image_classification.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Zs3Vxo5dmP7j"
   },
   "source": [
    "## Image classification\n",
    "\n",
    "### jey@aims.ac.za\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Module importation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 101
    },
    "id": "cRY4TeyJT1-c",
    "outputId": "342fcbcb-3a38-45b0-c326-c7e86cdf980e"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<p style=\"color: red;\">\n",
       "The default version of TensorFlow in Colab will soon switch to TensorFlow 2.x.<br>\n",
       "We recommend you <a href=\"https://www.tensorflow.org/guide/migrate\" target=\"_blank\">upgrade</a> now \n",
       "or ensure your notebook will continue to use TensorFlow 1.x via the <code>%tensorflow_version 1.x</code> magic:\n",
       "<a href=\"https://colab.research.google.com/notebooks/tensorflow_version.ipynb\" target=\"_blank\">more info</a>.</p>\n"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1.15.0\n"
     ]
    }
   ],
   "source": [
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Conv2D, MaxPooling2D, Flatten, Dropout\n",
    "from keras.datasets import cifar100\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from keras.utils import np_utils\n",
    "from sklearn.metrics import accuracy_score, confusion_matrix\n",
    "from keras.utils import to_categorical\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "from sklearn.model_selection import train_test_split\n",
    "import pandas as pd\n",
    "%matplotlib inline\n",
    "\n",
    "%tensorflow_version 1.x\n",
    "import tensorflow as tf\n",
    "print(tf.__version__)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "X1GTR35cmP7v"
   },
   "source": [
    "## Load the dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 54
    },
    "id": "uQ_KjFJomP7x",
    "outputId": "ac5338f0-5389-4d11-b177-a8b8327168fa"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Downloading data from https://www.cs.toronto.edu/~kriz/cifar-100-python.tar.gz\n",
      "169009152/169001437 [==============================] - 4s 0us/step\n"
     ]
    }
   ],
   "source": [
    "(X_train, Y_train), (X_test, Y_test) = cifar100.load_data(label_mode='fine')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "yEVqMquiWlTk"
   },
   "outputs": [],
   "source": [
    "# Your remaining code here.\n",
    "# Create additional cells and place your code neatly for easy readability. You can add a new cell by clicking on the `+ Code` button on the top left."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "REvIe0KwUjnB"
   },
   "source": [
    "# 1. Split the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "pvnnwtRfUfr1"
   },
   "outputs": [],
   "source": [
    "X_train, X_val, Y_train, Y_val = train_test_split(X_train, Y_train, test_size=.3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "GOvdHGmlUz-x"
   },
   "source": [
    "### View of shapes\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 54
    },
    "id": "pZ-GXR0JUxBM",
    "outputId": "420368f2-fae6-44a4-a027-09bec64e1c5c"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(35000, 32, 32, 3)\n",
      "(35000, 1)\n"
     ]
    }
   ],
   "source": [
    "print (X_train.shape)\n",
    "print (Y_train.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 54
    },
    "id": "SrFF2kTLUxGq",
    "outputId": "efbfae17-7cd1-4ea0-819d-995350310c98"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(15000, 32, 32, 3)\n",
      "(15000, 1)\n"
     ]
    }
   ],
   "source": [
    "print(X_val.shape)\n",
    "print (Y_val.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 54
    },
    "id": "JEdahRVFUxKr",
    "outputId": "288a4509-60c3-4ce8-a819-46c9efd35cf3"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(10000, 32, 32, 3)\n",
      "(10000, 1)\n"
     ]
    }
   ],
   "source": [
    "print(X_test.shape)\n",
    "print (Y_test.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "voDsrRnpjhW8"
   },
   "source": [
    "### Find the unique numbers from the train labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 126
    },
    "id": "VTtf3WF6UxEv",
    "outputId": "d3eb64da-2f4a-4d29-feae-72010320da14"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total number of outputs :  100\n",
      "Output classes :  [ 0  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19 20 21 22 23\n",
      " 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47\n",
      " 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71\n",
      " 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95\n",
      " 96 97 98 99]\n"
     ]
    }
   ],
   "source": [
    "classes = np.unique(Y_train)\n",
    "nClasses = len(classes)\n",
    "print('Total number of outputs : ', nClasses)\n",
    "print('Output classes : ', classes)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Nb1U09_ojt3b"
   },
   "source": [
    "### Plot some of the data\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 235
    },
    "id": "BXaytjzWUw_n",
    "outputId": "7665302f-efe2-4803-a200-8b8ff0cf0242"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 1.0, 'Ground Truth : [43]')"
      ]
     },
     "execution_count": 9,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXAAAADHCAYAAAAAoQhGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjIsIGh0\ndHA6Ly9tYXRwbG90bGliLm9yZy8li6FKAAAgAElEQVR4nO2deZRlV3Xev33vm1+9mrp6HjUTgUAS\nvaRgsA1mMCZeARKWDUkI8cKWbUIMMV5BgB3jAQdjA84yMUQYBRHmMESYBTYCowUyLEndmpDUdLfU\n6lZXqXqqud787t35470mdc7epXo1vaor9m8tLdU7fe65556773n3ne/svYmZYRiGYSSPYKM7YBiG\nYawMm8ANwzASik3ghmEYCcUmcMMwjIRiE7hhGEZCsQncMAwjodgEbhiGkVBsAl8niOgkEb1sA88/\nSkQvXqO2Pk1EDSJ6fI3au4mI5omIiejAWrRp9A6z7adt731EVCainjjYJHYCJ6LXE9HdncE61/n7\nLUREG923p4OIvtmZvOaJqNkxnoufP7bCNj9NRO9d4676/BkzX7bgnHuJ6O+IaIqIThPRb3h9+gQR\nHSOimIj+3cJ/Y+ZbAAyuc38Ti9m202bPbXvBuUeIaIKI7lxQdg0RHe7Y/SQRfYuInnXx35n5PQCe\nt879/QmJnMCJ6B0A/juAvwCwA8B2AL8F4IUAMoscE/asg08DM/8SM/cxcx+AzwD4wMXPzPxbfn0i\nSvW+l13xWQBHAWwD8C8BfICIfm7Bv9+P9j15cAP6lljMtjcVfwHgEa9sFMC/BjAMYCuAb6L9LGwI\niZvAiWgAwB8DeAszf4mZ57jN/cz8b5m53qn3SSL6KBF9g4jKAF5CRANE9CkiOk9Ep4jo94ko6NR/\nLxF9esF5DnR+4qc6n+8koj8hon8iornON+/Igvpv7LQ5QUTvWcX1vazzE/XdRHQGwMeJ6Ne9t4DU\nxeUHInoLgF8F8O7Om85XFzR3PRH9iIhmiOhzRJRdab+8Pg4CeBGA9zFzk5nvB/BVAL92sQ4zf4SZ\n/xFAfS3O+dOA2fbG2/aCfvwsgCsA/O+F5cw8xcwnuR2DhADEAC5fy3Mvh8RN4ABeACAL4PYu6v4b\nAO8DUAJwF4C/BjAA4FIAPw/g32PBpNNle7+G9ltnBsDvAQARXQ3gowDeCGAXgC0A9iyjXZ89APoA\n7APwlqeryMx/A+ALaP8M7GPm1y74518B8HK0r/f5nf4JiOgSIpomol3L7Cd5fz9nmccbLmbbC9go\n2+58sf01gLcCEGvZRBQS0TTaLycfAvDfum17rUniBD4C4AIzty4WENEPOjep6v2Mv52Z/4mZYwBN\nAK8H8K7Om81JAB/EIjd+Ef4XMx9j5iqALwK4tlP+OgBfZ+bvdd6S/gDtb+aV0gLwXmZudM61Uv6K\nmc8w8wSAry/orwMzP8HMg8z8VDeNMvM0gLsB/AERZYnoIIDXAiisoq+G2fZyWBfb7vCfAXyfmR9Y\npM2ImQfR/sJ8G9rLhRvCZl+D0pgAMEJEqYuGzsw/A7TVabhfSqcX/D0CIA3g1IKyUwB2L+PcZxb8\nXUH7TQJov5n85FzMXCaiiWW063OWmRurOP4ifn+H16DNi7wewP9Ae03wMQCfRvsnp7FyzLa7Z11s\nm4j2AvhtANcvVZeZL4qz54jois6XSU9J4hv4D9H+6fLqLuou/PlzAe03lf0LyvYBGOv8XYb7Brlj\nGX0aB7D34gciKqD9U3Ol+D/blupbz2MCd9YB/wUzb2XmF6D90/ueXvfjGYbZ9sbb9o0AdgL4cWed\n/oMAfqbzt0aA9pfdcpcf14TETeCdn+9/BOBviOh1RFQiooCIrgVQfJrjIrR/Gr6vc8x+AL+L9psj\nADwA4OeIaF9HTHrXMrr1JQC/TEQvIqIM2kLUWo7tgwCeS+0tTHkAf+j9+1m01wJ7BhFdTUR9nSWU\nNwF4CYC/WvDvGSLKob02niaiHNHm3ga30Zhtbwrb/jsAl6C9JHMt2vfjUOdvENEvEtHzOuvg/QA+\nDOAc2juyek7iJnAAYOYPoG2g/wXtG3wWwP8E8E4AP3iaQ/8T2t/4J9AWfj4L4NZOm3egLZg8BOAw\n2utq3fbnEQD/sdPeOIAptJcW1gRmfhTAnwG4E21D+Z5X5W8BPI/ae1O/tNz2iejSjsq/nLeIXwJw\nEsAkgF8H8IveT8h/BFAFcAPaY1xFeyuc8TSYbW+sbTNzvbO2foaZzwCYBdDo/A0AQ2h/Wc4AeBzt\nXz2vXKNloWVDlpHHWAoiuhVt1X+Mma9ag/Z+A+09tjkAVzHzqSUOMYx1YR1s+08A/A6ALDPnVtve\nkuezCdwwDCOZJHIJxTAMw7AJ3DAMI7GsagInolcS0VEieoyIbl6rThnGRmO2bSSBFa+BUzuAzjG0\n3VlHAdwL4A0dVXmxY2zB/RlNN7sE19cEmHnVWxVXYtvFYo6Hhkt+Z2RF/5VJqRMosalYGTe/jLT3\nMc1nUhkhUtqPY6/9QB6olXGsnNS/LcqO0ijSOquMTyCvM/bOScpFamOo7WxVb5tfTxnDmGX/tfa7\n20zrVpqanEO5XBNHrsYT8wYAjzHziXan6PNoOyAsauSdess+UbdHaEbof2XESmOs3mztrEv3hLT5\nQ/3eWrpMux59/uvunH6J+t1N2o8yxQi1c3LgfdROoD2kiuGvYJ5fQ0F+2bY9NFzC77z9NW5/opao\nF2Tda41i2ed8Wm75jqkpyhqRW5ZmGcmAG/I+haE2qUSirFqpOZ9zORkMMV2UXza1+Zoo48ibakIZ\ne2q+XBZlcSx35+XzeVFWr7gx00LlSzCK5RgGyjVFLWmPuSDtfE5l5NRZqVdEWTqTFmWBPxQpeb4o\ndvv/kQ/r4XFWs4SyG6477ygU111qB+8/RESHVnEuw+gly7btcllOWoax3qy7iMnMtzDzQWY+uN7n\nMoxestC2i8V13/JrGILVLKGMYUGMBLTDRI4tUndVhMpPY20xo7uFCm25ZDXfY96ygTakytoYSP68\n9tcNCnn5U7pWkwHc/PW/TmOyGzIwpnKUNj4Sdcz8n61dr5Zpa6Yb4th2kWXbNhEQ+L+WQ/nzOSL3\n/nEoly7qTflTXFsS8JcjG3UZej3Ncokgl5PLF/WWPDZMu32rN+ZFnVYg+5UvyC+zVtO3Imk/g4Py\nOFZyPtQa8tkpDXnHao+XMoZBVrYfpuT4+Lep2ZInYOW4dFqOP3tLJhlFR4i8td5AqQOs7g38XgBX\ndOLtZtCOTve1VbRnGJsFs20jEaz4DZyZW0T0VgD/ACAEcGsnboJhJBqzbSMprCoeODN/A8A31qgv\nhrFpMNs2koB5YhqGYSSUJGbkAaALa9oe71goaZpwp+xD7nIPtt+aLoh256DA3p5gjqTABWXfsPot\nrAi//lhoflWqr5W6v1rpP7vCDkfKcaE0OX3PfbIgArIpT9BmeWdS4Vbnczon9zRfmJR6qRQBgb5i\nn/M5Vvbdxy1pQ6w5nChTQZhxBThqKMKy3FqNGqRY2PL2VrPSVqzs09bE1aZi8Nms50MRSZvKKIJo\nGEsxsqXsy06xe2w9Uvblx3Ks44bsbIvda6oqvjHNhttWq6nMBbA3cMMwjMRiE7hhGEZCsQncMAwj\noSRiDTzSnEu0IDRdxPEgZf0vVMrSSlOh5rvirQXXI7lmpzuqyLJCwV1zDJTjUhnZiUZDWedUeuGH\nacnnpKPJYL90HrpwbkppTPYt63Utl5VroRNVuXbYUu5lLK5gc6+TEwiplHu9caysgafcRy7Fcoxy\nmgOQcs5m03XfT6VlW+m0bKteka1l8rJemHLtMVKcgup1GULAX78FIMaGlAesWlaCVCk6Qk557+Sq\npx8psV1ayijGSlupUNGGvABacqSBdCjHh5Q18HToOh1FiiNePu/WCRSHKcDewA3DMBKLTeCGYRgJ\nxSZwwzCMhGITuGEYRkJJhIgZa+KklhiiC4eTnOLsMJCVAs5AURH4SjLamJ8FZfT8nKhTq0nxxBcs\n22Vu+826FPwGBmTQ/ohl+5V5KYyUq67zxMiWPlFnoCgdS0qQwuy2ftmPbSVXaMkqguuxMRlp7+Gx\nGVE2J7rfXTaVDYMIoRAfZZ+b0bTzuaE4aKRS8riW5hTlOaZodgAlip2WxKCliNKNpmvLrbqWRUeK\nmLmc8rxm3P6LDDcAiop4yLPS0YbqirjvPYf9/TKyYV55zuvKnBEpuxUifx7JyParSgKPbEFOsX70\nUI7lvJLznKhCbQcF7A3cMAwjsdgEbhiGkVBsAjcMw0goq1oDJ6KTAObQ9jNoWdo045mC2baRBNZC\nxHwJM19Yg3YWJVDEoFApyyniT9FLmbRjSBHfFGEwGyieklIDQb3pCnx9OdnW3JwUAUt9sl695oog\ncUYROvNSqOpXvCfPn5NiYdUTcVKQ6dkyDSlKPXu3bP/K/dtFWY69dGFK5MQdW7eKsrnGE6LskbOu\ngLZm+eaXxzJt2xXmWixFrdjzBgwVB7t2EiCXVEa2FZErSmueyIFiQ4GSlb7VUMIKBt45pdYGKNEO\nY0U49Z/XQBFlcxk5GKms7H+lKlO7+RsYckpntxSUTQgp+VBXfZdlADPVsvO5qYi+vvAIAK1YPk+R\ndyiz4p3sRUlkNaKpLaEYhmEkltVO4AzgW0R0mIhuWosOGcYmwWzb2PSsdgnlRcw8RkTbANxBRD9m\n5u8trNAxfnsAjKSxLNseHi5tRB+Nn3JW9QbOzGOd/58D8FUANyh1bmHmgyYCGUliubbdV5IOUIax\n3qz4DZyIigACZp7r/P0KAH+8Zj1bgBbudbviFXnJdulZONLnihTNuhTuQkXMKypfbTlFSAgiV9zI\nFaWQ0a+In1o/chm3YloLK8rS0zOn5JIbTMu+9qW9UKaKUBvXpPg5pHidZViKM6GXKqpZk+JtKiPv\n2/49O0TZkfPuuMaql+36sBLb5jhG1Ru7hqI7hVn3XkWKJ2YaWmhXxTuw5YqYsZLSS2tfTStIsrOZ\nrC/KKsJ4TaZGC1JS9M6Te02ltCLUKnZMsXxOWpE859lzE+5xlbKoM6B4YeeUzQSDOaX/Xr0ZZayb\nsRRha0rKtth7zkHKcxK416imOsTqllC2A/hqx505BeCzzPz3q2jPMDYLZttGIljxBM7MJwA8bw37\nYhibArNtIynYNkLDMIyEsgHRCJf+zkh5a94jfq4uAFfuUiLpaVHQ6l6ku4ayZqesy2YCOTRZZQ0t\n5Xn31FJKerCmck4lRVLUcNtvNuTaWAy5ptZKS0eAFMl11FbNXU8MlZRnpKyPVubleuLkeVGE0IuG\nxyTH8MnRs6JsuirHv+Cl85pvymtk4d6zQe4+aK/R171UYqESZQ6hex2+wwYApElxaFHuSzY34Lbl\ne4gAMo8eZARNAGhpU4G3Jp1WHJP6s1K8HcnLHTkF39aUtfPZ6WlRNjUhDY2V9Wd4z+bYGWlnhay0\nj2xGXnepb0CUDW5zHdCGFaeg+YrUhYKcbN9fFw9DZez9Ne9FAm/aG7hhGEZCsQncMAwjodgEbhiG\nkVBsAjcMw0goPRYxCex9ZyjB2NDnRRV81m7NQUd2vRlJoaviCaJ5JZWZLAFqNdlWcUAKNo2W236t\nIdWGXFY6woSK2ImsK4xMK44wWvu1pnRGOHdeCkKlgjva1XnFiSQvxZlQEUQbUtfEzJzbXkXR1I4/\nNSvKMvl+UZbyxDdWRGX4otrGaZgAAsSRa0lZJYVX6EXqSysCVlpLFac4XaU9Z5h8XomG15LHRS0p\nRgaBImx6RUVFkGPF0SasyGcn9jYPzExMiDrTk1OirF6VhlavKecMXLtVNFKMn58UZdSSzkn9/dIe\n/fEvbRkWVTJKZMaWEiGy7tlArERmFIE8F7FtewM3DMNIKDaBG4ZhJBSbwA3DMBKKTeCGYRgJpcci\nJgsPI1JW54f7XVGuX4nwV51T0iop4k/G83jMKt9ZKcWDLQyUEIIsj23Mu0JjqIhGfYNS/Mwrwl0A\nNwpavSqFnuGhIVHWYikFtwaU6HU5LzKj4tD2xBnp+danREqdVtJaTdU9T8ScvG9zyjlDJTJj7Hl1\nap5oftFGaphhSBgacgcq0DzsYve+xC0p+IWhkq4rL4XwyNciY3kci0pAqHh6Bko6s2zaHeGgIW9e\ntSKFwakpmYUunXbbn5+ZEXUmzkl7byiCZV2JmBl7z+ZcVRpMdVaORSEjr2l2VlHovaHdq3gG50py\nswWFclOAn/ouUFLQUeTfD90V097ADcMwEopN4IZhGAnFJnDDMIyEsuQETkS3EtE5Inp4QdkwEd1B\nRMc7/5cLs4axyTHbNpJONyLmJwF8BMCnFpTdDOA7zPx+Irq58/mdXZ3RS91EwuUIGCy5gk1aWeQv\na+m6QiU8qtd8Ni8veUtpUJSFKSnAlf3QtAD6867geuXenaLOE0+OibJ6RfGC9EJbDpW2iDqVphRi\nmoq4mirIkJ5Hz7qC07mydFer1OT9COakkBQp54w9j81QESxJSZsVNZRQsV7zpHirBV5bWvKwJfgk\n1si2gwDI5V3bjhXP4DDjhQyuS+FReyZCJc1ayvPEbCmegEomNqj+z8pMEARue4GSdk1TwlstJQyy\nJ7RnMvL5KvbJsskJKYjOzkiRsRl56d+UjQnbt8lnYnjbiCibnpTewk+cOO18pkhe994rLxNl6RHp\nscmeCKuFFIYXinixjIJLvoF3MnH7UvOrAdzW+fs2AK9Zqh3D2GyYbRtJZ6Vr4NuZebzz9xm0cwga\nxjMBs20jMaxaxGRmxtNswSWim4joEBEdWu25DKOXLMe255RlJsNYb1Y6gZ8lop0A0Pn/ucUqMvMt\nzHyQmQ+u8FyG0UtWZNulkuLtZBjrzEo9Mb8G4E0A3t/5/+1dH+mtxofK6nw2dkWQfKh4oYXyuydU\nvo/27NrjfN42ID0g+3Ly4QtzMkTr2NlTomz3bneTwiW7dog6ZZb9Pz4mBdHyvCsqppQQn+empYBz\nUgkdO6OISzNeUUMRxsCyjJS8ioESJjPwhKpA8VbLKC+0Mrel0LoxFMo+1L2i6RWomAorsm0GI459\nUVjx3K17Zcr7PbMUAauK2Ok5N6oiJrPincnyvmiefq2G96tC8TIOskoIW0XgYy+ELSkiY3leeljO\nzMtfNtmifIYH+1wvyLqSh7beUnJWZqT35L7LtomyC2eecvtaVbxB0/J+T1ekx7Kv2UexvB9E7nhp\n9xHobhvh5wD8EMBVRDRKRG9G27hfTkTHAbys89kwEoXZtpF0lnwDZ+Y3LPJPL13jvhhGTzHbNpKO\neWIahmEklB5HIwQCb827qPRg7xZ3E3tOcSDYsnOXKMvniqLs2muucz5rF1yek+tZNWUdb9eeA6Ks\nb8hd354JlAh8GelA8NC4XE+vePv5cym5Hh1kZHSzeEA6IpGSNosn3bXylB9iDbrDQEpxkYmUslbg\nnjOOZR0tptpWJRXY87a4Y7Z/u1yrfKrpLgLfeWxUab03EANB7PanoaT1ajW9lGrKGjLSirNWQ67f\ncuyvp3f3OFfryvptS3Ee8syvVZP9yioJCTMZqR+Vy67mMz8tNaBzZ2VZSonoOLJVOriN7HQd6FKK\no9DxI8dEWdPPGwdgx67domxgcMA9ribXtlmJGFlRokGSr20FimbQRQlgb+CGYRiJxSZwwzCMhGIT\nuGEYRkKxCdwwDCOh9FTEJABZzwHnit1yU/6BLa4YWVQiA1551XNFWf+AFDf8lEZTU1OiTj2Q0dmm\n56TDTEpqFHhsdNz5/NAJKaQ9fEw6883WFa8TL4UVN6TIqB1GOSl2Rso1hZ7zRCGriIyK005GiY7H\nipMRPGebUlEKrlsG5f2+bETet8tT3rGBdHaY9gZDS03VKwgBQnLtNIbsc5B1Ra2IpNOO4rMESslH\nlTyRO60I6LHi0BIoadw4Je9x5KX1mp+UQmq1rFyj5vDj2WND6VcmI+14ZLsMRdM3LCMIhmnXXjJZ\n2dYlB/aLshkl/Vt5XgqUIyNuP1KB7ANnlAid83OyzEv9qKXeYy893mLxHOwN3DAMI6HYBG4YhpFQ\nbAI3DMNIKDaBG4ZhJJSeiphhQCh5nmfPvmSPqFeC6zE4VJReeC0lpdoMyVRISLuRBhtK2rVmRg7D\n2VnZ1rHDj4uyiRlXBKn5OdwA7Noq+3/NiBTzwpSXbq6lROBryvabSlTBC1NSiJkI3fY18WT/bunh\nuq0kPVyHSAqghcDrm9J+Rony2FAiJ56ecBPlzDWkR9tDZ92xL9e0KHu9gRnwtUEmRY1Mu3ZbLfsJ\ngYBsSo43paTdttj1XIyVqJeBIn+lsooorXg7T591RcvyhPTgTDVkWVER1X2Pynxe2oHmSVoalCnJ\n8v0yTWm95kYtnBg/K+oMFOS4BooY/9T4eVE2fta9zqsuk6kTtwzLZ7qQkzY5VXafzUJ6QNRpRP78\npgv09gZuGIaRUGwCNwzDSCg2gRuGYSSUbhI63EpE54jo4QVl7yWiMSJ6oPPfq9a3m4ax9phtG0mn\nGxHzkwA+AuBTXvmHmfkvl3U2BthL+7S1JAW+XMUVpyrT0nty6w7pVVVTUkqd8sK2npmR4t7pM1JI\nOj/+lCjLZ6SQtHuXK7Ls2C4FiVJehpPty8m2iN2+pQPpyRiGMv1bzLKtCxMyNOexKVeIue+ovMbR\n0+OirLRPCpvP2i890TKxK3qNKqLX8bEJUXZqUgrG4543XEURdGue12hrkbRTT8MnsUa23YojTHsh\nUxFK4TX0UmU1I3ldDHlc3NJShHllJB/ndEq+o2UUj81aVT47tUn3/oUtuXEgUPpaKsnwxjVv00GQ\nln3NFpX7pwjh9apMs3bqlGvL52Zlnasul89OsU8Kj9o4Ts2690kZLkQNWbh1SAquDW/MwlAe53tr\nano40MUbODN/D4Cc4Qwj4ZhtG0lnNWvgbyWihzo/Q+XXTAciuomIDhHRoVjLFmAYm49l23ZZSb5r\nGOvNSifwjwK4DMC1AMYBfHCxisx8CzMfZOaDwWK/Awxj87Ai2y72yZ/nhrHerMiRh5l/skueiD4O\n4OvdHBcQkE+767XDJel8kG25a3RjT8p101RhTJT175Pr4g8edZ1vTjwpj9s2JJ0Krrl8hyjbOSKd\nCnZ66Z2Kyob7tLKmliJtndN1NKhUZPQ3xMqXoOKMsHVERnG7Yp+7bn39s6TDxenx06LskSOPirJv\nz8k19sCLaHfyjNQuJuaVdGGiBGj4l6mMIfkvBEokxeWyUtuO0UKZ3dUYNU1Zze1jrKwhi7VtAGB5\nrxpVd8241pQRNPMFud7dn5faSnlK6hUtL8ef7AGQyynOQ4o9hl4UUnHvAFRr8ldMSomYmckpada2\nb3M+lwO5Xl9ryeP6i7L/KSWq5aVeir9MKK+xomg5W/rlGns17+p+VcUpLu05Q2ljCqzwDZyIFroh\nvRbAw4vVNYwkYbZtJIkl38CJ6HMAXgxghIhGAfwhgBcT0bVoh6k9CeA317GPhrEumG0bSWfJCZyZ\n36AUf2Id+mIYPcVs20g65olpGIaRUHoajRAAyIuOFtVlyqGml5IpDqR8Mj0vRZeH7zksyoLAFQh+\n+WU3iDrP3i8dVfqVaGktJRJgo+wKTmFZ2ZQfSQeFTEY6MKXyrsiYISliNmJZlskrQgxkWbrpCsjX\n3/izos7ohIy4eMWV+0TZF77/gCh78NEjbl+1baMknY4CRfgleOPIiueEYOO2qRIRwowrGNbK8l7F\ndVcYzKQ14VVeR1ZxyIm96I+FrLTZVEMe16oqqfrOSzE1aLr1cnkpiBJLwbUZy/bDrGuP6bQUUgNF\nqC6X5XMepuWOn4yXzmygpKQpU8a1pjjfFJVojdnQFUWbSuTLKJJ9HarIssG82/9pJa0be89JrKQ1\nBOwN3DAMI7HYBG4YhpFQbAI3DMNIKDaBG4ZhJJSeiphBQCjlXUEyn5FiWxi6dZ7zwhtFncFdB0TZ\nyGmZRum5z77abVtJefac664TZaR49Y2dOCXKZmdc77d6RYo6oZLGrb9fiVpYcr29OJJtzc5Kr9TK\nvPTAm52UURdbs66otmOX9NbMDcvrvvfYj0TZ6ZMnRFnaS/vVijRBVxRBDZGTsKgLxAFSkSvMFRXh\nsdnyvBsVUbfRkOJnU0k7R14UyiBWIvfNKim9JqSHbFSR9QY9D2WOZZ1mQ3o8RiXFGzntCqBZ5bkv\n5KWwX2lIQTSveJK2PMMaGZJjoXk216qy/7m07FvkGWlB6QO1pBDciuSY5T2xOWgo4qq38YFNxDQM\nw3hmYRO4YRhGQrEJ3DAMI6HYBG4YhpFQeipipoIAQyU3ZGompaSByrmeSrnBoqiz/bIDouz6G18k\nykaPuN6BM1My5GNdid7ZNyxThm2/RIoblwy54WQpJ9OnkeKZFigCH3v1Zi/I9Gbnj8r+VxVPz217\n9oqy/rwb2nL7FZeJOo9//5go27dPtvXsqy4XZaNTrlftsZNS9E2cOtklURRj3ktB1lLCydabrrFt\n2abli5D3s6mov826WxbU5fvY1BkpcNdnpZhXzEtv52bTFeBSSuqvSBHutFCxgReilZSQrVrI1HxB\n9iuTlcJvNnDLWPGwrJal13erpQQzTsuUcGHOFVhj5d23pUwkoZI6Lp1yr2mbkoJufM7dhKCNKWBv\n4IZhGInFJnDDMIyEYhO4YRhGQllyAieivUT0XSJ6lIgeIaK3dcqHiegOIjre+f+iyV8NYzNitm0k\nnW5EzBaAdzDzfURUAnCYiO4A8B8AfIeZ309ENwO4GcA7n66hIAiQz7geTFFDiiAFb5F/8sx5UeeG\nl0gvwnRBip2Zgis8BvPSC+3xozLnY2mrDDHbV5Lek0HB9aoqFGQOPARSdIEi3kZeLsS5hgxFOT0j\n+5/PyO9hPz9luz3X6+zEcXndSMuQoX0DUmQZGZLXefjHrncmb34XyzWzbSJCmHLF93pVEbXIFSjn\nZ2QeyFxK8fLTPB69nJVzU9L7duyEDFVaUHJKlgrynE1PcB1U8mtOz8lrjGLFI9HzJG22pMgYKB7L\nTUWgb0TSropeeNpULJ+dIJTXWGnKvl6YkiJvELhzy5aSEk5WCTGbTsmND/DE52JW9ivfcu1ikZSY\nS7+BM/M4M9/X+XsOwBEAuy42YNkAABDbSURBVAG8GsBtnWq3AXjNUm0ZxmbCbNtIOsvaRkhEBwBc\nB+BuANuZ+eI+tzMA5Ctx+5ibANwEALmU8iZqGJuA1dr2gLLV1TDWm65FTCLqA/BlAG9nZmczMrd/\nK6vRVpj5FmY+yMwHMzaBG5uQtbDtYlH+DDaM9aarN3AiSqNt4J9h5q90is8S0U5mHieinQDOLdkQ\nswg9R9riTuxO9AVWJv6yXHtrpOQaVHHEXbcOArm+JVfBgOHtW0QZhXK45iszzue6sranbcLP52Va\nqNkZt63z52V0xYwS4e6Esoafzck3wufc8FLn8/iZUVFndEpGOzyiOOQ89riMRjg3767BsnLdm2oF\nHGtn28RA2lubDetyfdtPrxfNyrXtSmNGlMWKw0zDc1aZuiDXwCtKWrdiXuoXKUWT8TLEoUXS9igl\n18XrdXlNDS8dYUaxT1baR0txalKc4AJPN0grz/mgss5fq8u55d7DMq3gNde4/d+3Z7+oU56X901z\nTmp694SVdIEUdJcesJtdKIR2pu4jzPyhBf/0NQBv6vz9JgC3d3VGw9gkmG0bSaebN/AXAngjgB8R\n0cVMtu8G8H4AXySiNwM4BeBX1qeLhrFumG0biWbJCZyZ78Liv3xfuki5YWx6zLaNpGOemIZhGAml\np9EI23K+K0o0FEGi5gk2Z44eF3Xq3/y2KGt5TjUAkAncaGw///xrZJ2iFCzTGSl4zNekOFObdyOc\nhSTrZJWN+qyJUjVX9MrnZWTDqbocr6lpGaHwiufsE2Vhnzs+YVOKbOfmZdl9R6WI+eS4dBD5aaZV\nb2Dq8ZNOWUMRtH1nn4xiZw3FzvzIgICMiJfNSkExm5HHaWm+pFQI+D40zaaSKi0jxfhqRdpQKuU2\n1ldUonZq6fZiORZzM9LeQe61D/UrQm1W9mvXZfI5yQ3KSKR79+5w6yj+ObPRSVHWqCqbJhru+Idp\nGXExFbn3VkvxCNgbuGEYRmKxCdwwDCOh2ARuGIaRUGwCNwzDSCg9FjEZjZYnSiiehbG3gF9XPJm/\n9PW/F2W79uwRZZcMuwLBDydkZMMXvOJfibK+UIoghYKUegb7XAE0JDmkcSz731JETJBbNjgio5gO\neKmdACCVlUJSYZsUZu8/6qZL++b//aqo8637j4iy83UpeuUU2UvKLN15kz0jYEbkpVBLpZX4KF5k\nSgql8EhpLXWZLGt5KmNfnxTxKxUpojUVEbOspAMjb3ogZcNBPi3velkRwqOWK/anA/mchBnZ/1hJ\njcaR7Ov8nHvOWkPaXqGopDdryLZmZ6VI+thjXvsl+fymW4pgqTw7Te+caeXJKQbuvBVaSjXDMIxn\nFjaBG4ZhJBSbwA3DMBKKTeCGYRgJpaciZjqdwradrkfTXFV6WoE9gUwRZ/ZeNizK9ozsEGU7YtcT\nM5qW53vyhIwWmt4mRYrB7VtFWbPqhvCcr8hUTsNbpKAYpqUIGJPr3jU9L/swMy1FkQcfGxdlj37v\nblF23/0POJ+HM4rIqIQHpaYsY5ahS/1QwdCEl2eorhmkMigN73bKFO0aLc+jkpRQyQFJ20illDRl\nnmgWxdI28nnp5deYL4uyWk0KcAVPVCw3peCneSy3ItnXqalJt+0+KbxDSxunhZJWaNSmnc9nx6V9\nblHSJD5xUqYoHB2dFmUHn3+187k8K0PObuvvLrTu3Kwr6PYPKmkY8/516w+OvYEbhmEkFJvADcMw\nEko3CR32EtF3iehRInqEiN7WKX8vEY0R0QOd/161/t01jLXDbNtIOt2sgbcAvIOZ7yOiEoDDRHRH\n598+zMx/uX7dM4x1xWzbSDTdJHQYBzDe+XuOiI4A2P30R+nMVxv4wcMnnbLzW6RA+auvuNH5/MQP\nD4s6g8NSBDl1+glRlt/nChf7dsnwkU9OToqyv/3935NtpeVw7exzhbprr32+qPP4pPRMG52UQknD\ny6H41BmZn3J6UoZxZUU0aoaKX6SXh/DknBRwpquyr8xKXj9FU2HPo4z0VMCyRK+47qylbQcUIpNz\nQ6RqHrggd3zrNSl612rKPVBFLHe841iKn4FiBwTF01PxsmwFrigXKyJpU7nGOJDCY8OrN6eI/Y2m\nzCkZkhRhs2kpnOayro2WCvJZHRgcFGVbdsj5Z9uQ9KDd5nleTlXlWJdKA6JsZk7ey9lpV2ANFW/0\nwBOQ2d8gcLGeWroIRHQAwHUALm5xeCsRPUREtxKR9Ps2jIRgtm0kka4ncCLqQzt799uZeRbARwFc\nBuBatN9iPrjIcTcR0SEiOqS+kRjGBrMWtl1WfrkYxnrT1QRORGm0DfwzzPwVAGDms8wcMXMM4OMA\nbtCOZeZbmPkgMx8MgsXSDxrGxrBWtl3MK/uaDWOdWXINnIgIwCcAHGHmDy0o39lZQwSA1wJ4eKm2\nmlGMp7z1n7yy/nPnYTdq3olxuV48OK/kX1LSjZ2bc50WXnvpPxN17rnnXlE2X5Xr4qhJR4atuw44\nnyuz0inozrvuF2UXlLbIi0YYQ64lkhLFbV5J2zSnOGtMeKm6qkqaLlKjnslxjVf8Xbx5foWtpW1H\ncYRyxR1z1gbJyxumLBcjk5XrvuWytBfpNyXPF8Xy3qWUk9YUe5nyIiUWC/K4ak1JDagIJM3I7dvc\nvHRwSaeUwWB53bHy2hk13OdiUFnvrs5Lh7ctofzi3danpKGrunpRX0F2ghQHrEpdPpsgL9qq4uxT\nbHpr84s8Nt3sQnkhgDcC+BERXXTlezeANxDRtZ2mTwL4zS7aMozNhNm2kWi62YVyF7RQz8A31r47\nhtE7zLaNpGOemIZhGAnFJnDDMIyE0tNohIAMUDejRCM8/OOTzud+JXJZJl8SZZWmFGLKnlhy26e/\nIOpogzC8a7soyxXkpv/0oFvvrkdOiDpxNifKQkWMnJp0BdCJshRAysp4aenZ1B2b3i4gXbA0VgbB\nX40JU0un14uk7oWBfhmdLlYcbebLrjNMpDQWK6nYtNveVFKvlSuuPWb37hR1ak1FxKwr/fCUx9kZ\n6cgzMCCvO6uMoXadjYa7jTPXL1MPZhWHsb6cHIw+ZW6ZnXXnlmxRCq6tSI6hctvQ8qI61rX0b752\nu4iIaW/ghmEYCcUmcMMwjIRiE7hhGEZCsQncMAwjofRcxIw9BWWmKj2tBra5sYPCjExVNDsrUybN\nz0gRM51xvdp2bpWp2Ao5KTK2svKcUSg95L5z7yPO5+NKBME5JZrf2UlZL/IEobqmMSoKVBBKQUWN\nXrZIRDNj9RARgsC1D9bUwtCNpBcpXniaAr1lSMbTqtdde2kqnsjEsg+xEr2ypXpsus9AU/G6rLbk\nM4dY1vMF3kpVejsGgRwL6pcbGGp1eWwq47Y/OSUjbZaUTQj+GAJAQalXLLkRCtNpOa6VOSnMxoq3\nc8WLQBmkpOBa9jw4Yz/N5MVj1VLDMAxj02MTuGEYRkKxCdwwDCOh2ARuGIaRUHouYsITVZpKLKGx\niTnnc3pAplDKKiE3tw9JMWBw2BV/BkvS26ulpKJCSrY/MSsFm8fHzjufR6dnRZ1yUwpEDHnOlDcU\npH2/KqKUGra0K73SRM21ggHEnqdrrHgM+uJyriDtkQNFjAwUe/FsNAikGO/3CQBma1Is1DwX83lX\nxJyZlun85hXhLpeR04qv50YtadvzSohZzbQ1Qa8Yu0J+Wkm7Njsjk25klQ0SfrhXABgoueNTLMnx\nipTn9dzUGVHW8C5qXgktnWm4ZYslw7E3cMMwjIRiE7hhGEZCWXICJ6IcEd1DRA8S0SNE9Eed8kuI\n6G4ieoyIvkBEym8Rw9i8mG0bSaebNfA6gF9g5vlO/sC7iOibAH4XwIeZ+fNE9DEAb0Y7GewSuN8Z\nmrNDzdv8Pjsn1wQv375HlG0dkM4Okbe+PXpWrknNKZv5x6ZmRNmM4nww560n1rT8AIGSfklxqom8\nNemw2yXqLh10Yq99ZTn9p401s20igLz0gK1ISQfGri3nCjL1lxZBMFCMoa/fjZpXbci14VosbbZv\nUK679yvPztSEu+Y9V5bOc+l0UZRFiqNQHLn9YCVdYFPRilqzco09TMvnqdVwjTmnOOellAigdXmL\nUFMcohC7+lda0eDyRekA1FLkNf8ya3NybT5T8Bx5NJ0OXbyBc5uLdy7d+Y8B/AKAL3XKbwPwmqXa\nMozNhNm2kXS6zUofdnIGngNwB4DHAUwz88Wv2lEAu9eni4axfphtG0mmqwmcmSNmvhbAHgA3AHhW\ntycgopuI6BARHVphHw1j3Vgr2y4rCREMY71Z1i4UZp4G8F0ALwAwSEQXF5X2ABhb5JhbmPkgMx9c\nVU8NYx1ZrW0XC3LN1TDWmyVFTCLaCqDJzNNElAfwcgB/jraxvw7A5wG8CcDtS59Opp1S8ZwPqsr6\n/ZQSGa0eSeFxwotKdmZaOuNUFdGoGknRSMpBQEzeECppm6A47WiwNzb6SK3C+cZES4e1tW0AgZcu\njaTF5PtcB5OmIjLWlQidaeULIu1FzCzHUvDLlKRDC+WlDZXL0gFtbsYVLXMF6bzSV5Di5/TEOVFG\nnmbpC74AkErL6aiqpF4j5Xn1XYBKDcU5JpQPQCuSZc1Y9q1Rdu9TY1zONfv2j4iywL9wAPW6K1pq\nkUOnp906kTIfAd3tQtkJ4DYiCtF+Y/8iM3+diB4F8Hki+lMA9wP4RBdtGcZmwmzbSDRLTuDM/BCA\n65TyE2ivGRpGIjHbNpKOeWIahmEkFJvADcMwEgqpqbfW62RE5wGcAjACQIY2Sw5J7n+S+w48ff/3\nM/PWXnbmImbbm4Ik9x1YgW33dAL/yUmJDiV5W2GS+5/kvgObv/+bvX9LkeT+J7nvwMr6b0sohmEY\nCcUmcMMwjISyURP4LRt03rUiyf1Pct+Bzd//zd6/pUhy/5Pcd2AF/d+QNXDDMAxj9dgSimEYRkLp\n+QRORK8koqOdbCc39/r8y4WIbiWic0T08IKyYSK6g4iOd/4vA0JsAohoLxF9l4ge7WSceVunfNP3\nP2nZcsyue0eS7RpYY9tm5p79ByBEO97ypQAyAB4EcHUv+7CCPv8cgOsBPLyg7AMAbu78fTOAP9/o\nfi7S950Aru/8XQJwDMDVSeg/2qG3+jp/pwHcDeCfA/gigNd3yj8G4Lc3QV/Nrnvb98Tadadva2bb\nve74CwD8w4LP7wLwro0e0C76fcAz9KMAdi4wpqMb3ccur+N2tCPuJar/AAoA7gNwI9qODinNnjaw\nf2bXG3sdibTrTj9XZdu9XkLZDeD0gs9JzXaynZnHO3+fAbB9IzvTDUR0AO3ATXcjIf1PULYcs+sN\nIol2DaydbZuIuUq4/XW5qbfyEFEfgC8DeDszO4GfN3P/eRXZcozVsZnt4iJJtWtg7Wy71xP4GIC9\nCz4vmu1kk3OWiHYCQOf/MoL9JqGTbf3LAD7DzF/pFCem/8DKsuX0GLPrHvNMsGtg9bbd6wn8XgBX\ndNTWDIDXA/haj/uwFnwN7UwtwHIytvQYIiK0kxEcYeYPLfinTd9/ItpKRIOdvy9myzmC/58tB9g8\nfTe77iFJtmtgjW17AxbtX4W2avw4gPdstIjQRX8/B2Ac7YxqowDeDGALgO8AOA7g2wCGN7qfi/T9\nRWj/jHwIwAOd/16VhP4DeC7a2XAeAvAwgP/aKb8UwD0AHgPwfwBkN7qvnX6ZXfeu74m1607/18y2\nzRPTMAwjoZiIaRiGkVBsAjcMw0goNoEbhmEkFJvADcMwEopN4IZhGAnFJnDDMIyEYhO4YRhGQrEJ\n3DAMI6H8PxmuFp2KPl4KAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x216 with 2 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "data_point = 15\n",
    "\n",
    "plt.figure(figsize=[6,3])\n",
    " \n",
    "# training data\n",
    "plt.subplot(121)\n",
    "plt.imshow(X_train[data_point,:,:], cmap='gray')\n",
    "plt.title(\"Ground Truth : {}\".format(Y_train[data_point]))\n",
    " \n",
    "# testing data\n",
    "plt.subplot(122)\n",
    "plt.imshow(X_test[data_point,:,:], cmap='gray')\n",
    "plt.title(\"Ground Truth : {}\".format(Y_test[data_point]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "l56aGMqTlAQl"
   },
   "source": [
    "### We don't need to reshape the data here because the deep is already in depth 3"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "Bz1ONG2alT0U"
   },
   "source": [
    "### Normalization\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "KPdbYW-lk_XM"
   },
   "outputs": [],
   "source": [
    "X_train = X_train / 255\n",
    "X_val = X_val / 255\n",
    "X_test = X_test / 255"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "id": "5v1GY23vpyBf",
    "outputId": "e157dff7-910e-436e-89a0-7eb4a4e01443"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.0"
      ]
     },
     "execution_count": 11,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.max()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "IrlCjjTaleLd"
   },
   "source": [
    "### One hot coding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "id": "KaoKz69BlpDe",
    "outputId": "2d34b177-1a64-4dcf-c359-24e318b0a208"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([59])"
      ]
     },
     "execution_count": 12,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Y_train[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "R0jD-YhOlh0I"
   },
   "outputs": [],
   "source": [
    "Y_train = np_utils.to_categorical(Y_train, 100)\n",
    "Y_val = np_utils.to_categorical(Y_val, 100)\n",
    "Y_test = np_utils.to_categorical(Y_test, 100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 145
    },
    "id": "XEskVsndlvOu",
    "outputId": "3b80af10-1287-4da6-ffa9-1d083629135a"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
       "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
       "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
       "       0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
       "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.,\n",
       "       0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
       "      dtype=float32)"
      ]
     },
     "execution_count": 14,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Y_train[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "ACsh1ftVl8qm"
   },
   "source": [
    "# 2. Create a CNN model\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "5xKbyiiBLMFV"
   },
   "source": [
    "### Network1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "4TEKP5ubl7dL"
   },
   "outputs": [],
   "source": [
    "def baseline_model1():\n",
    "    # create model\n",
    "    model = Sequential()\n",
    "    \n",
    "    model.add(Conv2D(filters=200, kernel_size=2, padding='same', activation='relu', input_shape=(32,32,3))) \n",
    "    model.add(MaxPooling2D(pool_size=2))\n",
    "    model.add(Dropout(rate = 0.4))\n",
    "    \n",
    "    model.add(Flatten())\n",
    "    \n",
    "    model.add(Dense(128, activation='relu'))\n",
    "    model.add(Dropout(rate = 0.4))\n",
    "    \n",
    "    model.add(Dense(100, activation='softmax'))\n",
    "    \n",
    "    # Compile the model\n",
    "    model.compile(loss='categorical_crossentropy',\n",
    "             optimizer='adam',\n",
    "             metrics=['accuracy'])\n",
    "    return model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "XuIj5KrgL_4y"
   },
   "source": [
    "**Network1** the accuracy of the prediction on the training data is 0.2597! too small and confusion matrix has shown too many prediction error!\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "NAdcpzhnLTf5"
   },
   "source": [
    "### Network2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "JMoad3QKLWOp"
   },
   "outputs": [],
   "source": [
    "def baseline_model2():\n",
    "    # create model\n",
    "    model = Sequential()\n",
    "    \n",
    "    model.add(Conv2D(filters=200, kernel_size=2, padding='same', activation='relu', input_shape=(32,32,3))) \n",
    "    model.add(MaxPooling2D(pool_size=2))\n",
    "    model.add(Dropout(rate = 0.4))\n",
    "\n",
    "    model.add(Conv2D(filters=64, kernel_size=2, padding='same', activation='relu'))\n",
    "    model.add(MaxPooling2D(pool_size=2))\n",
    "    model.add(Dropout(rate = 0.4))\n",
    "    \n",
    "    model.add(Flatten())\n",
    "    \n",
    "    model.add(Dense(128, activation='relu'))\n",
    "    model.add(Dense(64, activation='relu'))\n",
    "\n",
    "\n",
    "    model.add(Dropout(rate = 0.4))\n",
    "    \n",
    "    model.add(Dense(100, activation='softmax'))\n",
    "    \n",
    "    # Compile the model\n",
    "    model.compile(loss='categorical_crossentropy',\n",
    "             optimizer='adam',\n",
    "             metrics=['accuracy'])\n",
    "    return model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "1xuXmxUbURZm"
   },
   "source": [
    "**Network2** the accuracy of the prediction on the training data is 0.39! better then the previous one but still too small and confusion matrix has shown too many prediction error!\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "NJVx7Q8VLXWt"
   },
   "source": [
    "### Network3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "ZzNm-jauLa14"
   },
   "outputs": [],
   "source": [
    "def baseline_model3():\n",
    "    # create model\n",
    "    model = Sequential()\n",
    "    \n",
    "    model.add(Conv2D(filters=200, kernel_size=2, padding='same', activation='relu', input_shape=(32,32,3))) \n",
    "    model.add(MaxPooling2D(pool_size=2))\n",
    "    model.add(Dropout(rate = 0.4))\n",
    "\n",
    "    model.add(Conv2D(filters=64, kernel_size=2, padding='same', activation='relu'))\n",
    "    model.add(MaxPooling2D(pool_size=2))\n",
    "    model.add(Dropout(rate = 0.4))\n",
    "\n",
    "    model.add(Conv2D(filters=32, kernel_size=2, padding='same', activation='relu'))\n",
    "    model.add(MaxPooling2D(pool_size=2))\n",
    "    model.add(Dropout(rate = 0.4))\n",
    "    \n",
    "    model.add(Flatten())\n",
    "    \n",
    "    model.add(Dense(128, activation='relu'))\n",
    "    model.add(Dense(64, activation='relu'))\n",
    "    model.add(Dense(64, activation='linear'))\n",
    "\n",
    "\n",
    "    model.add(Dropout(rate = 0.4))\n",
    "    \n",
    "    model.add(Dense(100, activation='softmax'))\n",
    "    \n",
    "    # Compile the model\n",
    "    model.compile(loss='categorical_crossentropy',\n",
    "             optimizer='adam',\n",
    "             metrics=['accuracy'])\n",
    "    return model"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "vjpXKEeGVAT-"
   },
   "source": [
    "**Network3** more complex then the previous one! But the accuracy is still sthe same 3.954. So the model isn't good\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "YELgaLYiVG5s"
   },
   "source": [
    "# Testing the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "oTFzTNdOmdcu"
   },
   "outputs": [],
   "source": [
    "model = baseline_model3()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "X3dHMEUtmkvu"
   },
   "source": [
    "Determine the number of trainable parameters\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 563
    },
    "id": "HP3k9YTMmjkZ",
    "outputId": "8883e2ac-4680-4728-fdc2-b9fe72be2bad"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_2\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d_2 (Conv2D)            (None, 32, 32, 200)       2600      \n",
      "_________________________________________________________________\n",
      "max_pooling2d_2 (MaxPooling2 (None, 16, 16, 200)       0         \n",
      "_________________________________________________________________\n",
      "dropout_3 (Dropout)          (None, 16, 16, 200)       0         \n",
      "_________________________________________________________________\n",
      "conv2d_3 (Conv2D)            (None, 16, 16, 64)        51264     \n",
      "_________________________________________________________________\n",
      "max_pooling2d_3 (MaxPooling2 (None, 8, 8, 64)          0         \n",
      "_________________________________________________________________\n",
      "dropout_4 (Dropout)          (None, 8, 8, 64)          0         \n",
      "_________________________________________________________________\n",
      "flatten_2 (Flatten)          (None, 4096)              0         \n",
      "_________________________________________________________________\n",
      "dense_3 (Dense)              (None, 128)               524416    \n",
      "_________________________________________________________________\n",
      "dense_4 (Dense)              (None, 64)                8256      \n",
      "_________________________________________________________________\n",
      "dropout_5 (Dropout)          (None, 64)                0         \n",
      "_________________________________________________________________\n",
      "dense_5 (Dense)              (None, 100)               6500      \n",
      "=================================================================\n",
      "Total params: 593,036\n",
      "Trainable params: 593,036\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "a-QtfwxDmx5w"
   },
   "source": [
    "### The training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "id": "k86P2U8xmwMB",
    "outputId": "30913301-63cf-4e24-8d06-f331444f585c"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 35000 samples, validate on 10000 samples\n",
      "Epoch 1/150\n",
      "35000/35000 [==============================] - 9s 255us/step - loss: 4.3388 - acc: 0.0433 - val_loss: 3.9870 - val_acc: 0.0966\n",
      "Epoch 2/150\n",
      "35000/35000 [==============================] - 8s 237us/step - loss: 3.8593 - acc: 0.1015 - val_loss: 3.6029 - val_acc: 0.1589\n",
      "Epoch 3/150\n",
      "35000/35000 [==============================] - 8s 237us/step - loss: 3.5936 - acc: 0.1439 - val_loss: 3.3802 - val_acc: 0.2013\n",
      "Epoch 4/150\n",
      "35000/35000 [==============================] - 8s 238us/step - loss: 3.4076 - acc: 0.1769 - val_loss: 3.2262 - val_acc: 0.2388\n",
      "Epoch 5/150\n",
      "35000/35000 [==============================] - 8s 238us/step - loss: 3.2568 - acc: 0.2038 - val_loss: 3.0649 - val_acc: 0.2613\n",
      "Epoch 6/150\n",
      "35000/35000 [==============================] - 8s 238us/step - loss: 3.1497 - acc: 0.2251 - val_loss: 2.9883 - val_acc: 0.2765\n",
      "Epoch 7/150\n",
      "35000/35000 [==============================] - 8s 239us/step - loss: 3.0534 - acc: 0.2437 - val_loss: 2.8905 - val_acc: 0.2896\n",
      "Epoch 8/150\n",
      "35000/35000 [==============================] - 8s 239us/step - loss: 2.9780 - acc: 0.2560 - val_loss: 2.8425 - val_acc: 0.2997\n",
      "Epoch 9/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 2.9042 - acc: 0.2711 - val_loss: 2.7805 - val_acc: 0.3063\n",
      "Epoch 10/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 2.8416 - acc: 0.2838 - val_loss: 2.7474 - val_acc: 0.3193\n",
      "Epoch 11/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 2.7801 - acc: 0.2943 - val_loss: 2.7071 - val_acc: 0.3276\n",
      "Epoch 12/150\n",
      "35000/35000 [==============================] - 9s 244us/step - loss: 2.7230 - acc: 0.3063 - val_loss: 2.6246 - val_acc: 0.3408\n",
      "Epoch 13/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 2.6856 - acc: 0.3146 - val_loss: 2.6306 - val_acc: 0.3355\n",
      "Epoch 14/150\n",
      "35000/35000 [==============================] - 8s 239us/step - loss: 2.6444 - acc: 0.3219 - val_loss: 2.5991 - val_acc: 0.3468\n",
      "Epoch 15/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 2.5968 - acc: 0.3319 - val_loss: 2.5427 - val_acc: 0.3571\n",
      "Epoch 16/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 2.5708 - acc: 0.3374 - val_loss: 2.5290 - val_acc: 0.3561\n",
      "Epoch 17/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 2.5329 - acc: 0.3455 - val_loss: 2.5299 - val_acc: 0.3598\n",
      "Epoch 18/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 2.5022 - acc: 0.3520 - val_loss: 2.5272 - val_acc: 0.3594\n",
      "Epoch 19/150\n",
      "35000/35000 [==============================] - 8s 239us/step - loss: 2.4736 - acc: 0.3596 - val_loss: 2.4738 - val_acc: 0.3678\n",
      "Epoch 20/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 2.4494 - acc: 0.3641 - val_loss: 2.4810 - val_acc: 0.3689\n",
      "Epoch 21/150\n",
      "35000/35000 [==============================] - 8s 239us/step - loss: 2.4243 - acc: 0.3696 - val_loss: 2.4948 - val_acc: 0.3657\n",
      "Epoch 22/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 2.4061 - acc: 0.3722 - val_loss: 2.4489 - val_acc: 0.3766\n",
      "Epoch 23/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 2.3874 - acc: 0.3721 - val_loss: 2.4513 - val_acc: 0.3734\n",
      "Epoch 24/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 2.3634 - acc: 0.3799 - val_loss: 2.4168 - val_acc: 0.3793\n",
      "Epoch 25/150\n",
      "35000/35000 [==============================] - 8s 239us/step - loss: 2.3423 - acc: 0.3821 - val_loss: 2.4604 - val_acc: 0.3734\n",
      "Epoch 26/150\n",
      "35000/35000 [==============================] - 8s 239us/step - loss: 2.3270 - acc: 0.3846 - val_loss: 2.4175 - val_acc: 0.3828\n",
      "Epoch 27/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 2.2973 - acc: 0.3944 - val_loss: 2.4564 - val_acc: 0.3746\n",
      "Epoch 28/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 2.2752 - acc: 0.3967 - val_loss: 2.4148 - val_acc: 0.3825\n",
      "Epoch 29/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 2.2595 - acc: 0.4023 - val_loss: 2.4055 - val_acc: 0.3819\n",
      "Epoch 30/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 2.2422 - acc: 0.4035 - val_loss: 2.3884 - val_acc: 0.3838\n",
      "Epoch 31/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 2.2320 - acc: 0.4058 - val_loss: 2.4112 - val_acc: 0.3821\n",
      "Epoch 32/150\n",
      "35000/35000 [==============================] - 8s 239us/step - loss: 2.2078 - acc: 0.4127 - val_loss: 2.3677 - val_acc: 0.3913\n",
      "Epoch 33/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 2.2047 - acc: 0.4130 - val_loss: 2.3875 - val_acc: 0.3869\n",
      "Epoch 34/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 2.1844 - acc: 0.4158 - val_loss: 2.3923 - val_acc: 0.3887\n",
      "Epoch 35/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 2.1608 - acc: 0.4220 - val_loss: 2.3624 - val_acc: 0.3923\n",
      "Epoch 36/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 2.1454 - acc: 0.4240 - val_loss: 2.3673 - val_acc: 0.3909\n",
      "Epoch 37/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 2.1221 - acc: 0.4312 - val_loss: 2.4000 - val_acc: 0.3846\n",
      "Epoch 38/150\n",
      "35000/35000 [==============================] - 8s 242us/step - loss: 2.1233 - acc: 0.4327 - val_loss: 2.3774 - val_acc: 0.3895\n",
      "Epoch 39/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 2.1095 - acc: 0.4339 - val_loss: 2.3650 - val_acc: 0.3940\n",
      "Epoch 40/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 2.0800 - acc: 0.4388 - val_loss: 2.3562 - val_acc: 0.4009\n",
      "Epoch 41/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 2.0753 - acc: 0.4410 - val_loss: 2.3414 - val_acc: 0.4004\n",
      "Epoch 42/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 2.0519 - acc: 0.4450 - val_loss: 2.3545 - val_acc: 0.3936\n",
      "Epoch 43/150\n",
      "35000/35000 [==============================] - 8s 238us/step - loss: 2.0598 - acc: 0.4406 - val_loss: 2.3727 - val_acc: 0.3938\n",
      "Epoch 44/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 2.0293 - acc: 0.4480 - val_loss: 2.3486 - val_acc: 0.3971\n",
      "Epoch 45/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 2.0490 - acc: 0.4428 - val_loss: 2.3733 - val_acc: 0.3953\n",
      "Epoch 46/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 2.0162 - acc: 0.4532 - val_loss: 2.3577 - val_acc: 0.3955\n",
      "Epoch 47/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 2.0148 - acc: 0.4512 - val_loss: 2.3942 - val_acc: 0.3874\n",
      "Epoch 48/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 2.0009 - acc: 0.4558 - val_loss: 2.3550 - val_acc: 0.4006\n",
      "Epoch 49/150\n",
      "35000/35000 [==============================] - 8s 242us/step - loss: 1.9900 - acc: 0.4604 - val_loss: 2.3650 - val_acc: 0.3958\n",
      "Epoch 50/150\n",
      "35000/35000 [==============================] - 8s 239us/step - loss: 1.9671 - acc: 0.4593 - val_loss: 2.3758 - val_acc: 0.3907\n",
      "Epoch 51/150\n",
      "35000/35000 [==============================] - 8s 239us/step - loss: 1.9690 - acc: 0.4636 - val_loss: 2.3410 - val_acc: 0.3996\n",
      "Epoch 52/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 1.9470 - acc: 0.4701 - val_loss: 2.3417 - val_acc: 0.3984\n",
      "Epoch 53/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.9357 - acc: 0.4717 - val_loss: 2.3678 - val_acc: 0.3922\n",
      "Epoch 54/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.9311 - acc: 0.4717 - val_loss: 2.3460 - val_acc: 0.4000\n",
      "Epoch 55/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 1.9175 - acc: 0.4743 - val_loss: 2.3531 - val_acc: 0.3959\n",
      "Epoch 56/150\n",
      "35000/35000 [==============================] - 8s 242us/step - loss: 1.9131 - acc: 0.4766 - val_loss: 2.3471 - val_acc: 0.3956\n",
      "Epoch 57/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 1.9037 - acc: 0.4793 - val_loss: 2.3367 - val_acc: 0.4000\n",
      "Epoch 58/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 1.8974 - acc: 0.4795 - val_loss: 2.3638 - val_acc: 0.3943\n",
      "Epoch 59/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.8883 - acc: 0.4788 - val_loss: 2.3497 - val_acc: 0.3993\n",
      "Epoch 60/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 1.8728 - acc: 0.4819 - val_loss: 2.3548 - val_acc: 0.3987\n",
      "Epoch 61/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.8794 - acc: 0.4833 - val_loss: 2.3853 - val_acc: 0.3915\n",
      "Epoch 62/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 1.8578 - acc: 0.4884 - val_loss: 2.3547 - val_acc: 0.3966\n",
      "Epoch 63/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.8438 - acc: 0.4903 - val_loss: 2.3566 - val_acc: 0.3988\n",
      "Epoch 64/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.8330 - acc: 0.4913 - val_loss: 2.3533 - val_acc: 0.3965\n",
      "Epoch 65/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.8396 - acc: 0.4928 - val_loss: 2.3542 - val_acc: 0.4016\n",
      "Epoch 66/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 1.8224 - acc: 0.4929 - val_loss: 2.3789 - val_acc: 0.3976\n",
      "Epoch 67/150\n",
      "35000/35000 [==============================] - 8s 242us/step - loss: 1.8264 - acc: 0.4960 - val_loss: 2.3953 - val_acc: 0.3946\n",
      "Epoch 68/150\n",
      "35000/35000 [==============================] - 8s 239us/step - loss: 1.8272 - acc: 0.4918 - val_loss: 2.3618 - val_acc: 0.4023\n",
      "Epoch 69/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.7962 - acc: 0.5002 - val_loss: 2.3772 - val_acc: 0.3994\n",
      "Epoch 70/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.8048 - acc: 0.4992 - val_loss: 2.3835 - val_acc: 0.3923\n",
      "Epoch 71/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 1.7853 - acc: 0.5039 - val_loss: 2.3728 - val_acc: 0.4025\n",
      "Epoch 72/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.7957 - acc: 0.5022 - val_loss: 2.3724 - val_acc: 0.3999\n",
      "Epoch 73/150\n",
      "35000/35000 [==============================] - 8s 239us/step - loss: 1.7768 - acc: 0.5042 - val_loss: 2.3936 - val_acc: 0.3972\n",
      "Epoch 74/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 1.7775 - acc: 0.5085 - val_loss: 2.3665 - val_acc: 0.3974\n",
      "Epoch 75/150\n",
      "35000/35000 [==============================] - 8s 242us/step - loss: 1.7707 - acc: 0.5058 - val_loss: 2.3732 - val_acc: 0.3943\n",
      "Epoch 76/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.7658 - acc: 0.5075 - val_loss: 2.3515 - val_acc: 0.4008\n",
      "Epoch 77/150\n",
      "35000/35000 [==============================] - 8s 239us/step - loss: 1.7509 - acc: 0.5109 - val_loss: 2.3637 - val_acc: 0.3999\n",
      "Epoch 78/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 1.7443 - acc: 0.5114 - val_loss: 2.3723 - val_acc: 0.3999\n",
      "Epoch 79/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.7490 - acc: 0.5139 - val_loss: 2.3713 - val_acc: 0.3975\n",
      "Epoch 80/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 1.7339 - acc: 0.5165 - val_loss: 2.3832 - val_acc: 0.3944\n",
      "Epoch 81/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 1.7354 - acc: 0.5159 - val_loss: 2.3826 - val_acc: 0.3998\n",
      "Epoch 82/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.7305 - acc: 0.5189 - val_loss: 2.3690 - val_acc: 0.3975\n",
      "Epoch 83/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 1.7161 - acc: 0.5173 - val_loss: 2.3765 - val_acc: 0.3977\n",
      "Epoch 84/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 1.7048 - acc: 0.5197 - val_loss: 2.3877 - val_acc: 0.3968\n",
      "Epoch 85/150\n",
      "35000/35000 [==============================] - 8s 242us/step - loss: 1.7100 - acc: 0.5210 - val_loss: 2.3842 - val_acc: 0.3995\n",
      "Epoch 86/150\n",
      "35000/35000 [==============================] - 9s 243us/step - loss: 1.7016 - acc: 0.5192 - val_loss: 2.3953 - val_acc: 0.4003\n",
      "Epoch 87/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.6957 - acc: 0.5248 - val_loss: 2.3854 - val_acc: 0.3962\n",
      "Epoch 88/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 1.6912 - acc: 0.5259 - val_loss: 2.3832 - val_acc: 0.3984\n",
      "Epoch 89/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 1.6882 - acc: 0.5258 - val_loss: 2.3998 - val_acc: 0.3971\n",
      "Epoch 90/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 1.6859 - acc: 0.5249 - val_loss: 2.3847 - val_acc: 0.3980\n",
      "Epoch 91/150\n",
      "35000/35000 [==============================] - 8s 242us/step - loss: 1.6751 - acc: 0.5305 - val_loss: 2.4073 - val_acc: 0.3941\n",
      "Epoch 92/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 1.6728 - acc: 0.5296 - val_loss: 2.3855 - val_acc: 0.3995\n",
      "Epoch 93/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.6631 - acc: 0.5312 - val_loss: 2.3914 - val_acc: 0.3958\n",
      "Epoch 94/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.6628 - acc: 0.5344 - val_loss: 2.3998 - val_acc: 0.4015\n",
      "Epoch 95/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.6701 - acc: 0.5327 - val_loss: 2.3925 - val_acc: 0.4002\n",
      "Epoch 96/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 1.6436 - acc: 0.5386 - val_loss: 2.4078 - val_acc: 0.3955\n",
      "Epoch 97/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 1.6447 - acc: 0.5361 - val_loss: 2.3892 - val_acc: 0.3998\n",
      "Epoch 98/150\n",
      "35000/35000 [==============================] - 8s 239us/step - loss: 1.6426 - acc: 0.5349 - val_loss: 2.4273 - val_acc: 0.3946\n",
      "Epoch 99/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.6405 - acc: 0.5388 - val_loss: 2.4236 - val_acc: 0.3948\n",
      "Epoch 100/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.6440 - acc: 0.5361 - val_loss: 2.4607 - val_acc: 0.3946\n",
      "Epoch 101/150\n",
      "35000/35000 [==============================] - 8s 239us/step - loss: 1.6220 - acc: 0.5395 - val_loss: 2.4101 - val_acc: 0.3995\n",
      "Epoch 102/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.6204 - acc: 0.5418 - val_loss: 2.4160 - val_acc: 0.3901\n",
      "Epoch 103/150\n",
      "35000/35000 [==============================] - 8s 239us/step - loss: 1.6195 - acc: 0.5412 - val_loss: 2.4337 - val_acc: 0.3951\n",
      "Epoch 104/150\n",
      "35000/35000 [==============================] - 8s 239us/step - loss: 1.6091 - acc: 0.5446 - val_loss: 2.4542 - val_acc: 0.3913\n",
      "Epoch 105/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.6161 - acc: 0.5409 - val_loss: 2.4123 - val_acc: 0.3952\n",
      "Epoch 106/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 1.5935 - acc: 0.5475 - val_loss: 2.4270 - val_acc: 0.3962\n",
      "Epoch 107/150\n",
      "35000/35000 [==============================] - 8s 239us/step - loss: 1.5995 - acc: 0.5448 - val_loss: 2.4380 - val_acc: 0.3964\n",
      "Epoch 108/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.6015 - acc: 0.5433 - val_loss: 2.4278 - val_acc: 0.3966\n",
      "Epoch 109/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.5906 - acc: 0.5496 - val_loss: 2.4207 - val_acc: 0.3978\n",
      "Epoch 110/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.5897 - acc: 0.5501 - val_loss: 2.4551 - val_acc: 0.3949\n",
      "Epoch 111/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.5875 - acc: 0.5473 - val_loss: 2.4339 - val_acc: 0.4069\n",
      "Epoch 112/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.5858 - acc: 0.5502 - val_loss: 2.4247 - val_acc: 0.3938\n",
      "Epoch 113/150\n",
      "35000/35000 [==============================] - 8s 239us/step - loss: 1.5796 - acc: 0.5504 - val_loss: 2.4426 - val_acc: 0.3958\n",
      "Epoch 114/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 1.5728 - acc: 0.5527 - val_loss: 2.4504 - val_acc: 0.3945\n",
      "Epoch 115/150\n",
      "35000/35000 [==============================] - 8s 239us/step - loss: 1.5688 - acc: 0.5545 - val_loss: 2.4123 - val_acc: 0.3987\n",
      "Epoch 116/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 1.5739 - acc: 0.5520 - val_loss: 2.4254 - val_acc: 0.3959\n",
      "Epoch 117/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.5514 - acc: 0.5591 - val_loss: 2.4442 - val_acc: 0.4003\n",
      "Epoch 118/150\n",
      "35000/35000 [==============================] - 8s 242us/step - loss: 1.5648 - acc: 0.5523 - val_loss: 2.4539 - val_acc: 0.3962\n",
      "Epoch 119/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 1.5548 - acc: 0.5572 - val_loss: 2.4338 - val_acc: 0.3966\n",
      "Epoch 120/150\n",
      "35000/35000 [==============================] - 8s 239us/step - loss: 1.5575 - acc: 0.5563 - val_loss: 2.4744 - val_acc: 0.3923\n",
      "Epoch 121/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 1.5408 - acc: 0.5617 - val_loss: 2.4792 - val_acc: 0.3944\n",
      "Epoch 122/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.5350 - acc: 0.5641 - val_loss: 2.4645 - val_acc: 0.3977\n",
      "Epoch 123/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 1.5435 - acc: 0.5599 - val_loss: 2.4660 - val_acc: 0.3984\n",
      "Epoch 124/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.5351 - acc: 0.5620 - val_loss: 2.4857 - val_acc: 0.3897\n",
      "Epoch 125/150\n",
      "35000/35000 [==============================] - 8s 239us/step - loss: 1.5380 - acc: 0.5579 - val_loss: 2.4750 - val_acc: 0.3972\n",
      "Epoch 126/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.5272 - acc: 0.5656 - val_loss: 2.4848 - val_acc: 0.3939\n",
      "Epoch 127/150\n",
      "35000/35000 [==============================] - 8s 239us/step - loss: 1.5328 - acc: 0.5612 - val_loss: 2.4704 - val_acc: 0.3949\n",
      "Epoch 128/150\n",
      "35000/35000 [==============================] - 8s 239us/step - loss: 1.5325 - acc: 0.5611 - val_loss: 2.4729 - val_acc: 0.3981\n",
      "Epoch 129/150\n",
      "35000/35000 [==============================] - 8s 239us/step - loss: 1.5142 - acc: 0.5666 - val_loss: 2.4562 - val_acc: 0.3968\n",
      "Epoch 130/150\n",
      "35000/35000 [==============================] - 8s 239us/step - loss: 1.5311 - acc: 0.5640 - val_loss: 2.4566 - val_acc: 0.4001\n",
      "Epoch 131/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.5128 - acc: 0.5691 - val_loss: 2.4494 - val_acc: 0.3992\n",
      "Epoch 132/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.5151 - acc: 0.5658 - val_loss: 2.4776 - val_acc: 0.3901\n",
      "Epoch 133/150\n",
      "35000/35000 [==============================] - 8s 239us/step - loss: 1.5260 - acc: 0.5666 - val_loss: 2.4714 - val_acc: 0.3991\n",
      "Epoch 134/150\n",
      "35000/35000 [==============================] - 8s 239us/step - loss: 1.5046 - acc: 0.5681 - val_loss: 2.4736 - val_acc: 0.3961\n",
      "Epoch 135/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.5112 - acc: 0.5687 - val_loss: 2.4866 - val_acc: 0.3961\n",
      "Epoch 136/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.4991 - acc: 0.5704 - val_loss: 2.4793 - val_acc: 0.3949\n",
      "Epoch 137/150\n",
      "35000/35000 [==============================] - 8s 238us/step - loss: 1.5001 - acc: 0.5712 - val_loss: 2.4747 - val_acc: 0.3985\n",
      "Epoch 138/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.5072 - acc: 0.5693 - val_loss: 2.4772 - val_acc: 0.3975\n",
      "Epoch 139/150\n",
      "35000/35000 [==============================] - 8s 239us/step - loss: 1.4973 - acc: 0.5691 - val_loss: 2.4912 - val_acc: 0.3960\n",
      "Epoch 140/150\n",
      "35000/35000 [==============================] - 9s 255us/step - loss: 1.4920 - acc: 0.5727 - val_loss: 2.4927 - val_acc: 0.3974\n",
      "Epoch 141/150\n",
      "35000/35000 [==============================] - 8s 242us/step - loss: 1.4989 - acc: 0.5722 - val_loss: 2.4794 - val_acc: 0.4021\n",
      "Epoch 142/150\n",
      "35000/35000 [==============================] - 8s 242us/step - loss: 1.4816 - acc: 0.5767 - val_loss: 2.4928 - val_acc: 0.3968\n",
      "Epoch 143/150\n",
      "35000/35000 [==============================] - 9s 243us/step - loss: 1.4853 - acc: 0.5739 - val_loss: 2.4958 - val_acc: 0.3933\n",
      "Epoch 144/150\n",
      "35000/35000 [==============================] - 8s 242us/step - loss: 1.4770 - acc: 0.5737 - val_loss: 2.4810 - val_acc: 0.3992\n",
      "Epoch 145/150\n",
      "35000/35000 [==============================] - 8s 243us/step - loss: 1.4842 - acc: 0.5761 - val_loss: 2.5106 - val_acc: 0.3937\n",
      "Epoch 146/150\n",
      "35000/35000 [==============================] - 8s 243us/step - loss: 1.4821 - acc: 0.5744 - val_loss: 2.5056 - val_acc: 0.3904\n",
      "Epoch 147/150\n",
      "35000/35000 [==============================] - 8s 241us/step - loss: 1.4771 - acc: 0.5750 - val_loss: 2.4881 - val_acc: 0.3984\n",
      "Epoch 148/150\n",
      "35000/35000 [==============================] - 8s 242us/step - loss: 1.4680 - acc: 0.5812 - val_loss: 2.5064 - val_acc: 0.3976\n",
      "Epoch 149/150\n",
      "35000/35000 [==============================] - 8s 240us/step - loss: 1.4638 - acc: 0.5778 - val_loss: 2.5149 - val_acc: 0.3923\n",
      "Epoch 150/150\n",
      "35000/35000 [==============================] - 8s 242us/step - loss: 1.4684 - acc: 0.5762 - val_loss: 2.4837 - val_acc: 0.3954\n"
     ]
    }
   ],
   "source": [
    "history = model.fit(X_train, Y_train, validation_data=(X_test, Y_test), epochs=150, batch_size=200, verbose=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "OqS-A2tbnQVz"
   },
   "source": [
    "# 4. Plot the performance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 1000
    },
    "id": "VXOnHlOgnV6P",
    "outputId": "b40b3bed-bdeb-41cc-f27a-a6fb2c5f4b07"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'acc': [0.043314285895654135,\n",
       "  0.10148571420993124,\n",
       "  0.14394285742725646,\n",
       "  0.176942857333592,\n",
       "  0.20380000012261526,\n",
       "  0.22508571258613042,\n",
       "  0.24368571264403208,\n",
       "  0.25602857053279876,\n",
       "  0.2710571415083749,\n",
       "  0.2837999991859709,\n",
       "  0.2942571426289422,\n",
       "  0.3062571439572743,\n",
       "  0.3146285708461489,\n",
       "  0.3219142865283149,\n",
       "  0.33194285809993745,\n",
       "  0.3374000004359654,\n",
       "  0.34554285713604516,\n",
       "  0.35199999962534223,\n",
       "  0.3596285709312984,\n",
       "  0.36411428655896866,\n",
       "  0.3696285717827933,\n",
       "  0.372228570495333,\n",
       "  0.37214285816465104,\n",
       "  0.3799142857960292,\n",
       "  0.3821428576537541,\n",
       "  0.38462857059070044,\n",
       "  0.3944285718032292,\n",
       "  0.396742856332234,\n",
       "  0.40225714155605863,\n",
       "  0.40348571351596285,\n",
       "  0.40577142817633494,\n",
       "  0.4126857135977064,\n",
       "  0.4129999998637608,\n",
       "  0.4157999999182565,\n",
       "  0.42200000047683717,\n",
       "  0.42397142818995887,\n",
       "  0.4311999990258898,\n",
       "  0.4327428570815495,\n",
       "  0.4339142850467137,\n",
       "  0.4387714283806937,\n",
       "  0.44102857300213405,\n",
       "  0.44502857054982864,\n",
       "  0.440628571169717,\n",
       "  0.4480285714353834,\n",
       "  0.44280000056539265,\n",
       "  0.45317142844200137,\n",
       "  0.45117142694337026,\n",
       "  0.4558285708086831,\n",
       "  0.4603999999591282,\n",
       "  0.45925714203289575,\n",
       "  0.46360000031335014,\n",
       "  0.47008571437426977,\n",
       "  0.47171428424971446,\n",
       "  0.471742855821337,\n",
       "  0.47431428602763587,\n",
       "  0.47662857004574366,\n",
       "  0.47925714118140084,\n",
       "  0.4794857130731855,\n",
       "  0.47877142718860083,\n",
       "  0.4819428549494062,\n",
       "  0.4832857130255018,\n",
       "  0.4883714273997716,\n",
       "  0.49025714039802554,\n",
       "  0.4912571401255471,\n",
       "  0.4928285697528294,\n",
       "  0.49291428531919207,\n",
       "  0.49597142679350714,\n",
       "  0.4917999986239842,\n",
       "  0.5002285715511867,\n",
       "  0.4991714256150382,\n",
       "  0.5038571405410767,\n",
       "  0.5021714275223869,\n",
       "  0.5041999990599496,\n",
       "  0.5085142850875854,\n",
       "  0.505771427495139,\n",
       "  0.5075428548880986,\n",
       "  0.5109428560733795,\n",
       "  0.5113714259011405,\n",
       "  0.5139142828328269,\n",
       "  0.5165142859731402,\n",
       "  0.5158571406773158,\n",
       "  0.5188857121126992,\n",
       "  0.5172571405342647,\n",
       "  0.519742854493005,\n",
       "  0.5209999985354288,\n",
       "  0.519171427488327,\n",
       "  0.5247999984877451,\n",
       "  0.5259142848423549,\n",
       "  0.5257714256218502,\n",
       "  0.524942854813167,\n",
       "  0.5305428552627564,\n",
       "  0.5295999997002738,\n",
       "  0.531199996641704,\n",
       "  0.5344000004019056,\n",
       "  0.5327428564003536,\n",
       "  0.5385999994618552,\n",
       "  0.5361142829486302,\n",
       "  0.5349142851148333,\n",
       "  0.5387714307648795,\n",
       "  0.5360857166562761,\n",
       "  0.5395142836230141,\n",
       "  0.5418285698550088,\n",
       "  0.5411714271136693,\n",
       "  0.5446000008923667,\n",
       "  0.5409142855235508,\n",
       "  0.5475428567613874,\n",
       "  0.5448285741465432,\n",
       "  0.5433142856189183,\n",
       "  0.549599997145789,\n",
       "  0.5500857136930738,\n",
       "  0.5472571413857596,\n",
       "  0.5502285718917846,\n",
       "  0.5503714292390006,\n",
       "  0.5527142860208238,\n",
       "  0.55445714030947,\n",
       "  0.552000001158033,\n",
       "  0.5591142870698657,\n",
       "  0.5522857126167843,\n",
       "  0.5571999980722155,\n",
       "  0.5562857127189637,\n",
       "  0.5616571428094591,\n",
       "  0.5640571437563215,\n",
       "  0.5598857150759016,\n",
       "  0.5619714285646166,\n",
       "  0.5579142846379961,\n",
       "  0.565628571510315,\n",
       "  0.5611999978337969,\n",
       "  0.5610857144423893,\n",
       "  0.566600000006812,\n",
       "  0.5640000018051693,\n",
       "  0.5691428583008903,\n",
       "  0.5657714290278298,\n",
       "  0.5665714291163854,\n",
       "  0.5680857154301234,\n",
       "  0.568685714006424,\n",
       "  0.5704285705089569,\n",
       "  0.5712285678727286,\n",
       "  0.5693428540229797,\n",
       "  0.5690571440969194,\n",
       "  0.5726857158115932,\n",
       "  0.5721999997752053,\n",
       "  0.5767428585461207,\n",
       "  0.5739428571292332,\n",
       "  0.5736857153688159,\n",
       "  0.5760571444034577,\n",
       "  0.5743714279787882,\n",
       "  0.5749714274065835,\n",
       "  0.5812000017506735,\n",
       "  0.5778285712855202,\n",
       "  0.576171430179051],\n",
       " 'loss': [4.338766683850969,\n",
       "  3.8592524896349225,\n",
       "  3.5935626888275145,\n",
       "  3.4075786781311037,\n",
       "  3.25682049206325,\n",
       "  3.1497432490757533,\n",
       "  3.0534463242122105,\n",
       "  2.97796884400504,\n",
       "  2.9041671861921037,\n",
       "  2.8416313334873746,\n",
       "  2.780097955976214,\n",
       "  2.7230446311405725,\n",
       "  2.6856237397875105,\n",
       "  2.6444055366516115,\n",
       "  2.5967730958121162,\n",
       "  2.570779790878296,\n",
       "  2.5328608049665178,\n",
       "  2.502221132005964,\n",
       "  2.473613635471889,\n",
       "  2.4494252817971365,\n",
       "  2.424302964891706,\n",
       "  2.406090940747942,\n",
       "  2.3873796837670462,\n",
       "  2.363350075994219,\n",
       "  2.3422822652544295,\n",
       "  2.3269716480800082,\n",
       "  2.2973237235205515,\n",
       "  2.2752203587123327,\n",
       "  2.2594632993425643,\n",
       "  2.2422254725864956,\n",
       "  2.2319952392578126,\n",
       "  2.2078165163312637,\n",
       "  2.2046667643955775,\n",
       "  2.184435420036316,\n",
       "  2.160757863180978,\n",
       "  2.1453759738377163,\n",
       "  2.1221216222218104,\n",
       "  2.1233171776362827,\n",
       "  2.109502840042114,\n",
       "  2.0800159406661987,\n",
       "  2.0752711316517423,\n",
       "  2.051863097463335,\n",
       "  2.0598438256127496,\n",
       "  2.029342850957598,\n",
       "  2.0490145908083233,\n",
       "  2.0162280648095265,\n",
       "  2.014799100330898,\n",
       "  2.0009285954066685,\n",
       "  1.9899663427897862,\n",
       "  1.967094225202288,\n",
       "  1.9690499768938337,\n",
       "  1.9470453432628088,\n",
       "  1.9357037918908255,\n",
       "  1.9311125441959927,\n",
       "  1.9175469514301846,\n",
       "  1.9131185109274729,\n",
       "  1.9037425552095686,\n",
       "  1.8974343769890922,\n",
       "  1.8883454438618252,\n",
       "  1.8727786261694772,\n",
       "  1.8794354561397006,\n",
       "  1.8578098692212786,\n",
       "  1.843846583366394,\n",
       "  1.8330444199698313,\n",
       "  1.8395812892913819,\n",
       "  1.822406383923122,\n",
       "  1.8264439412525721,\n",
       "  1.8272337477547782,\n",
       "  1.7961809921264649,\n",
       "  1.8047945410864694,\n",
       "  1.785257319041661,\n",
       "  1.7957149028778077,\n",
       "  1.7767810692105974,\n",
       "  1.7775475222723824,\n",
       "  1.770680411883763,\n",
       "  1.76576954296657,\n",
       "  1.7508579983030046,\n",
       "  1.744327060835702,\n",
       "  1.7490253237315587,\n",
       "  1.7339329767227172,\n",
       "  1.7353643996374948,\n",
       "  1.7304917682920185,\n",
       "  1.7160818426949638,\n",
       "  1.7047508430480958,\n",
       "  1.7099956825801303,\n",
       "  1.701618629183088,\n",
       "  1.6957352924346925,\n",
       "  1.6911951528276716,\n",
       "  1.6882355764933994,\n",
       "  1.6858889859063284,\n",
       "  1.6750659554345266,\n",
       "  1.6728150340488979,\n",
       "  1.6631214584623064,\n",
       "  1.6628176144191198,\n",
       "  1.6700935384205409,\n",
       "  1.6436256034033638,\n",
       "  1.6447491897855486,\n",
       "  1.6425763811383929,\n",
       "  1.6404690544945852,\n",
       "  1.6439728580202375,\n",
       "  1.6219539192744663,\n",
       "  1.6204174423217774,\n",
       "  1.6195066520145962,\n",
       "  1.6090527282442366,\n",
       "  1.6161100455692836,\n",
       "  1.5935032606124877,\n",
       "  1.5995388398851667,\n",
       "  1.6015373870304652,\n",
       "  1.5905510234832763,\n",
       "  1.5896545389720371,\n",
       "  1.5874564218521119,\n",
       "  1.5857954250063215,\n",
       "  1.579612421308245,\n",
       "  1.5727974796295165,\n",
       "  1.5687688023703439,\n",
       "  1.5738669088908603,\n",
       "  1.5513739790235246,\n",
       "  1.5647602483204432,\n",
       "  1.554803284236363,\n",
       "  1.5575443015779769,\n",
       "  1.54083457333701,\n",
       "  1.535042051587786,\n",
       "  1.5434820079803466,\n",
       "  1.5351017802102225,\n",
       "  1.5380217068535942,\n",
       "  1.5271784142085485,\n",
       "  1.5328116737093245,\n",
       "  1.5325068398884365,\n",
       "  1.5141864102227347,\n",
       "  1.5310569899422781,\n",
       "  1.5127953999383108,\n",
       "  1.5151183557510377,\n",
       "  1.5260059717723302,\n",
       "  1.5046010187694006,\n",
       "  1.5111809641974312,\n",
       "  1.4990506294795445,\n",
       "  1.5001130608149937,\n",
       "  1.5071961641311646,\n",
       "  1.4973325245720999,\n",
       "  1.4920115389142719,\n",
       "  1.498929536002023,\n",
       "  1.481619438443865,\n",
       "  1.4853231246130807,\n",
       "  1.4770058863503592,\n",
       "  1.484224670955113,\n",
       "  1.4821236644472395,\n",
       "  1.4771349477767943,\n",
       "  1.4679661021913801,\n",
       "  1.463805764062064,\n",
       "  1.468374831335885],\n",
       " 'val_acc': [0.09660000026226044,\n",
       "  0.15890000015497208,\n",
       "  0.20129999965429307,\n",
       "  0.2387999987602234,\n",
       "  0.2612999975681305,\n",
       "  0.2765000000596046,\n",
       "  0.2895999997854233,\n",
       "  0.2996999987959862,\n",
       "  0.3062999987602234,\n",
       "  0.3193000018596649,\n",
       "  0.32759999990463257,\n",
       "  0.34079999804496763,\n",
       "  0.33549999833106997,\n",
       "  0.34680000245571135,\n",
       "  0.35710000157356264,\n",
       "  0.356100001335144,\n",
       "  0.35980000138282775,\n",
       "  0.35940000116825105,\n",
       "  0.36780000030994414,\n",
       "  0.3689000022411346,\n",
       "  0.3657000005245209,\n",
       "  0.376599999666214,\n",
       "  0.373400000333786,\n",
       "  0.3793000000715256,\n",
       "  0.3733999997377396,\n",
       "  0.3827999985218048,\n",
       "  0.3745999997854233,\n",
       "  0.3825,\n",
       "  0.3818999987840652,\n",
       "  0.3837999999523163,\n",
       "  0.38210000038146974,\n",
       "  0.39130000114440916,\n",
       "  0.3869000005722046,\n",
       "  0.3887000012397766,\n",
       "  0.39230000019073485,\n",
       "  0.39089999854564667,\n",
       "  0.38459999978542325,\n",
       "  0.38949999868869783,\n",
       "  0.39399999618530274,\n",
       "  0.4008999991416931,\n",
       "  0.40039999783039093,\n",
       "  0.39360000014305113,\n",
       "  0.3937999999523163,\n",
       "  0.3971000015735626,\n",
       "  0.39530000030994417,\n",
       "  0.39549999833106997,\n",
       "  0.38739999771118167,\n",
       "  0.4006000000238419,\n",
       "  0.3958000004291534,\n",
       "  0.39069999754428864,\n",
       "  0.39959999978542327,\n",
       "  0.39839999973773954,\n",
       "  0.39219999969005587,\n",
       "  0.4,\n",
       "  0.39589999854564667,\n",
       "  0.3955999994277954,\n",
       "  0.4,\n",
       "  0.39430000126361847,\n",
       "  0.3993000012636185,\n",
       "  0.39869999945163725,\n",
       "  0.3915000027418137,\n",
       "  0.39660000026226044,\n",
       "  0.3988000023365021,\n",
       "  0.39649999916553497,\n",
       "  0.4015999984741211,\n",
       "  0.39759999930858614,\n",
       "  0.3946000015735626,\n",
       "  0.40229999959468843,\n",
       "  0.39940000116825103,\n",
       "  0.39230000019073485,\n",
       "  0.4024999988079071,\n",
       "  0.3998999983072281,\n",
       "  0.39719999969005587,\n",
       "  0.39740000128746034,\n",
       "  0.39430000007152555,\n",
       "  0.400799999833107,\n",
       "  0.399899999499321,\n",
       "  0.3999000006914139,\n",
       "  0.3974999988079071,\n",
       "  0.3943999993801117,\n",
       "  0.399799998998642,\n",
       "  0.39749999821186066,\n",
       "  0.3976999992132187,\n",
       "  0.39680000305175783,\n",
       "  0.3994999986886978,\n",
       "  0.40030000030994417,\n",
       "  0.3961999988555908,\n",
       "  0.39839999794960024,\n",
       "  0.39709999978542326,\n",
       "  0.3979999989271164,\n",
       "  0.394099999666214,\n",
       "  0.3994999992847443,\n",
       "  0.39579999923706055,\n",
       "  0.40149999856948854,\n",
       "  0.4001999992132187,\n",
       "  0.3955000013113022,\n",
       "  0.3997999984025955,\n",
       "  0.3945999985933304,\n",
       "  0.3948000007867813,\n",
       "  0.3946000009775162,\n",
       "  0.39949999809265135,\n",
       "  0.3901000016927719,\n",
       "  0.39509999871253965,\n",
       "  0.39129999935626986,\n",
       "  0.3951999992132187,\n",
       "  0.39620000183582305,\n",
       "  0.3963999992609024,\n",
       "  0.3966000008583069,\n",
       "  0.39780000030994417,\n",
       "  0.3949000006914139,\n",
       "  0.40689999997615817,\n",
       "  0.3938000005483627,\n",
       "  0.39579999923706055,\n",
       "  0.39450000166893007,\n",
       "  0.39870000183582305,\n",
       "  0.39589999854564667,\n",
       "  0.4002999985218048,\n",
       "  0.39620000004768374,\n",
       "  0.3966000014543533,\n",
       "  0.3922999995946884,\n",
       "  0.3944000005722046,\n",
       "  0.39770000040531156,\n",
       "  0.39840000033378603,\n",
       "  0.3897000002861023,\n",
       "  0.39719999969005587,\n",
       "  0.3938999992609024,\n",
       "  0.3948999977111816,\n",
       "  0.3980999994277954,\n",
       "  0.3968000012636185,\n",
       "  0.40009999871253965,\n",
       "  0.39920000076293943,\n",
       "  0.3901000016927719,\n",
       "  0.3991000014543533,\n",
       "  0.39610000014305113,\n",
       "  0.3961000019311905,\n",
       "  0.3949000006914139,\n",
       "  0.398500000834465,\n",
       "  0.3975000023841858,\n",
       "  0.396000000834465,\n",
       "  0.39739999890327454,\n",
       "  0.4020999974012375,\n",
       "  0.39680000007152555,\n",
       "  0.3932999980449676,\n",
       "  0.39919999957084656,\n",
       "  0.3937000006437302,\n",
       "  0.39040000081062315,\n",
       "  0.39839999973773954,\n",
       "  0.39759999871253965,\n",
       "  0.3923000013828278,\n",
       "  0.39539999842643736],\n",
       " 'val_loss': [3.986970992088318,\n",
       "  3.6029069805145264,\n",
       "  3.3801504945755005,\n",
       "  3.2262374210357665,\n",
       "  3.064892430305481,\n",
       "  2.9883122968673708,\n",
       "  2.8905139064788816,\n",
       "  2.842545828819275,\n",
       "  2.780494031906128,\n",
       "  2.7474251794815063,\n",
       "  2.7070667695999147,\n",
       "  2.624639267921448,\n",
       "  2.6305639028549193,\n",
       "  2.5990951347351072,\n",
       "  2.542731204032898,\n",
       "  2.5289807510375977,\n",
       "  2.5299315547943113,\n",
       "  2.5272064638137817,\n",
       "  2.4737785339355467,\n",
       "  2.4809773349761963,\n",
       "  2.494754281044006,\n",
       "  2.448909516334534,\n",
       "  2.451329131126404,\n",
       "  2.416766633987427,\n",
       "  2.460357999801636,\n",
       "  2.417523503303528,\n",
       "  2.4564195680618286,\n",
       "  2.4148415899276734,\n",
       "  2.405515570640564,\n",
       "  2.3884424686431887,\n",
       "  2.411218867301941,\n",
       "  2.367688522338867,\n",
       "  2.3874737548828127,\n",
       "  2.3923077726364137,\n",
       "  2.362396602630615,\n",
       "  2.367335901260376,\n",
       "  2.3999851989746093,\n",
       "  2.3773967933654787,\n",
       "  2.3649623465538023,\n",
       "  2.356222357749939,\n",
       "  2.3413524770736696,\n",
       "  2.3545429706573486,\n",
       "  2.372682557106018,\n",
       "  2.348592176437378,\n",
       "  2.3732624673843383,\n",
       "  2.3576821088790894,\n",
       "  2.3942097663879394,\n",
       "  2.3549773263931275,\n",
       "  2.3649802017211914,\n",
       "  2.3758054065704344,\n",
       "  2.3410073471069337,\n",
       "  2.3416520023345946,\n",
       "  2.3678302907943727,\n",
       "  2.346007966995239,\n",
       "  2.353097929954529,\n",
       "  2.347121384143829,\n",
       "  2.3366755199432374,\n",
       "  2.3638392877578736,\n",
       "  2.349742169380188,\n",
       "  2.35477530002594,\n",
       "  2.385342631340027,\n",
       "  2.354685730934143,\n",
       "  2.3566407775878906,\n",
       "  2.353300304412842,\n",
       "  2.354178113937378,\n",
       "  2.3788746118545534,\n",
       "  2.3952711391448975,\n",
       "  2.36178530216217,\n",
       "  2.3771661520004272,\n",
       "  2.383466329574585,\n",
       "  2.3728191375732424,\n",
       "  2.3723943281173705,\n",
       "  2.3935772800445556,\n",
       "  2.3665035152435303,\n",
       "  2.373170428276062,\n",
       "  2.3514552545547485,\n",
       "  2.3637315702438353,\n",
       "  2.3722938776016234,\n",
       "  2.3713085508346556,\n",
       "  2.3831938457489015,\n",
       "  2.382637596130371,\n",
       "  2.3690213108062745,\n",
       "  2.3764523887634277,\n",
       "  2.38765341758728,\n",
       "  2.384237747192383,\n",
       "  2.395317540168762,\n",
       "  2.3853862190246584,\n",
       "  2.3832101917266844,\n",
       "  2.399822540283203,\n",
       "  2.3847218561172485,\n",
       "  2.4073392820358275,\n",
       "  2.3854942464828492,\n",
       "  2.3914389991760254,\n",
       "  2.3998101902008058,\n",
       "  2.3925049448013307,\n",
       "  2.407780966758728,\n",
       "  2.3891945695877075,\n",
       "  2.4272677278518677,\n",
       "  2.42355269908905,\n",
       "  2.460727558135986,\n",
       "  2.410112028121948,\n",
       "  2.415959806442261,\n",
       "  2.433730683326721,\n",
       "  2.45424485206604,\n",
       "  2.412322850227356,\n",
       "  2.4269680404663085,\n",
       "  2.4380426692962645,\n",
       "  2.4277776718139648,\n",
       "  2.420675072669983,\n",
       "  2.455080270767212,\n",
       "  2.4338896465301514,\n",
       "  2.4246929502487182,\n",
       "  2.442627778053284,\n",
       "  2.450378336906433,\n",
       "  2.41228018283844,\n",
       "  2.4253921270370484,\n",
       "  2.4441945648193357,\n",
       "  2.4538620710372925,\n",
       "  2.433817253112793,\n",
       "  2.4744163227081297,\n",
       "  2.4791871309280396,\n",
       "  2.464513273239136,\n",
       "  2.465978775024414,\n",
       "  2.4857235765457153,\n",
       "  2.4749646472930906,\n",
       "  2.484770169258118,\n",
       "  2.470418848991394,\n",
       "  2.4729389142990112,\n",
       "  2.45620313167572,\n",
       "  2.456598525047302,\n",
       "  2.449369287490845,\n",
       "  2.4776212310791017,\n",
       "  2.4714139127731323,\n",
       "  2.4735683679580687,\n",
       "  2.4865822172164918,\n",
       "  2.4793127250671385,\n",
       "  2.4746835088729857,\n",
       "  2.477233428955078,\n",
       "  2.491172742843628,\n",
       "  2.4926551246643065,\n",
       "  2.4794127225875853,\n",
       "  2.4928264236450195,\n",
       "  2.4958223867416383,\n",
       "  2.481039409637451,\n",
       "  2.5106150197982786,\n",
       "  2.5056432819366456,\n",
       "  2.488140425682068,\n",
       "  2.5064450025558473,\n",
       "  2.514942936897278,\n",
       "  2.4836502647399903]}"
      ]
     },
     "execution_count": 35,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "history.history"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "787kW39ungkm"
   },
   "source": [
    "\n",
    "### Plot the history\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 893
    },
    "id": "qJYMbbU-nffP",
    "outputId": "07c912c2-9142-459a-ef83-83c0ea453e7b"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA3gAAANsCAYAAAAJKQrDAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 1080x1080 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['acc'])\n",
    "plt.title('Training Performance')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.xlabel('Epochs')\n",
    "plt.legend(['Train'], loc='best')\n",
    "\n",
    "plt.draw()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 893
    },
    "id": "pW-6oinLnfAr",
    "outputId": "b145e062-1b9d-4610-b5aa-2f5d0a2e8154"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA34AAANsCAYAAAAEN3qEAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<Figure size 1080x1080 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(history.history['val_acc'])\n",
    "plt.title('Training Performance')\n",
    "plt.ylabel('Validation Accuracy')\n",
    "plt.xlabel('Epochs')\n",
    "plt.legend(['Train'], loc='best')\n",
    "\n",
    "plt.draw()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "QoIDj2Z6oRak"
   },
   "source": [
    "### Plot together"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 449
    },
    "id": "Ge8D-sSLoQyz",
    "outputId": "446f6633-ea9d-48dc-d4b6-956ad68b3b92"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA3gAAAGwCAYAAAD7Sse+AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n/OgRygjZAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 1080x1080 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig_size = plt.rcParams[\"figure.figsize\"]\n",
    "plt.rcParams[\"figure.figsize\"] = [15, 15]\n",
    "fig, axes = plt.subplots(nrows=1, ncols=1, sharex=True)\n",
    "\n",
    "plt.subplot(211)\n",
    "plt.plot(history.history['acc'])\n",
    "plt.plot(history.history['val_acc'])\n",
    "plt.title('Training Performance')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.xlabel('Epochs')\n",
    "plt.legend(['Train', 'Validation'], loc='best')\n",
    "\n",
    "plt.draw()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "KUkgNJVFvgXc"
   },
   "source": [
    "### Predict classes on validation data\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "AQ4EXEMZoQeS"
   },
   "outputs": [],
   "source": [
    "prediction_classes_val = model.predict_classes(X_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 126
    },
    "id": "Y-2a33b4vm-2",
    "outputId": "a44e2be2-546a-469b-9c28-386af0e0abac"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([72, 73, 84, 97, 40, 62, 88, 52, 10, 46, 97, 24, 71, 18, 14, 83, 95,\n",
       "       11, 48, 72, 86, 82, 10, 41, 70, 52, 25, 92, 90, 25, 88, 95, 56, 89,\n",
       "        9, 59, 33, 50, 42, 36, 41,  6, 83, 87,  2, 10, 58,  5, 94, 93, 47,\n",
       "        2, 14,  3, 79,  3, 40, 19, 90, 70, 14, 75, 40, 36, 36, 63,  3, 47,\n",
       "       20, 34, 98, 73, 84, 79,  0, 38, 90, 28, 65, 37, 89, 66, 54, 36, 46,\n",
       "       27, 15, 18, 95, 19, 10, 14, 45, 94,  8, 28,  4, 40, 80, 53])"
      ]
     },
     "execution_count": 40,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "prediction_classes_val[0:100]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "geTL9Zu4wJIK"
   },
   "source": [
    "#### Look at the correct validation values\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 126
    },
    "id": "irpBRVLpwFf2",
    "outputId": "7376fddb-5739-4cf3-d386-17d16fe78f05"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([49, 65, 40, 32, 35, 62, 74, 52, 13,  2, 74,  7, 71, 77, 44, 70, 73,\n",
       "       56, 48, 66, 86, 79, 10, 41, 92, 96, 25,  1, 90, 44, 26, 95, 56, 89,\n",
       "       20, 80, 74, 24, 42, 88, 41,  6, 83, 87,  2, 22, 58,  5, 94, 93, 59,\n",
       "       24, 19, 55, 31, 21, 10, 89, 90, 92, 14, 78, 22, 24,  5, 74, 38, 52,\n",
       "       20, 42, 13, 73,  5, 79,  0, 79, 90, 28, 35, 37, 89, 38, 54, 77, 35,\n",
       "       12, 97, 33, 95, 19, 83, 62, 26, 94,  8, 28, 55, 22, 67,  0])"
      ]
     },
     "execution_count": 41,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.argmax(Y_val,axis=1)[0:100]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "sdYM3OshwTTE"
   },
   "source": [
    "### The confusion matrix\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 145
    },
    "id": "H9EunIEVwFc5",
    "outputId": "6470c72d-dcac-42a8-baa1-87c37b8850e2"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[101,   1,   2, ...,   0,   0,   1],\n",
       "       [  0,  70,   0, ...,   0,   0,   0],\n",
       "       [  2,   1,  40, ...,   2,   7,   1],\n",
       "       ...,\n",
       "       [  0,   2,   0, ...,  51,   0,   0],\n",
       "       [  1,   1,  13, ...,   2,  35,   1],\n",
       "       [  0,   0,   1, ...,   2,   1,  42]])"
      ]
     },
     "execution_count": 42,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "confusion_matrix(np.argmax(Y_val,1), prediction_classes_val)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "AhCH-6M_wvLB"
   },
   "source": [
    "### Evaluate the model\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "P3EK3IARwlYB"
   },
   "outputs": [],
   "source": [
    "prediction_classes_test = model.predict_classes(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 145
    },
    "id": "dK4ElF07wld0",
    "outputId": "6708c421-2e8b-4175-c3d0-4ea8568b9975"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[69,  1,  0, ...,  0,  0,  0],\n",
       "       [ 1, 59,  1, ...,  1,  1,  0],\n",
       "       [ 1,  1, 32, ...,  0,  7,  0],\n",
       "       ...,\n",
       "       [ 0,  0,  0, ..., 31,  3,  0],\n",
       "       [ 0,  1,  5, ...,  1, 19,  2],\n",
       "       [ 0,  0,  0, ...,  1,  0, 34]])"
      ]
     },
     "execution_count": 44,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "confusion_matrix(np.argmax(Y_test,1), prediction_classes_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "id": "3eZe9RRVwllg",
    "outputId": "05f3f18c-76b9-48f0-b08b-5821898121ce"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.3954"
      ]
     },
     "execution_count": 45,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "accuracy_score(np.argmax(Y_test,1), prediction_classes_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "zqXbqpzTwljl"
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "name": "jérémie_bakambana_assign_3.ipynb",
   "provenance": []
  },
  "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": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}