;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;36m20\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m----> 2\u001b[1;33m     \u001b[0mdeep_sats\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mextract_layer_saturation\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mshallow\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mepoch\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mi\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m      3\u001b[0m     \u001b[0mplot_saturation_level\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mdeep_sats\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mNameError\u001b[0m: name 'shallow' is not defined"
     ]
    }
   ],
   "source": [
    "for i in range(20):\n",
    "    deep_sats = extract_layer_saturation(shallow, epoch=i)\n",
    "    plot_saturation_level(deep_sats)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "ename": "NameError",
     "evalue": "name 'medium' is not defined",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mNameError\u001b[0m                                 Traceback (most recent call last)",
      "\u001b[1;32m<ipython-input-21-804d055048fc>\u001b[0m in \u001b[0;36m<module>\u001b[1;34m\u001b[0m\n\u001b[0;32m      1\u001b[0m \u001b[1;32mfor\u001b[0m \u001b[0mi\u001b[0m \u001b[1;32min\u001b[0m \u001b[0mrange\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;36m20\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m----> 2\u001b[1;33m     \u001b[0mdeep_sats\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mextract_layer_saturation\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mmedium\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mepoch\u001b[0m\u001b[1;33m=\u001b[0m\u001b[0mi\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m      3\u001b[0m     \u001b[0mplot_saturation_level\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mdeep_sats\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mNameError\u001b[0m: name 'medium' is not defined"
     ]
    }
   ],
   "source": [
    "for i in range(20):\n",
    "    deep_sats = extract_layer_saturation(medium, epoch=i)\n",
    "    plot_saturation_level(deep_sats)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "cvd_csvs = []"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "cf_csvs = []"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Exception in thread Thread-6:\n",
      "Traceback (most recent call last):\n",
      "  File \"C:\\ProgramData\\Anaconda3\\lib\\threading.py\", line 917, in _bootstrap_inner\n",
      "    self.run()\n",
      "  File \"C:\\ProgramData\\Anaconda3\\lib\\site-packages\\matplotlib\\backends\\backend_webagg.py\", line 44, in run\n",
      "    tornado.ioloop.IOLoop.instance().start()\n",
      "  File \"C:\\ProgramData\\Anaconda3\\lib\\site-packages\\tornado\\ioloop.py\", line 221, in instance\n",
      "    return IOLoop.current()\n",
      "  File \"C:\\ProgramData\\Anaconda3\\lib\\site-packages\\tornado\\ioloop.py\", line 282, in current\n",
      "    loop = asyncio.get_event_loop()\n",
      "  File \"C:\\ProgramData\\Anaconda3\\lib\\asyncio\\events.py\", line 644, in get_event_loop\n",
      "    % threading.current_thread().name)\n",
      "RuntimeError: There is no current event loop in thread 'Thread-6'.\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/html": [
       "<!-- Within the kernel, we don't know the address of the matplotlib\n",
       "     websocket server, so we have to get in client-side and fetch our\n",
       "     resources that way. -->\n",
       "<script>\n",
       "  // We can't proceed until these Javascript files are fetched, so\n",
       "  // we fetch them synchronously\n",
       "  $.ajaxSetup({async: false});\n",
       "  $.getScript(\"http://\" + window.location.hostname + \":8988/_static/js/mpl_tornado.js\");\n",
       "  $.getScript(\"http://\" + window.location.hostname + \":8988/js/mpl.js\");\n",
       "  $.ajaxSetup({async: true});\n",
       "\n",
       "  function init_figure1(e) {\n",
       "    $('div.output').off('resize');\n",
       "\n",
       "    var output_div = $(e.target).find('div.output_subarea');\n",
       "    var websocket_type = mpl.get_websocket_type();\n",
       "    var websocket = new websocket_type(\n",
       "        \"ws://\" + window.location.hostname + \":8988/\" +\n",
       "        &#39;1&#39; + \"/ws\");\n",
       "\n",
       "    var fig = new mpl.figure(\n",
       "        &#39;1&#39;, websocket, mpl_ondownload, output_div);\n",
       "\n",
       "    // Fetch the first image\n",
       "    fig.context.drawImage(fig.imageObj, 0, 0);\n",
       "\n",
       "    fig.focus_on_mouseover = true;\n",
       "  }\n",
       "\n",
       "  // We can't initialize the figure contents until our content\n",
       "  // has been added to the DOM.  This is a bit of hack to get an\n",
       "  // event for that.\n",
       "  $('div.output').resize(init_figure1);\n",
       "</script>\n"
      ],
      "text/plain": [
       "<Figure size 864x576 with 20 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "matplotlib.use('webagg')\n",
    "\n",
    "\n",
    "cols = ['{} Filtersize'.format(col) for col in reversed(['Full','Halfed', 'Quartered', 'Eigth', 'Sixteenth'])]\n",
    "rows = ['{}'.format(row) for row in ['VGG11', 'VGG13', 'VGG16', 'VGG19']]\n",
    "\n",
    "fig, axes = plt.subplots(nrows=4, ncols=5, figsize=(12, 8))\n",
    "for ax, col in zip(axes[0], cols):\n",
    "    ax.set_title(col)\n",
    "\n",
    "for ax, row in zip(axes[:,0], rows):\n",
    "    ax.set_ylabel(row, rotation='vertical', size='large')\n",
    "\n",
    "    \n",
    "\n",
    "#fig.tight_layout()\n",
    "for i,net in enumerate([11, 13, 16, 19]):\n",
    "    for j,filterstack in enumerate(reversed(['', '_S', '_XS', '_XXS', '_XXXS'])):\n",
    "        filename = 'catdog_VGG{}{}_A0.csv'.format(net, filterstack)\n",
    "        df = extract_layer_saturation(pd.read_csv(filename, sep=';'), epoch=20)\n",
    "        plot_saturation_level_ax(df, axes[i][j])\n",
    "#plt.show()\n",
    "plt.savefig('collage_catdog.eps')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "ename": "RuntimeError",
     "evalue": "threads can only be started once",
     "output_type": "error",
     "traceback": [
      "\u001b[1;31m---------------------------------------------------------------------------\u001b[0m",
      "\u001b[1;31mRuntimeError\u001b[0m                              Traceback (most recent call last)",
      "\u001b[1;32mC:\\ProgramData\\Anaconda3\\lib\\site-packages\\IPython\\core\\formatters.py\u001b[0m in \u001b[0;36m__call__\u001b[1;34m(self, obj)\u001b[0m\n\u001b[0;32m    343\u001b[0m             \u001b[0mmethod\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mget_real_method\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mobj\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mprint_method\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    344\u001b[0m             \u001b[1;32mif\u001b[0m \u001b[0mmethod\u001b[0m \u001b[1;32mis\u001b[0m \u001b[1;32mnot\u001b[0m \u001b[1;32mNone\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 345\u001b[1;33m                 \u001b[1;32mreturn\u001b[0m \u001b[0mmethod\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    346\u001b[0m             \u001b[1;32mreturn\u001b[0m \u001b[1;32mNone\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    347\u001b[0m         \u001b[1;32melse\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32mC:\\ProgramData\\Anaconda3\\lib\\site-packages\\matplotlib\\figure.py\u001b[0m in \u001b[0;36m_repr_html_\u001b[1;34m(self)\u001b[0m\n\u001b[0;32m    408\u001b[0m                 'WebAgg' in self.canvas.__class__.__name__):\n\u001b[0;32m    409\u001b[0m             \u001b[1;32mfrom\u001b[0m \u001b[0mmatplotlib\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mbackends\u001b[0m \u001b[1;32mimport\u001b[0m \u001b[0mbackend_webagg\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 410\u001b[1;33m             \u001b[1;32mreturn\u001b[0m \u001b[0mbackend_webagg\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mipython_inline_display\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    411\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    412\u001b[0m     \u001b[1;32mdef\u001b[0m \u001b[0mshow\u001b[0m\u001b[1;33m(\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m,\u001b[0m \u001b[0mwarn\u001b[0m\u001b[1;33m=\u001b[0m\u001b[1;32mTrue\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32mC:\\ProgramData\\Anaconda3\\lib\\site-packages\\matplotlib\\backends\\backend_webagg.py\u001b[0m in \u001b[0;36mipython_inline_display\u001b[1;34m(figure)\u001b[0m\n\u001b[0;32m    296\u001b[0m     \u001b[0mWebAggApplication\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0minitialize\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    297\u001b[0m     \u001b[1;32mif\u001b[0m \u001b[1;32mnot\u001b[0m \u001b[0mwebagg_server_thread\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mis_alive\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 298\u001b[1;33m         \u001b[0mwebagg_server_thread\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mstart\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    299\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    300\u001b[0m     \u001b[0mfignum\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mfigure\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mnumber\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;32mC:\\ProgramData\\Anaconda3\\lib\\threading.py\u001b[0m in \u001b[0;36mstart\u001b[1;34m(self)\u001b[0m\n\u001b[0;32m    841\u001b[0m \u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    842\u001b[0m         \u001b[1;32mif\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0m_started\u001b[0m\u001b[1;33m.\u001b[0m\u001b[0mis_set\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[1;32m--> 843\u001b[1;33m             \u001b[1;32mraise\u001b[0m \u001b[0mRuntimeError\u001b[0m\u001b[1;33m(\u001b[0m\u001b[1;34m\"threads can only be started once\"\u001b[0m\u001b[1;33m)\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0m\u001b[0;32m    844\u001b[0m         \u001b[1;32mwith\u001b[0m \u001b[0m_active_limbo_lock\u001b[0m\u001b[1;33m:\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n\u001b[0;32m    845\u001b[0m             \u001b[0m_limbo\u001b[0m\u001b[1;33m[\u001b[0m\u001b[0mself\u001b[0m\u001b[1;33m]\u001b[0m \u001b[1;33m=\u001b[0m \u001b[0mself\u001b[0m\u001b[1;33m\u001b[0m\u001b[1;33m\u001b[0m\u001b[0m\n",
      "\u001b[1;31mRuntimeError\u001b[0m: threads can only be started once"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 20 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "cols = ['{} Filtersize'.format(col) for col in reversed(['Full','Halfed', 'Quartered', 'Eigth', 'Sixteenth'])]\n",
    "rows = ['{}'.format(row) for row in ['VGG11', 'VGG13', 'VGG16', 'VGG19']]\n",
    "\n",
    "fig, axes = plt.subplots(nrows=4, ncols=5, figsize=(12, 8))\n",
    "for ax, col in zip(axes[0], cols):\n",
    "    ax.set_title(col)\n",
    "\n",
    "for ax, row in zip(axes[:,0], rows):\n",
    "    ax.set_ylabel(row, rotation='vertical', size='large')\n",
    "\n",
    "    \n",
    "\n",
    "#fig.tight_layout()\n",
    "for i,net in enumerate([11, 13, 16, 19]):\n",
    "    for j,filterstack in enumerate(reversed(['', '_S', '_XS', '_XXS', '_XXXS'])):\n",
    "        filename = '10_VGG{}{}_A0.csv'.format(net, filterstack)\n",
    "        df = extract_layer_saturation(pd.read_csv(filename, sep=';'), epoch=20)\n",
    "        plot_saturation_level_ax(df, axes[i][j])\n",
    "#plt.show()\n",
    "plt.savefig('collage_cifar.eps')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Confusion matrix, without normalization\n",
      "[[13  0  0]\n",
      " [ 0 10  6]\n",
      " [ 0  0  9]]\n",
      "Normalized confusion matrix\n",
      "[[1.   0.   0.  ]\n",
      " [0.   0.62 0.38]\n",
      " [0.   0.   1.  ]]\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAUAAAAEYCAYAAAAtTS8wAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzt3XeYFFX2//H3hwyCiIIBMAFGUFEwx68RxbirYhYxLOq6uq7uz7Qr5ri6rmFdzGlFMYsBc1yVbEQBFSWpYCJIHM7vj3sHm3Zmumemu7qn+7x8+nG6uvrWqa7uww1Vt2RmOOdcOWpU6ACcc65QPAE658qWJ0DnXNnyBOicK1ueAJ1zZcsToHOubHkCzJKklpKelvSzpKH1KOdISS/kMrZCkbSjpM+KZXuS1pFkkpokFVNDIWmypN3j3+dJuj0P27hV0t9yXW4+qdTOA5R0BHAmsCEwBxgHXGZmb9Wz3KOB04DtzGxJvQMtcpIMWM/MJhU6lupImgycYGYvxefrAF8CTXN9jCTdDUw1swtyWW5S0j+rHJTXP5a3Qy7KK5SSqgFKOhP4J3A5sBqwFnALcEAOil8bmFAOyS8bXsvKH/9sE2RmJfEA2gJzgUNqWKc5IUFOj49/As3ja7sAU4G/AN8BM4Dj4msXAYuAxXEbxwODgPtTyl4HMKBJfN4f+IJQC/0SODJl+Vsp79sOGAn8HP+/XcprrwGXAG/Hcl4A2lezb5Xx/zUl/gOBfYAJwA/AeSnrbwW8A/wU170JaBZfeyPuy7y4v/1Syv9/wDfAfZXL4nu6xm1sEZ93BGYBu2Rx7O4B/hL/7hS3fUp83i2Wq7Tt3QcsBebHGP+acgyOBb6O2z8/y+O/3HGJyyxu/6R47BfFbT1dzX4YMBCYCPwI3MyvraxGwAXAV/H43Au0TfvuHB/jfiNl2XHAlFjeQGBL4IN43G5K2XZX4BXg+7jfDwArpbw+Gdg9/j2I+N2Nx31uymMJMCi+dg7wOeG79wlwUFy+EbAAqIjv+Skuvxu4NGWbJwKT4vF7CuiYzWeVaN4odOLK2Y5An3jwmtSwzsXAu8CqQAfgf8AlKQlkSVynKSFx/AK0S//SVPO88gvbBFgBmA1sEF9bA+ie/kMDVo4H/+j4vsPj81Xi66/FL+D6QMv4/Mpq9q0y/r/H+E8EZgL/BdoA3eOXtktcvxewTdzuOsB44Iz0H38V5V9FSCQtSUlIKV/48UArYDhwbZbHbgAxqQBHxH1+KOW1J1NiSN3eZOKPOu0Y3Bbj2wxYCGyUxfFfdlyq+gxI+3FXsx8GDANWIrQ+ZgJ9UvZjEtAFaA08BtyXFve9hO9Oy5RltwItgD3j8Xsixt+JkEh3jmV0A/aIx6YDIYn+s6rPirTvbso6PWPMm8fnhxD+IWtE+EdwHrBGDZ/Xss8I2JWQiLeIMd0IvJHNZ5Xko5SawKsAs6zmJuqRwMVm9p2ZzSTU7I5OeX1xfH2xmT1L+NdtgzrGsxToIamlmc0ws4+rWKcvMNHM7jOzJWb2IPApsF/KOneZ2QQzmw88TPiSVmcxob9zMTAEaA/cYGZz4vY/BjYFMLPRZvZu3O5k4D/Azlns04VmtjDGsxwzu43wL/p7hKR/fobyKr0O7CipEbATcDWwfXxt5/h6bVxkZvPN7H3gfUIihMzHPxeuNLOfzOxr4FV+PV5HAteZ2RdmNhc4Fzgsrbk7yMzmpX22l5jZAjN7gZCAHozxTwPeBDYHMLNJZvZiPDYzgevIfDyXkdSBkFxPM7OxscyhZjbdzJaa2UOEY7tVlkUeCdxpZmPMbGHc321jP22l6j6rxJRSAvweaJ+h/6QjoQlS6au4bFkZaQn0F8K/1rViZvMI/2IOBGZIekbShlnEUxlTp5Tn39Qinu/NrCL+Xfkj+jbl9fmV75e0vqRhkr6RNJvQb9q+hrIBZprZggzr3Ab0AG6MX/yMzOxzwj82PYEdCTWD6ZI2oG4JsLrPLNPxz4XabLsJoa+60pQqyks/ftUdz1UlDZE0LR7P+8l8PInvbQo8AvzXzIakLD9G0jhJP0n6iXBcsyqTtP2NSf976v7dzotSSoDvEJoIB9awznTCYEalteKyuphHaOpVWj31RTMbbmZ7EGpCnxISQ6Z4KmOaVseYauPfhLjWM7MVgfMI/Ww1qfGUAUmtCf1qdwCDJK1ci3heBw4m9ENOi8+PAdoRRvJrHU8Vajr+yx1PScsdzzpsK5ttL2H5hFafbVwR379pPJ5Hkfl4VrqR0M+3bIRb0tqE7+wfCV0yKwEfpZSZKdbl9lfSCoRWWhLf7ayVTAI0s58J/V83SzpQUitJTSXtLenquNqDwAWSOkhqH9e/v46bHAfsJGktSW0JVXwAJK0maf940BcSajcVVZTxLLC+pCMkNZHUD9iYUAPKtzaEfsq5sXZ6ctrr3xL6q2rjBmC0mZ0APEPovwJA0iBJr9Xw3tcJP7Y34vPXCKcdvZVSq01X2xhrOv7vA90l9ZTUgtBPVp9tVbXtP0taN/5DcTmhnzNXZxW0IQ5ISOoEnJ3NmyT9gVDLPsLMlqa8tAIhyc2M6x1HqAFW+hboLKlZNUX/Fzgufp7NCfv7XuxuKRolkwABzOw6wjmAFxAO3BTCj+qJuMqlwCjCKNqHwJi4rC7behF4KJY1muWTViPCaPJ0wgjYzsApVZTxPbBvXPd7wkjmvmY2qy4x1dJZhAGHOYR/6R9Ke30QcE9s/hyaqTBJBxAGogbGRWcCW0g6Mj5fkzCaXZ3XCT/iygT4FqFG9ka17wi1ngtijGdlipEajr+ZTSAMkrxE6OtKP2/0DmDjuK0nqL07CSPXbxDOClhASPC5chFhwOFnwj8+j2X5vsMJiX26pLnxcZ6ZfQL8g9Cy+hbYhOWP3yuEPuVvJP3m+2pmLwN/Ax4lnGXQFTisLjuWTyV3IrQrTpLGAbvFpO9cUfAE6JwrWyXVBHbOudrwBOicK1ueAJ1zZcsvuq4DNWlpatam0GEUxOYbrVXoEFzCxowZPcvMOuSirMYrrm225DcXEf2GzZ853Mz65GKbNfEEWAdq1obmG2Q8M6Qkvf3eTYUOwSWsZVOlX61UZ7Zkfla/nQXjbs72ipN68QTonEuOBI0aFzqKZTwBOueSpeIZevAE6JxLlrK9RDn/PAE65xLkTWDnXLkS3gR2zpUreRPYOVfGvAnsnCtP8iawc65MCa8BOufKVXHVAIsnEudceWikzI8MJN0p6TtJH6Usu0bSp5I+kPS4pJUyhlLPXXHOuexVNoEzPTK7m3ALhlQvAj3MbFNgAin36amOJ0DnXIJiEzjTIwMze4Nwv53UZS+k3GTqXaBzpnK8D9A5l6zszgNsL2lUyvPBZja4FlsZwG9v9PUbngCdc8nJfjaYWWbWu26b0PmEey4/kGldT4DOuWTlcRRY0rGEW83uZlnc8c0ToHMuWXm6FE5SH+D/ATub2S/ZvMcToHMuQbmZDUbSg8AuhL7CqcCFhFHf5sCLCkn2XTMbWFM5ngCdc8nJ0WwwZnZ4FYvvqG05ngCdcwny+QCdc+WsiC6F8wTonEuWzwfonCtLflc451w5k9cAnXPlSHgCdM6VKwllMd1VUopnOMb9xq0XHslXL1/BqKHnLVv291P6MuKhc3l3yDk8fcuprNGhbQEjTMYLw59n0+4b0H3Dblxz9ZWFDidRpbjvkjI+klISCVBSf0kdCx1Hrt339LsccOrNyy27/p6X2arfFWxz2JU89+ZHnHvS3gWKLhkVFRWc8adTefLp5xj7wScMHfIg4z/5pNBhJaJU990TYO71B0ouAb495nN++Hn5SxrnzFuw7O9WLZuTxfXeDdrIESPo2rUb63bpQrNmzTik32EMe/rJQoeViJLcd4EaKeMjKUWbACWtIOkZSe9L+khSP0m9JL0uabSk4ZLWkHQw0Bt4QNI4SS0l7SZprKQP49TZzWOZV0r6JE6ZfW1ctp+k9+L6L0larZD7nY1Bp+7HxOcu4bC9e3PJv58pdDh5NX36NDp3XnPZ806dOjNt2rQCRpScUtx3kbn25zXAoA8w3cw2M7MewPPAjcDBZtYLuBO4zMweAUYBR5pZT8AI02X3M7NNCAM9J0taGTgI6B6nzL40buctYBsz2xwYAvy1qmAknSRplKRRtmR+nnY5O4Nufpr19v4bQ54bxcB+OxU0lnyrqoZbTKOI+VSq+96oUaOMj8RiSWxLtfchsLukqyTtCKwJ9CDM9DAOuICqp7zeAPjSzCbE5/cAOwGzgQXA7ZJ+B1S2LTsDwyV9CJwNdK8qGDMbbGa9zay3mrTMzR7W08PPjeTA3XoWOoy86tSpM1OnTln2fNq0qXTsWHK9HVUq1X33GmAWYgLrRUiEVwC/Bz42s57xsYmZ7VnFW6v89OK9ArYCHgUOJNQoIdQqb4q1xT8ALXK7J7nVda0Oy/7uu/OmTJj8bQGjyb/eW27JpEkTmfzllyxatIihDw2h7777FzqsRJTkvivLR0KK9jzAOKr7g5ndL2kucBLQQdK2ZvaOpKbA+mb2MTAHaBPf+imwjqRuZjYJOBp4XVJroJWZPSvpXWBSXL8tUNmxcmxCu5eVe67oz4691qP9Sq2Z9PwlXHLrs/TZoTvrrb0qS5caX8/4gT9dNqTQYeZVkyZNuP6Gm9iv715UVFRwbP8BbNy9ykp6ySnFfRdKtImbiYp1FFHSXsA1wFJgMXAyYZ7/fxGSVhPgn2Z2m6TfA5cD84Ftge2Aa+M6I+N7VwaeJNTwBFxrZvdIOgC4npAE3wW2NLNdaoqtUatVrfkGh+Z0fxuKH0feVOgQXMJaNtXout6fI13T9l2t3f5XZFxv5l39crbNmhRtDdDMhgPDq3jpN73+ZvYooWlb6WVg87TVZhCawOnvfZKQGJ1zSSiicZyiTYDOuRIkiqoJ7AnQOZeoYjqVxxOgcy4xlSdCF4viqYs650pfji6Fi1d4fSfpo5RlK0t6UdLE+P92mcrxBOicS1SOToS+m3C1WKpzgJfNbD3CQOg5mQrxBOicS1QuaoBm9gbwQ9riAwhXfhH/f2CmcrwP0DmXqCxreO0ljUp5PtjMBmd4z2pmNgPAzGZIWjXTRjwBOucSU4sm7qyyPhHaOVea8nge4LeS1oi1vzWA7zLGkq9InHOuSvmbDOEpfr2e/1iyuMLLa4DOuUTl4jxASQ8CuxD6CqcCFwJXAg9LOh74GjgkUzmeAJ1ziZGgUQ6mvDezw6t5abfalOMJ0DmXoOK6EsQToHMuUUWU/zwBOucSlKMmcK54AnTOJUZ4AnTOlTFPgM658iTvA3TOlSnhE6I658qWvAnsnCtfXgN0zpUn7wN0zpUrPw3GOVfWvAnsnCtbRZT/PAHWxeYbrcXb791U6DAKYv0/P1XoEAri6H02LHQIJSFXs8HkiidA51yCfDYY51wZ8xqgc648+Wkwzrly5ZfCOefKWjE1gf2ucM65RFXeG7imRxZl/FnSx5I+kvSgpBZ1icUToHMuObEPMNOjxiKkTsCfgN5m1gNoDBxWl3CqbQJLWrGmN5rZ7Lps0DlXvpS72WCaAC0lLQZaAdPrWkh1PgaM5W9TXPncgLXqskHnXHlrVM9BEDObJulawr1/5wMvmNkLdSmr2gRoZmvWMT7nnKtWlvmvvaRRKc8Hm9ng8H61Aw4A1gV+AoZKOsrM7q9tLFmNAks6DOhiZpdL6gysZmaja7sx51x5k6Bxdk3gWWbWu5rXdge+NLOZoUw9BmwH1DoBZhwEkXQT8H/A0XHRL8Cttd2Qc85BTkaBvwa2kdRKYeXdgPF1iSWbGuB2ZraFpLEAZvaDpGZ12ZhzrryJnPQBvifpEWAMsAQYCwyuS1nZJMDFkhoRBj6QtAqwtC4bc865XAwCm9mFwIX1jiWLdW4GHgU6SLoIeAu4qr4bds6VoSyav0leKpexBmhm90oaTeh4BDjEzD7Kb1jOuVIksh4ESUS21wI3BhYTmsF+9Yhzrs6KaC6ErEaBzwceBDoCnYH/Sjo334E550pTg2oCA0cBvczsFwBJlwGjgSvyGZhzrvTU4jzARGSTAL9KW68J8EV+wnHOlbriSX81T4ZwPaHP7xfgY0nD4/M9CSPBzjlXaw1lQtTKkd6PgWdSlr+bv3Ccc6VMUsNoApvZHUkG4pwrD0VUAczcByipK3AZsDGwbNZVM1s/j3E550pQsZ0HmM05fXcDdxFi3xt4GBiSx5iccyWsmE6DySYBtjKz4QBm9rmZXUCYHcY552pNWTySks1pMAvjlDOfSxoITANWzW9YriovDH+es848nYqKCvoPOIGz/3pOoUPKm2uO6MluPVbj+zkL2eOK1wBo26optxzXm84rt2TqD/M55c5R/Dx/cWEDzbMFc2fzzA0XMPOrCUii7xmX03mjzQsdVp0V23mA2dQA/wy0JtyEZHvgRGBAPoOqiqSLJe2eec3fvG8XScPyEVOSKioqOONPp/Lk088x9oNPGDrkQcZ/8kmhw8qboe99zTG3LH/Cwal7rMfbE2ay8yWv8PaEmZyyR7cCRZecF/9zGV177cjAwc9zwk1P0n7NroUOqd4aVBPYzN4zszlm9rWZHW1m+5vZ2/kIRkGVMZnZ383spXxsNy2GorxX8sgRI+jatRvrdulCs2bNOKTfYQx7+slCh5U3Iz7/gZ9+WbTcsj02WZ1H3psCwCPvTWHPTdcoRGiJWfjLXL7+aCSb7XUwAI2bNqNF6xrvVdYg1PeucLlU04nQjxPnAKyKmf2uhvdeBXxlZrfE54OAOYSEeyjQHHjczC6UtA7wHPAqsC1wYJx2q3fc/p1mdr2ku4FhZvaIpC2BG4AVgIWEGWEXA/+O71sCnGlmr6bFtTJwJ9CFcIL3SWb2QYyvI7AOMAs4orp9K5Tp06fRufOvt2np1KkzI0a8V8CIkte+TXO+m70QgO9mL6R9m9Kel/enGVNo1XZlhl1/Lt998Smrd+vOHgPPp1mLVoUOrc4azHmAwE31KHcI8E/glvj8UOBKYAdgK0I/51OSdiJMb70BcJyZnSKpF9Ap3u8TSSulFhxno34I6GdmI+PtO+cDpwOY2SaSNgRekJR+qs5FwFgzO1DSrsC9QM/4Wi9gBzObX9UOSToJOAlgzbWSvyGe2W//LSqmM+pd7i2tWMI3kz5hz4F/o9OGm/HCrZfyzsOD2fmYMwodWr0U0/e2phOhX65roWY2VtKqkjoCHYAfgU0Jl9GNjau1BtYjJMCvzKyyw+cLoIukGwlXoKTf7m4DYIaZjYzbmg0gaQfgxrjsU0lfAekJcAfg93GdVyStIqltfO2p6pJfXH8wcdrtXr16V1szzpdOnTozdeqUZc+nTZtKx44dkw6joGbNWciqK4Za4KorNmfWnEWZ39SAtWm/Oiu2X51OG24GwIY79OGdoXWa+b2oFNN8evmM5RHgYKAfoUYo4Aoz6xkf3VKuNplX+SYz+xHYDHgNOBW4Pa3cyvsSp8vmn5Wq1qksa14VrxWN3ltuyaRJE5n85ZcsWrSIoQ8Noe+++xc6rES9+OE3HLx16AY4eOs1efHDbwocUX61XrkDbTqszvdTw9wjk8e9Q/u1GvYgSOWJ0JkeSclnh/8Q4DagPbAzsAlwiaQHzGyupE6EfrvlSGoPLDKzRyV9TjgRO9WnQEdJW8YmcBtCE/gN4Ejgldj0XQv4jNCvWKlynUsk7UK49d7sYqqSV6dJkyZcf8NN7Nd3LyoqKji2/wA27t690GHlzY39t2Dbbu1p17oZ7128B9c9+xm3vDiRfw/oTb9t1mL6j/MZeOeozAU1cHsN/BtPXn0WFUsW0271Nen754Y/C10RdQFmnwAlNTezhdmub2Yfx+Q0zcxmADMkbQS8ExPOXMJcgxVpb+0E3JUyGrzc5KtmtkhSP+BGSS0JyW93Qn/jrZI+JAyC9DezhWnJbVAs+wPCIMix2e5PMeiz9z702XufQoeRiNPuHlPl8sNveifhSAprta4bMeBfjxU6jJzJ5XmAcXzgdqAHoSU3wMxq9QXJ5lrgrYA7gLbAWpI2A04ws9MyvdfMNkl7fgNh9DZdj5R13ge2qKKs/il/jwS2qaKc/ukLzOw1QnMaM/uBcEf59HUGVRW/cy73ctjgugF43swOjoOjtR4ez6YP8F/AvsD3sCxB+aVwzrlaq7wvcKZHxnLC2R87ESpnmNkiM/uptvFkkwAbmdlXacvSm63OOZeVxsr8yEIXYCahS2uspNslrVDbWLJJgFNiM9gkNZZ0BjChthtyzjllUfuLNcD2kkalPE5KK6oJoavs32a2OeEsjlpfHJ/NIMjJhGbwWsC3wEtxmXPO1VqWfYCzzKx3Da9PBaaaWeXlUI+QjwRoZt8Bh9W2YOecSyegSQ5Ggc3sG0lTJG1gZp8RLoet9ewg2YwC30YVJx6bWXqV1DnnMsrhKPBpwANxBPgL4LjaFpBNEzh1BpYWwEHAlGrWdc656il3J0Kb2TjC5Cd1lk0T+KHU55LuA16sz0adc+VJQOMiuvKqLpfCrQusnetAnHPloUFdCifpR37tA2wE/EAdRluccw4ayHRYEGZoJszMMi0uWmpVTUznnHNZCNcCFzqKX9UYSkx2j5tZRXx48nPO1UsuLoXLWSxZrDNC0m8mJ3DOudoK8wFmfiSlpnuCNDGzJYRZlE+Mc/PNI05IamaeFJ1ztSQaJXrn35rV1Ac4gnCt3YEJxeKcK3Ei2bu+ZVJTAhSAmX2eUCzOuVKn3FwKlys1JcAOks6s7kUzuy4P8TjnSlhDqgE2Jty5rYjCdc41dEmO8mZSUwKcYWYXJxaJc67khUvhCh3FrzL2ATrnXM6o4VwJsltiUTjnykbxpL8aEmC8g5pzzuVMKcwG45xzdVZE+c8ToHMuOUJeA3TOla+GMgji3G9MuH7/QodQENtc+nKhQygZxZP+PAE65xIkFdcgSBFNTeicKweSMj6yLKexpLGShtU1Fk+AzrlEKYtHlk4HxtcnFk+AzrnEVJ4HmOmRsRypM9AXuL0+8XgfoHMuUVm2cNtLGpXyfLCZDU55/k/gr0Cb+sTiCdA5lyCh7Bq5s8ysypueS9oX+M7MRkvapT7ReAJ0ziUmR5fCbQ/sL2kfoAWwoqT7zeyo2hbkfYDOueQoNIEzPWpiZueaWWczWwc4DHilLskPvAbonEtYQ5kQ1TnnckpALm8JYmavAa/V9f2eAJ1zicpyECQRngCdc4nyJrBzrizluglcX54AnXMJyvo8wER4AnTOJUdeA3TOlanQBC6eDOgJ0DmXqOJJf54AnXMJ8ynxnXNlq4jynydA51yyPAE658pSmPG5eDKgJ0DnXHKymO0lSZ4AnXOJ8gTonCtTfiWIc66MeQ3QOVeWhCdA51wZK6YmsN8TpAF5YfjzbNp9A7pv2I1rrr6y0OEkplz3G+CIrdfkkVO25tFTtubIbdYsdDg5Ud97guRSwROgpI6SHqnD+26XtHGGdQZKOqbu0RWPiooKzvjTqTz59HOM/eAThg55kPGffFLosPKuXPcboOuqK/C7Xh056raRHHrrCHZcvz1rrdyy0GHVTw5uipRLBU+AZjbdzA5OXy6pxua5mZ1gZjX+EszsVjO7t74xFoORI0bQtWs31u3ShWbNmnFIv8MY9vSThQ4r78p1vwG6tF+BD6b+zILFS6lYaoye/CO7btSh0GHVm7L4LymJJkBJV0k6JeX5IEl/kfRRfN5f0lBJTwMvSGok6RZJH0saJulZSQfHdV+T1Dv+PVfSZZLel/SupNVSyj8r/t1N0ktxnTGSukpqLenl+PxDSQck+XnUxvTp0+jc+dcmUKdOnZk2bVoBI0pGue43wKTv5tJr7Xa0bdmEFk0bscN67VltxRaFDqteKmeEzvTIWI60pqRXJY2P+eH0usSTdA1wCNAv5fmhwMi0dbYFjjWzXYHfAesAmwAnxNeqsgLwrpltBrwBnFjFOg8AN8d1tgNmAAuAg8xsC+D/gH+omqkqJJ0kaZSkUTNnzcy4o7lmZlXFlHgcSSvX/Qb4ctYv3PXWZG49ZnNuPqonE76dQ8XS334eDY6yeGS2BPiLmW0EbAOcmqlLrCqJjgKb2VhJq0rqCHQAfgS+TlvtRTP7If69AzDUzJYC30h6tZqiFwHD4t+jgT1SX5TUBuhkZo/HOBbE5U2ByyXtBCwFOgGrAd9UEftgYDBAr169E/8WdurUmalTpyx7Pm3aVDp27Jh0GIkr1/2u9MTYGTwxdgYAp+3WlW9nLyhwRPWXiyaumc0gVGIwszmSxhN+v7XqIC5EH+AjwMGEmuCQKl6fl/J3tp/UYvu1qlDBbxN7deUcSUjEvcysJ/AtUJRtjN5bbsmkSROZ/OWXLFq0iKEPDaHvvvsXOqy8K9f9rtRuhaYArN62Obtu1IHnPvy2wBHVX5ZN4PaVLa74OKm68iStA2wOvFfbWApxHuAQ4DagPbAz0LyGdd8CjpV0DyFR7QL8t7YbNLPZkqZKOtDMnpDUHGgMtAW+M7PFkv4PWLu2ZSelSZMmXH/DTezXdy8qKio4tv8ANu7evdBh5V257nelfxy6KW1bNWVJxVKueOYz5ixYUuiQ6i+7as0sM+udsSipNfAocIaZza5tKIknQDP7ODZJp5nZjJi9q/MosBvwETCBkOF/ruOmjwb+I+liYDFwCKFf8GlJo4BxwKd1LDsRffbehz5771PoMBJXrvsNMOCu0YUOIadyOR1W7MJ6FHjAzB6rSxkFuRLEzDZJ+Xsy0CP+fTdwd8prSyWdZWZzJa0CjAA+jK/tkrJe65S/HyE0szGzQSnLJwK7VhFOdQMrzrlcy9Fd4eJg5R3AeDO7rq7lNIRL4YZJWgloBlxiZr8ZoHDONSC5qQBuT2jVfShpXFx2npk9W5tCij4Bptb0nHMNXW5OdDazt8hBKi36BOicKx2VJ0IXC0+AzrlkeQJ0zpWrRkV0JY8nQOdcooon/XkCdM4lye8K55wrV2FK/OLJgJ4AnXOJKp705wnQOZewIqoAegJ0ziXLm8DOubJVPOnPE6BzLkFJ3/QoE0+AzrlEeRPYOVe2iif9eQJ0ziVKfimcc648hROhCx3Frwp+Y3TnnCsUrwHpJ4SkAAAOcUlEQVQ65xLlTWDnXHny02Ccc+VK+Ciwc66MFdN5gD4I4pxLVOXVIDU9MpehPpI+kzRJ0jl1jcUToHMuUcriUeP7pcbAzcDewMbA4ZI2rkssngCdc4mSlPGRwVbAJDP7wswWAUOAA+oSi/cB1sGYMaNntWyqrwq0+fbArAJtu9DKdd8Lvd9r56qgsWNGD2/VTO2zWLWFpFEpzweb2eD4dydgSsprU4Gt6xKPJ8A6MLMOhdq2pFFm1rtQ2y+kct33UtpvM+uTg2KqqiJaXQryJrBzrqGZCqyZ8rwzML0uBXkCdM41NCOB9SStK6kZcBjwVF0K8iZwwzM48yolq1z3vVz3u0pmtkTSH4HhQGPgTjP7uC5lyaxOTWfnnGvwvAnsnCtbngCdc2XLE6Bzrmx5AnRFTcV05bwrOZ4AXdGSJIujdJKOlrRDoWNypcUTYAMiqaxOW0pJfn0I53p9VtiIkuM132SU1Q+qIZN0CrC1pMnAS2b2ZoFDSoSkrYABwPtmNjMuW1YzLEWV+ydpD2B9YKGZ3V7ouEqR1wAbAEmnAocANxFmwrhc0n6FjSo/qqj5fA98DWwqaXsINcNSriHF/dsH+CcwAfiHpCvjNFAuhzwBFjlJKwLtgP2B7eLie4CzJfUtWGB5kNbnt2/cvw7AhcA4YD9J28KvzeNSJGll4HSgH+E3OhHoA9wqyX+zOeQfZhGT1NPMZgM3Ah0JSfB3hOseGwOnSlqhhGpDApA0ELgc6A08BhwE3AAsBI6IzeKSUnkMJa1sZj8ARxBmOLk0zgSzD3A8cHEJHe+C8wRYpCSdTviydzaznwnHaj7hR7ELMArob2bzGnptSNKGsfa3VFJHwoDHEWZ2EaHmcwmwPfBvYAbwZeGizb2UPr99gQclrWFm3xP66L+W1JxQE74fGN7Qj3cx8UGQIiTpAEINYC8z+0nS6mY2XtI04GHCNOAHmtl3BQ00ByS1Bs4Clkr6g5lNjwM9LSQ1NrMPJP0F2NfMnpB0bZwFuGTE5Lc9cCnwJzObEV+aA3wD3EXo+z3ezN4s9UGgJHkNsIik9O+sDYwBukm6GBgm6X9m9gdgILCNmX1UqDhz7BfC4E4FodMfYBrwF6BtfL4K0Dx+PosTjzAPJK0mae+URZ2Bh83sDUktAczsC0Kt9zZCbf/1uNyTX474bDBFRFI7M/tRUjtCTa+CMODxDHA7cLmZjStkjLmSNuDRCNgIOBuYZmbnS7oVWJ1QC9oQOK6Ekj6Sfg98AMwE5hGa/aeY2bYp62wLVJjZiMJEWfo8ARYJSScRbuwyGRhnZrelvHYAcAWwW0rzqMFKS37rEio1k+Odvc4EvjGzCyR1J9z/YYKZTS5cxPkRR3svBt4xswck/RdYETgB6A78BzjJzF4pYJglzRNgEYi1gUGEE37XJwxyfA9cQBj1vQg4pJRqQACS/syvp3p8TOgDawWcASwBBpZacy8t+TcjJLuNgVeBYcAtwEqEGyFdZWbPFirWcuAJsADSO7ElHQesaGY3xP6fjQhJ4EJCH1kLMyvUXejyIjbvrgf2IIxu/xtYZGanSuoBnAhcYWbfFDDMvJC0IyHBfRoHt/oTBjleMLMn4jqV3SE+4JFHPgqcMElNCTW8F+O03h8BPwLnSnrBzMYDY2I/YHszG1m4aHOnih/yXMJgR1MzmxPP/XtP0vFmdoeks0tptFdSo3iaz5bAfcD/gMWSXjWzuyVVAAdIakM43eUn8AGPfPMEmLzGwEGSBhH6e/aL/V9dgBslXUo452tV6ninq2KT1uw7FhhLqNkuJFziNtbMfpb0GLAAoFSSn6TmZrYwJr/dCV0aB5rZOEn7A7+TREyCTYAxnvSS4wkwYWa2QNIQYE/gdWBK/OL/h9DvdRYhMZxoZtMKF2nupCS/U4GTgH5mNknSK8CfgImSFgKHEgaCSoKk9oSa/YVmNpfQtTEQeI5wad+bhBPbj5bUxCc8SJ73ASYs/iiaEpLdVYSm4OVm9o2kVmb2i6SmZtbgz3eTtArws4W7eK0BDAGOSe3PlLQnYaR3feAuM5tQmGjzI9bslwLtzGyspLOA84CtzWxi7OrYGfjSzN4vZKzlyBNggmINqC8wCRgP3Es4z28S4QTfgwgTHsxp6M0gSd0INbrrgEWEk5mfBvY0s9mSmpnZIkntzWxWIWPNh3gVS0X8++/AbsDpsel7NvBnYHcz+yTW/pYUMt5y5VeCJETSYYQprU4CVgZ2NrN5hNMg5sRlR5jZ7Iae/ADMbBJhZHcjYA8Lc/m9D1wff/CLJA0A7pPUotQu8DezCkndJG1tZhcT7mF7qaTNzewawukub0tagVBDdAXgNcAExOtd9wS+AnoBBwP7xKbhumb2ZanUAioTWUq/30XAOsAdhIkMTgN2JNQG9wOOLqXzG1MmNtiOcJJzK+BkM3tf0gXAlsAlZjZKUpd4uZsrEE+AeaYwk3NzwujmVcAIM9s9vnYi0A34u5ktLFyUuZE22nsQ8K2Z/S/+8DsCjxJO+D2EMAr8qZlNLFjAeSJpN8KVO1cSzmecCgw2s5FxlL8XoXtgbinU9hsyT4B5JOkPhDncDjKzaZKuIpz1fyqwL/AHQrP34wKGmXOSzgQOJwx4jI/LzgI2AB4CXi+FQZ7qSLoW+M7MrlaYyuoSYAvgL7EmuF4pJv6GyPsA8yRe0bE38DdgoaSTCQMdPQmXfO1CiSS/1P67eBXHwYTBnImSdpd0rJldS7jOeV/CKHjJkbSPwq0KxgBdJXWKNfvzCed2HiOpdRz9Lak+z4bKzwPMEzObL+lZQlNoKuGOZl8BDxIucVtcKn1+Kc3evoTR7emEU16+AVYDVpG0ipldFkd9fylcxPkhqSfwR+DvhES/E7CbpDcJFY0vgG0Ig2DXedO3OHgCzK97CVc9fG5mP0g6Evg9oeuhwSc/WG6wYw/CdFb9CJM3HA/8J17rehxhvjtK5ZQXhVlseprZ4/EcxzOApWY2Kr7+MrAtcCzhut+Dga0J/yC4IuF9gAlQmO/uOMKP5PBSGvUEkLQN8Dhwhpk9lPba8YQ+z6NLoblfSVIvQs3u03gt8wBCn+5gM7sjrrMysAKh62ML4GrCVTAl8zk0dN4HmIwWhHO9Di2F5FdF/9UYwmV9F8ZOfyS1lLQB4Z4ex5baj97MRgOzgFGSBpjZnYSZrbeRdHRc5wczm0I4A+Bkwj9+JfU5NHReA0xIFbOhNEhpfX57EWo44wjJ4HJgPcKo9y8K8901NrP5BQs4TyStSjjFZQbhZPbBcUKDIwl3cHvBzO5JWb9ZqUzwUEq8DzAhpZD8YLk+v7MIJzKPIkzgcG78/1XAq5J2KcXEl+J7YDPCFTwDgbskLbYws3NjQt9vqpI97ach8yawqzWFqet7mNnOhDn9ZgNvEX7k5xJmOelQuAjzR1JHSV3jdb6nECZxaEu4kflFko4xs3vN7MPU95XKP4ClxpvArlYUJvTcnnB526pAO2B/M1ss6VDgJQs39i458brdqwgj2k8CDxAmNZhiZv+NV4AsMrM3CximqwVvArusxcGPnQmnc4wANgH+GJNff8KtLN8qXIT5ZWbzJJ0HbEqY5WZ1wuexnqTRZvYylE5/bznwGqDLSspchU0IE3r+SDjBuwvwHaFWeGi5jHJK6ki4rHF/ws2sdjKzMYWNytWWJ0CXkaRdCTWdkWY2LJ703AN4ntAMXpkwlXtJ3bgpW5LWtxKbyLVceBPYZWMyoaZ3taT1CLNZHwC8bWavFzKwQlK80VFl8vOmb8PjNUCXNUnrA4cRpvc6FxgKHAUs8R++a4g8AbpaiVd6iHDO38Pe9HMNmSdAVyvezHOlxBOgc65s+ZUgzrmy5QnQOVe2PAE658qWJ0DnXNnyBOicK1ueAF21JFVIGifpI0lDJbWqR1m7SBoW/95f0jk1rLtSvJ9ybbcxKM5TmNXytHXulnRwLba1jqQGP7t3ufME6Goy38x6mlkPYBFh4s9lFNT6O2RmT5nZlTWsshJhrj3n8soToMvWm0C3WPMZL+kWwr1A1pS0p6R3JI2JNcXWAJL6SPpU0lvA7yoLktRf0k3x79UkPS7p/fjYDriScF/dcZKuieudLWmkpA8kXZRS1vmSPpP0EuHG6zWSdGIs531Jj6bVaneX9KakCZL2jes3lnRNyrb/UN8P0hUPT4AuozgF1t5A5SzHGwD3mtnmwDzgAmB3M9uCMEX+mZJaALcRps3fkTB3XlX+BbxuZpsR7pz2MXAO4VaiPc3sbEl7Eu41shXhxvK9JO0U78x2GLA5IcFumcXuPGZmW8btjSfcvrPSOoRZb/oCt8Z9OB742cy2jOWfqHBLTFcCfDYYV5OWksbFv98E7gA6Al+Z2btx+TaEefHejjeLawa8A2wIfGlmEwEk3U+4KXi6XYFjAOI08z9Lape2zp7xUXmfjdaEhNgGeLzyRuuSnspin3pIupTQzG4NDE957WEzWwpMlPRF3Ic9gU1T+gfbxm37NdAlwBOgq8l8M+uZuiAmuXmpi4AXzezwtPV6Arm6zlLAFWb2n7RtnFGHbdwNHGhm78dZrHdJeS29LIvbPs3MUhMlktap5XZdEfImsKuvd4HtJXWDMHN0nDbrU2BdSV3jeodX8/6XCffMrexvWxGYQ6jdVRoODEjpW+ykcFvKN4CDFO5B3IbQ3M6kDTBDUlPgyLTXDpHUKMbcBfgsbvvkuD6S1o/3BnElwGuArl7MbGasST0Yp8oCuMDMJkg6CXhG0izCvUJ6VFHE6cBgSccDFcDJZvaOpLfjaSbPxX7AjYB3Yg10LnCUmY2R9BDhvsRfEZrpmfwNeC+u/yHLJ9rPCDd4Xw0YaGYLJN1O6Bsco7DxmcCB2X06rtj5bDDOubLlTWDnXNnyBOicK1ueAJ1zZcsToHOubHkCdM6VLU+Azrmy5QnQOVe2/j+DZUusF020mQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from sklearn import svm, datasets\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import confusion_matrix\n",
    "from sklearn.utils.multiclass import unique_labels\n",
    "\n",
    "# import some data to play with\n",
    "iris = datasets.load_iris()\n",
    "X = iris.data\n",
    "y = iris.target\n",
    "class_names = iris.target_names\n",
    "\n",
    "# Split the data into a training set and a test set\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y, random_state=0)\n",
    "\n",
    "# Run classifier, using a model that is too regularized (C too low) to see\n",
    "# the impact on the results\n",
    "classifier = svm.SVC(kernel='linear', C=0.01)\n",
    "y_pred = classifier.fit(X_train, y_train).predict(X_test)\n",
    "\n",
    "\n",
    "def plot_confusion_matrix(y_true, y_pred, classes,\n",
    "                          normalize=False,\n",
    "                          title=None,\n",
    "                          cmap=plt.cm.Blues):\n",
    "    \"\"\"\n",
    "    This function prints and plots the confusion matrix.\n",
    "    Normalization can be applied by setting `normalize=True`.\n",
    "    \"\"\"\n",
    "    if not title:\n",
    "        if normalize:\n",
    "            title = 'Normalized confusion matrix'\n",
    "        else:\n",
    "            title = 'Confusion matrix, without normalization'\n",
    "\n",
    "    # Compute confusion matrix\n",
    "    cm = confusion_matrix(y_true, y_pred)\n",
    "    # Only use the labels that appear in the data\n",
    "    classes = classes[unique_labels(y_true, y_pred)]\n",
    "    if normalize:\n",
    "        cm = cm.astype('float') / cm.sum(axis=1)[:, np.newaxis]\n",
    "        print(\"Normalized confusion matrix\")\n",
    "    else:\n",
    "        print('Confusion matrix, without normalization')\n",
    "\n",
    "    print(cm)\n",
    "\n",
    "    fig, ax = plt.subplots()\n",
    "    im = ax.imshow(cm, interpolation='nearest', cmap=cmap)\n",
    "    ax.figure.colorbar(im, ax=ax)\n",
    "    # We want to show all ticks...\n",
    "    ax.set(xticks=np.arange(cm.shape[1]),\n",
    "           yticks=np.arange(cm.shape[0]),\n",
    "           # ... and label them with the respective list entries\n",
    "           xticklabels=classes, yticklabels=classes,\n",
    "           title=title,\n",
    "           ylabel='True label',\n",
    "           xlabel='Predicted label')\n",
    "\n",
    "    # Rotate the tick labels and set their alignment.\n",
    "    plt.setp(ax.get_xticklabels(), rotation=45, ha=\"right\",\n",
    "             rotation_mode=\"anchor\")\n",
    "\n",
    "    # Loop over data dimensions and create text annotations.\n",
    "    fmt = '.2f' if normalize else 'd'\n",
    "    thresh = cm.max() / 2.\n",
    "    for i in range(cm.shape[0]):\n",
    "        for j in range(cm.shape[1]):\n",
    "            ax.text(j, i, format(cm[i, j], fmt),\n",
    "                    ha=\"center\", va=\"center\",\n",
    "                    color=\"white\" if cm[i, j] > thresh else \"black\")\n",
    "    fig.tight_layout()\n",
    "    return ax\n",
    "\n",
    "\n",
    "np.set_printoptions(precision=2)\n",
    "\n",
    "# Plot non-normalized confusion matrix\n",
    "plot_confusion_matrix(y_test, y_pred, classes=class_names,\n",
    "                      title='Confusion matrix, without normalization')\n",
    "\n",
    "# Plot normalized confusion matrix\n",
    "plot_confusion_matrix(y_test, y_pred, classes=class_names, normalize=True,\n",
    "                      title='Normalized confusion matrix')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_0 = np.zeros(10000).astype(int)\n",
    "y_pred_0[:615] = 1\n",
    "y_pred_1 = np.ones(10000).astype(int)\n",
    "y_pred_1[:59] = 0\n",
    "y_pred = np.append(y_pred_0, y_pred_1)\n",
    "\n",
    "y_true_0 = np.zeros(10000).astype(int)\n",
    "y_true_1 = np.ones(10000).astype(int)\n",
    "y_true = np.append(y_true_0, y_true_1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred[:610] = 1\n",
    "y_true[-59:] = 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Normalized confusion matrix\n",
      "[[0.94 0.06]\n",
      " [0.01 0.99]]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x227d9b5dc88>"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAUQAAAEYCAYAAAAkpo9KAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzt3XecVcX5x/HPlwVEkA4WiiAioakgLaIiCipGAsbEiB2xxcRefvYauyZqosYYC2LvBcFgC7aIgqgoKAoIwmIBNCii4C7P74+ZhbvL7t0Lu3vP3d3n7eu+PGXOnDl3dx9m5syZIzPDOecc1Em6AM45lys8IDrnXOQB0TnnIg+IzjkXeUB0zrnIA6JzzkUeEGs5SZdIui8uby1phaS8Sj7HfElDKzPPDM55gqSv4vW0rEA+KyR1qsyyJUXSTEmDky5HLvOAWMViMPhKUqOUbcdImpxgsUplZp+b2WZmVph0WSpCUj3gr8De8XqWbWxe8fh5lVe6yidprKTLy0tnZj3MbHIWilRteUDMjrrAKRXNRIH/zMq3BdAAmJl0QXKBpLpJl6G68D+u7LgOOFNSs9J2Shooaaqk5fH/A1P2TZZ0haQ3gJVAp7jtckn/jU268ZJaSrpf0ncxj44pedwkaWHc946k3cooR0dJJqmupJ1j3kWfnyTNj+nqSDpH0lxJyyQ9IqlFSj6HS1oQ952f7ouRtKmkv8T0yyW9LmnTuG9EbOb9L15zt5Tj5ks6U9KMeNzDkhpI6gLMjsn+J+nl1Osq8b0eE5c7S3ol5rNU0sMp6UxS57jcVNI4SUtieS8o+gdK0uhY9uslfSvpM0n7prnu+ZLOiuX/QdKdkraQ9Jyk7yW9KKl5SvpHJX0Zy/iqpB5x+3HAocD/Ff0upOR/tqQZwA/xZ7q260LSREl/Scn/YUl3pftZ1Qpm5p8q/ADzgaHAE8DlcdsxwOS43AL4FjicUJM8OK63jPsnA58DPeL+enHbHGBboCkwC/gknqcuMA64O6UMhwEt474zgC+BBnHfJcB9cbkjYEDdEtdQdM6r4vqpwBSgHbAJ8E/gwbivO7ACGBT3/RUoAIaW8f3cEvNuC+QBA+NxXYAfgL3i+f8vXnP9lO/1baBN/A4/Av5Q2nWUdl3xnMfE5QeB8wkVhAbArinpDOgcl8cBTwONY56fAEfHfaOBn4Fj43WcACwGlOb3YgqhNtsW+BqYDvSO1/8ycHFK+jHxvJsANwLvpewbS/zdKpH/e0B7YNPU38W4vGU8556EgDoPaJz030vSn8QLUNM/rAuIPYHlQGuKB8TDgbdLHPMmMDouTwYuK7F/MnB+yvpfgOdS1n+d+gdTSpm+BXaMy5dQfkD8BzABqBPXPwKGpOzfKgaDusBFwEMp+xoBqyklIMYA9GNRWUrsuxB4pETafGBwyvd6WMr+a4HbSruO0q6L4gFxHHA70K6UchjQmRDkVgHdU/Ydn/JzHA3MSdnXMB67ZZrfi0NT1h8H/pGyfhLwVBnHNot5N43rYyk9II4p7XcxZf0AYCGwlJR/BGrzx5vMWWJmHwLPAueU2NUGWFBi2wJCraHIwlKy/Cpl+cdS1jcrWpF0hqSPYnPrf4RaZatMyi3peGAwcIiZrYmbOwBPxqbs/wgBspBQ22mTWl4z+wEo66ZGK0KNbG4p+4p9L/HcCyn+vXyZsrySlGveQP8HCHg7NtHHlFHW+hT/WZX8Oa0tj5mtjIvpypTRz1BSnqSrYxfFd4TAVlSmdEr7vUn1LCHQzzaz18tJWyt4QMyuiwlNqtQ/osWEAJNqa0JtqMhGT0kU+wvPBn4PNDezZoSaqjI89s/ASDNbnrJrIbCvmTVL+TQws3zgC0IzrSiPhoTmemmWAj8Rmv4lFfteJCnmm19K2vL8EP/fMGXblkULZvalmR1rZm0Itb5bi/oNS5T1Z4r/rEr+nKrKIcBIQkujKaHGC+t+hmX9fpT3e3MF4R+zrSQdXMEy1ggeELPIzOYADwMnp2yeCHSRdEjs+D6I0A/3bCWdtjGhD28JUFfSRUCT8g6S1D6W9Qgz+6TE7tuAKyR1iGlbSxoZ9z0GDJe0q6T6wGWU8XsWa313AX+V1CbWhHaWtAnwCLCfpCEKw2jOIDRZ/7tBVx/Os4QQuA6L5xhDShCWdKCkdnH1W0IgKSyRR2Es0xWSGsdrPx24b0PLsxEaE659GSGoX1li/1fABo2VlDQIOAo4In7+Lqlt+qNqPg+I2XcZoV8NAAtj5IYT/uCXEZpvw81saSWdbxLwHOEGwAJCjay8phTAEEIt6jGtu9NcNIzlJuAZ4HlJ3xNuDgyI1zMT+BPwAKG2+C2wKM15zgQ+AKYC3wDXEPoqZxNuBv2dUDv7NfBrM1ud4XWXdCxwFuE77kHxwNoPeEvSinhdp5jZZ6XkcRKhtjkPeD1eYzbuzI4j/OzyCTfQppTYfyfQPXZhPFVeZpKaxDxPNLP82Fy+E7g71sRrLcXOVeecq/W8huicc5EHROecizwgOudc5AHROecif+i7kqheQ1ODUh9VdlVgh+22SroItcrCzxewbOnSSrsDndekg1nBj2nT2I9LJpnZsMo6ZyY8IFYSNWjGJjsdn3Qxao2XJqSdM8JVsiGDBlRqflbwI5v84vdp0/z03i0ZPU1VmTwgOueyT4I6lToPcaXwgOicS0YOTu3pAdE5l4wcfCjGA6JzLgHeZHbOuUB4k9k55wKvITrn3Dreh+iccwDyJrNzzgGhD9GbzM45B15DdM65IgLyvIbonHOB31RxzjnwJrNzzqXymyrOOUdoLnuT2TnnIq8hOucceB+ic86l8iazc84RZ8zOvfCTeyVyztUOXkN0zrnI+xCdcw5/yZRzzhXjTWbnnIuzf9XxJrNzzsV3qiRdiPV5QHTOJUDIm8zOORd4k9k55yKvITrnHCEYqo4HROecA7yG6Jxza+ViQMy9Xk3nXM0nUB2l/ZSbhTRM0mxJcySdU8r+rSX9R9K7kmZI+lV5eXpAdM4lQlLaTznH5gG3APsC3YGDJXUvkewC4BEz6w2MAm4tr0zeZHbOZZ1QRYfd9AfmmNk8AEkPASOBWSlpDGgSl5sCi8vL1AOicy4ZFetCbAssTFlfBAwokeYS4HlJJwGNgKHlZepNZudc9imjJnMrSdNSPscVz2E9VmL9YGCsmbUDfgXcK6Wfc8xriM65RGTQZF5qZn3L2LcIaJ+y3o71m8RHA8MAzOxNSQ2AVsDXZZapvBI551xlE+lrhxkMyZkKbCdpG0n1CTdNnimR5nNgCICkbkADYEm6TD0gOueSoXI+aZhZAXAiMAn4iHA3eaakyySNiMnOAI6V9D7wIDDazEo2q4vxJrNzLvtU8ckdzGwiMLHEtotSlmcBu2xInl5DrKH26r8t79/7Jz68/yTOPGT934mtt2jKxL8eztt3/YFJNx5J29aNi+1v3LA+cx87jRtO2TdbRa72XnphEgN696Dfjl256S/Xrrd/1apVHH3kIfTbsSt77zGQzxfMX7tv5oczGLbnruzSb0d2G9CLn376KYslT0YFm8xVosoCoiST9JeU9TMlXbIBx28h6VlJ70uaJWli+UdllO9kSWV11JaWfrSkmyvj3NlSp4648dRfMfL/7qf3kbdw4JCedO3Qqliaq/64F/dPmkH/Mbdx5T2vcNlxQ4rtv/joPXnt/QXZLHa1VlhYyNlnnMzDT4znjakzeOKxh5j98axiae4fdxfNmjVj6vsf84c/ncKlF50HQEFBAScccyTX33QLb0x9n6cnvkS9evWSuIysquiTKlWhKmuIq4ADJLUqN2XpLgNeMLMdzaw7sN6jOa50/bq1ZW7+N8z/4n/8XLCGR1+eyfBduxZL07VDayZPnwfAK+/OZ/gu6/b37rIVmzdvxItT52a13NXZ9Glvs02nbem4TSfq16/Pb357EM89O75YmucmjGfUIYcDMGL/3/La5JcxM/7z0gt077k9PbffEYAWLVuSl5d7L2CqTOXVDmtcDREoAG4HTiu5Q1IHSS/F5wtfkrR1KcdvRbi1DoCZzYjH3itpZEpe90saEWtyT0kaL+kzSSdKOj0+xzhFUouUvA+T9F9JH0rqH/NpEY+fEdPvUEnfQ9a1adWYRV9/t3Y9f8l3tG1VvEn8wdyv2H9QeNJp5G5dadJoE1o02RQJrv7j3pz3jxeyWubq7osvFtOmbbu1623atuWLL/KLp1m8mLbtwkiRunXr0qRpU75Ztoy5cz5BEgfu/yv22LUff7vh+qyWPSm1LSBCeNbwUElNS2y/GRhnZjsA9wN/K+PYO+PD2edLahO33wEcBRDzHci6jtWewCGEx3quAFbG5xjfBI5IybuRmQ0E/gjcFbddCrwby3QeMG4jrzlxpf0ylby1du6tz7Nbrw68ecdx7NarI/lff0dB4RqO378fk976lEVLvlsvD1e20m5elvw5lJWmoKCQt978L7fdMY4Jz7/CxPFP8erkl6usrLkiF5vMVXqX2cy+kzQOOBn4MWXXzsABcfleYL0eaDObJKkTYWDlvsC7knqa2SuSbpG0eczjcTMriL98/zGz74HvJS0HitosHwCpNb4H4zleldREUjNgV+C3cfvLklqWEsiLiSPnw+j5TdImzar8Jd/RbvMma9fbtm7C4qXfF0vzxbIVjLrwEQAabVqP/Qd147sfVjGgRzt22aEDx43sR6NN61O/Xh4rflzNhbe/lNVrqG7atGnL4vy1DRoW5+ez5ZZtiqdp25b8RQtp07YdBQUFfLd8Oc1btKBN27YM3GU3WrYKvUtD99mX9997l0GD98zqNWRbUrXAdLJxl/lGwojxRmnSlDo2yMy+MbMHzOxwwkDMQXHXvcChhJri3SmHrEpZXpOyvobiwb/k+YzMHgUqWb7bzayvmfVVvYbpkmbVtI/z6dyuJR22bEa9unU4cM8eTHhjdrE0LZtuuva1uGcduhv3PPcuAEdd/iRdfn8jXUfdxLn/eJ4HJr3vwTADvfv0Y97cOSyY/xmrV6/myccfZth+w4ulGfar4Tz0wL0APPPU4+y2+x5IYs8hezNz5gesXLmSgoIC/vv6q/yia7ckLiNrpHDzL90nCVUeEM3sG+ARQlAs8l/CyHIIge31ksdJ2lNSw7jcGNiWMPIcYCxwasx/5kYU66CY767AcjNbDrway4KkwYTHhqplu7Gw0DjtxomMv/4w3hv3Jx7/zyw+mr+EC8cMZr+BXQAY1KsjM+47kRn3ncjmzRtxzb2vJVvoaq5u3bpcff1NHLj/fgzsuz0jDziQrt16cNXll/DchNBQOfSIMXz7zTf027Er/7j5Ri689AoAmjVvzgknnspeu+/M4IF92WHH3uw9rNyp+6q53LyponIGbm98xtIKM9ssLm8BfAZca2aXSOpI6LtrRXiU5igz+7zE8WcRaoAFhMB9t5mlDuP5N/CUmd0W10cDfc3sxLg+P64vTd0naTKhT3F3wtRAY8zs7XjT5W5gG2AlcJyZzSiZb1nqNG5jm+x0/MZ8VW4jLJpwftJFqFWGDBrAe9PfqbQo1WDLLrb1EaXdOljn0+v2fSfNs8xVosr6EIuCYVz+CmiYsj4fSNtBYmbXAdeVti/WHLcj9gXG9GMJNcei9Y6l7TOzwWWc7xvCfGoltxfL1zlXCWKTOddUuydVJA0FPgb+Hpu6zrlqRuRmH2K1e5bZzF4EShu36JyrRnLwJnP1C4jOuRogR5vMHhCdc1kncnMcogdE51wCkusnTMcDonMuEV5DdM45iC+ZSroQ6/OA6JzLuqJhN7nGA6JzLhHeZHbOOfBhN845VyQMu0m6FOvzgOicS0ByM9qk4wHROZcIbzI75xz4sBvnnCsSht3k3mRbHhCdc4nwGqJzzkXV6qaKpCZl7YPwRr3KL45zrjaQqt/kDjNZ/210ReuGT9LqnKuAHKwglh0Qzax9NgvinKtd6lQwIkoaBtwE5AF3mNnVpaT5PXAJoRL3vpkdki7PjPoQJY0COpnZlZLaAVuY2TsbWH7nnAPWvZd5449XHnALsBewCJgq6Rkzm5WSZjvgXGAXM/tW0ubl5VvufW9JNwN7AIfHTSuB2zb8Epxzbp06Sv8pR39gjpnNM7PVwEOs/9bMY4FbzOxbADP7utwyZVDugWZ2PPBTzPQboH4GxznnXJkyeOteK0nTUj7HpRzeFliYsr4obkvVBegi6Q1JU2ITO61Mmsw/S6pDaIMjqSWwJoPjnHOuVAJEudXApWleVF/awVZivS7h/e2DgXbAa5J6mtn/yjphJjXEW4DHgdaSLgVeB67J4DjnnCtTBZvMi4DUG7/tgMWlpHnazH42s8+A2YQAWaZya4hmNk7SO8DQuOlAM/uw3OI651xZKj4OcSqwnaRtgHxgFFDyDvJTwMHAWEmtCE3oeekyzfRJlTzgZ0KVNPceQHTOVSuiYsNuzKxA0onAJEJ8usvMZkq6DJhmZs/EfXtLmgUUAmeZ2bJ0+ZYbECWdT4i8T8breEDS/WZ21UZfjXOu1qvokypmNhGYWGLbRSnLBpwePxnJpIZ4GNDHzFYCSLoCeAfwgOic2yiqxtN/LSiRri7ltMOdc648FX1SpSqkm9zhBkKf4UpgpqRJcX1vwp1m55zbaNUqIAJFd5JnAhNStk+puuI452qDcFMl6VKsL93kDndmsyDOuVqkGk7/BYCkbYErgO5Ag6LtZtalCsvlnKvhcnGC2EzGFI4F7ibUcvcFHiE8SO2ccxulqMlcgSdVqkQmAbGhmU0CMLO5ZnYBYfYb55zbaHWktJ8kZDLsZpVC3XaupD8QHpMpd14x55wri1T97jIXOQ3YDDiZ0JfYFBhTlYVyztV8ORgPM5rc4a24+D3rJol1zrkKqVZ3mSU9yfrzi61lZgdUSYmcczWeSK6fMJ10NcSbs1aKGqB3lza88dLFSRej1mje78Ski1CrrJq9sPxEG6KC71SpKukGZr+UzYI452qXXJxHMNP5EJ1zrtKI3ByY7QHROZeIujlYRcw4IEraxMxWVWVhnHO1Q5gPMfdqiJm8l7m/pA+AT+P6jpL+XuUlc87VaHl10n+SkMlp/wYMB5YBmNn7+KN7zrkKKHqnSnV8dK+OmS0oUb0trKLyOOdqiRzsQswoIC6U1B8wSXnAScAnVVss51xNJom86jQOMcUJhGbz1sBXwItxm3PObbQcvKeS0bPMXxNeAu2cc5UmByuIGc2Y/S9KeabZzI6rkhI552o8QbVtMr+YstwA+A1QyQ82OudqlQRnxU4nkybzw6nrku4FXqiyEjnnajwBeTnYibgxj+5tA3So7II452qXallDlPQt6/oQ6wDfAOdUZaGcczVftXt0L75LZUegdfw0N7NOZvZINgrnnKuZpIo/uidpmKTZkuZIKrOSJul3kkxS3/LyTHtaMzPgSTMrjJ8yZ9B2zrkNUZFH9+JDIrcQXo3cHThYUvdS0jUmvA/qrZL7Si1TBmnelrRTJpk551wmwrCbCtUQ+wNzzGyema0mvCt+ZCnp/gxcC/yUSbnKPK2kov7FXQlBcbak6ZLelTQ9k8ydc650ok45H6CVpGkpn9Sxz20pPvxvUdy27gxSb6C9mT2baanS3VR5G9gJ2D/TzJxzLhNhxuxyky01s7L6/Uo7em2XnqQ6wA3A6A0pV7qAKAAzm7shGTrnXLkEdSs27mYR0D5lvR2wOGW9MdATmBzvZm8JPCNphJlNKyvTdAGxtaTTy9ppZn/NpNTOOVdShjXEdKYC20naBsgnzLdwSNFOM1sOtFp7PmkycGa6YAjpA2IesBmlV02dc65CKvIss5kVSDoRmESIVXeZ2UxJlwHTzOyZjck3XUD8wswu25hMnXMuHVHxCWLNbCIwscS2i8pIOziTPMvtQ3TOuUqXoy+ZShcQh2StFM65WqXaTe5gZt9ksyDOudol98Khv6jeOZeQHKwgekB0zmWfUPVqMjvnXFWqbjdVnHOuaojEXkafjgdE51zWVcY4xKrgAdE5lwhvMjvnXFQt36ninHOVLTSZcy8iekB0ziWg/NcEJMEDonMuETkYDz0gOueyz5vMzjlXRFAnB8fdeEB0ziVCXkN0zrnYZM69eOgB0TmXjFy8y5yDrXhXGZ6f9G926PELenTtzHXXXr3e/lWrVnHYIQfRo2tndhs4gAXz5wOwbNky9hm6B62abcapJ5+Y5VJXb3sN7Mb7T17Ih09fzJlH7bXe/q23as7E207i7YfPZdK/TqHt5s3W7rv85JFMe/Q8pj16Hr/be6dsFjsxKue/JOR0QJR0vqSZkmZIek/SgArmN1rSzRt4zIqKnDMJhYWFnHryn3h6/HO8O2MWjz70IB/NmlUszdi77qR5s+bM/HgOJ51yGuefdzYADRo04KJL/sxV11yfRNGrrTp1xI3n/J6RJ95K799ezoHD+tC105bF0lx12m+4f8Lb9D/oKq68/TkuO2kEAMN27UGvbu0ZMOpqBh1+PaceOZTGjRokcRlZUzT9V7pPEnI2IEraGRgO7GRmOwBDgYXJlqp6mPr222y7bWe26dSJ+vXrc+BBo3h2/NPF0jw7/mkOPfxIAA747e+Y/PJLmBmNGjVil113pUGDmv0HWdn69ezI3IVLmZ+/jJ8LCnl00nSGD96hWJqunbZi8luzAXhl6icMH7w9AN06bclr73xKYeEaVv60mg8+WcTeA7tl/RqySmEcYrpPEnI2IAJbAUvNbBWAmS0Fukl6siiBpL0kPRGXV0i6RtI7kl6U1F/SZEnzJI1Iybe9pH9Lmi3p4pS8Tpf0YfycmqVrrBKLF+fTrt26d3i3bduO/Pz89dO0D2nq1q1Lk6ZNWbZsWVbLWZO02bwpi776du16/lff0rZ102JpPvgkn/2H9AJg5J470mSzTWnRtBEzPslnn126s2mDerRs1ojd+3ah3ZbNs1r+JKicTxJy+abK88BFkj4BXgQeBl4GbpHU2syWAEcBd8f0jYDJZnZ2DJqXA3sB3YF7gKL3tPYHegIrgamSJgAW8xpA+Fm8JekVM3s3XQElHQccB9B+660r56orgZmtt63kzCKZpHGZK63Pq+Q3fO4NT3LD2Qdy2IgBvDF9DvlffUtBYSEvTfmYPj068J+xZ7D02xW8NeMzCgrWZKfgCal2L5lKmpmtkNQH2A3YgxAQzwHuBQ6TdDewM3BEPGQ18O+4/AGwysx+lvQB0DEl6xfMbBlArF3uSvjdfdLMfkjZvhuQNiCa2e3A7QB9+vRdP8IkpG3bdixatK53IT9/EW3atFk/zcKFtGvXjoKCAr5bvpwWLVpku6g1Rv7X/6PdFutqdW23aM7iJcuLpfliyXJGnXkHAI02rc/+Q3rx3YqfALj2zklce+ckAMZeOZo5C7/OUskTlHvxMKebzJhZoZlNNrOLgROB3xJqhIcBBwOPmllBTP6zrav2rAGKmtprKB74SwYuIyd/NBuvb79+zJnzKfM/+4zVq1fz6MMPsd/wEcXS7Dd8BPffew8ATzz+GLvvsafXECtg2swFdN66NR3atKRe3TwO3GcnJkyeUSxNy2aN1n7HZ43Zh3uengKEGzItmjYCoOd2bei5XRtefPPj7F5AAupIaT9JyNkaoqRfAGvM7NO4qRewwMwWS1oMXEBoEm+ovSS1AH4E9gfGEALoWElXE4Ljb4DDK3oNSalbty433HQzv95vHwoLCzly9Bi69+jBZZdcxE59+jL81yMYPeZoxow+nB5dO9O8eQvuvf+htcf/onNHvv/uO1avXs34Z57i2YnP06179wSvKPcVFq7htGseYfytfyKvjrjn6Sl8NO9LLjxhP6bP+pwJr3zAoL7bcdlJIzCD16fP4dSrHgGgXt08XrwrdFt/v+Inxpx/D4WFNbvJDLlZC1FpfUm5IDaX/w40AwqAOcBxZrZU0ijgVDP7ZUr6FWa2WVy+BFhhZten7pM0GvgVob+xM/CAmV0a05xOCI4Ad5jZjSXzTadPn772xlvTKuHKXSaa9/Mxktm0avYjrFn5daXFsG7b97Zxz0xOm6Z/p2bvmFnfyjpnJnK2hmhm7wADy9i9K/CvEuk3S1m+pLR9ZjYWGFvG+f4K/LWU7eUGQ+fchlGOvmQqp/sQSyPpHWAH4L6ky+Kc23gVHXYjaVgcPjdH0jml7D9d0qz4YMdLkjqUl2e1C4hm1sfMBhWNT3TOVUdCSv9Je7SUB9wC7EsYWnewpJId3e8CfeODHY8B15ZXqmoXEJ1zNUMFn1TpD8wxs3lmthp4CBiZmsDM/mNmK+PqFKBdeZl6QHTOZV15zeUMmsxtKf4o76K4rSxHA8+Vl2nO3lRxztVsGYx7bSUpdejG7fFhCCg9ZpY6ZEbSYUBfYPfyTugB0TmXiAyaxUvTDLtZBLRPWW8HLF7/HBoKnA/snsl9B28yO+cSUcEm81RgO0nbSKoPjGLdfAUhf6k38E9ghJll9Cyk1xCdc9mnik0mYmYFkk4EJgF5wF1mNlPSZcA0M3sGuA7YDHg0nutzMxtRZqZ4QHTOJUBUfM5DM5sITCyx7aKU5aEbmqcHROdcInLwQRUPiM65ZPhrSJ1zLvIaonPORR4QnXOOoqE1uRcRPSA657JPUCf34qEHROdcQjwgOucchAZz7kVED4jOuawT3mR2zrl1PCA651yQi+9U8YDonEtE7oVDD4jOuSRk9pqArPOA6JzLujDbTe5FRA+IzrlE5F449IDonEtIDlYQPSA655LhTWbnnItyLxx6QHTOJUDycYjOObdO7sVDD4jOuWTkYDz0gOicS4K8yeycc1A5ryGtCh4QnXOJ8IDonHORTxDrnHPgkzs451wR70N0zrkU3mR2zrnIa4jOORd5QHTOuSgXm8wys6TLUCNIWgIsSLocG6EVsDTpQtQi1fX77mBmrSsrM0n/JnwX6Sw1s2GVdc5MeECs5SRNM7O+SZejtvDvO7fVSboAzjmXKzwgOudc5AHR3Z50AWoZ/75zmPchOudc5DVE55yLPCA651zkAbGWktQw6TLURpL8by6H+Q+nFpLUDLhc0n5Jl6W2kLStpPZmtsaDYu7yH0zt1Aj4Ehguaa+kC1NLHADMldTBg2Lu8rvMtZSkrYGRwA5ZtQ9NAAAK60lEQVTAo2b2fMJFqvEkXQj8EdjZzOZLqmNma5Iul1vHA2ItIklmZin/bwgcC/TEg2JWSLoYOIEQFD/zoJhbfLabWiIlCP4KGCWpPnAHcBdwFHCApDwzey7RgtYgKd95d6AD8K6ZXSppBfCmpF96TTG3eECsJeIf5s7AlcD5wOaEgHgicCtwGvA7SVPNrDrOxpJz4nf+a+Ba4COgoaQJZvYXSXnAh5K2N7PPki2pK+IBsXbpBEw2swkAkhYC44BBwINAngfDyiOpCXAkcKiZTY/BcbCkYWZ2bdzfCfCAmCP8TlftsgTYSlL92Dx+EXgUaGxmi8ysOs7nmDOk9eaA/gnYDNgRwMzGA98AR8f1C8zspVKOcwnxgFgDSdpM0iZxeW9JZ0g6A3iZ0Cq4CeglaQ9gX8DvrFVQ/AfG4vLWkjqZ2WrgHqCDpN1j0peBHyQ1Kjq26DiXPL/LXMPEP7QHgPuAzwn9hLcBvwPyCX2GFwJNge2Aa8xsYjKlrRkktQaON7PLY+C7Ie4aDywDGgB7AAvj/88ws2cSKaxLywNiDSRpFHAMoW9qipndGbc/Bqw2s0PieivvM6w4SQOBI4AVQFfgVKCA8I9QA+B54EfC8KZPzOydojvQCRXZlcGbzDVI0dMPZvYQoVm8I9AnPqqHmf0OaCJp83jIskQKWsOY2X8JtXIIte7vzWw+oX+2P7C9mX1gZg+a2TvxGA+GOcgDYg0RaxxrJPWU1Dp24J8O9AD2ldRBUl/CH2w98D/Kikq9GWJmrwJ3A7OAsyRtHm9SvQh0lpTnN09ynzeZaxBJQwjDaN4A5gLXA9sDfwZWAfOA8TFYukoQh9IMBZab2UWSegLHAd0JN1ROAc41sxcSLKbLkAfEai7laYimhD++SYS+rN8SBl9fSqgV3gAcYWazvf+qckhqQ+gfvIYweUMLM9tdUmfgZsKwm+vM7I0Ei+k2gAfEGiA+jrcHMAA4yszmSuoFjCA8MnYu8LOZfZtgMWsUSb2BlkAXM7s1bnsaaGpmgyXtAPxoZp8mWU63YbwPsZqT1I/wKN40YFPCkBrM7D3gWcJQm9YeDCuuqA8w3lV+lFD7HiVpEICZjQR+ljTFzGZ4MKx+vIZYjUlqD9wIzDKzC+OEDc8D88xsTEzT2My+T7KcNUnsp90PuJNwl/6PhIrFv83s9ZimT9HdZFe9eA2xmorzGW5HuHkyRNJu8cmIvYDtJd0L4MGw8sRhTQMI4wzrmtmXwGNAIWG2oEEx6fSEiugqyANiNaOgKWEGFQPOA54BDpG0i5n9DAwEbkmwmDWOpG2BO83sSuCfwBOSGpnZDOAJwg2Ur8CHM1VnHhCriThdFBYsJzwTeyGh3/AxQk3x2FhT/NnMpiRX2pojZar/vLjeyMxOAP4DTJXUxMzeB640s9lJldNVDg+IOU5ScwAzK5S0naS+cSKB24EZQHszmwP8G/gY8JsnFSSpjaROcbUpgJl9QgiKt8T1Y4CpwHvxH6uVSZTVVS4PiDlMUldgoqR94h/dyYRBv09I2gbYEjgewMw+BP4e/+82UvzOXwB6S+oB3CPpbEkNgDOAbyR1AzCzI4ERZlboM17XDH6XOUdJ6kJ4PvY24F4zWyVpk/j/ywn/mPUmPJp3kJm96QOuK0ZSR8JQpRvM7M44gevWwCXAp4QXcuUBY+Pz4vh3XrN4QMxBcbzbP4HPzezyuK050Icw43VB/GPdHjgHeMHM/pZYgWsISUcBvczslPgz6AW0B1oDrwFDgNGAgFFmNi+psrqq4a8QyEHxUbw5wBJJWxCaytsAw4DnJd0Yb5q8Iel0YKyk+8zsmwSLXRPMA46RtA9wEOGG1Q7AFOCXhO6J5wljDzeP6V0N4n2IuWsR4fnYVwk1lfHALwhT0B+Rkm5LQsd/QbYLWANNJTyBcg3QhPDyrV0Jg9/rE8YezgXaEAZnuxrGa4g5yswekDQFaB4nFK1nZj9LehYYGTv5VxEC5Egz+y7RAtcAZrYSuFHSuNTadnwueVtgS0lfA42AhxMqpqtC3oeYg1TGe3ol7UaotZxT9OY8V3Uk1SM8+XMVcF7K2wrrmpnXyGsgryHmgJJ3KuNEr3lmVhj3Nwd+TXh38nlmNsHvblatGAz7EybZvaDoO4+7C5MrmatKXkNMWMp8hoMJk4oKuNvMVpYIitsDmNkHHgyzIwbFlmb2pX/ntYMHxBwgaS/C7NYPE4Jib2CAma1IDYrOuarlATEBkloBbeLEAEj6OzDdzO6O638j3FHez/uqnMseH3aTZXHOwqOB4yT1iZtXAi1Skl1AmNh1kywXz7lazQNilsU5CycDS4CD4vs3HgLOkzQyJtshfpokUkjnailvMmdJnN26m5k9H9c7AscSBvxeD3QG/kV4Y97OwNk+tMa57PKAmAWxmfwp4bnYfwFfEyYVbQDsRnjS5AZCjb0h0NDMZvmdTeeyywNilkjakTCz9bvAXYQXQ71NuKu8jNCPeIW/mMi55HgfYpbEWZVHALsT/iEaQHhGdg7QnPB8st9EcS5BXkPMMkn9CTOmnGtm/yh6TE/SNmb2WdLlc64280f3sszM3pY0lDAT9iZmdmPcNR98wlHnkuQ1xIRIGgC8SJjxeqEHQeeS5wExQfGNbT5tl3M5wm+qJOt7WPvKAOdcwryG6JxzkdcQnXMu8oDonHORB0TnnIs8ILqMSCqU9J6kDyU9KqlhBfIaHF+WhaQRks5Jk7aZpD9uxDkukXRmpttLpBkr6XcbcK6Okj7c0DK63OMB0WXqRzPrZWY9gdXAH1J3Ktjg3ycze8bMrk6TpBnhPcjOVTkPiG5jvAZ0jjWjjyTdCkwH2kvaW9KbkqbHmuRmAJKGSfpY0uuE900Tt4+WdHNc3kLSk5Lej5+BwNXAtrF2el1Md5akqZJmSLo0Ja/zJc2W9CJhxvG0JB0b83lf0uMlar1DJb0m6RNJw2P6PEnXpZz7+Ip+kS63eEB0G0RSXWBf4IO46RfAODPrDfxAmO17qJntBEwDTo/vkP4X4c2BuwFblpH934BXzGxHYCdgJnAOMDfWTs+StDewHeGNeL2APpIGxdnHRxHeR3MA0C+Dy3nCzPrF831EmMm8SEfCRBz7AbfFazgaWG5m/WL+x0raJoPzuGrCn2V2mdpU0ntx+TXgTqANsMDMpsTtvyRMZ/ZGHGteH3gT6Ap8VjS1maT7gONKOceehFl/iC/WWh5fwZpq7/h5N65vRgiQjYEn48vmkfRMBtfUU9LlhGb5ZsCklH2PxHdjfyppXryGvYEdUvoXm8Zzf5LBuVw14AHRZepHM+uVuiEGvR9SNwEvmNnBJdL1AirrCQABV5nZP0uc49SNOMdYYH8ze1/SaGBwyr6SeVk890lmlho4i2Y/dzWAN5ldZZoC7BLfE4OkhpK6AB8D20jaNqY7uIzjXwJOiMfmSWpCeLyxcUqaScCYlL7JtpI2B14FfiNpU0mNCc3z8jQGvojvXz60xL4DJdWJZe4EzI7nPiGmR1IXSY0yOI+rJryG6CqNmS2JNa0HJRVNdnuBmX0i6ThggqSlwOtAz1KyOAW4XdLRQCFwgpm9KemNOKzludiP2A14M9ZQVwCHmdl0SQ8D7wELCM368lwIvBXTf0DxwDsbeAXYAviDmf0k6Q5C3+L0+Pz5EmD/zL4dVx34s8zOORd5k9k55yIPiM45F3lAdM65yAOic85FHhCdcy7ygOicc5EHROeci/4feO06EgE6x9cAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_confusion_matrix(y_true, y_pred, np.asarray(['No Symbol', 'Symbol']), normalize=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Normalized confusion matrix\n",
      "[[0.99 0.01]\n",
      " [0.06 0.94]]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x227da178c88>"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAUQAAAEYCAYAAAAkpo9KAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAIABJREFUeJzt3XecFEX6x/HPd3dFJEcDQRBRSSJIOhEQFRUVQT09MSOm884cfmfOWe/UO/U8I+acQFCMmE4kKSgoSJRgAPRQRMFdnt8fXQuzw+7swO5Oz+4+b17zYrq7urp6ZvfZqu6qapkZzjnnICfuAjjnXLbwgOicc4EHROecCzwgOudc4AHROecCD4jOORd4QKzmJF0p6bHwfltJKyXllvMx5ksaUJ55pnHM0yR9F86ncRnyWSmpTXmWLS6SpkvqH3c5spkHxAoWgsF3kmonrDtJ0rgYi1UsM/vazOqYWUHcZSkLSZsB/wD2DeezfFPzCvvPLb/SlT9JIyRdW1o6M+toZuMyUKRKywNiZuQBZ5U1E0X8OyvdVkBNYHrcBckGkvLiLkNl4b9cmXELcL6kBsVtlNRb0kRJK8L/vRO2jZN0naQPgVVAm7DuWkn/DU26UZIaS3pc0k8hj9YJedwhaWHYNllS3xLK0VqSScqTtFvIu/D1m6T5IV2OpAslzZG0XNIzkhol5HOspAVh2yWpPhhJW0j6e0i/QtIHkrYI2waHZt7/wjm3T9hvvqTzJU0L+z0tqaakHYGZIdn/JL2deF5Jn+tJ4X1bSe+GfJZJejohnUlqG97Xl/SIpKWhvJcW/oGSNCyU/VZJP0qaJ2n/FOc9X9IFofy/SHpA0laSXpX0s6Q3JTVMSP+spG9DGd+T1DGsPwU4Gvi/wp+FhPz/Jmka8Ev4TtddupA0RtLfE/J/WtKDqb6rasHM/FWBL2A+MAB4Abg2rDsJGBfeNwJ+BI4lqkkeGZYbh+3jgK+BjmH7ZmHdbGB7oD4wA5gVjpMHPAI8lFCGY4DGYdt5wLdAzbDtSuCx8L41YEBe0jkUHvOGsHw2MB5oAWwO/Ad4MmzrAKwE+oVt/wDygQElfD53hbybA7lA77DfjsAvwD7h+P8XzrlGwuc6AWgWPsMvgD8Xdx7FnVc45knh/ZPAJUQVhJpAn4R0BrQN7x8BXgbqhjxnASeGbcOA34GTw3mcBiwBlOLnYjxRbbY58D0wBegazv9t4IqE9MPDcTcHbgc+Tdg2gvCzlZT/p0BLYIvEn8XwfutwzL2IAupcoG7cvy9xv2IvQFV/sT4gdgJWAE0pGhCPBSYk7fMRMCy8HwdcnbR9HHBJwvLfgVcTlg9K/IUppkw/AruE91dSekD8NzAayAnLXwB7J2zfJgSDPOBy4KmEbbWBNRQTEEMA+rWwLEnbLgOeSUq7GOif8Lkek7D9ZuCe4s6juPOiaEB8BLgXaFFMOQxoSxTkVgMdEradmvA9DgNmJ2yrFfbdOsXPxdEJy88D/05YPgN4qYR9G4S864flERQfEIcX97OYsHwosBBYRsIfger88iZzhpjZ58ArwIVJm5oBC5LWLSCqNRRaWEyW3yW8/7WY5TqFC5LOk/RFaG79j6hW2SSdcks6FegPHGVma8PqVsCLoSn7P6IAWUBU22mWWF4z+wUo6aZGE6Ia2ZxithX5XMKxF1L0c/k24f0qEs55I/0fIGBCaKIPL6GsNSj6XSV/T+vKY2arwttUZUrrO5SUK+nGcIniJ6LAVlimVIr7uUn0ClGgn2lmH5SStlrwgJhZVxA1qRJ/iZYQBZhE2xLVhgpt8pRE4Xrh34A/AQ3NrAFRTVVp7nsNMMTMViRsWgjsb2YNEl41zWwx8A1RM60wj1pEzfXiLAN+I2r6JyvyuUhSyHdxMWlL80v4v1bCuq0L35jZt2Z2spk1I6r13V143TCprL9T9LtK/p4qylHAEKKWRn2iGi+s/w5L+vko7efmOqI/ZttIOrKMZawSPCBmkJnNBp4GzkxYPQbYUdJR4cL3EUTX4V4pp8PWJbqGtxTIk3Q5UK+0nSS1DGU9zsxmJW2+B7hOUquQtqmkIWHbc8AgSX0k1QCupoSfs1DrexD4h6RmoSa0m6TNgWeAAyXtragbzXlETdb/btTZR8dZShS4jgnHGE5CEJZ0uKQWYfFHokBSkJRHQSjTdZLqhnM/F3hsY8uzCeoSnftyoqB+fdL274CN6ispqR9wAnBceP1LUvPUe1V9HhAz72qi62oAWNRHbhDRL/xyoubbIDNbVk7HGwu8SnQDYAFRjay0phTA3kS1qOe0/k5zYTeWO4CRwOuSfia6OdArnM904K/AE0S1xR+BRSmOcz7wGTAR+AG4ieha5Uyim0H/IqqdHQQcZGZr0jzvZCcDFxB9xh0pGlh7AB9LWhnO6ywzm1dMHmcQ1TbnAh+Ec8zEndlHiL67xUQ30MYnbX8A6BAuYbxUWmaS6oU8TzezxaG5/ADwUKiJV1sKF1edc67a8xqic84FHhCdcy7wgOicc4EHROecC3zQdzlR3hamGnXjLka10bX9tnEXoVpZsGA+y5YtK7c70Ln1Wpnl/5oyjf26dKyZDSyvY6bDA2I5UY26bL7Tn+IuRrXx4cd3xl2EamX3Xt3LNT/L/7XU35ffPr0rrdFU5ckDonMu8yTIKdd5iMuFB0TnXDyycGpPD4jOuXhk4aAYD4jOuRh4k9k55yLCm8zOORfxGqJzzq3n1xCdcw5A3mR2zjkguoboTWbnnAOvITrnXCEBuV5DdM65iN9Ucc458Cazc84l8psqzjlH1Fz2JrNzzgVeQ3TOOfBriM45l8ibzM45R5gxO/vCT/aVyDlXPXgN0TnnAr+G6Jxz+EOmnHOuCG8yO+dcmP0rx5vMzjkXnqkSdyE25AHRORcDIW8yO+dcxJvMzjkXeA3ROeeIgqFyPCA65xzgNUTnnFsnGwNi9l3VdM5VfQLlKOWr1CykgZJmSpot6cJitm8r6R1Jn0iaJumA0vL0gOici4WklK9S9s0F7gL2BzoAR0rqkJTsUuAZM+sKDAXuLq1M3mR2zmWcUFm73fQEZpvZXABJTwFDgBkJaQyoF97XB5aUlqkHROdcPMp2CbE5sDBheRHQKynNlcDrks4AagMDSsvUm8zOucxTWk3mJpImJbxOKZrDBixp+UhghJm1AA4AHpVSzznmNUTnXCzSaDIvM7PuJWxbBLRMWG7Bhk3iE4GBAGb2kaSaQBPg+xLLVFqJnHOuvInUtcM0uuRMBHaQtJ2kGkQ3TUYmpfka2BtAUnugJrA0VaYeEJ1z8VAprxTMLB84HRgLfEF0N3m6pKslDQ7JzgNOljQVeBIYZmbJzeoivMnsnMs8lX1yBzMbA4xJWnd5wvsZwO4bk6fXEKuofXq3Z+qLl/H5y1dw/gn7bLB9220aMuaeM5jw9EWMve8smm/ZYN22a88cwqRnL2bSsxdz2L67ZrLYldrrY1+jc8ed6NiuLbfcfOMG21evXs0xRx1Bx3Zt6du7Fwvmzwdg+fLl7DdgT5o0qMPZZ56e4VLHp4xN5gpRYQFRkkn6e8Ly+ZKu3Ij9t5L0iqSpkmZIGlP6XmnlO05SSRdqi0s/TNKd5XHsTMnJEbdf+CeGnH43Xf94LYcP7Ea7NlsXSXPDOYfw+OgJ9DziBq6/91WuPiNqZQzs05Eu7VvSa+iN9Dv2Vs4+fgB1a9eM4zQqlYKCAs4+86+8POpVPpk2g2efepIvZswokmbEgw/QsEFDpn85mzPOOodLLv4bADVr1uTyK6/hhptujaPosSnrSJWKUJE1xNXAoZKabOL+VwNvmNkuZtYB2GBojitej06tmbNwGfMXL+f3/AKeHTuFQf07F0nTrs02jPt4JgDvTpzFoP47A9C+zda8P/krCgrWsuq3NXw2axH79m6f8XOobCZOmMD227dluzZtqFGjBocfMZRXRr1cJM0ro17m6GOPB+DQPx7GuLffwsyoXbs2u/fpQ82a1ecPT2m1wypXQwTygXuBc5I3SGol6a0wvvAtSdsWs/82RLfWATCzaWHfRyUNScjrcUmDQ03uJUmjJM2TdLqkc8M4xvGSGiXkfYyk/0r6XFLPkE+jsP+0kL5oBKlEmm1Zn0Xf/bhuefF3P9K8af0iaT6btZiD9+4CwJC9dqFenS1oVL8202YtZr/dO7BFzc1o3KA2e3TfkRZbN8xo+SujJUsW06LF+l4gzZu3YPHixRumaRmlycvLo179+ixfvjyj5cwm1S0gQjTW8GhJ9ZPW3wk8YmadgceBf5aw7wNhcPYlkpqF9fcDJwCEfHuz/sJqJ+AoomE91wGrwjjGj4DjEvKubWa9gb8AD4Z1VwGfhDJdDDyyieccOxVziy751tpFt71I325t+ejJv9G3W1sWf/cj+QUFvDX+S177YAbvjDiPh284gY+nzSM/f21mCl6JFXfzMvmXOp001Uk2Npkr9C6zmf0k6RHgTODXhE27AYeG948CNxez71hJbYg6Vu4PfCKpk5m9K+kuSVuGPJ43s/zwg/WOmf0M/CxpBTAqZPcZkFjjezIc4z1J9SQ1APoAfwzr35bUuJhAXkToOR/1nt+sTnofSgYs/v5/tNhqfa2u+VYNWbJ0RZE03yxdwdDz7weg9hY1OHjvLvy08jcAbn5gLDc/MBaAEdcPY/bCEvuxuqB58xYsWrR+JNnixYto1qzZhmkWLqRFixbk5+fz04oVNGrUKDmraiMb/xhk4i7z7UQ9xmunSFNs3yAz+8HMnjCzY4k6YvYLmx4FjiaqKT6UsMvqhPdrE5bXUjT4Jx/PSG8oUHL57jWz7mbWXXlbpEqaUZOmL6Dttk1p1awxm+Xlcvh+uzJ63LQiaRo3qL3uB/KC4fvx8MvjgeiGTKP60VfVaYdmdNqhGW9+9GVmT6AS6t6jB7Nnf8X8efNYs2YNzz79FAcOGlwkzYGDBvP4ow8D8MLzz7HHnntlZVDIBCn6WUv1ikOF90M0sx8kPUMUFAubp/8l6lleGNg+SN5P0l7AeDNbJakusD1Rz3OAEcAE4Fszm74JxToCeEdSH2CFma2Q9F4oyzWS+hMNG/qpMv7AFhSs5ZybnmHU3X8lN0c8/PJ4vpj7LZeddiBTZnzN6Hc/o1/3Hbj6jMGYwQdTZnP2Dc8AsFleLm8+eDYAP6/8jeGXPExBgTeZS5OXl8dtd9zJQQfuR0FBAccPG06Hjh25+srL2bVbdwYdNJhhw09k+LBj6diuLQ0bNuLRx59at/9ObVvz808/sWbNGkaNfIlXxrxO+w7Js1lVJdn51D2V0nF70zOWVppZnfB+K2AecLOZXSmpNVFwbEI0lOYEM/s6af8LiGqA+UQ12YfMLLEbz2vAS2Z2T1geBnQ3s9PD8vywvCxxm6RxRNcU9yCaGmi4mU0IN10eArYDVgGnmNm05HxLklNrS9t8pz9tykflNsGPEytVT6hKb/de3Zk8eVK5RbCaW+9o2x5X3K2D9b66Zf/JKcYyV4gKqyEWBsPw/jugVsLyfGCvUva/BbiluG2SagE7EK4FhvQjiGqOhcuti9tmZv1LON4PRPOpJa8vkq9zrhyEJnO2qXQjVSQNAL4E/mVmK0pL75zLPqKaXkMsb2b2JlBcv0XnXCWShZcQK19AdM5VAVnaZPaA6JzLOJGd/RA9IDrnYhDfdcJUPCA652LhNUTnnIPwkKm4C7EhD4jOuYwr7HaTbTwgOudi4U1m55wD73bjnHOFom43cZdiQx4QnXMxyM7ZbjwgOudi4U1m55wD73bjnHOFom432TfZlgdE51wsvIbonHNBpbqpIqleqh3N7KfyL45zrjqQKt/kDtPZ8Gl0hcuGT9LqnCuDLKwglhwQzaxlJgvinKtecsoYESUNBO4AcoH7zezGYtL8CbiSqBI31cyOSpVnWtcQJQ0F2pjZ9ZJaAFuZ2eSNLL9zzgHrn8u86fsrF7gL2AdYBEyUNNLMZiSk2QG4CNjdzH6UtGVp+ZZ631vSncCewLFh1Srgno0/BeecWy9HqV+l6AnMNrO5ZrYGeIoNn5p5MnCXmf0IYGbfl1qmNMrd28xOBX4Lmf4A1EhjP+ecK1EaT91rImlSwuuUhN2bAwsTlheFdYl2BHaU9KGk8aGJnVI6TebfJeUQtcGR1BhYm8Z+zjlXLAGi1GrgshQPqi9uZ0taziN6fnt/oAXwvqROZva/kg6YTg3xLuB5oKmkq4APgJvS2M8550pUxibzIiDxxm8LYEkxaV42s9/NbB4wkyhAlqjUGqKZPSJpMjAgrDrczD4vtbjOOVeSsvdDnAjsIGk7YDEwFEi+g/wScCQwQlIToib03FSZpjtSJRf4nahKmn0DEJ1zlYooW7cbM8uXdDowlig+PWhm0yVdDUwys5Fh276SZgAFwAVmtjxVvqUGREmXEEXeF8N5PCHpcTO7YZPPxjlX7ZV1pIqZjQHGJK27POG9AeeGV1rSqSEeA3Qzs1UAkq4DJgMeEJ1zm0SVePqvBUnp8iilHe6cc6Up60iVipBqcofbiK4ZrgKmSxoblvclutPsnHObrFIFRKDwTvJ0YHTC+vEVVxznXHUQ3VSJuxQbSjW5wwOZLIhzrhqphNN/ASBpe+A6oANQs3C9me1YgeVyzlVx2ThBbDp9CkcADxHVcvcHniEaSO2cc5uksMlchpEqFSKdgFjLzMYCmNkcM7uUaPYb55zbZDlSylcc0ul2s1pR3XaOpD8TDZMpdV4x55wriVT57jIXOgeoA5xJdC2xPjC8IgvlnKv6sjAepjW5w8fh7c+snyTWOefKpFLdZZb0IhvOL7aOmR1aISVyzlV5Ir7rhKmkqiHembFSVAGd27Xkrfdvj7sY1UbDva6IuwjVyupZyVMNllEZn6lSUVJ1zH4rkwVxzlUv2TiPYLrzITrnXLkR2dkx2wOicy4WeVlYRUw7IEra3MxWV2RhnHPVQzQfYvbVENN5LnNPSZ8BX4XlXST9q8JL5pyr0nJzUr/ikM5h/wkMApYDmNlUfOiec64MCp+pUhmH7uWY2YKk6m1BBZXHOVdNZOElxLQC4kJJPQGTlAucAcyq2GI556oySeRWpn6ICU4jajZvC3wHvBnWOefcJsvCeyppjWX+nugh0M45V26ysIKY1ozZ91HMmGYzO6VCSuScq/IElbbJ/GbC+5rAIcDCiimOc65aiHFW7FTSaTI/nbgs6VHgjQorkXOuyhOQm4UXETdl6N52QKvyLohzrnqplDVEST+y/hpiDvADcGFFFso5V/VVuqF74VkquwBNw6uhmbUxs2cyUTjnXNUklX3onqSBkmZKmi2pxEqapMMkmaTupeWZ8rBmZsCLZlYQXiXOoO2ccxujLEP3wiCRu4gejdwBOFJSh2LS1SV6HtTHyduKLVMaaSZI2jWdzJxzLh1Rt5sy1RB7ArPNbK6ZrSF6VvyQYtJdA9wM/JZOuUo8rKTC64t9iILiTElTJH0iaUo6mTvnXPFETikvoImkSQmvxL7PzSna/W9RWLf+CFJXoKWZvZJuqVLdVJkA7AocnG5mzjmXjmjG7FKTLTOzkq77Fbf3ukt6knKA24BhG1OuVAFRAGY2Z2MydM65UgnyytbvZhHQMmG5BZD4JKy6QCdgXLibvTUwUtJgM5tUUqapAmJTSeeWtNHM/pFOqZ1zLlmaNcRUJgI7SNoOWEw038JRhRvNbAXQZN3xpHHA+amCIaQOiLlAHYqvmjrnXJmUZSyzmeVLOh0YSxSrHjSz6ZKuBiaZ2chNyTdVQPzGzK7elEydcy4VUfYJYs1sDDAmad3lJaTtn06epV5DdM65cpelD5lKFRD3zlgpnHPVSqWb3MHMfshkQZxz1Uv2hUN/UL1zLiZZWEH0gOicyzyhytVkds65ilTZbqo451zFELE9jD4VD4jOuYwrj36IFcEDonMuFt5kds65oFI+U8U558pb1GTOvojoAdE5F4PSHxMQBw+IzrlYZGE89IDonMs8bzI751whQU4W9rvxgOici4W8huicc6HJnH3x0AOicy4e2XiXOQtb8a48vPXGWHp17UiPzu244+83b7B99erVnHjcUfTo3I59+/fm6wXz122b/vk0Bu7Vh92770Lfnl347be0nvFd7e3Tsy1THzuDz584k/OP7rPB9m23qs+Y245nwkOnMfaOYTRvWq/I9rq1NmfO8+dx29kHZKrIsVIp/+KQ1QFR0iWSpkuaJulTSb3KmN8wSXdu5D4ry3LMOBQUFPC3c8/k6RdG8eGkabzw7FPM/GJGkTSPP/wgDRo0YOK0L/nzX8/iqssuBiA/P5/TTjyeW++4iw8nTeXlV99is802i+M0KpWcHHH7OQcy5ILH6HrcXRy+9860a9W0SJob/rIfj4/9lJ4n/JvrH36Xq08ZUGT7FSftxfufzs9gqeNTOP1XqlccsjYgStoNGATsamadgQHAwnhLVTlMmTSB7dpsT+vt2lCjRg0OOewIXh09qkiaV0ePYujRxwIw+JA/8v64tzEz3nnrDTp02plOO+8CQKPGjcnNzc34OVQ2Pdo3Z87iH5j/zY/8nl/As299zqA+7Yqkade6KeMmzwPg3SnzGNRnp3Xbuu64DVs2rM2bE6vJY9AV9UNM9YpD1gZEYBtgmZmtBjCzZUB7SS8WJpC0j6QXwvuVkm6SNFnSm5J6Shonaa6kwQn5tpT0mqSZkq5IyOtcSZ+H19kZOscK8c2SJTRr0WLdcrPmzflmyeIN0jRvET3nOy8vj3r16/PD8uXMmT0LSRw+5AD23L0H/7zt1oyWvbJq1qQei75fsW558dIVNG9at0iaz2Z/y8F7dABgSL/21Ktdk0b1tkASN/51Py7+9+sZLXPcVMorDtkcEF8nCl6zJN0taQ/gbaKgWNgWOQF4KLyvDYwzs27Az8C1wD7AIUDi41R7AkcDXYDDJXWX1C3k1Qv4A3CypK6lFVDSKZImSZq0fNmysp5vuTGzDdYlzyxSUpr8/AI+/ui/3PPAI4x+413GjHqJ9955u8LKWlUUV6NJ/ogvuvt1+nZpxUf3/5m+XVqz+PsV5Bes5dRDejB2/Fcs+v6nzBQ2CxQ+ZCrbmsxZe5fZzFaGQNUX2BN4GrgQeBQ4RtJDwG7AcWGXNcBr4f1nwGoz+13SZ0DrhKzfMLPlAKF22Qcw4EUz+yVhfV/gk1LKeC9wL0CXXbttGGFi0qx5c5YsWrRuecnixWy9TbMN0ixetJBmzVuQn5/PTytW0LBRI5o1a07vPn1p3KQJAAP23Z+pUz+h3557ZfQcKpvFS3+ixZb11y03b1qfJct+LpLmm+U/M/TSpwGovUUNDu7Xnp9+WU2vji3ZvfO2nHJwD2pvUYMam+Wy8tc1XPafNzN6DhmXfTeZs7qGiJkVmNk4M7sCOB34I1GN8BjgSOBZM8sPyX+39dWetUBhU3stRQN/cuAysvKr2XRdu/Vg7pzZLJg/jzVr1vDic08z8IBBRdIMPGAQTz3+KAAjX3yevnvsiST2GrAv0z//jFWrVpGfn89/P3iPndq1j+M0KpVJXy6hbYtGtNqmAZvl5XL43p0Y/eGXRdI0rl9rXU39gqP78vCY6O/tCdc8z46H30a7I27nortf54mxU6t+MCTqdpPqFYesrSFK2glYa2ZfhVVdgAVmtkTSEuBSoibxxtpHUiPgV+BgYDhRAB0h6Uai4HgIcGxZzyEueXl53Pj3Ozj84ANZW1DAUccOo12HjtxwzZV02bUb+x94EEcfP5y/nDSMHp3b0aBhQ+4b8TgADRo25LQzzmaffrshiQH7DWTfgdWjG0hZFBSs5ZzbxzDq1mPJzcnh4TGf8MX8pVw2fE+mzFzC6A9n0q9La64+dQBmxgdTF3D2baPjLnassrEWouKuJWWD0Fz+F9AAyAdmA6eY2TJJQ4GzzewPCelXmlmd8P5KYKWZ3Zq4TdIw4ACi641tgSfM7KqQ5lyi4Ahwv5ndnpxvKl127WZvvf9xOZy5S0eL/a+JuwjVyupP7mPtz0vKLYa137mrPTJyXMo0Pds0mGxm3cvrmOnI2hqimU0GepewuQ9wX1L6Ognvryxum5mNAEaUcLx/AP8oZn2pwdA5t3GUpQ+ZyupriMWRNBnoDDwWd1mcc5uurN1uJA0M3edmS7qwmO3nSpoRBna8JalVaXlWuoBoZt3MrF9h/0TnXGUkpNSvlHtLucBdwP5AB+BISR2Skn0CdA8DO54DNhzDmqTSBUTnXNVQxpEqPYHZZjbXzNYATwFDEhOY2TtmtiosjgdaUAoPiM65jCutuZxGk7k5RYfyLgrrSnIi8GppmWbtTRXnXNWWxnOZm0ialLB8bxgMAcXHzGK7zEg6BugO7FHaAT0gOudikUazeFmKbjeLgJYJyy2AJRseQwOAS4A90rnv4E1m51wsythkngjsIGk7STWAocDIIvlH8xH8BxhsZt+nUyavITrnMk9pNZlLZGb5kk4HxgK5wINmNl3S1cAkMxsJ3ALUAZ4Nx/razAaXmCkeEJ1zMRBln/PQzMYAY5LWXZ7wfsAGO5XCA6JzLhZZOFDFA6JzLh7+GFLnnAu8huicc4EHROeco7BrTfZFRA+IzrnME+RkXzz0gOici4kHROecg6jBnH0R0QOicy7jhDeZnXNuPQ+IzjkXycZnqnhAdM7FIvvCoQdE51wc0ntMQMZ5QHTOZVw02032RUQPiM65WGRfOPSA6JyLSRZWED0gOufi4U1m55wLsi8cekB0zsVA8n6Izjm3XvbFQw+Izrl4ZGE89IDonIuDvMnsnHNQPo8hrQgeEJ1zsfCA6JxzgU8Q65xz4JM7OOdcIb+G6JxzCbzJ7JxzgdcQnXMu8IDonHNBNjaZZWZxl6FKkLQUWBB3OTZBE2BZ3IWoRirr593KzJqWV2aSXiP6LFJZZmYDy+uY6fCAWM1JmmRm3eMuR3Xhn3d2y4m7AM45ly08IDrnXOAB0d0bdwGqGf+8s5hfQ3TOucBriM45F3hAdM65wANiNSWpVtxlqI4k+e9cFvMvpxqS1AC4VtKBcZelupC0vaSWZrbWg2L28i+meqoNfAsMkrRP3IWpJg4F5khq5UExe/ld5mpK0rbAEKAz8KyZvR5zkarfPO7OAAAK6UlEQVQ8SZcBfwF2M7P5knLMbG3c5XLreUCsRiTJzCzh/1rAyUAnPChmhKQrgNOIguI8D4rZxWe7qSYSguABwFBJNYD7gQeBE4BDJeWa2auxFrQKSfjMOwCtgE/M7CpJK4GPJP3Ba4rZxQNiNRF+MXcDrgcuAbYkCoinA3cD5wCHSZpoZpVxNpasEz7zg4CbgS+AWpJGm9nfJeUCn0va2czmxVtSV8gDYvXSBhhnZqMBJC0EHgH6AU8CuR4My4+kesDxwNFmNiUEx/6SBprZzWF7G8ADYpbwO13Vy1JgG0k1QvP4TeBZoK6ZLTKzyjifY9aQNpgD+jegDrALgJmNAn4ATgzLl5rZW8Xs52LiAbEKklRH0ubh/b6SzpN0HvA2UavgDqCLpD2B/QG/s1ZG4Q+MhffbSmpjZmuAh4FWkvYISd8GfpFUu3Dfwv1c/PwucxUTftGeAB4Dvia6TngPcBiwmOia4WVAfWAH4CYzGxNPaasGSU2BU83s2hD4bgubRgHLgZrAnsDC8P95ZjYylsK6lDwgVkGShgInEV2bGm9mD4T1zwFrzOyosNzErxmWnaTewHHASqAdcDaQT/RHqCbwOvArUfemWWY2ufAOdExFdiXwJnMVUjj6wcyeImoW7wJ0C0P1MLPDgHqStgy7LI+loFWMmf2XqFYOUa37ZzObT3R9tiews5l9ZmZPmtnksI8HwyzkAbGKCDWOtZI6SWoaLuCfC3QE9pfUSlJ3ol/YzcB/Kcsq8WaImb0HPATMAC6QtGW4SfUm0FZSrt88yX7eZK5CJO1N1I3mQ2AOcCuwM3ANsBqYC4wKwdKVg9CVZgCwwswul9QJOAXoQHRD5SzgIjN7I8ZiujR5QKzkEkZD1Cf65RtLdC3rj0Sdr68iqhXeBhxnZjP9+lX5kNSM6PrgTUSTNzQysz0ktQXuJOp2c4uZfRhjMd1G8IBYBYTheHsCvYATzGyOpC7AYKIhYxcBv5vZjzEWs0qR1BVoDOxoZneHdS8D9c2sv6TOwK9m9lWc5XQbx68hVnKSehANxZsEbEHUpQYz+xR4hairTVMPhmVXeA0w3FV+lqj2PVRSPwAzGwL8Lmm8mU3zYFj5eA2xEpPUErgdmGFml4UJG14H5prZ8JCmrpn9HGc5q5JwnfZA4AGiu/R/IapYvGZmH4Q03QrvJrvKxWuIlVSYz3AHopsne0vqG0ZG7APsLOlRAA+G5Sd0a+pF1M8wz8y+BZ4DCohmC+oXkk6JqYiujDwgVjKK1CeaQcWAi4GRwFGSdjez34HewF0xFrPKkbQ98ICZXQ/8B3hBUm0zmwa8QHQD5Tvw7kyVmQfESiJMF4VFVhCNib2M6Lrhc0Q1xZNDTfF3MxsfX2mrjoSp/nPDcm0zOw14B5goqZ6ZTQWuN7OZcZXTlQ8PiFlOUkMAMyuQtIOk7mEigXuBaUBLM5sNvAZ8CfjNkzKS1ExSm7BYH8DMZhEFxbvC8knARODT8MdqVRxldeXLA2IWk9QOGCNpv/BLdyZRp98XJG0HbA2cCmBmnwP/Cv+7TRQ+8zeArpI6Ag9L+pukmsB5wA+S2gOY2fHAYDMr8Bmvqwa/y5ylJO1IND72HuBRM1stafPw/7VEf8y6Eg3NO8LMPvIO12UjqTVRV6XbzOyBMIHrtsCVwFdED+TKBUaE8eL4Z161eEDMQqG/23+Ar83s2rCuIdCNaMbr/PDLujNwIfCGmf0ztgJXEZJOALqY2VnhO+gCtASaAu8DewPDAAFDzWxuXGV1FcMfIZCFwlC82cBSSVsRNZW3AwYCr0u6Pdw0+VDSucAISY+Z2Q8xFrsqmAucJGk/4AiiG1adgfHAH4guT7xO1Pdwy5DeVSF+DTF7LSIaH/seUU1lFLAT0RT0xyWk25rown9+pgtYBU0kGoFyE1CP6OFbfYg6v9cg6ns4B2hG1DnbVTFeQ8xSZvaEpPFAwzCh6GZm9rukV4Ah4SL/aqIAOcTMfoq1wFWAma0Cbpf0SGJtO4xL3h7YWtL3QG3g6ZiK6SqQX0PMQirhOb2S+hLVWi4sfHKeqziSNiMa+XMDcHHC0wrzzMxr5FWQ1xCzQPKdyjDRa66ZFYTtDYGDiJ6dfLGZjfa7mxUrBMOeRJPsXlr4mYfNBfGVzFUkryHGLGE+w/5Ek4oKeMjMViUFxZ0BzOwzD4aZEYJiYzP71j/z6sEDYhaQtA/R7NZPEwXFrkAvM1uZGBSdcxXLA2IMJDUBmoWJAZD0L2CKmT0Ulv9JdEf5QL9W5VzmeLebDAtzFp4InCKpW1i9CmiUkOxSooldN89w8Zyr1jwgZliYs3AcsBQ4Ijx/4yngYklDQrLO4VUvlkI6V015kzlDwuzW7c3s9bDcGjiZqMPvrUBb4D6iJ+btBvzNu9Y4l1keEDMgNJO/IhoXex/wPdGkojWBvkQjTW4jqrHXAmqZ2Qy/s+lcZnlAzBBJuxDNbP0J8CDRg6EmEN1VXk50HfE6fzCRc/Hxa4gZEmZVHgzsQfSHqBfRGNnZQEOi8cl+E8W5GHkNMcMk9SSaMeUiM/t34TA9SduZ2by4y+dcdeZD9zLMzCZIGkA0E/bmZnZ72DQffMJR5+LkNcSYSOoFvEk04/VCD4LOxc8DYozCE9t82i7nsoTfVInXz7DukQHOuZh5DdE55wKvITrnXOAB0TnnAg+IzjkXeEB0aZFUIOlTSZ9LelZSrTLk1T88LAtJgyVdmCJtA0l/2YRjXCnp/HTXJ6UZIemwjThWa0mfb2wZXfbxgOjS9auZdTGzTsAa4M+JGxXZ6J8nMxtpZjemSNKA6DnIzlU4D4huU7wPtA01oy8k3Q1MAVpK2lfSR5KmhJpkHQBJAyV9KekDoudNE9YPk3RneL+VpBclTQ2v3sCNwPahdnpLSHeBpImSpkm6KiGvSyTNlPQm0YzjKUk6OeQzVdLzSbXeAZLelzRL0qCQPlfSLQnHPrWsH6TLLh4Q3UaRlAfsD3wWVu0EPGJmXYFfiGb7HmBmuwKTgHPDM6TvI3pyYF9g6xKy/yfwrpntAuwKTAcuBOaE2ukFkvYFdiB6Il4XoJukfmH28aFEz6M5FOiRxum8YGY9wvG+IJrJvFBrook4DgTuCedwIrDCzHqE/E+WtF0ax3GVhI9ldunaQtKn4f37wANAM2CBmY0P6/9ANJ3Zh6GveQ3gI6AdMK9wajNJjwGnFHOMvYhm/SE8WGtFeARron3D65OwXIcoQNYFXgwPm0fSyDTOqZOka4ma5XWAsQnbngnPxv5K0txwDvsCnROuL9YPx56VxrFcJeAB0aXrVzPrkrgiBL1fElcBb5jZkUnpugDlNQJAwA1m9p+kY5y9CccYARxsZlMlDQP6J2xLzsvCsc8ws8TAWTj7uasCvMnsytN4YPfwnBgk1ZK0I/AlsJ2k7UO6I0vY/y3gtLBvrqR6RMMb6yakGQsMT7g22VzSlsB7wCGStpBUl6h5Xpq6wDfh+ctHJ207XFJOKHMbYGY49mkhPZJ2lFQ7jeO4SsJriK7cmNnSUNN6UlLhZLeXmtksSacAoyUtAz4AOhWTxVnAvZJOBAqA08zsI0kfhm4tr4briO2Bj0INdSVwjJlNkfQ08CmwgKhZX5rLgI9D+s8oGnhnAu8CWwF/NrPfJN1PdG1xShh/vhQ4OL1Px1UGPpbZOecCbzI751zgAdE55wIPiM45F3hAdM65wAOic84FHhCdcy7wgOicc8H/A1hkPO1Mso8HAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_confusion_matrix(y_pred, y_true, np.asarray(['No Symbol', 'Symbol']), normalize=True)"
   ]
  },
  {
   "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": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}