fashion_mnist_3layercnn.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n",
      "/opt/miniconda3/envs/trial_env/lib/python3.7/site-packages/tensorflow/python/framework/dtypes.py:526: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint8 = np.dtype([(\"qint8\", np.int8, 1)])\n",
      "/opt/miniconda3/envs/trial_env/lib/python3.7/site-packages/tensorflow/python/framework/dtypes.py:527: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint8 = np.dtype([(\"quint8\", np.uint8, 1)])\n",
      "/opt/miniconda3/envs/trial_env/lib/python3.7/site-packages/tensorflow/python/framework/dtypes.py:528: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint16 = np.dtype([(\"qint16\", np.int16, 1)])\n",
      "/opt/miniconda3/envs/trial_env/lib/python3.7/site-packages/tensorflow/python/framework/dtypes.py:529: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_quint16 = np.dtype([(\"quint16\", np.uint16, 1)])\n",
      "/opt/miniconda3/envs/trial_env/lib/python3.7/site-packages/tensorflow/python/framework/dtypes.py:530: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  _np_qint32 = np.dtype([(\"qint32\", np.int32, 1)])\n",
      "/opt/miniconda3/envs/trial_env/lib/python3.7/site-packages/tensorflow/python/framework/dtypes.py:535: FutureWarning: Passing (type, 1) or '1type' as a synonym of type is deprecated; in a future version of numpy, it will be understood as (type, (1,)) / '(1,)type'.\n",
      "  np_resource = np.dtype([(\"resource\", np.ubyte, 1)])\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "from keras.utils import to_categorical\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "# Load training and test data into dataframes\n",
    "data_train = pd.read_csv('fashion_data/fashion-mnist_train.csv')\n",
    "data_test = pd.read_csv('fashion_data/fashion-mnist_test.csv')\n",
    "\n",
    "# X forms the training images, and y forms the training labels\n",
    "X = np.array(data_train.iloc[:, 1:])\n",
    "y = to_categorical(np.array(data_train.iloc[:, 0]))\n",
    "\n",
    "# Here I split original training data to sub-training (80%) and validation data (20%)\n",
    "X_train, X_val, y_train, y_val = train_test_split(X, y, test_size=0.2, random_state=13)\n",
    "\n",
    "# X_test forms the test images, and y_test forms the test labels\n",
    "X_test = np.array(data_test.iloc[:, 1:])\n",
    "y_test = to_categorical(np.array(data_test.iloc[:, 0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "img_rows, img_cols = 28, 28\n",
    "input_shape = (img_rows, img_cols, 1)\n",
    "\n",
    "# Prepare the training images\n",
    "X_train = X_train.reshape(X_train.shape[0], img_rows, img_cols, 1)\n",
    "X_train = X_train.astype('float32')\n",
    "X_train /= 255\n",
    "\n",
    "# Prepare the test images\n",
    "X_test = X_test.reshape(X_test.shape[0], img_rows, img_cols, 1)\n",
    "X_test = X_test.astype('float32')\n",
    "X_test /= 255\n",
    "\n",
    "# Prepare the validation images\n",
    "X_val = X_val.reshape(X_val.shape[0], img_rows, img_cols, 1)\n",
    "X_val = X_val.astype('float32')\n",
    "X_val /= 255"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From /opt/miniconda3/envs/trial_env/lib/python3.7/site-packages/tensorflow/python/ops/resource_variable_ops.py:435: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Colocations handled automatically by placer.\n"
     ]
    }
   ],
   "source": [
    "import keras\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Dense, Dropout, Flatten\n",
    "from keras.layers import Conv2D, MaxPooling2D\n",
    "\n",
    "cnn3 = Sequential()\n",
    "cnn3.add(Conv2D(32, kernel_size=(3, 3), activation='relu', input_shape=input_shape))\n",
    "cnn3.add(MaxPooling2D((2, 2)))\n",
    "cnn3.add(Dropout(0.25))\n",
    "\n",
    "cnn3.add(Conv2D(64, kernel_size=(3, 3), activation='relu'))\n",
    "cnn3.add(MaxPooling2D(pool_size=(2, 2)))\n",
    "cnn3.add(Dropout(0.25))\n",
    "\n",
    "cnn3.add(Conv2D(128, kernel_size=(3, 3), activation='relu'))\n",
    "cnn3.add(Dropout(0.4))\n",
    "\n",
    "cnn3.add(Flatten())\n",
    "\n",
    "cnn3.add(Dense(128, activation='relu'))\n",
    "cnn3.add(Dropout(0.3))\n",
    "cnn3.add(Dense(10, activation='softmax'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "cnn3.compile(loss=keras.losses.categorical_crossentropy,\n",
    "              optimizer=keras.optimizers.Adam(),\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "conv2d_1 (Conv2D)            (None, 26, 26, 32)        320       \n",
      "_________________________________________________________________\n",
      "max_pooling2d_1 (MaxPooling2 (None, 13, 13, 32)        0         \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 13, 13, 32)        0         \n",
      "_________________________________________________________________\n",
      "conv2d_2 (Conv2D)            (None, 11, 11, 64)        18496     \n",
      "_________________________________________________________________\n",
      "max_pooling2d_2 (MaxPooling2 (None, 5, 5, 64)          0         \n",
      "_________________________________________________________________\n",
      "dropout_2 (Dropout)          (None, 5, 5, 64)          0         \n",
      "_________________________________________________________________\n",
      "conv2d_3 (Conv2D)            (None, 3, 3, 128)         73856     \n",
      "_________________________________________________________________\n",
      "dropout_3 (Dropout)          (None, 3, 3, 128)         0         \n",
      "_________________________________________________________________\n",
      "flatten_1 (Flatten)          (None, 1152)              0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 128)               147584    \n",
      "_________________________________________________________________\n",
      "dropout_4 (Dropout)          (None, 128)               0         \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 10)                1290      \n",
      "=================================================================\n",
      "Total params: 241,546\n",
      "Trainable params: 241,546\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "cnn3.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From /opt/miniconda3/envs/trial_env/lib/python3.7/site-packages/tensorflow/python/ops/math_ops.py:3066: to_int32 (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "Use tf.cast instead.\n",
      "Train on 48000 samples, validate on 12000 samples\n",
      "Epoch 1/25\n",
      "48000/48000 [==============================] - 48s 1ms/step - loss: 0.9118 - accuracy: 0.6564 - val_loss: 0.5545 - val_accuracy: 0.7811\n",
      "Epoch 2/25\n",
      "48000/48000 [==============================] - 48s 1ms/step - loss: 0.5502 - accuracy: 0.7934 - val_loss: 0.4399 - val_accuracy: 0.8383\n",
      "Epoch 3/25\n",
      "48000/48000 [==============================] - 49s 1ms/step - loss: 0.4715 - accuracy: 0.8259 - val_loss: 0.4084 - val_accuracy: 0.8496\n",
      "Epoch 4/25\n",
      "48000/48000 [==============================] - 49s 1ms/step - loss: 0.4234 - accuracy: 0.8443 - val_loss: 0.3606 - val_accuracy: 0.8713\n",
      "Epoch 5/25\n",
      "48000/48000 [==============================] - 48s 993us/step - loss: 0.3919 - accuracy: 0.8569 - val_loss: 0.3408 - val_accuracy: 0.8762\n",
      "Epoch 6/25\n",
      "48000/48000 [==============================] - 46s 954us/step - loss: 0.3676 - accuracy: 0.8649 - val_loss: 0.3223 - val_accuracy: 0.8829\n",
      "Epoch 7/25\n",
      "48000/48000 [==============================] - 45s 937us/step - loss: 0.3498 - accuracy: 0.8703 - val_loss: 0.3018 - val_accuracy: 0.8899\n",
      "Epoch 8/25\n",
      "48000/48000 [==============================] - 47s 970us/step - loss: 0.3375 - accuracy: 0.8759 - val_loss: 0.2949 - val_accuracy: 0.8926\n",
      "Epoch 9/25\n",
      "48000/48000 [==============================] - 46s 968us/step - loss: 0.3273 - accuracy: 0.8814 - val_loss: 0.2874 - val_accuracy: 0.8955\n",
      "Epoch 10/25\n",
      "48000/48000 [==============================] - 46s 956us/step - loss: 0.3108 - accuracy: 0.8857 - val_loss: 0.2774 - val_accuracy: 0.8986\n",
      "Epoch 11/25\n",
      "48000/48000 [==============================] - 46s 955us/step - loss: 0.3040 - accuracy: 0.8887 - val_loss: 0.2724 - val_accuracy: 0.9027\n",
      "Epoch 12/25\n",
      "48000/48000 [==============================] - 46s 960us/step - loss: 0.2950 - accuracy: 0.8919 - val_loss: 0.2657 - val_accuracy: 0.9014\n",
      "Epoch 13/25\n",
      "48000/48000 [==============================] - 47s 969us/step - loss: 0.2882 - accuracy: 0.8936 - val_loss: 0.2564 - val_accuracy: 0.9049\n",
      "Epoch 14/25\n",
      "48000/48000 [==============================] - 46s 962us/step - loss: 0.2830 - accuracy: 0.8947 - val_loss: 0.2673 - val_accuracy: 0.9028\n",
      "Epoch 15/25\n",
      "48000/48000 [==============================] - 47s 974us/step - loss: 0.2764 - accuracy: 0.8985 - val_loss: 0.2552 - val_accuracy: 0.9072\n",
      "Epoch 16/25\n",
      "48000/48000 [==============================] - 46s 959us/step - loss: 0.2722 - accuracy: 0.8990 - val_loss: 0.2499 - val_accuracy: 0.9087\n",
      "Epoch 17/25\n",
      "48000/48000 [==============================] - 48s 1ms/step - loss: 0.2690 - accuracy: 0.9015 - val_loss: 0.2464 - val_accuracy: 0.9075\n",
      "Epoch 18/25\n",
      "48000/48000 [==============================] - 47s 980us/step - loss: 0.2606 - accuracy: 0.9040 - val_loss: 0.2426 - val_accuracy: 0.9097\n",
      "Epoch 19/25\n",
      "48000/48000 [==============================] - 45s 935us/step - loss: 0.2608 - accuracy: 0.9045 - val_loss: 0.2454 - val_accuracy: 0.9084\n",
      "Epoch 20/25\n",
      "48000/48000 [==============================] - 44s 909us/step - loss: 0.2552 - accuracy: 0.9047 - val_loss: 0.2424 - val_accuracy: 0.9087\n",
      "Epoch 21/25\n",
      "48000/48000 [==============================] - 43s 890us/step - loss: 0.2495 - accuracy: 0.9064 - val_loss: 0.2458 - val_accuracy: 0.9076\n",
      "Epoch 22/25\n",
      "48000/48000 [==============================] - 46s 956us/step - loss: 0.2443 - accuracy: 0.9094 - val_loss: 0.2405 - val_accuracy: 0.9126\n",
      "Epoch 23/25\n",
      "48000/48000 [==============================] - 47s 987us/step - loss: 0.2425 - accuracy: 0.9096 - val_loss: 0.2374 - val_accuracy: 0.9097\n",
      "Epoch 24/25\n",
      "48000/48000 [==============================] - 45s 934us/step - loss: 0.2422 - accuracy: 0.9091 - val_loss: 0.2365 - val_accuracy: 0.9112\n",
      "Epoch 25/25\n",
      "48000/48000 [==============================] - 44s 910us/step - loss: 0.2338 - accuracy: 0.9128 - val_loss: 0.2272 - val_accuracy: 0.9159\n"
     ]
    }
   ],
   "source": [
    "history3 = cnn3.fit(X_train, y_train,\n",
    "          batch_size=256,\n",
    "          epochs=25,\n",
    "          verbose=1,\n",
    "          validation_data=(X_val, y_val))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Test loss: 0.2086141019642353\n",
      "Test accuracy: 0.9203000068664551\n"
     ]
    }
   ],
   "source": [
    "score3 = cnn3.evaluate(X_test, y_test, verbose=0)\n",
    "print('Test loss:', score3[0])\n",
    "print('Test accuracy:', score3[1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\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",
    "\n",
    "accuracy = history3.history['accuracy']\n",
    "val_accuracy = history3.history['val_accuracy']\n",
    "loss = history3.history['loss']\n",
    "val_loss = history3.history['val_loss']\n",
    "epochs = range(len(accuracy))\n",
    "\n",
    "plt.plot(epochs, accuracy, 'bo', label='Training accuracy')\n",
    "plt.plot(epochs, val_accuracy, 'b', label='Validation accuracy')\n",
    "plt.title('Training and validation accuracy')\n",
    "plt.legend()\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",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "gen = ImageDataGenerator(rotation_range=8, width_shift_range=0.08, shear_range=0.3,\n",
    "                               height_shift_range=0.08, zoom_range=0.08)\n",
    "batches = gen.flow(X_train, y_train, batch_size=256)\n",
    "val_batches = gen.flow(X_val, y_val, batch_size=256)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      "187/187 [==============================] - 63s 336ms/step - loss: 0.2865 - accuracy: 0.8923 - val_loss: 0.2896 - val_accuracy: 0.9049\n",
      "Epoch 2/50\n",
      "187/187 [==============================] - 70s 375ms/step - loss: 0.2904 - accuracy: 0.8923 - val_loss: 0.2170 - val_accuracy: 0.9022\n",
      "Epoch 3/50\n",
      "187/187 [==============================] - 65s 346ms/step - loss: 0.2881 - accuracy: 0.8930 - val_loss: 0.2412 - val_accuracy: 0.9022\n",
      "Epoch 4/50\n",
      "187/187 [==============================] - 66s 351ms/step - loss: 0.2909 - accuracy: 0.8919 - val_loss: 0.2890 - val_accuracy: 0.9014\n",
      "Epoch 5/50\n",
      "187/187 [==============================] - 384s 2s/step - loss: 0.2929 - accuracy: 0.8902 - val_loss: 0.1475 - val_accuracy: 0.9093\n",
      "Epoch 6/50\n",
      "187/187 [==============================] - 68s 363ms/step - loss: 0.2871 - accuracy: 0.8924 - val_loss: 0.2713 - val_accuracy: 0.9089\n",
      "Epoch 7/50\n",
      "187/187 [==============================] - 66s 354ms/step - loss: 0.2852 - accuracy: 0.8934 - val_loss: 0.2109 - val_accuracy: 0.9037\n",
      "Epoch 8/50\n",
      "187/187 [==============================] - 81s 435ms/step - loss: 0.2868 - accuracy: 0.8934 - val_loss: 0.2335 - val_accuracy: 0.9055\n",
      "Epoch 9/50\n",
      "187/187 [==============================] - 72s 386ms/step - loss: 0.2874 - accuracy: 0.8927 - val_loss: 0.2200 - val_accuracy: 0.9061\n",
      "Epoch 10/50\n",
      "187/187 [==============================] - 70s 375ms/step - loss: 0.2897 - accuracy: 0.8926 - val_loss: 0.2904 - val_accuracy: 0.9021\n",
      "Epoch 11/50\n",
      "187/187 [==============================] - 68s 366ms/step - loss: 0.2862 - accuracy: 0.8934 - val_loss: 0.2752 - val_accuracy: 0.9025\n",
      "Epoch 12/50\n",
      "187/187 [==============================] - 71s 379ms/step - loss: 0.2923 - accuracy: 0.8919 - val_loss: 0.2734 - val_accuracy: 0.9046\n",
      "Epoch 13/50\n",
      "187/187 [==============================] - 68s 366ms/step - loss: 0.2880 - accuracy: 0.8919 - val_loss: 0.3394 - val_accuracy: 0.9071\n",
      "Epoch 14/50\n",
      "187/187 [==============================] - 70s 374ms/step - loss: 0.2840 - accuracy: 0.8927 - val_loss: 0.2611 - val_accuracy: 0.9024\n",
      "Epoch 15/50\n",
      "187/187 [==============================] - 70s 372ms/step - loss: 0.2874 - accuracy: 0.8914 - val_loss: 0.3033 - val_accuracy: 0.9011\n",
      "Epoch 16/50\n",
      "187/187 [==============================] - 69s 367ms/step - loss: 0.2846 - accuracy: 0.8924 - val_loss: 0.3356 - val_accuracy: 0.9054\n",
      "Epoch 17/50\n",
      "187/187 [==============================] - 68s 365ms/step - loss: 0.2819 - accuracy: 0.8927 - val_loss: 0.3013 - val_accuracy: 0.9082\n",
      "Epoch 18/50\n",
      "187/187 [==============================] - 70s 377ms/step - loss: 0.2784 - accuracy: 0.8954 - val_loss: 0.2734 - val_accuracy: 0.9084\n",
      "Epoch 19/50\n",
      "187/187 [==============================] - 72s 383ms/step - loss: 0.2878 - accuracy: 0.8911 - val_loss: 0.2597 - val_accuracy: 0.9023\n",
      "Epoch 20/50\n",
      "187/187 [==============================] - 71s 378ms/step - loss: 0.2807 - accuracy: 0.8942 - val_loss: 0.2102 - val_accuracy: 0.9005\n",
      "Epoch 21/50\n",
      "187/187 [==============================] - 72s 386ms/step - loss: 0.2835 - accuracy: 0.8940 - val_loss: 0.1806 - val_accuracy: 0.9047\n",
      "Epoch 22/50\n",
      "187/187 [==============================] - 71s 382ms/step - loss: 0.2827 - accuracy: 0.8967 - val_loss: 0.3439 - val_accuracy: 0.9059\n",
      "Epoch 23/50\n",
      "187/187 [==============================] - 72s 387ms/step - loss: 0.2820 - accuracy: 0.8946 - val_loss: 0.2948 - val_accuracy: 0.9041\n",
      "Epoch 24/50\n",
      "187/187 [==============================] - 78s 417ms/step - loss: 0.2838 - accuracy: 0.8923 - val_loss: 0.2735 - val_accuracy: 0.9030\n",
      "Epoch 25/50\n",
      "187/187 [==============================] - 78s 416ms/step - loss: 0.2771 - accuracy: 0.8948 - val_loss: 0.2868 - val_accuracy: 0.8983\n",
      "Epoch 26/50\n",
      "187/187 [==============================] - 77s 411ms/step - loss: 0.2853 - accuracy: 0.8921 - val_loss: 0.2307 - val_accuracy: 0.9049\n",
      "Epoch 27/50\n",
      "187/187 [==============================] - 76s 409ms/step - loss: 0.2797 - accuracy: 0.8950 - val_loss: 0.2352 - val_accuracy: 0.9033\n",
      "Epoch 28/50\n",
      "187/187 [==============================] - 73s 393ms/step - loss: 0.2810 - accuracy: 0.8938 - val_loss: 0.2947 - val_accuracy: 0.9033\n",
      "Epoch 29/50\n",
      "187/187 [==============================] - 76s 406ms/step - loss: 0.2791 - accuracy: 0.8962 - val_loss: 0.2326 - val_accuracy: 0.9084\n",
      "Epoch 30/50\n",
      "187/187 [==============================] - 77s 409ms/step - loss: 0.2763 - accuracy: 0.8964 - val_loss: 0.2878 - val_accuracy: 0.9106\n",
      "Epoch 31/50\n",
      "187/187 [==============================] - 74s 398ms/step - loss: 0.2810 - accuracy: 0.8940 - val_loss: 0.2987 - val_accuracy: 0.9083\n",
      "Epoch 32/50\n",
      "187/187 [==============================] - 65s 347ms/step - loss: 0.2774 - accuracy: 0.8970 - val_loss: 0.1652 - val_accuracy: 0.9052\n",
      "Epoch 33/50\n",
      "187/187 [==============================] - 57s 304ms/step - loss: 0.2818 - accuracy: 0.8961 - val_loss: 0.2287 - val_accuracy: 0.9061\n",
      "Epoch 34/50\n",
      "187/187 [==============================] - 56s 301ms/step - loss: 0.2789 - accuracy: 0.8955 - val_loss: 0.2621 - val_accuracy: 0.9068\n",
      "Epoch 35/50\n",
      "187/187 [==============================] - 57s 307ms/step - loss: 0.2794 - accuracy: 0.8953 - val_loss: 0.2447 - val_accuracy: 0.9080\n",
      "Epoch 36/50\n",
      "187/187 [==============================] - 57s 306ms/step - loss: 0.2810 - accuracy: 0.8948 - val_loss: 0.2436 - val_accuracy: 0.9111\n",
      "Epoch 37/50\n",
      "187/187 [==============================] - 57s 304ms/step - loss: 0.2754 - accuracy: 0.8967 - val_loss: 0.2412 - val_accuracy: 0.9102\n",
      "Epoch 38/50\n",
      "187/187 [==============================] - 58s 312ms/step - loss: 0.2790 - accuracy: 0.8953 - val_loss: 0.1727 - val_accuracy: 0.9108\n",
      "Epoch 39/50\n",
      "187/187 [==============================] - 58s 311ms/step - loss: 0.2764 - accuracy: 0.8974 - val_loss: 0.2404 - val_accuracy: 0.9077\n",
      "Epoch 40/50\n",
      "187/187 [==============================] - 57s 305ms/step - loss: 0.2781 - accuracy: 0.8961 - val_loss: 0.3064 - val_accuracy: 0.9033\n",
      "Epoch 41/50\n",
      "187/187 [==============================] - 57s 307ms/step - loss: 0.2764 - accuracy: 0.8943 - val_loss: 0.2347 - val_accuracy: 0.9084\n",
      "Epoch 42/50\n",
      "187/187 [==============================] - 58s 309ms/step - loss: 0.2786 - accuracy: 0.8964 - val_loss: 0.2075 - val_accuracy: 0.9090\n",
      "Epoch 43/50\n",
      "187/187 [==============================] - 57s 307ms/step - loss: 0.2789 - accuracy: 0.8946 - val_loss: 0.2419 - val_accuracy: 0.9055\n",
      "Epoch 44/50\n",
      "187/187 [==============================] - 58s 309ms/step - loss: 0.2719 - accuracy: 0.8987 - val_loss: 0.2395 - val_accuracy: 0.9108\n",
      "Epoch 45/50\n",
      "187/187 [==============================] - 57s 307ms/step - loss: 0.2786 - accuracy: 0.8967 - val_loss: 0.2229 - val_accuracy: 0.9060\n",
      "Epoch 46/50\n",
      "187/187 [==============================] - 59s 314ms/step - loss: 0.2762 - accuracy: 0.8958 - val_loss: 0.2474 - val_accuracy: 0.9084\n",
      "Epoch 47/50\n",
      "187/187 [==============================] - 58s 310ms/step - loss: 0.2813 - accuracy: 0.8955 - val_loss: 0.2070 - val_accuracy: 0.9088\n",
      "Epoch 48/50\n",
      "187/187 [==============================] - 58s 312ms/step - loss: 0.2765 - accuracy: 0.8968 - val_loss: 0.2614 - val_accuracy: 0.9090\n",
      "Epoch 49/50\n",
      "187/187 [==============================] - 60s 322ms/step - loss: 0.2757 - accuracy: 0.8962 - val_loss: 0.2378 - val_accuracy: 0.9102\n",
      "Epoch 50/50\n",
      "187/187 [==============================] - 49s 264ms/step - loss: 0.2770 - accuracy: 0.8955 - val_loss: 0.2311 - val_accuracy: 0.9070\n"
     ]
    }
   ],
   "source": [
    "history3 = cnn3.fit_generator(batches, steps_per_epoch=48000//256, epochs=50,\n",
    "                    validation_data=val_batches, validation_steps=12000//256, use_multiprocessing=True)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYQAAAEICAYAAABfz4NwAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4zLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvIxREBQAAIABJREFUeJzsnXmYFNW5/z8vw77vyCIOKqIsI+CIa2TRGHeNMSrilqshakxMNPlJ1JjExHuNGjUabqLmukRQ3K6JN+KSBFCTuKEsioogICA7CoKsM7y/P94uuqanqruqu6e7hz6f5+mnu6qrTp3q5XzPu5xzRFVxOBwOh6NJsSvgcDgcjtLACYLD4XA4ACcIDofD4UjgBMHhcDgcgBMEh8PhcCRwguBwOBwOwAmCI4+ISIWIbBaRvvk8tpiIyP4ikvfcbBE5TkSW+Lbni8hXohybxbX+KCLXZXt+mnJ/JSIP5btcR/FoWuwKOIqHiGz2bbYGtgO1ie3vqOrkOOWpai3QNt/HlgOqOiAf5YjIpcD5qjrKV/al+SjbsefjBKGMUdXdDXKiB3qpqv497HgRaaqqNYWom8PhKDzOZeQIJeESeFxEHhORTcD5InKEiLwuIhtEZKWI3C0izRLHNxURFZHKxPakxPvPi8gmEXlNRPrFPTbx/oki8pGIbBSRe0TkXyJycUi9o9TxOyKyUEQ+F5G7fedWiMidIrJeRD4GTkjz+dwgIlNS9k0UkTsSry8VkQ8S9/NxovceVtZyERmVeN1aRB5J1G0ecEjAdRclyp0nIqcl9g8Bfgd8JeGOW+f7bH/uO/+yxL2vF5E/i0jPKJ9NJkTkjER9NojINBEZ4HvvOhFZISJfiMiHvns9XETeSexfLSK3Rb2eowFQVfdwD4AlwHEp+34F7ABOxToPrYBDgcMw63Jf4CPgysTxTQEFKhPbk4B1QDXQDHgcmJTFsd2BTcDpifeuBnYCF4fcS5Q6/gXoAFQCn3n3DlwJzAP6AF2AV+xvEnidfYHNQBtf2WuA6sT2qYljBBgDbAWqEu8dByzxlbUcGJV4fTswA+gE7AO8n3Ls2UDPxHdyXqIOPRLvXQrMSKnnJODnidfHJ+o4FGgJ/DcwLcpnE3D/vwIeSrw+KFGPMYnv6LrE594MGAR8AuyVOLYfsG/i9VvA2MTrdsBhxf4vlPPDWQiOTPxTVf9PVXep6lZVfUtV31DVGlVdBNwHjExz/lOqOlNVdwKTsYYo7rGnALNV9S+J9+7ExCOQiHX8L1XdqKpLsMbXu9bZwJ2qulxV1wO3pLnOIuA9TKgAvgpsUNWZiff/T1UXqTEN+AcQGDhO4WzgV6r6uap+gvX6/dd9QlVXJr6TRzExr45QLsA44I+qOltVtwETgJEi0sd3TNhnk45zgWdVdVriO7oFaI8Jcw0mPoMSbsfFic8OTNj7i0gXVd2kqm9EvA9HA+AEwZGJZf4NETlQRJ4TkVUi8gVwE9A1zfmrfK+3kD6QHHZsL389VFWxHnUgEesY6VpYzzYdjwJjE6/Pw4TMq8cpIvKGiHwmIhuw3nm6z8qjZ7o6iMjFIjIn4ZrZABwYsVyw+9tdnqp+AXwO9PYdE+c7Cyt3F/Yd9VbV+cA12PewJuGC3Ctx6LeAgcB8EXlTRE6KeB+OBsAJgiMTqSmX92K94v1VtT1wI+YSaUhWYi4cAEREqNuApZJLHVcCe/u2M6XFPg4cl+hhn44JBCLSCngK+C/MndMReCliPVaF1UFE9gV+D1wOdEmU+6Gv3EwpsiswN5RXXjvMNfVphHrFKbcJ9p19CqCqk1T1KMxdVIF9LqjqfFU9F3ML/gZ4WkRa5lgXR5Y4QXDEpR2wEfhSRA4CvlOAa/4VGC4ip4pIU+AqoFsD1fEJ4Aci0ltEugDXpjtYVVcD/wQeBOar6oLEWy2A5sBaoFZETgGOjVGH60Sko9g4jSt977XFGv21mDZeilkIHquBPl4QPYDHgEtEpEpEWmAN86uqGmpxxajzaSIyKnHtH2NxnzdE5CARGZ243tbEoxa7gQtEpGvCotiYuLddOdbFkSVOEBxxuQa4CPuz34v1kBuURKN7DnAHsB7YD5iFjZvIdx1/j/n638UCnk9FOOdRLEj8qK/OG4AfAs9ggdmzMGGLws8wS2UJ8DzwJ1+5c4G7gTcTxxwI+P3ufwMWAKtFxO/68c5/AXPdPJM4vy8WV8gJVZ2Hfea/x8TqBOC0RDyhBXArFvdZhVkkNyROPQn4QCyL7XbgHFXdkWt9HNkh5o51OBoPIlKBuSjOUtVXi10fh2NPwVkIjkaBiJwgIh0SboefYpkrbxa5Wg7HHoUTBEdj4WhgEeZ2OAE4Q1XDXEYOhyMLnMvI4XA4HICzEBwOh8ORoFFNbte1a1etrKwsdjUcDoejUfH222+vU9V0qdpAREEQkROA32IDSv6oqrekvH8Z8F0st3gzMF5V3xebuOwDYH7i0NdV9bLEOYcAD2Hz40wFrtIM/qvKykpmzpwZpcoOh8PhSCAimUbcAxFcRokUv4nAidgQ87EiMjDlsEdVdYiqDsXyje/wvfexqg5NPC7z7f89MB7on3iEzirpcDgcjoYnSgxhBLAwMUnXDmAKycm8gN3zoXi0IcPw+cR0u+1V9bWEVfAn4IxYNXc4HA5HXokiCL2pO9HWcgLmkRGR7ybmj78V+L7vrX4iMktEXpbkEoG9qTs5WWCZiXLHi8hMEZm5du3aCNV1OBwORzZEiSEETcZVzwJQ1YnARBE5DxuWfhGJofGquj4RM/iziAyKWmai3Puw6Yuprq52ObIORwHZuXMny5cvZ9u2bcWuiiMCLVu2pE+fPjRrFjaVVXqiCMJy6s682AebNiCMKVh8gMTAoe2J128nLIgDEmX651/PVKbD4SgCy5cvp127dlRWVmKTzDpKFVVl/fr1LF++nH79+mU+IYAoLqO3sAUs+olIcxILYfgPEJH+vs2Tscm1EJFuiaC0N21vf2CRqq4ENiWWzxPgQmyVJofDUUJs27aNLl26ODFoBIgIXbp0ycmay2ghqGqNiFwJvIilnT6gqvNE5CZgpqo+C1wpIsdhqx99jrmLAI4BbhKRGiwl9TJV/Szx3uUk006fTzwcDkeJ4cSg8ZDrdxVpHIKqTsXGCvj33eh7fVXIeU8DT4e8NxMYHLmmDkeO7NwJf/oTXHwxVFQUuzYOR+nhpq5wlA0vvwyXXgrTphW7Jo6orF+/nqFDhzJ06FD22msvevfuvXt7x45oyyZ861vfYv78+WmPmThxIpMnT057TFSOPvpoZs+enZeyCk2jmrrC4ciFTZvsefHi4tZjT2byZLj+eli6FPr2hZtvhnE5LL/TpUuX3Y3rz3/+c9q2bcuPfvSjOseoKqpKkybB/dsHH3ww43W++93vZl/JPQhnITjKhi1b7HnJkqJWY49l8mQYPx4++QRU7Xn8eNufbxYuXMjgwYO57LLLGD58OCtXrmT8+PFUV1czaNAgbrrppt3Hej32mpoaOnbsyIQJEzj44IM54ogjWLNmDQA33HADd9111+7jJ0yYwIgRIxgwYAD//ve/Afjyyy/5xje+wcEHH8zYsWOprq7OaAlMmjSJIUOGMHjwYK677joAampquOCCC3bvv/vuuwG48847GThwIAcffDDnn39+3j+zKDgLwVE2bN1qz04QGobrr0+KrseWLbY/FyshjPfff58HH3yQP/zhDwDccsstdO7cmZqaGkaPHs1ZZ53FwIF1Z9nZuHEjI0eO5JZbbuHqq6/mgQceYMKECfXKVlXefPNNnn32WW666SZeeOEF7rnnHvbaay+efvpp5syZw/Dhw9PWb/ny5dxwww3MnDmTDh06cNxxx/HXv/6Vbt26sW7dOt59910ANmzYAMCtt97KJ598QvPmzXfvKzTOQnDs5vPP4dBD4f33i12ThsFZCA3L0qXx9ufKfvvtx6GHHrp7+7HHHmP48OEMHz6cDz74gPcDfsitWrXixBNPBOCQQw5hSciP4cwzz6x3zD//+U/OPfdcAA4++GAGDRqUtn5vvPEGY8aMoWvXrjRr1ozzzjuPV155hf3335/58+dz1VVX8eKLL9KhQwcABg0axPnnn8/kyZOzHliWK04QHLuZNw9mzoTXXy92TRoGJwgNS9++8fbnSps2bXa/XrBgAb/97W+ZNm0ac+fO5YQTTgjMx2/evPnu1xUVFdTU1ASW3aJFi3rHxF1MLOz4Ll26MHfuXI4++mjuvvtuvvOd7wDw4osvctlll/Hmm29SXV1NbW1trOvlAycIjt2sWmXPe+qUUZ7LaOXK5GtH/rj5Zmjduu6+1q1tf0PzxRdf0K5dO9q3b8/KlSt58cUX836No48+mieeeAKAd999N9AC8XP44Yczffp01q9fT01NDVOmTGHkyJGsXbsWVeWb3/wmv/jFL3jnnXeora1l+fLljBkzhttuu421a9eyJdX/VgBcDMGxG08Q1q0rbj0aCv//a+lSGDCgeHXZE/HiBPnMMorK8OHDGThwIIMHD2bfffflqKOOyvs1vve973HhhRdSVVXF8OHDGTx48G53TxB9+vThpptuYtSoUagqp556KieffDLvvPMOl1xyCaqKiPDrX/+ampoazjvvPDZt2sSuXbu49tpradeuXd7vISNeylZjeBxyyCHqaDiuv14VVC++OPey7rlH9cADVXftyr2sfHHFFXZ/oPrCC8WuTePg/fffL3YVSoadO3fq1q1bVVX1o48+0srKSt25c2eRa1WfoO8Mm1UiYxvrLATHbvJpITz/PHz4ofUU99kn9/LywZYt0KyZjVh2cQRHXDZv3syxxx5LTU0Nqsq9995L06Z7VhO6Z92NIyfyKQizZtnz3LmlIwhbt1pdPvnECYIjPh07duTtt98udjUaFBdUduwmX4KwerUFbsEEoVTYsgXatjXfthMEh6M+zkJw7CZfWUaedQClJwitW0OXLk4QHI4gnIXgAGDXLuvZV1TAxo3mZ88WTxBGjSotQdi6FVq1gspKJwgORxBOEBwAfPYZ1NRA/8RSR+vXZ1/WrFnQrx8ccwx89FHp5Px7FkJlpVlDpVIvh6NUcILgAJLuosGJFSpyiSPMmgXDh0NVlVkepTIVhl8QwILLjtJm1KhR9QaZ3XXXXVxxxRVpz2vbti0AK1as4Kyzzgote+bMmWnLueuuu+oMEDvppJPyMs/Qz3/+c26//facy8k3kQRBRE4QkfkislBE6s0EJSKXici7IjJbRP4pIgMT+78qIm8n3ntbRMb4zpmRKHN24tE9f7fliIsnCEOG2HO2gvDFF7BwIQwbZoIApeM28ruMwLmNGgNjx45lypQpdfZNmTKFsWPHRjq/V69ePPXUU1lfP1UQpk6dSseOHbMur9TJKAiJNZEnAicCA4GxXoPv41FVHaKqQ4FbgTsS+9cBp6rqEGxZzUdSzhunqkMTjzW53IgjN/JlIcyZY8/DhsG++1qP3NtXbFItBCcIpc9ZZ53FX//6V7Zv3w7AkiVLWLFiBUcfffTucQHDhw9nyJAh/OUv9ZdlX7JkCYMTP+qtW7dy7rnnUlVVxTnnnMNWn8/w8ssv3z119s9+9jMA7r77blasWMHo0aMZPXo0AJWVlaxL/DnuuOMOBg8ezODBg3dPnb1kyRIOOuggvv3tbzNo0CCOP/74OtcJYvbs2Rx++OFUVVXx9a9/nc8//3z39QcOHEhVVdXuSfVefvnl3QsEDRs2jE3eIh95IkqW0QhgoaouAhCRKcDpwG5HgKp+4Tu+DaCJ/b58E+YBLUWkhapuz7XijvziCYI3gWO2mUZeQHnYMAtQDx5cOhaCJwi9etkANScI8fjBDyDfC4ENHQqJtjSQLl26MGLECF544QVOP/10pkyZwjnnnIOI0LJlS5555hnat2/PunXrOPzwwznttNNC1xX+/e9/T+vWrZk7dy5z586tM331zTffTOfOnamtreXYY49l7ty5fP/73+eOO+5g+vTpdO3atU5Zb7/9Ng8++CBvvPEGqsphhx3GyJEj6dSpEwsWLOCxxx7j/vvv5+yzz+bpp59Ou77BhRdeyD333MPIkSO58cYb+cUvfsFdd93FLbfcwuLFi2nRosVuN9Xtt9/OxIkTOeqoo9i8eTMtW7aM8WlnJorLqDewzLe9PLGvDiLyXRH5GLMQvh9QzjeAWSli8GDCXfRTCfkWRWS8iMwUkZlr99RZ10qA1avNndKvn21nayHMmgU9ekDPnrZdVWWCEHOiyLyjmnQZNWliA9ScIDQO/G4jv7tIVbnuuuuoqqriuOOO49NPP2X16tWh5bzyyiu7G+aqqiqqPJ8m8MQTTzB8+HCGDRvGvHnzMk5c989//pOvf/3rtGnThrZt23LmmWfy6quvAtCvXz+GDh0KpJ9iG2x9hg0bNjBy5EgALrroIl555ZXddRw3bhyTJk3aPSL6qKOO4uqrr+buu+9mw4YNeR8pHaW0oIa63t9bVScCE0XkPOAGzEVkBYgMAn4NHO87ZZyqfioi7YCngQuAPwWUex9wH0B1dXWRm5U9l1WrYK+9oHlzaN8+N0EYNiy5XVUFf/yjDVTr1Ss/dc0GbyZkbzZOl3oan3Q9+YbkjDPO4Oqrr+add95h69atu3v2kydPZu3atbz99ts0a9aMysrKwCmv/QT1OxcvXsztt9/OW2+9RadOnbj44oszlqNpejje1Nlg02dnchmF8dxzz/HKK6/w7LPP8stf/pJ58+YxYcIETj75ZKZOncrhhx/O3//+dw488MCsyg8iioWwHNjbt90HWJHm+CnAGd6GiPQBngEuVNWPvf2q+mnieRPwKOaachQJTxAAunXLThC2b7c1FVIFAYrvNvLigk4QGh9t27Zl1KhR/Md//EedYPLGjRvp3r07zZo1Y/r06XySIW3smGOOYXJiPc/33nuPuYkf5RdffEGbNm3o0KEDq1ev5vnnn999Trt27QL99Mcccwx//vOf2bJlC19++SXPPPMMX/nKV2LfW4cOHejUqdNu6+KRRx5h5MiR7Nq1i2XLljF69GhuvfVWNmzYwObNm/n4448ZMmQI1157LdXV1Xz44Yexr5mOKBbCW0B/EekHfAqcC5znP0BE+qvqgsTmycCCxP6OwHPAT1T1X77jmwIdVXWdiDQDTgH+nuvNOLJn1arkGISuXbMThHnzbCxDmCCccELu9cwWr5PWqpU9V1aam8xzIzlKm7Fjx3LmmWfWyTgaN24cp556KtXV1QwdOjRjT/nyyy/nW9/6FlVVVQwdOpQRI6wPevDBBzNs2DAGDRpUb+rs8ePHc+KJJ9KzZ0+mT5++e//w4cO5+OKLd5dx6aWXMmzYsLTuoTAefvhhLrvsMrZs2cK+++7Lgw8+SG1tLeeffz4bN25EVfnhD39Ix44d+elPf8r06dOpqKhg4MCBu1d/yxtRpkQFTgI+Aj4Grk/suwk4LfH6t1jQeDYwHRiU2H8D8GViv/fojgWe3wbmJs77LVCRqR5u+uuGo2tX1csvt9cnn6w6fHj8Mv74R5taesGCuvv33lt13Ljc65gL8+db3SZNsu1Jk2zbze6cHjf9deOjwae/VtWpwNSUfTf6Xl8Vct6vgF+FFHtIlGs7Gp6dO80i8FxGXbtm5+KZNQvatbN0Uz9eYLmYBLmMwNxGBx1UjBo5HKWHG6nsYE1iBIhfELJxGc2aZWmETVJ+VVVV8MEHsGNHbvXMhVSXkZdN5eIIDkcSJwiO3WMQ/IKwdWvdJSczUVtrA9D88QOPqiqLLeQ5/hWLVAvBy6hygpAZLXbOsCMyuX5XThAc9QShWzd7jmMlLFwIX34ZLghQXLdRqiC4sQjRaNmyJevXr3ei0AhQVdavX5/TYDW3HoIj0EIAE4S+faOV4R+hnMoBB1hvvJiCkOoyApd6GoU+ffqwfPly3KDQxkHLli3p06dP1uc7QXDsFoQePezZE4Q4bcCsWdboD0yd5Qpo2tSmxCjmnEapFgKYIARMf+Pw0axZM/p5ARfHHo9zGTlYtQo6dQJvgKXfQojKrFk2b1GzZsHvFzvTKEwQ1qyJFytxOPZknCA4WLUqaR1AfEFQrT9lRSpVVXadNUWa0zbMZQTObeRweDhBcNSZtgLMWmjSJLogfPqpHZtJEADefTf7euaCZwU4QXA4wnGC4KgnCE2a2EL0UQUhXUDZo9iZRlu2mDvL79JyYxEcjro4QXDUEwSINzht1iwQSTb6QXTvbm6pYgnC1q114wdg9WnRwgmCw+HhBKHM2bzZHkGCEDXLaNYsSy1NLGMbSjEDy1u21J/Ezo1FcDjq4gShzPHWE8nVQkjnLvKoqkrOiFpovNXSUnFjERyOJE4QypwwQYi6JsJnn8Enn0QThIMPtjUTFizIfGy+CXIZgROExsR778H/+3/FX31vT8YJQpmTOkrZw7MQMv35vDV2o1oIUBy3UZDLCEwQ1q61aTccpc2TT8Jtt2W/mp8jM04Qypx0glBbCxs3pj8/SoaRx4EH2qjlYglCmIUAZuU4Shvvt1qssSzlgBOEMmfVKguueoPRPKIOTps1C/r0qX9+EC1amCgUQxDSuYwAFi8uaHUcWeAEoeGJJAgicoKIzBeRhSIyIeD9y0TkXRGZLSL/FJGBvvd+kjhvvoh8LWqZjsKwapWlhFZU1N0fdT6jefNgyJDo16uqKs6cRmEuIzcWofHgBKHhySgIIlIBTAROBAYCY/0NfoJHVXWIqg4FbgXuSJw7EFuDeRBwAvDfIlIRsUxHAQgagwDRLARVCxAPGBD9elVVsGwZfP55vHrmSpjLqEcPaNnSCUJjwBMELxHCkX+iWAgjgIWqukhVdwBTgNP9B6jqF77NNoAXijwdmKKq21V1MbAwUV7GMh2FIUwQoqyJsHKlBWP7949+vaFD7fmdd6Kfkw/CBEHEjUVoDKg6C6EQRBGE3sAy3/byxL46iMh3ReRjzEL4foZzI5WZKHe8iMwUkZluTvb8kzqxnUcUC+Gjj+z5gAOiX2/ECHt+/fXo5+SDrVuDXUbgUk8bAxs2JJdgdYLQcEQRBAnYVy8ZUVUnqup+wLXADRnOjVRmotz7VLVaVau7ed1WR17wel1BFkKbNhYETicI3niCOBZCp062qP1rr8Wra66EWQiw5wjC1Kmw//6ZM8MaI551AE4QGpIogrAc2Nu33QdYkeb4KcAZGc6NW6ajAfj8c9i5M1gQRDJPX7FggYnG3nuHHxPE4YebhVCoAUY7d9ro6HSCsG6dTeHRWNm+Hb73Pfj4Y4vR7Gl4gtC8uROEhiSKILwF9BeRfiLSHAsSP+s/QET8fcSTAW8s6rPAuSLSQkT6Af2BN6OU6Wh4wsYgeGSavuKjj6xH2iRm8vIRR8D69YUbsRy0FoKfPWEswt13w6JF9npPHGTn/VYHDnSC0JBk/Curag1wJfAi8AHwhKrOE5GbROS0xGFXisg8EZkNXA1clDh3HvAE8D7wAvBdVa0NKzPP9+bIQCZByDR9xYIF8dxFHkccYc+FiiMErZbmx0s9baxjEdasgV/+EnononCN2dIJw/utVlU5QWhIIvXtVHWqqh6gqvup6s2JfTeq6rOJ11ep6iBVHaqqo/2Nu6renDhvgKo+n67MxsrkydbLbNLEnidPLnaNwvHX9ZxzbF82FkJtLSxcmJ0gDBwI7dsXLo6QSRA8C2HhwoJUpw6qMGoUPP109mXceKNZQbfdZtt7qiA0b24pzps2Ja0+R35xI5VzZPJkGD/e3A2q9jx+fGmKQmpdvcb+lVeCj08nCMuWWdZHnAwjjyZNLNuoUIKQyWXUvbtZCdOmFaY+ftavh5dfhunTszv/3Xfh/vvh8svhkENs354qCHvtlcyIcwmHDUPZCkI2vfqgc66/vv4i7Vu22P5SI6iuADeH2Gddu1rgOWi66mwyjPwccYQ1Zps2ZXd+HDJZCCJw8snwj3/Atm0NXx8/KxKpFCtXxj9XFa65Bjp0gJ/9LLkexZ4sCN2727ZzGzUMZSkI2fTqw84JC0QuXdowdQ+qV1RhC6tTWFaKNxZh/fr672UzBsHPEUfArl0wc2Z258chkyAAnHSSHffyyw1fHz+ffmrPK7LIsZs6Ff72NxODLl2cIDhypywFIV2vPqyBDTsndQ4gj759813r+sQVtrA6eftT7/3DD21/kNtowQJrgMLiD5k47DB7LoTbKJPLCMyP36oVPPdcw9fHT7YWws6dZh0ccABccYXta9PGnp0gOLKlLAThqqvgwguT22E9ZX+vP7WBDTuntrZ+z7N1a3PD5DPYnA931c03169rRUWyrqn3fv/9dkyQIHz0kfVK+/XL7v46d7YAYSZByMdnGMVCaNUKxowxQSjkAiyehbByZbzr/v73MH8+3H47NGtm+yoq7D72tLTTmhqLGThBKACq2mgehxxyiMZl0iTV1q1VQXWffWx7n31sO/VRURG8f599ws/xlymS3PZf13u0bm37M9U3allB9QE7N0r5TZqojhlj+8PuD1Sfeqp+OT161P+8otyfn4svVu3aVXXXrvC6ZvMZpvLII3buRx+lP27iRDvuww/jlZ8L3/lO8t7WrYt2zvr1qp06qR57bP3PrmtX1csvz389i8mKFfb5/Pd/23br1qrXXFPcOjU2gJkaoY0teiMf5xFXEMIalMsvj9/ATpqk2qJF9MYproCkq2+XLvEFLOwaHjt32ns33mjbIuH3/4c/1D13x47wY/fZJ/r3c++9ds6CBfE/wzjcd5+dt2xZ+uMWL7bj7rgjXvkemT7zIE45JXlf774b7To//7ldY86c+u9VVqpecEGcWpc+77xjn8///q9t74n32NBEFYQ92mUU5lKZOhXuu89mufRmu/S2g+jbF8aNg29/O7nPO2fcuOBzsnFLhdU3KKgL5q5q2bLuvpYtLUCaKbawdq2958UA0sU8Ul1G6QZwLV0a3c3jDVALcxuFfYZxA/ZRXEZgdR040H4fmfjlL+Hcc5Pb2aYfr1hhYzK811F47z2LHXhLkvpp23bPiyGkDqDs3t25jBqMKKpRKo+4FkJYrzfMpZLJRXHLLbbvhz/MfO1s3FLpeulh59x8c919p54arWed2usKu/e0AYa2AAAgAElEQVSWLVW/9rW6Pd9rrgmvU5cu0d08NTWq7dqFuziysRCCeun/+Z923tatmb+3H/9YtVkz1S++CO/xr1mj2qqVaps2SZdNttZMjx6qo0fbsQ89lLl+qqqHHWbuoiAOP1z1q1+NVk5j4cEH7fNZtMi2TzlFddiwolap0YGzEDJn1aQyblyw5eBZAV7POErQLiiA27q19eqDWLo0vF5duoQHrr3BSP/8J5x2GrzxRrRU2NReV9i9t28Pf/973Z7v3XfbOalZO14dowa6KyrSD1AL+wzDxk2E9dLffNPuacCAzFbLSSdZBs+NN4b3+O+5xzKXvvwSvkisBJKNNbNzp/V0q6ttO6qFsGxZ+ISCe7KF4A1KcxZCAxJFNUrlka8YQtygpMfxx1sZ550X/fqpPcxMsYWw+ob1Vv/0J93th5861V537Zq5t/rAA3V7XWE0bx5cVpMmFqxNrVM6qyzoHm64wcravDn6ZxhG2GfbsmX9ff7fgf8affta779t2+Cy9t7bArrt29v2e++lv3Y6C2HpUjvm3ntVO3RQ/d730n8Xqha/EbHPLYgzzlCtqspcTup9R415FIOrrrLP22PCBLPiwpIRSoFS+2xxQWVj0iTVXr2SDWUuX0z//lbO6adnX0YmkYr7Q7rtNitj40bV2loLuB10UGYh9NwoX36Zvvx0LqsgwhrGMFfSj35kr6dPj/b5eQR9Ttm43IK+jzC3nv/xm9/Y8wsvJOsTt/Px2mt2XLdu0Y5XVf3kk6SIBH0eoNq0abRstnx1lhq68TvnHNUDDkhu33GH1ffzz/N7nXyR745oPnCC4GP9ervTO+/M6nRVtcbW6y0fd1z25ajm9w/04x9bD9jrLf3Xf1kdb701/TW+//26va4w2rQJbhDbtAm/tziZUn362PN//mf0e457jbCH99nEOaeiQnXkyGTDfP/9desV9pkHvXfVVfXLz2QR/utfdtzUqZk/j2ysqSgZalG+i3w2fiNHqh5zTN1rQmHTg+OQr+y4fOIEwUdtrf2Rr7suq9NV1VIWvS/2iCOyLyffXHihuTk8Vq82c/qqq9Kfd/bZdXtdYZx4YvCP+xvfCD8nTu9dxOoxbFjurqEwKyTM7ZUpkN+sWd1tr5ypU5Oum5/9LPNnGNZohqU6pwvMT5li23PnZv480jVA6e47jtsy28YvjugMGKD6zW8mt196ya7xyivZXSNuhyzu8XGTWQqBE4QU9tpL9dJLsz5dX3kl+eeI6qMtBCecoFpdXXffueeqduyY3h10zDF1e11heBbH3nvbD7pnT9t+9NF49UzXcHzlK8GNUtgfL26cYtiw+ud45YfVq3lzcxH6Ywu9etl371ljPXuqXnJJ9vce97HPPkkXod9dkqkBitOQh7nLwkQqrK6ZBkfGsSo6dLD4nXcP3m8waMBkpmuEjUEKu3a+ra9i4QQhhaoq1dNOy/p0ffhh+7RGjFDdb7/sy8k3w4apnnxy3X0zZlhdH3ww/LwDDjArIRP3329lLV1q2//7v7Y9c2a8eqb7Y3XuHO8PFPcPd+yxdRv3KIMBv/51C3Z7o4f//GfdLYT+BrVly+x7jHEfIpYC7JWXS6JC3MGZYY90adRhxHFXeckPqdYa2Eh37zvMVfCy/a3FnaWgoa2TMPIqCMAJwHxgITAh4P2rsVXR5gL/APZJ7B8NzPY9tgFnJN57CFjse29opnrkIgjHHWc52tnijQ696CLLHS8VevdW/da36u7btcsCyyNGhJ/Xvr3FETLxzDP2K3nnHdv+9a9te+PG+HUN+3HH7WXG7bUdcUT6uE9QvV5/PSkAu3bZb6dfP+sY5KvH2KyZiY5/X8uW4bGQLl2CpwsJatxbtYomFlEb0nSPuEkScdxVQRli3qNDh/DfQtx7CPutZbJG4zT8+Tw+LnkTBKAC+BjYF2gOzAEGphwzGmideH058HhAOZ2Bz3zHPQScFaWS3iMXQTjvvNx69hddZI3vD39oKYmlwK5d1qhMmFD/vd/+1r7dt9+u/96XX9p7UQK5r75qx770km1fcolq9+651TuVvn3DG60w4vScDj44vnVYW2vZP+PGqb78stVn4sTs3AFhf+xevVQPPbRumRMmxA+aBzXmt99u187X4Mwo126oqVjSPeJaAvm0EOL+FuLGvtJ95nHJpyAcAbzo2/4J8JM0xw8D/hWwfzww2bddUEG46iobFZstxxyjevTRlv8tUho50J99Zt9g0Nw7n39uvcRx41S3bav73qJFdt4DD2S+xgcf2LFezMD7HPLJpEn1e8r5zFQ54ABLXYzLBRfYn/JrXzNx2LIl+4BhUKPZrl3SSvvoIyvn4YfDj49ybc+lN2uWbWcrYPnorcZtAOOKQevW8YPj+YwhxP0t5NN1GJd8CsJZwB992xcAv0tz/O+AGwL2TwNO8W0/lHBDzQXuBFpkqksuguBN8bBlS3bn7723NRBekDXbcvLJhx9aXSZPDn7/iivs/TZtzCf+wAOWhfTvf9v+557LfI21a+3Y3/7Wtvfaq76LKh+cfnrdxiqfaYt9+mRXZy+jB1R/9Svbl6+A4Rdf2Hm//rVtb9pk27fcEn5OFEvKy8B59VXbLuZ4g7jB/3TuqlTXUZMmqgcemF36bL78+PmyEOI+im0hfDNAEO4JOfZ84PXUxh3oCawFmqXsE6AF8DBwY0iZ44GZwMy+/vzKmHjB0U8+iX/u9u3JmUHvucfKWbs266rkDc+V8be/Bb+/Y4c1+pddZu4u789YWWmvg9xJqdTU2J/vpz9NNmJxxgxE5f/+z8qeMSP/ZXfurPrd78Y/77PPzMXQtq29Vg1uYFu0iN/AemL+yCPJfe3apU8Xvuuu+o1DauPuif3zzyf35XPcSxziNphBn23Tplbvhx6qew8jRpggFFPw4l47rgstzrxgmSi4ywg4DvgA6B7w3lXAfWmuMQr4a6a65GIh/OUvdrdvvRX/3AUL7NwHH0xmPSxZknVV8saTT1pd/DnpYezaZYHhX/zC0lR79lTdsCHadbw59r0J8dKl+2XL+vWW6nnFFfkvu1UrG8CXDT/4QdI68kjtzX772/HLnTbNzp02LbkvNd8+FS+e0717eKM1d64d8+ST8euUb7JprFMb5TFjzF2Xyne/a0IfdE62YpBN45sPayOfWUlh5FMQmgKLgH6+oPKglGOGJQLP/UPKeB0YnbKvZ+JZgLuAWzLVJRdB8KYJ8I/wjIpnhs+Yofr44/bam8MmE59+qvr++/GvGYXf/c7qsmpVw5TvceCB1lB5LpSgefjzwX/8hzXeUReKiUJtrdXZW/chn2zfbmX/4hfxz/UW7fGPth01Kn185tFHM//2vPhQ1JlTG5pcG7TTT1cdMqT+/l/8wu5zx4581LL4Ywca2oqLKggZZztV1RrgSuDFhAXwhKrOE5GbROS0xGG3AW2BJ0Vktog8650vIpXA3kDq8uWTReRd4F2gK/CrTHXJhW7d7DmbWRK9WU779UuuWxt1mcKf/AS+/vX414zCmjU2i2eXLg1TvkfXrrYmwoIFtr3//g1znauvtllEf//7/JW5bZs9Z1oLIRuaN7cZOJcvj3+uN7Npr17Jfb16pZ/x1LtO2EynYLOdQunMeDpuHCxZArt22XPY+iFheGspp+ItpRm0vGs25GvtjWzJ9XPKF02jHKSqU4GpKftu9L0+Ls25S4DeAfvHRK5lHshlLdZFi2zd2t697TVEF4S1a+0LVrXGO5+sWWONddNI32L2dO0KH39s6yj36dMwjSvAoEE2/fQ998CPflR/8Z9s2LrVnhuqzn36ZCcIn34K7drZw6Nnz+TaykG/lWXL7HhvQZ0gvA5LqQhCrqxaZYsBpeL/P/fsmft1+vYNnjY+3cJReyJ79HoIftq2tQYmWwuhb1+bvz+uhbBpE2zfDp9/Hv+6mVi9OvnHaEi6dTNhW7AA+vdv2Gtdc419R5Mm5ac8b22G1LUb8kW2grBihXUw/PTsaQK2cWPwOcuXp7cOwO5TZM8QBNXMFkK+1kWIu/bGnkrZCIKI/YjWro1/7uLF5i6CpEkeVRC8P2bUxU/isGZNctGQhsRzGX30UXBvLZ+MHg3DhsFvfmPmc65EXT4zW/be23rucfn007ruIkhur1wZfM6yZSZA6RDZcxbJ2bjROlOFEIRMi2OVC2UjCGA93WwtBE8QsrEQwBqAfFMoC6FrV6ipgc8+a3gLQcTcRR9+CM8/n3t5DS0IffrAhg3xG+BPPw22ECC885BupTQ/e4ogpK7q5yffggCl48cvJmUlCNksvbd5s/WOUwUh6h/OE4TGbiF4NLSFAPDNb1pDe/vtuZflxRAa0mUE8QR/1y6zAlIFIZ2FsGOHdQDKSRBWr7bnIEHo0MHiem4pzfziBCED/gwjyN5CyLcgbNtm6/kWykLwaGgLAeyP/oMfwIwZ8PbbuZVVCAsB4sUR1q2z9ZRTXUaehRAkCJ7gZHIZgQlC1N9nKZO6lrIfzwXsBCG/lKUg2PCHaKQKQosWtlB7lD9cbW2yh5pvQfD+CIWwELyU3SZNYN99G/56AN/+tmXT/OY3uZVTioLg/RZSLYR27awxD/qteHGKKBZCmzZ7hoWQzmUEThAagrIThO3b4/1ZUgVBxP5wUQTBf518C4JnThfSQqistNz7QtC+PYwfD088EZwOGJWGdhl5jXqcwLLX20+1ECCZeppKlDEIHvlwGdXWWtyomKxaZdZip07B7ztByD9lJQjZDE5bvNh6l965EN0k99xF0HAWQiEFoRDuIj/f/74J8G9/m30ZDW0htGxpv418WAhggpDOQojqMspVEC64AE45JbcycmXVKrOAm4S0Uo1VEN55B+bOLXYtgikrQcgmM8HLMPIPFIpqIXiC0KpV/rOMCukyatvWHgcd1PDX8rP33nDOOXD//ZbJkw0NLQgQfyzCp5/a7ynIFdKrV7iF0KFD3YFsYeRDEKZPhxdfTFrIxSBsDIJH9+5mKcdxAZcCl10GF15Y7FoE4wQhA/6UU4+oPlrvmAMOsB93bW3062aikC4jEfjHP+C66xr+Wqlcc419jv/zP9md39AuI4gvCCtW2PfWrFn99zwLIbWRizIGwSNXQVi9Oum/f/TR7MvJlSiCsHVr4wugL1kCc+aUpnVTloIQdXCaqglCaiA1roUwYICJQTaD4sJYs8b++A3Z8/UzYkRdt1mhGDbMBPX117M7v6FHKkN2FkKQuwjMQtiypa67EaKPQYDcs4zmzLHn9u1h8uTi9cCjCAKUZsMaxrZtyXZgxoyiViWQshKEuDGE9eutpxVkIcQRBC93P59xhEINSisF9t3XelXZsGWLBcIbcr6nvfe234onPpkIGqXsEZZ6GmXaCo82bWzcwo4d0Y5PxROEa6+FDz6A2bOzKycXamvtf5pOEDx3aWMSBH/H4R//KF49wigrQWjZ0nywUX9AqRlGHtlYCJBfQSjUoLRSYJ99sheErVsb1jqA+IPTguYx8vCEwv9b2b7dvu84LiPI3kqYPdvE5zvfMSGdPDm7cnJh3TobwLenWQheckCnTk4QSoI4mQlhghA3y+jAA+3ZWQjZUVlpDUQ2fvEtWxrerRZnLML27eYyiGMhxEk5hdynwJ4zBw4+2KZVP/FEeOyx/Ma/opBpDAI0bkE47zybQTiXlOqGwAlCGnK1ELw/5P77W2A2n5lG5WQhVFbaczZ/nlITBK+hi2MhFFIQtm2zeaSGDrXtceOsLi+nrmbSwEQRhFzWOCkWniBcdJE9l5qVUJaCEDW4u3ix9ZJSU/2iZhlt2mRC0L69Nd75shB27bJ7KCcLAbJzGxXCZeQ17lEEId2gNLDfWuvWdS2EOGMQIDdBmDfPrIGDD7btU0+18grtNooiCC1b2n+rsQlC165QXW3/XycIRSbOjKdBKadggrBlS+bsi02b7NgmTTKvhhWH9etNFJwgZKYQFkLr1tZxiDJaOd2gNLAOROrgNE9oChFD8ALKniC0bg1nnglPPZVcfa4QpJvHyE9jG5zmZYuJwJgxMG1aaY2jiCQIInKCiMwXkYUiMiHg/atF5H0RmSsi/xCRfXzv1SaW1UxdWrOfiLwhIgtE5HERKcikCJ6FEGWu/XSCoJrMcQ9j06akdZFPQSjkoLRSoEcP6w2WqssIoqeeZrIQvPdSLYSOHZMNfSZysRBmz7bf9377JfeNG2cTKT73XPzysmXVquSAyHQ0VkEAOPZYu88PPihunfxkFAQRqQAmAicCA4GxIjIw5bBZQLWqVgFPAbf63tuqqkMTj9N8+38N3Kmq/YHPgUtyuI/IdO9uJnGmFcxqa61HGiQIUXtgmzc3jCAUclBaKeAtWFKqLiOILggrVtgEienWwU61EOKMQYDcltGcMweqqupOFzFmjIlyId1GmcYgeDRmQRiTWES4lNxGUSyEEcBCVV2kqjuAKcDp/gNUdbqqelnYrwNpjVsREWAMJh4ADwNnxKl4tkQdnLZihU1RHGYhQGZBSLUQ1qyxMnOl3CwEyF4QStFC6NUr/fraqRZCnDEIkL2FoGqC4AWUPZo2hbFjzUJoiKVgg9gTBWHzZpuCxfsu993X3KGNTRB6A37v6PLEvjAuAfxrXbUUkZki8rqIeI1+F2CDqnrzKYaWKSLjE+fPXJuHob5RU9XCMowgniB4f07PZxy2PGIcys1CAPvjlLogrF2b2c+eblCaR8+e1nh4actxpq2A7AXhk09s2UovfuBn3Dgb6Pb00/HKzBZvYrtMxHEBF5ugKcyPPdZGLBd7ZlmPKIIQ1JcJDIOIyPlANXCbb3dfVa0GzgPuEpH94pSpqveparWqVnfLw9wJUVPVoghCpj9cqoUA+XEbrVljvbawaYH3RCor7Y8fN1BaKJeR9yfPlFqcblCah3/ltG3bbAxGIVxGqQFlP4ccYiPuC+U2imMh7Nply7sWi1deiTZOI0wQNm6EWbMapm5xiSIIywH/z7EPUK9ZE5HjgOuB01R1u7dfVVcknhcBM4BhwDqgo4h4EwoEltkQxLEQPN91Ktm6jCB/gtCtW/i0wHsi2Y5FKKSFAOndRqrRLQQwQYg7BgFsqo7mzeMLwuzZ9psfMqT+eyJmJcyYEW/th2zYts1cK1EFAYrnNvrXv2DkSHj22czHBglCqcURojQpbwH9E1lBzYFzgTq3LyLDgHsxMVjj299JRFokXncFjgLeV1UFpgNnJQ69CPhLrjcTBW9u/yiC0KuXBQBTiSoIqUFlyI8grF5dXvEDyD71tJQEYdMm+81EtRBWrIg/BsEjmwnu5syxNS+833cq551nz489Fq/cuKRbSzmVYgvC3/9uz++9l/nYZctMWP3ff48eMHhw6QhCxim/VLVGRK4EXgQqgAdUdZ6I3ATMVNVnMRdRW+BJixezNJFRdBBwr4jswsTnFlV9P1H0tcAUEfkVlqWU5QTH8Wja1DI8MoUjwlJOIXqWkT+G0LWrXTtfFkI5xQ8gO0HYudNM+UK4jKIMTouScgp1LQTPtxzHQoDsltGcPRsOPTT8/f33h8MOg//+b5u6e8gQe+S7c9KYBGH6dHuePz/zscuW2T2lTns+Zgzcd59ZRi1b5r+OcYg0B6SqTgWmpuy70ff6uJDz/g0EGKC7XUgjItc0j0TJTFi8GEaPDn4vioVQW2u9U89CaNIkfDWsuKxeXfjVy4pNjx5mrcVxGRVicRyPtm1trEA6d0qmQWkeHTpYw7BiRTJInY2FEEcQNm603/yll6Y/7mc/g299C66+OrmvWzcThmHDLBvpkEPi1TWVKKOUPYopCNu2JadljyoIQcJ+7LFw993w2mvhbU6hKCMvdJJMo5W3b7feXJiFEEUQvD+jf9qL3r3zM59ROc1j5NGkSfzU00IKAtifPYqFkEkQRJKpp8uWQefO8e8hriB4SzoGBZT9nHiiNdirV5u75M47bXqLTZvgd7+zKRkOOQT+8AcbzJYNcQShc2f7bRRDEF57zdqKffc1Qcg04jhMEEaOtHuYNq1h6hmHshSETBbC0qX25WYShHR/uCBByMfgtM2braErN5cRxBeEQqyW5ifTWISoLiNIWpNxxyB4xBUEL8ModQxCGN27W8/2Bz+w1ezefNMa8okTzTq+/HK7h0sugTfeiFd3TxCi/MYrKswdG1cQrr8ezj4bXngh+5lcZ8ywhvySS0wQvXoHoRouCB06mKuuFOIIZSsI6WIICxfac+pKaR7Nm9sPMZ2F4OWQ51sQynFQmkfcsQiFthAyCcKKFeZWilIfv4UQ110E2QlCly7RxCqMjh3hiisshfLNNy0I/fjjcPjhNhdSVFatsro0jziZjbe2clRUzZp58kmzePr1g5//PH4G24wZMHx4Mu6Szm20YYO1F2Hifuyx9plla1Xli7IVhPXrwweDvPqqBYCHDw9+XyTzFNieIPjnYunVy34YUVfWCqIcB6V5VFaaIEb9/IohCKtXh69UFiXl1KNnz2TaabYWQpwso9mzzTpIN4I6KiLWSN5/f1IgP/ww+vlRB6V59OgRz0JYutQa3rvvNlEYOBBuusmE4Wtfi7a05datFj8YNSq5AFY6QQhKOfUzZoxZKq+8Ev0+GoKyFARvcNq6dcHvT5tmawinm1gr0x8uzEKA3EYrl7uFANF7csVwGUF4nCjKoDSPXr2s0Vq/PjtBiJNlVFNjaZOZ4gfZ0LGjPeL04KMOSvOIO32FFy+proazzjK30ZIlFjCfNw9OOSXzZ/faayb8o0fb996qVXrRyyQIRx5pSRPFdhuVpSCky0z44guYOTM5YCSMTBZCWAwBcnMbeXUuVwsBoruNihFUhnC3UVwLwaOhXUYffWQZMw0hCGCNexxBiDvOJltBGDw4ua9vXxOExx+3/3WmKTqmTze38dFHWxzhgANysxBatYKjjoIXXyz86nR+nCCk8Oqr9oVkSv+K6jJKzTKC3DKNyt1lBKUrCOkGp+3aZZZhVAvBLwi5BJWjzLUfN6Aclx494gtCXAvhiy+ir9cwd67FB1MXvgLrqe+/Pzz8cPoyZsywbKr27W17wIDMgtC0afr7uvBCmwr78suLt0ZCWQtCUGB5+nQLZh1xRPoyMpnkYTEEyN1C6NgxesBtT2Kvvey+o7qMPEEotMsoSBDWrLGORlQLwX9cthZClDU7wOIHzZol1/7ON3EEYfNm62jFtRAg+kqIc+faFN9BiFjDPH16+O9syxbLnBo1KrlvwADrqGzfHnzOsmX2nVZUhNfroovgJz+x2Mt110W5k/xT1oIQZCFMm2a9hEyNSDYWQocOVm4ugrB6dXlaBxB/LILXGBbKQmjXznqMqYJQUwN33GGvg+bGCiIfLiOI5jaaMwcGDWq4TkYcQfCOy0YQoriNtm41F1mYIABccIE9P/JI8Pv//reNgvd7EQYMMCvQy1BMJeqaFjffDN/5DtxyC9x+e+bj801ZCkLHjqbUqT+gzz6z3lKU0YJRBcE/L4w34CidIKjC88+H+xHLcVCanziCUGiXEdif3j9aeckSOOYYuO02G+H7ta9FK6dTJwsydu2anYUTZxnNOXMaLn4A9nvdsCGaSyfOtBUenshGyWR6/31ruNMJQmWl9f7/9Kdg140XPzjqqOQ+z7oKcxtFFQQRG8tx9tnw4x/DAw9kPieflKUgNGkSPFr55ZftB5ApoAyZs4w2b7ZjUmckzSQIL7wAJ51kc5sEUc4WAsQbi1AMQfCPRXj8cfPLz5tnE8I98ED9eWzC8NZWzsY6gOhTYK9ebVk9DS0IEK0HH3UtZT+DBtl/7bXXMh/rBZTTCQKY+2bBguAyZ8ywtFq/9X/AAfYcJAiq8dKHKyrMOvna1+Db34b//d9o5+WDshQECB6cNn26NR4jIsywFMVCCApaRREEMBdDkJVQ7hZCZaU1YlF841u3WsNayHhLnz4mWJdeCueeCwcdZFbnuefGL+voo+ErX8muHlFdRg0dUIbk7zWK2ygbl1HTpjb47d//znzs3Ln2Hw8bdOrxjW/YcQ89VHf/5s02gMwfPwD7r/fqFSwIa9dabCHuFOZPP22TCY4dW7h01LIWhNQey/Tp9ieM0oBEEYSgcQzefEZhWQQvvWQurYUL4S8pE4LX1FheerlbCBAtsOxNfZ2PwVZR8VZOe+ABCwy+8kr4FCiZeOQRGzyVDVEFYfZsey6EhRBVEESSY4WicuSRJm6Z7nfuXEs3TRfcBWvgv/ENs/L8nY9//9v+h6mCAOGZRplSTsNo0wb++lezPk4/PSneDYkThARr1tjgnKizDXqCELZ0XzoLYcuW4CHqy5aZH/S666wRSQ0qeRZNuVsIEM1tVKi1EPwcf7wNePr73y1AGNVFlG/iWAh7722TxDUUcQTBm7Yi7ud25JH2X3zzzfBjvDWjM7mLPC66yP6n/o7Z9OlmkfjjBx6eIKR29rIVBLDv5aWX4PzzCzPDcdkKQmoMwRuuHkcQINx14V8cx0+61NO//c2eTzrJphd+7TVbkcmjnAelecQRhEItn+nnyCPhrbeixaEakjgWQkNaBxDfQsimw3PYYWZZpHMbrVplFnZUQRg92hpx/5gEL34QZP0PGACff17fFZ2LIIDFkv7wh8J0bspWELp3t168l/kwbZo14FHncs80BXY6CwGCBeGll+z9gQMtI6Vz57pWQjb+1T2Nnj2t91iqFkKpECXLSNUCpwcd1LB1adnS0nGjCkKcDCOPjh0tuJxOEKIGlD2aNLEU1Jdesv/rpk0m9mGdxrA5jZYts4yxPCwJ3+BEEgQROUFE5ovIQhGZEPD+1SLyvojMFZF/iMg+if1DReQ1EZmXeO8c3zkPichiEZmdeDRgWKs+qYNZpk+39MCmkZYMyvyHC4shhAlCba1ZCF/9anLyvCuuMHP1o4/sGGchJMcixIkhlCNRLITPPrN8+lxmOI1K1LEIcSe283PkkWZVh7lxPUEIWrlW3sYAABcFSURBVDM6jAsvtPImTzZrvbY2OH4A6QWhT5/CxrKyJaMgiEgFMBE4ERgIjBWRgSmHzQKqVbUKeAq4NbF/C3Chqg4CTgDuEpGOvvN+rKpDE4/ZOd5LLPyDWT791BrdOGZ+vi2EWbPsD3r88cl9V15pvWFvUFM5T1vhJ+pYhGK4jEoFTwjTCYI3yaJ/EFxDEVUQclkv/MgjbbxD2HiEuXMtqSNOvGTAAJu14OGHrdPYrJldJ4h99jFLIEgQsnUXFZooFsIIYKGqLlLVHcAU4HT/Aao6XVW9SYlfB/ok9n+kqgsSr1cAa4CSMJz8guCtixpn+bpMghAWQ2jTxkYsp85n9NJL9nycbzHSHj2sh/Lww1bPNWvsB+fNn1KuRB2LUM4WQpMmdu+NSRC8xZ+ycRlBsqH2x938pJuyIh0XXWRjSR580FLS/YNN/VRU2DxIe7og9Ab8K8UuT+wL4xLg+dSdIjICaA587Nt9c8KVdKeItAgqTETGi8hMEZm5NupkJRHw/HmeIHTqFC+4lm7gz65dJhRBggDBYxFeesnWpE3t/V9zjcU5Jk5MDkprDKZnQ1JZaa6FTGMRylkQIPOMp6UmCNkMSvOz//42sjsojrBjh00cl40gnHOOdcTWrs3caTzwwLqCUFtrnb89SRCCmp/ALHoROR+oBm5L2d8TeAT4lqp6Hr6fAAcChwKdgWuDylTV+1S1WlWru+UxKuOPIUybZn7B1FHF6UhnIQRNfe0nVRA2b7Yfsd9d5HHggbZm7cSJ1isu54Cyh5dptHRp+uPK2WUEpScIn30WvngQ5J40IWJWQpAgzJ9v8ZJsBKFjRxsHAOHxA48BA2DRIrsWmMjV1u5ZgrAc8N9OH6BejoyIHAdcD5ymqtt9+9sDzwE3qOrr3n5VXanGduBBzDVVMNq2teyHN96whjaOuwjSC0LQTKd+UgXh5ZftB/TVrwYf/+MfW7rcyy+7+AFETz11FkJmQWjbNv1CUPkiyvQV2cxjlMqRR1o8MHXxq3fftedsBAHg2mttoFrQ+AM/AwbYwLVFi2w715TTQhNFEN4C+otIPxFpDpwLPOs/QESGAfdiYrDGt7858AzwJ1V9MuWcnolnAc4A3svlRuIiYo3rc8/Zdty88XRZRkEznfrxBMEbwPLSS8kFMoI4+ujkdBrOQnCCEJVM823FXZksF6KMRcjVZQTJOELqHERz51pA2MsEisvw4bYudMuW6Y9LzTTyBKFv3+yuW2gyCoKq1gBXAi8CHwBPqOo8EblJRE5LHHYb0BZ4MpFC6gnG2cAxwMUB6aWTReRd4F2gK/Cr/N1WNLp1M7dC9+6W+x+HXF1GO3darx9MEEaODP+xicCPfmSvnYUQfSxCubuMMq3ZsXJlYdxFkBSedIKQ7bQVfqqr7beR6jaaO9f+4w09ctwTBC/TqbFZCJGy7lV1KjA1Zd+NvtfH1TvJ9k8CJoW8V+SxnMnGdfTo+IHadEHlKBYCWLBpyxb78Ywfn/56Z55pE6adfnr648qBigrrcaUbi7BrlwXjy91CSLd+98qVlshQCKJYCKtXW1A46ligIFq1st58kCAUYvR4x47WrvgthLZtLbOwMVC2I5WhriDEpXlz++FmE0PwllFcsSI5XUVQQNlPRYWtpJRpJbdyIdNYhEIvjlOKRIkhFMpCiOoyyodL9MgjbU4jL7C7fr11vrKNH8TFP8mdl3LaWDIDnSCQfc8hbMbTqBaCJwjedBWO6GQai+AJQjm7jNIJwubN9iiUILRubfXJZCHkSxC2bUvO5JprQDkuQYLQWChrQTjnHJgwwfKXsyFMEDLFEDx/6rJldaercESnstJ6uGGrcBVjcZxSI50gFDLl1CPTWIRs5zFKxQsse26juHMY5cqAAZbl9NlnThAaFYccAv/1X9k3xmFZHJksBG9pxOeeqz9dhSMamcYiOEGw3+e2bcELLZWaIKjmz2XUq5e5FP2C0K1b4TL0vMDye+/ZPTlBKBMyuYzChriD/WhnzrTXxwWG5B3pyJR66lxG6TPhvBTPQqWdQnpB2LzZvrN8NdpHHmlTWKgmp6wolBXuCcL06XZ9JwhlQlha36ZN9l66kc9eHCFougpHZjIJgrMQ0s94WmoWQj4Gpfk58kgLJC9ZYj31QrmLwBa3atbMFkkCJwhlQzoLIcxd5OFlGjl3UXb06mVZXk4QwskkCM2a2epkhaJHD8v4qamp/14+BqX58eIIjzxilkchBaFZM9hvP3g9MS+DE4QyIV1QOZMgeBZC2HQVjvRUVNgfLWwsgnMZZRaEvfYqbDJDjx7mQgmaozLfiz9VVVln4P77k9uFxJvCApwglA25WAhf/apZB0cf3TB1KwcqK5NzxqTiLITMglBIdxGkH4uQb5dR06a2rOby5ea6LXRatxdH6NQpfSyx1HCCkAPpsowyTRj2la/Aiy9axpEjOwYNshzzoCwaJwjp59sqpiB47iE/3rQVXbvm73qe22jAgMxzEOUbTxAak3UAThByIhcLwZE7I0bY5//BB/Xfcy6j9NOrlJqFsGpV7tNWpOIJQqHdReAEoSxp08Z6oqlruEaJIThy59BD7fmtt+q/5yyEcJfRjh0W3C0lQcjXoDQ/RxxhFrj3OykkThDKEK8H5jU+Hs5CKAwHHGDLib75Zv33vO+knC2EMEHIt78+Tn1atQoXhHwPHOvUyZa+vPLK/JYbha5d4ZJL4OtfL/y1cyGPBlr54R/4448ZRIkhOHKnSROb7jjIQti61SYgrKgofL1KhTBBKMYYBLAYQdhYhFWrMi8+kw377Zf/MqPyxz8W79rZ4iyEHAgaCZppPWVHfhkxAubMqT+nUbkvjgPmLqmoKB1BgGBBUG0Yl5EjPk4QciAoi8N77QShMBx6qOV7z5lTd78TBOuRB2XClZog5HvaCkf2RBIEETlBROaLyEIRmRDw/tUi8r6IzBWRf4jIPr73LhKRBYnHRb79h4jIu4ky704spdmoCLIQMk1s58gv3tKiqXGELVvKO37gETTj6cqVSfdNodlrr/qCkO9Ryo7sySgIIlIBTAROBAYCY0UkdZjHLKBaVauAp4BbE+d2Bn4GHAaMAH4mIp0S5/weGA/0TzxOyPluCkxQWp8ThMLSu7c1MqlxhK1bnYUAwfNtrVpls3/mM8UzKj162NTQ/rEjxQpyO+oTxUIYASxU1UWqugOYAtRZyFFVp6uql2vzOtAn8fprwN9U9TNV/Rz4G3CCiPQE2qvqa6qqwJ+AM/JwPwUlnYXggsqFQcSshCALwQlCuIVQDHcRmCDs2mWi4JHvaSsc2RNFEHoDy3zbyxP7wrgEeD7Dub0TrzOWKSLjRWSmiMxcGzQJShFxLqPS4NBDbYWqjRuT+5zLyAgThGL1xoPGIjiXUekQRRCCfPsaeKDI+UA1cFuGcyOXqar3qWq1qlZ369YtQnULR5AgZFotzZF/vDiCt74EOJeRRylaCFBXEFavthTiEvt7lyVRBGE54B9v1wdYkXqQiBwHXA+cpqrbM5y7nKRbKbTMUicoy8hZCIWnutqe/XEE5zIyUgVh1y5rgIstCP75jFavtoFc5TxmpFSIIghvAf1FpJ+INAfOBZ71HyAiw4B7MTFY43vrReB4EemUCCYfD7yoqiuBTSJyeCK76ELgL3m4n4LiYgilQefOti62P47gBMFITTtdt87SdIstCKkuI+cuKg0y5hmoao2IXIk17hXAA6o6T0RuAmaq6rOYi6gt8GQie3Spqp6mqp+JyC8xUQG4SVU/S7y+HHgIaIXFHJ6nkdGsmT1cllHxOfRQePXV5PbWrS6GAPWzjIo5BgFsqpEWLeq7jFyGUWkQKfFMVacCU1P23eh7HboqsKo+ADwQsH8mMDhyTUuU1BlPvT+fsxAKy4gR8NhjSf+4sxAMz2WkahlZnqumWIIQNH3F6tXQv39x6uOoixupnCOpghBlPWVH/vHPfKrqBMGjbVtzEe3YYdvFthCgriCoOpdRKeGarRwJEgTnLio8w4ZZUPLNN2HnTgueOpdR/QnuPEEopovGLwibNtk8VM5lVBo4QciR1KCdm+m0OLRuDYMHm4Xg1kJIEiQI7dsX97PxC4IblFZaOEHIEWchlA4jRpggeN+HE4T6qdHFHIPg0aMHrF1rVpwblFZaOEHIkdQsDrdaWvE49FD4/HN47z3bdi6j+vNtlYog1Nbaqm1uHqPSwglCjjgLoXTwRizPmGHPzkIIdhmVgiCAiYFzGZUWThByJEgQXAyhOAwaZFbByy/bthOEuoLgZfSUkiCsWmUZeV27FrdODsMJQo44C6F0aNoUhg9PTmHhXEZ1BWHTJgu4l5IgrF5tcxi5aStKAycIOZKaZeRiCMXFW0ENnIUAdQWhFMYgQDJesGqVCYJzF5UORVgiY8+iTRubJqG21kZhOkEoLl4cAZwgQN0so1IYgwDQsSM0b550GTlBKB2chZAjXhbHli1JS8HFEIqHN2IZnMsI6mYZlYqFIALduyddRsUWKEcSJwg54p/x1E1sV3z22w86JRZpdRaCxVVatCgtQYDk4DTnMiotnCDkiF8Q3OI4xUckaSU4QTC8Ce5WrjRx6Nix2DUyEViwwKatcIJQOjhByBFnIZQeRx0FLVs6QfDwBMFLOZWg9QoLTI8e8PHH9tq5jEoHJwg54g/aOUEoDX70I5vkrnnzYtekNPBbCKXgLoK6VoGzEEoHJwg5EmQhuKBycWndGoYMKXYtSgcvNdoJgiMTkQRBRE4QkfkislBEJgS8f4yIvCMiNSJylm//aBGZ7XtsE5EzEu89JCKLfe8Nzd9tFQ5/FoezEByliN9CKBX3jF8ESqVOjgjjEESkApgIfBVYDrwlIs+q6vu+w5YCFwM/8p+rqtOBoYlyOgMLgZd8h/xYVZ/K5QaKjd9C8NJOnSA4Sok2bWD5cpv4r9QshCZNoEuX4tbFkSTKwLQRwEJVXQQgIlOA04HdgqCqSxLv7UpTzlnA86q6JevaliAuqOwoddq2hcWL7XWpCYKbtqK0iOIy6g0s820vT+yLy7nAYyn7bhaRuSJyp4i0CDpJRMaLyEwRmbl27dosLtuwBAWVPZFwOEqBtm2TS2iWmiA4d1FpEUUQgpLUNM5FRKQnMAR40bf7J8CBwKFAZ+DaoHNV9T5VrVbV6m7dusW5bEFItRBat3Y9Hkdp4U9yKBVB6NzZ/icuoFxaRBGE5cDevu0+wIqY1zkbeEZVd3o7VHWlGtuBBzHXVKOjaVNLb/QGpjl3kaPUKEVBaNIE9tnHHo7SIUoM4S2gv4j0Az7FXD/nxbzOWMwi2I2I9FTVlSIiwBnAezHLLBm8VdPc1NeOUsQThCZNzGdfKjz/fHKaEUdpkNFCUNUa4ErM3fMB8ISqzhORm0TkNAAROVRElgPfBO4VkXne+SJSiVkYL6cUPVlE3gXeBboCv8r9doqDtyaCEwRHKeK5Nbt3Ly135gEHlJZAOSJOf62qU4GpKftu9L1+C3MlBZ27hIAgtKqOiVPRUsYvCG5QmqPU8H6TpeIucpQubqRyHvBGgroYgqMUcYLgiIoThDzgXEaOUsYJgiMqThDygBMERynjBMERFScIecCfZeRiCI5SwwmCIypOEPKAJwguhuAoRfbbD8aMgdGji10TR6kTKcvIkZ42bWDNGnvtBMFRarRpA//4R7Fr4WgMOAshD7RtC9u322snCA6Ho7HiBCEP+CezczEEh8PRWHGCkAf8guAsBIfD0VhxgpAHnCA4HI49AScIecAJgsPh2BNwgpAHnCA4HI49AScIecAfSHZBZYfD0VhxgpAHnIXgcDj2BJwg5AGXdupwOPYEnCDkAU8Q3HrKDoejMRNJEETkBBGZLyILRWRCwPvHiMg7IlIjImelvFcrIrMTj2d9+/uJyBsiskBEHheR5rnfTnHwBMFZBw6HozGTURBEpAKYCJwIDATGisjAlMOWAhcDjwYUsVVVhyYep/n2/xq4U1X7A58Dl2RR/5LAEwQXP3A4HI2ZKBbCCGChqi5S1R3AFOB0/wGqukRV5wK7olxURAQYAzyV2PUwcEbkWpcYThAcDseeQBRB6A0s820vJ2CN5DS0FJGZIvK6iHiNfhdgg6rWZCpTRMYnzp+5du3aGJctHE2bQosWThAcDkfjJsr01xKwT2Nco6+qrhCRfYFpIvIu8EXUMlX1PuA+gOrq6jjXLSht2jhBcDgcjZsoFsJyYG/fdh9gRdQLqOqKxPMiYAYwDFgHdBQRT5BilVmKtGnjgsoOh6NxE0UQ3gL6J7KCmgPnAs9mOAcAEekkIi0Sr7sCRwHvq6oC0wEvI+ki4C9xK19K3HQTXH55sWvhcDgc2ZNREBJ+/iuBF4EPgCdUdZ6I3CQipwGIyKEishz4JnCviMxLnH4QMFNE5mACcIuqvp9471rgahFZiMUU/iefN1ZoLr4YRo0qdi0cDocje8Q6642D6upqnTlzZrGr4XA4HI0KEXlbVaszHedGKjscDocDcILgcDgcjgROEBwOh8MBOEFwOBwORwInCA6Hw+EAnCA4HA6HI4ETBIfD4XAAjWwcgoisBT7J8vSu2JQZ5Ya77/KiXO8byvfeo9z3PqraLVNBjUoQckFEZkYZmLGn4e67vCjX+4byvfd83rdzGTkcDocDcILgcDgcjgTlJAj3FbsCRcLdd3lRrvcN5XvvebvvsokhOBwOhyM95WQhOBwOh+P/t3c/IVqVURzHvz9GI6nAlIxwDAtc6KKmjQi2sCFiSskWBUWBi6BNC4MkrI0ouHBTbdqV5KJ/klrSqsGMWlmZhoZFFlIy4rsoqTaK+XPxnKGXYfqzuHduPfd8YLj3OXMZzmGeec+9z33fuX8jG0JKKSWgJw1B0oSkbyWdlrS163zaImm3pIGkk0OxRZImJX0X2xu7zLENkpZJOizplKSvJW2OeNW1S7pW0meSvoq6t0f8NklHou534kmH1ZE0IumYpA9iXH3dks5IOiHpuKQvItbYPK++IUgaAV4B7gdWAY9JWtVtVq15HZiYEdsKHLK9AjgU49pcBp61vRJYAzwdv+Paa78IjNu+ExgDJiStAXYBL0XdvwBPdphjmzZTnuI4rS9132N7bOizB43N8+obArAaOG37B9uXgLeBjR3n1ArbnwA/zwhvBPbE/h7goTlNag7YPmf7y9j/jfIisZTKa3fxewznx5eBceDdiFdXN4CkUWA98GqMRQ/q/guNzfM+NISlwE9D47MR64ubbZ+D8sIJLOk4n1ZJWg7cBRyhB7XHsslxYABMAt8DF+JZ6FDvfH8ZeA64EuPF9KNuAx9KOirpqYg1Ns/nNZDgf51mieV7bSsk6XpgH/CM7V/LSWPdbP8BjElaCBwAVs522Nxm1S5JG4CB7aOS1k2HZzm0qrrDWttTkpYAk5K+afKH9+EK4SywbGg8Ckx1lEsXzku6BSC2g47zaYWk+ZRm8Ibt/RHuRe0Ati8AH1PuoSyUNH2yV+N8Xws8KOkMZQl4nHLFUHvd2J6K7YByArCaBud5HxrC58CKeAfCNcCjwMGOc5pLB4FNsb8JeL/DXFoR68evAadsvzj0raprl3RTXBkgaQFwL+X+yWHg4TisurptP2971PZyyt/zR7Yfp/K6JV0n6YbpfeA+4CQNzvNefFJZ0gOUM4gRYLftnR2n1ApJbwHrKP8O9zywDXgP2AvcCvwIPGJ75o3n/zVJdwOfAif4c035Bcp9hGprl3QH5SbiCOXkbq/tHZJup5w5LwKOAU/Yvthdpu2JJaMttjfUXnfUdyCG84A3be+UtJiG5nkvGkJKKaV/1oclo5RSSv9CNoSUUkpANoSUUkohG0JKKSUgG0JKKaWQDSGllBKQDSGllFK4ClpWRuJieeFJAAAAAElFTkSuQmCC\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",
    "\n",
    "accuracy = history3.history['accuracy']\n",
    "val_accuracy = history3.history['val_accuracy']\n",
    "loss = history3.history['loss']\n",
    "val_loss = history3.history['val_loss']\n",
    "epochs = range(len(accuracy))\n",
    "\n",
    "plt.plot(epochs, accuracy, 'bo', label='Training accuracy')\n",
    "plt.plot(epochs, val_accuracy, 'b', label='Validation accuracy')\n",
    "plt.title('Training and validation accuracy')\n",
    "plt.legend()\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",
    "plt.show()"
   ]
  },
  {
   "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": "3.7.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}