dog_breed_tf_lite.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "executionInfo": {
     "elapsed": 635,
     "status": "ok",
     "timestamp": 1607980917799,
     "user": {
      "displayName": "Mahmut Can Kurt",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GiR0Vz-Xi8c64gs5AldT12G3fhmeowAL59isdw9=s64",
      "userId": "17273613824855319138"
     },
     "user_tz": -180
    },
    "id": "EvQeZEtxbhfC"
   },
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "assert tf.__version__.startswith('2')\n",
    "\n",
    "import os\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 35
    },
    "executionInfo": {
     "elapsed": 930,
     "status": "ok",
     "timestamp": 1607980918100,
     "user": {
      "displayName": "Mahmut Can Kurt",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GiR0Vz-Xi8c64gs5AldT12G3fhmeowAL59isdw9=s64",
      "userId": "17273613824855319138"
     },
     "user_tz": -180
    },
    "id": "bbrALc_ybnCH",
    "outputId": "4919f31e-270a-4cdd-a195-b08a182222c5"
   },
   "outputs": [
    {
     "data": {
      "application/vnd.google.colaboratory.intrinsic+json": {
       "type": "string"
      },
      "text/plain": [
       "'2.3.0'"
      ]
     },
     "execution_count": 19,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tf.__version__"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 925,
     "status": "ok",
     "timestamp": 1607980918100,
     "user": {
      "displayName": "Mahmut Can Kurt",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GiR0Vz-Xi8c64gs5AldT12G3fhmeowAL59isdw9=s64",
      "userId": "17273613824855319138"
     },
     "user_tz": -180
    },
    "id": "xX_cHGsgbovo",
    "outputId": "2e49528c-cc07-4139-aac6-f65746918a99"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Drive already mounted at /content/gdrive; to attempt to forcibly remount, call drive.mount(\"/content/gdrive\", force_remount=True).\n"
     ]
    }
   ],
   "source": [
    "from google.colab import drive\n",
    "drive.mount('/content/gdrive')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "executionInfo": {
     "elapsed": 918,
     "status": "ok",
     "timestamp": 1607980918101,
     "user": {
      "displayName": "Mahmut Can Kurt",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GiR0Vz-Xi8c64gs5AldT12G3fhmeowAL59isdw9=s64",
      "userId": "17273613824855319138"
     },
     "user_tz": -180
    },
    "id": "VgGBgLtlbrk5"
   },
   "outputs": [],
   "source": [
    "image_dir = '/content/gdrive/My Drive/dataset/Images'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 1949,
     "status": "ok",
     "timestamp": 1607980919137,
     "user": {
      "displayName": "Mahmut Can Kurt",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GiR0Vz-Xi8c64gs5AldT12G3fhmeowAL59isdw9=s64",
      "userId": "17273613824855319138"
     },
     "user_tz": -180
    },
    "id": "mL-_qA6VbyAk",
    "outputId": "da5b514b-17c4-4afa-e122-1266d93c96ff"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 16516 images belonging to 120 classes.\n",
      "Found 4074 images belonging to 120 classes.\n"
     ]
    }
   ],
   "source": [
    "IMAGE_SIZE = 224\n",
    "BATCH_SIZE = 64\n",
    "\n",
    "datagen = tf.keras.preprocessing.image.ImageDataGenerator(\n",
    "    rescale=1./255, \n",
    "    validation_split=0.2)\n",
    "\n",
    "train_generator = datagen.flow_from_directory(\n",
    "    image_dir,\n",
    "    target_size=(IMAGE_SIZE, IMAGE_SIZE),\n",
    "    batch_size=BATCH_SIZE, \n",
    "    subset='training')\n",
    "\n",
    "val_generator = datagen.flow_from_directory(\n",
    "    image_dir,\n",
    "    target_size=(IMAGE_SIZE, IMAGE_SIZE),\n",
    "    batch_size=BATCH_SIZE, \n",
    "    subset='validation')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 28240,
     "status": "ok",
     "timestamp": 1607980945433,
     "user": {
      "displayName": "Mahmut Can Kurt",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GiR0Vz-Xi8c64gs5AldT12G3fhmeowAL59isdw9=s64",
      "userId": "17273613824855319138"
     },
     "user_tz": -180
    },
    "id": "4FFaOwvsdNUV",
    "outputId": "7175ef1b-c6a3-4854-f7bc-ccb8d5723b67"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "((64, 224, 224, 3), (64, 120))"
      ]
     },
     "execution_count": 23,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "for image_batch, label_batch in train_generator:\n",
    "  break\n",
    "image_batch.shape, label_batch.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 28237,
     "status": "ok",
     "timestamp": 1607980945437,
     "user": {
      "displayName": "Mahmut Can Kurt",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GiR0Vz-Xi8c64gs5AldT12G3fhmeowAL59isdw9=s64",
      "userId": "17273613824855319138"
     },
     "user_tz": -180
    },
    "id": "mzYrNpCUFmUs",
    "outputId": "29ff5526-1cfa-4e88-c112-e37166ed8045"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'n02085620-Chihuahua': 0, 'n02085782-Japanese_spaniel': 1, 'n02085936-Maltese_dog': 2, 'n02086079-Pekinese': 3, 'n02086240-Shih-Tzu': 4, 'n02086646-Blenheim_spaniel': 5, 'n02086910-papillon': 6, 'n02087046-toy_terrier': 7, 'n02087394-Rhodesian_ridgeback': 8, 'n02088094-Afghan_hound': 9, 'n02088238-basset': 10, 'n02088364-beagle': 11, 'n02088466-bloodhound': 12, 'n02088632-bluetick': 13, 'n02089078-black-and-tan_coonhound': 14, 'n02089867-Walker_hound': 15, 'n02089973-English_foxhound': 16, 'n02090379-redbone': 17, 'n02090622-borzoi': 18, 'n02090721-Irish_wolfhound': 19, 'n02091032-Italian_greyhound': 20, 'n02091134-whippet': 21, 'n02091244-Ibizan_hound': 22, 'n02091467-Norwegian_elkhound': 23, 'n02091635-otterhound': 24, 'n02091831-Saluki': 25, 'n02092002-Scottish_deerhound': 26, 'n02092339-Weimaraner': 27, 'n02093256-Staffordshire_bullterrier': 28, 'n02093428-American_Staffordshire_terrier': 29, 'n02093647-Bedlington_terrier': 30, 'n02093754-Border_terrier': 31, 'n02093859-Kerry_blue_terrier': 32, 'n02093991-Irish_terrier': 33, 'n02094114-Norfolk_terrier': 34, 'n02094258-Norwich_terrier': 35, 'n02094433-Yorkshire_terrier': 36, 'n02095314-wire-haired_fox_terrier': 37, 'n02095570-Lakeland_terrier': 38, 'n02095889-Sealyham_terrier': 39, 'n02096051-Airedale': 40, 'n02096177-cairn': 41, 'n02096294-Australian_terrier': 42, 'n02096437-Dandie_Dinmont': 43, 'n02096585-Boston_bull': 44, 'n02097047-miniature_schnauzer': 45, 'n02097130-giant_schnauzer': 46, 'n02097209-standard_schnauzer': 47, 'n02097298-Scotch_terrier': 48, 'n02097474-Tibetan_terrier': 49, 'n02097658-silky_terrier': 50, 'n02098105-soft-coated_wheaten_terrier': 51, 'n02098286-West_Highland_white_terrier': 52, 'n02098413-Lhasa': 53, 'n02099267-flat-coated_retriever': 54, 'n02099429-curly-coated_retriever': 55, 'n02099601-golden_retriever': 56, 'n02099712-Labrador_retriever': 57, 'n02099849-Chesapeake_Bay_retriever': 58, 'n02100236-German_short-haired_pointer': 59, 'n02100583-vizsla': 60, 'n02100735-English_setter': 61, 'n02100877-Irish_setter': 62, 'n02101006-Gordon_setter': 63, 'n02101388-Brittany_spaniel': 64, 'n02101556-clumber': 65, 'n02102040-English_springer': 66, 'n02102177-Welsh_springer_spaniel': 67, 'n02102318-cocker_spaniel': 68, 'n02102480-Sussex_spaniel': 69, 'n02102973-Irish_water_spaniel': 70, 'n02104029-kuvasz': 71, 'n02104365-schipperke': 72, 'n02105056-groenendael': 73, 'n02105162-malinois': 74, 'n02105251-briard': 75, 'n02105412-kelpie': 76, 'n02105505-komondor': 77, 'n02105641-Old_English_sheepdog': 78, 'n02105855-Shetland_sheepdog': 79, 'n02106030-collie': 80, 'n02106166-Border_collie': 81, 'n02106382-Bouvier_des_Flandres': 82, 'n02106550-Rottweiler': 83, 'n02106662-German_shepherd': 84, 'n02107142-Doberman': 85, 'n02107312-miniature_pinscher': 86, 'n02107574-Greater_Swiss_Mountain_dog': 87, 'n02107683-Bernese_mountain_dog': 88, 'n02107908-Appenzeller': 89, 'n02108000-EntleBucher': 90, 'n02108089-boxer': 91, 'n02108422-bull_mastiff': 92, 'n02108551-Tibetan_mastiff': 93, 'n02108915-French_bulldog': 94, 'n02109047-Great_Dane': 95, 'n02109525-Saint_Bernard': 96, 'n02109961-Eskimo_dog': 97, 'n02110063-malamute': 98, 'n02110185-Siberian_husky': 99, 'n02110627-affenpinscher': 100, 'n02110806-basenji': 101, 'n02110958-pug': 102, 'n02111129-Leonberg': 103, 'n02111277-Newfoundland': 104, 'n02111500-Great_Pyrenees': 105, 'n02111889-Samoyed': 106, 'n02112018-Pomeranian': 107, 'n02112137-chow': 108, 'n02112350-keeshond': 109, 'n02112706-Brabancon_griffon': 110, 'n02113023-Pembroke': 111, 'n02113186-Cardigan': 112, 'n02113624-toy_poodle': 113, 'n02113712-miniature_poodle': 114, 'n02113799-standard_poodle': 115, 'n02113978-Mexican_hairless': 116, 'n02115641-dingo': 117, 'n02115913-dhole': 118, 'n02116738-African_hunting_dog': 119}\n"
     ]
    }
   ],
   "source": [
    "print (train_generator.class_indices)\n",
    "\n",
    "labels = '\\n'.join(sorted(train_generator.class_indices.keys()))\n",
    "\n",
    "with open('labels.txt', 'w') as f:\n",
    "  f.write(labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 28233,
     "status": "ok",
     "timestamp": 1607980945438,
     "user": {
      "displayName": "Mahmut Can Kurt",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GiR0Vz-Xi8c64gs5AldT12G3fhmeowAL59isdw9=s64",
      "userId": "17273613824855319138"
     },
     "user_tz": -180
    },
    "id": "OYPcBr5_Fvw9",
    "outputId": "01aa3449-d01b-4115-bd57-379e936f03ef"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "n02085620-Chihuahua\n",
      "n02085782-Japanese_spaniel\n",
      "n02085936-Maltese_dog\n",
      "n02086079-Pekinese\n",
      "n02086240-Shih-Tzu\n",
      "n02086646-Blenheim_spaniel\n",
      "n02086910-papillon\n",
      "n02087046-toy_terrier\n",
      "n02087394-Rhodesian_ridgeback\n",
      "n02088094-Afghan_hound\n",
      "n02088238-basset\n",
      "n02088364-beagle\n",
      "n02088466-bloodhound\n",
      "n02088632-bluetick\n",
      "n02089078-black-and-tan_coonhound\n",
      "n02089867-Walker_hound\n",
      "n02089973-English_foxhound\n",
      "n02090379-redbone\n",
      "n02090622-borzoi\n",
      "n02090721-Irish_wolfhound\n",
      "n02091032-Italian_greyhound\n",
      "n02091134-whippet\n",
      "n02091244-Ibizan_hound\n",
      "n02091467-Norwegian_elkhound\n",
      "n02091635-otterhound\n",
      "n02091831-Saluki\n",
      "n02092002-Scottish_deerhound\n",
      "n02092339-Weimaraner\n",
      "n02093256-Staffordshire_bullterrier\n",
      "n02093428-American_Staffordshire_terrier\n",
      "n02093647-Bedlington_terrier\n",
      "n02093754-Border_terrier\n",
      "n02093859-Kerry_blue_terrier\n",
      "n02093991-Irish_terrier\n",
      "n02094114-Norfolk_terrier\n",
      "n02094258-Norwich_terrier\n",
      "n02094433-Yorkshire_terrier\n",
      "n02095314-wire-haired_fox_terrier\n",
      "n02095570-Lakeland_terrier\n",
      "n02095889-Sealyham_terrier\n",
      "n02096051-Airedale\n",
      "n02096177-cairn\n",
      "n02096294-Australian_terrier\n",
      "n02096437-Dandie_Dinmont\n",
      "n02096585-Boston_bull\n",
      "n02097047-miniature_schnauzer\n",
      "n02097130-giant_schnauzer\n",
      "n02097209-standard_schnauzer\n",
      "n02097298-Scotch_terrier\n",
      "n02097474-Tibetan_terrier\n",
      "n02097658-silky_terrier\n",
      "n02098105-soft-coated_wheaten_terrier\n",
      "n02098286-West_Highland_white_terrier\n",
      "n02098413-Lhasa\n",
      "n02099267-flat-coated_retriever\n",
      "n02099429-curly-coated_retriever\n",
      "n02099601-golden_retriever\n",
      "n02099712-Labrador_retriever\n",
      "n02099849-Chesapeake_Bay_retriever\n",
      "n02100236-German_short-haired_pointer\n",
      "n02100583-vizsla\n",
      "n02100735-English_setter\n",
      "n02100877-Irish_setter\n",
      "n02101006-Gordon_setter\n",
      "n02101388-Brittany_spaniel\n",
      "n02101556-clumber\n",
      "n02102040-English_springer\n",
      "n02102177-Welsh_springer_spaniel\n",
      "n02102318-cocker_spaniel\n",
      "n02102480-Sussex_spaniel\n",
      "n02102973-Irish_water_spaniel\n",
      "n02104029-kuvasz\n",
      "n02104365-schipperke\n",
      "n02105056-groenendael\n",
      "n02105162-malinois\n",
      "n02105251-briard\n",
      "n02105412-kelpie\n",
      "n02105505-komondor\n",
      "n02105641-Old_English_sheepdog\n",
      "n02105855-Shetland_sheepdog\n",
      "n02106030-collie\n",
      "n02106166-Border_collie\n",
      "n02106382-Bouvier_des_Flandres\n",
      "n02106550-Rottweiler\n",
      "n02106662-German_shepherd\n",
      "n02107142-Doberman\n",
      "n02107312-miniature_pinscher\n",
      "n02107574-Greater_Swiss_Mountain_dog\n",
      "n02107683-Bernese_mountain_dog\n",
      "n02107908-Appenzeller\n",
      "n02108000-EntleBucher\n",
      "n02108089-boxer\n",
      "n02108422-bull_mastiff\n",
      "n02108551-Tibetan_mastiff\n",
      "n02108915-French_bulldog\n",
      "n02109047-Great_Dane\n",
      "n02109525-Saint_Bernard\n",
      "n02109961-Eskimo_dog\n",
      "n02110063-malamute\n",
      "n02110185-Siberian_husky\n",
      "n02110627-affenpinscher\n",
      "n02110806-basenji\n",
      "n02110958-pug\n",
      "n02111129-Leonberg\n",
      "n02111277-Newfoundland\n",
      "n02111500-Great_Pyrenees\n",
      "n02111889-Samoyed\n",
      "n02112018-Pomeranian\n",
      "n02112137-chow\n",
      "n02112350-keeshond\n",
      "n02112706-Brabancon_griffon\n",
      "n02113023-Pembroke\n",
      "n02113186-Cardigan\n",
      "n02113624-toy_poodle\n",
      "n02113712-miniature_poodle\n",
      "n02113799-standard_poodle\n",
      "n02113978-Mexican_hairless\n",
      "n02115641-dingo\n",
      "n02115913-dhole\n",
      "n02116738-African_hunting_dog"
     ]
    }
   ],
   "source": [
    "!cat labels.txt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "executionInfo": {
     "elapsed": 29767,
     "status": "ok",
     "timestamp": 1607980946978,
     "user": {
      "displayName": "Mahmut Can Kurt",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GiR0Vz-Xi8c64gs5AldT12G3fhmeowAL59isdw9=s64",
      "userId": "17273613824855319138"
     },
     "user_tz": -180
    },
    "id": "BoGHYkdsFvzQ"
   },
   "outputs": [],
   "source": [
    "IMG_SHAPE = (IMAGE_SIZE, IMAGE_SIZE, 3)\n",
    "\n",
    "# Create the base model from the pre-trained model MobileNet V2\n",
    "base_model = tf.keras.applications.MobileNetV2(input_shape=IMG_SHAPE,\n",
    "                                              include_top=False, \n",
    "                                              weights='imagenet')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "executionInfo": {
     "elapsed": 29762,
     "status": "ok",
     "timestamp": 1607980946980,
     "user": {
      "displayName": "Mahmut Can Kurt",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GiR0Vz-Xi8c64gs5AldT12G3fhmeowAL59isdw9=s64",
      "userId": "17273613824855319138"
     },
     "user_tz": -180
    },
    "id": "fAI65cPwFv1h"
   },
   "outputs": [],
   "source": [
    "base_model.trainable = False"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "executionInfo": {
     "elapsed": 30030,
     "status": "ok",
     "timestamp": 1607980947253,
     "user": {
      "displayName": "Mahmut Can Kurt",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GiR0Vz-Xi8c64gs5AldT12G3fhmeowAL59isdw9=s64",
      "userId": "17273613824855319138"
     },
     "user_tz": -180
    },
    "id": "QGLZMeT5Fv32"
   },
   "outputs": [],
   "source": [
    "model = tf.keras.Sequential([\n",
    "  base_model,\n",
    "  tf.keras.layers.Conv2D(32, 3, activation='relu'),\n",
    "  tf.keras.layers.Dropout(0.2),\n",
    "  tf.keras.layers.GlobalAveragePooling2D(),\n",
    "  tf.keras.layers.Dense(120, activation='softmax')\n",
    "])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "executionInfo": {
     "elapsed": 30024,
     "status": "ok",
     "timestamp": 1607980947254,
     "user": {
      "displayName": "Mahmut Can Kurt",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GiR0Vz-Xi8c64gs5AldT12G3fhmeowAL59isdw9=s64",
      "userId": "17273613824855319138"
     },
     "user_tz": -180
    },
    "id": "ZRyBUIwtFv6h"
   },
   "outputs": [],
   "source": [
    "model.compile(optimizer=tf.keras.optimizers.Adam(), \n",
    "              loss='categorical_crossentropy', \n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 30018,
     "status": "ok",
     "timestamp": 1607980947254,
     "user": {
      "displayName": "Mahmut Can Kurt",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GiR0Vz-Xi8c64gs5AldT12G3fhmeowAL59isdw9=s64",
      "userId": "17273613824855319138"
     },
     "user_tz": -180
    },
    "id": "qQG7ulf_F8q6",
    "outputId": "a8eef9c6-22c0-4ab1-b876-737f83881eef"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "mobilenetv2_1.00_224 (Functi (None, 7, 7, 1280)        2257984   \n",
      "_________________________________________________________________\n",
      "conv2d_1 (Conv2D)            (None, 5, 5, 32)          368672    \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 5, 5, 32)          0         \n",
      "_________________________________________________________________\n",
      "global_average_pooling2d_1 ( (None, 32)                0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 120)               3960      \n",
      "=================================================================\n",
      "Total params: 2,630,616\n",
      "Trainable params: 372,632\n",
      "Non-trainable params: 2,257,984\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 30014,
     "status": "ok",
     "timestamp": 1607980947255,
     "user": {
      "displayName": "Mahmut Can Kurt",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GiR0Vz-Xi8c64gs5AldT12G3fhmeowAL59isdw9=s64",
      "userId": "17273613824855319138"
     },
     "user_tz": -180
    },
    "id": "4GV8h1M_F-dF",
    "outputId": "c34c0c99-cf6d-4e65-fe30-de016cc24bed"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of trainable variables = 4\n"
     ]
    }
   ],
   "source": [
    "print('Number of trainable variables = {}'.format(len(model.trainable_variables)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 10218911,
     "status": "ok",
     "timestamp": 1607991136161,
     "user": {
      "displayName": "Mahmut Can Kurt",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GiR0Vz-Xi8c64gs5AldT12G3fhmeowAL59isdw9=s64",
      "userId": "17273613824855319138"
     },
     "user_tz": -180
    },
    "id": "BsIExHewGAmg",
    "outputId": "39298f3a-f17b-45cd-a403-9e877ce8f0c9"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/25\n",
      "259/259 [==============================] - 7886s 30s/step - loss: 2.5469 - accuracy: 0.3896 - val_loss: 1.4975 - val_accuracy: 0.5955\n",
      "Epoch 2/25\n",
      "259/259 [==============================] - 95s 366ms/step - loss: 1.2183 - accuracy: 0.6529 - val_loss: 1.2843 - val_accuracy: 0.6414\n",
      "Epoch 3/25\n",
      "259/259 [==============================] - 95s 365ms/step - loss: 0.9456 - accuracy: 0.7197 - val_loss: 1.2935 - val_accuracy: 0.6483\n",
      "Epoch 4/25\n",
      "259/259 [==============================] - 95s 368ms/step - loss: 0.8166 - accuracy: 0.7462 - val_loss: 1.2359 - val_accuracy: 0.6654\n",
      "Epoch 5/25\n",
      "259/259 [==============================] - 95s 367ms/step - loss: 0.6903 - accuracy: 0.7796 - val_loss: 1.2470 - val_accuracy: 0.6723\n",
      "Epoch 6/25\n",
      "259/259 [==============================] - 94s 364ms/step - loss: 0.6006 - accuracy: 0.8074 - val_loss: 1.2420 - val_accuracy: 0.6728\n",
      "Epoch 7/25\n",
      "259/259 [==============================] - 95s 366ms/step - loss: 0.5511 - accuracy: 0.8208 - val_loss: 1.3049 - val_accuracy: 0.6699\n",
      "Epoch 8/25\n",
      "259/259 [==============================] - 95s 366ms/step - loss: 0.5030 - accuracy: 0.8331 - val_loss: 1.3151 - val_accuracy: 0.6797\n",
      "Epoch 9/25\n",
      "259/259 [==============================] - 94s 363ms/step - loss: 0.4373 - accuracy: 0.8552 - val_loss: 1.3445 - val_accuracy: 0.6684\n",
      "Epoch 10/25\n",
      "259/259 [==============================] - 94s 362ms/step - loss: 0.3987 - accuracy: 0.8670 - val_loss: 1.4762 - val_accuracy: 0.6605\n",
      "Epoch 11/25\n",
      "259/259 [==============================] - 95s 368ms/step - loss: 0.3559 - accuracy: 0.8775 - val_loss: 1.4017 - val_accuracy: 0.6730\n",
      "Epoch 12/25\n",
      "259/259 [==============================] - 94s 364ms/step - loss: 0.3240 - accuracy: 0.8886 - val_loss: 1.4677 - val_accuracy: 0.6792\n",
      "Epoch 13/25\n",
      "259/259 [==============================] - 94s 363ms/step - loss: 0.2952 - accuracy: 0.8992 - val_loss: 1.4975 - val_accuracy: 0.6662\n",
      "Epoch 14/25\n",
      "259/259 [==============================] - 95s 366ms/step - loss: 0.2544 - accuracy: 0.9128 - val_loss: 1.7121 - val_accuracy: 0.6529\n",
      "Epoch 15/25\n",
      "259/259 [==============================] - 94s 362ms/step - loss: 0.2560 - accuracy: 0.9117 - val_loss: 1.6411 - val_accuracy: 0.6706\n",
      "Epoch 16/25\n",
      "259/259 [==============================] - 93s 361ms/step - loss: 0.2349 - accuracy: 0.9186 - val_loss: 1.6326 - val_accuracy: 0.6748\n",
      "Epoch 17/25\n",
      "259/259 [==============================] - 94s 364ms/step - loss: 0.2004 - accuracy: 0.9336 - val_loss: 1.8116 - val_accuracy: 0.6564\n",
      "Epoch 18/25\n",
      "259/259 [==============================] - 94s 363ms/step - loss: 0.1952 - accuracy: 0.9361 - val_loss: 1.7910 - val_accuracy: 0.6620\n",
      "Epoch 19/25\n",
      "259/259 [==============================] - 94s 364ms/step - loss: 0.1722 - accuracy: 0.9432 - val_loss: 1.8632 - val_accuracy: 0.6586\n",
      "Epoch 20/25\n",
      "259/259 [==============================] - 94s 362ms/step - loss: 0.1657 - accuracy: 0.9427 - val_loss: 1.8228 - val_accuracy: 0.6613\n",
      "Epoch 21/25\n",
      "259/259 [==============================] - 95s 365ms/step - loss: 0.1643 - accuracy: 0.9438 - val_loss: 2.0060 - val_accuracy: 0.6551\n",
      "Epoch 22/25\n",
      "259/259 [==============================] - 94s 362ms/step - loss: 0.1523 - accuracy: 0.9480 - val_loss: 1.9895 - val_accuracy: 0.6610\n",
      "Epoch 23/25\n",
      "259/259 [==============================] - 94s 363ms/step - loss: 0.1653 - accuracy: 0.9428 - val_loss: 2.0874 - val_accuracy: 0.6465\n",
      "Epoch 24/25\n",
      "259/259 [==============================] - 95s 367ms/step - loss: 0.1314 - accuracy: 0.9554 - val_loss: 2.0069 - val_accuracy: 0.6627\n",
      "Epoch 25/25\n",
      "259/259 [==============================] - 94s 363ms/step - loss: 0.1152 - accuracy: 0.9609 - val_loss: 2.2994 - val_accuracy: 0.6519\n",
      "CPU times: user 34min 54s, sys: 2min 1s, total: 36min 56s\n",
      "Wall time: 2h 49min 48s\n"
     ]
    }
   ],
   "source": [
    "%%time\n",
    "\n",
    "epochs = 25\n",
    "\n",
    "history = model.fit(train_generator,\n",
    "                    steps_per_epoch=len(train_generator), \n",
    "                    epochs=epochs, \n",
    "                    validation_data=val_generator, \n",
    "                    validation_steps=len(val_generator))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 513
    },
    "executionInfo": {
     "elapsed": 10219329,
     "status": "ok",
     "timestamp": 1607991136586,
     "user": {
      "displayName": "Mahmut Can Kurt",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GiR0Vz-Xi8c64gs5AldT12G3fhmeowAL59isdw9=s64",
      "userId": "17273613824855319138"
     },
     "user_tz": -180
    },
    "id": "aEsGJIY5GEGY",
    "outputId": "ee2b99dc-465d-4a34-9999-186536efeb71"
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAfQAAAHwCAYAAABQR52cAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdd3hc5ZX48e9Rt7pkuanZcu9ykQummZZQbTqYathAQkJYyKaQhASWhIVN2F9CAmGXAAkEB4eQQOxgSsCAqcYF21jGNsaWVdzVZXXN+f1xr6SRrDKSNR6V83me+8ztc+Z6rDPve9/7vqKqGGOMMaZvCwp0AMYYY4w5fpbQjTHGmH7AEroxxhjTD1hCN8YYY/oBS+jGGGNMP2AJ3RhjjOkHLKGbAUVEXhWRG3t630ASkRwROdsP531HRL7mzl8rIm/4sm833iddRCpEJLi7sRpjLKGbPsD9Y984eUSkymv52q6cS1XPU9Vnenrf3khE7haRNW2sTxKRWhGZ6uu5VHWZqn6lh+Jq8QNEVXNVNVpVG3ri/G28n4jIbhHZ5o/zG9NbWEI3vZ77xz5aVaOBXOAir3XLGvcTkZDARdkrPQcsEJGMVuuvBj5T1a0BiCkQTgOGAqNFZM6JfGP7TpoTyRK66bNEZKGI5IvID0TkAPAHEUkQkX+KyGERKXbnU72O8a5GXioi74vIw+6+e0TkvG7umyEia0SkXETeFJHHROS5duL2JcaficgH7vneEJEkr+3Xi8heESkUkR+3d31UNR9YDVzfatMNwLOdxdEq5qUi8r7X8jkisl1ESkXkUUC8to0RkdVufEdEZJmIxLvb/gSkAyvdGpbvi8goEdHG5CciySKyQkSKRGSXiNzide77ROQFEXnWvTbZIpLV3jVw3Qj8A1jlznt/riki8i/3vQ6KyI/c9cEi8iMR+dJ9nw0iktY6Vnff1t+TD0TkVyJSCNzX0fVwj0kTkb+7/w6FIvKoiIS5MU3z2m+oiFSKyJBOPq8ZoCyhm75uOJAIjARuxflO/8FdTgeqgEc7OH4esANIAn4BPCUi0o19/wx8AgwG7uPYJOrNlxivAW7CKVmGAd8FEJHJwOPu+ZPd92szCbue8Y5FRCYAM9x4u3qtGs+RBPwduAfnWnwJnOy9C/CgG98kIA3nmqCq19OyluUXbbzFciDfPf5y4L9E5Eyv7YvcfeKBFR3FLCKR7jmWudPVIhLmbosB3gRec99rLPCWe+h3gCXA+UAscDNQ2eGFaTYP2A0MAx7o6HqI027gn8BeYBSQAixX1Vr3M17ndd4lwFuqetjHOMxAo6o22dRnJiAHONudXwjUAhEd7D8DKPZafgf4mju/FNjltS0SUGB4V/bFSYb1QKTX9ueA53z8TG3FeI/X8jeB19z5n+L8wW/cFuVeg7PbOXckUAYscJcfAP7RzWv1vjt/A/Cx136Ck4C/1s55LwY+bevf0F0e5V7LEJxk1wDEeG1/EPijO38f8KbXtslAVQfX9jrgsHvuCKAUuMTdtsQ7rlbH7QAWt7G+KdYOrlNuJ//eTdcDOKkxvjb2m4fz40fc5fXAlYH8/2dT756shG76usOqWt24ICKRIvJ/bpV0GbAGiJf2W1AfaJxR1cYSWHQX900GirzWAeS1F7CPMR7wmq/0iinZ+9yqehQobO+93Jj+Ctzg1iZcCzzbhTja0joG9V4WkWEislxECtzzPodTkvdF47Us91q3F6fk2qj1tYmQ9u9V3wi8oKr17vfkbzRXu6fh1C60paNtnWnxb9/J9UgD9qpqfeuTqOpanM+3UEQm4tQgrOhmTGYAsIRu+rrWwwX+BzABmKeqsTgNosDrHq8f7AcS3erdRmkd7H88Me73Prf7noM7OeYZ4ErgHCAGWHmccbSOQWj5ef8L599lmnve61qds6MhHvfhXMsYr3XpQEEnMR3DbQ9wJnCdiBwQp53F5cD57m2DPGB0O4fnAWPaWH/UffX+tx7eap/Wn6+j65EHpHfwg+QZd//rgRe9f7wa05oldNPfxODcCy4RkUTgXn+/oaruxakOvc9tzHQScJGfYnwRuFBETnHvBd9P5/+P3wNKgCdovj97PHG8AkwRkUvdRHQHLZNaDFABlIpICvC9VscfpJ1Eqqp5wIfAgyISISLTgX/DKdV21fXATpwfLTPcaTzO7YElOPeuR4jInSISLiIxIjLPPfZJ4GciMk4c00VksDr3rwtwfiQEi8jNtJ34vXV0PT7B+YH0kIhEuZ/Zuz3Cc8AlOEn92W5cAzOAWEI3/c2vgUHAEeBjnAZPJ8K1OPdDC4GfA38BatrZt9sxqmo28C2cRm37gWKcBNXRMYqTDEbSMil0Kw5VPQJcATyE83nHAR947fKfwCyc+9Wv4DSg8/YgcI+IlIjId9t4iyU496r3AS8B96rqm77E1sqNwO9U9YD3BPwvcKNbrX8Ozo+vA8AXwBnusf8PeAF4A6cNwlM41wrgFpykXAhMwfkB0pF2r4c6z95fhFOdnovzb3mV1/Y8YCNOCf+9rl8CM5A0NrYwxvQgEfkLsF1V/V5DYPo3EXka2Keq9wQ6FtO7WUI3pgeI02FJEbAH+ArwMnCSqn4a0MBMnyYio4BNwExV3RPYaExv57cqdxF5WkQOiUibvVG596V+I07HEVtEZJa/YjHmBBiO8/hSBfAb4DZL5uZ4iMjPgK3ALy2ZG1/4rYQuIqfh/HF7VlWP6TNaRM4Hvo3TccM84BFVndd6P2OMMcZ0zm8ldFVdg1MF2Z7FOMleVfVjnOdfR/grHmOMMaY/C2Qr9xRadsCQT8vOI4wxxhjjoz4xEpCI3IrTTzdRUVGzJ06cGOCIjDHGmBNjw4YNR1S100F5ApnQC2jZu1Qq7fQGpapP4HSKQVZWlq5fv97/0RljjDG9gIjs9WW/QFa5r8DtX1pE5gOlqro/gPEYY4wxfZbfSugi8jzOaFhJIpKP061kKICq/i/O2MTnA7twBiC4yV+xGGOMMf2d3xK6qi7pZLvidGFpjDHGmONkfbkbY4wx/YAldGOMMaYfsIRujDHG9AOW0I0xxph+wBK6McYY0w9YQjfGGGP6gT7R9asxxhjTW6gqJZV1HCyv5mBZDQfLqjlU1jzf4FGeWjrnhMdlCd0YY4xxlVfXcbCshkNl1RzwStKHWiTvGmobPMccGx8ZyrCYCJLjI1BVROSExm4J3RhjzIDU4FF2HChnw94i1u8tZsPeYvKLq47ZLyY8hKGx4QyLjWDOqERnPiaCYbERDHPXD4kJJyI0OACfopkldGOMMQNCRU09m3JLWL+3iA17i/k0t4SKmnoAhsSEkzUygWvnjSQ5PoKhMU6yHhobQXR430iVfSNKY4wxfZKqcrS2geKjtRQdraWospaSylqKjtY56yprKT5aS0VNPQmRYQyJCXemaOd1aKwznxAZRlCQ71XYqkpBSRUb3JL3+pxith8ow6MgAhOGxXDxzGRmj0wga2QiqQmDTngVeU+zhG6MMX2Mx6McLK+moLiK2noPQUFCkAjBQSAiBIuzHBQEwe42Z/JaDnKXRaj3KPUNSm2Dh3qPh/oGpa7BQ73HfW1Q6j0e6hq0aVvj9voGD7UNSmllY3Kuo+hoLcWV7nS0rs37zeDEkhAZSkJkGJHhIeQWVXKorIaquoY2902KDmuR7IfEhDM0JqJpPjhI2JxX4lSf5xRzoKwagMiwYGamx3P7mePIGpnAjPR4YiNC/fpvFAiW0I0xphcqrawjr7iS3KJK8ooq3fkq8osqyS+uajdJBooIJESGER8ZSmJkGGmJkWSmxpMQFUZilJO0EyLD3OUwEiPDiIkIabPUfbSmnsPlNRyuqHFey2s4VF7dNH+4ooZt+8s4UlFLg0ePOT4lfhBzMhLJGpnA7JEJTBweQ0hw/39K2xK6McacYKpKZW0DB8qqnWRdVElecRV5Rc0JvKy6vsUx8ZGhpCVEMmlELOdMGUZ6YiSpCZGEhwThUcXjAY8qDaqoKg3ussfjrPOoU7L3qNLgUVRx1yvBIoQEBxEaLIQGBxES5L4GCyFBzvqQVutDg9zX4CDCgoOIjgghuAtV4h2JCg8hKjyEUUlRHe7n8SjFlbVNib+qtoFpqXGMiBvUI3H0NZbQjTGmCzweZdv+MgpKqqisraeytoHKmgaO1tZTVeu8VtY0UNk4X9vgTvUcrWmgqraeyroGtFXBMiwkiLSEQaQlRjJ7ZAJpCZGkJUaSluis649VxMcrKEgYHB3O4OhwJg4PdDSBZwndGGM6caC0mjVfHOa9L47w/heHKa6sa3O/iNAgosJCGBQW3PwaHkxiVBiRYcFEhoUQFRZMZFgwg8JCGBYbTlpiJOmJkQyJDu9Soy9jWrOEbowxrVTVNvBJThFrdh7mvS8Os/NgBeA82nTGxKGcNm4IY4dGO1XDYcFEhocwKDS4x6qcjekOS+jGmAFPVdl+oJz3vjjMmp1H+CSniNp6D2EhQczLSOTy2amcOm4IE4fH9PlHm0z/ZQndGDMgHamo4f0vjjRVpR8urwFg/LBorp8/ktPGD2HuqEQGhQW29y9jfGUJ3RjTb6kqRUdryS+ucifn8a9Pc0vI3lcGQEJkKKeMG8Kp45I4bdwQhsdFBDhqY7rHEroxps9qK2G3fK06ppOSuEGhTBgew/e+OoFTxyUxNTnOGqOZfsESujHGb+obPGzbX8ba3UWs3VNITmElIUHS8vnmoMbnnZufhQ4JCmreLziI0KDm56Araxs6TdipCYMYPSSK08YPITVhEKkJkaQmDCIlYZA9/mX6Lb8mdBE5F3gECAaeVNWHWm0fCTwNDAGKgOtUNd+fMRlj/KeuwcPWglLW7ili7e5C1uUUNw1+kZEUxYRhMXhUnS5D3W5D6xuU6vqGpi5FG9xt3l2ONnZNWtfgITwkiLTESEvYxrTit4QuIsHAY8A5QD6wTkRWqOo2r90eBp5V1WdE5EzgQeB6f8VkjOlZtfUePiso4ePdRazdU8T6nCIqa50S85ghUSyekcy80YOZn5HI0Fi7N22MP/mzhD4X2KWquwFEZDmwGPBO6JOB77jzbwMv+zEeY8xxqqlvYHNeKWt3F7J2jzMEZWOV9/hh0Vw+O5V5GYOZm5HIkJjwAEdrzMDiz4SeAuR5LecD81rtsxm4FKda/hIgRkQGq2qhH+MyxvhAVcktqmRrQRlb95WyKbeEjbnF1NQ7g4JMHB7DVXPSmD86kTmjEhkcbQncmEAKdKO47wKPishSYA1QABwzbp6I3ArcCpCenn4i4zNmQGjwKLsPV7B1XylbC8rI3ldK9r4yyt0BQkKDhQnDY7h23kjmj05kbkYi8ZFhAY7aGOPNnwm9AEjzWk511zVR1X04JXREJBq4TFVLWp9IVZ8AngDIyso6dqw8Y4zPaus9fHGonGy35L21oJTP95c3VZ2HhwQxaUQsi2ckMzU5jqkpcYwbFk14iHWwYkxv5s+Evg4YJyIZOIn8auAa7x1EJAkoUlUP8EOcFu/GmB5SU9/A9v3lbCkoJbvAKXXvOFDeNJZ2dHgIk5NjWTI3nakpsUxNiWN0UtSAGDvamP7GbwldVetF5HbgdZzH1p5W1WwRuR9Yr6orgIXAgyKiOFXu3/JXPMb0d3UNHnYeLOez/FI255fyWUEJOw6UU9fgVGrFR4YyLSWOm04Z1VTyHpkYaZ2qGNNPiLYelLeXy8rK0vXr1wc6DGMCqsGj7DpUwZb8Ej4rKGVLfinb9pdR6zZYi40IYXpqPNNS48hMdZJ3SvwgG1jEmD5IRDaoalZn+wW6UZwxphMej7Kn8Cif5TuJ+7OCErYWlDXd844KC2ZqShw3njSS6anxTE+NIz0x0pK3MQOMJXRjepG6Bg+7DlWQva+5pfnn+8ood3tbiwgNYkpyHFfNSSMzLY5pKfGMToqyanNjjCV0YwKlsraez/eXs81N3Nn7ythxsLyp2jwi1G1tPjOZ6SnxTE+LY+yQaGuwZoxpkyV0Y06AksraFqXu7H1l7D5cgcdtwhIfGcqU5FiWLhjFlORYpiTHkpEUTbCVvI0xPrKEbkwPK6+uY0t+KZvyStic54y7XVBS1bQ9OS6CyclxXDBthJO8U+JIjouwe97GmONiCd2Y41Df4GHHwXI25ZWwKbeETXkl7DpcQePDI6OTopg9MoEbThrJlOQ4JifHkhhlPawZY3qeJXRjfKSq7C+tdpK3m8A/Kyhtam2eEBnKjLR4LpyezIz0eGakxhMXaUN5GmNODEvoxrSjoqaeLfklLUrfh8prAAgLDmJycixXzUljZno8M9Li7VExY0xAWUI3xouq8smeIpatzeW1rQeaukgdNTiSBWMGMyMtnhnpCUwaEWN9mxtjehVL6MYApZV1/G1jPn/+JJddhyqIiQjhmnnpnD5hCDNS40mw+97GmF7OEroZsFSVTXklLFuby8rN+6ip9zAjLZ5fXj6dC6cnMyjMSuDGmL7DEroZcCpq6nn50wL+vDaXbfvLiAoL5rLZqVwzN52pKXGBDs8YY7rFEroZMLYWlPLnT3L5x6cFHK1tYNKIWB64ZCqLZ6QQHW7/FYwxfZv9FTP9WlVtAyu37GPZ2lw255UQHhLERZnJXDsvnRlp8dYq3RjTb1hCN/2Ox6NsP1DOC+vz+NvGfMqr6xk7NJp7L5rMpTNT7dlwY0y/ZAnd9Hnl1XVszitlw95iNuYW82luMWXV9YQGC+dNHcG189KZm5FopXFjTL9mCd30KarKniNH2Zhbwoa9TvLecbAcVRCB8UNjuGD6CGamJ3DmxKEkRYcHOmRjjDkhLKGbXu1oTT2b80v4NLeEjW4JvLiyDoCYiBBmpidw7tThzB6ZQGZaPLERVp1ujBmYLKGbXmV/aRWf7ClifY6TvLcfKKfBHWN07NBozpk8jFnpCcwamcDYIdEE2fCixhgDWEI3AZZXVMnaPUWs3V3I2j1F5BZVAhAVFszM9AS+tXAMM0cmMDMtnvhI663NGGPaYwndnDCqyt7CStbuKWTt7iLW7ilqGic8PjKUuaMSuXHBKOZlJDJpRCzBVvo2xhifWUI3fqOqfHn4qFcCL+RgmTNaWVJ0GHMzErn1tNHMG53I+KExVn1ujDHHwa8JXUTOBR4BgoEnVfWhVtvTgWeAeHefu1V1lT9jMv6VW1jJOzsPNZXAj1Q4CXxoTDjzRg9mXkYi80cnMmZItD1GZowxPchvCV1EgoHHgHOAfGCdiKxQ1W1eu90DvKCqj4vIZGAVMMpfMRn/2Vt4lEfe+oKXPy3Ao5AcF8Gp45KYl5HIvNGDGTXYxgo3xhh/8mcJfS6wS1V3A4jIcmAx4J3QFYh15+OAfX6Mx/hBXlElj67exYsb8wkJEv7tlAyunz+K9MGRgQ7NGGMGFH8m9BQgz2s5H5jXap/7gDdE5NtAFHC2H+MxPWhfSRWPvr2LF9blERQkXD9/JN9cOIahsRGBDs0YYwakQDeKWwL8UVX/R0ROAv4kIlNV1eO9k4jcCtwKkJ6eHoAwTaODZdU89vYuln+Sh6IsmZvON88Yw4i4QYEOzRhjBjR/JvQCIM1rOdVd5+3fgHMBVPUjEYkAkoBD3jup6hPAEwBZWVnqr4BN+w6VV/O/7+zmubV78XiUK7LSuP3MsaTEWyI3xpjewJ8JfR0wTkQycBL51cA1rfbJBc4C/igik4AI4LAfYzJdVFhRw/+t2c2zH+VQ16BcNiuFb585jrREu0dujDG9id8SuqrWi8jtwOs4j6Q9rarZInI/sF5VVwD/AfxeRO7CaSC3VFWtBN4LFB+t5Yn3dvPMhzlU1zVw8YwUvn3WODKSogIdmjHGmDb49R66+0z5qlbrfuo1vw042Z8xmK4prazjqfd38/QHORytrefC6cn8+1njGDs0OtChGWOM6UCgG8WZXqKmvoEn39vD/777JeXV9Zw/bTj/ftZ4JgyPCXRoxhhjfGAJ3fDBriP85B9b2X34KGdPGsp3zpnA5OTYzg80xhjTa1hCH8AOlVXz81c+Z8XmfaQnRvKHm+ZwxoShgQ7LGGNMN1hCH4DqGzw89/Fe/ueNndTUe/j3s8Zx28IxRIQGBzo0Y4wx3WQJfYD5NLeYe17eSva+Mk4dl8T9i6day3VjjOkHLKEPECWVtfz3aztYvi6XoTHhPHbNLM6fNtwGTDHGmH7CEno/p6q8uCGfB1/dTmlVHTefnMGdZ48jJiI00KEZY4zpQZbQ+7HtB8r4yctbWZdTzKz0eH5+8TRrvW6MMf2UJfR+6GhNPY+89QVPvb+H2IgQfnHZdC6fnUpQkFWvG2NMf2UJvR9RVV7beoD7/7mN/aXVXD0njR+cO5GEqLBAh2aMMcbPLKH3E7sPV/CfK7fx7s7DTBoRy6PXzGL2yIRAh2WMMeYE6TShi8hFwCutxyg3vcPOg+U89vYuVm7ex6DQYH5y4WRuPGkkIcFBgQ7NGGPMCeRLCf0q4Nci8jecEdO2+zkm44OtBaU8unoXr2UfIDIsmFtOHc2/nZrB0JiIQIdmjDEmADpN6Kp6nYjEAktwxi1X4A/A86pa7u8ATUsb9hbz6OoveHvHYWIiQrjjzLHcdHKG3Sc3xpgBzqd76KpaJiIvAoOAO4FLgO+JyG9U9bf+DNA4jd0+3l3Eb1d/wYdfFpIQGcr3vjqB608aSaw9T26MMQbf7qEvAm4CxgLPAnNV9ZCIRALbAEvofqKqvLvzMI+u3sX6vcUkRYfz4/Mncc28dKLCrT2jMcaYZr5khcuAX6nqGu+VqlopIv/mn7AGNo9H+dfnB3l09S4+KyglOS6C+xdP4cqsNBtAxRhjTJt8Sej3AfsbF0RkEDBMVXNU9S1/BTYQNXiUVZ/t57G3d7H9QDnpiZH892XTuGRmKmEh1mrdGGNM+3xJ6H8FFngtN7jr5vglogGovsHDy5v28bu3d7H7yFHGDIniV1dlctH0ZHv8zBhjjE98SeghqlrbuKCqtSJiTap7iMejfHPZRt7YdpBJI2L53bWzOHfKcOum1RhjTJf4ktAPi8giVV0BICKLgSP+DWvg+NWbO3lj20HuPm8iXz9ttA1naowxplt8SejfAJaJyKOAAHnADX6NaoB4Zct+frt6F1dmpVoyN8YYc1x86VjmS2C+iES7yxW+nlxEzgUeAYKBJ1X1oVbbfwWc4S5GAkNVNd7X8/dl2ftK+e5fNzN7ZAI/u3iqJXNjjDHHxaeHmUXkAmAKENGYeFT1/k6OCQYeA84B8oF1IrJCVbc17qOqd3nt/21gZlc/QF90pKKGW5/dQHxkKI9fN4vwEHsUzRhjzPHptAm1iPwvTn/u38apcr8CGOnDuecCu1R1t9uobjmwuIP9lwDP+3DePq223sNtz23gSEUNT1yfZX2vG2OM6RG+PBO1QFVvAIpV9T+Bk4DxPhyXgnO/vVG+u+4YIjISyABW+3DePktVuXdFNutyivnF5dOZlhoX6JCMMcb0E74k9Gr3tVJEkoE6YEQPx3E18KKqNrS1UURuFZH1IrL+8OHDPfzWJ85zH+/l+U9yuW3hGBbPaPO3jTHGGNMtviT0lSISD/wS2AjkAH/24bgCIM1rOdVd15ar6aC6XVWfUNUsVc0aMmSID2/d+3z45RHuW7mNMycO5btfmRDocIwxxvQzHTaKE5Eg4C1VLQH+JiL/BCJUtdSHc68DxolIBk4ivxq4po33mAgkAB91Nfi+Iq+okm8t20hGUhSPXD2DYOs0xhhjTA/rsISuqh6cluqNyzU+JnNUtR64HXgd+Bx4QVWzReR+dwS3RlcDy1VVuxx9H3C0pp5bnl1Pg0f5/Q1ZxNhwp8YYY/zAl8fW3hKRy4C/dzXpquoqYFWrdT9ttXxfV87Zl3g8ynde2MTOg+X88aa5ZCRFBTokY4wx/ZQvCf3rwHeAehGpxnl0TVU11q+R9QO/fusLXs8+yE8unMxp4/vmvf8+o6EOjnwBB7Ph4FY4vANihkHyLEiZBUMmQfAAGUO+pgLKCqA033kt2wc15c51SF8AsT3dptUY0xv40lNczIkIpL959bP9/OatL7h8dio3nzwq0OH0H6pQcdBJ2ge3uQk8Gw5vB0+ds09QKAweC7kfwoY/OutCImD4dCepNSb5xDEQ1MdGs6utdBJ0WT6UFrgJu6B5vrQAalrfFRMIDoOGGmcxfiSMXADpJzmvg8fCQOipsK4KKguhsgiihtgPG9PvSGe16CJyWlvrVXWNXyLqRFZWlq5fvz4Qb+2zbfvKuOzxD5k4Ioblt87v3T3B1VVBSR6U5ELJXqguBfU4iVMb3HkPeLzmW08ttjVAcDiEx7hTLIRHey3HQJj3fFT7yaSuyknUjUn74FbntbKweZ/YFBg2xZmGuq9J4yA41PkMRbth36dQsBH2bYT9m6Gu0jk2PA6SM50EnzzTSfJxab0judWUQ/56yFsL+za5pe18qCo+dt+oIRCbDLGpEJfiXJPYlOb5mBEgQXBgC+R+BHs/hNyPofJI8/Hp853S+8iTYNi03l2boer8GzYm58pC57q0WHZfKwuh0t1WX9XyPOknwZRLYPJiiBkemM9ijA9EZIOqZnW6nw8JfaXXYgROD3AbVPXM4wuxe3p7Qi+sqGHRox/Q4FFW3H4yQ2MD3BNcXbWTDEr2ulNu81S8F44e6vwcEuROwc3zQcFO4mtrW0MNVJc1l5g7O3dYTMukHxbllEILdzk/EgBCBsGwyW7ynuom8MkQmdi169FQD0d2NCf4fZ/Cga3NsUYmtSzFJ42H+HTn8/qLKpTmQd4nTqLN+9j54aIeQGDIBEjIcJJ2XIqTuBvnY5IhtBvfMVXn+u79sDnJl+x1toVFQ9pcJ8Gnz4fULAgd1KMfucuK9sDnK+DzlbB/S3NtQ1si4p3vReRgZxqU6C676wYlOLdksl+CQ9sAgZEnw5SLneQePfSEfSxjfNFjCb2NE6cBv1bVy7ob3PHozQm9tt7DdU+tZXNeCX/9xklMTz1B48yowoHPnATVOmFXHGi5b1AoxKU6SSo+3al+jU+HBPd1UEKrxH0cVdL1Nc793Joyp8TZONW2sa71fjHDm0vew6ZCwij/JdX6Gqf0X7CxuTR/eDvg/t8ICnXeP3E0DLJ2aKsAACAASURBVB7j9TrGuZZdjauhzvn3ylvrTLlroXyfsy00ykmg6fOdpJo6ByJOUI+CZfu8EvxHbrJT5/Mnz3Sq50eeDOnzTkxMh7Y7SXzbCjj4mbNuRCaMOhWikryS9eDmZB0R37XahUPbncSe/Xc4stP5zo88GaZeCpMWOe9jTID5M6ELkK2qk7sb3PHozQn9xy99xrK1uTxy9Qz/9wTXUO+U5D7/J2x/BUpznfVBIW0n7Mb5mOH+LW32FzUVTtIt3AVFX0Lhl071fdHu5ip7cO5NJ4xykvvgMZCY0Twfm+r8IKoqdqrPcz92EnjBhuZzxKU5iTttvpMoh07pPdXdVcXOj43cj5ypYKNbkyEwfJqT+EYucKaeSHyqzi2Rz1c6ifzITmd92jwnuU660LnW/qAKhz5vTu6Fu5wfthmnOtXykxZ1vTaoPbWVUL4fyg84/xcj4pqn0MgTe8tH1flBGxLeO241BVLtUdj9rvNvP2wyjJjRa37Q9WSV+29pKqoQBMwAclT1uuOOsht6a0L/08d7+cnLW/n66aP54XmT/PMmdVXw5dtOAt/5qnNfMDgcxpwBEy+EjNO6V2I0vlN1/hA3JfnG1z1Osve+TxscDtHDmn9sSbCTCNPnO0kqbZ5Tbd5X1FZCwXqnFL/3A8hb1/x5kyY0l+BHLvD9c3k8zjm3/cNJ5CV7nVLyqFOcJDrxwhPfeE3VueWR/XcnwRftdv7tRi90k/uFTk1Waw11UHHITdb7oWx/83zT8oE2Gi16CQppmeDbnOLdKQ4iYp33ra1wfoTWVrQzf7SNbe469Tg1Q23d0vGeP55amYa65mtQVuDO72s51ZRCymznOo9e6LTl8Hej1ZJc2Pm6M+1Zc+ytnNhUSJ7h1AyNcF9jhvk3pjb0ZEK/0WuxHieZf3Cc8XVbb0zoH+8u5Lon13LquCSevHFOz/YEV1UMO9+A7Sth11tOyS48DsZ/FSZeAGPPdu4/m8DzeJw/VEVuab7wS+cP1ZCJTuk7ZbbTPqC/qK+F/Zuc5N7Y0K6mzNmWMKplCT4ho7kE2FDvPIGwbQVs/6dzzYJCnT/ikxfBhAsganCAPlQrqk5jwuyXnKk4x4l1zBlOg0PvZH30MM1lH1dQCEQPd2rGYkc4DRSbJjcxVJc2T1UlLZdbT60b9nUkNNJpDxEe7XzvGtuqhEU56xu3hUQ4jQm9n5woP3DsZwmLcZN8stvoMrV5Pnqoc47y/c2PSnon74pDx54vZJBzTWLdc4ZEODVYh7c72wclwujTIeN057uRmOH7Z2+PpwHy18HO15wkfsgdzTtxNIw/z/m7OmyKs37fJqfGaP8mp9TeKGZEc3JvTPYxI/xaw9GTCT0KqG4cOMUd5zxcVSs7PNBPeltCzy+uZNGjH5AQGcpL3zqZ2J7oCa60AHasckosez8AT73zR2HiBU7pYOQpEBJ2/O9jTE/yNDhtEfZ+CDnvO69VRc62mBFOYg8dBDvc2qWQQTD2LKch2vivnri2At2l6vxx3/p3p0ah9qjzuWJHOAnbO1k3Ju/IpJ4tZda7DU6rS6HaTf7BoW5yjvFK1lHHV1PXUOck9WMei8x3k3WB8/hoeyLi3WTvTjHJrZZHuO112kiCZfthz7uw+x1nKt/vrI8f2Vx6zzjd9x99VcVOYWjn67DrX85yUIjzlMP4c50paWzH56guc27B7XeT/L5N7i0hN39GDW1Zkk+e4fxQ6aEk35MJ/WPgbFWtcJejgTdUdUGPRNpFvS2hP/jq5zz9/h5ev/M0Rg/pZklZ1flyfL7SqU7ft9FZP3ick8AnXui0uu5rz0ybgU3VaU3eWILf+4FT3Tv+q05JfOzZ/avGYqCpr20ukVccctoYND4mGRbZM++h6nQY1Zjcc95rrgUaPq05wacvaH7PxmMaS+G5HzmP00YOhnFfcb5/Y848/h+QNRXOD9h9m5oT/eHt7i2MSPhhfo/d/uzJhL5JVWd0tu5E6W0J/eonPqKytoEVt5/S8Y5VxS1boHu3RC/JhdpyZ7+U2U5JfOJFMMSXYeeN6SMa/9YM9MZXpvsa6p3kufttpwFb3lpoqHUap6bOdRqj7lkDxXuc/YdNdRL4+HOdv63+bl9UW+m0vygrcB6D7CG+JnRfmtMeFZFZqrrRPfFsoAs3cvovj0fZWlDGxTOTnaov7wTdemrdECYsurkF+qhTnGeNx5/btxpJGdMVlsjN8QoOcR7rTM2C077nJNDcj5pL8J+9CKNOhgW3w7ivQnxaZ2fsWWGRkDYHmHNi39flS0K/E/iriOzD6cd9OHCVX6PqI3YfLuf6+r9zZ/Yq2FTWcmNoVPOz3SNPOvYxsvbuHxljjPFNWKTTDmPsWYGOpFfwpS/3de6Y5RPcVTtU1YcuwPq5umpCV3yTH4SupCL5LMLHL2yZtCMTLWEbY4w5YTpN6CLyLWCZqm51lxNEZImq/s7v0fVWFYfhL9cysmAtj3iu5PYb/g+CrcGaMcaYwPElC92iqiWNC6paDNziv5B6uYPb4MkzYf8W/jv2h3yQcjPBlsyNMcYEmC+ZKNjt7hVoeg59YD4EvfMNeOorUF9L3Q2v8FRhJpmpvfzZWWOMMQOCLwn9NeAvInKWiJwFPA+86t+wehlV+OgxeP4qp7eiW1azPWgstQ2eEzcAizHGGNMBX1q5/wC4FfiGu7wFp6X7wNBQB6/8B2x8BiZdBJf8H4RFsflzZ6jJTEvoxhhjegFfWrl7RGQtMAa4EkgC/ubvwHqFyiJ44Qand6JT/wPOuKept7Yt+SUkRIaSlhjgcaKNMcYYOkjoIjIeWOJOR4C/AKjqGScmtAA7sgv+fCWU5jml8syrW2zekl/K9NR4xB5NM8YY0wt0VELfDrwHXKiquwBE5K4TElWg7X7HKZkHhcKNK53hLr1U1taz82A5X5l84ofRM8YYY9rSUaO4S4H9wNsi8nu3QVyXiqMicq6I7BCRXSJydzv7XCki20QkW0T+3JXz+8X6p+FPlzqjA92y+phkDrC1oAyPYg3ijDHG9BrtltBV9WXgZXf41MU4XcAOFZHHgZdU9Y2OTuw+3vYYcA6QD6wTkRWqus1rn3HAD4GTVbVYRIYe9yfqroZ6eOMeWPu4MyLPZU9BRGybu27Jdx7Ln55mj6wZY4zpHTp9bE1Vj6rqn1X1IiAV+BSn5Xtn5gK7VHW3qtYCy3F+GHi7BXjM7awGVT3Upeh7SnUpPH+1k8znfxOWLG83mQNszi9lRFwEQ2MiTmCQxhhjTPu61MWZqhar6hOq6ktP+ClAntdyvrvO23hgvIh8ICIfi8i5XYmnRxTnwFNfdYbju/BXcO6DnQ6xtyW/hOnWoYwxxphexJfn0P39/uOAhTil/zUiMs27q1kAEbkV51l40tPTe+7dcz+G5deApx6u+zuMPr3TQ0oqa9lbWMmVWSd4WD5jjDGmA/7shLwA8M56qe46b/nAClWtU9U9wE6cBN+CWyuQpapZQ4YM6bkI66shagh8bbVPyRycx9UAZqRZgzhjjDG9hz8T+jpgnIhkiEgYcDWwotU+L+OUzhGRJJwq+N1+jKml0QvhGx9A0lifD2lsEDc1xarcjTHG9B5+S+iqWg/cDrwOfA68oKrZInK/iCxyd3sdKBSRbcDbwPdUtdBfMbUpuGt3HTbnlzI6KYq4QaF+CsgYY4zpOr/eQ1fVVcCqVut+6jWvwHfcqU/YnFfCgjGDAx2GMcYY04IN5N0FB0qrOVReYx3KGGOM6XUsoXfBZvf+eaZ1KGOMMaaXsYTeBVvySwgOEiaPsIRujDGmd7GE3gVb8kuZMCyGQWEddzxjjDHGnGiW0H2kqmzJL7XqdmOMMb2SJXQf7S2spLSqzhrEGWOM6ZUsofuosUGc9eFujDGmN7KE7qPNeaWEhwQxflhMoEMxxhhjjmEJ3Udb8kuYkhxLaLBdMmOMMb2PZScf1Dd42Lqv1O6fG2OM6bUsofvgi0MVVNd5bIQ1Y4wxvZYldB9ssQZxxhhjejlL6D7YnF9KTEQIowZHBToUY4wxpk2W0H2wJb+E6alxBAVJoEMxxhhj2mQJvRPVdQ1s319uDeKMMcb0apbQO7Ftfxn1HiXT7p8bY4zpxSyhd2JLXmODOCuhG2OM6b0soXdiS34pQ2LCGREXEehQjDHGmHZZQu/E5vwSMlPjELEGccYYY3ovS+gdKK+uY/eRo1bdbowxptezhN6BzwpKUbUOZYwxxvR+ltA7sCW/FLAGccYYY3o/vyZ0ETlXRHaIyC4RubuN7UtF5LCIbHKnr/kznq7anFdCWuIgEqPCAh2KMcYY06EQf51YRIKBx4BzgHxgnYisUNVtrXb9i6re7q84jseW/FJmpFvp3BhjTO/nzxL6XGCXqu5W1VpgObDYj+/Xo45U1FBQUsUMq243xhjTB/gzoacAeV7L+e661i4TkS0i8qKIpPkxni6xEdaMMcb0JYFuFLcSGKWq04F/Ac+0tZOI3Coi60Vk/eHDh09IYJvzSgkSmJpiCd0YY0zv58+EXgB4l7hT3XVNVLVQVWvcxSeB2W2dSFWfUNUsVc0aMmSIX4JtbUt+CWOHRhMV7rdmBsYYY0yP8WdCXweME5EMEQkDrgZWeO8gIiO8FhcBn/sxHp+pKlvyS+1xNWOMMX2G34qfqlovIrcDrwPBwNOqmi0i9wPrVXUFcIeILALqgSJgqb/i6Yr84ioKj9baCGvGGGP6DL/WJ6vqKmBVq3U/9Zr/IfBDf8bQHdahjDHGmL7GbhC3YUt+CWHBQUwcERPoUIwx/VRdXR35+flUV1cHOhTTS0RERJCamkpoaGi3jreE3obN+SVMGhFDeEhwoEMxxvRT+fn5xMTEMGrUKBvN0aCqFBYWkp+fT0ZGRrfOEejH1nodj0fZWlBm1e3GGL+qrq5m8ODBlswNACLC4MGDj6vGxhJ6K7uPVFBRU28dyhhj/M6SufF2vN8HS+itbM5zGsRlplkJ3RjTfxUWFjJjxgxmzJjB8OHDSUlJaVqura3t8Nj169dzxx13dPoeCxYs6KlwAbjzzjtJSUnB4/H06Hn7C7uH3sqW/BIiw4IZMyQ60KEYY4zfDB48mE2bNgFw3333ER0dzXe/+92m7fX19YSEtJ0isrKyyMrK6vQ9Pvzww54JFvB4PLz00kukpaXx7rvvcsYZZ/TYub119Ll7Oyuht7Ipv5SpKXEEB1lVmDFmYFm6dCnf+MY3mDdvHt///vf55JNPOOmkk5g5cyYLFixgx44dALzzzjtceOGFgPNj4Oabb2bhwoWMHj2a3/zmN03ni46Obtp/4cKFXH755UycOJFrr70WVQVg1apVTJw4kdmzZ3PHHXc0nbe1d955hylTpnDbbbfx/PPPN60/ePAgl1xyCZmZmWRmZjb9iHj22WeZPn06mZmZXH/99U2f78UXX2wzvlNPPZVFixYxefJkAC6++GJmz57NlClTeOKJJ5qOee2115g1axaZmZmcddZZeDwexo0bR2O35B6Ph7Fjx3Kiuin31jd/hvhJbb2Hz/eVceOCkYEOxRgzgPznymy27Svr0XNOTo7l3oumdPm4/Px8PvzwQ4KDgykrK+O9994jJCSEN998kx/96Ef87W9/O+aY7du38/bbb1NeXs6ECRO47bbbjnn06tNPPyU7O5vk5GROPvlkPvjgA7Kysvj617/OmjVryMjIYMmSJe3G9fzzz7NkyRIWL17Mj370I+rq6ggNDeWOO+7g9NNP56WXXqKhoYGKigqys7P5+c9/zocffkhSUhJFRUWdfu6NGzeydevWphbmTz/9NImJiVRVVTFnzhwuu+wyPB4Pt9xyS1O8RUVFBAUFcd1117Fs2TLuvPNO3nzzTTIzMzlR3ZR7sxK6lx0Hyqlt8Nj9c2PMgHXFFVcQHOw8sltaWsoVV1zB1KlTueuuu8jOzm7zmAsuuIDw8HCSkpIYOnQoBw8ePGafuXPnkpqaSlBQEDNmzCAnJ4ft27czevTopiTaXkKvra1l1apVXHzxxcTGxjJv3jxef/11AFavXs1tt90GQHBwMHFxcaxevZorrriCpKQkABITEzv93HPnzm3xuNhvfvMbMjMzmT9/Pnl5eXzxxRd8/PHHnHbaaU37NZ735ptv5tlnnwWcHwI33XRTp+/nD1ZC97LZHTI10x5ZM8acQN0pSftLVFRU0/xPfvITzjjjDF566SVycnJYuHBhm8eEh4c3zQcHB1NfX9+tfdrz+uuvU1JSwrRp0wCorKxk0KBB7VbPtyckJKSpQZ3H42nR+M/7c7/zzju8+eabfPTRR0RGRrJw4cIOHydLS0tj2LBhrF69mk8++YRly5Z1Ka6eYiV0L1vyS0iIDCU1YVCgQzHGmIArLS0lJSUFgD/+8Y89fv4JEyawe/ducnJyAPjLX/7S5n7PP/88Tz75JDk5OeTk5LBnzx7+9a9/UVlZyVlnncXjjz8OQENDA6WlpZx55pn89a9/pbCwEKCpyn3UqFFs2LABgBUrVlBXV9fm+5WWlpKQkEBkZCTbt2/n448/BmD+/PmsWbOGPXv2tDgvwNe+9jWuu+66FjUcJ5oldC+NI6zZs6HGGAPf//73+eEPf8jMmTO7VKL21aBBg/jd737Hueeey+zZs4mJiSEurmUfIJWVlbz22mtccMEFTeuioqI45ZRTWLlyJY888ghvv/0206ZNY/bs2Wzbto0pU6bw4x//mNNPP53MzEy+853vAHDLLbfw7rvvkpmZyUcffdSiVO7t3HPPpb6+nkmTJnH33Xczf/58AIYMGcITTzzBpZdeSmZmJldddVXTMYsWLaKioiJg1e0A0tjSsK/IysrS9evX9/h5K2vrmXrv69x+xli+85UJPX5+Y4zx9vnnnzNp0qRAhxFwFRUVREdHo6p861vfYty4cdx1112BDqvL1q9fz1133cV77713XOdp63shIhtUtdPnBK2E7tpaUIZHbYQ1Y4w5kX7/+98zY8YMpkyZQmlpKV//+tcDHVKXPfTQQ1x22WU8+OCDAY3DGsW5trgN4qanWZevxhhzotx11119skTu7e677+buu+8OdBhWQm+0Ob+U5LgIhsZEBDoUY4wxpsssobu25JdYdbsxxpg+yxI6UFJZy97CSqtuN8YY02dZQsd5XA2sQxljjDF9lyV0mhvETU2xEroxZmA444wzmrpPbfTrX/+6qRvVtixcuJDGx4bPP/98SkpKjtnnvvvu4+GHH+7wvV9++WW2bdvWtPzTn/6UN998syvhd2igDrNqCR2nQdzopCjiBoV2vrMxxvQDS5YsYfny5S3WLV++vMMBUrytWrWK+Pju1Wq2Tuj3338/Z599drfO1VrrYVb9xR8d7RwvS+jA5rwSpqda6dwYM3BcfvnlvPLKK039mefk5LBv3z5OPfVUbrvtNrKyspgyZQr33ntvm8ePGjWKI0eOAPDAAw8wfvx4TjnllKYhVsF5xnzOnDlkZmZy2WWXUVlZyYcffsiKFSv43ve+x4wZM/jyyy9bDGv61ltvMXPmTKZNm8bNN99MTU1N0/vde++9zJo1i2nTprF9+/Y24xrIw6z69Tl0ETkXeAQIBp5U1Yfa2e8y4EVgjqr2fDdwHThQWs2h8hobYc0YEziv3g0HPuvZcw6fBue1+ScXcEYKmzt3Lq+++iqLFy9m+fLlXHnllYgIDzzwAImJiTQ0NHDWWWexZcsWpk+f3uZ5NmzYwPLly9m0aRP19fXMmjWL2bNnA3DppZdyyy23AHDPPffw1FNP8e1vf5tFixZx4YUXcvnll7c4V3V1NUuXLuWtt95i/Pjx3HDDDTz++OPceeedACQlJbFx40Z+97vf8fDDD/Pkk08eE89AHmbVbyV0EQkGHgPOAyYDS0Rkchv7xQD/Dqz1VywdaRxhzR5ZM8YMNN7V7t7V7S+88AKzZs1i5syZZGdnt6geb+29997jkksuITIyktjYWBYtWtS0bevWrZx66qlMmzaNZcuWtTv8aqMdO3aQkZHB+PHjAbjxxhtZs2ZN0/ZLL70UgNmzZzcN6OJtoA+z6s8S+lxgl6ruBhCR5cBioPU342fAfwPf82Ms7dqSX0JIkDAlOTYQb2+MMR2WpP1p8eLF3HXXXWzcuJHKykpmz57Nnj17ePjhh1m3bh0JCQksXbq0w6FDO7J06VJefvllMjMz+eMf/8g777xzXPE2DsHa3vCrA32YVX/eQ08B8ryW8911TURkFpCmqq/4MY4ObckvZfywGCJCAzPcnTHGBEp0dDRnnHEGN998c1PpvKysjKioKOLi4jh48CCvvvpqh+c47bTTePnll6mqqqK8vJyVK1c2bSsvL2fEiBHU1dW1SF4xMTGUl5cfc64JEyaQk5PDrl27APjTn/7E6aef7vPnGejDrAasUZyIBAH/D/gPH/a9VUTWi8j6nmxAoKpsyS8l0zqUMcYMUEuWLGHz5s1NCT0zM5OZM2cyceJErrnmGk4++eQOj581axZXXXUVmZmZnHfeecyZM6dp289+9jPmzZvHySefzMSJE5vWX3311fzyl79k5syZfPnll03rIyIi+MMf/sAVV1zBtGnTCAoK4hvf+IZPn8OGWfXj8KkichJwn6p+1V3+IYCqPuguxwFfAhXuIcOBImBRRw3jenL41JwjR1n48Ds8eOk0lsxN75FzGmOML2z41IGps2FWj2f4VH/eQ18HjBORDKAAuBq4pnGjqpYCSY3LIvIO8N0T2cq9uUGcldCNMcb410MPPcTjjz/e4/fOG/mtyl1V64HbgdeBz4EXVDVbRO4XkUUdH31ijBkSza2njWb8sJhAh2KMMaafu/vuu9m7dy+nnHKKX87v1+fQVXUVsKrVup+2s+9Cf8bSlqkpcdbdqzHGmH7BeoozxpgA8VcbJtM3He/3wRK6McYEQEREBIWFhZbUDeAk88LCQiIiIrp9Dr9WuRtjjGlbamoq+fn5PdqXt+nbIiIiSE1N7fbxltCNMSYAQkNDW3Qhaszxsip3Y4wxph+whG6MMcb0A5bQjTHGmH7Ab12/+ouIHAb29uApk4AjPXg+Y9e0p9n17Hl2TXueXdOe5X09R6pqpwOn97mE3tNEZL0vfeQa39k17Vl2PXueXdOeZ9e0Z3XnelqVuzHGGNMPWEI3xhhj+gFL6PBEoAPoh+ya9iy7nj3PrmnPs2vas7p8PQf8PXRjjDGmP7ASujHGGNMPDOiELiLnisgOEdklIncHOp7+QERyROQzEdkkIusDHU9fIyJPi8ghEdnqtS5RRP4lIl+4rwmBjLGvaeea3iciBe73dJOInB/IGPsSEUkTkbdFZJuIZIvIv7vr7XvaDR1czy5/RwdslbuIBAM7gXOAfGAdsERVtwU0sD5ORHKALFW151G7QUROAyqAZ1V1qrvuF0CRqj7k/vBMUNUfBDLOvqSda3ofUKGqDwcytr5IREYAI1R1o4jEABuAi4Gl2Pe0yzq4nlfSxe/oQC6hzwV2qepuVa0FlgOLAxyTGeBUdQ1Q1Gr1YuAZd/4ZnP/sxkftXFPTTaq6X1U3uvPlwOdACvY97ZYOrmeXDeSEngLkeS3n082LaFpQ4A0R2SAitwY6mH5imKrud+cPAMMCGUw/cruIbHGr5K16uBtEZBQwE1iLfU+PW6vrCV38jg7khG784xRVnQWcB3zLre40PUSde2QD8z5Zz3ocGAPMAPYD/xPYcPoeEYkG/gbcqapl3tvse9p1bVzPLn9HB3JCLwDSvJZT3XXmOKhqgft6CHgJ59aGOT4H3ftsjffbDgU4nj5PVQ+qaoOqeoDfY9/TLhGRUJzks0xV/+6utu9pN7V1PbvzHR3ICX0dME5EMkQkDLgaWBHgmPo0EYlyG3UgIlHAV4CtHR9lfLACuNGdvxH4RwBj6RcaE4/rEux76jMREeAp4HNV/X9em+x72g3tXc/ufEcHbCt3APcxgF8DwcDTqvpAgEPq00RkNE6pHCAE+LNd064RkeeBhTgjLR0E7gVeBl4A0nFGGrxSVa2Rl4/auaYLcaoyFcgBvu51/9d0QEROAd4DPgM87uof4dz3te9pF3VwPZfQxe/ogE7oxhhjTH8xkKvcjTHGmH7DEroxxhjTD1hCN8YYY/oBS+jGGGNMP2AJ3RhjjOkHLKEbY4wx/YAldGOMMaYfsIRujA9E5FURubHzPbu2byC5Y9ef7YfzviMiX3PnrxWRN3zZtxvvky4iFe5QyMYMeJbQTb/l/rFvnDwiUuW1fG1XzqWq56nqM53v2bV9eyMRuVtE1rSxPklEakVkqq/nUtVlqvqVHoqrxQ8QVc1V1WhVbeiJ87d6LxWRsT19XmP8yRK66bfcP/bRqhoN5AIXea1b1rifiIQELspe6TlggYhktFp/NfCZqlq/58b0QpbQzYAjIgtFJF9EfiAiB4A/iEiCiPxTRA6LSLE7n+p1jHc18lIReV9EHnb33SMi53Vz3wwRWSMi5SLypog8JiLPtRO3LzH+TEQ+cM/3hogkeW2/XkT2ikihiPy4veujqvnAauD6VptuAJ7tLI5WMS8Vkfe9ls8Rke0iUioijwLitW2MiKx24zsiIstEJN7d9iecPsJXujUs3xeRUW5JOsTdJ1lEVohIkYjsEpFbvM59n4i8ICLPutcmW0Sy2rsG7RGROPcch91reY+IBLnbxorIu+5nOyIif3HXi4j8SkQOiUiZiHzWlVoOY3xlCd0MVMOBRGAkcCvO/4U/uMvpQBXwaAfHzwN24Az48QvgKRGRbuz7Z+ATYDBwH8cmUW++xHgNcBMwFAgDvgsgIpNxxle+Hkh236/NJOx6xjsWEZmAM1DEn32M4xjuj4u/A/fgXIsvgZO9dwEedOObhDO88X0Aqno9LWtZftHGWywH8t3jLwf+S0TO9Nq+yN0nHmdksE5jbsNvgThgNHA6zo+cm9xtPwPeABJwru1v3fVfAU4DfCHBywAAIABJREFUxrvHXgkUduO9jemQJXQzUHmAe1W1RlWrVLVQVf+mqpWqWg48gPMHuz17VfX37v3bZ4ARwLCu7Csi6cAc4KeqWquq79PBEL4+xvgHVd2pqlU4I1/NcNdfDvxTVdeoag3wE5pHdmrLS26MC9zlG4BXVfVwN65Vo/OBbFV9UVXrcEY6POD1+Xap6r/+f3v3HR9Vme9x/PNLhxRa6KH3ElooFtRgFxREdBXhruhar2WbZd2rK+p6dVd3dV3boq7uellRsSwqihURC1UIEnoIEAglARIghLTn/jFDNmJCJskMk0y+79crr8ycOefMbw6j3zzPOed5vP8me4A/+7hfzKwTnj8O7nLOFTrnVgAveOs+aqFzbq733+EVYLAv+67wHuF4Tjvc7Zw74JzLBP7Ef/7wKcbzR04Hbw0LKyyPB/rimRBrjWZ2k0BQoEtjtcc5V3j0iZk1NbO/ebtR84EFQHOr+grqikFU4H0YV8N1OwB7KywD2FZVwT7WuLPC44IKNXWouG/n3CGO00r01vQG8FNvb8IU4J81qKMyx9bgKj43s7ZmNsvMtnv3+394WvK+OHosD1RYtgXoWOH5sccmxmp2/UQiEOndb2XvcSeeXobF3i79awCcc5/h6Q14GthtZjPMLKEG7yviEwW6NFbHzhv8a6APMMo5l4CnixQqnOMNgGygpZk1rbCs03HWr0uN2RX37X3PVtVs8w883cPn4GlhvlvHOo6twfjh5/1fPP8uyd79Tj1mn8eb63kHnmMZX2FZZ2B7NTXVRA7/aYX/6D2cczudc9c55zoANwDPmPdKeefck865FKA/nq73O/xYlwigQBc5Kh7PueD9ZtYSuC/Qb+ic2wIsBaabWZSZnQxcFKAaZwMXmtloM4sCHqD6//6/BPYDM4BZzrmiOtbxPjDAzC7xtoxvw3Mtw1HxwEEgz8w68uPQ24Xn3PWPOOe2AV8DD5tZjJkNAn6Gp5VfW1HefcWYWYx32evAQ2YWb2ZdgF8dfQ8zu6zCxYH78PwBUmZmI8xslJlFAoeAQo5/ukOkVhToIh5PAE3wtMK+BT48Qe87BTgZT/f374HXgCNVrFvrGp1zq4Gb8VzUlo0ncLKq2cbh6Wbv4v1dpzqccznAZcAjeD5vL+CrCqvcDwwD8vCE/1vH7OJh4B4z229mt1fyFpOBrnha62/juUbiE19qq8JqPH+4HP25GrgVTyhnAAvxHM+/e9cfASwys4N4roX4uXMuA0gAnsdzzLfg+eyP1qEukUqZ579ZEakPvLc6rXXOBbyHQERCi1roIkHk7Y7tYWZhZnY+MAF4J9h1iUjDE7BAN7O/ewdSqHRUKe9gC096B4BIM7NhgapFpB5rB8zHc+74SeAm59x3Qa1IRBqkgHW5m9npeP4n9U/n3I9GRTKzsXjOR43FM/DGX5xzowJSjIiISIgLWAvdObcA2HucVSbgCXvnnPsWz32s7QNVj4iISCgL5jn0jvxwEI0sfjgIhIiIiPioQcwyZWbX4xlvm9jY2JS+ffsGuaK6W7fzANERYXRNjA12KSIiUo8tW7YsxznXurr1ghno2/nhKFFJVDGqk3NuBp7BLRg+fLhbunRp4KsLsIfeT+flrzP5/N5ziI+JDHY5IiJST5nZlurXCm6X+xy840Sb2UlAXmOasODcAe0oLnXMX7cn2KWIiEgICORta68C3wB9zDP39M/M7EYzu9G7ylw8oy1txDOK0n8Hqpb6aFjnFrSKjeKj9F3BLkVEREJAwLrcnXOTq3nd4RmKslEKDzPO6d+W99KyOVJSSnREdRNViYiIVK1BXBQXqs4d0JZZS7bxzaZcUvu0CXY5IhJCiouLycrKorCwsPqVpV6IiYkhKSmJyMjaXVelQA+iU3okEhsVzkfpuxToIuJXWVlZxMfH07VrVzwz1Up95pwjNzeXrKwsunXrVqt9aCz3IIqJDCe1Txs+Tt9FWZkmyRER/yksLKRVq1YK8wbCzGjVqlWdelQU6EF27oC27DlwhO+27Q92KSISYhTmDUtd/70U6EGW2qcNEWHGR+k7g12KiIjf5ObmMmTIEIYMGUK7du3o2LFj+fOioqLjbrt06VJuu+22at/jlFNO8Uut8+fP58ILL/TLvoJJ59CDrFmTSE7u0Yq5q7L59Tl9iIrQ31gi0vC1atWKFStWADB9+nTi4uK4/fbby18vKSkhIqLyCBo+fDjDhw+v9j2+/vpr/xQbIpQe9cBVJ3dl297DPDt/U7BLEREJmGnTpnHjjTcyatQo7rzzThYvXszJJ5/M0KFDOeWUU1i3bh3wwxbz9OnTueaaa0hNTaV79+48+eST5fuLi4srXz81NZVLL72Uvn37MmXKFI7OJDp37lz69u1LSkoKt912W41a4q+++irJyckMHDiQu+66C4DS0lKmTZvGwIEDSU5O5vHHHwfgySefpH///gwaNIgrrrii7gerFtRCrwfO7t+WCUM68NfPNnBO/7b075AQ7JJERAIiKyuLr7/+mvDwcPLz8/nyyy+JiIjgk08+4be//S1vvvnmj7ZZu3Ytn3/+OQcOHKBPnz7cdNNNP7q167vvvmP16tV06NCBU089la+++orhw4dzww03sGDBArp168bkyccdHuUHduzYwV133cWyZcto0aIF5557Lu+88w6dOnVi+/btfP/99wDs3++5/umRRx5h8+bNREdHly870RTo9cT0iwbw1cYc7pi9knduPpXIcHWeiIh/3P/uatJ35Pt1n/07JHDfRQNqvN1ll11GeLhnIK28vDyuuuoqNmzYgJlRXFxc6Tbjxo0jOjqa6Oho2rRpw65du0hKSvrBOiNHjixfNmTIEDIzM4mLi6N79+7lt4FNnjyZGTNm+FTnkiVLSE1NpXVrz5woU6ZMYcGCBdx7771kZGRw6623Mm7cOM4991wABg0axJQpU7j44ou5+OKLa3xc/EGpUU+0iI3i9xcPZPWOfJ5T17uIhKjY2P/MMHnvvfcyZswYvv/+e959990qb9mKjo4ufxweHk5JSUmt1vGHFi1asHLlSlJTU3nuuee49tprAXj//fe5+eabWb58OSNGjAjY+x+PWuj1yPkD23PhoPY8+dkGzh3Qjj7t4oNdkoiEgNq0pE+EvLw8OnbsCMDLL7/s9/336dOHjIwMMjMz6dq1K6+99prP244cOZLbbruNnJwcWrRowauvvsqtt95KTk4OUVFRTJo0iT59+jB16lTKysrYtm0bY8aMYfTo0cyaNYuDBw/SvHlzv3+m41ELvZ65f/wAEmIiuf2NlZSUlgW7HBGRgLnzzju5++67GTp0aEBatE2aNOGZZ57h/PPPJyUlhfj4eJo1a1bpup9++ilJSUnlP5mZmTzyyCOMGTOGwYMHk5KSwoQJE9i+fTupqakMGTKEqVOn8vDDD1NaWsrUqVNJTk5m6NCh3HbbbSc8zAHs6JWADUWozId+PHNXZfPfM5dzx3l9uHlMz2CXIyIN0Jo1a+jXr1+wywi6gwcPEhcXh3OOm2++mV69evHLX/4y2GVVqbJ/NzNb5pyr9j4+tdDrobHJ7RmX3J6/fLKB9bsOBLscEZEG6/nnn2fIkCEMGDCAvLw8brjhhmCXFDAK9Hrq/gkDiIuJ4A51vYuI1Novf/lLVqxYQXp6OjNnzqRp06bBLilgFOj1VGJcNPePH8DKrDxeWLg52OWIiEg9p0Cvxy4c1J7zB7Tjzx+vZ+Nudb2LiEjVFOj1mJnx4MUDaRoVzu1vpFGqKVZFRKQKCvR6rnW8p+t9xbb9vLgwI9jliIhIPaVAbwDGD+7Auf3b8thH69m4+2CwyxERqdaYMWOYN2/eD5Y98cQT3HTTTVVuk5qaytHbkseOHVvpmOjTp0/nscceO+57v/POO6Snp5c//93vfscnn3xSk/IrVd+nWVWgNwBmxu8nDqRJZDh3zl6prncRqfcmT57MrFmzfrBs1qxZPk+QMnfu3FoPznJsoD/wwAOcffbZtdpXQ6JAbyDaxMcwfXx/lm/dz0tf6ap3EanfLr30Ut5//32KiooAyMzMZMeOHZx22mncdNNNDB8+nAEDBnDfffdVun3Xrl3JyckB4KGHHqJ3796MHj26fIpV8NxjPmLECAYPHsykSZMoKCjg66+/Zs6cOdxxxx0MGTKETZs2MW3aNGbPng14RoQbOnQoycnJXHPNNRw5cqT8/e677z6GDRtGcnIya9eu9fmz1pdpVhXoDcjFQzpydr82PDpvHZtzDgW7HBGRKrVs2ZKRI0fywQcfAJ7W+U9+8hPMjIceeoilS5eSlpbGF198QVpaWpX7WbZsGbNmzWLFihXMnTuXJUuWlL92ySWXsGTJElauXEm/fv148cUXOeWUUxg/fjyPPvooK1asoEePHuXrFxYWMm3aNF577TVWrVpFSUkJzz77bPnriYmJLF++nJtuuqnabv2jjk6z+tlnn7FixQqWLFnCO++8w4oVK8qnWV21ahVXX3014Jlm9bvvviMtLY3nnnuuRse0OpqcpQExMx6amMw5f/6CO95YyWs3nEx4mAW7LBGp7z74Dexc5d99tkuGCx457ipHu90nTJjArFmzePHFFwF4/fXXmTFjBiUlJWRnZ5Oens6gQYMq3ceXX37JxIkTyweEGT9+fPlr33//Pffccw/79+/n4MGDnHfeecetZ926dXTr1o3evXsDcNVVV/H000/zi1/8AvD8gQCQkpLCW2+95cNBqF/TrKqF3sC0TYjhvosGsHTLPv7xdWawyxERqdKECRP49NNPWb58OQUFBaSkpLB582Yee+wxPv30U9LS0hg3blyV06ZWZ9q0aTz11FOsWrWK++67r9b7OeroFKz+mH41GNOsqoXeAF0yrCPvpe3gj/PWcmbfNnRNjK1+IxFpvKppSQdKXFwcY8aM4Zprrim/GC4/P5/Y2FiaNWvGrl27+OCDD0hNTa1yH6effjrTpk3j7rvvpqSkhHfffbd8PPYDBw7Qvn17iouLmTlzZvlUrPHx8Rw48OPBuPr06UNmZiYbN26kZ8+evPLKK5xxxhl1+oz1aZpVBXoDZGY8fMkgznn8C+58M41Z151EmLreRaQemjx5MhMnTiy/4n3w4MEMHTqUvn370qlTJ0499dTjbj9s2DAuv/xyBg8eTJs2bRgxYkT5aw8++CCjRo2idevWjBo1qjzEr7jiCq677jqefPLJ8ovhAGJiYnjppZe47LLLKCkpYcSIEdx44401+jxHp1k96o033iifZtU5x7hx45gwYQIrV67k6quvpqzMMxdHxWlW8/LycM75fZpVTZ/agL2+dBt3zk7j/vEDuOqUrsEuR0TqEU2f2jBp+tRG6rKUJM7o3ZpHPljL1tyCYJcjIiJBpEBvwDxd78lEhBl3vrmSMg04IyLSaCnQG7gOzZtwz4X9+DZjLzMXbQl2OSIiEiQBDXQzO9/M1pnZRjP7TSWvdzazz83sOzNLM7OxgawnVP1keCdO65XIw+p6F5EKGto1Uo1dXf+9AhboZhYOPA1cAPQHJptZ/2NWuwd43Tk3FLgCeCZQ9YQyM+MPkwYRbsbts9X1LiKeK7pzc3MV6g2Ec47c3FxiYmJqvY9A3rY2EtjonMsAMLNZwAQgvcI6DkjwPm4G7AhgPSGtQ/Mm/O6i/twxO42Xvs7kZ6O7BbskEQmipKQksrKy2LNnT7BLER/FxMT84Ja4mgpkoHcEtlV4ngWMOmad6cBHZnYrEAuE/nQ4AXRpShIffr+TP364ltQ+renROi7YJYlIkERGRtKtm/6wb0yCfVHcZOBl51wSMBZ4xcx+VJOZXW9mS81sqf7arNrRq95jIsP59esrKSktC3ZJIiJyggQy0LcDnSo8T/Iuq+hnwOsAzrlvgBgg8dgdOedmOOeGO+eGHx0AXyrXJiGGBy8eyIpt+5nxZUawyxERkRMkkIG+BOhlZt3MLArPRW9zjllnK3AWgJn1wxPoaoLX0UWD2jM2uR2Pf7yetTvzg12OiIicAAELdOdcCXALMA9Yg+dq9tVm9oCZHZ3/7tfAdWa2EngVmOZ0SWadmRkPThhIQkwkv359JcXqehcRCXkayz2EzVu9kxteWcbPz+rFL8/pHexyRESkFjSWu3DegHZMHNqRpz/fyKqsvGCXIyIiAaRAD3HTLxpAq7gofv3GCo6UlAa7HBERCRAFeohr1jSSRyYNYv2ugzz+8YZglyMiIgGiQG8ExvRpwxUjOjFjwSaWbdkX7HJERCQAFOiNxP+M60f7Zk24/Y2VHC5S17uISKhRoDcS8TGRPHrpIDbnHOKP89YGuxwREfEzBXojckrPRK46uQsvfZXJN5tyg12OiIj4kQK9kbnrgr50bdWUO2av5OCRkmCXIyIifqJAb2SaRkXw2GWD2b7/MP87d02wyxERET9RoDdCw7u25LrTuvOvRVv5Yr2GzhcRCQUK9EbqV+f0pmebOO6anUbe4eJglyMiInWkQG+kYiLD+dNlg9lz8AgPvJse7HJERKSOFOiN2OBOzfnv1B68uTyLj9N3BbscERGpAwV6I3frmb3o1z6Bu99axb5DRcEuR0REakmB3shFRYTxp8sGk3e4iHv//X2wyxERkVpSoAv9OyRw25m9eC8tm/fSdgS7HBERqQUFugBwU2oPBndqzh1vpLE0c2+wyxERkRpSoAsAEeFhPP/TFNo1i+Hql5bw/fa8YJckIiI1oECXcm3iY5h57SgSmkTyXy8uYv2uA8EuSUREfKRAlx/o0LwJM68dRWR4GFNeWERmzqFglyQiIj5QoMuPdE2MZea1oygpLWPKC4vYvv9wsEsSEZFqVBvoZnarmbU4EcVI/dGrbTyv/GwU+YXFTHn+W3bnFwa7JBEROQ5fWuhtgSVm9rqZnW9mFuiipH4Y2LEZL189gt0HjjD1xUUaeEZEpB6rNtCdc/cAvYAXgWnABjP7XzPrEeDapB5I6dKSF346nMzcAn7698XkF2oiFxGR+sinc+jOOQfs9P6UAC2A2Wb2xwDWJvXEKT0TeW7qMNZk53PNS0soKCoJdkkiInIMX86h/9zMlgF/BL4Ckp1zNwEpwKQA1yf1xJl92/KXK4ayfOs+rv/nMgqLS4NdkoiIVOBLC70lcIlz7jzn3BvOuWIA51wZcGFAq5N6Zdyg9vzx0sEs3JjDLf9aTnFpWbBLEhERL1/Ood8HtDKz27xXvA+r8NqagFYn9c6lKUk8OGEAn6zZzS9fW0FpmQt2SSIigm9d7vcC/wBaAYnAS2Z2T6ALk/rrv07uyt0X9OW9tGx+82YaZQp1EZGgi/BhnanAYOdcIYCZPQKsAH4fyMKkfrvhjB4UFJXyl083EBsdwX0X9Ud3NIqIBI8v59B3ADEVnkcD233Zufe+9XVmttHMflPFOj8xs3QzW21m//Jlv1I//OLsXlx3Wjde/jqTR+etC3Y5IiKNmi8t9DxgtZl9DDjgHGCxmT0J4Jy7rbKNzCwceNq7fhaewWnmOOfSK6zTC7gbONU5t8/M2tTp08gJZWb8dmw/CopKeWb+JmKjI7h5TM9glyUi0ij5Euhve3+Omu/jvkcCG51zGQBmNguYAKRXWOc64Gnn3D4A59xuH/ct9YSZ8eCEgRwuKuXReetoEhnONaO7BbssEZFGp9pAd879w8yigN7eReuO3rpWjY7AtgrPs4BRx6zTG8DMvgLCgenOuQ992LfUI2Fhxh8vHURBUSkPvJfO4eJSrj+9O5HhmvtHRORE8eUq91RgA57u82eA9WZ2up/ePwLPsLKpwGTgeTNrXkkN15vZUjNbumfPHj+9tfhTRHgYT04eygUD2/HovHWc/8QC5q9Th4uIyIniSxPqT8C5zrkznHOnA+cBj/uw3XagU4XnSfz4YrosYI5zrtg5txlYjyfgf8A5N8M5N9w5N7x169Y+vLUEQ1REGM9MGcaLVw2ntMwx7aUlXPPyEjL2HAx2aSIiIc+XQI90zpVfwuycWw9E+rDdEqCXmXXzdtlfAcw5Zp138LTOMbNEPF3wGT7sW+opM+Osfm356Jdn8NuxfVm8eS/nPbGAh95P18QuIiIB5EugLzOzF8ws1fvzPLC0uo2ccyXALcA8YA3wunNutZk9YGbjvavNA3LNLB34HLjDOZdbu48i9UlURBjXn96Dz29P5ZKhSbywcDNjHp3PrMVbNbqciEgAmGciteOsYBYN3AyM9i76EnjGOXckwLVVavjw4W7p0mr/npB6ZlVWHve/u5qlW/YxsGMC9100gBFdWwa7LBGRes/Mljnnhle73vEC3Xsv+WrnXF9/FlcXCvSGyznHu2nZPDx3Ddl5hVw0uAO/uaAvHZs3CXZpIiL1lq+Bftwud+dcKbDOzDr7rTJptMyM8YM78NmvU/n5Wb34aPVOzvrTfJ74ZD2HizQdq4hIXfjS5b4AGAosBg4dXe6cG1/lRgGkFnroyNpXwCMfrOW9tGw6NIvh7rH9uHBQe40JLyJSgV+63L07OqOy5c65L2pZW50o0EPP4s17uf/d1azekc+Iri2476IBDOzYLNhliYjUC37pcvca65z7ouIPMLbuJYp4jOzWkjm3jOaRS5LJ2HOIi55ayF2z09idXxjs0kREGgxfAv2cSpZd4O9CpHELDzOuGNmZz+9I5drR3XjruyxSH5vPXz/doPPrIiI+qDLQzewmM1sF9DGztAo/m4FVJ65EaUwSYiL5n3H9+eRXZ3BG79b86eP1nPmn+by1PIsy3b8uIlKlKs+hm1kzoAXwMFBxLvMDzrm9J6C2SukceuOyePNefv9+OmlZeSR3bMY94/oxqnurYJclInLC+O2iOO/OwoG2VJidzTm3tU4V1pICvfEpK3P8e+V2/vjhOrLzCjl/QDt+c0FfuibGBrs0EZGA8zXQq50+1cxuAaYDu4Ay72IHDKpLgSK+CgszJg5N4vwB7XlxYQbPzN/Ep2t3cdXJXbn1zF40a+rL1AIiIqHNl9vWNgKj6ssY62qhy+78Qv788XpeW7qNZk0i+flZvZh6UhfNvy4iIcmft61tA/LqXpKIf7RJiOGRSYN4/9bTGNAhgfvfTee8xxfwcfoufDmFJCISinxpob8I9AHeB8onZHHO/TmwpVVOLXSpyDnH5+t289D7a9i05xAnd2/F/4zrp4FpRCRk+LOFvhX4GIgC4iv8iASdmXFm37Z8+IvTeXDCANbuzOeipxZy+xsryTkYlAkBRUSCwqer3H+0kVmEd77zE04tdDmevMPFPPP5Rl76KpPmTSN5esowTdMqIg1anVvoZrawwuNXjnl5cR1qEwmYZk0iuXtsP/59y6k0jQrnihnf8vyCDJ1bF5GQd7wu94o3+Q485jVNhyX1Wr/2Ccy5dTTn9GvLQ3PXcOP/LSO/sDjYZYmIBMzxAt1V8biy5yL1TkJMJM9OHcY94/rxyZrdjP/rQtJ35Ae7LBGRgDheoDc3s4lmNsn7+BLvzyRAlxBLg2BmXHtad2ZdfxKHi0uZ+MxXvLF0W7DLEhHxu+ON5f7S8TZ0zl0dkIqqoYvipLb2HDjCz2d9x9ebcrl8eCfunzCAmMjwYJclInJcdR76NViBLRIoreOjeeVno3j84/U89flGVm3P49mpw+jSSmPCi0jDp7EypVEJDzNuP68Pf582nO37D3PhXxcyb/XOYJclIlJnCnRplM7s25b3bh1N11ax3PDKMh6eu4aS0rLqNxQRqacU6NJodWrZlDduPJkpozrztwUZXPn8InbnFwa7LBGRWqk20M3sMjOL9z6+x8zeMrNhgS9NJPBiIsN5aGIyj18+mFXb8xj75EK+2VQvJhYUEakRX1ro9zrnDpjZaOBs4EXg2cCWJXJiTRyaxDs3n0pCkwimvPAtz87fRFmZhlsQkYbDl0Av9f4eB8xwzr2PZ6IWkZDSp108c24ZzQXJ7fnDh2u5/pWl7MxTF7yINAy+BPp2M/sbcDkw18yifdxOpMGJi47gqclDmX5Rf75Yv4fRf/iMX72+gjXZGmFOROo3X+ZDbwqcD6xyzm0ws/ZAsnPuoxNR4LE0sIycKNv2FvDiws28vnQbBUWlnNYrkWtP687pvRIx03QGInJi+DqwjC+B3gPIcs4dMbNUYBDwT+fcfr9UWkMKdDnR8gqKmbl4Cy9/lcnuA0fo2y6en43uxvghHYiO0EhzIhJY/gz0FcBwoCswF/g3MMA5N9YPddaYAl2CpaikjDkrd/D8ggzW7TpAm/horjqlK1NHdaFZ08hglyciIarO86FXUOacKwEuAf7qnLsDaO9jEeeb2Toz22hmvznOepPMzJlZtQWLBEtURBiXpiTx4S9O45/XjKRPu3genbeOkx/5lOlzVrM1tyDYJYpII1blWO4VFJvZZOCnwEXeZdU2R8wsHHgaOAfIApaY2RznXPox68UDPwcW1aRwkWAxM07v3ZrTe7cmfUc+LyzMYOaiLfzzm0wuGNiea0/rxtDOLYJdpog0Mr600K8GTgYecs5tNrNuwCs+bDcS2Oicy3DOFQGzgAmVrPcg8AdA9wdJg9O/QwJ//skQvrzzTK4/vQcLNuxh4jNfc9lzXzNv9U5KdS+7iJwg1Qa6t0V9O7DKzAbiuUDuDz7suyNQceLpLO+yct4R5zp5720XabDaNYvhNxf05Zu7z+J3F/YnO6+QG15Zxtl//oLXl26jWOPEi0iA+TL0ayqwAU/3+TPAejM7va5vbGZhwJ+BX/uw7vVmttTMlu7Zs6euby0SMHHREVwzuhvzb0/lqSuHEhsdzp2z0zjrTwp2EQksX65yXwZc6Zxb533eG3jVOZdSzXYnA9Odc+d5n98N4Jx72Pu8GbAJOOjdpB2wFxjvnKvyMnZd5S4NiXOOT9fs5olP1/P99nw6t2zKLWf2ZOLQjkSGa3wmEameP69yjzwa5gDOufX4cFEcsAToZWbdzCwKuAKYU2E/ec4LSXhmAAASpklEQVS5ROdcV+dcV+BbqglzkYbGzDi7f1vevWU0L/x0OAlNItRiF5GA8CXQl5nZC2aW6v15Hqg2dL23ut0CzAPWAK8751ab2QNmNr5uZYs0LAp2EQk0X7rco4GbgdHeRV8CzzjnjgS4tkqpy11CgbriRcRXfhkpznsv+WrnXF9/FlcXCnQJJQp2EamOX86hO+dKgXVm1tlvlYlIOXXFi4i/+NLlvgAYCiwGDh1d7pwLynlwtdAllFXaYh/Tk/FDOhATqYlgRBojf07OckZly51zX9SytjpRoEtjcGywx0dHcOHg9kwalkRKlxaavlWkEfE10Kscy93MegJtjw1uMxsNZNe9RBGpytGu+LP6teGbjFzeXLadd77bwauLt9EtMZZLhnbkkpQkOjZvEuxSRaSeqLKFbmbvAXc751YdszwZ+F/n3EWVbhhgaqFLY3XwSAkfrMrmzeVZfJuxFzM4uXsrLk1J4vyB7Wga5ctcSyLS0NS5y93MljjnRlTx2irnXHIda6wVBboIbNtbwFvLt/Pm8iy27i0gNiqcscntmZSSxMiuLQkLU5e8SKjwR6BvcM71quK1jc65nnWssVYU6CL/4ZxjSeY+Zi/bxvtp2RwqKqVTyyZcMjSJScOS6NyqabBLFJE68kegvwp85px7/pjl1wLnOOcu90ulNaRAF6lcQVEJ81bv5M1l2/lqUw7OwchuLbk0JYmxye2Ji1aXvEhD5I9Abwu8DRQBy7yLhwNRwETn3E4/1VojCnSR6u3Yf5i3v9vO7GVZbM45RExkGGf1bcu4Qe0Z06cNTaJ0C5xIQ+HP29bGAAO9T1c75z7zQ321pkAX8Z1zjuVb9/P2d1l8+P1Ocg4W0TQqnLP6tWVccntS+7TW/e0i9ZzfAr2+UaCL1E5pmWNRRi7vrcrmw+93svdQEbFR4ZzTvy3jBnXg9N6JREco3EXqGwW6iFSppLSMbzJyeT8tmw9X72R/QTHx0RGcM6AtFw5qz+ierYmK0FjyIvWBAl1EfFJcWsZXG3N4Py2beat3kl9YQkJMBOcNaMe4Qe05tWeiJooRCSIFuojUWFFJGQs37uG9tGw+Xr2LA0dKaN40kvMHtGP84A6c3KOVhp0VOcEU6CJSJ0dKSlmwPof303bwcfouDhWVMrBjAjen9uS8Ae00eI3ICaJAFxG/KSwu5d8rtvPs/E1k5hbQo3Us/53qmQVO3fEigaVAFxG/Ky1zzF2VzdOfb2TtzgN0bN6EG8/ozmXDO+n2N5EAUaCLSMA45/hs7W6e+nwj323dT2JcNNee1o2pJ3XRiHQifqZAF5GAc87xbcZenpm/kS835JAQE8G0U7tx9SldaREbFezyREKCAl1ETqiV2/bzzPyNzFu9i6ZR4Vw5sjPXnd6dtgkxwS5NpEFToItIUKzfdYBn529izsodhJsxKSWJm87ooZnfRGpJgS4iQbU1t4DnFmxi9tIsSsrKGD+4A5NSkgg3o7jMUVJaRnGpo6SsjJJSR3FpGaVlrvy1klJHsfe1kqPLyhzNm0YysmtLBiU112h20igo0EWkXtiVX8gLX2Ywc9FWCopKa72fiDAjItwoLC4DICYyjGGdWzCqWytGdW/JkE7NdaW9hCQFuojUK/sLili9I5/wMCMy3IgICyMi3IgMD/Ms8z6PCK/wuPy3lY9Ql3vwCEsy9/Jtxl4Wbd7L2p35OAdREWEM6dSck7q1ZFT3Vgzr3ELTxEpIUKCLSKOQV1DM4sy9LMrIZdHmvazekUeZg8hwY1BSc0Z5Az6lSwvdUicNkgJdRBql/MJilmXuY9HmvSzanMuqrDxKyhzhYcbADgmM6t6KM/u2YVS3lhqXXhoEBbqICHDoSAnLt+5jUYYn4Fds209xqaN761iuHNmZS1OSaN5U98xL/aVAFxGpxOGiUuauymbmoi0s37qfqIgwxiW3Z8qozqR0aaFWu9Q7CnQRkWqsyc7nX4u28vZ32zl4pITebeO4cmRnJg5LolmTyGCXJwIo0EVEfHboSAnvrtzBvxZvJS0rj5jIMC4a1IErR3VmSKfmarVLUNWLQDez84G/AOHAC865R455/VfAtUAJsAe4xjm35Xj7VKCLSCCtysrjX4u38O8VOygoKqV/+wSuHNWZi4d21FXyEhRBD3QzCwfWA+cAWcASYLJzLr3COmOARc65AjO7CUh1zl1+vP0q0EXkRDhQWMy/V+xg5qKtrMnOJzYqnPFDOjJlVGcGdmwW7PKkEfE10AP55+ZIYKNzLsNb0CxgAlAe6M65zyus/y0wNYD1iIj4LD4mkqkndWHKqM6s2LafmYu28vZ3Wby6eCuDkppxSo9EEppEEB8TSUJMBAlNvL9jIkloEkl8TARNIsPVXS8nTCADvSOwrcLzLGDUcdb/GfBBZS+Y2fXA9QCdO3f2V30iItUyM4Z2bsHQzi2498L+vL08i1lLtvHiwgyKS4/fwxkRZuXh7gn6COKjPb8TYiLp3S6e03ol0r5ZkxP0aSSU1YsTQmY2FRgOnFHZ6865GcAM8HS5n8DSRETKNWsSybRTuzHt1G445zhSUkb+4WLyC0vILywm/3AxB8ofl3CgsJj8Qu8y73p7Dhwk/3AJ+w8XlY9L36N1LKf1as3onomM6t6S+BhdYS81F8hA3w50qvA8ybvsB8zsbOB/gDOcc0cCWI+IiN+YGTGR4cREhtMmoebbO+dYu/MACzfk8OXGHGYt2crLX2cSEWYM6dSc0b0SOa1XIoOSmhMZrlnlpHqBvCguAs9FcWfhCfIlwJXOudUV1hkKzAbOd85t8GW/uihORELRkZJSlm3Zx8INOSzcmMOq7Xk4B3HREZzUvRWn9UpkdK9EuifG6rx8IxP0q9y9RYwFnsBz29rfnXMPmdkDwFLn3Bwz+wRIBrK9m2x1zo0/3j4V6CLSGOw7VMQ3Gbl8uSGHhRv3sG3vYQA6NIvh1J6ecD+1ZyKJcdFBrlQCrV4EeiAo0EWkMdqSe4gvN+Tw1UbPT35hCQAJMRG0iI2iRdMoWjSNrOJxFC1iI2nZNIrmTaOIilAXfkNSH25bExERP+nSKpYurWKZelIXSsscq7bn8c2mXHbmHWZfQTH7CorYc/AI63cdZF9BEQVFpVXuKy46guZNI2nRNIrmTSOJjggnIswI9849Hx529HcYEWFWPif90eflr3uXR4aH0aF5E7q2iqVzy6YnbB76g0dK2LznEBk5B9m05xAHC0sYm9yu0Y7Jr0AXEWlgwr0Xzg3p1LzKdQqLS9nvDfp9h4rYV1DM3oIi9nsf7yso8v4UU1xSRGmZo6SszPvb/fB36Y+XH0+7hBi6tGpK11axdEn0/m7VlC6tYms82l5JaRlZ+w6TkXOQjD2HyMg5RMYez+PdB/5zHbUZRIaH8fevNtOrTRxXjOzMJUM70iK28cykpy53ERGpEeccZQ6KvUFfWFzK9v2HycwtYEvOIc/vXM/vnIM/vHkpMS7K29vwn6Dv2iqWNgnRbN93mIw9h9jkDe/NOYfYknvoB/f7N28aSffEWLq3jqNbYiw9Wnsed27ZlNIyx3tpO3h18TZWbPPMpHfBwHZcMaIzJ3Vv2WBb7TqHLiIiQXfwSAlbcg+xNbegQtAfYktuAdl5hZVuExUeRpdWTene+pjgTozzucW9JjufWYu38tZ32zlQWEL3xFguH9GJSSlJDe5CQgW6iIjUa4XFpWzdW0Bmjqf7vGOLJvRIjKNjiyaEh/mnNX24qJS5q7KZtWQrSzL3ERlunNu/HVeM7MSpPRIJ89P7BJICXUREpIINuw4wa8k23lqexb6CYjq1bMIVIzpzWUoSbRJigl1elRToIiIilSgsLmXe6p3MWryNbzJyCQ8zzuzbhitHdub03q391jvgL7ptTUREpBIxkeFMGNKRCUM6sjnnELOWbOXNZVl8nL6LxLgokjs2o3+HBAZ0aEb/9gl0btm0YXTNq4UuIiKNXVFJGZ+u2cXH6btIz85nw+6DlHpvz4uLjqBf+3j6t/eGfIcEerWNIzrixNxvry53ERGRWiosLmXDroOs3pFHenY+q3fksyY7v3zAnogwo2ebuPKAH9AhgX7tE2jWxP8z5anLXUREpJZiIsNJTmpGclKz8mVlZY7M3EPlAZ++I58v1u/hzeVZ5et0atmEgR2a8dSVw074uXgFuoiIiA/CwozurePo3jqOCwd1KF+++0BhecCn78gnv7A4KBfWKdBFRETqoE18DG36xDCmT5ug1qEpd0REREKAAl1ERCQEKNBFRERCgAJdREQkBCjQRUREQoACXUREJAQo0EVEREKAAl1ERCQEKNBFRERCgAJdREQkBCjQRUREQoACXUREJAQo0EVEREKAAl1ERCQEKNBFRERCgAJdREQkBCjQRUREQkBAA93MzjezdWa20cx+U8nr0Wb2mvf1RWbWNZD1iIiIhKqABbqZhQNPAxcA/YHJZtb/mNV+BuxzzvUEHgf+EKh6REREQlkgW+gjgY3OuQznXBEwC5hwzDoTgH94H88GzjIzC2BNIiIiISmQgd4R2FbheZZ3WaXrOOdKgDygVQBrEhERCUkRwS7AF2Z2PXC99+lBM1vnx90nAjl+3J/omPqbjqf/6Zj6n46pf1U8nl182SCQgb4d6FTheZJ3WWXrZJlZBNAMyD12R865GcCMQBRpZkudc8MDse/GSsfUv3Q8/U/H1P90TP2rNsczkF3uS4BeZtbNzKKAK4A5x6wzB7jK+/hS4DPnnAtgTSIiIiEpYC1051yJmd0CzAPCgb8751ab2QPAUufcHOBF4BUz2wjsxRP6IiIiUkMBPYfunJsLzD1m2e8qPC4ELgtkDT4ISFd+I6dj6l86nv6nY+p/Oqb+VePjaerhFhERafg09KuIiEgIaNSBXt3QtFJzZpZpZqvMbIWZLQ12PQ2Nmf3dzHab2fcVlrU0s4/NbIP3d4tg1tjQVHFMp5vZdu/3dIWZjQ1mjQ2JmXUys8/NLN3MVpvZz73L9T2theMczxp/Rxttl7t3aNr1wDl4Br1ZAkx2zqUHtbAGzswygeHOOd2PWgtmdjpwEPinc26gd9kfgb3OuUe8f3i2cM7dFcw6G5Iqjul04KBz7rFg1tYQmVl7oL1zbrmZxQPLgIuBaeh7WmPHOZ4/oYbf0cbcQvdlaFqRE8o5twDPHR8VVRwi+R94/mMXH1VxTKWWnHPZzrnl3scHgDV4Rv3U97QWjnM8a6wxB7ovQ9NKzTngIzNb5h3hT+qurXMu2/t4J9A2mMWEkFvMLM3bJa/u4VrwzpA5FFiEvqd1dszxhBp+RxtzoEtgjHbODcMzy97N3u5O8RPvwEuN8zyZfz0L9ACGANnAn4JbTsNjZnHAm8AvnHP5FV/T97TmKjmeNf6ONuZA92VoWqkh59x27+/dwNt4Tm1I3ezynmc7er5td5DrafCcc7ucc6XOuTLgefQ9rREzi8QTPjOdc295F+t7WkuVHc/afEcbc6D7MjSt1ICZxXov6sDMYoFzge+Pv5X4oOIQyVcB/w5iLSHhaPB4TUTfU595p7h+EVjjnPtzhZf0Pa2Fqo5nbb6jjfYqdwDvbQBP8J+haR8KckkNmpl1x9MqB88ohP/SMa0ZM3sVSMUz09Iu4D7gHeB1oDOwBfiJc04XefmoimOaiqcr0wGZwA0Vzv/KcZjZaOBLYBVQ5l38WzznffU9raHjHM/J1PA72qgDXUREJFQ05i53ERGRkKFAFxERCQEKdBERkRCgQBcREQkBCnQREZEQoEAXEb8xs1Qzey/YdYg0Rgp0ERGREKBAF2mEzGyqmS32zrP8NzMLN7ODZva4d07mT82stXfdIWb2rXeSiLePThJhZj3N7BMzW2lmy82sh3f3cWY228zWmtlM70hYIhJgCnSRRsbM+gGXA6c654YApcAUIBZY6pwbAHyBZ0Q1gH8CdznnBuEZzero8pnA0865wcApeCaQAM9sUb8A+gPdgVMD/qFEhIhgFyAiJ9xZQAqwxNt4boJnIo0y4DXvOv8HvGVmzYDmzrkvvMv/AbzhHbO/o3PubQDnXCGAd3+LnXNZ3ucrgK7AwsB/LJHGTYEu0vgY8A/n3N0/WGh27zHr1XZc6CMVHpei/8+InBDqchdpfD4FLjWzNgBm1tLMuuD5/8Gl3nWuBBY65/KAfWZ2mnf5fwFfOOcOAFlmdrF3H9Fm1vSEfgoR+QH95SzSyDjn0s3sHuAjMwsDioGbgUPASO9ru/GcZwfPVJjPeQM7A7jau/y/gL+Z2QPefVx2Aj+GiBxDs62JCABmdtA5FxfsOkSkdtTlLiIiEgLUQhcREQkBaqGLiIiEAAW6iIhICFCgi4iIhAAFuoiISAhQoIuIiIQABbqIiEgI+H8cr0mafjwXVwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 576x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "acc = history.history['accuracy']\n",
    "val_acc = history.history['val_accuracy']\n",
    "\n",
    "loss = history.history['loss']\n",
    "val_loss = history.history['val_loss']\n",
    "\n",
    "plt.figure(figsize=(8, 8))\n",
    "plt.subplot(2, 1, 1)\n",
    "plt.plot(acc, label='Training Accuracy')\n",
    "plt.plot(val_acc, label='Validation Accuracy')\n",
    "plt.legend(loc='lower right')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.ylim([min(plt.ylim()),1])\n",
    "plt.title('Training and Validation Accuracy')\n",
    "\n",
    "plt.subplot(2, 1, 2)\n",
    "plt.plot(loss, label='Training Loss')\n",
    "plt.plot(val_loss, label='Validation Loss')\n",
    "plt.legend(loc='upper right')\n",
    "plt.ylabel('Cross Entropy')\n",
    "plt.ylim([0,1.0])\n",
    "plt.title('Training and Validation Loss')\n",
    "plt.xlabel('epoch')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "executionInfo": {
     "elapsed": 10219324,
     "status": "ok",
     "timestamp": 1607991136587,
     "user": {
      "displayName": "Mahmut Can Kurt",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GiR0Vz-Xi8c64gs5AldT12G3fhmeowAL59isdw9=s64",
      "userId": "17273613824855319138"
     },
     "user_tz": -180
    },
    "id": "729b4V-pGJJP"
   },
   "outputs": [],
   "source": [
    "base_model.trainable = True"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 10219320,
     "status": "ok",
     "timestamp": 1607991136588,
     "user": {
      "displayName": "Mahmut Can Kurt",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GiR0Vz-Xi8c64gs5AldT12G3fhmeowAL59isdw9=s64",
      "userId": "17273613824855319138"
     },
     "user_tz": -180
    },
    "id": "y6_MeLFcGJLy",
    "outputId": "bfd62cbf-6e04-47df-e6c9-1cc27b979f25"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of layers in the base model:  155\n"
     ]
    }
   ],
   "source": [
    "# Let's take a look to see how many layers are in the base model\n",
    "print(\"Number of layers in the base model: \", len(base_model.layers))\n",
    "\n",
    "# Fine tune from this layer onwards\n",
    "fine_tune_at = 100\n",
    "\n",
    "# Freeze all the layers before the `fine_tune_at` layer\n",
    "for layer in base_model.layers[:fine_tune_at]:\n",
    "  layer.trainable =  False"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "executionInfo": {
     "elapsed": 10219315,
     "status": "ok",
     "timestamp": 1607991136588,
     "user": {
      "displayName": "Mahmut Can Kurt",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GiR0Vz-Xi8c64gs5AldT12G3fhmeowAL59isdw9=s64",
      "userId": "17273613824855319138"
     },
     "user_tz": -180
    },
    "id": "k5GwINrXGJOR"
   },
   "outputs": [],
   "source": [
    "model.compile(loss='categorical_crossentropy',\n",
    "              optimizer = tf.keras.optimizers.Adam(1e-5),\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 10219309,
     "status": "ok",
     "timestamp": 1607991136589,
     "user": {
      "displayName": "Mahmut Can Kurt",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GiR0Vz-Xi8c64gs5AldT12G3fhmeowAL59isdw9=s64",
      "userId": "17273613824855319138"
     },
     "user_tz": -180
    },
    "id": "f5pDLR9LGJQk",
    "outputId": "16da7ac9-0e91-4fcb-e84f-f136239e7c35"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "mobilenetv2_1.00_224 (Functi (None, 7, 7, 1280)        2257984   \n",
      "_________________________________________________________________\n",
      "conv2d_1 (Conv2D)            (None, 5, 5, 32)          368672    \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 5, 5, 32)          0         \n",
      "_________________________________________________________________\n",
      "global_average_pooling2d_1 ( (None, 32)                0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 120)               3960      \n",
      "=================================================================\n",
      "Total params: 2,630,616\n",
      "Trainable params: 2,235,224\n",
      "Non-trainable params: 395,392\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 10219303,
     "status": "ok",
     "timestamp": 1607991136589,
     "user": {
      "displayName": "Mahmut Can Kurt",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GiR0Vz-Xi8c64gs5AldT12G3fhmeowAL59isdw9=s64",
      "userId": "17273613824855319138"
     },
     "user_tz": -180
    },
    "id": "B8RHIh6PGJTg",
    "outputId": "07f7aca2-3898-495d-c56b-de4665cc035a"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of trainable variables = 60\n"
     ]
    }
   ],
   "source": [
    "print('Number of trainable variables = {}'.format(len(model.trainable_variables)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 10707463,
     "status": "ok",
     "timestamp": 1607991624753,
     "user": {
      "displayName": "Mahmut Can Kurt",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GiR0Vz-Xi8c64gs5AldT12G3fhmeowAL59isdw9=s64",
      "userId": "17273613824855319138"
     },
     "user_tz": -180
    },
    "id": "g9ydUwUpGJVj",
    "outputId": "51cb60bd-4c2b-42c9-ce2e-02b2f6071049"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/5\n",
      "259/259 [==============================] - 97s 373ms/step - loss: 1.5929 - accuracy: 0.6611 - val_loss: 2.1902 - val_accuracy: 0.6613\n",
      "Epoch 2/5\n",
      "259/259 [==============================] - 97s 376ms/step - loss: 0.7944 - accuracy: 0.7735 - val_loss: 2.0721 - val_accuracy: 0.6625\n",
      "Epoch 3/5\n",
      "259/259 [==============================] - 96s 372ms/step - loss: 0.5816 - accuracy: 0.8181 - val_loss: 1.9699 - val_accuracy: 0.6640\n",
      "Epoch 4/5\n",
      "259/259 [==============================] - 96s 372ms/step - loss: 0.4514 - accuracy: 0.8542 - val_loss: 1.8915 - val_accuracy: 0.6632\n",
      "Epoch 5/5\n",
      "259/259 [==============================] - 97s 373ms/step - loss: 0.3667 - accuracy: 0.8776 - val_loss: 1.8548 - val_accuracy: 0.6674\n"
     ]
    }
   ],
   "source": [
    "history_fine = model.fit(train_generator, \n",
    "                         steps_per_epoch=len(train_generator), \n",
    "                         epochs=5, \n",
    "                         validation_data=val_generator, \n",
    "                         validation_steps=len(val_generator))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 27101,
     "status": "ok",
     "timestamp": 1607991908205,
     "user": {
      "displayName": "Mahmut Can Kurt",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GiR0Vz-Xi8c64gs5AldT12G3fhmeowAL59isdw9=s64",
      "userId": "17273613824855319138"
     },
     "user_tz": -180
    },
    "id": "gQvJaoCVGVib",
    "outputId": "0c6a9119-d10d-45de-d9e2-f74dfc75fa58"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: save/fine_tuning/assets\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Assets written to: save/fine_tuning/assets\n"
     ]
    }
   ],
   "source": [
    "saved_model_dir = 'save/fine_tuning'\n",
    "tf.saved_model.save(model, saved_model_dir)\n",
    "\n",
    "converter = tf.lite.TFLiteConverter.from_saved_model(saved_model_dir)\n",
    "tflite_model = converter.convert()\n",
    "\n",
    "with open('model.tflite', 'wb') as f:\n",
    "  f.write(tflite_model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 17
    },
    "executionInfo": {
     "elapsed": 1169,
     "status": "ok",
     "timestamp": 1607991988821,
     "user": {
      "displayName": "Mahmut Can Kurt",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GiR0Vz-Xi8c64gs5AldT12G3fhmeowAL59isdw9=s64",
      "userId": "17273613824855319138"
     },
     "user_tz": -180
    },
    "id": "nYBF4wUdyoN0",
    "outputId": "bc1148ce-4d07-4c7c-aba8-3d64b76c535e"
   },
   "outputs": [
    {
     "data": {
      "application/javascript": [
       "\n",
       "    async function download(id, filename, size) {\n",
       "      if (!google.colab.kernel.accessAllowed) {\n",
       "        return;\n",
       "      }\n",
       "      const div = document.createElement('div');\n",
       "      const label = document.createElement('label');\n",
       "      label.textContent = `Downloading \"${filename}\": `;\n",
       "      div.appendChild(label);\n",
       "      const progress = document.createElement('progress');\n",
       "      progress.max = size;\n",
       "      div.appendChild(progress);\n",
       "      document.body.appendChild(div);\n",
       "\n",
       "      const buffers = [];\n",
       "      let downloaded = 0;\n",
       "\n",
       "      const channel = await google.colab.kernel.comms.open(id);\n",
       "      // Send a message to notify the kernel that we're ready.\n",
       "      channel.send({})\n",
       "\n",
       "      for await (const message of channel.messages) {\n",
       "        // Send a message to notify the kernel that we're ready.\n",
       "        channel.send({})\n",
       "        if (message.buffers) {\n",
       "          for (const buffer of message.buffers) {\n",
       "            buffers.push(buffer);\n",
       "            downloaded += buffer.byteLength;\n",
       "            progress.value = downloaded;\n",
       "          }\n",
       "        }\n",
       "      }\n",
       "      const blob = new Blob(buffers, {type: 'application/binary'});\n",
       "      const a = document.createElement('a');\n",
       "      a.href = window.URL.createObjectURL(blob);\n",
       "      a.download = filename;\n",
       "      div.appendChild(a);\n",
       "      a.click();\n",
       "      div.remove();\n",
       "    }\n",
       "  "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "application/javascript": [
       "download(\"download_e2923780-27af-41bb-b07b-6aef779cfe4c\", \"model.tflite\", 10393928)"
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "application/javascript": [
       "\n",
       "    async function download(id, filename, size) {\n",
       "      if (!google.colab.kernel.accessAllowed) {\n",
       "        return;\n",
       "      }\n",
       "      const div = document.createElement('div');\n",
       "      const label = document.createElement('label');\n",
       "      label.textContent = `Downloading \"${filename}\": `;\n",
       "      div.appendChild(label);\n",
       "      const progress = document.createElement('progress');\n",
       "      progress.max = size;\n",
       "      div.appendChild(progress);\n",
       "      document.body.appendChild(div);\n",
       "\n",
       "      const buffers = [];\n",
       "      let downloaded = 0;\n",
       "\n",
       "      const channel = await google.colab.kernel.comms.open(id);\n",
       "      // Send a message to notify the kernel that we're ready.\n",
       "      channel.send({})\n",
       "\n",
       "      for await (const message of channel.messages) {\n",
       "        // Send a message to notify the kernel that we're ready.\n",
       "        channel.send({})\n",
       "        if (message.buffers) {\n",
       "          for (const buffer of message.buffers) {\n",
       "            buffers.push(buffer);\n",
       "            downloaded += buffer.byteLength;\n",
       "            progress.value = downloaded;\n",
       "          }\n",
       "        }\n",
       "      }\n",
       "      const blob = new Blob(buffers, {type: 'application/binary'});\n",
       "      const a = document.createElement('a');\n",
       "      a.href = window.URL.createObjectURL(blob);\n",
       "      a.download = filename;\n",
       "      div.appendChild(a);\n",
       "      a.click();\n",
       "      div.remove();\n",
       "    }\n",
       "  "
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "application/javascript": [
       "download(\"download_7769f1bc-8fdb-4dd2-b109-30fd75934063\", \"labels.txt\", 2910)"
      ],
      "text/plain": [
       "<IPython.core.display.Javascript object>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from google.colab import files\n",
    "\n",
    "files.download('model.tflite')\n",
    "files.download('labels.txt')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 513
    },
    "executionInfo": {
     "elapsed": 1076,
     "status": "ok",
     "timestamp": 1607992077044,
     "user": {
      "displayName": "Mahmut Can Kurt",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14GiR0Vz-Xi8c64gs5AldT12G3fhmeowAL59isdw9=s64",
      "userId": "17273613824855319138"
     },
     "user_tz": -180
    },
    "id": "56f_ncUPy9rJ",
    "outputId": "274735d3-e9ed-46ac-f8eb-988bbb8f42f1"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light",
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "acc = history_fine.history['accuracy']\n",
    "val_acc = history_fine.history['val_accuracy']\n",
    "\n",
    "loss = history_fine.history['loss']\n",
    "val_loss = history_fine.history['val_loss']\n",
    "\n",
    "plt.figure(figsize=(8, 8))\n",
    "plt.subplot(2, 1, 1)\n",
    "plt.plot(acc, label='Training Accuracy')\n",
    "plt.plot(val_acc, label='Validation Accuracy')\n",
    "plt.legend(loc='lower right')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.ylim([min(plt.ylim()),1])\n",
    "plt.title('Training and Validation Accuracy')\n",
    "\n",
    "plt.subplot(2, 1, 2)\n",
    "plt.plot(loss, label='Training Loss')\n",
    "plt.plot(val_loss, label='Validation Loss')\n",
    "plt.legend(loc='upper right')\n",
    "plt.ylabel('Cross Entropy')\n",
    "plt.ylim([0,1.0])\n",
    "plt.title('Training and Validation Loss')\n",
    "plt.xlabel('epoch')\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "authorship_tag": "ABX9TyOa9UB1kO8PXXZgMX16aZ5m",
   "name": "dog_breed_tf_lite.ipynb",
   "provenance": []
  },
  "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": "3.8.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}