D_Feature_Extraction_VGG19_1.ipynb

{
 "cells": [
  {
   "cell_type": "raw",
   "metadata": {},
   "source": [
    "Entreno de un ConvNet VGG19 preentrenado con Imagenet y\n",
    "entrenando conjunto de datos pequeño\n",
    "Conjunto de datos de 4 clases\n",
    "Utilizaremos un conjunto de datos reducido:\n",
    "- Conjunto de entrenamiento: 1.000 muestras x clase \n",
    "- Conjunto de validación: 300 muestras x clase\n",
    "- Conjunto de test: 300 muestras x clase\n",
    "El conjunto de datos: balanced_dataset\n",
    "contiene tres carpetas, train, validation y test, cada una con 4 clases"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/usr/local/lib/python3.5/dist-packages/h5py/__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.\n",
      "  from ._conv import register_converters as _register_converters\n",
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "from  keras.preprocessing.image  import ImageDataGenerator, img_to_array, load_img \n",
    "from  keras.models  import Sequential\n",
    "from  keras.layers  import Dropout,  Flatten,  Dense\n",
    "from  keras.applications.vgg19 import  preprocess_input,  VGG19 \n",
    "from keras.optimizers import RMSprop\n",
    "from  keras.utils.np_utils import to_categorical \n",
    "import time\n",
    "import os\n",
    "import numpy as np \n",
    "import math\n",
    "import  matplotlib.pyplot  as  plt\n",
    "\n",
    "img_width = 224\n",
    "img_height = 224\n",
    "batch_size = 64 \n",
    "epochs = 80\n",
    "nb_train = 4000\n",
    "nb_validation = 1200\n",
    "nb_test = 1200\n",
    "nb_classes = 4\n",
    "\n",
    "PATH_TO_FE = \"VGG\" \n",
    "train_dir  =  \"balanced_dataset/train\"\n",
    "validation_dir = \"balanced_dataset/validation\" \n",
    "test_dir  =  \"balanced_dataset/test\" \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Load Pretrained model\n",
    "model = VGG19(include_top=False,  weights='imagenet') "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 4000 images belonging to 4 classes.\n",
      "Found 1200 images belonging to 4 classes.\n",
      "Found 1200 images belonging to 4 classes.\n"
     ]
    }
   ],
   "source": [
    "#PREPROCESSING + FEATURE EXTRACTION\n",
    "\n",
    "datagen  =  ImageDataGenerator(rescale=1. / 255)\n",
    "\n",
    "train_generator = datagen.flow_from_directory( \n",
    "    train_dir,\n",
    "    target_size = (img_width, img_height), \n",
    "    batch_size = batch_size,\n",
    "    class_mode  =  None,  \n",
    "    shuffle = False, \n",
    "    interpolation  =  'lanczos')\n",
    "\n",
    "validation_generator = datagen.flow_from_directory( \n",
    "    validation_dir, \n",
    "    target_size=(img_width, img_height), \n",
    "    batch_size=batch_size, \n",
    "    class_mode=None,\n",
    "    shuffle=False, \n",
    "    interpolation  =  'lanczos')\n",
    "\n",
    "test_generator = datagen.flow_from_directory(\n",
    "    test_dir,\n",
    "    target_size = (img_width, img_height),\n",
    "    batch_size = batch_size,\n",
    "    class_mode = None,\n",
    "    shuffle=False, \n",
    "    interpolation  =  'lanczos')\n",
    "\n",
    "###########Feature Extraction (FE)#########################\n",
    "#Train\n",
    "max_size_train = int(math.ceil(nb_train / batch_size)) \n",
    "train_features = model.predict_generator(train_generator, max_size_train)\n",
    "np.save('VGG/VGG19_FE_train_1.npy',  train_features)\n",
    "#Validation\n",
    "max_size_validation = int(math.ceil(nb_validation / batch_size)) \n",
    "validation_features  =  model.predict_generator(validation_generator,  max_size_validation) \n",
    "np.save('VGG/VGG19_FE_validation_1.npy',  validation_features)\n",
    "#Test\n",
    "max_size_test = int(math.ceil(nb_test / batch_size)) \n",
    "test_features  =  model.predict_generator(test_generator,  max_size_test)\n",
    "\n",
    "#Labels Extraction\n",
    "train_labels  =  train_generator.classes  \n",
    "train_labels = to_categorical(train_labels, num_classes=nb_classes) \n",
    "validation_labels  =  validation_generator.classes  \n",
    "validation_labels  =  to_categorical(validation_labels,  num_classes=nb_classes) \n",
    "test_labels  =  test_generator.classes  \n",
    "test_labels  =  to_categorical(test_labels,  num_classes=nb_classes) \n",
    "\n",
    "#Save Features\n",
    "train_data = np.load('VGG/VGG19_FE_train_1.npy') \n",
    "validation_data  =  np.load('VGG/VGG19_FE_validation_1.npy')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Top Model Classificator\n",
    "model = Sequential() \n",
    "model.add(Flatten(input_shape=train_data.shape[1:]))\n",
    "model.add(Dense(256,  activation='relu')) \n",
    "model.add(Dropout(0.5))\n",
    "model.add(Dense(nb_classes,  activation='softmax'))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "#Compile models\n",
    "model.compile(\n",
    "    optimizer = RMSprop(lr=2e-5),\n",
    "    loss='categorical_crossentropy',  \n",
    "    metrics=['accuracy'])\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train on 4000 samples, validate on 1200 samples\n",
      "Epoch 1/80\n",
      "4000/4000 [==============================] - 1s 243us/step - loss: 1.3044 - acc: 0.4020 - val_loss: 1.0828 - val_acc: 0.5608\n",
      "Epoch 2/80\n",
      "4000/4000 [==============================] - 1s 148us/step - loss: 1.0595 - acc: 0.5485 - val_loss: 0.9776 - val_acc: 0.5908\n",
      "Epoch 3/80\n",
      "4000/4000 [==============================] - 1s 153us/step - loss: 0.9589 - acc: 0.5962 - val_loss: 0.9162 - val_acc: 0.6283\n",
      "Epoch 4/80\n",
      "4000/4000 [==============================] - 1s 170us/step - loss: 0.8960 - acc: 0.6332 - val_loss: 0.8728 - val_acc: 0.6483\n",
      "Epoch 5/80\n",
      "4000/4000 [==============================] - 1s 155us/step - loss: 0.8389 - acc: 0.6715 - val_loss: 0.8499 - val_acc: 0.6558\n",
      "Epoch 6/80\n",
      "4000/4000 [==============================] - 1s 155us/step - loss: 0.7885 - acc: 0.6857 - val_loss: 0.7907 - val_acc: 0.6733\n",
      "Epoch 7/80\n",
      "4000/4000 [==============================] - 1s 166us/step - loss: 0.7558 - acc: 0.7020 - val_loss: 0.8024 - val_acc: 0.6708\n",
      "Epoch 8/80\n",
      "4000/4000 [==============================] - 1s 144us/step - loss: 0.7225 - acc: 0.7248 - val_loss: 0.7725 - val_acc: 0.6858\n",
      "Epoch 9/80\n",
      "4000/4000 [==============================] - 1s 148us/step - loss: 0.6970 - acc: 0.7368 - val_loss: 0.7379 - val_acc: 0.7050\n",
      "Epoch 10/80\n",
      "4000/4000 [==============================] - 1s 165us/step - loss: 0.6726 - acc: 0.7422 - val_loss: 0.7461 - val_acc: 0.6950\n",
      "Epoch 11/80\n",
      "4000/4000 [==============================] - 1s 169us/step - loss: 0.6538 - acc: 0.7580 - val_loss: 0.6938 - val_acc: 0.7217\n",
      "Epoch 12/80\n",
      "4000/4000 [==============================] - 1s 149us/step - loss: 0.6254 - acc: 0.7730 - val_loss: 0.6939 - val_acc: 0.7233\n",
      "Epoch 13/80\n",
      "4000/4000 [==============================] - 1s 174us/step - loss: 0.6010 - acc: 0.7857 - val_loss: 0.6735 - val_acc: 0.7383\n",
      "Epoch 14/80\n",
      "4000/4000 [==============================] - 1s 166us/step - loss: 0.5855 - acc: 0.7850 - val_loss: 0.7207 - val_acc: 0.6975\n",
      "Epoch 15/80\n",
      "4000/4000 [==============================] - 1s 140us/step - loss: 0.5614 - acc: 0.7955 - val_loss: 0.6622 - val_acc: 0.7392\n",
      "Epoch 16/80\n",
      "4000/4000 [==============================] - 1s 150us/step - loss: 0.5498 - acc: 0.7987 - val_loss: 0.6309 - val_acc: 0.7517\n",
      "Epoch 17/80\n",
      "4000/4000 [==============================] - 1s 158us/step - loss: 0.5362 - acc: 0.8115 - val_loss: 0.6332 - val_acc: 0.7533\n",
      "Epoch 18/80\n",
      "4000/4000 [==============================] - 1s 164us/step - loss: 0.5221 - acc: 0.8160 - val_loss: 0.6223 - val_acc: 0.7567\n",
      "Epoch 19/80\n",
      "4000/4000 [==============================] - 1s 157us/step - loss: 0.5004 - acc: 0.8235 - val_loss: 0.6105 - val_acc: 0.7533\n",
      "Epoch 20/80\n",
      "4000/4000 [==============================] - 1s 158us/step - loss: 0.4931 - acc: 0.8282 - val_loss: 0.6238 - val_acc: 0.7617\n",
      "Epoch 21/80\n",
      "4000/4000 [==============================] - 1s 164us/step - loss: 0.4822 - acc: 0.8275 - val_loss: 0.6004 - val_acc: 0.7708\n",
      "Epoch 22/80\n",
      "4000/4000 [==============================] - 1s 161us/step - loss: 0.4716 - acc: 0.8347 - val_loss: 0.6308 - val_acc: 0.7517\n",
      "Epoch 23/80\n",
      "4000/4000 [==============================] - 1s 163us/step - loss: 0.4546 - acc: 0.8482 - val_loss: 0.6039 - val_acc: 0.7550\n",
      "Epoch 24/80\n",
      "4000/4000 [==============================] - 1s 168us/step - loss: 0.4389 - acc: 0.8488 - val_loss: 0.5919 - val_acc: 0.7492\n",
      "Epoch 25/80\n",
      "4000/4000 [==============================] - 1s 159us/step - loss: 0.4340 - acc: 0.8503 - val_loss: 0.5857 - val_acc: 0.7750\n",
      "Epoch 26/80\n",
      "4000/4000 [==============================] - 1s 151us/step - loss: 0.4224 - acc: 0.8615 - val_loss: 0.6138 - val_acc: 0.7475\n",
      "Epoch 27/80\n",
      "4000/4000 [==============================] - 1s 158us/step - loss: 0.4127 - acc: 0.8615 - val_loss: 0.6065 - val_acc: 0.7542\n",
      "Epoch 28/80\n",
      "4000/4000 [==============================] - 1s 163us/step - loss: 0.4037 - acc: 0.8760 - val_loss: 0.5684 - val_acc: 0.7817\n",
      "Epoch 29/80\n",
      "4000/4000 [==============================] - 1s 141us/step - loss: 0.3868 - acc: 0.8762 - val_loss: 0.5540 - val_acc: 0.7775\n",
      "Epoch 30/80\n",
      "4000/4000 [==============================] - 1s 142us/step - loss: 0.3808 - acc: 0.8725 - val_loss: 0.5664 - val_acc: 0.7850\n",
      "Epoch 31/80\n",
      "4000/4000 [==============================] - 1s 172us/step - loss: 0.3749 - acc: 0.8775 - val_loss: 0.5636 - val_acc: 0.7808\n",
      "Epoch 32/80\n",
      "4000/4000 [==============================] - 1s 149us/step - loss: 0.3675 - acc: 0.8810 - val_loss: 0.5540 - val_acc: 0.7917\n",
      "Epoch 33/80\n",
      "4000/4000 [==============================] - 1s 138us/step - loss: 0.3604 - acc: 0.8880 - val_loss: 0.5551 - val_acc: 0.7725\n",
      "Epoch 34/80\n",
      "4000/4000 [==============================] - 1s 149us/step - loss: 0.3462 - acc: 0.8907 - val_loss: 0.5484 - val_acc: 0.7775\n",
      "Epoch 35/80\n",
      "4000/4000 [==============================] - 1s 151us/step - loss: 0.3419 - acc: 0.8965 - val_loss: 0.5526 - val_acc: 0.7867\n",
      "Epoch 36/80\n",
      "4000/4000 [==============================] - 1s 149us/step - loss: 0.3354 - acc: 0.8960 - val_loss: 0.5471 - val_acc: 0.7808\n",
      "Epoch 37/80\n",
      "4000/4000 [==============================] - 1s 154us/step - loss: 0.3209 - acc: 0.9065 - val_loss: 0.5490 - val_acc: 0.7783\n",
      "Epoch 38/80\n",
      "4000/4000 [==============================] - 1s 155us/step - loss: 0.3192 - acc: 0.9050 - val_loss: 0.5370 - val_acc: 0.7808\n",
      "Epoch 39/80\n",
      "4000/4000 [==============================] - 1s 154us/step - loss: 0.3058 - acc: 0.9070 - val_loss: 0.5238 - val_acc: 0.7950\n",
      "Epoch 40/80\n",
      "4000/4000 [==============================] - 1s 132us/step - loss: 0.3017 - acc: 0.9098 - val_loss: 0.5564 - val_acc: 0.7708\n",
      "Epoch 41/80\n",
      "4000/4000 [==============================] - 1s 162us/step - loss: 0.3013 - acc: 0.9113 - val_loss: 0.5309 - val_acc: 0.7900\n",
      "Epoch 42/80\n",
      "4000/4000 [==============================] - 1s 166us/step - loss: 0.2870 - acc: 0.9147 - val_loss: 0.5261 - val_acc: 0.7867\n",
      "Epoch 43/80\n",
      "4000/4000 [==============================] - 1s 162us/step - loss: 0.2844 - acc: 0.9177 - val_loss: 0.5286 - val_acc: 0.7850\n",
      "Epoch 44/80\n",
      "4000/4000 [==============================] - 1s 167us/step - loss: 0.2795 - acc: 0.9215 - val_loss: 0.5486 - val_acc: 0.7733\n",
      "Epoch 45/80\n",
      "4000/4000 [==============================] - 1s 163us/step - loss: 0.2750 - acc: 0.9185 - val_loss: 0.5139 - val_acc: 0.7933\n",
      "Epoch 46/80\n",
      "4000/4000 [==============================] - 1s 165us/step - loss: 0.2664 - acc: 0.9245 - val_loss: 0.5198 - val_acc: 0.7992\n",
      "Epoch 47/80\n",
      "4000/4000 [==============================] - 1s 141us/step - loss: 0.2593 - acc: 0.9265 - val_loss: 0.5196 - val_acc: 0.8008\n",
      "Epoch 48/80\n",
      "4000/4000 [==============================] - 1s 152us/step - loss: 0.2525 - acc: 0.9305 - val_loss: 0.5465 - val_acc: 0.7758\n",
      "Epoch 49/80\n",
      "4000/4000 [==============================] - 1s 158us/step - loss: 0.2512 - acc: 0.9300 - val_loss: 0.5216 - val_acc: 0.7867\n",
      "Epoch 50/80\n",
      "4000/4000 [==============================] - 1s 146us/step - loss: 0.2451 - acc: 0.9323 - val_loss: 0.5278 - val_acc: 0.7892\n",
      "Epoch 51/80\n",
      "4000/4000 [==============================] - 1s 134us/step - loss: 0.2393 - acc: 0.9333 - val_loss: 0.5298 - val_acc: 0.7792\n",
      "Epoch 52/80\n",
      "4000/4000 [==============================] - 1s 159us/step - loss: 0.2300 - acc: 0.9363 - val_loss: 0.5088 - val_acc: 0.8058\n",
      "Epoch 53/80\n",
      "4000/4000 [==============================] - 1s 165us/step - loss: 0.2267 - acc: 0.9382 - val_loss: 0.5270 - val_acc: 0.7800\n",
      "Epoch 54/80\n",
      "4000/4000 [==============================] - 1s 153us/step - loss: 0.2218 - acc: 0.9405 - val_loss: 0.5393 - val_acc: 0.7967\n",
      "Epoch 55/80\n",
      "4000/4000 [==============================] - 1s 151us/step - loss: 0.2242 - acc: 0.9373 - val_loss: 0.5689 - val_acc: 0.7717\n",
      "Epoch 56/80\n",
      "4000/4000 [==============================] - 1s 147us/step - loss: 0.2185 - acc: 0.9415 - val_loss: 0.5133 - val_acc: 0.8058\n",
      "Epoch 57/80\n",
      "4000/4000 [==============================] - 1s 156us/step - loss: 0.2094 - acc: 0.9427 - val_loss: 0.5097 - val_acc: 0.8058\n",
      "Epoch 58/80\n",
      "4000/4000 [==============================] - 1s 150us/step - loss: 0.2037 - acc: 0.9500 - val_loss: 0.4991 - val_acc: 0.8100\n",
      "Epoch 59/80\n",
      "4000/4000 [==============================] - 1s 165us/step - loss: 0.2027 - acc: 0.9472 - val_loss: 0.5097 - val_acc: 0.7992\n",
      "Epoch 60/80\n",
      "4000/4000 [==============================] - 1s 166us/step - loss: 0.1959 - acc: 0.9543 - val_loss: 0.4987 - val_acc: 0.8008\n",
      "Epoch 61/80\n",
      "4000/4000 [==============================] - 1s 133us/step - loss: 0.1934 - acc: 0.9517 - val_loss: 0.5143 - val_acc: 0.7967\n",
      "Epoch 62/80\n",
      "4000/4000 [==============================] - 1s 132us/step - loss: 0.1857 - acc: 0.9550 - val_loss: 0.5266 - val_acc: 0.7917\n",
      "Epoch 63/80\n",
      "4000/4000 [==============================] - 1s 158us/step - loss: 0.1909 - acc: 0.9507 - val_loss: 0.5283 - val_acc: 0.7825\n",
      "Epoch 64/80\n",
      "4000/4000 [==============================] - 1s 168us/step - loss: 0.1744 - acc: 0.9617 - val_loss: 0.5391 - val_acc: 0.7833\n",
      "Epoch 65/80\n",
      "4000/4000 [==============================] - 1s 147us/step - loss: 0.1818 - acc: 0.9575 - val_loss: 0.5094 - val_acc: 0.8117\n",
      "Epoch 66/80\n",
      "4000/4000 [==============================] - 1s 145us/step - loss: 0.1721 - acc: 0.9620 - val_loss: 0.5326 - val_acc: 0.8083\n",
      "Epoch 67/80\n",
      "4000/4000 [==============================] - 1s 149us/step - loss: 0.1746 - acc: 0.9597 - val_loss: 0.5058 - val_acc: 0.7967\n",
      "Epoch 68/80\n",
      "4000/4000 [==============================] - 1s 155us/step - loss: 0.1644 - acc: 0.9622 - val_loss: 0.5078 - val_acc: 0.7942\n",
      "Epoch 69/80\n",
      "4000/4000 [==============================] - 1s 152us/step - loss: 0.1574 - acc: 0.9680 - val_loss: 0.5053 - val_acc: 0.7950\n",
      "Epoch 70/80\n",
      "4000/4000 [==============================] - 1s 158us/step - loss: 0.1664 - acc: 0.9620 - val_loss: 0.4986 - val_acc: 0.8000\n",
      "Epoch 71/80\n",
      "4000/4000 [==============================] - 1s 151us/step - loss: 0.1530 - acc: 0.9675 - val_loss: 0.5049 - val_acc: 0.7908\n",
      "Epoch 72/80\n",
      "4000/4000 [==============================] - 1s 137us/step - loss: 0.1519 - acc: 0.9670 - val_loss: 0.5476 - val_acc: 0.7833\n",
      "Epoch 73/80\n",
      "4000/4000 [==============================] - 1s 146us/step - loss: 0.1501 - acc: 0.9705 - val_loss: 0.5732 - val_acc: 0.7725\n",
      "Epoch 74/80\n",
      "4000/4000 [==============================] - 1s 152us/step - loss: 0.1453 - acc: 0.9667 - val_loss: 0.5371 - val_acc: 0.8008\n",
      "Epoch 75/80\n",
      "4000/4000 [==============================] - 1s 161us/step - loss: 0.1412 - acc: 0.9730 - val_loss: 0.5101 - val_acc: 0.7950\n",
      "Epoch 76/80\n",
      "4000/4000 [==============================] - 1s 142us/step - loss: 0.1435 - acc: 0.9655 - val_loss: 0.5212 - val_acc: 0.7900\n",
      "Epoch 77/80\n",
      "4000/4000 [==============================] - 1s 152us/step - loss: 0.1332 - acc: 0.9745 - val_loss: 0.5124 - val_acc: 0.7933\n",
      "Epoch 78/80\n",
      "4000/4000 [==============================] - 1s 165us/step - loss: 0.1336 - acc: 0.9742 - val_loss: 0.5338 - val_acc: 0.7800\n",
      "Epoch 79/80\n",
      "4000/4000 [==============================] - 1s 152us/step - loss: 0.1349 - acc: 0.9695 - val_loss: 0.5228 - val_acc: 0.7875\n",
      "Epoch 80/80\n",
      "4000/4000 [==============================] - 1s 146us/step - loss: 0.1280 - acc: 0.9772 - val_loss: 0.5425 - val_acc: 0.7900\n"
     ]
    }
   ],
   "source": [
    "#Train model\n",
    "epochs = 80\n",
    "history = model.fit(\n",
    "    train_data, \n",
    "    train_labels,\n",
    "    epochs=epochs,  \n",
    "    batch_size=batch_size, \n",
    "    validation_data=(validation_data, validation_labels))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# h5py \n",
    "model.save_weights('VGG/VGG19_FE_model_1.h5')\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def print_graphics (history):\n",
    "    #Mostramos otro tipo de grafico\n",
    "    acc = history.history['acc']\n",
    "    val_acc = history.history['val_acc']\n",
    "    loss = history.history['loss']\n",
    "    val_loss = history.history['val_loss']\n",
    "\n",
    "    epochs = range(1, len(acc)+ 1)\n",
    "\n",
    "    plt.plot(epochs, acc, 'bo', label='Training acc')\n",
    "    plt.plot(epochs, val_acc, 'b', label='Validation acc')\n",
    "    #plt.tittle('Trainning and validation accuracy')\n",
    "    plt.legend()\n",
    "\n",
    "    plt.figure()\n",
    "    plt.plot(epochs, loss, 'bo', label='Training loss')\n",
    "    plt.plot(epochs, val_loss, 'b', label='Validation loss')\n",
    "    #plt.tittle('Trainning and validation loss')\n",
    "    plt.legend()\n",
    "\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def print_evaluate_model(model, history):\n",
    "    #Evaluate model\n",
    "    (loss, acc) = model.evaluate(\n",
    "        test_features, test_labels, \n",
    "        batch_size=batch_size, \n",
    "        verbose=0)\n",
    "\n",
    "    print(\"acc: {0:.2f}% - loss: {1:f}\".format(acc * 100, loss))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "acc: 76.67% - loss: 0.598405\n"
     ]
    }
   ],
   "source": [
    "print_evaluate_model(model, history)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAD8CAYAAACMwORRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvNQv5yAAAIABJREFUeJzt3Xl4VNX5B/DvmxjAAMoWFwgkaFEISyCkoAVURAVBsSKtYGoRqrS4FmwtCILaqr/Wta3Uii1FJUoprUoURRAUrcUSVlkEUwENawiLQEAT5v398c4kk8lsmcxklnw/zzNP5t65c++ZJe+c+55zzxFVBRERJZakaBeAiIjCj8GdiCgBMbgTESUgBnciogTE4E5ElIAY3ImIEhCDOxFRAmJwJyJKQAGDu4jMFpH9IrLRx+MiIn8QkSIR2SAiOeEvJhER1cZpQWwzB8CzAF7y8fjVADo5b30BPOf861ebNm00MzMzqEISEZFZvXr1AVVNC7RdwOCuqitEJNPPJtcBeEltHIOVItJCRM5V1T3+9puZmYnCwsJAhyciIjcisjOY7cKRc28H4Cu35WLnOm+FGi8ihSJSWFJSEoZDExGRN/XaoKqqs1Q1V1Vz09ICnlUQEVGIwhHcdwFo77ac7lxHRERREkyDaiALAdwpIvNgDalHAuXbfSkvL0dxcTFOnjwZhmJRpDRp0gTp6elISUmJdlGIyIeAwV1EXgVwGYA2IlIMYAaAFABQ1T8DWARgKIAiAGUAxoZamOLiYjRv3hyZmZkQkVB3QxGkqigtLUVxcTE6duwY7eIQkQ8B0zKqOlpVz1XVFFVNV9W/quqfnYEdau5Q1fNVtbuqhtwF5uTJk2jdujUDewwTEbRu3ZpnV0QhyM8HMjOBpCT7m58fuWPF3BWqDOyxj58RkfEM1rff7jt45+cD48cDO3cCqvZ3/PjIBfiYC+5ERLHCX03bW7B+7jnfwXvqVKCsrPr+y8psfSQwuLspLS1Fz5490bNnT5xzzjlo165d5fK3334b1D7Gjh2LrVu3+t1m5syZyI/k+RhRA1LXVIev5weqaXsL1p7KyoAxY2zfO31cevTll7Urb9BUNSq33r17q6fNmzfXWOfP3LmqGRmqIvZ37txaPd2vGTNm6OOPP15jvcPh0FOnToXvQHGqtp8VUSTMnauamqpq4dduqanVY4FnnJgwoWq5dWvVRo2qPz8lxda7r3O/JSfbc309XttbRkbtXjOAQg0ixsZtzb0+81dFRUXIyspCXl4eunbtij179mD8+PHIzc1F165d8fDDD1du279/f6xbtw4VFRVo0aIFJk+ejOzsbFx88cXYv38/AGDatGl45plnKrefPHky+vTpgwsvvBAff/wxAOD48eO44YYbkJWVhZEjRyI3Nxfr1q2rUbYZM2bgu9/9Lrp164af/exnsM8e2LZtGy6//HJkZ2cjJycHO3bsAAA8+uij6N69O7KzszE1UueDRHUUKB3iemzMGO+pDldtuU0bYNw436mT0lLA86S8vNzW+3LqlD03HFJTgUceCc++agjmFyASt7rW3DMywvMr6It7zf3zzz9XEdFVq1ZVPl5aWqqqquXl5dq/f3/dtGmTqqr269dP165dq+Xl5QpAFy1apKqqEydO1Mcee0xVVadOnapPP/105fb33Xefqqq+8cYbOnjwYFVVfeyxx/T2229XVdV169ZpUlKSrl27tkY5XeVwOBw6atSoyuPl5OTowoULVVX1xIkTevz4cV24cKH2799fy8rKqj03FKy5U134O+v2Vht3r02Hs9YcrVtdsg1I9Jq7rzxVpPJX559/PnJzcyuXX331VeTk5CAnJwdbtmzB5s2bazzn9NNPx9VXXw0A6N27d2Xt2dOIESNqbPPRRx9h1KhRAIDs7Gx07drV63Pfe+899OnTB9nZ2fjggw+wadMmHDp0CAcOHMC1114LwC46Sk1NxdKlSzFu3DicfvrpAIBWrVrV/o0gqiNvZ91jx1ot21dt3L02Ha5acziJABkZwIQJ9lcESE72vm1GBuBwADt2AHl5kStT3Ab3Dh1qt76umjZtWnn/888/x+9//3ssW7YMGzZswJAhQ7z2+27UqFHl/eTkZFRUVHjdd+PGjQNu401ZWRnuvPNOvPbaa9iwYQPGjRvH/udUb4JNnXg+5q0h0hW8VS3tEYuCCdZ/+pP9dTiAF1+0tIu7iKZhPMRtcH/kkei9cV9//TWaN2+OM844A3v27MHixYvDfox+/fph/vz5AIBPP/3U65nBiRMnkJSUhDZt2uDo0aP45z//CQBo2bIl0tLSUFBQAMAuDisrK8OVV16J2bNn48SJEwCAgwcPhr3c1DB4q33ffLPVWL3lud1r5r56jdRFcrL/2rI/KSlA69b2/NatAbc6GQCLK3Pn1j5Y5+UBs2ZV1eQzMmw5krV1d3Eb3KP5xuXk5CArKwudO3fGj3/8Y/Tr1y/sx7jrrruwa9cuZGVl4aGHHkJWVhbOPPPMatu0bt0aY8aMQVZWFq6++mr07Vs1R0p+fj6efPJJ9OjRA/3790dJSQmuueYaDBkyBLm5uejZsyeefvrpsJebEkdta9+udIm/RspIpFRSUy3w+qotuwdvz9RJRgbwt78BBw7Y8w8cAGbP9h5XQok5eXlVNflIp2FqCCYxH4lbOLpCJrLy8nI9ceKEqqpu27ZNMzMztby8PMqlqsLPKrH5a9SsrwZNV5dDb90VXWXw1igZyS7SsQBBNqiGY1RIioBjx45h0KBBqKiogKri+eefx2mn8eOiyMnPtxr5l19abd0z9x2oi2AoRIBWrYCjR6vX9lNTq9eK3cvWoYOlQnzVgl217IYubtMyia5FixZYvXo11q9fjw0bNuCqq66KdpEozgS6ctP9cc88eX00aroaIv2lQlyimt6IU6wKEsUxzxrt0KHAokUWpEWqctyuRs177gEOHqxZW65rjdz9WIDluc84w/uxgJoNkaxthx9r7kRxxF9t2/3qS6Bm46V7o6a3Rs9QZGTY/l5+ObRGSooc1tyJoshXzdvbcrhr28FITrYAHaj2HajmzZp5/WPNnShKghkyNtA4KJHk3sWQte/4w+DuZuDAgTUuSHrmmWcwYcIEv89r1qwZAGD37t0YOXKk120uu+wyFBb6n6TqmWeeQZlb5+GhQ4fi8OHDwRSdYpivCR1+9KPAQ8bWJ8/+4GzUjG8M7m5Gjx6NefPmVVs3b948jB49Oqjnt23bFgsWLAj5+J7BfdGiRWjRokXI+6P6EeqEDvXBNWmWtysvPYO5e56cwTv+Mbi7GTlyJN56663KiTl27NiB3bt3Y8CAAZX9znNyctC9e3e88cYbNZ6/Y8cOdOvWDYANDTBq1Ch06dIF119/feUl/wAwYcKEyuGCZ8yYAQD4wx/+gN27d2PgwIEYOHAgACAzMxMHDhwAADz11FPo1q0bunXrVjlc8I4dO9ClSxfcdttt6Nq1K6666qpqx3EpKChA37590atXL1xxxRXYt28fAOtLP3bsWHTv3h09evSoHL7gnXfeQU5ODrKzszFo0KCwvLeJKpRBsMIp0NWXL79s5fKWVmEwT3DBXOkUiVugK1TvuUf10kvDe7vnnsBXfw0bNkxff/11VbVhd++9915VtStGjxw5oqqqJSUlev7556vD4VBV1aZNm6qq6vbt27Vr166qqvrkk0/q2LFjVVV1/fr1mpycXDlksGuo3YqKCr300kt1/fr1qqqakZGhJSUllWVxLRcWFmq3bt302LFjevToUc3KytI1a9bo9u3bNTk5uXIo4B/84Af68ssv13hNBw8erCzrCy+8oJMmTVJV1fvuu0/vcXtTDh48qPv379f09HT94osvqpXVU0O+QtX9Csjk5Pq5WtPbVaKJePUlBYZEH/I3UtxTM+4pGVXF/fffjx49euCKK67Arl27KmvA3qxYsQI/+tGPAAA9evRAjx49Kh+bP38+cnJy0KtXL2zatMnroGDuPvroI1x//fVo2rQpmjVrhhEjRuDDDz8EAHTs2BE9e/YE4HtY4eLiYgwePBjdu3fH448/jk2bNgEAli5dijvuuKNyu5YtW2LlypW45JJL0LFjRwAcFhiov4t9vNW8/Y2Dwto2+ROzXSGdmYd6d91112HixIlYs2YNysrK0Lt3bwA2EFdJSQlWr16NlJQUZGZmhjS87vbt2/HEE09g1apVaNmyJW655ZY6DdPrGi4YsCGDvaVl7rrrLkyaNAnDhw/H+++/jwcffDDk4yUq9y6Jrt+zSFzs443npfZE4cCau4dmzZph4MCBGDduXLWG1CNHjuCss85CSkoKli9fjp0BWsQuueQSvPLKKwCAjRs3YsOGDQBsuOCmTZvizDPPxL59+/D2229XPqd58+Y4evRojX0NGDAAr7/+OsrKynD8+HG89tprGDBgQNCv6ciRI2jXrh0A4MUXX6xcf+WVV2LmzJmVy4cOHcJFF12EFStWYPv27QASd1hgf7Xx0tLwX+zjGpLWW22cgZ0igcHdi9GjR2P9+vXVgnteXh4KCwvRvXt3vPTSS+jcubPffUyYMAHHjh1Dly5dMH369MozgOzsbPTq1QudO3fGTTfdVG244PHjx2PIkCGVDaouOTk5uOWWW9CnTx/07dsXt956K3r16hX063nwwQfxgx/8AL1790abNm0q10+bNg2HDh1Ct27dkJ2djeXLlyMtLQ2zZs3CiBEjkJ2djRtvvDHo48QSX90PfQXzugZwV/D2NR64q7+454QOTK1QxASTmAcwBMBWAEUAJnt5PAPAewA2AHgfQHqgfXLI3/gWy5+Vt+FqI3lLTa05B2giDzlL0YVwNaiKSDKAmQCuBpAFYLSIZHls9gSAl1S1B4CHATwWht8dopB4m0ginHixD8WDYNIyfQAUqeoXqvotgHkArvPYJgvAMuf95V4eJworf2mXcF8gxIt9KB4FE9zbAfjKbbnYuc7degAjnPevB9BcRFqHUiA766BYFu3PKNCYLKHwnEeTwZziXbgaVH8B4FIRWQvgUgC7ANToASwi40WkUEQKS0pKauykSZMmKC0tjXrwIN9UFaWlpWjSpEm9Hte9ph6Oqz791cYPHGAwp/gXTD/3XQDauy2nO9dVUtXdcNbcRaQZgBtUtcaIV6o6C8AsAMjNza0RwdPT01FcXAxvgZ9iR5MmTZCenh7x47j6nntOPFGbC4dEvA+l62+aNqJEEExwXwWgk4h0hAX1UQBuct9ARNoAOKiqDgBTAMwOpTApKSmVV0ZSw+ZKvbhq6KGczGVkWM2bqCEKmJZR1QoAdwJYDGALgPmquklEHhaR4c7NLgOwVUS2ATgbwCNed0bkRzhTL57TuBE1NBKt/HZubq4GGt+cGg7PmnowXLMEMe1CDYmIrFbV3EDb8QpVCht/3RM9xzn33L62NXVe9UnkH2vuFBbB1LxTUoAzzvA+IFcwXI2qGRmsmVPDFWzNPWZHhaT4EsxVoeXlVaMqBju6onvqhQGdKHgM7hQWX34Z/n1yKFyi0DHnTmHRoUN49uM+NC4DO1HoGNypVjzHQXfNFXrsWM2hbmvLs5GUgZ0odAzuFDTPMV08J7VQ9T1Zs7dxzgONrkhEoWPOnYIWqNG0vBxo1szGZfHGfSo7NpASRRa7QlLQkpICDwMgYmkVIooMXsREIfF3IVJSEN+WcDWsElHdMC3TwLmnSjwvLHKNk+4SaDRGjudCFDsY3Bswz6tKa3thUatWtnzwIHPoRLGGwb0BC3WuUYeDeXWiWMecewMW6lWlzKsTxT4G9wbGvcE0mAZST8yrE8UHBvcGxPMiJG8NpJ4XFrlfiMQLjYjiB4N7gvHsyug+hrqvHLv7eC7uE0VznHSi+MWLmBKItzHV3cdQ9/VR88IjovjBi5gaEFdt/Uc/qlkzd42h7u83nA2kRImHXSHjXChzj7pjAylRYmLNPc6F2ledDaREiY019zjkPmRAKE0mGRnWOEpEiYs19zjgOUHGuHFV3Rn98TaGOtMwRA0Dg3uM8zZBhmtgL19SU4G5c61L4+zZ7KdO1BCxK2SMy8y0wB4MEQ7gReHlCg8ikT/W7NnAoUPANdcAF1xQP8eMR2HtCikiQ0Rkq4gUichkL493EJHlIrJWRDaIyNBQCk01BTv+S0YGLzSi8FAF1qwB7r4bOOssC7ah9sYK1u7ddob6i18AnTtbcJ84Edi1K7LHTWQBG1RFJBnATABXAigGsEpEFqrqZrfNpgGYr6rPiUgWgEUAMiNQ3ganQ4fANXfm0Ru2Z58FBgwAsrNr/9zCQuCFF6pq6KrAypXAxo1A48bAZZcBb78NDBsGFBTYNIqR8Le/2XAYy5YBn30GvPmmzSVQUAB88AHQrl1kjpvQVNXvDcDFABa7LU8BMMVjm+cB/Mpt+48D7bd3795K3s2dq5qRoSqi2rq1aqNGqvZvZ7eUFFsvYtvNnRvtElO0rF1r34lLLw3t+UOG2Pfr3HOrbt/7nupzz6kePGjbzJ2rmpxs6w8fDlvRK506pZqZqTpwYPX1n3yi2ry5aufOqvv2hf+48QpAoQaIr2pfi4DBfSSAv7gt3wzgWY9tzgXwKaxmfwhA70D7ZXCvwmCe2BwO1aVLVZ98UvWnP7Ug1rWr6pYtdd/3rbdWfU/Wrq3dcw8eVD3tNNX77gu87YIFtm1uruozz6jefrvqFVeoXnih6m23qX70kb3OUCxebOV/9dWaj61YoXr66ao9eqiWlga3vxMnVP/yFyvnqVO1L8+XX6r27av661+rfv117Z8fafUd3CcBuFerau6bASR52dd4AIUACjt06FBPb0VsmztXNTW1ejD3dsvIiHZJqbYcDtVFi1Rzcqo+x9atVS++2ALWmDF12//Bg7afkSPtO3TLLTW3OXlS9fXXVcvLaz42Z46V6b//De54BQWqjRvbc848U7VPH9VrrlFt2tTWfec7qg8/rLp3b+1ex8iR9r6cPOn98SVL7Li5uarvvqt65Ij37fbtU50xQzUtrer9Hj7c9/a+3H9/1fPT0lSfesp+MGJFOIN7MGmZTQDauy1/AeAsf/tlzd1kZAQO7IDV2il89u1TvfJKCxyR8PHHqv362WfXsaMF0gMHqh6/8047I9u1K/RjPPlkVY19wgQ74/NMX0ycaNs8/3zN519zjWqHDrWrcR84oLp/f/XnHD1qr2/gQDtWo0aq48apfvpp4P3t3WtnBBMn+t+uoEC1SZOq/4Vu3VTz8lSvv171kkvsTMj1w3PttarLl6v+4Q+WTsrKUi0qCu71ffutpaaGDVNdudLOTgDVtm1Vf/c71UOHqm/vcKhu2qS6bVtw+w+HcAb305zBuiOARgDWA+jqsc3bAG5x3u8CYDec3Sx93RpacPdMvbjSLMEE9oZScz9yxNIWH38c2eOUl6tedpm9r9nZ3oObw6G6Z09oqYYdOywQtW1ruetvvqm5TVGRff7331/7/atauuH88+0HRNVSPIDVnF1WrLBjJCVZ+sQ9RXH4sAXhSZNCO74vW7dayub00608w4bZ++jLb39r223eHHjfhw9bCufBB1UHD1Zt396C+iWXWJCfNEn1s8+qP2fpUtVWrVRbtrTnBvLaa1aeN96oWrdsWdUPV9OmqnffbWdDd91lP9yAarNmquvXB96/w2HvUUlJ4G19CVtwt31hKIBtAP4HYKpz3cMAhjvvZwH4tzPwrwNwVaB9NqTgHmzqxdctNTXx8+ylparf/a5Wnt77OkX3FErwnTTJjvP979vfd96puc1zz1XVuu+6ywJDsGW66SYL7jt3+t9uxAgLOseO1Xxs1y7Lc997rzVkjh5dPb2waJHWyFMPHqx6zjn2Y3LsmOp559lt1izbtqCgatu5c21dpH5IDxxQ/c1v7Lubnq5aWFhzG4fDPuv+/SNTBpeiIvsRAFR/9jP/efShQ+1H2Vsaa+1a1R//2M64APuMr7lG9Y9/tOd06OA9JVVYaO/FsGFWqQPs+xWqsAb3SNwaUnAPNvXSEBpQP/lEdeHC6rXZ/futBt2okTXuAVaj8+ebb6yW36pV7Rq+XnnF9n/33baPdu1q9tI4cMCCbk6O/fO60gEpKdZzY/hw1V/8wmqF3l4foDp1auCy/Pvftu2zz1atq6iw1+X6LjRqZLnt5GTVnj1Vd++27YYOrQrkLq6AP3eupX0A1Q8+sFRDerqdrbh8//v22kNpcKyNtWst6DVpYu+9u2XLrIwvvhjZMqiqlpXZj7qI9cx5772a2+zcaWc5gT67Xbvs+cePV60rLLSzlYsuqsrPHz1q6TLXZ9mli6WrXngh8A+/PwzuMSTY9EsiBPOTJy1XWVFRff3Ro3a67t6weOedlvPOyrJ/DNdp8/DhdprrKx+9b5/V9oCq2n4wDV/r19txBgywgKeq+sQTWqNRccIEC6YbN9ry8eOqb76p+qtfWW27WzfL74pUD1gOh6VJzj47uB8bh8N6ZZx/vr1f5eWWRwbsbGHlyqqzhbfftpRARobqW2/ZsadPr76/U6dUL7jAAjmges89VY89/ritW73ayta4sf3A1Yd9++w9B1Svu84+X9f71KKFBd768tFHqp06WVkeeKD6md+MGfa+fvFFaPv+5z9tv6NHW/DPzLT9TZwYfE+fYDC4R5l7jj05OXBgT5Sc+gMP2Os591zVX/7SAuTy5ZbeEFH9+c8tUN54Y1UDWNOmqu+/X7WPoiKrsd58c839e6sJujd8NW9u9x94wGqyb72lOm2a6qBBdpy2bavngI8csZ4fI0dW7T8pKXDgO37ccr3Jyar/+petW7DAyjBrVvDv1/z59py//93KAKg++qj3bVetUj3rLNvmtNO8//jNnKmVqS33muXhw/be5OWpzptn26xYEXw56+qbb+wHKz1dtXt3O4u44Yaatfn6cPy46tix9h5MnGgBvqLCcvhXXVW3fT/6aNX/9He+o/rhh+EpszsG9yiqbY49UXLq33xjtbE+fax2dtpp1b/onsHk0CHV2bNV16ypuS9XdzRXTnjXLtUpU/zncJcvt1p3z54WoF3HTk5W7dXLzhw2bar5vClT7Idn61arYbZpU3UBjz9ff20175QUa2A77zyr1XuetfhTXm41PNd79fTT/rcvKrIzndtu8/74sWN26r96dc3HJk604/TrZymd2pQz0Tgc9gMO2HemoMDu/+Mfdd/vlCl2luf+4xpODO5R5CvHnpxcs7dMrKdhysosz/ynPwXe1lULfestW963zy4kefTR2n/Rjx61nHDPnlaDT0mxgD1ypP/eF+7PX7bMAr63Bkt3e/faWUTnzrWveR86ZD8crs/YW+NsIH/6kz03mPdY1QJIKA3J27dX/ejdfnvtn59oHA4Lwq4zvrQ07z2bYg2DexT5yrFHq6/6li2Wjw6lpvbCC1b2zMzAjW+XX27bhatGmJ9vx27WzPLH//tfePbrjasRs1ev2pe/pMSed8MNoR3b4ai/y+t/+EN7ncuW1c/xYp3DYV0rAQv08YDBPYp81dyjkVcvKLBaSSjdrxwOSzO4UkyLFvne9rPP1G++OBQOhwUhzwtHIuGLL6y73MqVoT3f4Yh8z5Nw+N//rD2iIadkvHFvvI51DO5R5C3nXt95dYdD9bHH7GwhJ8f6SrdqVbtW+/fes7I//7zl0ocP973txImWOqntpedEVDvBBnfOxBQm7lPhTZ0KjBkTvRmQjh+3Y02ZAtx4I/DhhzZ86uHDwIwZwe/n978H0tKAH/8Y+MlPbBjWr76qud2JE8CcOcCIEcDZZ4ftZRBRXQTzCxCJWyLV3GOhpu6ycaP1phCxmrt7w9sdd1iD2oYNgffjujx+2jRb3r7dlh94oOa2rgGoli8PxysgIn8QZM2d0+yFga+p8DIybGakcCottckTliwBzj3XZsm5+GI7Y5g9G7jrLuCMM2wO1SuuqP7cgweBTp2AHj1sUgTXNGYnTtj9Jk2qtv35z4GZM+11tW1r64YNA9autXUpKVXbXnyxnRVs3syp0YgiLdhp9gLOxESB+ZoKL9gp8nw5fBjYts1uW7cCy5cD//mPTafXpo09/tvfAq1a2dRkH38MDBpkgf2cc2rur1Ur4De/AW6/3dI0IjbTzbJlFth/+Uvgnnts/7NnAz/8YVVgB4AJE4BrrwUWLgRuuMFmzvnzn23mnmeeYWAniinBVO8jcUuktEy4e8dUVNjgU+77SkqyhtHp0+1S+VOn7KrD+fNtMKNOnWxwokC9ICoqbOID137PO88u5rj2Wq28jH/4cLv/ySc1n9u+vV0B+vHHVeOUX3ZZbE5qQJSIwN4ykRVo9qRQc+4Oh41cB9il+m+8Yf3Uw3lxxWef2VjXW7ZUz8n/5z92mT5gE0p48+tfV73Gtm1tVMJQZ+AhotpjcI8gbw2o4RrJ0RU8J08Oa5FrZeVK1a++8v7Y3r1W8//Vr+wqUCKqX8EGd+bcg5Sfb10cv/zSGi9Pnar+eHm5zQx/4EDox/jLX4AHHrCuh48+Wrfy1kXfvr4fO/tsYP36+isLEYWG/dyDkJ8PjB9vvURUawZ2F38NqBUVwOuv219vli0DfvYzYPBgC/JsnCSiumBwD8LUqUBZWeDtOnTw/di0acD11wMLFnh//LnngLPOssfduxkSEYWCwT0IwXRpTE0FHnnE+2Nvv21dFgFg8eKaj586Bbz3ntXamzULvZxERC4M7j64DyeQ5ONdSk4OPLxAcbHl0Hv0sD7i775rqR13a9YAhw4BV14Z9pdBRA0Ug7sXweTYU1OBF1+0C3527PAe2CsqgNGj7QrQ+fOB4cOB3bvtSk53S5bYX88rSomIQsXeMl74yrEnJ1sw79DBUjC+BgJTBfbvB373O+Cjj+yK0QsvBE4/3R5fvBjo2rVq+yVLgOxsy7kTEYUDg7sXvnLsDofdfHnsMeBf/7LhAr7+2tbdemvVj0CHDjZMwLvvApMm2brjx4F//9su+yciChcGdy86dPA+EJi/3jAbNgD33w/k5gI33wxccAHQpQtw+eXVt7vqKsvPnzxp47l8+KH1kWe+nYjCiTl3Lx55xHLq7vz1hgFsBMUmTSzl8uyzwN13W8BOTq6+3eDBFtg/+siWlywLw0FOAAAQDklEQVQBGjcGBgwI72sgooaNwd2LvDyrXQc72cbhw5ZXz8uzkRf9ufRS68fu6hK5ZAnQv39VPp6IKBwY3H3Iy7NeMP56w7jMmWMNsHfcEXi/TZtaMH/3XWDvXuDTT5mSIaLwCyq4i8gQEdkqIkUiMtnL40+LyDrnbZuIHA5/USPLvV97ZqYtB8PhsJTM974H9OoV3HMGD7Yc/dy5tszgTkThFjC4i0gygJkArgaQBWC0iGS5b6OqE1W1p6r2BPBHAP+KRGEjxbNf+86dtuwZ4KdPt7TKnj1V6959FygqCq7W7nLVVfb30UeB1q2Bnj3r/hqIiNwFU3PvA6BIVb9Q1W8BzANwnZ/tRwN4NRyFqy/e+rWXldl6l/XrrUF1xQqbVu6zz2z9zJk2UuLIkcEfLzvbJp4+dMhmTvJ1BSwRUaiCCSvtALjPeV/sXFeDiGQA6AhgWd2LVn8CTZOnav3QW7YEli61K0779bO0yltvWS2/UaPgj5eUVFV7Z0qGiCIh3HXGUQAWqKrXQXFFZLyIFIpIYUlJSZgPHTpf/ddd6//xD+CDD6zmPmiQzWPaurX1Z09KAn7609of84YbrIfMkCGhl5uIyJdggvsuAO3dltOd67wZBT8pGVWdpaq5qpqblpYWfCkjxNWIunNnzfHTXf3ajx8HfvELy4vfeqs9dt55dlXpoEE2Bns7r+cx/l1/vU3skZ5e55dBRFRDMFeorgLQSUQ6woL6KAA3eW4kIp0BtATwn7CWMEJcjaiuXLuqBXhV69fuGjtm+nTgq6+AV16pfkFSWpqlaOrC80IpIqJwCRjcVbVCRO4EsBhAMoDZqrpJRB6GzeW30LnpKADznHP8xTxvjaiuwL5jhy1v326Df40ebX3TiYjihUQrFufm5mphYWFUjg1YrtzbSxexvusVFTYEb2Gh9Yxh+oSIYoGIrFbV3EDbNdhOeIEaUR96yBpRn3uOgZ2I4k+DCu7uV6EeO1az+6KrEXXJEvs7dqz1iCEiijcNZshfzwbU0lIbwKtVK7uYyDUBx8CB1jOmSxfgj3+MbpmJiELVYIK7twbU8nLg4EGgfXugb1+gpMQaT48fB95/3wb5IiKKRw0muPu6ChWwQb/+8x+b5xQA/vY3ICvL9/ZERLGuwQR3X7MrnXUWMG+e3d+9G9i1C/jud+u3bERE4dZgGlS9za6UnAw89VTVctu2DOxElBgaTHB3za7k6uqYlGTT4fmbhIOIKF41mLQMYIH84EGb3/SllxjYiShxNZjg7nAAzz8P/OpXNhPSTTVGxyEiShwNIrjv2AH85CfAsmU2pMCcOTVHgSQiSiQJn3N/6SWge3fgv/+1mvu77wLnnBPtUhERRVZC19y3bAHGjbNZk156yUZ8JCJqCBI6uP/yl3aV6YIFNv46EVFDkbBpmaVLbX7TpCSbwDoz08aXISJqCBKy5n7qlDWgigCHD9u6nTtt4DCAXSCJKPElZM19zhwbS8ZzMo6yMhtAjIgo0SVccD92DJg2zffj/gYQIyJKFAkX3B9/HNi713d3R18zMBERJZKECu6qwF//CgwbBjzxRM2BwlwzLRERJbqECu7r1tmQvSNHVg0UlpFhDasZGbbMxlQiaggSqrdMQYEF8qFDbTkvj8GciBqmhKq5v/mmTZd31lnRLgkRUXQlTHDfswdYtQq45ppol4SIKPoSJri/9Zb9vfba6JaDiCgWBBXcRWSIiGwVkSIRmexjmx+KyGYR2SQir4S3mIG9+aZ1c+zevb6PTEQUewI2qIpIMoCZAK4EUAxglYgsVNXNbtt0AjAFQD9VPSQi9Zr1PnkSWLIEuOUWjtNORAQEV3PvA6BIVb9Q1W8BzANwncc2twGYqaqHAEBV94e3mP4tX25DCzRtagOEJSVxoDAiatiCCe7tAHzltlzsXOfuAgAXiMi/RWSliAwJVwGDUVAANG5sE17v3GkXM7kGCmOAJ6KGKFwNqqcB6ATgMgCjAbwgIi08NxKR8SJSKCKFJSUlYTmwquXbk5KAEyeqP8aBwoiooQomuO8C0N5tOd25zl0xgIWqWq6q2wFsgwX7alR1lqrmqmpuWphmz9iwAfjqq5qB3YUDhRFRQxRMcF8FoJOIdBSRRgBGAVjosc3rsFo7RKQNLE3zRRjL6VNBgf1NT/f+OAcKI6KGKGBwV9UKAHcCWAxgC4D5qrpJRB4WkeHOzRYDKBWRzQCWA/ilqpZGqtBVZQNeecXmSP2//+NAYURELkGNLaOqiwAs8lg33e2+ApjkvNWbwkKbBNt9QLCpUy0V06GDBXaOLUNEDVFcDxw2Zw7QpAnwwx/aMgcKIyIycTv8wDffAK++CowYAZx5ZrRLQ0QUW+I2uBcUAIcOAWPGRLskRESxJ26D+5w5QLt2wKBB0S4JEVHsicvgvncv8M47wM03A8nJ0S4NEVHsicvgnp8PnDrFlAwRkS9xF9xVLSVz0UVA587RLg0RUWyKu+C+Zg2wcSNr7URE/sRdcH/zTRsB8sYbo10SIqLYFXfBffp0q7m3bBntkhARxa64C+4iwHe+Y/fz8zk5BxGRN3E7/EB+vk3GUVZmy67JOQAOQUBEFHc1d5epU6sCuwsn5yAiMnEb3H1NwsHJOYiI4ji4+5qEg5NzEBHFcXB/5BFOzkFE5EvcBve8PJukIyPDetBkZFSftIOIqCGL294yACfnICLyJW5r7kRE5BuDOxFRAmJwJyJKQAzuREQJiMGdiCgBMbgTESUgBnciogTE4E5ElICCCu4iMkREtopIkYhM9vL4LSJSIiLrnLdbw19UIiIKVsArVEUkGcBMAFcCKAawSkQWqupmj03/rqp3RqCMRERUS8HU3PsAKFLVL1T1WwDzAFwX2WIREVFdBBPc2wH4ym252LnO0w0iskFEFohIe287EpHxIlIoIoUlJSUhFJeIiIIRrgbVAgCZqtoDwBIAL3rbSFVnqWququampaWF6dBEROQpmOC+C4B7TTzdua6Sqpaq6jfOxb8A6B2e4hERUSiCCe6rAHQSkY4i0gjAKAAL3TcQkXPdFocD2BK+IhIRUW0F7C2jqhUicieAxQCSAcxW1U0i8jCAQlVdCOBuERkOoALAQQC3RLDMREQUgKhqVA6cm5urhYWFUTk2EVG8EpHVqpobaDteoUpElIAY3ImIEhCDOxFRAmJwJyJKQAzuREQJiMGdiCgBMbgTESUgBnciogTE4E5ElIAY3ImIEhCDOxFRAmJwJyJKQAzuREQJiMGdiCgBMbgTESUgBnciogTE4E5ElIAY3ImIEhCDOxFRAmJwJyJKQAzuREQJiMGdiCgBMbgTESUgBnciogTE4E5ElICCCu4iMkREtopIkYhM9rPdDSKiIpIbviJWyc8HMjOBpCT7m58fiaMQEcW/0wJtICLJAGYCuBJAMYBVIrJQVTd7bNccwD0APolEQfPzgfHjgbIyW96505YBIC8vEkckIopfwdTc+wAoUtUvVPVbAPMAXOdlu18D+C2Ak2EsX6WpU6sCu0tZma0nIqLqggnu7QB85bZc7FxXSURyALRX1bfCWLZqvvyyduuJiBqyOjeoikgSgKcA3BvEtuNFpFBECktKSmp1nA4dareeiKghCya47wLQ3m053bnOpTmAbgDeF5EdAC4CsNBbo6qqzlLVXFXNTUtLq1VBH3kESE2tvi411dYTEVF1wQT3VQA6iUhHEWkEYBSAha4HVfWIqrZR1UxVzQSwEsBwVS0MZ0Hz8oBZs4CMDEDE/s6axcZUIiJvAvaWUdUKEbkTwGIAyQBmq+omEXkYQKGqLvS/h/DJy2MwJyIKRsDgDgCqugjAIo91031se1ndi0VERHXBK1SJiBIQgzsRUQJicCciSkAM7kRECUhUNToHFikBsDPIzdsAOBDB4tRFrJYtVssFsGyhiNVyAbFbtlgtF1C3smWoasALhaIW3GtDRApVNSIjTdZVrJYtVssFsGyhiNVyAbFbtlgtF1A/ZWNahogoATG4ExEloHgJ7rOiXQA/YrVssVougGULRayWC4jdssVquYB6KFtc5NyJiKh24qXmTkREtRDzwT3Y+VvrqSyzRWS/iGx0W9dKRJaIyOfOvy2jUK72IrJcRDaLyCYRuSeGytZERP4rIuudZXvIub6jiHzi/Fz/7hxxtN6JSLKIrBWRN2OsXDtE5FMRWScihc51sfB5thCRBSLymYhsEZGLY6RcFzrfK9ftaxH5eYyUbaLzu79RRF51/k9E/HsW08Hdbf7WqwFkARgtIllRLNIcAEM81k0G8J6qdgLwnnO5vlUAuFdVs2Dj6d/hfJ9ioWzfALhcVbMB9AQwREQugk3J+LSqfgfAIQA/iULZAJv3d4vbcqyUCwAGqmpPty5zsfB5/h7AO6raGUA27L2LerlUdavzveoJoDeAMgCvRbtsItIOwN0AclW1G2xk3VGoj++ZqsbsDcDFABa7LU8BMCXKZcoEsNFteSuAc533zwWwNQbetzdgE5rHVNkApAJYA6Av7AKO07x9zvVYnnTYP/zlAN4EILFQLuexdwBo47Euqp8ngDMBbIezrS5WyuWlnFcB+HcslA1V05S2go3C+yaAwfXxPYvpmjuCmL81Bpytqnuc9/cCODuahRGRTAC9AHyCGCmbM/WxDsB+AEsA/A/AYVWtcG4Src/1GQD3AXA4l1vHSLkAQAG8KyKrRWS8c120P8+OAEoA/M2ZyvqLiDSNgXJ5GgXgVef9qJZNVXcBeALAlwD2ADgCYDXq4XsW68E9rqj9DEet+5GINAPwTwA/V9Wv3R+LZtlU9ZTa6XI6gD4AOkejHO5E5BoA+1V1dbTL4kN/Vc2BpSTvEJFL3B+M0ud5GoAcAM+pai8Ax+GR5oiB/4FGAIYD+IfnY9EomzPHfx3sh7EtgKaomdqNiFgP7oHmb40F+0TkXABw/t0fjUKISAossOer6r9iqWwuqnoYwHLYaWgLEXFNFhONz7UfgOHOeX/nwVIzv4+BcgGorPFBVffDcsd9EP3PsxhAsap+4lxeAAv20S6Xu6sBrFHVfc7laJftCgDbVbVEVcsB/Av23Yv49yzWg7vf+VtjxEIAY5z3x8Dy3fVKRATAXwFsUdWnYqxsaSLSwnn/dFhbwBZYkB8ZrbKp6hRVTVeb93cUgGWqmhftcgGAiDQVkeau+7Ac8kZE+fNU1b0AvhKRC52rBgHYHO1yeRiNqpQMEP2yfQngIhFJdf6fut6zyH/PotnwEWSDxFAA22B52qlRLsursLxZOawW8xNYnvY9AJ8DWAqgVRTK1R92urkBwDrnbWiMlK0HgLXOsm0EMN25/jwA/wVQBDuFbhzFz/UyAG/GSrmcZVjvvG1yfe9j5PPsCaDQ+Xm+DqBlLJTLWbamAEoBnOm2LuplA/AQgM+c3/+XATSuj+8Zr1AlIkpAsZ6WISKiEDC4ExElIAZ3IqIExOBORJSAGNyJiBIQgzsRUQJicCciSkAM7kRECej/ATcfecys5vVeAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f1a98f04e10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7f1a98ee01d0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "print_graphics (history)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}