Exercise_4_Multi_class_classifier.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "wYtuKeK0dImp"
   },
   "outputs": [],
   "source": [
    "# ATTENTION: Please do not alter any of the provided code in the exercise. Only add your own code where indicated\n",
    "# ATTENTION: Please do not add or remove any cells in the exercise. The grader will check specific cells based on the cell position.\n",
    "# ATTENTION: Please use the provided epoch values when training.\n",
    "\n",
    "import csv\n",
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
    "from os import getcwd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "4kxw-_rmcnVu"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(27455, 28, 28)\n",
      "(27455,)\n",
      "(7172, 28, 28)\n",
      "(7172,)\n"
     ]
    }
   ],
   "source": [
    "def get_data(filename):\n",
    "    with open(filename) as training_file:\n",
    "        reader = csv.reader(training_file, delimiter=',')    \n",
    "        imgs = []\n",
    "        labels = []\n",
    "\n",
    "        next(reader, None)\n",
    "        \n",
    "        for row in reader:\n",
    "            label = row[0]\n",
    "            data = row[1:]\n",
    "            img = np.array(data).reshape((28, 28))\n",
    "\n",
    "            imgs.append(img)\n",
    "            labels.append(label)\n",
    "\n",
    "        images = np.array(imgs).astype(float)\n",
    "        labels = np.array(labels).astype(float)\n",
    "            \n",
    "    return images, labels\n",
    "path_sign_mnist_train = f\"{getcwd()}/../tmp2/sign_mnist_train.csv\"\n",
    "path_sign_mnist_test = f\"{getcwd()}/../tmp2/sign_mnist_test.csv\"\n",
    "training_images, training_labels = get_data(path_sign_mnist_train)\n",
    "testing_images, testing_labels = get_data(path_sign_mnist_test)\n",
    "\n",
    "# Keep these\n",
    "print(training_images.shape)\n",
    "print(training_labels.shape)\n",
    "print(testing_images.shape)\n",
    "print(testing_labels.shape)\n",
    "\n",
    "# Their output should be:\n",
    "# (27455, 28, 28)\n",
    "# (27455,)\n",
    "# (7172, 28, 28)\n",
    "# (7172,)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "awoqRpyZdQkD"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(27455, 28, 28, 1)\n",
      "(7172, 28, 28, 1)\n"
     ]
    }
   ],
   "source": [
    "# In this section you will have to add another dimension to the data\n",
    "# So, for example, if your array is (10000, 28, 28)\n",
    "# You will need to make it (10000, 28, 28, 1)\n",
    "# Hint: np.expand_dims\n",
    "\n",
    "training_images = np.expand_dims(training_images, axis=3)\n",
    "testing_images = np.expand_dims(testing_images, axis=3)\n",
    "\n",
    "# Create an ImageDataGenerator and do Image Augmentation\n",
    "train_datagen = ImageDataGenerator(\n",
    "                                   rescale = 1.0/255.,\n",
    "                                   rotation_range=40,\n",
    "                                   width_shift_range=0.2,\n",
    "                                   height_shift_range=0.2,\n",
    "                                   shear_range=0.2,\n",
    "                                   zoom_range=0.2,\n",
    "                                   horizontal_flip=True,\n",
    "                                   fill_mode='nearest'\n",
    "    )\n",
    "\n",
    "validation_datagen = ImageDataGenerator(\n",
    "    rescale = 1.0/255.\n",
    "   )\n",
    "    \n",
    "# Keep These\n",
    "print(training_images.shape)\n",
    "print(testing_images.shape)\n",
    "    \n",
    "# Their output should be:\n",
    "# (27455, 28, 28, 1)\n",
    "# (7172, 28, 28, 1)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "Rmb7S32cgRqS"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/15\n",
      "429/429 [==============================] - 55s 127ms/step - loss: 2.7531 - accuracy: 0.1633 - val_loss: 2.0260 - val_accuracy: 0.3153\n",
      "Epoch 2/15\n",
      "429/429 [==============================] - 49s 113ms/step - loss: 2.1144 - accuracy: 0.3313 - val_loss: 1.3244 - val_accuracy: 0.5545\n",
      "Epoch 3/15\n",
      "429/429 [==============================] - 48s 113ms/step - loss: 1.7237 - accuracy: 0.4441 - val_loss: 1.0459 - val_accuracy: 0.6488\n",
      "Epoch 4/15\n",
      "429/429 [==============================] - 48s 112ms/step - loss: 1.4539 - accuracy: 0.5265 - val_loss: 0.8088 - val_accuracy: 0.7295\n",
      "Epoch 5/15\n",
      "429/429 [==============================] - 46s 108ms/step - loss: 1.2788 - accuracy: 0.5804 - val_loss: 0.6842 - val_accuracy: 0.7603\n",
      "Epoch 6/15\n",
      "429/429 [==============================] - 46s 106ms/step - loss: 1.1652 - accuracy: 0.6149 - val_loss: 0.5682 - val_accuracy: 0.7978\n",
      "Epoch 7/15\n",
      "429/429 [==============================] - 47s 109ms/step - loss: 1.0392 - accuracy: 0.6568 - val_loss: 0.3777 - val_accuracy: 0.8728\n",
      "Epoch 8/15\n",
      "429/429 [==============================] - 47s 110ms/step - loss: 0.9508 - accuracy: 0.6838 - val_loss: 0.3627 - val_accuracy: 0.8748\n",
      "Epoch 9/15\n",
      "429/429 [==============================] - 47s 109ms/step - loss: 0.8870 - accuracy: 0.7046 - val_loss: 0.3164 - val_accuracy: 0.8951\n",
      "Epoch 10/15\n",
      "429/429 [==============================] - 47s 109ms/step - loss: 0.8266 - accuracy: 0.7237 - val_loss: 0.3161 - val_accuracy: 0.8864\n",
      "Epoch 11/15\n",
      "429/429 [==============================] - 46s 107ms/step - loss: 0.7708 - accuracy: 0.7434 - val_loss: 0.2971 - val_accuracy: 0.8943\n",
      "Epoch 12/15\n",
      "429/429 [==============================] - 46s 107ms/step - loss: 0.7450 - accuracy: 0.7517 - val_loss: 0.2435 - val_accuracy: 0.9202\n",
      "Epoch 13/15\n",
      "429/429 [==============================] - 47s 109ms/step - loss: 0.7087 - accuracy: 0.7611 - val_loss: 0.2618 - val_accuracy: 0.9016\n",
      "Epoch 14/15\n",
      "429/429 [==============================] - 46s 108ms/step - loss: 0.6698 - accuracy: 0.7733 - val_loss: 0.2994 - val_accuracy: 0.8970\n",
      "Epoch 15/15\n",
      "429/429 [==============================] - 47s 109ms/step - loss: 0.6313 - accuracy: 0.7867 - val_loss: 0.2096 - val_accuracy: 0.9176\n"
     ]
    }
   ],
   "source": [
    "# Define the model\n",
    "# Use no more than 2 Conv2D and 2 MaxPooling2D\n",
    "model = tf.keras.models.Sequential([\n",
    "    tf.keras.layers.Conv2D(64, (3,3), activation='relu', input_shape=(28, 28, 1)),\n",
    "    tf.keras.layers.MaxPooling2D(2,2),\n",
    "    tf.keras.layers.Conv2D(128, (3,3), activation='relu'),\n",
    "    tf.keras.layers.MaxPooling2D(2,2), \n",
    "    tf.keras.layers.Flatten(),\n",
    "    tf.keras.layers.Dropout(0.5),\n",
    "    \n",
    "    tf.keras.layers.Dense(512, activation='relu'),\n",
    "    tf.keras.layers.Dense(26, activation='softmax')])\n",
    "    \n",
    "\n",
    "# Compile Model. \n",
    "model.compile(loss = 'sparse_categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
    "\n",
    "train_gen = train_datagen.flow(\n",
    "    training_images,\n",
    "    training_labels,\n",
    "    batch_size=64\n",
    ")\n",
    "\n",
    "val_gen = validation_datagen.flow(\n",
    "    testing_images,\n",
    "    testing_labels,\n",
    "    batch_size=64\n",
    ")\n",
    "\n",
    "# Train the Model\n",
    "history = model.fit_generator(\n",
    "    train_gen,\n",
    "    epochs=15,\n",
    "    validation_data=val_gen,\n",
    ")\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {},
    "colab_type": "code",
    "id": "_Q3Zpr46dsij"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAEICAYAAABPgw/pAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8QZhcZAAAgAElEQVR4nO3deXxU1f3/8dcnC0QIsiOyGQQVAgSIKYsUFREFVBBKERQUFFHUqri0Ea3ytVr9WUXAfQNrRZQiKBZcqGCVUkFA9kWQNYAsUfZFAuf3x2dChpBlkkxyZ/k8H495ZJabez+TwDtnzj33HHHOYYwxJvzFeF2AMcaY4LBAN8aYCGGBbowxEcIC3RhjIoQFujHGRAgLdGOMiRAW6CZPIhIrIgdEpEEwt/WSiDQWkaCP0xWRy0Vko9/jNSLSMZBti3GsN0VkRHG/v4D9PiEibwd7v6ZsxXldgAkOETng97ACcBQ47nt8m3NuQlH255w7DiQGe9to4Jy7IBj7EZEhwADn3KV++x4SjH2byGSBHiGccycD1dcCHOKc+3d+24tInHMuqyxqM8aUDetyiRK+j9QfiMhEEdkPDBCR9iLyrYjsEZHtIjJWROJ928eJiBORJN/jd32vfyoi+0XkfyLSsKjb+l7vJiI/iMheEXlBRP4rIoPyqTuQGm8TkXUi8ouIjPX73lgReV5EMkVkPdC1gJ/PwyLyfq7nXhKRUb77Q0Rkle/9/OhrPee3rwwRudR3v4KI/MNX2wrgwlzbPiIi6337XSEiPXzPtwBeBDr6urN2+/1sR/p9/+2+954pIh+JyNmB/GwKIyK9fPXsEZFZInKB32sjRGSbiOwTkdV+77WdiCzyPb9DRP4W6PFMkDjn7BZhN2AjcHmu554AfgWuQf+QnwH8BmiLflI7F/gBuMu3fRzggCTf43eB3UAaEA98ALxbjG1rAfuBnr7X7gOOAYPyeS+B1PgxUBlIAn7Ofu/AXcAKoB5QHfha/8nneZxzgQNARb997wTSfI+v8W0jwGXAYSDF99rlwEa/fWUAl/ruPwt8BVQFzgFW5tq2L3C273dyva+Gs3yvDQG+ylXnu8BI3/0rfDW2AhKAl4FZgfxs8nj/TwBv++439dVxme93NAJY47vfDNgE1PZt2xA413f/O6C/734loK3X/xei7WYt9Ogyxzn3iXPuhHPusHPuO+fcPOdclnNuPfA6cEkB3z/ZObfAOXcMmIAGSVG3vRpY7Jz72Pfa82j45ynAGp9yzu11zm1EwzP7WH2B551zGc65TODpAo6zHliO/qEB6AL84pxb4Hv9E+fceqdmAV8CeZ74zKUv8IRz7hfn3Ca01e1/3EnOue2+38l76B/jtAD2C3AD8KZzbrFz7giQDlwiIvX8tsnvZ1OQfsA059ws3+/oafSPQlsgC/3j0czXbbfB97MD/cN8nohUd87td87NC/B9mCCxQI8uW/wfiEgTEZkuIj+JyD7gcaBGAd//k9/9QxR8IjS/bev41+Gcc2iLNk8B1hjQsdCWZUHeA/r77l/ve5xdx9UiMk9EfhaRPWjruKCfVbazC6pBRAaJyBJf18YeoEmA+wV9fyf355zbB/wC1PXbpii/s/z2ewL9HdV1zq0B7kd/Dzt9XXi1fZsOBpKBNSIyX0S6B/g+TJBYoEeX3EP2XkNbpY2dc2cCj6JdCqVpO9oFAoCICKcGUG4lqXE7UN/vcWHDKicBl4tIXbSl/p6vxjOAycBTaHdIFeCLAOv4Kb8aRORc4BVgGFDdt9/VfvstbIjlNrQbJ3t/ldCuna0B1FWU/cagv7OtAM65d51zHdDullj054Jzbo1zrh/arfYc8KGIJJSwFlMEFujRrRKwFzgoIk2B28rgmP8CUkXkGhGJA+4BapZSjZOAe0WkrohUB/5U0MbOuZ+AOcDbwBrn3FrfS+WBcsAu4LiIXA10LkINI0Skiug4/bv8XktEQ3sX+rftVrSFnm0HUC/7JHAeJgK3iEiKiJRHg/Ub51y+n3iKUHMPEbnUd+wH0fMe80SkqYh08h3vsO92An0DA0Wkhq9Fv9f33k6UsBZTBBbo0e1+4Cb0P+tr6MnLUuWc2wFcB4wCMoFGwPfouPlg1/gK2te9DD1hNzmA73kPPcl5srvFObcHGA5MRU8s9kH/MAXiMfSTwkbgU+Adv/0uBV4A5vu2uQDw73eeCawFdoiIf9dJ9vd/hnZ9TPV9fwO0X71EnHMr0J/5K+gfm65AD19/enngGfS8x0/oJ4KHfd/aHVglOorqWeA659yvJa3HBE60C9MYb4hILPoRv49z7huv6zEmnFkL3ZQ5Eenq64IoD/wZHR0x3+OyjAl7FujGC78F1qMf568Eejnn8utyMcYEyLpcjDEmQlgL3RhjIoRnk3PVqFHDJSUleXV4Y4wJSwsXLtztnMtzqK9ngZ6UlMSCBQu8OrwxxoQlEcn3imfrcjHGmAhhgW6MMRHCAt0YYyKErVhkTIQ7duwYGRkZHDlyxOtSTBEkJCRQr1494uPzm8rndBboxkS4jIwMKlWqRFJSEjq5pQl1zjkyMzPJyMigYcOGhX+Dj3W5GBPhjhw5QvXq1S3Mw4iIUL169SJ/qrJANyYKWJiHn+L8zsIv0NesgXvvhWPHvK7EGGNCSvgF+rp1MGYMTA5kamtjjJcyMzNp1aoVrVq1onbt2tStW/fk419/DWyq9MGDB7NmzZoCt3nppZeYMGFCMErmt7/9LYsXLw7Kvspa+J0U7dYNzj8fnn8e+vUD+yhpTMiqXr36yXAcOXIkiYmJPPDAA6dsc3LF+pi825fjx48v9Dh33nlnyYuNAOHXQo+JgXvuge++g2+/9boaY0wxrFu3juTkZG644QaaNWvG9u3bGTp0KGlpaTRr1ozHH3/85LbZLeasrCyqVKlCeno6LVu2pH379uzcuROARx55hNGjR5/cPj09nTZt2nDBBRcwd+5cAA4ePMjvfvc7kpOT6dOnD2lpaQG3xA8fPsxNN91EixYtSE1N5euvvwZg2bJl/OY3v6FVq1akpKSwfv169u/fT7du3WjZsiXNmzdnchn2JoRfCx3gxhvh4Ye1ld6+vdfVGBM+7r0Xgt2d0KoV+MK0KFavXs0777xDWloaAE8//TTVqlUjKyuLTp060adPH5KTk0/5nr1793LJJZfw9NNPc9999zFu3DjS09NP27dzjvnz5zNt2jQef/xxPvvsM1544QVq167Nhx9+yJIlS0hNTQ241rFjx1K+fHmWLVvGihUr6N69O2vXruXll1/mgQce4LrrruPo0aM45/j4449JSkri008/PVlzWQm/FjpAYiLceit8+CFsyneeGmNMCGvUqNHJMAeYOHEiqamppKamsmrVKlauXHna95xxxhl069YNgAsvvJCNGzfmue/evXufts2cOXPo168fAC1btqRZs2YB1zpnzhwGDBgAQLNmzahTpw7r1q3joosu4oknnuCZZ55hy5YtJCQkkJKSwmeffUZ6ejr//e9/qVy5csDHKanwbKED3HUXjBoFL70EzzzjdTXGhIditKRLS8WKFU/eX7t2LWPGjGH+/PlUqVKFAQMG5DkGu1y5cifvx8bGkpWVlee+y5cvX+g2wTBw4EDat2/P9OnT6dq1K+PGjePiiy9mwYIFzJgxg/T0dLp168aIESNKrQZ/4dlCB2jQAHr3hjfegAMHvK7GGFMC+/bto1KlSpx55pls376dzz//POjH6NChA5MmTQK07zuvTwD56dix48lRNKtWrWL79u00btyY9evX07hxY+655x6uvvpqli5dytatW0lMTGTgwIHcf//9LFq0KOjvJT/h20IH7Q/85z/hnXfgjju8rsYYU0ypqakkJyfTpEkTzjnnHDp06BD0Y/zhD3/gxhtvJDk5+eQtv+6QK6+88uQcKh07dmTcuHHcdttttGjRgvj4eN555x3KlSvHe++9x8SJE4mPj6dOnTqMHDmSuXPnkp6eTkxMDOXKlePVV18N+nvJj2driqalpbkSL3DhHLRtC3v3wqpVOgLGGHOKVatW0bRpU6/L8FxWVhZZWVkkJCSwdu1arrjiCtauXUtcXOi2a/P63YnIQudcWl7bh+47CYSIttJvuAE++wy6d/e6ImNMiDpw4ACdO3cmKysL5xyvvfZaSId5cYT/u/n97+HBB/VkjwW6MSYfVapUYeHChV6XUarCv48iPl5HvMycCStWeF2NMcZ4JvwDHWDoUEhI0DlejDEmSkVGoFevrleP/uMfsHu319UYY4wnIiPQQed3OXIEXnvN60qMMcYTkRPoyclwxRV65WiA03IaY0pXp06dTrtIaPTo0QwbNqzA70tMTARg27Zt9OnTJ89tLr30Ugob+jx69GgOHTp08nH37t3Zs2dPIKUXaOTIkTz77LMl3k+wRU6ggw5h3L5dLzYyxniuf//+vP/++6c89/7779O/f/+Avr9OnTolmq0wd6DPmDGDKlWqFHt/oa7QQBeR+iIyW0RWisgKEbknj20uFZG9IrLYd3u0dMotxJVXQpMmOoTRowumjDE5+vTpw/Tp008uZrFx40a2bdtGx44dT44LT01NpUWLFnz88cenff/GjRtp3rw5oFPY9uvXj6ZNm9KrVy8OHz58crthw4adnHr3scceA3SGxG3bttGpUyc6deoEQFJSErt959lGjRpF8+bNad68+cmpdzdu3EjTpk259dZbadasGVdcccUpxylMXvs8ePAgV1111cnpdD/44AMA0tPTSU5OJiUl5bQ54osrkHHoWcD9zrlFIlIJWCgiM51zuSdC+MY5d3VQqiqu7LnShw2DuXOhFC4fNiaclfXsudWqVaNNmzZ8+umn9OzZk/fff5++ffsiIiQkJDB16lTOPPNMdu/eTbt27ejRo0e+a2m+8sorVKhQgVWrVrF06dJTpr998sknqVatGsePH6dz584sXbqUu+++m1GjRjF79mxq1Khxyr4WLlzI+PHjmTdvHs452rZtyyWXXELVqlVZu3YtEydO5I033qBv3758+OGHJ2daLEh++1y/fj116tRh+vTpgE6nm5mZydSpU1m9ejUiEpRuIAighe6c2+6cW+S7vx9YBdQNytFLw8CBULVqSM0qZ0w08+928e9ucc4xYsQIUlJSuPzyy9m6dSs7duzIdz9ff/31yWBNSUkhJSXl5GuTJk0iNTWV1q1bs2LFikIn3pozZw69evWiYsWKJCYm0rt3b7755hsAGjZsSKtWrYCCp+gNdJ8tWrRg5syZ/OlPf+Kbb76hcuXKVK5cmYSEBG655RamTJlChQoVAjpGYYp0paiIJAGtgXl5vNxeRJYA24AHnHOnXeUjIkOBoQANGjQoaq2BqVhRx6X/7W86V/o555TOcYwJQ160c3r27Mnw4cNZtGgRhw4d4sILLwRgwoQJ7Nq1i4ULFxIfH09SUlKeU+YWZsOGDTz77LN89913VK1alUGDBhVrP9myp94FnX63KF0ueTn//PNZtGgRM2bM4JFHHqFz5848+uijzJ8/ny+//JLJkyfz4osvMmvWrBIdB4pwUlREEoEPgXudc/tyvbwIOMc51xJ4Afgor3045153zqU559Jq1qxZ3JoLd+edOs/Liy+W3jGMMQFJTEykU6dO3HzzzaecDN27dy+1atUiPj6e2bNns6mQxWouvvhi3nvvPQCWL1/O0qVLAZ16t2LFilSuXJkdO3acXCkIoFKlSuzfv/+0fXXs2JGPPvqIQ4cOcfDgQaZOnUrHjh1L9D7z2+e2bduoUKECAwYM4MEHH2TRokUcOHCAvXv30r17d55//nmWLFlSomNnC6iFLiLxaJhPcM5Nyf26f8A752aIyMsiUsM5581VPvXrQ58+Olf6Y4/pCkfGGM/079+fXr16nTLi5YYbbuCaa66hRYsWpKWl0aRJkwL3MWzYMAYPHkzTpk1p2rTpyZZ+y5Ytad26NU2aNKF+/fqnTL07dOhQunbtSp06dZg9e/bJ51NTUxk0aBBt2rQBYMiQIbRu3Trg7hWAJ5544uSJT4CMjIw89/n555/z4IMPEhMTQ3x8PK+88gr79++nZ8+eHDlyBOcco0aNCvi4BSl0+lzRMxR/B352zt2bzza1gR3OOScibYDJaIs9350HZfrcgnz7ra43+sILOteLMVHKps8NX0WdPjeQLpcOwEDgMr9hid1F5HYRud23TR9gua8PfSzQr6AwLxPt2ulc6WPGwIkTnpZijDFlodAuF+fcHCDvcUQ527wIhF6H9fDh0K8fzJgBV3s7otIYY0pbZF0pmlvv3lCvng1hNFHP6w/MpuiK8zuL7EDPniv9yy9h2TKvqzHGEwkJCWRmZlqohxHnHJmZmSQkJBTp+8J7TdFA/PyzttKvvx7efLP0j2dMiDl27BgZGRklGpttyl5CQgL16tU7uVh1tshdUzQQ1arBTTfB+PHw1FNQmuPfjQlB8fHxNGzY0OsyTBmI7C6XbHffDUeP2lzpxpiIFh2B3rQpdO1qc6UbYyJadAQ66BDGn34C39SVxhgTaaIn0Lt00Zb688/bXOnGmIgUPYEuopNBf/89zJnjdTXGGBN00RPoAAMG6KgXu9DIGBOBoivQK1SA226Djz6CDRu8rsYYY4IqugId4I47dKk6myvdGBNhoi/Q69WD3/9erxrNY+J7Y4wJV9EX6KAnR/ftg7ff9roSY4wJmugM9DZt4KKLbK50Y0xEic5AB22l//gjTJ/udSXGGBMU0RvovXrp2qPPP+91JcYYExTRG+hxcfCHP8Ds2RCkFbeNMcZL0RvoAEOG6Nj0MWO8rsQYY0osugO9alUYNAgmTICdO72uxhhjSiS6Ax10rvRff4VXX/W6EmOMKREL9AsugKuugpdf1kUwjDEmTFmggw5h3LHD5ko3xoQ1C3SAzp2hWTP429/g2DGvqzHGmGKxQAedK/0vf4Hly2HkSK+rMcaYYrFAz9arF9x8Mzz1FPznP15XY4wxRWaB7m/MGGjcWBfC+Plnr6sxxpgisUD3l5gI772ni0nfdputPWqMCSsW6LmlpcETT8DkyTB+vNfVGGNMwMIy0HfvLuUDPPggXHaZzvXyww+lfDBjjAmOQgNdROqLyGwRWSkiK0Tknjy2EREZKyLrRGSpiKSWTrkwcSKcfTasX19aR0CXqHvnHUhIgOuv1ytJjTEmxAXSQs8C7nfOJQPtgDtFJDnXNt2A83y3ocArQa3Sz8UX65oUb75ZWkfwqVtXD7JwITz6aCkfzBhjSq7QQHfObXfOLfLd3w+sAurm2qwn8I5T3wJVROTsoFeL5uzVV8O4cWVwDVCvXjB0KDzzDMyaVcoHM8aYkilSH7qIJAGtgXm5XqoLbPF7nMHpoY+IDBWRBSKyYNeuXUWr1M/QoXql/iefFHsXgRs1Cs4/HwYOhMzMMjigMcYUT8CBLiKJwIfAvc65fcU5mHPudedcmnMurWbNmsXZBQBdu0K9evD668XeReAqVtSO+1274NZbbSijMSZkBRToIhKPhvkE59yUPDbZCtT3e1zP91ypiI3VtSm++AI2bCito/hp3VqvIJ06Fd54owwOaIwxRRfIKBcB3gJWOedG5bPZNOBG32iXdsBe59z2INZ5mptv1ilY3nqrNI/iZ/hw6NJFZ2ZcvbqMDmqMMYELpIXeARgIXCYii3237iJyu4jc7ttmBrAeWAe8AdxROuXmqF8funcvo5OjoEMZ335bl6y7/nqbO90YE3LiCtvAOTcHkEK2ccCdwSoqUEOHQo8eMH06XHttGRywTh39C9KzJzzyiE63a4wxISIsrxTN1q2bDmMsk5Oj2Xr0gGHD4NlnYebMMjywMcYULKwDPS5OT45+9hls2lSGB372WWjaFG66qQzmITDGmMCEdaCDBydHQfvRJ07Ucem33GJDGY0xISHsA71BA+16eestyMoqwwO3bAn/7//BtGnw6qtleGBjjMlb2Ac66MnRbdv05GiZuvtuuPJKuO8+WLmyjA9ujDGniohA795dB6CU6clRyBnKWKmSDmU8cqSMCzDGmBwREehxcdqV/emnsHlzGR+8dm1dCGPJEnjooTI+uDHG5IiIQAcNdCjjk6PZrroK7roLRo/WITfGGOOBiAn0c87RSbvK/ORotmeegebNYdAg2LnTgwKMMdEuYgId9OTo1q3a9VLmzjhDF5jeswcGD7ahjMaYMhdRgX7VVbo8XZmfHM3WooVOBzBjBrz0kkdFGGOiVUQFeny8Xmg0YwZs2VL49qXirrt02M0DD8Dy5R4VYYyJRhEV6JBz4ea4cR4VIKKjXqpUgf794fBhjwoxxkSbiAv0hg3hiit0fefjxz0qolYtHZ++fDn86U8eFWGMiTYRF+igJ0czMjweQdi1K9xzD7zwQhktfmqMiXYRGejXXANnneXhydFsTz+ty9f17Wvj040xpS4iAz375Oi//qXDGD2TkKALnzZtqotifPyxh8UYYyJdRAY66DzpJ054eHI0W40a8OWX2lLv0wcmTfK4IGNMpIrYQD/3XF3T2dOTo9mqVtWWevv2OvLlnXc8LsgYE4kiNtBBT45u3gyff+51JcCZZ+olrJddpisded7Bb4yJNBEd6D17hsjJ0WwVK+qIl6uugttugzFjvK7IGBNBIjrQ4+N1WhXPT476S0iAKVOgd2+4914dCWOMMUEQ0YEOenL0+HG9eDNklCsHH3ygi2I89BA89phN5mWMKbGID/RGjeDyy0Pk5Ki/uDg9OXrzzfD443pFqYW6MaYEIj7QQU+ObtoEM2d6XUkusbHwxhtwxx06S+Pdd+tYS2OMKYY4rwsoCz17Qs2aenK0a1evq8klJgZefFHnU3/uOV2X9NVXNeyNMaYIoiLQy5XTk6PPPQfbt+uc6SFFRFvoZ5wBTzyhoT5+vHbLGGNMgKKiywVC9OSoPxH4y1800N99Vy9A+vVXr6syxoSRqAn0887Ta3reeCPEu6kffhhGjYLJk3WqgCNHvK7IGBMmCg10ERknIjtFJM/ld0TkUhHZKyKLfbdHg19mcAwdChs3wr//7XUlhRg+HF5+WS9C6tEDDh3yuiJjTBgIpIX+NlDYqcRvnHOtfLfHS15W6bj2Wp0rK2SuHC3IsGE6s9i//61L2u3f73VFxpgQV2igO+e+Bn4ug1pKXfnyMGiQzmL7009eVxOAwYNhwgSYM0eXYdqzx+uKjDEhLFh96O1FZImIfCoizYK0z1Jx662QlaUrxIWF/v3hn/+EhQuhc2fIzPS6ImNMiApGoC8CznHOtQReAD7Kb0MRGSoiC0Rkwa5du4Jw6KI7/3y49NIwODnqr1cv+OgjWLFCi9+xw+uKjDEhqMSB7pzb55w74Ls/A4gXkRr5bPu6cy7NOZdWs2bNkh662IYOhfXrdd2JsNG9O0yfroVfckkIzTZmjAkVJQ50EaktIuK738a3z5DuF+jVC6pXD5OTo/46d9a1Sbdtg4svhpUrva7IGBNCAhm2OBH4H3CBiGSIyC0icruI3O7bpA+wXESWAGOBfs6F9ixTCQm6xsRHH4Vh70XHjjopzd690KoV/PnPcPiw11UZY0KAeJW9aWlpbsGCBZ4cG2D1al27+emndaLDsLNrF9x/P/zjH9C4Mbzyik4raYyJaCKy0DmXltdrUXOlaG5NmmivRVidHPVXs6ZOv5t9lVSXLjBwoAa9MSYqRW2gg64C9+OPMHu215WUQOfOsGyZdr188IH+pRo3zuZWNyYKRXWg9+4N1aqF4cnR3BISdJGMxYshORluuUWHN65a5XVlxpgyFNWBnn1ydOpU2LnT62qCIDkZ/vMfXZ5p2TJo2RIefdQm+DImSkR1oINeOXrsGPz9715XEiQxMdpCX70a+vbVKXlTUmDWLK8rM8aUsqgP9KZNdSTg669HWLdzrVo6r/oXX+hZ386d9eOInTQ1JmJFfaCDXjm6bh189ZXXlZSCLl20++Xhh+G99/Qv2NtvR9hfL2MMWKAD8LvfQdWqOpQ7ImUvbbd4sY6CGTwYOnWCNWu8rswYE0QW6GjeDR2qkxo+95zX1ZSiZs3g66+1f2nJEu1bHzkSjh71ujJjTBBYoPv85S/w+9/DAw/As896XU0pionRM8GrVulHk//7Pw32iOxvMia6WKD7xMdrF3PfvvDgg/C3v3ldUSmrXVvf8Gef6TCfTp20K2b3bq8rM8YUkwW6n7g4XSDouuvgj3+EZ57xuqIycOWVsHw5pKfrqJjzz9e/aGvXel2ZMaaILNBziYvTXOvXTyftevpprysqAxUqwFNPwaJF2lJ//nkN9i5d4MMPtQVvjAl5Fuh5iIvTSQyvvx4eekizLiq0aKEBvnmzTiWwejX06QMNGuhcMZs3e12hMaYAFuj5iIvTq0evvx5GjIC//tXrispQnToa4Bs2wLRpkJoKTz4JDRtCjx4wYwYcP+51lcaYXCzQCxAXpzPU3nCDXpfz5JNeV1TG4uLgmmtylr5LT4f58+Gqq6BRI/0r99NPXldpjPGxQC9EbKy21AcMgEce0etzolJSkv5F27wZJk3SQH/4YahfX4cGzZplV58a4zEL9ADExurV8gMHak/E4497XZGHypXTAftffql97HffrYtsdO6sV6GOGgWZIb2krDERywI9QLGxMH68zm/12GN6PU7Uu+ACvbR261btm6pRQ5fFq1sXbrwR5s61VrsxZcgCvQhiY+Gtt2DQIL1ifuRIjwsKFWecoR9f/vtfnVLgllt0Be4OHXRO9pdfhn37vK7SmIhngV5EsbG6fsTgwdpKt1DPJSUFXnoJtm3TOWPi4uDOO7XVft99sGmT1xUaE7Es0Ishd6g/9pj1LJwmMVHnjFm4EObN0+GOY8fqydR+/eC777yu0JiIY4FeTDExGuq33KInSR991EI9TyLQpo3OqbB+Pdx7r45jb9MGLrlEx7mfOOF1lcZEBAv0EoiJ0V6FIUN0OOOf/2yhXqAGDXQqy4wMPZm6cSP07KmLbrz2Ghw+7HWFxoQ1C/QSionRLLr1Vh2m/cgjFuqFOvNM7U9ft05nfKxUCW6/XQP/scciZMVuY8qeBXoQxMTAq6/qIhl//ateb2OhHoD4eOjfX/vTv1zN6hUAABMqSURBVPoK2rXT/qsGDfSHuXq11xUaE1Ys0IMkJkaXsLvtNp3Ma8QIC/WAiWh/+ief6MIbN92k49qbNtWpB776yn6YxgTAAj2IYmJ0yPXtt+u0u+nplkNF1qSJ9mFt3qzdL99+q1P6pqVp94xN5WtMvizQgyw71IcN0wUy/vQnC/ViqVVLB/lv3qwBf+CAzpLWqJGeULULlYw5jQV6KRDRa2vuuEOXsrvnHpttttiyV/BetUqHOJ57ri78Wq+efv3hB68rNCZkFBroIjJORHaKyPJ8XhcRGSsi60RkqYikBr/M8CMCL76ogzleeAGuvRb27/e6qjAWE5PTn/7ddzqF7+jROp9Ms2Y6vGjhQvs4ZKJaIC30t4GuBbzeDTjPdxsKvFLysiKDiPYOvPwyfPop/Pa3sGWL11VFgLQ0mDhRx7GPGaPdM089pc8nJelHoq++gqwsjws1pmwVGujOua+BnwvYpCfwjlPfAlVE5OxgFRgJhg3TNSI2btQLJO2q9yCpV0+n7509G3bsgHHjdDKw117TE6m1a+v8DNOm2UVLJioEow+9LuDf7szwPXcaERkqIgtEZMGuXbuCcOjwceWVOptsQoKO0Js82euKIkyNGjnhvXs3/POf+kOfOlWvRq1ZU9dHnTAB9uzxulpjSkWZnhR1zr3unEtzzqXVrFmzLA8dEpo103mqWrXSNSKeesq6fEtFYmJOeO/cCZ9/njO974ABGu5XXqlXg23f7nW1xgRNMAJ9K1Df73E933MmD7Vq6Wpt/fvrxUeDB8Ovv3pdVQQrVw6uuEKv+tq6VT8mDR+uE4UNG6bT+l50kY4xXbvW62qNKZFgBPo04EbfaJd2wF7nnDV7CpCQoI3HkSN1vdIuXWzVtjIREwPt22t4//ADLFumUw0cPaoXDJx/PjRvrrOszZoFBw96XbExRSKukM/8IjIRuBSoAewAHgPiAZxzr4qIAC+iI2EOAYOdcwsKO3BaWppbsKDQzSLexInaSq9fH/71Lx2FZzywaZOusjR1KnzzjU7pGxurJ1k7dNBWfIcO+osyxkMistA5l5bna4UFemmxQM8xd66OUz92DKZM0QEaxkO//KJTDsydq/3u8+bBoUP6Wr16OeF+0UUa+PHx3tZroooFehjYsAGuvlp7Al59VRfOMCEiK0vXSp07Nyfksy8oqFBBx6Jmh3y7dlCtmrf1mohmgR4m9u6Fvn3hiy/gwQd1gq8Ym5whNG3ZcmrAL16cM79DcrIGfHbIn3eeXmVmTBBYoIeRrCy90PHll7Ub5t13oWJFr6syhTp4EObPzwn5uXNzxrvXqJET8G3b6hWtiYne1mvClgV6mHFO538ZPlzHrE+bpqPrTBg5cUIX6Pjvf3MCPnsisZgYHU3Ttq3e2rXTaYNjY72t2YQFC/QwNX069OunK7Z98gmk2rRn4W33bm3Fz5unJ13nz89pxVeqBL/5TU7At20LZ53lbb0mJFmgh7GlS/VkaWamru/Qs6fXFZmgOXFCL2bKDvh58/QXnj2p2Dnn5IR727b6Fz0hwduajecs0MPc9u0a5AsW6DUx999v59gi1uHDsGhRTsDPm6eLfADExWkfXHbAt21rJ1yjkAV6BDh0SJfanDwZhgzRk6Y2/DlKbN+eE+7z5ul0nQcO6GvVqmlXzYUX5twaNLCQj2AW6BHixAm9Kv2vf4XLLoO337YLF6PS8eOwcuWpAb9iRU5XTfXqpwb8hRdq942FfESwQI8wf/+7rsp2/LiOW7/vPh0JZ6LYkSPa/75wYc5t+fJTQz419dSQT0qykA9DFugRaNMmGDsW3nhDl7br2FGD/ZprbPSb8TlyRCcgyw74BQtODflq1U4P+YYNLeRDnAV6BNu3D956S1di27QJGjWCe++FQYPs2hWTh9whn92SP3ZMX69aNSfkGzfWoZO1a+vtrLOgfHlv6zcW6NEgK0snCnzuOe1WrVIFbrsN7rpL55MyJl9Hj54e8suW5YS8vypVcsLd/2vu+zVr2ln7UmKBHmX+9z8YNUpnboyJgeuu0+4YuzDJBOzXX3Wd1p9+0lv2/bye278/733UqJF34CcladdOw4ba7WNdPEVigR6lNmzQfvY339RRbpdcosF+9dU26ZcJokOHCg58/+dyL9ZdqRKce25OwPvfT0rS2SzNKSzQo9zevRrqY8boJIHnnaf97DfdZBN/mTLknJ702bhRWxsbNuhSgNn3N2w4PfBr18477Bs21L7EuDhP3oqXLNANkLOAxnPP6dDlqlXh9tu1n71OHa+rM1HPOW3N+we8f+Bv2ZIzRTFomDdokBP2jRrpidxGjfRWqZJ376UUWaCbUzink/+NGqUnUuPidBKw++7TK8uNCUnHjkFGxumt+vXr9bZr16nb16qVE/C5v1avHrZ99xboJl8//qj97G+9pVN6d+qkLfYePaLy06wJZ/v2abCvW6f/sNety7mfkaEtmWyVK+cd9I0bw9lnh/RJJgt0U6hfftF+9hde0E+2devqsMchQ/TftzFh7cgRbc37h332140bcy62Ap3RMrvbxr8Lp3Fj7eLxeDimBboJWFYWzJgBL72kS+HFxUHv3nDHHXDxxWH7KdWY/GVl6YyWuYN+3Tpt8fufqI2N1Xlxcod99q0MRuVYoJtiWbtWF6weN07XYWjWTIN9wABddMOYiOecznbpH/bZt3Xr9KOtv7PPzjvsGzcO2uLhFuimRA4dgg8+0Fb7woU6pcDAgRruzZt7XZ0xHvrll1MD3v/+tm2nblulSk7A//730KdPsQ5pgW6CZv58nYv9/ff1ivGOHeHOO6FXLyhXzuvqjAkhhw5pv31erfubb4aHHirWbi3QTdBlZsL48fDKK9rNeNZZcOutOq2vzdFuTOkpKNBDd2yOCWnVq8MDD2g/+4wZumjOk0/q1dq9e8O//60Lchhjyo4FuimRmBjo1g0++URb6n/8I3zzDXTpAk2bwujRp583MsaUDgt0EzRJSfDUU3oNx7vvait++HAd03733flPymeMCQ4LdBN05cvDDTfo9AKLFkH//jpCJiUF/vMfr6szJnJZoJtS1bq1TivwzTd6kdKll2qrPfekesaYkgso0EWkq4isEZF1IpKex+uDRGSXiCz23YYEv1QTzi66CBYv1nliRo/WoJ8/3+uqjIkshQa6iMQCLwHdgGSgv4gk57HpB865Vr7bm0Gu00SAihV1rpiZM3WIbvv28MgjujiOMabkAmmhtwHWOefWO+d+Bd4HepZuWSaSXX65Lll500061LFNG1i61OuqjAl/gQR6XWCL3+MM33O5/U5ElorIZBHJ89ISERkqIgtEZMGu3HMXm6hSubLOEfPxx7oyWVqajpDxn/TOGFM0wTop+gmQ5JxLAWYCf89rI+fc6865NOdcWs2aNYN0aBPOevSA5cvh2mthxAj47W9hzRqvqzImPAUS6FsB/xZ3Pd9zJznnMp1zR30P3wQuDE55JhrUqAGTJun8MGvX6gnTsWPtSlNjiiqQQP8OOE9EGopIOaAfMM1/AxHxXwKhB7AqeCWaaHHdddpav+wyuOce6NxZ1x4wxgSm0EB3zmUBdwGfo0E9yTm3QkQeF5Eevs3uFpEVIrIEuBsYVFoFm8h29tk6jcBbb+lUvSkpet+jOeSMCSs226IJWRs36iyjs2dD9+7wxhtQp47XVRnjLZtt0YSlpCSdtXHsWA315s21n91a68bkzQLdhLSYGPjDH/Qq0wsu0HlhrrsOdu/2ujJjQo8FugkL55+v88E89RR89JGubzptWuHfZ0w0ifO6AGMCFRcH6enan37jjdCzp3bLnHnmqbdKlQq/n/24UiXdrzGRwP4pm7CTkqITe40dq1MG7Nunc63v3KlLN+7fr88dPBjY/ipUODXsK1fWdXybNYPkZL3VqQMipfu+jCkpG+ViItbx43DggIZ79i077Au6/8sverVqZmbOvipX1mD3D/lmzXTxDgt6U5YKGuViLXQTsWJjNYgrVy7e9+/cCStXwooV+nXlSp175k2/uUTPPDMn4P0Dv359C3pT9qyFbkwR7dqVE/D+gb9jR842iYmnh3yjRqf31/uHfu4/AIG+Vr48VK0K5cqV7H2Z8FBQC90C3Zggycw8PeRXrNDZJMtCxYoa7NWq6ddA7lerpp9gYmPLpkZTctblYkwZqF4dOnbUm7+ff4ZVq2DDhpyLonK3owp6XNi2hw9rv/8vv+ixsu+vW5dz/9ChgmuvXPn0wG/cWCdKa90azj1XrwmIRM7pH+Mff4T163O+rl+vJ8t79dJZQWvU8LrSwlkL3ZgocPRo3qGffT/3c5mZGmjZ89NXqgStWuUEfOvW2o0UH+/t+wrUsWOweXPeof3jj3pS3F+dOtCwIWzdqlNQxMTAJZdA794a8HXzWhGijFiXizGmyI4e1S6jRYvg++/1tmRJTmu/XDlo0eLUkE9J0a4fL+zZk3dYr1+vYX78eM625ctrYDdqpJ8+/L8mJelQVtDW+5Il8OGHMGWKdqMBtGuXE+6NG5ft+7RAN8YExfHjOmd9dsB//70G/s8/6+sxMXpVr3/It26t3VGBcC5n6OiePTmfGnI/zv3azp361V/NmqeHdfbXs88uXhfS6tUwdaoG/MKF+lxKioZ7794631Bpj26yQDfGlBrnYMuWU0P+++/1uWz160NqqgbesWP5B/aePQUvbBITA1Wq5PT1V62qj2vUODW0GzbUIaWladMmDfcpU2DOHP05NG4Mv/udhvtvflM64W6Bbowpc7t366Rq/l02P/yg/e7+YZw7nP0f536uUqXQHN+/Y4fOMTRlCsyapece6tXTLpnevXVpxWBNMWGBbowJCVlZOkQyFEM5WH75Bf71L+2W+fxzOHJEP0H07Kmt98su0z784rL50I0xISEuLrLDHPSTxMCB2mLftQv++U/o0kXXze3eHWrVgueeK51j2zh0Y4wpJYmJ0KeP3o4ehS+/1JZ7/fqlczwLdGOMKQPly2sLvXv30juGdbkYY0yEsEA3xpgIYYFujDERwgLdGGMihAW6McZECAt0Y4yJEBboxhgTISzQjTEmQng2l4uI7AI2FfPbawC7g1hOaQunesOpVgivesOpVgivesOpVihZvec452rm9YJngV4SIrIgv8lpQlE41RtOtUJ41RtOtUJ41RtOtULp1WtdLsYYEyEs0I0xJkKEa6C/7nUBRRRO9YZTrRBe9YZTrRBe9YZTrVBK9YZlH7oxxpjThWsL3RhjTC4W6MYYEyHCLtBFpKuIrBGRdSKS7nU9+RGR+iIyW0RWisgKEbnH65oCISKxIvK9iPzL61oKIiJVRGSyiKwWkVUi0t7rmgoiIsN9/w6Wi8hEEUnwuiZ/IjJORHaKyHK/56qJyEwRWev7WtXLGrPlU+vffP8WlorIVBGp4mWN/vKq1++1+0XEiUiNYBwrrAJdRGKBl4BuQDLQX0SSva0qX1nA/c65ZKAdcGcI1+rvHmCV10UEYAzwmXOuCdCSEK5ZROoCdwNpzrnmQCzQz9uqTvM20DXXc+nAl86584AvfY9DwducXutMoLlzLgX4AXiorIsqwNucXi8iUh+4AtgcrAOFVaADbYB1zrn1zrlfgfeBnh7XlCfn3Hbn3CLf/f1o4NT1tqqCiUg94CrgTa9rKYiIVAYuBt4CcM796pzb421VhYoDzhCROKACsM3jek7hnPsa+DnX0z2Bv/vu/x24tkyLykdetTrnvnDOZfkefgvUK/PC8pHPzxbgeeCPQNBGpoRboNcFtvg9ziDEQxJARJKA1sA8bysp1Gj0H9gJrwspRENgFzDe1z30pohU9Lqo/DjntgLPoi2x7cBe59wX3lYVkLOcc9t9938CzvKymCK4GfjU6yIKIiI9ga3OuSXB3G+4BXrYEZFE4EPgXufcPq/ryY+IXA3sdM4t9LqWAMQBqcArzrnWwEFCpzvgNL6+557oH6I6QEURGeBtVUXjdHxzyI9xFpGH0e7OCV7Xkh8RqQCMAB4N9r7DLdC3AvX9HtfzPReSRCQeDfMJzrkpXtdTiA5ADxHZiHZlXSYi73pbUr4ygAznXPYnnslowIeqy4ENzrldzrljwBTgIo9rCsQOETkbwPd1p8f1FEhEBgFXAze40L7AphH6x32J7/9bPWCRiNQu6Y7DLdC/A84TkYYiUg49sTTN45ryJCKC9vGucs6N8rqewjjnHnLO1XPOJaE/11nOuZBsRTrnfgK2iMgFvqc6Ays9LKkwm4F2IlLB9++iMyF8EtfPNOAm3/2bgI89rKVAItIV7S7s4Zw75HU9BXHOLXPO1XLOJfn+v2UAqb5/1yUSVoHuO+lxF/A5+h9iknNuhbdV5asDMBBt6S723bp7XVQE+QMwQUSWAq2Av3pcT758nyQmA4uAZej/u5C6VF1EJgL/Ay4QkQwRuQV4GugiImvRTxlPe1ljtnxqfRGoBMz0/V971dMi/eRTb+kcK7Q/mRhjjAlUWLXQjTHG5M8C3RhjIoQFujHGRAgLdGOMiRAW6MYYEyEs0I0xJkJYoBtjTIT4/ywF/QJKAK4LAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot the chart for accuracy and loss on both training and validation\n",
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "acc = history.history['accuracy']\n",
    "val_acc = history.history['val_accuracy']\n",
    "loss =  history.history['loss']\n",
    "val_loss =  history.history['val_loss']\n",
    "\n",
    "epochs = range(len(acc))\n",
    "\n",
    "plt.plot(epochs, acc, 'r', label='Training accuracy')\n",
    "plt.plot(epochs, val_acc, 'b', label='Validation accuracy')\n",
    "plt.title('Training and validation accuracy')\n",
    "plt.legend()\n",
    "plt.figure()\n",
    "\n",
    "plt.plot(epochs, loss, 'r', label='Training Loss')\n",
    "plt.plot(epochs, val_loss, 'b', label='Validation Loss')\n",
    "plt.title('Training and validation loss')\n",
    "plt.legend()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Submission Instructions"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Now click the 'Submit Assignment' button above."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# When you're done or would like to take a break, please run the two cells below to save your work and close the Notebook. This will free up resources for your fellow learners. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%javascript\n",
    "<!-- Save the notebook -->\n",
    "IPython.notebook.save_checkpoint();"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%javascript\n",
    "IPython.notebook.session.delete();\n",
    "window.onbeforeunload = null\n",
    "setTimeout(function() { window.close(); }, 1000);"
   ]
  }
 ],
 "metadata": {
  "colab": {
   "name": "Exercise 8 - Question.ipynb",
   "provenance": []
  },
  "coursera": {
   "course_slug": "convolutional-neural-networks-tensorflow",
   "graded_item_id": "8mIh8",
   "launcher_item_id": "gg95t"
  },
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}