06_batchnorm.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<div style=\"direction:rtl;text-align:right;font-family:Tahoma\">\n",
    "اگر روی گوگل کولب اجرا میکنید این خطوط را از حالت کامنت خارج نمائید.\n",
    "</div>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "#!wget https://raw.githubusercontent.com/Alireza-Akhavan/SRU-deeplearning-workshop/master/dataset.py\n",
    "#!mkdir dataset\n",
    "#!wget https://github.com/Alireza-Akhavan/SRU-deeplearning-workshop/raw/master/dataset/Data_hoda_full.mat -P dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 1. Import libraries and modules\n",
    "from tensorflow import keras\n",
    "from keras import layers\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Dropout, BatchNormalization\n",
    "import numpy as np\n",
    "from dataset import load_hoda\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "# Load pre-shuffled HODA data into train and test sets\n",
    "X_train, y_train, X_test, y_test = load_hoda(\n",
    "                                            training_sample_size=3500,\n",
    "                                            test_sample_size=400,size=28)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Preprocess input data\n",
    "'''standard our data values to the range [-1, 1]'''\n",
    "X_train = X_train.astype('float32')\n",
    "X_test = X_test.astype('float32')\n",
    "X_train = (X_train - 127.5) / 127.5\n",
    "X_test = (X_test - 127.5) / 127.5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Reshape to original image shape (n x 784)  ==> (n x 28 x 28 x 1)\n",
    "X_train = X_train.reshape(-1,28,28,1)\n",
    "X_test = X_test.reshape(-1,28,28,1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 5. Define model architecture\n",
    "model = Sequential()\n",
    "model.add(layers.Conv2D(32, (3, 3), activation='relu',\n",
    "                        input_shape=(28, 28, 1)))\n",
    "model.add(BatchNormalization())\n",
    "model.add(layers.MaxPooling2D((2, 2)))\n",
    "model.add(layers.Conv2D(64, (3, 3), activation='relu'))\n",
    "model.add(BatchNormalization())\n",
    "model.add(layers.MaxPooling2D((2, 2)))\n",
    "model.add(layers.Conv2D(64, (3, 3), activation='relu'))\n",
    "model.add(BatchNormalization())\n",
    "model.add(layers.Flatten())\n",
    "model.add(layers.Dense(64, activation='relu'))\n",
    "model.add(BatchNormalization())\n",
    "model.add(Dropout(0.5))\n",
    "model.add(layers.Dense(10, activation='softmax'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/200\n",
      "11/11 [==============================] - 35s 296ms/step - loss: 2.0855 - sparse_categorical_accuracy: 0.3933 - val_loss: 2.0395 - val_sparse_categorical_accuracy: 0.7671\n",
      "Epoch 2/200\n",
      "11/11 [==============================] - 2s 191ms/step - loss: 0.5649 - sparse_categorical_accuracy: 0.8258 - val_loss: 1.8229 - val_sparse_categorical_accuracy: 0.7729\n",
      "Epoch 3/200\n",
      "11/11 [==============================] - 2s 188ms/step - loss: 0.4052 - sparse_categorical_accuracy: 0.8676 - val_loss: 1.8079 - val_sparse_categorical_accuracy: 0.3757\n",
      "Epoch 4/200\n",
      "11/11 [==============================] - 2s 182ms/step - loss: 0.2747 - sparse_categorical_accuracy: 0.9229 - val_loss: 1.9082 - val_sparse_categorical_accuracy: 0.2729\n",
      "Epoch 5/200\n",
      "11/11 [==============================] - 2s 185ms/step - loss: 0.2077 - sparse_categorical_accuracy: 0.9428 - val_loss: 2.0446 - val_sparse_categorical_accuracy: 0.2357\n",
      "Epoch 6/200\n",
      "11/11 [==============================] - 2s 184ms/step - loss: 0.1637 - sparse_categorical_accuracy: 0.9559 - val_loss: 2.1738 - val_sparse_categorical_accuracy: 0.2171\n",
      "Epoch 7/200\n",
      "11/11 [==============================] - 2s 190ms/step - loss: 0.1370 - sparse_categorical_accuracy: 0.9657 - val_loss: 2.3156 - val_sparse_categorical_accuracy: 0.1771\n",
      "Epoch 8/200\n",
      "11/11 [==============================] - 2s 185ms/step - loss: 0.1086 - sparse_categorical_accuracy: 0.9779 - val_loss: 2.4793 - val_sparse_categorical_accuracy: 0.1471\n",
      "Epoch 9/200\n",
      "11/11 [==============================] - 2s 189ms/step - loss: 0.0932 - sparse_categorical_accuracy: 0.9779 - val_loss: 2.5912 - val_sparse_categorical_accuracy: 0.1486\n",
      "Epoch 10/200\n",
      "11/11 [==============================] - 2s 185ms/step - loss: 0.0916 - sparse_categorical_accuracy: 0.9805 - val_loss: 2.6542 - val_sparse_categorical_accuracy: 0.1857\n",
      "Epoch 11/200\n",
      "11/11 [==============================] - 2s 184ms/step - loss: 0.0662 - sparse_categorical_accuracy: 0.9855 - val_loss: 2.8096 - val_sparse_categorical_accuracy: 0.1371\n",
      "Epoch 12/200\n",
      "11/11 [==============================] - 2s 182ms/step - loss: 0.0582 - sparse_categorical_accuracy: 0.9892 - val_loss: 2.8554 - val_sparse_categorical_accuracy: 0.1443\n",
      "Epoch 13/200\n",
      "11/11 [==============================] - 2s 184ms/step - loss: 0.0491 - sparse_categorical_accuracy: 0.9959 - val_loss: 2.9937 - val_sparse_categorical_accuracy: 0.1371\n",
      "Epoch 14/200\n",
      "11/11 [==============================] - 2s 187ms/step - loss: 0.0515 - sparse_categorical_accuracy: 0.9920 - val_loss: 3.0735 - val_sparse_categorical_accuracy: 0.1400\n",
      "Epoch 15/200\n",
      "11/11 [==============================] - 2s 184ms/step - loss: 0.0344 - sparse_categorical_accuracy: 0.9967 - val_loss: 3.0534 - val_sparse_categorical_accuracy: 0.1471\n",
      "Epoch 16/200\n",
      "11/11 [==============================] - 2s 182ms/step - loss: 0.0311 - sparse_categorical_accuracy: 0.9981 - val_loss: 2.9961 - val_sparse_categorical_accuracy: 0.1457\n",
      "Epoch 17/200\n",
      "11/11 [==============================] - 2s 183ms/step - loss: 0.0307 - sparse_categorical_accuracy: 0.9959 - val_loss: 2.8793 - val_sparse_categorical_accuracy: 0.1657\n",
      "Epoch 18/200\n",
      "11/11 [==============================] - 2s 187ms/step - loss: 0.0255 - sparse_categorical_accuracy: 0.9978 - val_loss: 2.7236 - val_sparse_categorical_accuracy: 0.1929\n",
      "Epoch 19/200\n",
      "11/11 [==============================] - 2s 185ms/step - loss: 0.0261 - sparse_categorical_accuracy: 0.9968 - val_loss: 2.6140 - val_sparse_categorical_accuracy: 0.2000\n",
      "Epoch 20/200\n",
      "11/11 [==============================] - 2s 185ms/step - loss: 0.0235 - sparse_categorical_accuracy: 0.9985 - val_loss: 2.5463 - val_sparse_categorical_accuracy: 0.2200\n",
      "Epoch 21/200\n",
      "11/11 [==============================] - 2s 186ms/step - loss: 0.0223 - sparse_categorical_accuracy: 0.9982 - val_loss: 2.4188 - val_sparse_categorical_accuracy: 0.2557\n",
      "Epoch 22/200\n",
      "11/11 [==============================] - 2s 187ms/step - loss: 0.0223 - sparse_categorical_accuracy: 0.9966 - val_loss: 2.0729 - val_sparse_categorical_accuracy: 0.3614\n",
      "Epoch 23/200\n",
      "11/11 [==============================] - 2s 188ms/step - loss: 0.0190 - sparse_categorical_accuracy: 0.9989 - val_loss: 1.9620 - val_sparse_categorical_accuracy: 0.3714\n",
      "Epoch 24/200\n",
      "11/11 [==============================] - 2s 190ms/step - loss: 0.0196 - sparse_categorical_accuracy: 0.9995 - val_loss: 1.8949 - val_sparse_categorical_accuracy: 0.3986\n",
      "Epoch 25/200\n",
      "11/11 [==============================] - 2s 190ms/step - loss: 0.0170 - sparse_categorical_accuracy: 0.9984 - val_loss: 1.5978 - val_sparse_categorical_accuracy: 0.4829\n",
      "Epoch 26/200\n",
      "11/11 [==============================] - 2s 186ms/step - loss: 0.0154 - sparse_categorical_accuracy: 0.9985 - val_loss: 1.4202 - val_sparse_categorical_accuracy: 0.5371\n",
      "Epoch 27/200\n",
      "11/11 [==============================] - 2s 189ms/step - loss: 0.0143 - sparse_categorical_accuracy: 0.9990 - val_loss: 1.3249 - val_sparse_categorical_accuracy: 0.5529\n",
      "Epoch 28/200\n",
      "11/11 [==============================] - 2s 188ms/step - loss: 0.0144 - sparse_categorical_accuracy: 0.9988 - val_loss: 1.1857 - val_sparse_categorical_accuracy: 0.5886\n",
      "Epoch 29/200\n",
      "11/11 [==============================] - 2s 212ms/step - loss: 0.0128 - sparse_categorical_accuracy: 0.9992 - val_loss: 0.9866 - val_sparse_categorical_accuracy: 0.6457\n",
      "Epoch 30/200\n",
      "11/11 [==============================] - 2s 210ms/step - loss: 0.0102 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.8345 - val_sparse_categorical_accuracy: 0.7114\n",
      "Epoch 31/200\n",
      "11/11 [==============================] - 2s 202ms/step - loss: 0.0111 - sparse_categorical_accuracy: 0.9995 - val_loss: 0.7266 - val_sparse_categorical_accuracy: 0.7600\n",
      "Epoch 32/200\n",
      "11/11 [==============================] - 2s 198ms/step - loss: 0.0110 - sparse_categorical_accuracy: 0.9994 - val_loss: 0.6016 - val_sparse_categorical_accuracy: 0.8043\n",
      "Epoch 33/200\n",
      "11/11 [==============================] - 2s 186ms/step - loss: 0.0100 - sparse_categorical_accuracy: 0.9987 - val_loss: 0.4966 - val_sparse_categorical_accuracy: 0.8457\n",
      "Epoch 34/200\n",
      "11/11 [==============================] - 2s 186ms/step - loss: 0.0105 - sparse_categorical_accuracy: 0.9995 - val_loss: 0.4134 - val_sparse_categorical_accuracy: 0.8757\n",
      "Epoch 35/200\n",
      "11/11 [==============================] - 2s 187ms/step - loss: 0.0106 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.3374 - val_sparse_categorical_accuracy: 0.8986\n",
      "Epoch 36/200\n",
      "11/11 [==============================] - 2s 188ms/step - loss: 0.0086 - sparse_categorical_accuracy: 0.9993 - val_loss: 0.3152 - val_sparse_categorical_accuracy: 0.9014\n",
      "Epoch 37/200\n",
      "11/11 [==============================] - 2s 183ms/step - loss: 0.0174 - sparse_categorical_accuracy: 0.9949 - val_loss: 0.3963 - val_sparse_categorical_accuracy: 0.8814\n",
      "Epoch 38/200\n",
      "11/11 [==============================] - 2s 185ms/step - loss: 0.0234 - sparse_categorical_accuracy: 0.9955 - val_loss: 0.2742 - val_sparse_categorical_accuracy: 0.9114\n",
      "Epoch 39/200\n",
      "11/11 [==============================] - 2s 187ms/step - loss: 0.0112 - sparse_categorical_accuracy: 0.9999 - val_loss: 0.2298 - val_sparse_categorical_accuracy: 0.9257\n",
      "Epoch 40/200\n",
      "11/11 [==============================] - 2s 184ms/step - loss: 0.0125 - sparse_categorical_accuracy: 0.9979 - val_loss: 0.1820 - val_sparse_categorical_accuracy: 0.9457\n",
      "Epoch 41/200\n",
      "11/11 [==============================] - 2s 185ms/step - loss: 0.0080 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.1666 - val_sparse_categorical_accuracy: 0.9486\n",
      "Epoch 42/200\n",
      "11/11 [==============================] - 2s 188ms/step - loss: 0.0079 - sparse_categorical_accuracy: 0.9997 - val_loss: 0.1526 - val_sparse_categorical_accuracy: 0.9571\n",
      "Epoch 43/200\n",
      "11/11 [==============================] - 2s 186ms/step - loss: 0.0075 - sparse_categorical_accuracy: 0.9993 - val_loss: 0.1369 - val_sparse_categorical_accuracy: 0.9671\n",
      "Epoch 44/200\n",
      "11/11 [==============================] - 2s 190ms/step - loss: 0.0081 - sparse_categorical_accuracy: 0.9995 - val_loss: 0.1224 - val_sparse_categorical_accuracy: 0.9671\n",
      "Epoch 45/200\n",
      "11/11 [==============================] - 2s 189ms/step - loss: 0.0053 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.1147 - val_sparse_categorical_accuracy: 0.9700\n",
      "Epoch 46/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "11/11 [==============================] - 2s 184ms/step - loss: 0.0068 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.1049 - val_sparse_categorical_accuracy: 0.9714\n",
      "Epoch 47/200\n",
      "11/11 [==============================] - 2s 182ms/step - loss: 0.0059 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0982 - val_sparse_categorical_accuracy: 0.9700\n",
      "Epoch 48/200\n",
      "11/11 [==============================] - 2s 188ms/step - loss: 0.0051 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0952 - val_sparse_categorical_accuracy: 0.9729\n",
      "Epoch 49/200\n",
      "11/11 [==============================] - 2s 184ms/step - loss: 0.0046 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0934 - val_sparse_categorical_accuracy: 0.9729\n",
      "Epoch 50/200\n",
      "11/11 [==============================] - 2s 186ms/step - loss: 0.0052 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0900 - val_sparse_categorical_accuracy: 0.9700\n",
      "Epoch 51/200\n",
      "11/11 [==============================] - 2s 191ms/step - loss: 0.0071 - sparse_categorical_accuracy: 0.9993 - val_loss: 0.0895 - val_sparse_categorical_accuracy: 0.9743\n",
      "Epoch 52/200\n",
      "11/11 [==============================] - 2s 189ms/step - loss: 0.0050 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0881 - val_sparse_categorical_accuracy: 0.9714\n",
      "Epoch 53/200\n",
      "11/11 [==============================] - 2s 183ms/step - loss: 0.0045 - sparse_categorical_accuracy: 0.9999 - val_loss: 0.0878 - val_sparse_categorical_accuracy: 0.9729\n",
      "Epoch 54/200\n",
      "11/11 [==============================] - 2s 191ms/step - loss: 0.0046 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0866 - val_sparse_categorical_accuracy: 0.9729\n",
      "Epoch 55/200\n",
      "11/11 [==============================] - 2s 188ms/step - loss: 0.0035 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0819 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 56/200\n",
      "11/11 [==============================] - 2s 183ms/step - loss: 0.0041 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0804 - val_sparse_categorical_accuracy: 0.9743\n",
      "Epoch 57/200\n",
      "11/11 [==============================] - 2s 186ms/step - loss: 0.0040 - sparse_categorical_accuracy: 0.9999 - val_loss: 0.0781 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 58/200\n",
      "11/11 [==============================] - 2s 184ms/step - loss: 0.0037 - sparse_categorical_accuracy: 0.9996 - val_loss: 0.0798 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 59/200\n",
      "11/11 [==============================] - 2s 183ms/step - loss: 0.0035 - sparse_categorical_accuracy: 0.9999 - val_loss: 0.0810 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 60/200\n",
      "11/11 [==============================] - 2s 186ms/step - loss: 0.0043 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0745 - val_sparse_categorical_accuracy: 0.9786\n",
      "Epoch 61/200\n",
      "11/11 [==============================] - 2s 184ms/step - loss: 0.0051 - sparse_categorical_accuracy: 0.9990 - val_loss: 0.0795 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 62/200\n",
      "11/11 [==============================] - 2s 186ms/step - loss: 0.0042 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0758 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 63/200\n",
      "11/11 [==============================] - 2s 183ms/step - loss: 0.0034 - sparse_categorical_accuracy: 0.9996 - val_loss: 0.0679 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 64/200\n",
      "11/11 [==============================] - 2s 187ms/step - loss: 0.0030 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0692 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 65/200\n",
      "11/11 [==============================] - 2s 185ms/step - loss: 0.0030 - sparse_categorical_accuracy: 0.9999 - val_loss: 0.0759 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 66/200\n",
      "11/11 [==============================] - 2s 186ms/step - loss: 0.0027 - sparse_categorical_accuracy: 0.9999 - val_loss: 0.0781 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 67/200\n",
      "11/11 [==============================] - 2s 192ms/step - loss: 0.0033 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0752 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 68/200\n",
      "11/11 [==============================] - 2s 187ms/step - loss: 0.0029 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0743 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 69/200\n",
      "11/11 [==============================] - 2s 188ms/step - loss: 0.0028 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0742 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 70/200\n",
      "11/11 [==============================] - 2s 188ms/step - loss: 0.0026 - sparse_categorical_accuracy: 0.9995 - val_loss: 0.0767 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 71/200\n",
      "11/11 [==============================] - 2s 182ms/step - loss: 0.0031 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0811 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 72/200\n",
      "11/11 [==============================] - 2s 185ms/step - loss: 0.0039 - sparse_categorical_accuracy: 0.9997 - val_loss: 0.0724 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 73/200\n",
      "11/11 [==============================] - 2s 204ms/step - loss: 0.0028 - sparse_categorical_accuracy: 0.9997 - val_loss: 0.0653 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 74/200\n",
      "11/11 [==============================] - 2s 202ms/step - loss: 0.0028 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0654 - val_sparse_categorical_accuracy: 0.9786\n",
      "Epoch 75/200\n",
      "11/11 [==============================] - 2s 197ms/step - loss: 0.0022 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0701 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 76/200\n",
      "11/11 [==============================] - 2s 192ms/step - loss: 0.0023 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0735 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 77/200\n",
      "11/11 [==============================] - 2s 187ms/step - loss: 0.0023 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0736 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 78/200\n",
      "11/11 [==============================] - 2s 184ms/step - loss: 0.0024 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0716 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 79/200\n",
      "11/11 [==============================] - 2s 188ms/step - loss: 0.0021 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0689 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 80/200\n",
      "11/11 [==============================] - 2s 186ms/step - loss: 0.0023 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0686 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 81/200\n",
      "11/11 [==============================] - 2s 186ms/step - loss: 0.0023 - sparse_categorical_accuracy: 0.9997 - val_loss: 0.0687 - val_sparse_categorical_accuracy: 0.9743\n",
      "Epoch 82/200\n",
      "11/11 [==============================] - 2s 186ms/step - loss: 0.0018 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0666 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 83/200\n",
      "11/11 [==============================] - 2s 193ms/step - loss: 0.0025 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0662 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 84/200\n",
      "11/11 [==============================] - 2s 186ms/step - loss: 0.0021 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0658 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 85/200\n",
      "11/11 [==============================] - 2s 186ms/step - loss: 0.0017 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0664 - val_sparse_categorical_accuracy: 0.9786\n",
      "Epoch 86/200\n",
      "11/11 [==============================] - 2s 189ms/step - loss: 0.0015 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0674 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 87/200\n",
      "11/11 [==============================] - 2s 194ms/step - loss: 0.0014 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0659 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 88/200\n",
      "11/11 [==============================] - 2s 199ms/step - loss: 0.0020 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0696 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 89/200\n",
      "11/11 [==============================] - 2s 187ms/step - loss: 0.0019 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0695 - val_sparse_categorical_accuracy: 0.9786\n",
      "Epoch 90/200\n",
      "11/11 [==============================] - 2s 190ms/step - loss: 0.0016 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0701 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 91/200\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "11/11 [==============================] - 2s 193ms/step - loss: 0.0017 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0706 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 92/200\n",
      "11/11 [==============================] - 2s 189ms/step - loss: 0.0020 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0710 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 93/200\n",
      "11/11 [==============================] - 2s 182ms/step - loss: 0.0019 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0743 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 94/200\n",
      "11/11 [==============================] - 2s 216ms/step - loss: 0.0019 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0750 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 95/200\n",
      "11/11 [==============================] - 2s 223ms/step - loss: 0.0016 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0750 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 96/200\n",
      "11/11 [==============================] - 2s 194ms/step - loss: 0.0015 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0760 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 97/200\n",
      "11/11 [==============================] - 2s 186ms/step - loss: 0.0016 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0736 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 98/200\n",
      "11/11 [==============================] - 2s 203ms/step - loss: 0.0016 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0767 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 99/200\n",
      "11/11 [==============================] - 2s 201ms/step - loss: 0.0013 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0793 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 100/200\n",
      "11/11 [==============================] - 2s 196ms/step - loss: 0.0013 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0776 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 101/200\n",
      "11/11 [==============================] - 2s 198ms/step - loss: 0.0013 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0772 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 102/200\n",
      "11/11 [==============================] - 2s 194ms/step - loss: 0.0011 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0767 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 103/200\n",
      "11/11 [==============================] - 2s 194ms/step - loss: 0.0010 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0755 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 104/200\n",
      "11/11 [==============================] - 2s 199ms/step - loss: 0.0015 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0770 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 105/200\n",
      "11/11 [==============================] - 2s 196ms/step - loss: 0.0014 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0773 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 106/200\n",
      "11/11 [==============================] - 2s 195ms/step - loss: 0.0014 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0771 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 107/200\n",
      "11/11 [==============================] - 2s 192ms/step - loss: 0.0015 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0775 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 108/200\n",
      "11/11 [==============================] - 2s 196ms/step - loss: 0.0011 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0741 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 109/200\n",
      "11/11 [==============================] - 2s 194ms/step - loss: 0.0012 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0738 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 110/200\n",
      "11/11 [==============================] - 2s 195ms/step - loss: 0.0012 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0731 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 111/200\n",
      "11/11 [==============================] - 2s 199ms/step - loss: 0.0011 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0732 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 112/200\n",
      "11/11 [==============================] - 2s 195ms/step - loss: 0.0010 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0742 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 113/200\n",
      "11/11 [==============================] - 2s 192ms/step - loss: 0.0013 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0754 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 114/200\n",
      "11/11 [==============================] - 2s 195ms/step - loss: 0.0010 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0757 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 115/200\n",
      "11/11 [==============================] - 2s 218ms/step - loss: 0.0010 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0730 - val_sparse_categorical_accuracy: 0.9786\n",
      "Epoch 116/200\n",
      "11/11 [==============================] - 2s 206ms/step - loss: 0.0011 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0673 - val_sparse_categorical_accuracy: 0.9786\n",
      "Epoch 117/200\n",
      "11/11 [==============================] - 2s 199ms/step - loss: 0.0011 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0648 - val_sparse_categorical_accuracy: 0.9800\n",
      "Epoch 118/200\n",
      "11/11 [==============================] - 2s 195ms/step - loss: 8.6929e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0667 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 119/200\n",
      "11/11 [==============================] - 2s 189ms/step - loss: 0.0012 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0684 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 120/200\n",
      "11/11 [==============================] - 2s 188ms/step - loss: 0.0013 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0695 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 121/200\n",
      "11/11 [==============================] - 2s 187ms/step - loss: 9.0241e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0698 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 122/200\n",
      "11/11 [==============================] - 2s 196ms/step - loss: 0.0013 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0711 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 123/200\n",
      "11/11 [==============================] - 2s 188ms/step - loss: 0.0013 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0726 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 124/200\n",
      "11/11 [==============================] - 2s 195ms/step - loss: 9.6683e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0723 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 125/200\n",
      "11/11 [==============================] - 2s 203ms/step - loss: 8.8805e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0735 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 126/200\n",
      "11/11 [==============================] - 2s 198ms/step - loss: 7.6072e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0738 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 127/200\n",
      "11/11 [==============================] - 2s 198ms/step - loss: 0.0010 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0706 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 128/200\n",
      "11/11 [==============================] - 2s 191ms/step - loss: 8.9438e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0702 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 129/200\n",
      "11/11 [==============================] - 2s 194ms/step - loss: 0.0012 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0704 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 130/200\n",
      "11/11 [==============================] - 2s 193ms/step - loss: 9.5157e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0704 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 131/200\n",
      "11/11 [==============================] - 2s 195ms/step - loss: 0.0016 - sparse_categorical_accuracy: 0.9996 - val_loss: 0.0724 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 132/200\n",
      "11/11 [==============================] - 2s 190ms/step - loss: 7.9028e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0730 - val_sparse_categorical_accuracy: 0.9743\n",
      "Epoch 133/200\n",
      "11/11 [==============================] - 2s 188ms/step - loss: 0.0010 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0734 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 134/200\n",
      "11/11 [==============================] - 2s 198ms/step - loss: 9.7406e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0740 - val_sparse_categorical_accuracy: 0.9786\n",
      "Epoch 135/200\n",
      "11/11 [==============================] - 2s 191ms/step - loss: 9.2616e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0749 - val_sparse_categorical_accuracy: 0.9757\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 136/200\n",
      "11/11 [==============================] - 2s 190ms/step - loss: 0.0011 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0766 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 137/200\n",
      "11/11 [==============================] - 2s 184ms/step - loss: 6.8437e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0780 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 138/200\n",
      "11/11 [==============================] - 2s 197ms/step - loss: 8.7062e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0794 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 139/200\n",
      "11/11 [==============================] - 2s 194ms/step - loss: 8.6511e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0815 - val_sparse_categorical_accuracy: 0.9743\n",
      "Epoch 140/200\n",
      "11/11 [==============================] - 2s 199ms/step - loss: 7.8327e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0788 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 141/200\n",
      "11/11 [==============================] - 2s 186ms/step - loss: 0.0012 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0812 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 142/200\n",
      "11/11 [==============================] - 2s 191ms/step - loss: 6.8832e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0823 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 143/200\n",
      "11/11 [==============================] - 2s 202ms/step - loss: 0.0011 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0832 - val_sparse_categorical_accuracy: 0.9743\n",
      "Epoch 144/200\n",
      "11/11 [==============================] - 2s 211ms/step - loss: 9.1377e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0853 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 145/200\n",
      "11/11 [==============================] - 2s 190ms/step - loss: 0.0010 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0786 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 146/200\n",
      "11/11 [==============================] - 2s 188ms/step - loss: 5.7712e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0777 - val_sparse_categorical_accuracy: 0.9786\n",
      "Epoch 147/200\n",
      "11/11 [==============================] - 2s 187ms/step - loss: 7.9366e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0759 - val_sparse_categorical_accuracy: 0.9786\n",
      "Epoch 148/200\n",
      "11/11 [==============================] - 2s 186ms/step - loss: 6.5362e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0766 - val_sparse_categorical_accuracy: 0.9786\n",
      "Epoch 149/200\n",
      "11/11 [==============================] - 2s 189ms/step - loss: 0.0011 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0785 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 150/200\n",
      "11/11 [==============================] - 2s 184ms/step - loss: 7.0239e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0797 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 151/200\n",
      "11/11 [==============================] - 2s 190ms/step - loss: 4.7795e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0800 - val_sparse_categorical_accuracy: 0.9786\n",
      "Epoch 152/200\n",
      "11/11 [==============================] - 2s 189ms/step - loss: 6.6449e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0790 - val_sparse_categorical_accuracy: 0.9786\n",
      "Epoch 153/200\n",
      "11/11 [==============================] - 2s 187ms/step - loss: 7.1095e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0781 - val_sparse_categorical_accuracy: 0.9786\n",
      "Epoch 154/200\n",
      "11/11 [==============================] - 2s 191ms/step - loss: 6.6458e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0788 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 155/200\n",
      "11/11 [==============================] - 2s 192ms/step - loss: 8.1770e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0793 - val_sparse_categorical_accuracy: 0.9786\n",
      "Epoch 156/200\n",
      "11/11 [==============================] - 2s 189ms/step - loss: 6.3006e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0795 - val_sparse_categorical_accuracy: 0.9800\n",
      "Epoch 157/200\n",
      "11/11 [==============================] - 2s 186ms/step - loss: 5.8295e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0797 - val_sparse_categorical_accuracy: 0.9800\n",
      "Epoch 158/200\n",
      "11/11 [==============================] - 2s 187ms/step - loss: 5.1507e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0796 - val_sparse_categorical_accuracy: 0.9800\n",
      "Epoch 159/200\n",
      "11/11 [==============================] - 2s 186ms/step - loss: 7.1311e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0798 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 160/200\n",
      "11/11 [==============================] - 2s 190ms/step - loss: 5.5583e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0825 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 161/200\n",
      "11/11 [==============================] - 2s 193ms/step - loss: 6.0440e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0813 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 162/200\n",
      "11/11 [==============================] - 2s 187ms/step - loss: 4.8695e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0800 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 163/200\n",
      "11/11 [==============================] - 2s 192ms/step - loss: 5.2438e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0809 - val_sparse_categorical_accuracy: 0.9786\n",
      "Epoch 164/200\n",
      "11/11 [==============================] - 2s 186ms/step - loss: 5.4935e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0830 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 165/200\n",
      "11/11 [==============================] - 2s 186ms/step - loss: 5.3565e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0842 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 166/200\n",
      "11/11 [==============================] - 2s 190ms/step - loss: 4.3594e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0837 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 167/200\n",
      "11/11 [==============================] - 2s 188ms/step - loss: 5.0513e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0786 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 168/200\n",
      "11/11 [==============================] - 2s 182ms/step - loss: 5.3888e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0775 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 169/200\n",
      "11/11 [==============================] - 2s 189ms/step - loss: 3.6740e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0771 - val_sparse_categorical_accuracy: 0.9786\n",
      "Epoch 170/200\n",
      "11/11 [==============================] - 2s 191ms/step - loss: 4.4554e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0774 - val_sparse_categorical_accuracy: 0.9786\n",
      "Epoch 171/200\n",
      "11/11 [==============================] - 2s 182ms/step - loss: 5.1266e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0776 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 172/200\n",
      "11/11 [==============================] - 2s 190ms/step - loss: 3.8428e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0787 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 173/200\n",
      "11/11 [==============================] - 2s 185ms/step - loss: 5.4393e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0790 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 174/200\n",
      "11/11 [==============================] - 2s 186ms/step - loss: 4.5834e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0790 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 175/200\n",
      "11/11 [==============================] - 2s 188ms/step - loss: 4.3555e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0763 - val_sparse_categorical_accuracy: 0.9757\n",
      "Epoch 176/200\n",
      "11/11 [==============================] - 2s 192ms/step - loss: 5.0432e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0741 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 177/200\n",
      "11/11 [==============================] - 2s 186ms/step - loss: 4.6758e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0724 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 178/200\n",
      "11/11 [==============================] - 2s 187ms/step - loss: 5.0601e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0715 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 179/200\n",
      "11/11 [==============================] - 2s 191ms/step - loss: 4.9869e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0743 - val_sparse_categorical_accuracy: 0.9771\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 180/200\n",
      "11/11 [==============================] - 2s 190ms/step - loss: 4.3862e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0787 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 181/200\n",
      "11/11 [==============================] - 2s 190ms/step - loss: 6.0768e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0775 - val_sparse_categorical_accuracy: 0.9786\n",
      "Epoch 182/200\n",
      "11/11 [==============================] - 2s 188ms/step - loss: 4.8166e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0766 - val_sparse_categorical_accuracy: 0.9786\n",
      "Epoch 183/200\n",
      "11/11 [==============================] - 2s 186ms/step - loss: 4.6304e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0777 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 184/200\n",
      "11/11 [==============================] - 2s 185ms/step - loss: 4.6769e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0794 - val_sparse_categorical_accuracy: 0.9771\n",
      "Epoch 185/200\n",
      "11/11 [==============================] - 2s 188ms/step - loss: 0.0014 - sparse_categorical_accuracy: 0.9989 - val_loss: 0.0809 - val_sparse_categorical_accuracy: 0.9743\n",
      "Epoch 186/200\n",
      "11/11 [==============================] - 2s 186ms/step - loss: 9.8040e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0805 - val_sparse_categorical_accuracy: 0.9786\n",
      "Epoch 187/200\n",
      "11/11 [==============================] - 2s 188ms/step - loss: 0.0015 - sparse_categorical_accuracy: 0.9996 - val_loss: 0.0865 - val_sparse_categorical_accuracy: 0.9814\n",
      "Epoch 188/200\n",
      "11/11 [==============================] - 2s 187ms/step - loss: 0.0019 - sparse_categorical_accuracy: 0.9997 - val_loss: 0.1019 - val_sparse_categorical_accuracy: 0.9786\n",
      "Epoch 189/200\n",
      "11/11 [==============================] - 2s 185ms/step - loss: 0.0018 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0800 - val_sparse_categorical_accuracy: 0.9786\n",
      "Epoch 190/200\n",
      "11/11 [==============================] - 2s 185ms/step - loss: 0.0014 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.1047 - val_sparse_categorical_accuracy: 0.9743\n",
      "Epoch 191/200\n",
      "11/11 [==============================] - 2s 188ms/step - loss: 0.0011 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0933 - val_sparse_categorical_accuracy: 0.9800\n",
      "Epoch 192/200\n",
      "11/11 [==============================] - 2s 186ms/step - loss: 0.0010 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0812 - val_sparse_categorical_accuracy: 0.9786\n",
      "Epoch 193/200\n",
      "11/11 [==============================] - 2s 186ms/step - loss: 7.1902e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0764 - val_sparse_categorical_accuracy: 0.9814\n",
      "Epoch 194/200\n",
      "11/11 [==============================] - 2s 187ms/step - loss: 7.1979e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0800 - val_sparse_categorical_accuracy: 0.9800\n",
      "Epoch 195/200\n",
      "11/11 [==============================] - 2s 183ms/step - loss: 5.6721e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0837 - val_sparse_categorical_accuracy: 0.9786\n",
      "Epoch 196/200\n",
      "11/11 [==============================] - 2s 182ms/step - loss: 5.7618e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0849 - val_sparse_categorical_accuracy: 0.9800\n",
      "Epoch 197/200\n",
      "11/11 [==============================] - 2s 189ms/step - loss: 8.6442e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0814 - val_sparse_categorical_accuracy: 0.9800\n",
      "Epoch 198/200\n",
      "11/11 [==============================] - 2s 199ms/step - loss: 0.0011 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0819 - val_sparse_categorical_accuracy: 0.9814\n",
      "Epoch 199/200\n",
      "11/11 [==============================] - 2s 186ms/step - loss: 6.2654e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0786 - val_sparse_categorical_accuracy: 0.9814\n",
      "Epoch 200/200\n",
      "11/11 [==============================] - 2s 185ms/step - loss: 7.8258e-04 - sparse_categorical_accuracy: 1.0000 - val_loss: 0.0766 - val_sparse_categorical_accuracy: 0.9814\n"
     ]
    }
   ],
   "source": [
    "# 6. Compile model\n",
    "model.compile(loss='sparse_categorical_crossentropy',\n",
    "              optimizer='adam',\n",
    "              metrics=['sparse_categorical_accuracy'])\n",
    "\n",
    "\n",
    "# 7. Fit model on training data\n",
    "history = model.fit(X_train, y_train,\n",
    "          epochs=200, batch_size=256, validation_split=0.2)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAEICAYAAABPgw/pAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAAgAElEQVR4nO3deXhU5d3/8feXJCSEsEkAkQABBREKxBBRERUfqSJaEcUqooK2AqLto/5staI+thZrrX304nKhWHHBWBAXqo+oLVqxdYOIYZPFgAHCGsIW9oTcvz/uM8lkmEkmMDNnzuT7uq65zsyZM+d8c2bymXvus4kxBqWUUt7XxO0ClFJKRYYGulJKJQgNdKWUShAa6EoplSA00JVSKkFooCulVILQQE9gIvKBiIyN9LRuEpFiERkahfkaETnNuT9NRB4KZ9rjWM4YEfnH8dapVF1E90OPLyKyz+9hOnAYOOo8nmCMyY99VfFDRIqBnxtj5kd4vgboYYwpitS0IpIN/ACkGGMqI1GnUnVJdrsAVZsxJsN3v67wEpFkDQkVL/TzGB+0y8UjRGSIiJSIyH0ishV4SUTaiMj/iUipiOxy7mf5veZTEfm5c3+ciPxHRJ50pv1BRC47zmm7ichnIlIuIvNF5FkReS1E3eHU+KiIfO7M7x8ikun3/E0isl5EykRkch3r5xwR2SoiSX7jRorIUuf+QBH5UkR2i8gWEXlGRJqGmNfLIvJ7v8e/cl6zWURuDZj2chH5VkT2ishGEXnE7+nPnOFuEdknIuf61q3f6weJyCIR2eMMB4W7bhq4nk8SkZecv2GXiMz1e26EiBQ6f8NaERnmjK/VvSUij/jeZxHJdrqefiYiG4BPnPFznPdhj/MZ6eP3+mYi8mfn/dzjfMaaicj7IvKLgL9nqYhcFexvVaFpoHvLycBJQFdgPPb9e8l53AU4CDxTx+vPBlYDmcATwIsiIscx7evAQqAt8AhwUx3LDKfGG4BbgPZAU+BeABHpDTzvzP8UZ3lZBGGM+QrYD/xXwHxfd+4fBe52/p5zgYuBSXXUjVPDMKeeHwM9gMD++/3AzUBr4HLgdr8gusAZtjbGZBhjvgyY90nA+8BU52/7X+B9EWkb8Dccs26CqG89z8R24fVx5vWUU8NA4FXgV87fcAFQHGp9BHEhcAZwqfP4A+x6ag8sBvy7CJ8EBgCDsJ/jXwNVwCvAjb6JRKQ/0AmY14A6FIAxRm9xesP+Yw117g8BjgBpdUyfA+zye/wptssGYBxQ5PdcOmCAkxsyLTYsKoF0v+dfA14L828KVuODfo8nAR869x8GZvk919xZB0NDzPv3wAznfgts2HYNMe1dwDt+jw1wmnP/ZeD3zv0ZwON+0/X0nzbIfJ8GnnLuZzvTJvs9Pw74j3P/JmBhwOu/BMbVt24asp6BjtjgbBNkur/46q3r8+c8fsT3Pvv9bd3rqKG1M00r7BfOQaB/kOlSgZ3Y7RJgg/+5WP+/JcJNW+jeUmqMOeR7ICLpIvIX5yfsXuxP/Nb+3Q4BtvruGGMOOHczGjjtKcBOv3EAG0MVHGaNW/3uH/Cr6RT/eRtj9gNloZaFbY1fLSKpwNXAYmPMeqeOnk43xFanjsewrfX61KoBWB/w950tIv9yujr2ABPDnK9v3usDxq3Htk59Qq2bWupZz52x79muIC/tDKwNs95gqteNiCSJyONOt81ealr6mc4tLdiyjDGHgTeAG0WkCTAa+4tCNZAGurcE7pL0/4DTgbONMS2p+YkfqhslErYAJ4lIut+4znVMfyI1bvGft7PMtqEmNsZ8hw3Ey6jd3QK262YVthXYEnjgeGrA/kLx9zrwLtDZGNMKmOY33/p2IduM7SLx1wXYFEZdgepazxux71nrIK/bCJwaYp77sb/OfE4OMo3/33gDMALbLdUK24r31bADOFTHsl4BxmC7wg6YgO4pFR4NdG9rgf0Zu9vpj/2faC/QafEWAI+ISFMRORf4SZRqfBO4QkQGOxswf0f9n9nXgV9iA21OQB17gX0i0gu4Pcwa3gDGiUhv5wslsP4W2NbvIac/+ga/50qxXR3dQ8x7HtBTRG4QkWQRuQ7oDfxfmLUF1hF0PRtjtmD7tp9zNp6miIgv8F8EbhGRi0WkiYh0ctYPQCFwvTN9HjAqjBoOY39FpWN/BflqqMJ2X/2viJzitObPdX5N4QR4FfBntHV+3DTQve1poBm29fMV8GGMljsGu2GxDNtvPRv7jxzMcddojFkB3IEN6S3ALqCknpf9Dbu94RNjzA6/8fdiw7YceMGpOZwaPnD+hk+AImfobxLwOxEpx/b5v+H32gPAFOBzsXvXnBMw7zLgCmzrugy7kfCKgLrDVd96vgmowP5K2Y7dhoAxZiF2o+tTwB5gATW/Gh7Ctqh3Ab+l9i+eYF7F/kLaBHzn1OHvXmAZsAjbZ/5HamfQq0Bf7DYZdRz0wCJ1wkRkNrDKGBP1XwgqcYnIzcB4Y8xgt2vxKm2hqwYTkbNE5FTnJ/owbL/p3Ppep1QoTnfWJGC627V4mQa6Oh4nY3ep24fdh/p2Y8y3rlakPEtELsVub9hG/d06qg7a5aKUUglCW+hKKZUgXDs5V2ZmpsnOznZr8Uop5UnffPPNDmNMu2DPuRbo2dnZFBQUuLV4pZTyJBEJPLq4mna5KKVUgtBAV0qpBKGBrpRSCUIDXSmlEoQGulJKJYh6A11EZojIdhFZHuJ5EZGpIlLkXDYqN/Jluic/H7KzoUkTO8zPb9hzIvZ5EXtLSrLDzEzIyDh2vG8+/vMOnDZSN98y/euLxk2Xo8vR5Rz7uuyAzIiI+q6AgT0NaS6wPMTzw7Gn5hTgHODrcK6sMWDAAOOm114zpmtXY0SMadvWmObNjQG96U1veovdLT3dZlFDAAXGHOcVi4wxn2FPdRnKCOBVZ1lfYa+S0vGEvmWiKD/ftnhvvBHWr7ertawM9u93uzKlVGNz4ABMDnnp84aLRB96J2pfoquE2pfQqiYi40WkQEQKSktLI7Dohpk0CW66yQa4UkrFgw0bIjevSAS6BBlngk1ojJlujMkzxuS1axf0yNWoyc+HadNsi1wppeJFl8CLGp6ASAR6CbWvuZiFvVZi3MjPh7FjNcyVUvElPR2mTInc/CIR6O8CNzt7u5wD7DH2Goau8t/L5MYb4ehRtyuKP02cd1+C/cbS5ehydDlRWY5v+q5dYfp0GDMmcjXVe3IuEfFdozFTREqwF59NATDGTMNe6HY49nqLB7DXJ3RVfj6MH283OJyoJk2gqgraOteaLyuzb4ivtd+8OaSlHTve97quXe03sO9Ny8+3G0E2bLA/tXzP+cavX293azp6tGaZO3fWnjYWli+H11+H9u3h7LMhJwe2bIG1a+3z3bvbW30f5spK+O47u166dIG//Q02bYKzzrLz7RjG5vNDh2DZMmjVClq0gFdegYoK+/qzzrLL+PprWLgQTjrJ/hpr0+bE1wHA1q22/qNHba1nnGHfn507YelSW0c4duyAb76B8nLYtQu+/RZ69oRbbrF/F0DTptC/P7RuXfM6/+W0bQt9+0JKSv3LKy+HDz+0yzz9dBg92n5OG6q8HBYtsuu3uLhmfLt2tvZTT4WSEpgxw76vPllZ0KOHfd++/toOKyvtc2lpcPfdcNddkJx87HJ27YLcXPv3ZmTYdZKeDocP2/X29dewapX9/8rKsp+DgQNrrzewz69cCZtD9Bds327Xz/79djkDBti/q3VryMuzy3rgAfjoI/ve/OQntqbzzoMRI2zGLF9uP9fz59u/r3lzO5+WLe3rDx6EvXtr/naw/wv33QePP97w96M+rl3gIi8vz0TrbIuZmeFv+MzIsG9orAMzWnbvhq++sh/m3FwbbNOm2RD46U9h3TobLps3w1//agPjjDNgzZqaD74x9h/M96UUSnp6zT9kKIcP25s///lmZNS0dPxlZNgvke3bYcmS8ILTN9/kZFvbiaqqgn37ao9LTbX/3OXlDZ9fWpoNi/R06NfPvk9btx47XYsWNY2DwOWkpECzZqGXkZEBP/oRfPll7fcwNdXe/HXsaMN+5UrYtq1mvO+LY+3ami9jsF/uvvdqxw77JdeiRc06at/eDo2x75sx9suvXz/7WfTVvWYN/OMftT8/5eU1y2naFI4cqaknKckG5YEDNcF40kl2XfiW47/efIJ99gI1a2a/UHfvtg0Hn9NOs/8r7drBHXdAaSnMnWtrKCuzyzpwAC6+GH74AS6/3P6Nvi+lfftssLdqZcN9wABbs0+XLrZBdDxE5BtjTF7Q5xIt0PPzbRdLfZKSbEvPywFeWAgvvACffWY/2H362FaqbxfMlBTo0MG2oILp1cveVq60rUX/FneXLnDzzfYfYuFC29o45RQb/mBbSKtW1b9dIiXFtrBEYPVqGD7c/oP7WlqhtvCXlcHixfYfytcS373bfulcf70Nj4ICW1tysp1mwAAoKoLZs23LKBI6d7ZfLE2b2n/cpUttqHToYP+BmzcPbz4ZGdC7d+3W9ZEj9m/0dQfu22dbjP47gPkvZ9Mmu978gydQWZn9XPTrBxMm2JbrF1/A++/X7nY0xv49q1fb97Rr15r3ftMm26Lu3r2m9TtwYO1A2rLF/v9s22YbDTffbLs4ffbssYF4+unHfrkaA3//OyxYUDOuTZua5bRoAStW2M9xWZldJ3v31rR+Bw6ETp1qluP7HGzfXns5ycn2iynUL8mWLe3fnpxc80uyvNyuk1desf9Pjz1Wu+VfVWV/jTz/PPzhD3DJJaHfi2hpNIHu2/hZX3+5CMyc6c0wX78eLroINm60H8JmzeDCC+0//tKlNjxvuskG0Jtv1vxsbN3atop697Y/U5s1sx/2aPczKqUiq65Ad+0CF5Hm6zcPJ8wnTvRmmBsDP/uZbcH96le2xTxmTOj+4sGDaz8+++zo16iUck/CBPrkyfVvBA3cQOk106bBxx/b4YQJblejlIo3CRPo69eHfi49PfK7B8VaQYHdM+DSS+0vEaWUCpQQp8/Nzw/dF5yU5O0wr6yEt96Cq6+2/eSvvab93kqp4BIi0CdPDr63hYj392SZOBFGjbJfTO+8Y3fJVEqpYBIi0EPt+maMt8N8/357cM/NN9vd8XIT6kzzSqlI83yg5+cHPzAF7EZQL3v/fbs/9S232Ba6UkrVxdOBXteuipE+6Y0b5syx/ebnn+92JUopL/B0oIfaVdHrG0LBdre8/z5cc422zpVS4fF0oIfqO6+q8naYg92b5eBBe5i7UkqFw9OBHurE8JE8YbwbKivhiSfsOSsCj/ZUSqlQPB3oU6Yce+KfROg7f+ste2Kj++7Tfc6VUuHzdKCPGWP7yn1niovGCePd8Je/2HNJjxjhdiVKKS/x/KH/Y8Z4P8D9VVTY82SPH68bQ5VSDePpFnoiMAZefRUGDYKHH7anwD14EM491+3KlFJe4/kWutfNmmXP4Z6WZk/q77scmQa6UqqhPNtC910EukkTO8zPd7ui4/POO/ZSYG+9Za/K8uc/2/Ocd+7sdmVKKa/xZKD7jhBdv952Waxfbx97LdQrKuxVhIYPh6FD7VWFtmyxrXPdu0Up1VCeDPRgR4geOGDHe8kXX9hrIl5+ub1k3MiRdrx2tyiljocnAz3UEaKhxser99+3Fw0eOtQ+vvlm24V08cXu1qWU8iZPbhTt0iX4FYq8doTohx/CBRfYq5wDDBlir3Luf5VxpZQKlydb6IlwhOju3bBsmQ1xfxrmSqnj5clAT4QjRL/+2g61v1wpFSme7HIB7x8h+uWXtr984EC3K1FKJQpPttATwZdfwo9+VNN/rpRSJ0oD3QVVVbbLRbtblFKR5MlA9/pRoitX2v3PNdCVUpHkuT5031GivgOLfEeJgnf61BcutMNzznG3DqVUYvFcCz0RjhJduRJSU+G009yuRCmVSDwX6IlwlOiqVdCzp57vXCkVWZ4L9ES4juiqVdCrl9tVKKUSjecC3etHiR4+bK8XqoGulIo0zwW6148SXbsWjh7VQFdKRZ7n9nIBbx8lumqVHWqgK6UizXMtdK/zBXrPnu7WoZRKPGEFuogME5HVIlIkIvcHeb6ViLwnIktEZIWI3BL5UhPDqlWQlQUZGW5XopRKNPUGuogkAc8ClwG9gdEi0jtgsjuA74wx/YEhwJ9FpGmEa00IuoeLUipawmmhDwSKjDHrjDFHgFnAiIBpDNBCRATIAHYClRGtNAEYo4GulIqecAK9E7DR73GJM87fM8AZwGZgGfDfxpiqwBmJyHgRKRCRgtLS0uMs2bu2bIHycg10pVR0hBPowa4/bwIeXwoUAqcAOcAzItLymBcZM90Yk2eMyWvXrl2Di/U63cNFKRVN4QR6CdDZ73EWtiXu7xbgbWMVAT8AGlsBVq+2Qw10pVQ0hBPoi4AeItLN2dB5PfBuwDQbgIsBRKQDcDqwLpKFJoJVq+zeLaec4nYlSqlEVO+BRcaYShG5E/gISAJmGGNWiMhE5/lpwKPAyyKyDNtFc58xZkcU6/Yk3wZRCdaJpZRSJyisI0WNMfOAeQHjpvnd3wxcEtnS6vbJJ9C2LfTvH8ulnphVq+D8892uQimVqDx3pKgx8Ic/wMUXw8MPu11N+Pbvt6f41f5zpVS0eC7QX3wRHnjA3j982N1aGmLNGjvUQFdKRYvnTs514422lf7SS1DpoUOXdJdFpVS0ea6FnpYGt90GKSlQUeF2NeFbs8ZuDNXLzimlosVzge6TkuKtFnpRkT0pV1qa25UopRKVZwM9Odlbgb52LZx6qttVKKUSmQZ6jKxdq90tSqno8myge6kPvbwctm/XFrpSKro8G+heaqGvc06CoIGulIomDfQYWLvWDjXQlVLR5OlA90qXS1GRHWqgK6WiybOB7qXdFteuhcxMaNXK7UqUUonMs4HutS4XbZ0rpaLN04HulS4XDXSlVCx4NtC90uVSWWnPstitm9uVKKUSnWcD3StdLtu3Q1WVPexfKaWiSQM9yjY7V1/Vy84ppaLNs4HulSNFNdCVUrHi2UBPTrZdGVVVbldSNw10pVSseDrQAY4edbeO+mzeDE2aQPv2bleilEp0ng/0eO922bwZOnSoqVcppaLFs4GekmKH8b5hdPNm7W5RSsWGZwPd1+LVQFdKKcvzge6FLhcNdKVULHg20L3Q5XLkCJSWaqArpWLDs4HuhS6XrVvtUANdKRULGuhRpPugK6ViybOB7utyiec+dA10pVQseTbQtYWulFK1aaBH0aZNts7MTLcrUUo1Bp4NdC90uWzYAJ0720P/lVIq2jwbNV5ooW/YAF26uF2FUqqx0ECPoo0bbQtdKaViwfOBHq9dLkePQkmJttCVUrHj2UCP9yNFt2yxoa6BrpSKFc8Gerx3uWzYYIca6EqpWNFAjxINdKVUrIUV6CIyTERWi0iRiNwfYpohIlIoIitEZEFkyzxWvO+26At03SiqlIqVeq+jIyJJwLPAj4ESYJGIvGuM+c5vmtbAc8AwY8wGEYn6Bde80EJv3RpatnS7EqVUYxFOC30gUGSMWWeMOQLMAkYETHMD8LYxZgOAMWZ7ZMs8VrwH+saN2t2ilIqtcAK9E7DR73GJM85fT6CNiHwqIt+IyM3BZiQi40WkQEQKSktLj69ihxe6XLS7RSkVS+EEugQZZwIeJwMDgMuBS4GHRKTnMS8yZroxJs8Yk9euXbsGF1trgXHeQtejRJVSsRbOtehLAP+2ZhawOcg0O4wx+4H9IvIZ0B9YE5Eqg4jnQN+zB3buhK5d3a5EKdWYhNNCXwT0EJFuItIUuB54N2CavwPni0iyiKQDZwMrI1tqbfF8YNHSpXbYt6+7dSilGpd6W+jGmEoRuRP4CEgCZhhjVojIROf5acaYlSLyIbAUqAL+aoxZHtXC4/jQ/8JCO8zJcbcOpVTjEk6XC8aYecC8gHHTAh7/CfhT5Eo7Vn4+TJ5s+6ezsuy4eGyhFxZCu3bQsaPblSilGpOwAj0e5OfD+PFw4IB9vNHZ76agwL2aQikstK1zCbY5WSmlosQzh/5PnlwT5v4++ij2tdSlogKWL9fuFqVU7Hkm0H2H0gfasye2ddRn1So4ckQDXSkVe54J9FD7dMfbofVLlthh//7u1qGUanw8E+hTpkB6+rHjBw+OfS11+fJLSEuD0093uxKlVGPjmUAfMwamT7cH64jYYVoa9DzmeFT37NsHr70GV11Vs1ulUkrFimcCHWyoFxdDVZUdNmsWX7stvvYa7N0Lv/iF25UopRojTwV6oJSU+Al0Y+CZZyA3F8491+1qlFKNkacDPTk5fo4ULS6GFSvg1lt1/3OllDs8H+jx0kIvLrbDXr1cLUMp1Yh5OtDjqcvFF+jZ2W5WoZRqzDwd6PHU5bJ+ve1q0YtaKKXc4vlAj6cW+imnQNOmbleilGqsNNAjZP167W5RSrnL04GekhJfXS56hSKllJs8Hejx0kI/etSezldb6EopN2mgR8DmzbYObaErpdzk6UCPl90WdZdFpVQ88HSgx8tui+vX26G20JVSbvJ8oMdTCz3UOduVUioWPB3o8dLlsn49dOhgz/6olFJu8XSgx0uXy/btcPLJblehlGrsPB/o8dBC37EDMjPdrkIp1dh5OtDjpcultFQDXSnlPk8Herx0uWgLXSkVDzwf6G630CsrYdcuaNfO3TqUUkoD/QSVldmhttCVUm7zdKDHQx/6jh12qIGulHKbpwM9HvrQfYGuXS5KKbd5PtC1ha6UUpanAz0eulxKS+1QA10p5TZPB7qvy8UY92rwtdDbtnWvBqWUggQIdICqKvdq2LEDWraE1FT3alBKKfB4oKek2KGb3S56lKhSKl54OtB9LXQ393TRo0SVUvEiIQLdzRa6BrpSKl5ooJ+g0lLdB10pFR/CCnQRGSYiq0WkSETur2O6s0TkqIiMilyJocVDH7q20JVS8aLeQBeRJOBZ4DKgNzBaRHqHmO6PwEeRLjIUt/vQDxyAgwc10JVS8SGcFvpAoMgYs84YcwSYBYwIMt0vgLeA7RGsr05ud7noQUVKqXgSTqB3Ajb6PS5xxlUTkU7ASGBaXTMSkfEiUiAiBaW+NDwBbne5LFtmh6ed5s7ylVLKXziBLkHGBR6b+TRwnzHmaF0zMsZMN8bkGWPy2kVgS6LbXS4LFkDTpnD22e4sXyml/CWHMU0J0NnvcRawOWCaPGCWiABkAsNFpNIYMzciVYbgdpfLggU2zJs1c2f5SinlL5wW+iKgh4h0E5GmwPXAu/4TGGO6GWOyjTHZwJvApGiHObjb5VJeDosXw4UXxn7ZSikVTL0tdGNMpYjcid17JQmYYYxZISITnefr7DePJje7XD7/HI4e1UBXSsWPcLpcMMbMA+YFjAsa5MaYcSdeVnh8J8Q6dChWS6yxYIH9Qjn33NgvWymlgvH0kaKtW9vh7t2xXe62bfDCCzBkCDRvHttlK6VUKJ4O9DZt7HDXrtgt0xiYNAn27YOpU2O3XKWUqk9YXS7xyo0W+rJl8PbbMGUKnHFG7JarlFL18XQLvVUrO4xlC33LFjscMiR2y1RKqXB4OtCTkmyox7KF7luW79eBUkrFC08HOthgjWULXQNdKRWvPB/obdpooCulFCRIoMeyy2XXLnv+Fj3cXykVbzwf6G50ubRuDRLslGVKKeUizwe6G10u2t2ilIpHng/01q1jv5eLBrpSKh55PtDbtLGXgjtyJDbL00BXSsWrhAh0iF23y65dGuhKqfjk+UCP9eH/u3fXfIkopVQ88Xygx7KFbox2uSil4lfCBHosWuiHDtm+eg10pVQ88nyg+8I1Fi10PUpUKRXPPB/osexy0UBXSsUzzwd6LDeK+r40NNCVUvHI84GelmZv2kJXSjV2ng90iN3h/75A190WlVLxKCECPVaH/2sLXSkVzxIi0Nu0gZ07o78cX6D7Ln2nlFLxJCECvXt3WLMm+svZvbumz14ppeJNQgR6v36waROUlUV3OXoeF6VUPEuYQAdYtiy6y9HD/pVS8SzZ7QIiwT/QhwyJ3nJ27tRAV4mhoqKCkpISDh065HYpKoS0tDSysrJISUkJ+zUJEegnnwyZmbB0aXSXs3Il/PjH0V2GUrFQUlJCixYtyM7ORvR6inHHGENZWRklJSV069Yt7NclRJeLiG2lRzPQt2+HLVsgJyd6y1AqVg4dOkTbtm01zOOUiNC2bdsG/4JKiEAHG+jLl8PRo9GZ/5Ildti/f3Tmr1SsaZjHt+N5fxIq0A8cgHXrojP/wkI71EBXSsWrhAn0AQPs8NNPozP/wkLo3Bnato3O/JWKZ/n5kJ0NTZrYYX7+ic2vrKyMnJwccnJyOPnkk+nUqVP14yP1XCC4oKCAX/7yl/UuY9CgQSdWpAclxEZRgL594fTT7QftttsiP//CQu0/V41Tfj6MH29/AQOsX28fA4wZc3zzbNu2LYXOz95HHnmEjIwM7r333urnKysrSU4OHk95eXnk5eXVu4wvvvji+IrzsIRpoYvAjTfCggWwYUNk533wIKxapYGuGqfJk2vC3OfAATs+ksaNG8c999zDRRddxH333cfChQsZNGgQZ555JoMGDWL16tUAfPrpp1xxxRWA/TK49dZbGTJkCN27d2fq1KnV88vIyKiefsiQIYwaNYpevXoxZswYjDEAzJs3j169ejF48GB++ctfVs/XX3FxMeeffz65ubnk5ubW+qJ44okn6Nu3L/379+f+++8HoKioiKFDh9K/f39yc3NZu3ZtZFdUHRKmhQ5www3w0EPw+uvgrNuIWLYMqqo00FXjFKqBFOmGE8CaNWuYP38+SUlJ7N27l88++4zk5GTmz5/PAw88wFtvvXXMa1atWsW//vUvysvLOf3007n99tuP2Xf722+/ZcWKFZxyyimcd955fP755+Tl5TFhwgQ+++wzunXrxujRo4PW1L59e/75z3+SlpbG999/z+jRoykoKOCDDz5g7ty5fP3116Snp7PTOaHUmDFjuP/++xk5ciSHDh2iqqoq8isqhIQK9O7dYdAgePPNyAb6++/bvsPzzovcPJXyii5dbDdLsPGRdu2115KUlIo72qIAAA6SSURBVATAnj17GDt2LN9//z0iQkVFRdDXXH755aSmppKamkr79u3Ztm0bWVlZtaYZOHBg9bicnByKi4vJyMige/fu1ft5jx49munTpx8z/4qKCu68804KCwtJSkpijXPiqPnz53PLLbeQnp4OwEknnUR5eTmbNm1i5MiRgD04KJbC6nIRkWEislpEikTkmKgUkTEistS5fSEiru0Lcu65sGJF5HZfNAbeeAMuuAA6dIjMPJXykilTwMmsaunpdnykNW/evPr+Qw89xEUXXcTy5ct57733Qu6TnZqaWn0/KSmJysrKsKbxdbvU56mnnqJDhw4sWbKEgoKC6o22xphjdi0Md57RUm+gi0gS8CxwGdAbGC0ivQMm+wG40BjTD3gUOPZrLkb69IFDhyK3++KKFbb//Kc/jcz8lPKaMWNg+nTo2tVuq+ra1T4+3g2i4dqzZw+dOnUC4OWXX474/Hv16sW6desoLi4GYPbs2SHr6NixI02aNGHmzJkcdVqLl1xyCTNmzOCAs4Fh586dtGzZkqysLObOnQvA4cOHq5+PhXBa6AOBImPMOmPMEWAWMMJ/AmPMF8YY3zWDvgKycEmfPna4YkVk5jdnju1uufrqyMxPKS8aMwaKi+22pOLi6Ic5wK9//Wt+85vfcN5551WHaCQ1a9aM5557jmHDhjF48GA6dOhAqyAXO5g0aRKvvPIK55xzDmvWrKn+FTFs2DCuvPJK8vLyyMnJ4cknnwRg5syZTJ06lX79+jFo0CC2bt0a8dpDkfp+IojIKGCYMebnzuObgLONMXeGmP5eoJdv+lDy8vJMQUHB8VVdh/JyaNnS/hx84IETm9fq1TB4sD1o6eOPI1OfUvFg5cqVnHHGGW6X4bp9+/aRkZGBMYY77riDHj16cPfdd7tdVrVg75OIfGOMCbrfZjgt9GDHnwb9FhCRi4CfAfeFeH68iBSISEFpaWkYi264Fi3sxpoTbaFv3WpPxNWkCTz/fGRqU0rFlxdeeIGcnBz69OnDnj17mDBhgtslnZBw9nIpATr7Pc4CNgdOJCL9gL8Clxljgl5qwhgzHad/PS8vL2pbD/r0OfFA/9vfYONGWLQIevaMTF1Kqfhy9913x1WL/ESF00JfBPQQkW4i0hS4HnjXfwIR6QK8DdxkjInBxeDq1qeP3ZDp3+32wgswfHj481iwAE49FcI4IE0ppeJCvYFujKkE7gQ+AlYCbxhjVojIRBGZ6Ez2MNAWeE5ECkUk8p3jDdC7Nxw+XHtPl7lz4YMPYP/++l9fVQX//jdceGH0alRKqUgL68AiY8w8YF7AuGl+938O1LkRNJZ8e7osWwY9etj7vnOlr1tnz/tSl+XL7dWJNNCVUl6SMOdy8de3LyQnw8KF9vHOnVBSYu+Hc1qFBQvsUANdKeUlCRnozZrBmWfCl1/ax/4Xjw430Lt2tTelVOQNGTKEjz76qNa4p59+mkmTJtX5Gt+uzsOHD2f37t3HTPPII49U7w8eyty5c/nuu++qHz/88MPMnz+/IeXHrYQMdLCnAFi0CCoqarpbmjatP9ArKuCTT+Cii6Jfo1KN1ejRo5k1a1atcbNmzQp5gqxA8+bNo/VxXrE9MNB/97vfMXTo0OOaV7xJqJNz+Tv3XJg61Yb5kiX2ItLZ2VBUVPfrPv4Ydu3SI0NV43HXXTVX5IqUnBx4+unQz48aNYoHH3yQw4cPk5qaSnFxMZs3b2bw4MHcfvvtLFq0iIMHDzJq1Ch++9vfHvP67OxsCgoKyMzMZMqUKbz66qt07tyZdu3aMcC52s0LL7zA9OnTOXLkCKeddhozZ86ksLCQd999lwULFvD73/+et956i0cffZQrrriCUaNG8fHHH3PvvfdSWVnJWWedxfPPP09qairZ2dmMHTuW9957j4qKCubMmUOvXr1q1VRcXMxNN93EfmfPi2eeeab6IhtPPPEEM2fOpEmTJlx22WU8/vjjFBUVMXHiREpLS0lKSmLOnDmceuqpJ7TeE7qFDvDVVzbU+/WzuyHW10J/4w17pOkll0S/RqUaq7Zt2zJw4EA+/PBDwLbOr7vuOkSEKVOmUFBQwNKlS1mwYAFL67j6+zfffMOsWbP49ttvefvtt1m0aFH1c1dffTWLFi1iyZIlnHHGGbz44osMGjSIK6+8kj/96U8UFhbWCtBDhw4xbtw4Zs+ezbJly6isrOR5v6MKMzMzWbx4MbfffnvQbh3faXYXL17M7Nmzq6+q5H+a3SVLlvDrX/8asKfZveOOO1iyZAlffPEFHTt2PLGVSgK30Lt0gY4d4e9/t3utTJgAzZvbU+tWVEDA6ZI5csTu6jh3LowYAX4nZ1MqodXVko4mX7fLiBEjmDVrFjNmzADgjTfeYPr06VRWVrJlyxa+++47+vXrF3Qe//73vxk5cmT1KWyvvPLK6ueWL1/Ogw8+yO7du9m3bx+XXnppnfWsXr2abt260dM5knDs2LE8++yz3HXXXYD9ggAYMGAAb7/99jGvj4fT7CZsoIvY87DMmWMfDxli93Y5ehT+8x/brTJypJ3uu+/gv/4LSkvtPujXXutq6Uo1CldddRX33HMPixcv5uDBg+Tm5vLDDz/w5JNPsmjRItq0acO4ceNCnjbXJ/AUtj7jxo1j7ty59O/fn5dffplP67ngcH3ntfKdgjfUKXr9T7NbVVVVHdKxPM1uwna5ADz1FLz3HmzebFvdvl9Xl10G11wDjz9u94T58Y9tsN9zj71WYj1f5EqpCMjIyGDIkCHceuut1RtD9+7dS/PmzWnVqhXbtm3jgw8+qHMeF1xwAe+88w4HDx6kvLyc9957r/q58vJyOnbsSEVFBfl+V7Vu0aIF5eXlx8yrV69eFBcXU+RsaJs5cyYXNmDf5Xg4zW7CttABOnWyNx9foKemwtChNWdjzMy0e7bUd8CRUiqyRo8ezdVXX129x0v//v0588wz6dOnD927d+e8ei4Tlpuby3XXXUdOTg5du3bl/PPPr37u0Ucf5eyzz6Zr16707du3OsSvv/56brvtNqZOncqbb75ZPX1aWhovvfQS1157bfVG0YkTJx6zzFAmTZrENddcw5w5c7joootqnWa3sLCQvLw8mjZtyvDhw3nssceYOXMmEyZM4OGHHyYlJYU5c+bQvXv3sJcXTL2nz42WaJ0+ty7GwIMPwpVXQm4u/M//2MC/+WZ7lkalGgs9fa43NPT0uQndQg8kUvuyWY895l4tSikVaQndh66UUo2JBrpSjZTbFzRWdTue90cDXalGKC0tjbKyMg31OGWMoaysrMH7pzeqPnSllJWVlUVJSQnRuhSkOnFpaWlkZWU16DUa6Eo1QikpKXTr1s3tMlSEaZeLUkolCA10pZRKEBroSimVIFw7UlRESoH1x/nyTGBHBMuJpHitTetqmHitC+K3Nq2rYY63rq7GmHbBnnAt0E+EiBSEOvTVbfFam9bVMPFaF8RvbVpXw0SjLu1yUUqpBKGBrpRSCcKrgT7d7QLqEK+1aV0NE691QfzWpnU1TMTr8mQfulJKqWN5tYWulFIqgAa6UkolCM8FuogME5HVIlIkIve7WEdnEfmXiKwUkRUi8t/O+EdEZJOIFDq34S7UViwiy5zlFzjjThKRf4rI986wjQt1ne63XgpFZK+I3OXGOhORGSKyXUSW+40LuY5E5DfOZ261iETtqrMh6vqTiKwSkaUi8o6ItHbGZ4vIQb/1Ni3GdYV832K1vuqobbZfXcUiUuiMj8k6qyMfovsZM8Z45gYkAWuB7kBTYAnQ26VaOgK5zv0WwBqgN/AIcK/L66kYyAwY9wRwv3P/fuCPcfBebgW6urHOgAuAXGB5fevIeV+XAKlAN+czmBTDui4Bkp37f/SrK9t/OhfWV9D3LZbrK1RtAc//GXg4luusjnyI6mfMay30gUCRMWadMeYIMAsY4UYhxpgtxpjFzv1yYCXQqe5XuWoE8Ipz/xXgKhdrAbgYWGuMOd6jhU+IMeYzYGfA6FDraAQwyxhz2BjzA1CE/SzGpC5jzD+MMZXOw6+Ahp1TNUp11SFm66u+2kREgJ8Cf4vW8kPUFCofovoZ81qgdwI2+j0uIQ5CVESygTOBr51Rdzo/j2e40bUBGOAfIvKNiIx3xnUwxmwB+2ED2rtQl7/rqf1P5vY6g9DrKJ4+d7cCH/g97iYi34rIAhE5P9SLoijY+xZP6+t8YJsx5nu/cTFdZwH5ENXPmNcCXYKMc3W/SxHJAN4C7jLG7AWeB04FcoAt2J97sXaeMSYXuAy4Q0QucKGGkESkKXAlMMcZFQ/rrC5x8bkTkclAJZDvjNoCdDHGnAncA7wuIi1jWFKo9y0u1pdjNLUbDjFdZ0HyIeSkQcY1eJ15LdBLgM5+j7OAzS7VgoikYN+sfGPM2wDGmG3GmKPGmCrgBaL4UzMUY8xmZ7gdeMepYZuIdHTq7ghsj3Vdfi4DFhtjtkF8rDNHqHXk+udORMYCVwBjjNPp6vw8L3Puf4Ptd+0Zq5rqeN9cX18AIpIMXA3M9o2L5ToLlg9E+TPmtUBfBPQQkW5OK+964F03CnH65l4EVhpj/tdvfEe/yUYCywNfG+W6motIC9997Aa15dj1NNaZbCzw91jWFaBWq8ntdeYn1Dp6F7heRFJFpBvQA1gYq6JEZBhwH3ClMeaA3/h2IpLk3O/u1LUuhnWFet9cXV9+hgKrjDElvhGxWmeh8oFof8aivbU3CluPh2O3GK8FJrtYx2DsT6KlQKFzGw7MBJY5498FOsa4ru7YreVLgBW+dQS0BT4GvneGJ7m03tKBMqCV37iYrzPsF8oWoALbOvpZXesImOx85lYDl8W4riJs/6rvczbNmfYa5z1eAiwGfhLjukK+b7FaX6Fqc8a/DEwMmDYm66yOfIjqZ0wP/VdKqQThtS4XpZRSIWigK6VUgtBAV0qpBKGBrpRSCUIDXSmlEoQGulJKJQgNdKWUShD/HydU/BW8c9gSAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAEICAYAAABPgw/pAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAAgAElEQVR4nO3de3xU1bnw8d9DCARIuCYqJpCARZBrghGQm3jpqShVa7WKOSrailhbq/ZYqbbK0fp+3ra+Hg+16sFaLxWLfWvLqxZai4KoLWoE5FKCAoJEESFCCHcSnvePtYdMhplkkszMnpk8389nf/bMnjV7P7Nn5pk1a6+9tqgqxhhjUl87vwMwxhgTG5bQjTEmTVhCN8aYNGEJ3Rhj0oQldGOMSROW0I0xJk1YQjdhichCEbk21mX9JCKbReS8OKxXReQr3u3HReSn0ZRtwXbKROTVlsbZyHoniUhlrNdrEq+93wGY2BGRvUF3OwOHgDrv/o2qOjfadanq5HiUTXeqOiMW6xGRIuBjIFNVa711zwWifg9N22MJPY2oanbgtohsBr6jqotCy4lI+0CSMMakD2tyaQMCf6lF5E4R+Rx4SkR6iMgrIrJDRHZ5twuCnrNERL7j3Z4mIm+JyINe2Y9FZHILy/YTkaUiUiMii0Tk1yLyXIS4o4nxfhF521vfqyKSG/T41SKyRUSqROTuRvbPGBH5XEQygpZ9Q0RWebdHicg/RWS3iGwTkUdEpEOEdT0tIj8Lun+H95zPROT6kLIXisgKEdkjIltFZFbQw0u9+W4R2SsiZwb2bdDzx4rIeyJS7c3HRrtvGiMip3nP3y0ia0XkoqDHLhCRf3nr/FRE/sNbnuu9P7tF5EsReVNELL8kmO3wtuMkoCdQCEzHvfdPeff7AgeARxp5/mhgPZAL/AJ4UkSkBWWfB94FegGzgKsb2WY0MV4FXAecAHQAAglmMPCYt/6Tve0VEIaqLgP2AeeErPd573YdcJv3es4EzgW+20jceDGc78XzVWAAENp+vw+4BugOXAjcJCKXeI9N9ObdVTVbVf8Zsu6ewF+A2d5rewj4i4j0CnkNx+2bJmLOBF4GXvWe931grogM9Io8iWu+ywGGAq97y38IVAJ5wInAXYCNK5JgltDbjqPAvap6SFUPqGqVqr6oqvtVtQZ4ADirkedvUdUnVLUOeAbojfviRl1WRPoCZwD3qOphVX0LeCnSBqOM8SlV/VBVDwB/AIq95ZcBr6jqUlU9BPzU2weR/B6YCiAiOcAF3jJU9X1VXaaqtaq6GfifMHGE8y0vvjWqug/3Axb8+pao6mpVPaqqq7ztRbNecD8AH6nq77y4fg9UAF8PKhNp3zRmDJAN/G/vPXodeAVv3wBHgMEi0lVVd6nq8qDlvYFCVT2iqm+qDRSVcJbQ244dqnowcEdEOovI/3hNEntwf/G7Bzc7hPg8cENV93s3s5tZ9mTgy6BlAFsjBRxljJ8H3d4fFNPJwev2EmpVpG3hauOXikhH4FJguapu8eI41WtO+NyL43/hautNaRADsCXk9Y0WkcVek1I1MCPK9QbWvSVk2RYgP+h+pH3TZMyqGvzjF7zeb+J+7LaIyBsicqa3/JfABuBVEdkkIjOjexkmliyhtx2htaUfAgOB0aralfq/+JGaUWJhG9BTRDoHLevTSPnWxLgteN3eNntFKqyq/8Ilrsk0bG4B13RTAQzw4rirJTHgmo2CPY/7h9JHVbsBjwett6na7We4pqhgfYFPo4irqfX2CWn/PrZeVX1PVS/GNcfMx9X8UdUaVf2hqvbH/Uu4XUTObWUsppksobddObg26d1ee+y98d6gV+MtB2aJSAevdvf1Rp7Smhj/CEwRkfHeAcz7aPrz/jxwC+6H4/+GxLEH2Csig4CboozhD8A0ERns/aCExp+D+8dyUERG4X5IAnbgmoj6R1j3AuBUEblKRNqLyBXAYFzzSGu8g2vb/5GIZIrIJNx7NM97z8pEpJuqHsHtkzoAEZkiIl/xjpUElteF34SJF0vobdfDQCdgJ7AM+GuCtluGO7BYBfwMeAHXXz6cFseoqmuBm3FJehuwC3fQrjG/ByYBr6vqzqDl/4FLtjXAE17M0cSw0HsNr+OaI14PKfJd4D4RqQHuwavtes/djztm8LbXc2RMyLqrgCm4fzFVwI+AKSFxN5uqHgYuwv1T2Qk8ClyjqhVekauBzV7T0wzg373lA4BFwF7gn8CjqrqkNbGY5hM7bmH8JCIvABWqGvd/CMakO6uhm4QSkTNE5BQRaed167sY1xZrjGklO1PUJNpJwJ9wBygrgZtUdYW/IRmTHqzJxRhj0oQ1uRhjTJrwrcklNzdXi4qK/Nq8McakpPfff3+nquaFe8y3hF5UVER5eblfmzfGmJQkIqFnCB9jTS7GGJMmLKEbY0yasIRujDFpwvqhG9OGHDlyhMrKSg4ePNh0YeOrrKwsCgoKyMzMjPo5ltCNaUMqKyvJycmhqKiIyNcnMX5TVaqqqqisrKRfv35RP8+aXIxpQw4ePEivXr0smSc5EaFXr17N/idlCd2YNsaSeWpoyfvU5hP6zp3wzDNQW+t3JMYY0zptPqH/7GcwbRp8/euwZ4/f0RiT3qqqqiguLqa4uJiTTjqJ/Pz8Y/cPHz7c6HPLy8u55ZZbmtzG2LFjYxLrkiVLmDJlSkzWlShtOqGrwssvQ2EhLFoEt97qd0TGJJe5c6GoCNq1c/O5c1u3vl69erFy5UpWrlzJjBkzuO22247d79ChA7WN/FUuLS1l9uzZTW7jH//4R+uCTGFNJnQRyRKRd0XkAxFZKyL/GaaMiMhsEdkgIqtEZGR8wo2t9eth0ya480644QZ4/nnXBGOMccl7+nTYssVVfrZscfdbm9RDTZs2jdtvv52zzz6bO++8k3fffZexY8dSUlLC2LFjWb9+PdCwxjxr1iyuv/56Jk2aRP/+/Rsk+uzs7GPlJ02axGWXXcagQYMoKysjMLrsggULGDRoEOPHj+eWW25psib+5ZdfcskllzB8+HDGjBnDqlWrAHjjjTeO/cMoKSmhpqaGbdu2MXHiRIqLixk6dChvvvlmbHdYI6LptngIOEdV94pIJvCWiCxU1WVBZSbjLkE1ABiNu6ju6JhHG2OveFdfnDIFJkyAxx6Dp56CO+7wNy5jksHdd8P+/Q2X7d/vlpeVxXZbH374IYsWLSIjI4M9e/awdOlS2rdvz6JFi7jrrrt48cUXj3tORUUFixcvpqamhoEDB3LTTTcd12d7xYoVrF27lpNPPplx48bx9ttvU1payo033sjSpUvp168fU6dObTK+e++9l5KSEubPn8/rr7/ONddcw8qVK3nwwQf59a9/zbhx49i7dy9ZWVnMmTOHr33ta9x9993U1dWxP3QnxlGTNXR19np3M70pdBD1i4FnvbLLgO4i0ju2ocbeyy/DiBHQpw8MHQoTJ7qkXmeXtjWGTz5p3vLWuPzyy8nIyACgurqayy+/nKFDh3Lbbbexdu3asM+58MIL6dixI7m5uZxwwgls3779uDKjRo2ioKCAdu3aUVxczObNm6moqKB///7H+ndHk9Dfeustrr76agDOOeccqqqqqK6uZty4cdx+++3Mnj2b3bt30759e8444wyeeuopZs2axerVq8nJyWnpbmm2qNrQRSRDRFYCXwB/V9V3QorkA1uD7ld6y0LXM11EykWkfMeOHS2NOSYOHoS334bzz69fdv318PHHsGaNf3EZkyz69m3e8tbo0qXLsds//elPOfvss1mzZg0vv/xyxL7YHTt2PHY7IyMjbPt7uDItuahPuOeICDNnzuQ3v/kNBw4cYMyYMVRUVDBx4kSWLl1Kfn4+V199Nc8++2yzt9dSUSV0Va1T1WKgABglIkNDioTrMHncHlDVOapaqqqleXlhh/NNmIoKVxMfGdTaHzg4/v77/sRkTDJ54AHo3Lnhss6d3fJ4qq6uJj/f1QeffvrpmK9/0KBBbNq0ic2bNwPwwgsvNPmciRMnMtc7eLBkyRJyc3Pp2rUrGzduZNiwYdx5552UlpZSUVHBli1bOOGEE7jhhhv49re/zfLly2P+GiJpVi8XVd0NLAHOD3moEugTdL8A+KxVkcVZoBY+NOin6ZRToFs3sGHajXHt5HPmuF5gIm4+Z07s289D/ehHP+LHP/4x48aNoy4O7Z+dOnXi0Ucf5fzzz2f8+PGceOKJdOvWrdHnzJo1i/LycoYPH87MmTN55plnAHj44YcZOnQoI0aMoFOnTkyePJklS5YcO0j64osv8oMf/CDmryGSJq8pKiJ5wBFV3S0inYBXgZ+r6itBZS4EvgdcgDsYOltVRzW23tLSUvXzAhczZ8JDD8G+fRB8HOXcc6GmBt5917fQjImbdevWcdppp/kdhu/27t1LdnY2qsrNN9/MgAEDuO222/wO6zjh3i8ReV9VS8OVj6aG3htYLCKrgPdwbeiviMgMEZnhlVkAbAI2AE8A323pC0iUNWtg0KCGyRygtBQ++ACaOMfBGJPCnnjiCYqLixkyZAjV1dXceOONfocUE012W1TVVUBJmOWPB91W4ObYhhZfa9bUt5kHKy11yXz1ajj99MTHZYyJv9tuuy0pa+St1SbPFK2pcSdJDA09tItL6GDt6MaY1NMmE/q//uXm4RJ6URH07GkJ3RiTetpkQg/XwyVAxC1fty6xMRljTGu1yYT+wguQl+dq4+EMHOjGeTHGmFTS5hL6G2/A3//uui22i/DqBw50g3R9+WViYzMm3U2aNIm//e1vDZY9/PDDfPe7kTvGTZo0iUAX5wsuuIDdu3cfV2bWrFk8+OCDjW57/vz5/CvQ3grcc889LFq0qDnhh5VMw+y2uYT+k5/AySfDTTdFLjNwoJtbLd2Y2Jo6dSrz5s1rsGzevHlRjacCbpTE7t27t2jboQn9vvvu47zzzmvRupJVm0ro27fDW2/BLbdAp06Ry1lCNyY+LrvsMl555RUOHToEwObNm/nss88YP348N910E6WlpQwZMoR777037POLiorY6Y1x/cADDzBw4EDOO++8Y0PsgutjfsYZZzBixAi++c1vsn//fv7xj3/w0ksvcccdd1BcXMzGjRuZNm0af/zjHwF47bXXKCkpYdiwYVx//fXH4isqKuLee+9l5MiRDBs2jIqKikZfn9/D7EYzfG7aCLwXJcf1qm+oXz93wpEldJPObr0VVq6M7TqLi+HhhyM/3qtXL0aNGsVf//pXLr74YubNm8cVV1yBiPDAAw/Qs2dP6urqOPfcc1m1ahXDhw8Pu57333+fefPmsWLFCmpraxk5ciSneyeOXHrppdxwww0A/OQnP+HJJ5/k+9//PhdddBFTpkzhsssua7CugwcPMm3aNF577TVOPfVUrrnmGh577DFu9a54k5uby/Lly3n00Ud58MEH+c1vfhPx9fk9zG6bqqEHEnSgBh5J+/ZuXBdL6MbEXnCzS3Bzyx/+8AdGjhxJSUkJa9eubdA8EurNN9/kG9/4Bp07d6Zr165cdNFFxx5bs2YNEyZMYNiwYcydOzfi8LsB69evp1+/fpx66qkAXHvttSxduvTY45deeikAp59++rEBvSLxe5jdNlVDX7/eNbX06dN0WevpYtJdYzXpeLrkkku4/fbbWb58OQcOHGDkyJF8/PHHPPjgg7z33nv06NGDadOmRRw2N0Ak3CCv7gpI8+fPZ8SIETz99NMsWbKk0fU0NZ5VYAjeSEP0NrWuwDC7F154IQsWLGDMmDEsWrTo2DC7f/nLX7j66qu54447uOaaaxpdf1PaXA19wIDIvVuCDRwIGzZAE++fMaaZsrOzmTRpEtdff/2x2vmePXvo0qUL3bp1Y/v27SxcuLDRdUycOJE///nPHDhwgJqaGl5++eVjj9XU1NC7d2+OHDlybMhbgJycHGpqao5b16BBg9i8eTMbNmwA4He/+x1nnXVWi16b38PstrkaelPt5wEDB7oxXT7+2P0IGGNiZ+rUqVx66aXHml5GjBhBSUkJQ4YMoX///owbN67R548cOZIrrriC4uJiCgsLmTBhwrHH7r//fkaPHk1hYSHDhg07lsSvvPJKbrjhBmbPnn3sYChAVlYWTz31FJdffjm1tbWcccYZzJgx47htRmPWrFlcd911DB8+nM6dOzcYZnfx4sVkZGQwePBgJk+ezLx58/jlL39JZmYm2dnZMbkQRpPD58ZLoofPPXTIDc5/111w//1Nl1+xwl384vnnIcoeVcYkPRs+N7XEY/jctLBxIxw92vQB0YBhw1x7+7JlTZc1xphk0GYSeuAA56BB0ZVv3x7OOAPeCb16qjHGJKk2l9C9nklRGT3aNb145xgYkxb8amY1zdOS96nNJPTXXnN9y7t2jf45Y8a4A6OxPvnCGL9kZWVRVVVlST3JqSpVVVVkZWU163ltopfLJ5+4hH7PPc173ujRbr5sWf1tY1JZQUEBlZWV7Nixw+9QTBOysrIoKCho1nPaREJ/9llQhWuvbd7z8vOhoADefhsSeOFuY+ImMzOTfv36+R2GiZO0b3JRhaefhrPPdmO0NNfkybBgAcRgmAVjjImrtE/o69a5LotXXdWy5191FezbB0EnohljTFJK+4T+7rtuPn58y54/YYJrenn++djFZIwx8ZD2Cf2dd6Bbt+Z1VwyWkQFXXAELF9oVjIwxya3JhC4ifURksYisE5G1InLc4UERmSQi1SKy0pua2Z8kft59150gFM2AXJF84xtw5AjEYPx5Y4yJm2jSXC3wQ1U9DRgD3Cwig8OUe1NVi73pvphG2UIHDsCqVTBqVOvWU1wMIvDBB7GJyxhj4qHJhK6q21R1uXe7BlgH5Mc7sFhYscINf9vahJ6dDV/5iiV0Y0xya1ZDhIgUASVAuBFOzhSRD0RkoYgMifD86SJSLiLliTixITAOS2sTOrhaup0xaoxJZlEndBHJBl4EblXVPSEPLwcKVXUE8Ctgfrh1qOocVS1V1dK8vLyWxhy15ctdD5XevVu/ruJi2LQJqqtbvy5jjImHqBK6iGTikvlcVf1T6OOqukdV93q3FwCZIpIb00hb4MMPIVZDPxcXu7l3EW9jjEk60fRyEeBJYJ2qPhShzEleOURklLfeqlgG2lyqLqHH6mpDI0a4ubWjG2OSVTRjuYwDrgZWi0igFfkuoC+Aqj4OXAbcJCK1wAHgSvV5OLeqKti9O3YJ/eSTITfX2tGNMcmryYSuqm8B4S+vXV/mEeCRWAUVCx995OaxSugiMGQIVFTEZn3GGBNraXumaCCht/QM0XDy8uxsUWNM8krrhJ6R0bIRFiPp0QN27Yrd+owxJpbSNqF/+CEUFUFmZuzW2bOnq6HbxV6MMckobRP6Rx/Frv08oGdPd0k6GxvdGJOM0jKhq8Ynoffo4ebW7GKMSUZpmdC3b4e9e+NTQwc7MGqMSU5pmdD/9S83HzQotuu1GroxJpmlZUJfvdrNhw2L7Xqthm6MSWZpm9Bzc+HEE2O7XkvoxphklpYJfc0aVzuXRs9vbT5rcjHGJLO0S+hHj9Yn9FjLzob27a2GboxJTmmX0Ddvhn374pPQRepPLjLGmGSTdgk9cEB06ND4rN9O/zfGJKu0TehDwl4Er/Wshm6MSVZpmdD79YOcnPis32roxphklXYJPV4HRAOshm6MSVZpldAPHYL16y2hG2PaprRK6BUVUFcX34TeowdUV7vtGGNMMkmrhB7vHi5Qf7bo7t3x24YxxrRE2iX0zMzYXnYuVOBsUWt2McYkm7RK6GvWwGmnxfYqRaECNXTr6WKMSTZpldBXr45v+zlAr15uvn17fLdjjDHNlZIJ/fBhN2ZLsN27YevW+Cf0IUPcEADLl8d3O8YY01xNJnQR6SMii0VknYisFZEfhCkjIjJbRDaIyCoRGRmfcOGFFyArCzZubLh8zRo3j+cBUXAnLA0ZAu+8E9/tGGNMc0VTQ68FfqiqpwFjgJtFZHBImcnAAG+aDjwW0yiD5OW5a4ZWVjZcHq+LWoQzahS8+66LwxhjkkWTCV1Vt6nqcu92DbAOyA8pdjHwrDrLgO4i0jvm0QIFBW4emtDXrIFu3aBPn3hstaHRo6Gq6vh/CcYY46dmtaGLSBFQAoQ2OOQDW4PuV3J80kdEpotIuYiU79ixo3mReiIl9NWrXXNLrC9qEc7o0W5uzS7GmGQSdUIXkWzgReBWVd0T+nCYpxzXIKGqc1S1VFVL8/Lymhepp3Nn13UwOKGrJqaHS8CQIS4OS+jGmGQSVUIXkUxcMp+rqn8KU6QSCG7sKAA+a3144RUUNEzon37qernE+4BoQPv2UFoKb7+dmO0ZY0w0ounlIsCTwDpVfShCsZeAa7zeLmOAalXdFsM4GwhN6Ik8IBpw4YWu66K1oxtjkkU0NfRxwNXAOSKy0psuEJEZIjLDK7MA2ARsAJ4AvhufcJ3QhB7ospjIhD51qps//3zitmmMMY1p31QBVX2L8G3kwWUUuDlWQTWloAC++MINl9uxo6uh5+fXj7OSCH36wMSJMHcu/OQniTkYa4wxjUnJM0UDPV0+81rpy8uhuDjxcZSVufHXV6xI/LaNMSZUSif0ykrYuRPWrYPx4xMfx1e/6uYrVyZ+28YYE6rJJpdkFDh5qLLSneADMGFC4uM44QQ3b2GXemOMiamUTOj53ilLW7e6UQ87dnTdCBOtc2c3rszOnYnftjHGhErJhJ6T407zr6yEZcvc2CodOyY+DhE3tozV0I0xySAl29ABCgvhpZdcX3A/mlsCcnOthm6MSQ4pm9B/9St3yn9dnes+6BeroRtjkkXKJvSJE2HtWvjLX+Df/s2/OKyGboxJFinZhh6QnQ0XXOBvDFZDN8Yki5StoSeL3FyoqXFnrRpjjJ8sobdSYBTgQH94Y4zxiyX0VsrNdXNrdjHG+M0SeisFauh2YNQY4zdL6K1kNXRjTLKwhN5KVkM3xiQLS+it1LOnGwLAaujGGL9ZQm+ljAyX1K2GbozxmyX0GMjNtRq6McZ/ltBjIC/PaujGGP9ZQo8Bq6EbY5KBJfQY6N0bPv3U7yiMMW2dJfQY6N8fdu2C3bv9jsQY05ZZQo+B/v3dfNMmf+MwxrRtTSZ0EfmtiHwhImsiPD5JRKpFZKU33RP7MJObJXRjTDKIZjz0p4FHgGcbKfOmqk6JSUQpyBK6MSYZNFlDV9WlwJcJiCVlde3qerps3Oh3JMaYtixWbehnisgHIrJQRIZEKiQi00WkXETKd6RZP7/+/a2GbozxVywS+nKgUFVHAL8C5kcqqKpzVLVUVUvzAqNapQlL6MYYv7U6oavqHlXd691eAGSKSG6rIwtj7lwoKoJ27dx87tx4bKVl+veHLVugttbvSIwxbVWrE7qInCQi4t0e5a0z5hdkmzsXpk93SVPVzadPT56k3r8/1NXBJ5/4HYkxpq2Kptvi74F/AgNFpFJEvi0iM0RkhlfkMmCNiHwAzAauVFWNdaB33w379zdctn+/W54MTjnFza3ZxRjjlya7Larq1CYefwTXrTGuItV8k6VGHEjoFRVw3nn+xmKMaZtS5kzRvn2btzzRCgrg5JPhrbf8jsQY01alTEJ/4AHo3Lnhss6d3fJkIAKTJsEbb7g2fmOMSbSUSehlZTBnDhQWuuRZWOjul5X5HVm9SZPg88/hww/9jsQY0xZFc+p/0igrS64EHmrSJDdfsgQGDvQzEmNMW5QyNfRU8JWvuHb0JUv8jsQY0xZZQo8hETjrLNeObowxiWYJPcZKS2HbNrvGqDEm8Syhx9jQoW6+dq2/cRhj2h5L6DEWSOhrwl4OxBhj4scSeoz17g3du1tCN8YkniX0GBNxtXRL6MaYRLOEHgeBhG5njBpjEskSehwMHQq7d7veLsYYkyiW0ONgiHcRPmt2McYkkiX0OLCui8YYP1hCj4NevSAnBzZv9jsSY0xbYgk9DgKjQW7Z4nckxpi2xBJ6nPTtmzxXUzLGtA2W0OPEaujGmESzhB4nffvCl1/C3r1+R2KMaSssocdJYaGbW7OLMSZRLKHHSSChW7OLMSZRLKHHSd++bm41dGNMojSZ0EXktyLyhYiEPe9RnNkiskFEVonIyNiHmXp694b27a2GboxJnGhq6E8D5zfy+GRggDdNBx5rfVipLyMDCgqshm6MSZwmE7qqLgW+bKTIxcCz6iwDuotI71gFmMqs66IxJpFi0YaeD2wNul/pLWvz7OQiY0wixSKhS5hlYUcCF5HpIlIuIuU7duyIwaaTW1ERVFbC4cN+R2KMaQtikdArgT5B9wuAz8IVVNU5qlqqqqV5eXkx2HRyO/VUOHoUNm70OxJjTFsQi4T+EnCN19tlDFCtqnZpB2DQIDevqPA3DmNM29C+qQIi8ntgEpArIpXAvUAmgKo+DiwALgA2APuB6+IVbKoZONDNLaEbYxKhyYSuqlObeFyBm2MWURrJyYH8fEvoxpjEsDNF42zQIEvoxpjEsIQeZ4GErmH7/RhjTOxYQo+zQYNgzx74/HO/IzHGpDtL6HFmPV2MMYliCT3OLKEbYxLFEnqc5edDt26wcqXfkRhj0p0l9DgTgVGj4J13/I7EGJPuLKEnwJlnwurVdn1RY0x8WUJPgDFj3Jgu773ndyTGmHRmCT0BRo9282XL/I3DGJPeLKEnQM+eblwXS+jGmHiyhJ4gY8a4hG5njBpj4sUSeoKMHg1ffGFXMDLGxI8l9AQpKXFz649ujIkXS+gJMnw4tGsHK1b4HYkxJl1ZQk+Qzp3dJeksoRtj4sUSegKVlFhCN8bEjyX0BCopga1boarK70iMMenIEnoC2YFRY0w8WUJPoOJiN7dmF2NMPFhCT6DcXOjb18Z0McbEhyX0BDvzTPjnP/2OwhiTjiyhJ9jYse7A6NatfkdijEk3KZnQ586FoiJ3ok5RkbufKsaOdXOrpRtjYi2qhC4i54vIehHZICIzwzw+SUSqRWSlN90T+1CduXNh+nTYssUNdLVli7ufKkl9xAjo1An+8Q+/IzHGpJsmE7qIZAC/BiYDg4GpIjI4TNE3VbXYm+6LcZzH3H037N/fcNn+/W55KsjMdJeksx+Qw7MAAA7PSURBVIRujIm1aGroo4ANqrpJVQ8D84CL4xtWZJFGK0ylUQzHjnVdF/ft8zsSY0w6iSah5wPBh/AqvWWhzhSRD0RkoYgMCbciEZkuIuUiUr5jx44WhOu6/TVneTI66yyorYWlS/2OxBiTTqJJ6BJmWehlGpYDhao6AvgVMD/cilR1jqqWqmppXl5e8yL1PPCAG+gqWOfObnmqmDjRtaMvXOh3JMaYdBJNQq8E+gTdLwA+Cy6gqntUda93ewGQKSK5MYsySFkZzJkDhYUg4uZz5rjlqaJTJzj7bEvoxpjYiiahvwcMEJF+ItIBuBJ4KbiAiJwkIuLdHuWtN25DUJWVwebNcPSom6dSMg+YPBk2bICPPvI7EmNMumjfVAFVrRWR7wF/AzKA36rqWhGZ4T3+OHAZcJOI1AIHgCtV7eqZjZk82c0XLoQBA/yNxRiTHsSvvFtaWqrl5eW+bDtZDBzomoxefdXdV3W9X0aO9DcuY0zyEpH3VbU03GMpeaYopPbZogGXXAKLF8Pu3e7+yy/D6afD22/7G5cxJjWlZEJP9bNFAy65xHVfDBwcne/1DXrnHf9iMsakrpRM6Kl+tmjA6NFw4okukR89CgsWuOXLl/sblzEmNaVkQk+Hs0XBNRddfLFL5G+9Bdu3Q1aWJXRjTMukZEJPh7NFA/79390QAFOmuH713/42VFTYsADGmOZLyYSeDmeLBkyY4Nr+DxyAMWPga19zxwVWrfI7MmNMqmmyH3oyCpxIdPfd7oBoRkbDNvRUO9Fo6lQYPBhycqBDB7ds+XJ3dSNjjIlWSiZ0qE/a06fXHyAN9HYJfjxVjBjh5qqQl2ft6MaY5kvJJpeAdOntEkwExo1zfdL37vU7GmNMKknphJ4uvV1C/ehHsGMHPPKI35EYY1JJSif0SL1a2rVLvZOMgp15Jlx4IfziF1Bd7Xc0xphUkdIJPVxvF4C6utQ8czTY/ffDrl3wX//ldyTGmFSR0gk9MDZ6Rsbxj6V6W3pJCVx2GTz0EFTFbSBiY0w6SemEDi6pHz0a/rEtWxIbS6z953+6A6P33BP5NRpjTEDKJ3Ro/AzR3NzUbXoZPBhuuAEefRQmTYJ16/yOyBiTzNIioT/wgOvuF05VVWq3pz/2GDzxBKxe7fqq33svHDzod1TGmGSUFgm9rMydkBPJ/v3wgx8kLp5YatcOvvMdN77Lt74F993nEvvixX5HZoxJNmmR0MFd+acxVVWuFp+qTTAnngjPPeeublRXB+ec4xL8yy/D4cN+R2eMSQZpk9AjdWEMVVUF11+fmkkd4Ktfdc0vd90Ff/87XHSRO4Ywcya8+Sbs2eN3hMakn337XGXq6FHXnfjnP0/OC7yn1TVF5851TSvRdvPr1Qv++79Tb9yXgMOHYdEiePxxN6Z6XZ1b3q2bS/J9+sDJJ0OXLtCpk5s6d3bznBwYMsRd17RLl4bHIAIfiUjHJZpDNTbrMbGn6o7HZGW59+joUXc5xF273Ly6GrKz3TVu28dh1KejR91ntrbWdT3u0MHFdPiwayYNTAcOuM9z9+6xj6Ep1dUukd95J3z8sRtIb8MGeO891xx61VWue/SgQa68Knz6qZv27YPevd1UV+e+l7HYj41dUzStEnpAbm7z+26nenLfvRuWLHG1hq1b3fAHn3wC27bVfymOHAn/3Kwst88yM90HOPBFHj/ePX/nTjj1VDj3XNfbZt8+d0GOd95xj3fpAvn5bj2ZmW6d27a5dv+tW+Gss+Dyy90l93r2hC++gDVr3LRtm/sSdO3qHuvZ070X2dluPYEfn5wc92VQddOhQ+715OS4175pk1tWW1ufJGprXcLassVt8/BhNx054qbu3d36P/nElamqcj94Xbq4uUj9j5FI/baPHnWTiEs0J54Yfr/W1bntHzzoyubnQ3GxOxO4Tx9XZteu+qm62r1PBw6496y62u3rQLy1tW77+flum6Gx7d8PNTXuX9qBA+61Hjrk5gcOuG18+WX9tGuXizEwyueuXeG7x3btCiedBB07uikrq2Hs+/a5hJyZ6eaB/aZa/14EvyeBKTT1dOjglkfqotu3r/tcZGW5snV19fsldMrKcjEHphNPdK9z3z5Yvx4+/NANrxF4vzMzG/6I7N/vygbGijr1VDj/fJg9230On3zS/VN+9FFX5pRTXML+8MPIYzC1a+eSe58+cN119QMJNlebS+iBa46GDtzVWqme9Gtr6xPGrl3wwQewebNL2Dt3ui9+9+5u+vxzl7QLCtwXYu1aWLmyfl0i7uBsfr77AH/2WX2yVHXPOeUU9wH+618j/z3t1MnNDxyI3+vOyHA/WB07ukTQoYP7Uu7a5bbbp49LFnl57n7gixz4agQSuYj7Ugam2lq3/wLHZ0K1a+deX1aWK1tZ6RJsc4m4hBOo3UXzuQ68zsBrzsqq/8EMnrp0ce/1kSPu892rl1se/DlYvNjtq0OH3HTwoNsfPXq4KTu7YcIOTint27v937798VPw8tpa90PUoYNLssFTx44uUa5b57Yd+LEKt87AtH+/uwLY55+7KXAhdnCf2YEDXZIPvN+HDzf8MQ/Me/Rwg+WNHetimz/f/cCdc45b144d8PTTsGyZW8/AgW4qLHTr+PRTF0dGhvuOVVa6Ss7ll8ONNzb/s+A+D20soUPzm1+SSbt29TXApt6eRP7IbNnifgRycmD4cLftaAQu2PHqqy4h9OjhmnuGDHFJFFxC2b27vvZYU+OWHzhQX+usq6uv/XXs6L64NTUunlNOcV/A0ITRoYP7UQn8c/BTXZ3bD8uXuy96u3b1SbFHD1fDCzSJderkEmqXLsefCb1rl0sk0PBHp0uX+n8z4c6ebssOH3YJNzOz/t9fqmosoaOqTU7A+cB6YAMwM8zjAsz2Hl8FjGxqnaeffromwnPPqfbqFahj2RQ8tWvn5iK2HduObSdR2wk8r7DQ5afmAspVw+fVJnu5iEgG8GtgMjAYmCoig0OKTQYGeNN04LHm/urES1mZ+6vz3HPJUUtLJoG2SlXbjm3HtpOo7QSeF7ggTyx73EXTbXEUsEFVN6nqYWAecHFImYuBZ70fkGVAdxHpHbswW6+sDJ56KvpmAmOMibdYDyIYTULPB7YG3a/0ljW3DCIyXUTKRaR8R6ARMIECtXVVV2O35G6M8VssL8gTTUIP14s49E9GNGVQ1TmqWqqqpXmBo2E+seRujEkGjQ0u2FzRJPRKoE/Q/QLgsxaUSVrByT3S9Nxz9cMLxPtEmXbtErMdY4y/Ond2Z7nHSjQJ/T1ggIj0E5EOwJXASyFlXgKuEWcMUK2q22IXpv/KylyfY1V3UCOex8/r6qLbTmv/WSTqh8O2Y9ux7Rz/vMJCd4GeWHY5bvJEVFWtFZHvAX8DMoDfqupaEZnhPf44sAC4ANdtcT9wXexCNJGUlaXuSU7GmNiLamQBVV2AS9rByx4Puq3AzbENzRhjTHOkzWiLxhjT1llCN8aYNGEJ3Rhj0oQldGOMSRO+jbYoIjuALS18ei6wM4bhxFKyxmZxNU+yxgXJG5vF1TwtjatQVcOemelbQm8NESnXSMNH+ixZY7O4midZ44Lkjc3iap54xGVNLsYYkyYsoRtjTJpI1YQ+x+8AGpGssVlczZOscUHyxmZxNU/M40rJNnRjjDHHS9UaujHGmBCW0I0xJk2kXEIXkfNFZL2IbBCRmT7G0UdEFovIOhFZKyI/8JbPEpFPRWSlN13gQ2ybRWS1t/1yb1lPEfm7iHzkzXv4ENfAoP2yUkT2iMitfuwzEfmtiHwhImuClkXcRyLyY+8zt15EvpbguH4pIhUiskpE/iwi3b3lRSJyIGi/PR55zXGJK+L7lqj91UhsLwTFtVlEVnrLE7LPGskP8f2MRbp6dDJOuOF7NwL9gQ7AB8Bgn2LpDYz0bucAH+Iuoj0L+A+f99NmIDdk2S+Amd7tmcDPk+C9/Bwo9GOfAROBkcCapvaR975+AHQE+nmfwYwExvVvQHvv9s+D4ioKLufD/gr7viVyf0WKLeTx/wPck8h91kh+iOtnLNVq6NFcsDohVHWbqi73btcA6whzHdUkcjHwjHf7GeASH2MBOBfYqKotPVu4VVR1KfBlyOJI++hiYJ6qHlLVj3Hj/o9KVFyq+qqq1np3l+GuCJZQEfZXJAnbX03FJiICfAv4fby2HyGmSPkhrp+xVEvoUV2MOtFEpAgoAd7xFn3P+3v8Wz+aNnDXc31VRN4XkeneshPVu4qUNz/Bh7iCXUnDL5nf+wwi76Nk+txdDywMut9PRFaIyBsiMsGHeMK9b8m0vyYA21X1o6BlCd1nIfkhrp+xVEvoUV2MOpFEJBt4EbhVVfcAjwGnAMXANtzfvUQbp6ojgcnAzSIy0YcYIhJ3KcOLgP/rLUqGfdaYpPjcicjdQC0w11u0DeirqiXA7cDzItI1gSFFet+SYn95ptKw4pDQfRYmP0QsGmZZs/dZqiX0pLoYtYhk4t6suar6JwBV3a6qdap6FHiCOP7VjERVP/PmXwB/9mLYLiK9vbh7A18kOq4gk4HlqrodkmOfeSLtI98/dyJyLTAFKFOv0dX7e17l3X4f1+56aqJiauR9831/AYhIe+BS4IXAskTus3D5gTh/xlItoUdzweqE8NrmngTWqepDQct7BxX7BrAm9LlxjquLiOQEbuMOqK3B7adrvWLXAv8vkXGFaFBr8nufBYm0j14CrhSRjiLSDxgAvJuooETkfOBO4CJV3R+0PE9EMrzb/b24NiUwrkjvm6/7K8h5QIWqVgYWJGqfRcoPxPszFu+jvXE4enwB7ojxRuBuH+MYj/tLtApY6U0XAL8DVnvLXwJ6Jziu/rij5R8AawP7COgFvAZ85M17+rTfOgNVQLegZQnfZ7gflG3AEVzt6NuN7SPgbu8ztx6YnOC4NuDaVwOfs8e9st/03uMPgOXA1xMcV8T3LVH7K1Js3vKngRkhZROyzxrJD3H9jNmp/8YYkyZSrcnFGGNMBJbQjTEmTVhCN8aYNGEJ3Rhj0oQldGOMSROW0I0xJk1YQjfGmDTx/wHAMBkWHBO8VwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "acc = history.history['sparse_categorical_accuracy']\n",
    "val_acc = history.history['val_sparse_categorical_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, 'bo', label='Training acc')\n",
    "plt.plot(epochs, val_acc, 'b', label='Validation acc')\n",
    "plt.title('Training and validation accuracy')\n",
    "plt.legend()\n",
    "\n",
    "plt.figure()\n",
    "\n",
    "plt.plot(epochs, loss, 'bo', 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()"
   ]
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "tf2.5",
   "language": "python",
   "name": "tensorflow_25"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.10"
  },
  "nbpresent": {
   "slides": {
    "300ee14f-a043-486e-b274-7ff253907cd7": {
     "id": "300ee14f-a043-486e-b274-7ff253907cd7",
     "prev": "cb74e0bc-4513-4d13-b7f1-14c3078a7927",
     "regions": {
      "26dc3f39-a230-447c-af4c-f5e5b2fb7835": {
       "attrs": {
        "height": 0.8,
        "width": 0.8,
        "x": 0.1,
        "y": 0.1
       },
       "content": {
        "cell": "c58440a5-3f8f-4f37-9c79-6bf766209406",
        "part": "whole"
       },
       "id": "26dc3f39-a230-447c-af4c-f5e5b2fb7835"
      }
     }
    },
    "878aa53a-1444-4100-8f50-7a408191c579": {
     "id": "878aa53a-1444-4100-8f50-7a408191c579",
     "prev": null,
     "regions": {
      "a6c6843a-5ea6-4fbc-b890-3b4b8ae475b3": {
       "attrs": {
        "height": 0.8,
        "width": 0.8,
        "x": 0.1,
        "y": 0.1
       },
       "content": {
        "cell": "588ee1fa-64b5-453b-ade7-8e6b2515821c",
        "part": "whole"
       },
       "id": "a6c6843a-5ea6-4fbc-b890-3b4b8ae475b3"
      }
     }
    },
    "96ffe88e-7b50-43de-afdd-942e564f4e3e": {
     "id": "96ffe88e-7b50-43de-afdd-942e564f4e3e",
     "prev": "878aa53a-1444-4100-8f50-7a408191c579",
     "regions": {
      "b7e52e12-489a-468d-b10c-af2024fd2856": {
       "attrs": {
        "height": 0.8,
        "width": 0.8,
        "x": 0.1,
        "y": 0.1
       },
       "content": {
        "cell": "de829a92-1fb6-44ad-a2c6-fc1001e1f6e1",
        "part": "whole"
       },
       "id": "b7e52e12-489a-468d-b10c-af2024fd2856"
      }
     }
    },
    "cb74e0bc-4513-4d13-b7f1-14c3078a7927": {
     "id": "cb74e0bc-4513-4d13-b7f1-14c3078a7927",
     "prev": "96ffe88e-7b50-43de-afdd-942e564f4e3e",
     "regions": {
      "444878ee-68f3-4abb-acff-a7079b21e86d": {
       "attrs": {
        "height": 0.8,
        "width": 0.8,
        "x": 0.1,
        "y": 0.1
       },
       "content": {
        "cell": "25f3f538-1ee8-4d98-a6bb-14cbeb7a702d",
        "part": "whole"
       },
       "id": "444878ee-68f3-4abb-acff-a7079b21e86d"
      }
     }
    }
   },
   "themes": {}
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}