Image Classifier with Image Augmentation.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense, Dropout, Activation\n",
    "from tensorflow.keras.layers import Flatten, Conv2D, MaxPooling2D\n",
    "from tensorflow.keras.optimizers import SGD\n",
    "from tensorflow.keras.callbacks import TensorBoard\n",
    "import pickle\n",
    "import time\n",
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "from keras.callbacks import ReduceLROnPlateau\n",
    "from keras.layers.normalization import BatchNormalization\n",
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "#NAME = 'CatsDogs-CNN{}'.format(int(time.time()))\n",
    "#tensorboard = TensorBoard(log_dir='logs/{}'.format(NAME))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = pickle.load(open('X.pickle', 'rb'))\n",
    "y = pickle.load(open('y.pickle', 'rb'))\n",
    "\n",
    "X = X/255.0\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "#CNN model\n",
    "model = Sequential()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d_8 (Conv2D)            (None, 50, 50, 32)        320       \n",
      "_________________________________________________________________\n",
      "conv2d_9 (Conv2D)            (None, 50, 50, 32)        9248      \n",
      "_________________________________________________________________\n",
      "max_pooling2d_4 (MaxPooling2 (None, 25, 25, 32)        0         \n",
      "_________________________________________________________________\n",
      "dropout_6 (Dropout)          (None, 25, 25, 32)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_10 (Conv2D)           (None, 25, 25, 64)        18496     \n",
      "_________________________________________________________________\n",
      "conv2d_11 (Conv2D)           (None, 25, 25, 64)        36928     \n",
      "_________________________________________________________________\n",
      "max_pooling2d_5 (MaxPooling2 (None, 12, 12, 64)        0         \n",
      "_________________________________________________________________\n",
      "dropout_7 (Dropout)          (None, 12, 12, 64)        0         \n",
      "_________________________________________________________________\n",
      "flatten_2 (Flatten)          (None, 9216)              0         \n",
      "_________________________________________________________________\n",
      "dense_4 (Dense)              (None, 256)               2359552   \n",
      "_________________________________________________________________\n",
      "dropout_8 (Dropout)          (None, 256)               0         \n",
      "_________________________________________________________________\n",
      "dense_5 (Dense)              (None, 1)                 257       \n",
      "=================================================================\n",
      "Total params: 2,424,801\n",
      "Trainable params: 2,424,801\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.add(Conv2D(32,(3,3), padding='same',\n",
    "                 activation='relu', input_shape = X.shape[1:]))\n",
    "model.add(Conv2D(32,(3,3), padding='same',\n",
    "                activation='relu'))\n",
    "model.add(MaxPooling2D(pool_size=(2,2)))\n",
    "model.add(Dropout(0.4))\n",
    "\n",
    "model.add(Conv2D(64,(3,3), padding='same',\n",
    "                activation='relu'))\n",
    "model.add(Conv2D(64,(3,3), padding='same',\n",
    "                activation='relu'))\n",
    "model.add(MaxPooling2D(pool_size=(2,2)))\n",
    "model.add(Dropout(0.4))\n",
    "\n",
    "model.add(Flatten())\n",
    "model.add(Dense(256, activation='relu'))\n",
    "model.add(Dropout(0.4))\n",
    "model.add(Dense(1, activation='softmax'))\n",
    "\n",
    "model.summary()\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "model.compile(optimizer='SGD', loss='binary_crossentropy',\n",
    "             metrics=['accuracy'])\n",
    "\n",
    "learning_rate_reduction = ReduceLROnPlateau(monitor='val_acc',\n",
    "                                           patience=3,\n",
    "                                           verbose=1,\n",
    "                                           factor=0.5,\n",
    "                                           min_lr=0.00001)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "epochs = 5\n",
    "batch_size = 64"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "#data augmentation\n",
    "datagen = ImageDataGenerator(\n",
    "            featurewise_center=True,\n",
    "            featurewise_std_normalization=True,\n",
    "            rotation_range=20,\n",
    "            width_shift_range=0.2,\n",
    "            height_shift_range=0.2)\n",
    "datagen.fit(X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP4AAAD8CAYAAABXXhlaAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvIxREBQAAIABJREFUeJztnXuQJfdV37+n+77vvJ87szOzsy9pJVmyhFdCjo1tjG2EDbZDHAoHiHCJMgmQmEAK21AkuJKq2KHAQAKmFNuFMC78JEhxjF2SLFkYhKzVw5a1K+1Tuzuz89h5z9yZ++xf/ti7O3PO+c1jV9Ld2fT5VKm0v57fr/vXfft3+57T53wPOedgGEa8CK72BAzDaDy28A0jhtjCN4wYYgvfMGKILXzDiCG28A0jhtjCN4wYYgvfMGLIy1r4RHQXEb1IRMeJ6COv1KQMw3h1oSuN3COiEMBRAG8HMALgSQDvd84dXm9MmM+7ZHvHpXZqLtL7LZVZ26VT/O+Vmmcyol3z9InEeSYSvB3InQCoVvW2tXNLJfVUxGGihPhu9RzGiS61tOjkGUPi0jlPn82Q+wCAQJ6yuD+ilD6QOvYVzCWo6G01/tGra5tY1veuvHZybnIfABCJW0HOn3y3nLz+oe4jkXPxnbOcn7w3fMdZO7/y4gyqK4VNPwF5ypfDHQCOO+dOAgARfQHAewCsu/CT7R0Y+Pf/4VJ7+IFlPaEXz7J29bpBvo/xOTXGJcTVmPH0KZZYO+jp4n/PpvWEJ6fFTsQiGOpTQ0j0KXdmWbuW0j+y5GKa28s/lkh/vyAhLl0to/vIRRyJy5Qs6FWQm+R3dFDjfRYH9C1TlccWp6gWFvTCyZ/Tc1kY5tdFnk/PU/whAQBz+/i3hbx2gR6CUgdvOzHfxJIek1jh7UqT7iORn1FuTJ9zoij2m+PtUrte0+nZ1f0c/conN58IXt5P/Z0A1q7Skfo2wzC2Oa+6c4+IPkhEh4joUK1QeLUPZxjGFng5P/VHAaz9HT5Q38Zwzt0L4F4AaE10u32feGH1b7v0T2W3wn/rRGn++9Ql9ZSpIn4DNuV1n5Zm1i7u5j/104eOqTEQNjy1trB2ODWvhlT72nk7x+dPVf3zbvomfk7SppR2HgDUhGXi+zntsyHXklrUc5E/7YPK5j4gaTIUu/jP0eSCHhOJ+Rc79E/Y3ATf79IQ/7u8toA+p7n9/O9hWR9H/vyvpuT5qCFIzW/sh0ku6jHVPN+v72f7sjAHQm6hIvD4G5Z2rf47Sum/+3g5T/wnAewnot1ElALwswAeeBn7MwyjQVzxE985VyWiXwPwTQAhgM86555/xWZmGMarxsv5qQ/n3NcBfP0VmothGA3CIvcMI4a8rCf+ZUMAwjXfNZ7gIervZe3pG7m3IzWo37d3fuM430dG93FL3Nuy0s0di8nrhOcIgBPBN+Hhl1QfhXDuhSX+wtr3Hl86guS7Z5/DRm7zBePI98bS2ecNIBF+0lLL5pEpjriTKrGVlzdivokVfS9UMzIYh/dZGNZzy4o4hGRBRsDoqax9Dw4AtSw/bm5cj5HOvHKz+LPnOK38NsWS5+V3dlI4NAfl37VDcK2j0Rds5MOe+IYRQ2zhG0YMsYVvGDGksTZ+IgH0rkZDUNFjZC5yA7EibKeUJ24arSJQuqgDsinFjeJyi4gDL4hICQC0zIOJXE8n7zCnI1Oowm3MxCI/x1qn9j/IIBkX8rl5k0TENl8AT0r4DmTQT1jWhqhMAlnu4c8Gny8hFEE/87t4u/mkfr7IwJRKk7ZdV3r4flpFjJWMsQcAF2wcPFTs1mOW+za+3lWebgFA5w1If0lFx5Ch2Cl8Fp7HrsyVyMzwMTJ/AeDJTFtN1rInvmHEEFv4hhFDbOEbRgxprI1fi0Dza4x0j9BFZV8/a8t33J1PzagxtMST08vX96s+iXluVKoEFdLGkctz484Fwt7taFNjZOJONMj9AlTTdnWU5MeW7+hT83pM+1GhL1DUjgAZh0A1EVOQ1R9/+swsazc9xtvo1OdcHuSxCz1PcN/Iwn7hqAGw1C+Slzx+jLSwb0v8MF57loTgSkkk/6Q8CUPShpe+kKTIkQeAlV5+LbOT/FqXW/VnFi7zudSadJ/FQX5dwpI4n049Jjux5j2+xwfjw574hhFDbOEbRgyxhW8YMcQWvmHEkAYn6RCwRkEn6mpVXYISd/hlp7m3IsroKQfCSRgu6QCeYEE4APP82KU+7YCqZoX6jzh08yElOITKEJdrqTbxjJvcsSk1JvcPXNSzctte1Uci1Xtruc0/SkfSAaXHZE5yL1tU4k5E5pytk5JOT6GI1PYd7ZBtE4lUUauOeDl/kH9GMshHOr4AHZgVCsdcYUB7vxJLfP4yGEc6+wAgNcfHyGCcSos+Tn6Ed6qUPME4UplIxIwFOs4MS0Orx2qEAo9hGNcotvANI4bYwjeMGNJwG3+tSi6VdQBPIAJtWoRYR3DynN5vt8jWCPT32amf40E9+RG+30RBJwzJYBuZPFO8focaI/eTnhSVL6Z1sQ+p3ltuFuq+HsESGfQjxTAAILHC7cxqdvPv+dk7uEBJ6/MiuuWMvv60wO1+1ySqQESeiklyQ1X36f3WGGtPvJXPzZfYI5OVVKGRSI8pt4voIdElM6kFP6o5cf8s8UGZcT1mpYe3Sx0ef0Nh44QhX3GVtf4GE+IwDGNdbOEbRgyxhW8YMcQWvmHEkMY69ySy9BWgHEHBPHeOeRV053kKX3lIZ5ANf2mStc++m3taCgNaZmVxJ3fQ9D/MA1HovA5MiQa4xAutiGAiWZ4bwMr1XFk4WeDXpZrxOJfSUhlHOwBLbSIASTitUovauZRa5N6hci8PrElXhIcKgJvmGXxUEo5SzznLzz7Kaa+Vy4pt4hRVSW8AySXeqbBTON2mtXMvM8Wv5cxt/Lr4HGa1AR4ZlDzC759y28ZZdIB2Fl/YKDI108KJuKzHrM04lCrJ62FPfMOIIbbwDSOG2MI3jBhyVW18l9P2OgmV3Vovt9df+oAOmqkOcXtr4IvaJnZZnr0gkzey53ViT8uzPNjGCf8CeYJmgkW+Yyqs8H206WSgzCiXhSnv4H1qnsCb9Bw35krt+qOMxGVQJZY9JmZR7Ecq8SY9n1m4JLYJpVuvVI4od+5CfY5VoRCUEEk56QXto5AKNhmREyWTXgCt+Nv/Ld6eOaDHpI9ym15WwFGBQwCWhvkHkD/tCfLZwc9JJgNVWrTvgNn9prJrGMZ62MI3jBhiC98wYkhjbfyoxhI6onyX6uKa+XvjxCgXqdj3P8/r/Yr3xK5FizrQDFe/TS5xkYfEnEdKVSLfu/reT8uEGuEHkNV5AGD5NTyBaKWL77fzcV2utdrDE3tKbXousmJryzEe7yD9EQAQtW6cYEOVLWSBCDVfqVYMAJHwFUhFYED7Laqiiu30LZ538uL2UAIZrdovEAr3zswN3PZeGdAvx6ks3rcn+YGqeU+5XDHdwpC+llmZ3CN2k5nS57xWeVeKxayHPfENI4bYwjeMGLLpwieizxLRJBH9YM22DiJ6kIiO1f/fvtE+DMPYXmzlif8XAO4S2z4C4GHn3H4AD9fbhmFcI2zqCnDOPUZEw2LzewC8pf7v+wA8CuDDmx4tCIDcqqMnnNH1jCr9XE0ncVQr1iikwkvFk8gjAmekGimd0oq50d5B3hYBJVT01E8Wc3EZcSCPSm32DHc8Lgxrp6ckXORRJ9Wspy6zTGpZ4mMqvdxBCAClDp4YExZFqSgxVwC6/JhsJ3SgSq2FX5dip5aHrWY2LivddkRPpbCTt5f7REDMvH7WTfww79Nykv89M63nv7SLt8OVzYOWqiKRylfuaqVPBPmc5ccudmmnYdPp1WMFOg7Ny5Xa+L3OuYu6SOMAejfqbBjG9uJlO/eccw7q2bIKEX2QiA4R0aFybXm9boZhNJArXfgTRNQHAPX/T67X0Tl3r3PuoHPuYCrMrdfNMIwGcqUBPA8AuBvAx+v/v/9KdhK164SVSiu39ZJZbkdXhQ8AAKI0P43kjP5lsTLI7dmW0zwoo7Yo6nEDKO3gX1SpeW5ARU2eJCOhzBtMcz+GkwksgAp4aTvGj1Pp08Iiy30iYchjL0qV3cWbeIZK00l9zis9KdGHi42Ud+q5pBf59XbN/LpFUlADQKmdb8uO6WCikbv5OaaErZ2Z1NdSimbI6jWldp9Ahnj+iS6Lw2qIUsOtipLXyQU9N1X2W9/KSpxDCmvkxvR+F3evfs6+qj8+tvI6768BPA7geiIaIaJ7cGHBv52IjgF4W71tGMY1wla8+u9f508/9grPxTCMBmGRe4YRQxqbpOMcF1n0VIjJjHC70y3wdtjapMaEBW4T+6rKhKWNk0so1O9q03/3JGvP/cLrWbvpnH5pGi5zo4yaxLv+UR27gFbu60iP83Oeul2rRzSN8mOHJc+78hS/DjVRfWf6tbpaccsZ/q7fpbnNH5R9ypN8W6WTxxT4RCWbvnOCtc/cc73qExaE8KqoLusT2yx1CLEOYVenFj3iKeJjLAk3Ri3rqfLDbw2V2OMT4shM87kt7vVUGBKVftYm4ABAuUefdLigP/vNsCe+YcQQW/iGEUNs4RtGDLGFbxgxpKHOPZcImWqurJIDQCV0uKpwZiS1I4NGJlg78CjjpIoiQUUExVTfeIsaEz72PdZu+9zj/Div0fKr8zfx/cpKzgmdC6RUeVyWR2Fkp7VDbaWHB8Akip5S2gnhDBPBRQmP6FAgnKBOXO/kqK4eJEkd46W0oy5P1rZw7HZ9X5cpn98tEoaEym65TTvqep4SlXT6eJ+y9mcCwhEng4Ck0i0AFHpFyfRuPqj5lL5P524QZdeb9OeamOTnXO7gfbJndDDUSt+aNRKsGz3PsCe+YcQQW/iGEUNs4RtGDGmojU+RQ1BYtbVdXkc5BFNc6IG6efCKk9VnAVBaZCZ4AniiNh5UsjTAxzSf5hVvACDcN7zhfs++U2dZDN7PExVr7SIjMeuJ7BABMPI4iWVtC5ba+Hd2UNa2Xf4s96EUu/ixZTVdAEjM8etAUjjEc21VIJYIhqIJrpQMAJThc8kd133CEvcNTN+UFn/XU5FBS4HoU27RQTPpWX4taxl+Pj7FXBlokz/Dz7kw4MmaEmPCaW2vVzpFANgy32/rCb3faI1Pi6pbK6VjT3zDiCG28A0jhtjCN4wY0tj3+GGA6hqbN1zQRpqsglMcEBVjWrVd2npYbKtqm7jcwW3K8bdxW6rlf2k7TopRpkb4O+yBb86qMdHJM6xNTlSi6e1RYyASYajC55ae0AKdUYrPLVHQyRu1HP94M+eFHyPQYqE1IS6SmOVJRVGv9mvIpKigxN/J06KO14hmxLWb0yKebjc/VttJvt+xO/XtW+gXG8R77fyoftZVhE5pJEzvyJOkgwUh8CGSg5pP6eMs7OH78e03Nc7PqWlE7GNY2/DlztX73SrpGIaxLrbwDSOG2MI3jBhiC98wYkjDA3jWOvSiJl095djPc4fT4De506T1sKeyzsQUP05ey3hnvsedbn0de1i7ltGOFqk2Ux7izqbkmEdNJyPUb/tFrRFPmezCDd2snZ4WVXLyOtBDOvOSU9oB6EQgjbzeUVI7imRSjgw4WtynlZEz09zpljrMa1W7Fq2apIKWPApI2e+f5UMGuWM00DlSKI3zz94l+f2TmtW3fKmHzyUzJtR8p/WYSGwq7uSfx0JOn48srhO26mC02rK43rtkNSHthObbLEnHMIx1sIVvGDHEFr5hxJCGq+zSGtsuMaYDYA58bExtY7vY2a037uDbXEJ/n63s3MHaQVWKOuhL4Yjb1s3PiUphiwU1hgb4ccq9QkH3tBacSC5w+zASdvZyj7bxZeWcliVPmVRxHajCbdnkgkcxVyTcRM3c5+Kr2JP6AbfFKcft1OgcF0rxEujPLBJBPaEIdOr9LL/WAHD6p/kEpQLtSp/nBJr49U8U+ZiKdmugmuPXKTPCP6NKs0dBVxj50ZQueyNN+GqrCAAreZJwruDxbU98w4ghtvANI4bYwjeMGNJQGz9Khij1rSaXuIRWPkzN8Pfc4Tj3A1Tz+t0/Im4YVVq0TTxzgG/LnxO2oEesMlHktp8UDgkqnlIuSzwRJmwWwhtFnZiUOsPjEOYPykwTjRScKO7QsQupOW73SxszLOr5u3DjZ0FuxCOQKgRR3RL3fVBKf2bRMt8PVbTvQ77bj6Z4klTuhN5vOMff9SeWhdhmu/Zr0By/N2pit8VB7T+Rx5ZVeFOzHj/TIL9OmXN6+RX7RZ9Ofj8V5z1+gaU1+5HBAutgT3zDiCG28A0jhtjCN4wYYgvfMGJIY5N0ag7JxVVHSTDnCYBZEs4jURVHBqEAOhAoUdbOmJbm3awtq8xQzZP8IMo7R2k+F9cj6ikDCBaESm1FqK706KoywQI/56QoD73S6UkSEc695LLvO1yUuBZzQeBJ0hGbXIIH8KSePaXHrAhlH5lw4wnOUWXJPUk6aoxQ+iGP0tK+L/IS4yfex6NvqF3fG4kz3AFbaeX3glS6BYBiD7+W+TP8HJM/yh22AOCe7OL7GNZzCZL8nKKjPMEp4wvguXn1nCnpCVDyYE98w4ghtvANI4ZsuvCJaJCIHiGiw0T0PBF9qL69g4geJKJj9f97KiMahrEd2YqNXwXwm865p4moGcBTRPQggF8E8LBz7uNE9BEAHwHw4Q33FABRatVeKl7XqbpkxrlNGYxwUQeKtC1e3ckFMmpZfVpNp7hQxdgbefBQat6TsCJsyuUBHiQTVPRcmqa5jRkIl4XL6OAil+TzzT3PE5WWuwf1GFE5p/lFrVJLEbf3qm382iZP82sLAK5NZKSc44lJ0ZLHLyODfoRN7zxBS5Tc/NZzwldDLVxZWF43AKxSEwDs/0u+jxd/SasE13ZzH0U0LYJkWnRwUUcnv5+mWvl167xf39vFO8R+POY4jQt/wyA/n3RWz6VyZlUm2JW39iN+017OuTHn3NP1fy8COAJgJ4D3ALiv3u0+AO/d0hENw7jqXJaNT0TDAG4D8ASAXufcxUfTOIDedYYZhrHN2PLCJ6ImAF8F8OvOOSY255xzWEfsi4g+SESHiOhQpaJ/JhqG0Xi2tPCJKIkLi/7zzrm/qW+eIKK++t/7AEz6xjrn7nXOHXTOHUwm874uhmE0mE09LHQhauIzAI445/5wzZ8eAHA3gI/X/3//pkeLuHJtakb/SKgJJdhQBPAEK9q5IZVhA48aTbVTBKIs8GPXsvo7cPY6vl+putJ2VM+/9M+4KkzrUf4rJzHFnX8A4DKejMM1tB/2qPlKx9DpUb1foWQbCmdl5AmswQKfnwqa8WXvyf2I4wae0uBybkp1F0CQ487U2m5+bb3XMiFKdAtV4+ZTPIgGAJZL/N5IiFus1KSDZqZneGBN+iz/DKd/yOMsllQ9gU2y4vg4dzSWcto5zFbx1pLztuTVfwOAXwDwHBE9W9/227iw4L9ERPcAOA3gZ7Z2SMMwrjabLnzn3Hew/vfIj72y0zEMoxFY5J5hxJAGJ+lECOZXI1oC0j8kFm7mgQ+Ttw3zDp53B8t9fGNuXO83WeB9Eiu8vbRDJ2KkZ3mfcEXYyJ68klIH79P2vDDGPbasQgSmBB6lHAl16IQhJ6v2ROLYXTqYhQo8mMWVhL/EkwCldyKuU8kTwCOScijhuRXltZJJRXJuABCI4Bsxl51fPqGGvPSBvay90seP67lNkcmJY1e4HyO/Q7/Bqj7Hg8Z8SrzVLu5gyJzmvoP0jL7pCvvXzMVTaceHPfENI4bYwjeMGGIL3zBiSENt/GpTAtN3rqqgdjytK+kkhAhFdkqopDZrg0va9OQxo2tp3kfa+JHn9WiUlqoUvCnfuQJavbfcLRJjkvq7NnGOq8e6Vv6O2HmMzEo3f8edyOiPMhznF8KVhE3p9AlEIkmHhB1NHVoZGUL9VibyKNENAJTmtrhMyAEA6uOKueEsz3hyOR0fQGXxEl4oACPHPw8fTUM8bmLxnC6lE7zAE4YK+/j82xPaL1MVN0zU7lEWLonYkev5Obe169iFwujlJ8baE98wYogtfMOIIbbwDSOG2MI3jBjSUOdeolBF51OrjiBfOetahm9b6RZqLlv5qvI43eS4SCjoOs+VkCWhE8t8xwu7tdNt8CEeAFPoF0kWrbrUVesMV3OBcFAFKzoAprxPqNF4FHODBe78qg5zZ1mpQycHpRb4sZPn+UUoDvDjAkBWzFc62GSiDwCdyNOmnYaVXr4tWBbXZdFTzksmDKU8XlvBwMPcYXaylQdDhQl9QxWGufOOhFNudlxfJxoSjtJ5PbdELz+nyhL/jMZXPCrNa0toRVZCyzCMdbCFbxgxxBa+YcSQhtr4Lhmi3L2qwpOc0jZadoInlsztFao9HhNG2uJee32T3BjnSbiRY6p5fvCUFrbFcp+06fl3a2rRI94xxJNlMie4mJErimQbAOlZYYtPLqk+ytYWVYh8KsGJOX6sWpMuyyxxczzgJchzP0a0oINOgm6ejFXp1wlDa0VbACASQUqBp8KNDHaiovwQdaBQOMsDjqqtPMjntpt09aAjE1wUZHfXNGsff3yXGlPt58d2aZ2kUylwm56K/Bydx9+QKKyes1wL62FPfMOIIbbwDSOG2MI3jBjSUBu/liLM7161Gd1ebT8mRfKMFMOQQhcXdiSO4zFLE1KTQrxCDTz5EvLdfyhep/ts5LC4cfJPsV3PP7XAD1QRlYESR7T9njw8wtqU1u/kpYhnTQg1Zk/y5BpAC3EE4j14cESPQYbHC7hlUTF4oE8NqXTxRCSf2kW5nX+Q4YpIOsrqc6Zl/iEpIVNPYpJk/+f4Pl66ZZ/qU30TP8cXR3hZCZ8mKc3xa9l3QAtTTzzH93Pg9pdY+8SjvOozAJSGV+frkibEYRjGOtjCN4wYYgvfMGKILXzDiCGNDeAJgNoaP1B6XjsiZABCaon3KbdpJ5AMvgl94qvCeVcRwTiR70qI6SWXNlftSQgHVC3Fd+wNFBLHqeb4GNo/oMaEp8b5fj1BPhCOrdToHP+7x9HlstyhRkWRWOJJeqEMH1O5bidrByWtRhOKhJtyh1bGceKjjoR6kaygBOjS4JDqP55zpir/zBLz/FrueEhXMqJ38evQnedBQC8W+DUAgKRQyB070a36uDyfy/HHhlm7tFN7obdYPIdhT3zDiCG28A0jhtjCN4wY0nAbv7Km8qhPPbblLLcHp27mU1zZobNtup7i31+5SW1TTt3CbTIVGJTxZf/wpgweSnuq/c7tE0EzwnQlT1GcmRv43DJTfL9Bt7arO85K+9bjPBD2ufQDSKVbAICwkZ2wf6lFK85GeX6SYUE4WTx2dUmoBPtEWSQukD4WPSaQQhyiei4qng9AKAlXhQBIYlqPWfkur7r74i4uvBEu6blV2vm1zPfqaju5NJ/LdAsPdEp4KuXs3XH+0r9nUp5INA/2xDeMGGIL3zBiiC18w4ghDbXxgwqQH1u1IQv9+ntn+kYxJWHSZCa1LbvSI9vaJs6PieSfBWG7el6wy6o9KmnHk6STn+D7newQ9vusHhOJPJL0orAFT2khC8jqssnNP0ryZY5IakLIQtrMHhu51rJxhdpKk54b1YQfo+ypHJvnn0koD+3LR9ksCccn/CliExLTIilKxgYA2P2XZ1n78Ed5IpLb4akQLIQwr+vSSTrP/SNPCIpa+b0w+IDn/KbWJHWNbG1J2xPfMGKILXzDiCGbLnwiyhDRd4noe0T0PBF9rL59NxE9QUTHieiLRKSTow3D2JZs5YlfAvBW59xrAdwK4C4iuhPAJwB80jm3D8AsgHtevWkahvFKsqknwDnnAFz0diTr/zkAbwXwr+rb7wPwewA+tdG+ohSwOLj6XbOVqjgygSWtK2sr51jgSdKRjrpaijuOSh5lHKmyK5NGSi16TKGPO4rk/AOP2m92gjuPcqNCBWdKy/nKZBqVjAKA5rlT0ImS0ZT1lIyWDkBZdtqTpJN4kTu60C4q4BR19aDFPTwwJbmsHWhScaeW5eeYmtMOtE2dnDOeayna1CRUglv0/GmaJ+7s+zy/TkO/f0yNeeTF61j72VODqs/wt/h+Uuf5veB7VK91lNIWFIbW2Y1nx0QhET0LYBLAgwBOAJhzzl28k0YA6HQkwzC2JVta+M65mnPuVgADAO4AcGCrByCiDxLRISI6VF3WIYqGYTSey/LqO+fmADwC4PUA2ojo4u+qAQCj64y51zl30Dl3MJHL+7oYhtFgNrXxiagbQMU5N0dEWQBvxwXH3iMA3gfgCwDuBnD/5R488ORLSCVbZb97gmaaR7ktOLtPn1Z6TiR4iC4JLWSLkihMKudb8yT2SD9Aam5zmyu5IHYsAkaiLl1JNpgTE17SVYlcq0ioERVq3aL+BSZFNRQln8oJt71JVPcNa9p+bxIiGuU2fdzUNE8qqrYKkZCq3q8TwidUFn4NjxpxNL+wYZ9Sr35gZYWNnxrh6sPn/s2QGnNgiTuopMgJAPUZSd+Na/E8PNfa9Vu08bcS5tMH4D4iCnHhF8KXnHNfI6LDAL5ARP8VwDMAPrOlIxqGcdXZilf/+wBu82w/iQv2vmEY1xgWuWcYMcQWvmHEkIZm51EVyEyvOh985bBkUI8vGEey2M9Po5bRfUpCnVc6ESuiohOg1XqX9vPgitwpHcwi99s8wp01oScLzYV8bkvDnskIgip33OVGtXMvWBHBNyI4x+vIi2Q6pOhT1RFIFArVHlmeWzqsAIRTfC5BTl/LWpNwxInbhUoetRkxf1lmy5ehSCLTMZrlasSJRa62A+jSXHIudJarIPvmhpwngEpkQ1YG+bET8yKgBwAtrHHSehypPuyJbxgxxBa+YcQQW/iGEUMaauMDvFKOrzR1VeZDCLNIqu0AQLWZ25AtR3XCiqx64+SZe4RZ5JjUeREc4jGnZInuoLq50kxqhAd25ESpapff3Basdmv12+UBfjGVko8n4caJ4JVgmQfRqLLTAEj0kdV2ZNlsAMoWzRzTN4Nr4ue9squNtaOcJxPcp7CzhvKwttcTT29cYSg5Mq3nJpOkRAJUtKSDowKpauzxfURdPGosEAFIvvOdmQfPAAATH0lEQVSb+Ildl/5d+d9by463J75hxBBb+IYRQ2zhG0YMaWwlnQRQ7Fy1UaTQBaDtZp2ko20cqbwrxwBAVbzbV8fxVL4NRQHa1LwQ8/CMkUIbMwd4p/6HtGIuCeVaJ1Vdp4UNCi2qkZjVFV2b5jpYe3mYV3vJvaT9DWWRkJKc49e22qzf/Uvxh7DA7fXwvEf8QghmuHmdJSWTk8ISv7iVVh2wkR7j1yESPolwyRMYUhHv4FN8TNTJrxsAUEV80CIWQO4D0J+ZPC4ABOJzPPrvdrG2S+gknFrT6lxq33wFhTgMw/j/C1v4hhFDbOEbRgyxhW8YMaSxATyOB+34HGoJEetREV9NKY9sX1n4Xla6fWWq+LZEgTvqyr1aDijI8m2trSIR5rFONaawmzt9kjMiMcajYCMdXZDBHz36ODTFFV/coqfMVoHvJ5Pdw9pVj3MsOc+zjGSij++GUUo4soTWTu5kBKACs4J2nZhUzfMbJCjyaxulPYFaMqhHzCUcOQ+Jk2rD3Xy+taxHWVg49+bu5Fqz5+7aocZQUajpZLR3OzHNj1UTpbjSWe0QrC2scbhuHL90CXviG0YMsYVvGDHEFr5hxJDGB/B0rRp3tZzHFs9wezHRwm3i2gmdsFLu5LZSuksnhVRn+LimF7i9teipwFIRpuncAk8SCXs8wRJ57heoNPM+5FPDbeH2bTTMSy4HiyKSCFB2P814Pkoh/BCMTvF2i8eu7uABPCTUcL3KtiJhSAa3JE6M6bkJsQ63s1d1SYoElZqw32tp/dwKSiIwKCEEPzxBM64k/BpCbCQ54QlAEoIeYz/J95tMa59RdZnfg8m89vc0dXNfzewID2KKPPvNtKzOn0IL4DEMYx1s4RtGDLGFbxgxpLE2fgBUm9bYiK2eJIVpYccJZYtot7Z3aZr3KU1rP0DuLD/VuTcJsYuq/g4Mz/O5tN/ABRmmJnXyhhyjxDp8lU6kIIOwmV1OJ8YEopJLbVArlIQznvJAa/FUjg3PTbA2DfXztieZRgqFqKQjmZzi2ear8holxHv6gL+krub0e/ywxC94+hgXvXQVPRcptunSouLxkvYZzb+Ov7dvbePXcrmok3SCovA3HNVVceb38HPq38P9Msslvd9SZc38ZXnmdbAnvmHEEFv4hhFDbOEbRgyxhW8YMeQqJOmsUeAZ006rai8PaghmhWKrzwlXFMo4HiXS9jdxJ8/EszxgpNrmkQMKuKOkKc0DPaY8jpTkMHd+ZR7h6rfO59yTCjsiYeXoL2on4nW/c5q1a3u6VZ9yGw/yyZ4UarEicAUASF67Ka4AvHjnsBqTO8OTgYJFcS136LmR2G+pVwcThUVPHfU1pOc8CSsiqMcVhUpwWQfNUCu/vrLUNjzKwufeIby2M9xRF6b1/VRr4+eTLHgqMSX4uLFJHjSWzevP7EqwJ75hxBBb+IYRQ2zhG0YMaWy13FSEYOdqkkp4xJMkIpQ3onZuxwUJT/kaYTan5vT32dgLPMAltZfb4tVZLUpRy/Fjjf0jD9pI36DFL4rz3G8x9LQIeGnWQRskkmkKg7wCDnmUhY/93i2sff0fnFR9ol6eZbR0E78G2XFdfSc4wwN4ZKJP/qHn1RjsG2JNJyrsOo+QRfGHuHpsLaM/s6yw8YMSb1fz+jOTATwQATtBO7eZLwwSgUDj3BdSumlQj6nyz6SrTyTXLMiSUEBbD++zIAVAAKTF/U3E59/fptWUT02s+nJctDUlDnviG0YMsYVvGDFkywufiEIieoaIvlZv7yaiJ4joOBF9kYi2Vq3PMIyrzuXY+B8CcATAxZeenwDwSefcF4jozwHcA+BTG+3ARYTqGqGEqEm/05YVafvvOLfpxE7PcmHDKKPtHCc2VU4LWzuvfQfZbi6aUZ3nNvFgh05yOf/EAGuH00IUM6NjFyIhdpGe4X6N1Ly2ZYv7+Pvc07+4V/XZdR+3+9NN/Njldj2XJPH4hnBWvKMP9bPCnRlX29Yiq+cCQCbi7/aXhrW/RwplSlGNydv0sya6ndvAKzPX8w6Bvudu/Bifv6yCE67oeIL2ndx3s1IWIpmed/S1HP/M8k064WxxjN9jYSuPO5hZ1n6Bf3HDs5f+/bmspzKxhy098YloAMC7AHy63iYAbwXwlXqX+wC8d0tHNAzjqrPVn/p/BOC3AFx8LHYCmHPOXfwqHAGw0zeQiD5IRIeI6FBt0aONbRhGw9l04RPRTwKYdM49dSUHcM7d65w76Jw7GHpeZRmG0Xi2YuO/AcC7ieidADK4YOP/MYA2IkrUn/oDAEZfvWkahvFKsunCd859FMBHAYCI3gLgPzrnfo6IvgzgfQC+AOBuAPdvtq90qoK9A6uVTM5mdTBFSTiyTh/ljrvEgv6RktgtKsZkdPJGWiQ/zP+AJ7B0DPGkEQA4f7adtftv5cEtJ89o1ZvB49wRFLWKQA5PAhFEWeyT/1IowLRoh40rcYdgYVg7oE78292snZ3kx+444knSqfG5VHZwlVf0izaApChNrc5Rlv0GEMxzx2ktrYOJXILvJxDBOUrdCEBTVjjQ+rlzbPqErurjWjb+JZpY0E649hz/TE6e5E7RfJdWUw5FUteuNn3PDQ0cZ+2/H+VO25nzOmHrqy+8/tK/Zxf+Sf3dx8t5j/9hAL9BRMdxweb/zMvYl2EYDeSyQnadc48CeLT+75MA7njlp2QYxquNRe4ZRgxpaJJOuZrA6clVG+vGfh34cTrN7eqVErd3y806MMLNcb9ANatPa1GW3W3lNv95j2Lu8F5u0790nNtxQZP2JWQm+bbwPA/ykYq0PsJ2bqdSoI3Zni5uV4+Ptas+JaF/UR7i57zcpwNgyp3cd5CY421ZdRgAgjL3deTGRCVi7RbAzr/ndnNy2WOwi01OqOzuuOusGjK1xO31uVH+uV5/sx4TpbivqdbEr0tQ0aIak98UFYz3cB9Lc1b7BRaW+X16bknfc8enu1g7neT30/tv+64a83dnbrz079BTgdeHPfENI4bYwjeMGGIL3zBiSENt/CBwyK9JVHj+8T2qz57buQ02tIO/6yxUtV36xBM8EeOn73hG9Xl+nlegHVvg9lU+rUUYz0zwd74/cusLrP30375GjYnSwrYTVWHhSdIpDnIjOHmY96nepEOdWfUUAEHSY9uJbdcPcJ/F0WkuoAEATSf5fmt3cF+Ck9lOAMqj3K5efJ3wUXhESVNf5e+5XeBJ0hGPJZmks1zR/p75WT6Xt72OC4eUIl19Z6rMj+0qvE+5TX9m7Ue5TV8Y5tfF53ORPosdw1rIRVbKWSxwv8AT08NqzJvXvPsfT3oqK3uwJ75hxBBb+IYRQ2zhG0YMsYVvGDGkoc69KCIUllcdJeGeTco4A6g6/t30+AtaaYbS3Hn0le/ervrsHOblhtfOA/CXNXYzfNvRDh6oUr5VO92iZ7jDqXgDlylILujEmJkbRJnv63kCyE39QvkWwPHzPNDD1fR3eFMbd6AdOcVLXuem9RhRlRwrIjiqvUc7pDK7uQNzfp4nJu3/Yx3opBJ3PAWGnEj2CYXK7o/uOKbG/N/STaz90DO87TvOgSS/D6nGOzntD0RY4n3+7MfvY+1vzN+sxjzw3GtZ+5eHvq36/NXY61l7TxO/b+9/6jY1pja8+jkue5zfPuyJbxgxxBa+YcQQW/iGEUMaauP3ZhfwG6996FL7c2d+WPU5Pc0DH45/j6vWJvu0KEWmk9va/S262siJcZ6x0tPB+2QSWshiXIh3DLXwYKJapL83R+8R9qGobPK+A99XY2RQRpvY7w0tOpnphXM8YciV9VxyKW5bS+tcVhAGgPHn+H4TM9xnUe3Sx1la4IlH6ZPcL0BFLUpRbeN+gGpOG9LJAv9MZPLM55/W98+eoUnWLrTxMZHnMxt5G0/S6XyeXzfyxEaN/zC/Lo8X9rF2U6h9OV3d/J778KM/o/pct3eMtb/2Le6vys7pAKozldXPrFzSQU0+7IlvGDHEFr5hxBBb+IYRQ2zhG0YMaahzb3y5Bb//1DtWD57SXpPXDfDsvIkOrr5aKOsAhfMzvM/Jw1q99+ffxYMlHjvPnTH5pM7Oa8tzR+KTP+DBQ1TzlOoSCig7d/ISWt8au06NWSryqJnCxOb1B2pV4QDcr9XNXxzljrqb9vA+gSdrbrSJzz/VwbO9/uTmL6oxT65wNd8/D97E59qqS4AtDfBttbS+ljKQJrkkSl57MhLv2sGz8b6d4Nf7zJy+NwpiLuVWviy6btcBVP3CGfy5Z7mj0RdQ1d7F3av9Q9OqT2ua33N/9s8/zdq//Pf/Wo1J5VadkeQrI+/BnviGEUNs4RtGDLGFbxgxpKE2fjJRw47uVdXZXFInbxRrfEq/u/trrP3fz9ylxvzKwUdZ+7M9b1R9/m70RtaeOsKTXLpvPA+JtIF37OI2WXtGBxN1pHmwykKF24+TBa0009XEA5DaRJWWhaJWgPmpG5/j+y3qSjSDPTzg6Kywb30BSDuG+TlOP8sTk34t+X415mOv+T+s/brhM6x97nc91Xf+lPtqlnbqWzFZ4PNLijigFk+FoU8degtr9wsFp7eISjUA8FyeJy9d18qDgLKhvk/vf/KHWPv1N/OEoRemdZWle/b9I2v/wcPvVH0e++mvsPYvnX0zawdJbcN3tKzeP+dCs/ENw1gHW/iGEUNs4RtGDGmojV+pJjA2uWpn9vfMqT7pkL8ffWTpBtYendf24r0l/t74ls5zqs/j54b5cYTCqc9eP3KMi2iEizyRZHZQC3GstPMkifOL3Kb/ieHDasx3JrjacEuKJ3iMTelz/nqBC0z8yi1a1OHpgKvoDjTx7/n/Mfh1NeZPZg6ydsfQIdb+hzkthPLjOW4Tv3nXA6z93yZ/RI15aB+vglz1hC6UxXxz49znkvMoI6d7+P0zs8STgb4xx+8nAMhn+X4+sPc7rP3xs9oWf8MtR1lbCsbMzeoTKkWiCnKbnv+Bv/pV3meIOzak/wQA+rOrfrOXEjo5yIc98Q0jhtjCN4wYYgvfMGKILXzDiCENde7lUiUc3HP6UtuXJFKucQfakUXuBPKIpOLnh55g7U8dfZPqkwhFiejzvITWQl4794Z3c6fV2Cwf8xs3P6zG/MH338bau7t5QIx05AFAqyi7JYOA/uiNX1JjjlV4ANJvPqnVXH7/dh4M8lM5rgAzHemr+Z+7ufOx5nhAyNvyR9SYitjNd4o8OejJ/8IdhgDQM8uvdxTqZ1ChjzvDlnfoQCbJuwd4YNODEwdY+6URft0AoK+XK9n+wj/dw9ovvPmzaszfFngwVCESpbU9d+rxFR7U8+GD31R9Rstcgeq3u55l7Q+89A5IPtz96KV/fzuhVZB92BPfMGKILXzDiCG28A0jhpBzPqv5VToY0XkApwF0AZjapPt24VqaK3Btzfdamitwbcx3l3Oue7NODV34lw5KdMg5pz0+25Braa7AtTXfa2muwLU3342wn/qGEUNs4RtGDLlaC//eq3TcK+Famitwbc33WporcO3Nd12uio1vGMbVxX7qG0YMaejCJ6K7iOhFIjpORB9p5LG3AhF9logmiegHa7Z1ENGDRHSs/v/2jfbRKIhokIgeIaLDRPQ8EX2ovn27zjdDRN8lou/V5/ux+vbdRPRE/Z74IhHpwglXCSIKiegZIvpavb1t53q5NGzhE1EI4E8B/ASAGwG8n4hu3HhUw/kLAFLN8yMAHnbO7QfwcL29HagC+E3n3I0A7gTwq/XruV3nWwLwVufcawHcCuAuIroTwCcAfNI5tw/ALIB7NthHo/kQgLXJCdt5rpdFI5/4dwA47pw76ZwrA/gCgPc08Pib4px7DMCM2PweAPfV/30fgPc2dFLr4Jwbc849Xf/3Ii7coDuxfefrnHNL9Way/p8D8FYAF7OJts18iWgAwLsAfLreJmzTuV4JjVz4OwGsrY81Ut+23el1zl0sWj4OoHejzlcDIhoGcBuAJ7CN51v/6fwsgEkADwI4AWDOOXdRL2s73RN/BOC3AFxMT+zE9p3rZWPOvcvAXXgFsq1egxBRE4CvAvh15xzLu91u83XO1ZxztwIYwIVfgAc2GXJVIKKfBDDpnHvqas/l1aKR+fijAAbXtAfq27Y7E0TU55wbI6I+XHhabQuIKIkLi/7zzrm/qW/etvO9iHNujogeAfB6AG1ElKg/SbfLPfEGAO8moncCyABoAfDH2J5zvSIa+cR/EsD+umc0BeBnATywyZjtwAMA7q7/+24A91/FuVyibnN+BsAR59wfrvnTdp1vNxG11f+dBfB2XPBLPALgffVu22K+zrmPOucGnHPDuHCffss593PYhnO9YpxzDfsPwDsBHMUF2+53GnnsLc7vrwGMAajggg13Dy7Ydg8DOAbgIQAdV3ue9bm+ERd+xn8fwLP1/965jed7C4Bn6vP9AYD/VN++B8B3ARwH8GUA6as9VzHvtwD42rUw18v5zyL3DCOGmHPPMGKILXzDiCG28A0jhtjCN4wYYgvfMGKILXzDiCG28A0jhtjCN4wY8v8ArXqDfnor9jwAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#split training and validation set for the fitting\n",
    "random_seed = 2\n",
    "X_train, X_val, y_train, y_val = train_test_split(X,\n",
    "                                                 y,\n",
    "                                                 test_size=0.1,\n",
    "                                                 random_state=random_seed)\n",
    "g = plt.imshow(X_train[1100][:,:,0])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/5\n",
      " - 200s - loss: 7.9686 - acc: 0.5002\n",
      "Epoch 2/5\n",
      " - 201s - loss: 7.9637 - acc: 0.5005\n",
      "Epoch 3/5\n",
      " - 197s - loss: 7.9773 - acc: 0.4996\n",
      "Epoch 4/5\n",
      " - 196s - loss: 7.9725 - acc: 0.4999\n",
      "Epoch 5/5\n",
      " - 191s - loss: 7.9596 - acc: 0.5007\n"
     ]
    }
   ],
   "source": [
    "'''history = model.fit(X, y, batch_size=batch_size,\n",
    "                      epochs=epochs, \n",
    "                      validation_data = (X_val, y_val))'''\n",
    "\n",
    "\n",
    "history = model.fit_generator(datagen.flow(X, y, batch_size = batch_size),\n",
    "                   epochs = epochs, \n",
    "                   verbose = 2,\n",
    "                   validation_data = (X_val, y_val),\n",
    "                   steps_per_epoch = X.shape[0] // batch_size)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}