9특성추출.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "83d1b5ea",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1000\n",
      "1000\n",
      "500\n"
     ]
    }
   ],
   "source": [
    "#디렉터리 생성 후 경로 변수만 필요햘 떄\n",
    "import os\n",
    "original_dataset_dir = '/Users/jw1010110/AI/dogs-vs-cats/train'\n",
    "base_dir = '/Users/jw1010110/AI/dogs-vs-cats/cats_and_dogs_small'\n",
    "\n",
    "train_dir = os.path.join(base_dir,'train')\n",
    "validation_dir = os.path.join(base_dir,'validation')\n",
    "test_dir = os.path.join(base_dir,'test')\n",
    "train_cats_dir = os.path.join(train_dir,'cats')\n",
    "train_dogs_dir = os.path.join(train_dir,'dogs')\n",
    "validation_cats_dir = os.path.join(validation_dir,'cats')\n",
    "validation_dogs_dir = os.path.join(validation_dir,'dogs')\n",
    "test_cats_dir = os.path.join(test_dir,'cats')\n",
    "test_dogs_dir = os.path.join(test_dir,'dogs')\n",
    "\n",
    "train_dir = os.path.join(base_dir,'train')\n",
    "validation_dir=os.path.join(base_dir,'validation')\n",
    "test_dir = os.path.join(base_dir,'test')\n",
    "\n",
    "print(len(os.listdir(train_dogs_dir)))\n",
    "print(len(os.listdir(train_cats_dir)))\n",
    "print(len(os.listdir(validation_dogs_dir)))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "29c50781",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Metal device set to: Apple M1 Pro\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-05-18 16:26:27.026884: I tensorflow/core/common_runtime/pluggable_device/pluggable_device_factory.cc:305] Could not identify NUMA node of platform GPU ID 0, defaulting to 0. Your kernel may not have been built with NUMA support.\n",
      "2022-05-18 16:26:27.027009: I tensorflow/core/common_runtime/pluggable_device/pluggable_device_factory.cc:271] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 0 MB memory) -> physical PluggableDevice (device: 0, name: METAL, pci bus id: <undefined>)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"vgg16\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " input_1 (InputLayer)        [(None, 150, 150, 3)]     0         \n",
      "                                                                 \n",
      " block1_conv1 (Conv2D)       (None, 150, 150, 64)      1792      \n",
      "                                                                 \n",
      " block1_conv2 (Conv2D)       (None, 150, 150, 64)      36928     \n",
      "                                                                 \n",
      " block1_pool (MaxPooling2D)  (None, 75, 75, 64)        0         \n",
      "                                                                 \n",
      " block2_conv1 (Conv2D)       (None, 75, 75, 128)       73856     \n",
      "                                                                 \n",
      " block2_conv2 (Conv2D)       (None, 75, 75, 128)       147584    \n",
      "                                                                 \n",
      " block2_pool (MaxPooling2D)  (None, 37, 37, 128)       0         \n",
      "                                                                 \n",
      " block3_conv1 (Conv2D)       (None, 37, 37, 256)       295168    \n",
      "                                                                 \n",
      " block3_conv2 (Conv2D)       (None, 37, 37, 256)       590080    \n",
      "                                                                 \n",
      " block3_conv3 (Conv2D)       (None, 37, 37, 256)       590080    \n",
      "                                                                 \n",
      " block3_pool (MaxPooling2D)  (None, 18, 18, 256)       0         \n",
      "                                                                 \n",
      " block4_conv1 (Conv2D)       (None, 18, 18, 512)       1180160   \n",
      "                                                                 \n",
      " block4_conv2 (Conv2D)       (None, 18, 18, 512)       2359808   \n",
      "                                                                 \n",
      " block4_conv3 (Conv2D)       (None, 18, 18, 512)       2359808   \n",
      "                                                                 \n",
      " block4_pool (MaxPooling2D)  (None, 9, 9, 512)         0         \n",
      "                                                                 \n",
      " block5_conv1 (Conv2D)       (None, 9, 9, 512)         2359808   \n",
      "                                                                 \n",
      " block5_conv2 (Conv2D)       (None, 9, 9, 512)         2359808   \n",
      "                                                                 \n",
      " block5_conv3 (Conv2D)       (None, 9, 9, 512)         2359808   \n",
      "                                                                 \n",
      " block5_pool (MaxPooling2D)  (None, 4, 4, 512)         0         \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 14,714,688\n",
      "Trainable params: 14,714,688\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "from tensorflow.keras.applications import VGG16 #합성곱층 13개 + 완전연결층 3개\n",
    "\n",
    "conv_base = VGG16(weights = 'imagenet', #모델 초기화할 체크포인트 지정\n",
    "         include_top = False, #네트워크 최상위 완전 연결 분류기 보함여부 cf.True라면 input_shape이 원본과 동일한(244,244,3)이어야함\n",
    "                  input_shape = (150,150,3)) #네트워크에 주입할 이미지 텐서의 크기\n",
    "\n",
    "conv_base.summary()\n",
    "#number of weight parameter of the cnn = Size(width) of kernels(filter) used in the Conv layer ^2 \n",
    "#                                        * Number of channels of the input image\n",
    "#                                        * Number of kernels\n",
    "\n",
    "#Number of biases of the Conv layer = Number of kernels(filter)\n",
    "\n",
    "#Number of parameters of the Conv layer = number of weight parameter of the cnn \n",
    "#                                        + Number of biases of the Conv layer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "870295a7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 2000 images belonging to 2 classes.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-05-18 16:26:27.374777: W tensorflow/core/platform/profile_utils/cpu_utils.cc:128] Failed to get CPU frequency: 0 Hz\n",
      "2022-05-18 16:26:27.427125: I tensorflow/core/grappler/optimizers/custom_graph_optimizer_registry.cc:113] Plugin optimizer for device_type GPU is enabled.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 1000 images belonging to 2 classes.\n",
      "Found 1000 images belonging to 2 classes.\n"
     ]
    }
   ],
   "source": [
    "#사전 훈련된 합성곱 기반 층을 사용한 특성 추출\n",
    "import numpy as np\n",
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "\n",
    "datagen = ImageDataGenerator(rescale = 1./255)\n",
    "batch_size = 20\n",
    "\n",
    "def extract_features(directory,sample_count):\n",
    "    features = np.zeros(shape=(sample_count,4,4,512))\n",
    "    labels = np.zeros(shape=(sample_count))\n",
    "    #flow_from_directory 반환은 (x,y)튜플  \n",
    "    #x는 (배치 크기, *표적 크기, 채널)의 형태의 이미지 배치로 구성된 numpy 배열  \n",
    "    #y는 그에 대응하는 라벨로 이루어진 numpy 배열\n",
    "    generator = datagen.flow_from_directory(\n",
    "        directory,\n",
    "        target_size=(150,150),\n",
    "        batch_size = batch_size,\n",
    "        class_mode='binary')\n",
    "    \n",
    "    i=0\n",
    "    for inputs_batch, labels_batch in generator:\n",
    "        features_batch = conv_base.predict(inputs_batch)\n",
    "        features[i*batch_size : (i+1) * batch_size] = features_batch\n",
    "        labels[i * batch_size : (i+1) * batch_size] = labels_batch\n",
    "        i+=1\n",
    "        if i * batch_size >= sample_count:\n",
    "            break\n",
    "        return features,labels\n",
    "    \n",
    "train_features, train_labels = extract_features(train_dir,2000)\n",
    "validation_features, validation_labels = extract_features(validation_dir,1000)\n",
    "test_features, test_labels = extract_features(test_dir,1000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "14adb4e9",
   "metadata": {},
   "outputs": [],
   "source": [
    "#추출된 특성 크기는 (samples,4,4,512)이다. cf.4*4*512=8192. 완전 연결 분휴기에 주입하기 위해 (samples,8192)로 펼쳐야함\n",
    "\n",
    "train_features = np.reshape(train_features,(2000,4*4*512))\n",
    "validation_features = np.reshape(validation_features,(1000,4*4*512))\n",
    "test_features = np.reshape(test_features,(1000,4*4*512))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "1292e24d",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/30\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/jw1010110/miniforge3/envs/tensorflow_gpu/lib/python3.9/site-packages/keras/optimizer_v2/rmsprop.py:130: UserWarning: The `lr` argument is deprecated, use `learning_rate` instead.\n",
      "  super(RMSprop, self).__init__(name, **kwargs)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      " 11/100 [==>...........................] - ETA: 0s - loss: 0.6887 - acc: 1.0000 "
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-05-18 16:26:28.032922: I tensorflow/core/grappler/optimizers/custom_graph_optimizer_registry.cc:113] Plugin optimizer for device_type GPU is enabled.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "100/100 [==============================] - 1s 8ms/step - loss: 0.6903 - acc: 0.9960 - val_loss: 0.6881 - val_acc: 0.9860\n",
      "Epoch 2/30\n",
      " 10/100 [==>...........................] - ETA: 0s - loss: 0.6849 - acc: 1.0000"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-05-18 16:26:28.762505: I tensorflow/core/grappler/optimizers/custom_graph_optimizer_registry.cc:113] Plugin optimizer for device_type GPU is enabled.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "100/100 [==============================] - 1s 6ms/step - loss: 0.6821 - acc: 0.9945 - val_loss: 0.6824 - val_acc: 0.9850\n",
      "Epoch 3/30\n",
      "100/100 [==============================] - 1s 6ms/step - loss: 0.6714 - acc: 0.9970 - val_loss: 0.6762 - val_acc: 0.9850\n",
      "Epoch 4/30\n",
      "100/100 [==============================] - 1s 6ms/step - loss: 0.6627 - acc: 0.9970 - val_loss: 0.6687 - val_acc: 0.9850\n",
      "Epoch 5/30\n",
      "100/100 [==============================] - 1s 6ms/step - loss: 0.6528 - acc: 0.9990 - val_loss: 0.6586 - val_acc: 0.9860\n",
      "Epoch 6/30\n",
      "100/100 [==============================] - 1s 6ms/step - loss: 0.6421 - acc: 0.9995 - val_loss: 0.6516 - val_acc: 0.9860\n",
      "Epoch 7/30\n",
      "100/100 [==============================] - 1s 6ms/step - loss: 0.6330 - acc: 0.9995 - val_loss: 0.6454 - val_acc: 0.9860\n",
      "Epoch 8/30\n",
      "100/100 [==============================] - 1s 7ms/step - loss: 0.6228 - acc: 1.0000 - val_loss: 0.6367 - val_acc: 0.9860\n",
      "Epoch 9/30\n",
      "100/100 [==============================] - 1s 6ms/step - loss: 0.6120 - acc: 1.0000 - val_loss: 0.6234 - val_acc: 0.9860\n",
      "Epoch 10/30\n",
      "100/100 [==============================] - 1s 6ms/step - loss: 0.6014 - acc: 1.0000 - val_loss: 0.6145 - val_acc: 0.9860\n",
      "Epoch 11/30\n",
      "100/100 [==============================] - 1s 6ms/step - loss: 0.5903 - acc: 1.0000 - val_loss: 0.6035 - val_acc: 0.9870\n",
      "Epoch 12/30\n",
      "100/100 [==============================] - 1s 6ms/step - loss: 0.5794 - acc: 1.0000 - val_loss: 0.5969 - val_acc: 0.9860\n",
      "Epoch 13/30\n",
      "100/100 [==============================] - 1s 6ms/step - loss: 0.5685 - acc: 0.9995 - val_loss: 0.5826 - val_acc: 0.9890\n",
      "Epoch 14/30\n",
      "100/100 [==============================] - 1s 6ms/step - loss: 0.5570 - acc: 1.0000 - val_loss: 0.5721 - val_acc: 0.9880\n",
      "Epoch 15/30\n",
      "100/100 [==============================] - 1s 6ms/step - loss: 0.5453 - acc: 1.0000 - val_loss: 0.5584 - val_acc: 0.9890\n",
      "Epoch 16/30\n",
      "100/100 [==============================] - 1s 6ms/step - loss: 0.5335 - acc: 1.0000 - val_loss: 0.5526 - val_acc: 0.9880\n",
      "Epoch 17/30\n",
      "100/100 [==============================] - 1s 6ms/step - loss: 0.5217 - acc: 1.0000 - val_loss: 0.5422 - val_acc: 0.9880\n",
      "Epoch 18/30\n",
      "100/100 [==============================] - 1s 6ms/step - loss: 0.5098 - acc: 1.0000 - val_loss: 0.5281 - val_acc: 0.9880\n",
      "Epoch 19/30\n",
      "100/100 [==============================] - 1s 6ms/step - loss: 0.4979 - acc: 1.0000 - val_loss: 0.5177 - val_acc: 0.9880\n",
      "Epoch 20/30\n",
      "100/100 [==============================] - 1s 6ms/step - loss: 0.4855 - acc: 1.0000 - val_loss: 0.5029 - val_acc: 0.9900\n",
      "Epoch 21/30\n",
      "100/100 [==============================] - 1s 6ms/step - loss: 0.4731 - acc: 1.0000 - val_loss: 0.4923 - val_acc: 0.9890\n",
      "Epoch 22/30\n",
      "100/100 [==============================] - 1s 6ms/step - loss: 0.4607 - acc: 1.0000 - val_loss: 0.4850 - val_acc: 0.9880\n",
      "Epoch 23/30\n",
      "100/100 [==============================] - 1s 6ms/step - loss: 0.4483 - acc: 1.0000 - val_loss: 0.4684 - val_acc: 0.9900\n",
      "Epoch 24/30\n",
      "100/100 [==============================] - 1s 6ms/step - loss: 0.4360 - acc: 1.0000 - val_loss: 0.4568 - val_acc: 0.9900\n",
      "Epoch 25/30\n",
      "100/100 [==============================] - 1s 6ms/step - loss: 0.4236 - acc: 1.0000 - val_loss: 0.4475 - val_acc: 0.9900\n",
      "Epoch 26/30\n",
      "100/100 [==============================] - 1s 6ms/step - loss: 0.4112 - acc: 1.0000 - val_loss: 0.4393 - val_acc: 0.9880\n",
      "Epoch 27/30\n",
      "100/100 [==============================] - 1s 6ms/step - loss: 0.3989 - acc: 1.0000 - val_loss: 0.4285 - val_acc: 0.9880\n",
      "Epoch 28/30\n",
      "100/100 [==============================] - 1s 6ms/step - loss: 0.3866 - acc: 1.0000 - val_loss: 0.4166 - val_acc: 0.9880\n",
      "Epoch 29/30\n",
      "100/100 [==============================] - 1s 6ms/step - loss: 0.3744 - acc: 1.0000 - val_loss: 0.4017 - val_acc: 0.9890\n",
      "Epoch 30/30\n",
      "100/100 [==============================] - 1s 6ms/step - loss: 0.3622 - acc: 1.0000 - val_loss: 0.3889 - val_acc: 0.9900\n"
     ]
    }
   ],
   "source": [
    "#완전 연결 분류기를 정의하고 훈련\n",
    "from keras import models\n",
    "from keras import layers\n",
    "from tensorflow.keras import optimizers\n",
    "\n",
    "model = models.Sequential()\n",
    "model.add(layers.Dense(256,activation='relu',input_dim=4*4*512))\n",
    "model.add(layers.Dropout(0.5))\n",
    "model.add(layers.Dense(1,activation='sigmoid'))\n",
    "model.compile(optimizer=optimizers.RMSprop(lr=2e-5),loss='binary_crossentropy',metrics=['acc'])\n",
    "history = model.fit(\n",
    "    train_features,\n",
    "    train_labels,\n",
    "    epochs=30,\n",
    "    batch_size=20,\n",
    "    validation_data=(validation_features,validation_labels))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "31280bca",
   "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",
    "\n",
    "acc = history.history['acc']\n",
    "val_acc = history.history['val_acc']\n",
    "loss = history.history['loss']\n",
    "val_loss = history.history['val_loss']\n",
    "\n",
    "epochs = range(1,len(acc) + 1)\n",
    "\n",
    "plt.plot(epochs,acc,'bo',label='Training acc')\n",
    "plt.plot(epochs,val_acc,'b',label='validation acc')\n",
    "plt.title('Training and validation accuracy')\n",
    "\n",
    "plt.legend()\n",
    "\n",
    "plt.figure()\n",
    "\n",
    "plt.plot(epochs,loss,'bo',label='Training acc')\n",
    "plt.plot(epochs,val_loss,'b',label='validation acc')\n",
    "plt.title('Training and validation loss')\n",
    "plt.legend()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6523a031",
   "metadata": {},
   "source": [
    "데이터증식을 사용한 특성 추출"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "fc10e11f",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential_1\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " vgg16 (Functional)          (None, 4, 4, 512)         14714688  \n",
      "                                                                 \n",
      " flatten (Flatten)           (None, 8192)              0         \n",
      "                                                                 \n",
      " dense_2 (Dense)             (None, 256)               2097408   \n",
      "                                                                 \n",
      " dense_3 (Dense)             (None, 1)                 257       \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 16,812,353\n",
      "Trainable params: 16,812,353\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "#합성곱 기반 층 위에 완전 연결 분류기 추가하기\n",
    "\n",
    "model = models.Sequential()\n",
    "model.add(conv_base)\n",
    "model.add(layers.Flatten())\n",
    "model.add(layers.Dense(256,activation='relu'))\n",
    "model.add(layers.Dense(1,activation='sigmoid'))\n",
    "\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "0669a766",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "conv_base 동결하기 전 훈련되는 가중치 수 30\n",
      "conv_base 동결한 후 훈련되는 가중치 수 4\n"
     ]
    }
   ],
   "source": [
    "print('conv_base 동결하기 전 훈련되는 가중치 수',len(model.trainable_weights))\n",
    "conv_base.trainable = False\n",
    "print('conv_base 동결한 후 훈련되는 가중치 수',len(model.trainable_weights))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "be6dc00e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 2000 images belonging to 2 classes.\n",
      "Found 1000 images belonging to 2 classes.\n",
      "Epoch 1/30\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/var/folders/pn/m2bqkk014sv8108lk7tl4xy80000gn/T/ipykernel_19487/372503109.py:30: UserWarning: `Model.fit_generator` is deprecated and will be removed in a future version. Please use `Model.fit`, which supports generators.\n",
      "  history = model.fit_generator(\n",
      "2022-05-18 17:26:23.379375: I tensorflow/core/grappler/optimizers/custom_graph_optimizer_registry.cc:113] Plugin optimizer for device_type GPU is enabled.\n",
      "2022-05-18 17:26:32.854854: I tensorflow/core/grappler/optimizers/custom_graph_optimizer_registry.cc:113] Plugin optimizer for device_type GPU is enabled.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "100/100 - 14s - loss: 0.0401 - acc: 0.9925 - val_loss: 0.3578 - val_acc: 0.9270 - 14s/epoch - 136ms/step\n",
      "Epoch 2/30\n",
      "100/100 - 12s - loss: 0.0168 - acc: 0.9950 - val_loss: 0.4046 - val_acc: 0.9270 - 12s/epoch - 124ms/step\n",
      "Epoch 3/30\n",
      "100/100 - 12s - loss: 0.0357 - acc: 0.9950 - val_loss: 0.3754 - val_acc: 0.9200 - 12s/epoch - 125ms/step\n",
      "Epoch 4/30\n",
      "100/100 - 12s - loss: 0.0166 - acc: 0.9955 - val_loss: 0.4090 - val_acc: 0.9230 - 12s/epoch - 125ms/step\n",
      "Epoch 5/30\n",
      "100/100 - 13s - loss: 0.0098 - acc: 0.9970 - val_loss: 0.3936 - val_acc: 0.9330 - 13s/epoch - 125ms/step\n",
      "Epoch 6/30\n",
      "100/100 - 12s - loss: 0.0078 - acc: 0.9975 - val_loss: 0.4685 - val_acc: 0.9210 - 12s/epoch - 125ms/step\n",
      "Epoch 7/30\n",
      "100/100 - 12s - loss: 0.0024 - acc: 0.9995 - val_loss: 0.3819 - val_acc: 0.9400 - 12s/epoch - 125ms/step\n",
      "Epoch 8/30\n",
      "100/100 - 13s - loss: 0.0056 - acc: 0.9980 - val_loss: 0.3935 - val_acc: 0.9310 - 13s/epoch - 126ms/step\n",
      "Epoch 9/30\n",
      "100/100 - 13s - loss: 0.0146 - acc: 0.9970 - val_loss: 0.4060 - val_acc: 0.9300 - 13s/epoch - 125ms/step\n",
      "Epoch 10/30\n",
      "100/100 - 13s - loss: 0.0057 - acc: 0.9975 - val_loss: 0.4378 - val_acc: 0.9300 - 13s/epoch - 125ms/step\n",
      "Epoch 11/30\n",
      "100/100 - 13s - loss: 0.0172 - acc: 0.9955 - val_loss: 0.4099 - val_acc: 0.9220 - 13s/epoch - 126ms/step\n",
      "Epoch 12/30\n",
      "100/100 - 12s - loss: 0.0048 - acc: 0.9985 - val_loss: 0.3692 - val_acc: 0.9340 - 12s/epoch - 125ms/step\n",
      "Epoch 13/30\n",
      "100/100 - 12s - loss: 0.0149 - acc: 0.9980 - val_loss: 0.4300 - val_acc: 0.9320 - 12s/epoch - 125ms/step\n",
      "Epoch 14/30\n",
      "100/100 - 13s - loss: 0.0124 - acc: 0.9975 - val_loss: 0.4425 - val_acc: 0.9300 - 13s/epoch - 125ms/step\n",
      "Epoch 15/30\n",
      "100/100 - 12s - loss: 0.0052 - acc: 0.9990 - val_loss: 0.4526 - val_acc: 0.9360 - 12s/epoch - 125ms/step\n",
      "Epoch 16/30\n",
      "100/100 - 12s - loss: 0.0031 - acc: 0.9995 - val_loss: 0.4874 - val_acc: 0.9260 - 12s/epoch - 125ms/step\n",
      "Epoch 17/30\n",
      "100/100 - 12s - loss: 0.0079 - acc: 0.9975 - val_loss: 0.5060 - val_acc: 0.9280 - 12s/epoch - 125ms/step\n",
      "Epoch 18/30\n",
      "100/100 - 13s - loss: 0.0190 - acc: 0.9930 - val_loss: 0.4585 - val_acc: 0.9310 - 13s/epoch - 125ms/step\n",
      "Epoch 19/30\n",
      "100/100 - 12s - loss: 0.0026 - acc: 0.9990 - val_loss: 0.5324 - val_acc: 0.9180 - 12s/epoch - 125ms/step\n",
      "Epoch 20/30\n",
      "100/100 - 13s - loss: 0.0046 - acc: 0.9980 - val_loss: 0.4830 - val_acc: 0.9220 - 13s/epoch - 127ms/step\n",
      "Epoch 21/30\n",
      "100/100 - 12s - loss: 0.0038 - acc: 0.9980 - val_loss: 1.2644 - val_acc: 0.8720 - 12s/epoch - 125ms/step\n",
      "Epoch 22/30\n",
      "100/100 - 12s - loss: 0.0258 - acc: 0.9940 - val_loss: 0.4504 - val_acc: 0.9260 - 12s/epoch - 125ms/step\n",
      "Epoch 23/30\n",
      "100/100 - 12s - loss: 0.0085 - acc: 0.9980 - val_loss: 0.3827 - val_acc: 0.9280 - 12s/epoch - 125ms/step\n",
      "Epoch 24/30\n",
      "100/100 - 13s - loss: 0.0065 - acc: 0.9975 - val_loss: 0.4265 - val_acc: 0.9360 - 13s/epoch - 125ms/step\n",
      "Epoch 25/30\n",
      "100/100 - 13s - loss: 0.0051 - acc: 0.9975 - val_loss: 0.5810 - val_acc: 0.9240 - 13s/epoch - 125ms/step\n",
      "Epoch 26/30\n",
      "100/100 - 12s - loss: 0.0043 - acc: 0.9980 - val_loss: 0.6176 - val_acc: 0.9180 - 12s/epoch - 125ms/step\n",
      "Epoch 27/30\n",
      "100/100 - 13s - loss: 0.0050 - acc: 0.9980 - val_loss: 0.5473 - val_acc: 0.9240 - 13s/epoch - 125ms/step\n",
      "Epoch 28/30\n",
      "100/100 - 12s - loss: 0.0026 - acc: 0.9995 - val_loss: 0.6964 - val_acc: 0.9130 - 12s/epoch - 125ms/step\n",
      "Epoch 29/30\n",
      "100/100 - 13s - loss: 0.0082 - acc: 0.9980 - val_loss: 0.4390 - val_acc: 0.9290 - 13s/epoch - 125ms/step\n",
      "Epoch 30/30\n",
      "100/100 - 13s - loss: 0.0037 - acc: 0.9980 - val_loss: 0.5328 - val_acc: 0.9240 - 13s/epoch - 125ms/step\n"
     ]
    }
   ],
   "source": [
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "from tensorflow.keras import optimizers\n",
    "\n",
    "train_datagen = ImageDataGenerator(\n",
    "    rescale=1./255,\n",
    "    rotation_range=20,\n",
    "    width_shift_range=0.1,\n",
    "    height_shift_range=0.1,\n",
    "    shear_range=0.1,\n",
    "    zoom_range=0.1,\n",
    "    horizontal_flip=True,\n",
    "    fill_mode='nearest')\n",
    "\n",
    "test_datagen = ImageDataGenerator(rescale=1./255)\n",
    "\n",
    "train_generator = train_datagen.flow_from_directory(\n",
    "    train_dir,\n",
    "    target_size=(150,150),\n",
    "    batch_size=20,\n",
    "    class_mode='binary')\n",
    "\n",
    "validation_generator = test_datagen.flow_from_directory(\n",
    "    validation_dir,\n",
    "    target_size=(150,150),\n",
    "    batch_size=20,\n",
    "    class_mode='binary')\n",
    "\n",
    "model.compile(loss='binary_crossentropy',optimizer=optimizers.RMSprop(learning_rate=2e-5),metrics=['acc'])\n",
    "\n",
    "history = model.fit_generator(\n",
    "        train_generator,\n",
    "        steps_per_epoch=100,\n",
    "        epochs=30,\n",
    "        validation_data=validation_generator,\n",
    "        validation_steps=50,\n",
    "        verbose=2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "b4b5a204",
   "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",
    "\n",
    "acc = history.history['acc']\n",
    "val_acc = history.history['val_acc']\n",
    "loss = history.history['loss']\n",
    "val_loss = history.history['val_loss']\n",
    "\n",
    "epochs = range(1,len(acc) + 1)\n",
    "\n",
    "plt.plot(epochs,acc,'bo',label='Training acc')\n",
    "plt.plot(epochs,val_acc,'b',label='validation acc')\n",
    "plt.title('Training and validation accuracy')\n",
    "\n",
    "plt.legend()\n",
    "\n",
    "plt.figure()\n",
    "\n",
    "plt.plot(epochs,loss,'bo',label='Training acc')\n",
    "plt.plot(epochs,val_loss,'b',label='validation acc')\n",
    "plt.title('Training and validation loss')\n",
    "plt.legend()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "a0c40e3f",
   "metadata": {},
   "outputs": [],
   "source": [
    "conv_base.trainable = True\n",
    "\n",
    "set_trainable = False\n",
    "for layer in conv_base.layers:\n",
    "    if layer.name == 'block5_conv1':\n",
    "        set_trainable = True\n",
    "    if set_trainable:\n",
    "        layer.trainable = True\n",
    "    else:\n",
    "        layer.trainable = False\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "2b99b414",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/jw1010110/miniforge3/envs/tensorflow_gpu/lib/python3.9/site-packages/keras/optimizer_v2/rmsprop.py:130: UserWarning: The `lr` argument is deprecated, use `learning_rate` instead.\n",
      "  super(RMSprop, self).__init__(name, **kwargs)\n",
      "/var/folders/pn/m2bqkk014sv8108lk7tl4xy80000gn/T/ipykernel_19487/959284415.py:2: UserWarning: `Model.fit_generator` is deprecated and will be removed in a future version. Please use `Model.fit`, which supports generators.\n",
      "  history = model.fit_generator(\n",
      "2022-05-18 16:53:34.121388: I tensorflow/core/grappler/optimizers/custom_graph_optimizer_registry.cc:113] Plugin optimizer for device_type GPU is enabled.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "100/100 [==============================] - ETA: 0s - loss: 0.1875 - acc: 0.9245"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2022-05-18 16:53:43.674799: I tensorflow/core/grappler/optimizers/custom_graph_optimizer_registry.cc:113] Plugin optimizer for device_type GPU is enabled.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "100/100 [==============================] - 13s 128ms/step - loss: 0.1875 - acc: 0.9245 - val_loss: 0.2584 - val_acc: 0.8970\n",
      "Epoch 2/100\n",
      "100/100 [==============================] - 13s 126ms/step - loss: 0.1595 - acc: 0.9425 - val_loss: 0.2278 - val_acc: 0.9130\n",
      "Epoch 3/100\n",
      "100/100 [==============================] - 13s 126ms/step - loss: 0.1252 - acc: 0.9520 - val_loss: 0.2124 - val_acc: 0.9260\n",
      "Epoch 4/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.1265 - acc: 0.9505 - val_loss: 0.2279 - val_acc: 0.9190\n",
      "Epoch 5/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.1015 - acc: 0.9580 - val_loss: 0.2527 - val_acc: 0.9210\n",
      "Epoch 6/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0923 - acc: 0.9640 - val_loss: 0.2492 - val_acc: 0.9110\n",
      "Epoch 7/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0882 - acc: 0.9635 - val_loss: 0.2118 - val_acc: 0.9260\n",
      "Epoch 8/100\n",
      "100/100 [==============================] - 13s 126ms/step - loss: 0.0728 - acc: 0.9740 - val_loss: 0.2216 - val_acc: 0.9260\n",
      "Epoch 9/100\n",
      "100/100 [==============================] - 13s 126ms/step - loss: 0.0616 - acc: 0.9780 - val_loss: 0.2291 - val_acc: 0.9240\n",
      "Epoch 10/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0574 - acc: 0.9795 - val_loss: 0.2525 - val_acc: 0.9140\n",
      "Epoch 11/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0455 - acc: 0.9835 - val_loss: 0.4604 - val_acc: 0.8770\n",
      "Epoch 12/100\n",
      "100/100 [==============================] - 13s 126ms/step - loss: 0.0432 - acc: 0.9870 - val_loss: 0.2770 - val_acc: 0.9170\n",
      "Epoch 13/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0338 - acc: 0.9870 - val_loss: 0.2461 - val_acc: 0.9330\n",
      "Epoch 14/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0474 - acc: 0.9790 - val_loss: 0.2324 - val_acc: 0.9260\n",
      "Epoch 15/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0297 - acc: 0.9905 - val_loss: 0.2570 - val_acc: 0.9240\n",
      "Epoch 16/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0319 - acc: 0.9900 - val_loss: 0.2602 - val_acc: 0.9210\n",
      "Epoch 17/100\n",
      "100/100 [==============================] - 13s 126ms/step - loss: 0.0356 - acc: 0.9900 - val_loss: 0.2711 - val_acc: 0.9170\n",
      "Epoch 18/100\n",
      "100/100 [==============================] - 13s 126ms/step - loss: 0.0299 - acc: 0.9890 - val_loss: 0.2557 - val_acc: 0.9260\n",
      "Epoch 19/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0230 - acc: 0.9925 - val_loss: 0.2389 - val_acc: 0.9270\n",
      "Epoch 20/100\n",
      "100/100 [==============================] - 13s 128ms/step - loss: 0.0282 - acc: 0.9910 - val_loss: 0.2429 - val_acc: 0.9300\n",
      "Epoch 21/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0244 - acc: 0.9930 - val_loss: 0.2337 - val_acc: 0.9260\n",
      "Epoch 22/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0219 - acc: 0.9910 - val_loss: 0.2476 - val_acc: 0.9310\n",
      "Epoch 23/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0171 - acc: 0.9950 - val_loss: 0.2488 - val_acc: 0.9290\n",
      "Epoch 24/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0202 - acc: 0.9930 - val_loss: 0.3525 - val_acc: 0.9130\n",
      "Epoch 25/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0112 - acc: 0.9980 - val_loss: 0.2683 - val_acc: 0.9260\n",
      "Epoch 26/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0184 - acc: 0.9935 - val_loss: 0.3196 - val_acc: 0.9260\n",
      "Epoch 27/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0116 - acc: 0.9965 - val_loss: 0.2817 - val_acc: 0.9260\n",
      "Epoch 28/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0178 - acc: 0.9930 - val_loss: 0.2749 - val_acc: 0.9200\n",
      "Epoch 29/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0099 - acc: 0.9970 - val_loss: 0.3121 - val_acc: 0.9240\n",
      "Epoch 30/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0128 - acc: 0.9960 - val_loss: 0.2832 - val_acc: 0.9280\n",
      "Epoch 31/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0110 - acc: 0.9960 - val_loss: 0.2895 - val_acc: 0.9280\n",
      "Epoch 32/100\n",
      "100/100 [==============================] - 13s 126ms/step - loss: 0.0131 - acc: 0.9960 - val_loss: 0.2707 - val_acc: 0.9330\n",
      "Epoch 33/100\n",
      "100/100 [==============================] - 13s 128ms/step - loss: 0.0084 - acc: 0.9970 - val_loss: 0.2997 - val_acc: 0.9320\n",
      "Epoch 34/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0141 - acc: 0.9955 - val_loss: 0.4976 - val_acc: 0.9090\n",
      "Epoch 35/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0049 - acc: 0.9990 - val_loss: 0.3374 - val_acc: 0.9320\n",
      "Epoch 36/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0105 - acc: 0.9950 - val_loss: 0.2927 - val_acc: 0.9290\n",
      "Epoch 37/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0103 - acc: 0.9975 - val_loss: 0.2893 - val_acc: 0.9290\n",
      "Epoch 38/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0057 - acc: 0.9975 - val_loss: 0.3868 - val_acc: 0.9240\n",
      "Epoch 39/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0111 - acc: 0.9965 - val_loss: 0.2818 - val_acc: 0.9300\n",
      "Epoch 40/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0072 - acc: 0.9980 - val_loss: 0.3307 - val_acc: 0.9260\n",
      "Epoch 41/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0062 - acc: 0.9980 - val_loss: 0.3208 - val_acc: 0.9340\n",
      "Epoch 42/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0044 - acc: 0.9985 - val_loss: 0.3179 - val_acc: 0.9300\n",
      "Epoch 43/100\n",
      "100/100 [==============================] - 13s 126ms/step - loss: 0.0033 - acc: 0.9995 - val_loss: 0.3212 - val_acc: 0.9310\n",
      "Epoch 44/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0080 - acc: 0.9975 - val_loss: 0.2935 - val_acc: 0.9300\n",
      "Epoch 45/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0046 - acc: 0.9985 - val_loss: 0.3559 - val_acc: 0.9230\n",
      "Epoch 46/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0080 - acc: 0.9980 - val_loss: 0.5393 - val_acc: 0.9050\n",
      "Epoch 47/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0075 - acc: 0.9975 - val_loss: 0.3597 - val_acc: 0.9220\n",
      "Epoch 48/100\n",
      "100/100 [==============================] - 13s 126ms/step - loss: 0.0042 - acc: 0.9985 - val_loss: 0.3429 - val_acc: 0.9300\n",
      "Epoch 49/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0095 - acc: 0.9960 - val_loss: 0.3011 - val_acc: 0.9290\n",
      "Epoch 50/100\n",
      "100/100 [==============================] - 13s 126ms/step - loss: 0.0055 - acc: 0.9980 - val_loss: 0.3980 - val_acc: 0.9170\n",
      "Epoch 51/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0010 - acc: 1.0000 - val_loss: 0.3235 - val_acc: 0.9320\n",
      "Epoch 52/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0052 - acc: 0.9985 - val_loss: 0.3691 - val_acc: 0.9290\n",
      "Epoch 53/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0064 - acc: 0.9980 - val_loss: 0.3488 - val_acc: 0.9290\n",
      "Epoch 54/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0033 - acc: 0.9995 - val_loss: 0.3663 - val_acc: 0.9250\n",
      "Epoch 55/100\n",
      "100/100 [==============================] - 13s 126ms/step - loss: 0.0054 - acc: 0.9990 - val_loss: 0.3239 - val_acc: 0.9300\n",
      "Epoch 56/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0014 - acc: 1.0000 - val_loss: 0.3614 - val_acc: 0.9320\n",
      "Epoch 57/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0045 - acc: 0.9985 - val_loss: 0.9125 - val_acc: 0.8820\n",
      "Epoch 58/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0038 - acc: 0.9985 - val_loss: 0.3490 - val_acc: 0.9280\n",
      "Epoch 59/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0074 - acc: 0.9985 - val_loss: 0.3605 - val_acc: 0.9290\n",
      "Epoch 60/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0048 - acc: 0.9995 - val_loss: 0.3500 - val_acc: 0.9310\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 61/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0037 - acc: 0.9995 - val_loss: 0.4603 - val_acc: 0.9190\n",
      "Epoch 62/100\n",
      "100/100 [==============================] - 13s 126ms/step - loss: 0.0022 - acc: 0.9995 - val_loss: 0.3990 - val_acc: 0.9260\n",
      "Epoch 63/100\n",
      "100/100 [==============================] - 13s 126ms/step - loss: 0.0016 - acc: 0.9995 - val_loss: 0.4055 - val_acc: 0.9270\n",
      "Epoch 64/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0044 - acc: 0.9985 - val_loss: 0.3613 - val_acc: 0.9310\n",
      "Epoch 65/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0063 - acc: 0.9970 - val_loss: 0.3498 - val_acc: 0.9290\n",
      "Epoch 66/100\n",
      "100/100 [==============================] - 13s 126ms/step - loss: 0.0045 - acc: 0.9990 - val_loss: 0.3637 - val_acc: 0.9360\n",
      "Epoch 67/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0043 - acc: 0.9995 - val_loss: 0.4110 - val_acc: 0.9320\n",
      "Epoch 68/100\n",
      "100/100 [==============================] - 13s 126ms/step - loss: 0.0047 - acc: 0.9985 - val_loss: 0.4324 - val_acc: 0.9220\n",
      "Epoch 69/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0092 - acc: 0.9975 - val_loss: 0.3729 - val_acc: 0.9310\n",
      "Epoch 70/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0012 - acc: 0.9990 - val_loss: 0.3339 - val_acc: 0.9360\n",
      "Epoch 71/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0037 - acc: 0.9980 - val_loss: 0.3529 - val_acc: 0.9350\n",
      "Epoch 72/100\n",
      "100/100 [==============================] - 13s 126ms/step - loss: 0.0043 - acc: 0.9975 - val_loss: 0.3661 - val_acc: 0.9310\n",
      "Epoch 73/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0065 - acc: 0.9970 - val_loss: 0.3871 - val_acc: 0.9280\n",
      "Epoch 74/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 7.2979e-04 - acc: 1.0000 - val_loss: 0.3613 - val_acc: 0.9320\n",
      "Epoch 75/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0024 - acc: 0.9985 - val_loss: 0.3306 - val_acc: 0.9400\n",
      "Epoch 76/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0047 - acc: 0.9985 - val_loss: 0.3448 - val_acc: 0.9390\n",
      "Epoch 77/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0023 - acc: 0.9995 - val_loss: 0.4137 - val_acc: 0.9290\n",
      "Epoch 78/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0024 - acc: 0.9990 - val_loss: 0.3730 - val_acc: 0.9330\n",
      "Epoch 79/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0019 - acc: 0.9995 - val_loss: 0.3648 - val_acc: 0.9380\n",
      "Epoch 80/100\n",
      "100/100 [==============================] - 13s 128ms/step - loss: 0.0011 - acc: 0.9995 - val_loss: 0.4493 - val_acc: 0.9300\n",
      "Epoch 81/100\n",
      "100/100 [==============================] - 13s 128ms/step - loss: 0.0014 - acc: 0.9990 - val_loss: 0.3971 - val_acc: 0.9340\n",
      "Epoch 82/100\n",
      "100/100 [==============================] - 13s 128ms/step - loss: 0.0084 - acc: 0.9985 - val_loss: 0.3722 - val_acc: 0.9400\n",
      "Epoch 83/100\n",
      "100/100 [==============================] - 13s 126ms/step - loss: 0.0011 - acc: 0.9995 - val_loss: 0.3964 - val_acc: 0.9360\n",
      "Epoch 84/100\n",
      "100/100 [==============================] - 13s 126ms/step - loss: 0.0057 - acc: 0.9980 - val_loss: 0.3776 - val_acc: 0.9330\n",
      "Epoch 85/100\n",
      "100/100 [==============================] - 13s 128ms/step - loss: 0.0022 - acc: 0.9990 - val_loss: 0.3724 - val_acc: 0.9290\n",
      "Epoch 86/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 4.2272e-04 - acc: 1.0000 - val_loss: 0.4262 - val_acc: 0.9270\n",
      "Epoch 87/100\n",
      "100/100 [==============================] - 13s 128ms/step - loss: 0.0014 - acc: 0.9995 - val_loss: 0.4161 - val_acc: 0.9320\n",
      "Epoch 88/100\n",
      "100/100 [==============================] - 13s 130ms/step - loss: 0.0038 - acc: 0.9990 - val_loss: 0.4578 - val_acc: 0.9270\n",
      "Epoch 89/100\n",
      "100/100 [==============================] - 13s 132ms/step - loss: 0.0011 - acc: 1.0000 - val_loss: 0.4259 - val_acc: 0.9300\n",
      "Epoch 90/100\n",
      "100/100 [==============================] - 13s 130ms/step - loss: 0.0028 - acc: 0.9990 - val_loss: 0.4145 - val_acc: 0.9320\n",
      "Epoch 91/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 0.0050 - acc: 0.9980 - val_loss: 0.4331 - val_acc: 0.9290\n",
      "Epoch 92/100\n",
      "100/100 [==============================] - 13s 131ms/step - loss: 0.0015 - acc: 0.9995 - val_loss: 0.3781 - val_acc: 0.9310\n",
      "Epoch 93/100\n",
      "100/100 [==============================] - 13s 129ms/step - loss: 0.0027 - acc: 0.9990 - val_loss: 0.4080 - val_acc: 0.9390\n",
      "Epoch 94/100\n",
      "100/100 [==============================] - 13s 129ms/step - loss: 4.2907e-04 - acc: 1.0000 - val_loss: 0.3974 - val_acc: 0.9310\n",
      "Epoch 95/100\n",
      "100/100 [==============================] - 13s 126ms/step - loss: 0.0026 - acc: 0.9985 - val_loss: 0.4329 - val_acc: 0.9310\n",
      "Epoch 96/100\n",
      "100/100 [==============================] - 13s 127ms/step - loss: 7.6725e-04 - acc: 0.9995 - val_loss: 0.4549 - val_acc: 0.9330\n",
      "Epoch 97/100\n",
      "100/100 [==============================] - 13s 129ms/step - loss: 0.0028 - acc: 0.9990 - val_loss: 0.4581 - val_acc: 0.9260\n",
      "Epoch 98/100\n",
      "100/100 [==============================] - 14s 135ms/step - loss: 0.0032 - acc: 0.9990 - val_loss: 0.3724 - val_acc: 0.9370\n",
      "Epoch 99/100\n",
      "100/100 [==============================] - 13s 130ms/step - loss: 0.0027 - acc: 0.9990 - val_loss: 0.4088 - val_acc: 0.9310\n",
      "Epoch 100/100\n",
      "100/100 [==============================] - 13s 126ms/step - loss: 0.0015 - acc: 0.9995 - val_loss: 0.4282 - val_acc: 0.9260\n"
     ]
    }
   ],
   "source": [
    "model.compile(loss='binary_crossentropy',optimizer=optimizers.RMSprop(lr=1e-5),metrics=['acc'])\n",
    "history = model.fit_generator(\n",
    "        train_generator,\n",
    "        steps_per_epoch=100,\n",
    "        epochs=100,\n",
    "        validation_data=validation_generator,\n",
    "        validation_steps=50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "ee012420",
   "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": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAEICAYAAABPgw/pAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAAsTAAALEwEAmpwYAAA9H0lEQVR4nO2deXgUVdbG30MgBAyyg0DYRED2AJEdBUFFdFAQFAYXYGYQcMZRx0HQcWRkmBlRkY9RVEQUAQVFQEFwQUFUHCUoomxhl8geIlvYQs73x+mbrnR6qU66esv5PU8/3VV9q+rc6uS9p849915iZiiKoiixT6lIG6AoiqKEBhV0RVGUOEEFXVEUJU5QQVcURYkTVNAVRVHiBBV0RVGUOEEFXfEKEa0gontCXTaSENEeIurtwHmZiK5wfX6JiB63U7YI1xlKRB8X1U4/5+1BRJmhPq8SfkpH2gAldBDRKctmeQDnAFx0bd/LzPPsnouZb3SibLzDzKNCcR4iagBgN4AyzJzrOvc8ALZ/Q6XkoYIeRzBzsvlMRHsA/J6ZV3qWI6LSRiQURYkfNORSAjCP1ET0CBEdBPAaEVUmomVEdISIsl2fUyzHrCai37s+DyOiL4noGVfZ3UR0YxHLNiSiNUR0kohWEtELRDTXh912bJxIRF+5zvcxEVWzfH8XEe0loiwieszP/elERAeJKMGyrz8RbXR97kBEXxPRr0R0gIieJ6JEH+d6nYj+adn+q+uY/UQ0wqPsTUT0PRGdIKJ9RDTB8vUa1/uvRHSKiDqbe2s5vgsRrSOi4673LnbvjT+IqJnr+F+JaBMR9bN815eINrvO+QsRPezaX831+/xKRMeI6AsiUn0JM3rDSw6XAagCoD6AkZDf/jXXdj0AZwA87+f4jgC2AagGYDKAV4mIilD2TQDfAqgKYAKAu/xc046NvwUwHEANAIkAjMA0B/Ci6/y1XddLgReY+X8ATgO41uO8b7o+XwTwoKs+nQH0AjDGj91w2dDHZc91ABoD8IzfnwZwN4BKAG4CMJqIbnV9d7XrvRIzJzPz1x7nrgLgAwDTXHWbAuADIqrqUYdC9yaAzWUALAXwseu4PwGYR0RNXUVehYTvKgBoCeAz1/6/AMgEUB1ATQCPAtB5RcKMCnrJIQ/AE8x8jpnPMHMWM7/LzDnMfBLAJADX+Dl+LzO/wswXAcwGUAvyj2u7LBHVA3AVgL8z83lm/hLA+74uaNPG15g5g5nPAHgbQKpr/0AAy5h5DTOfA/C46x744i0AQwCAiCoA6OvaB2Zez8z/Y+ZcZt4D4GUvdnjjdpd9PzHzaUgDZq3famb+kZnzmHmj63p2zgtIA7Cdmee47HoLwFYAv7GU8XVv/NEJQDKA/7h+o88ALIPr3gC4AKA5EV3KzNnM/J1lfy0A9Zn5AjN/wTpRVNhRQS85HGHms2aDiMoT0cuukMQJyCN+JWvYwYOD5gMz57g+JgdZtjaAY5Z9ALDPl8E2bTxo+Zxjsam29dwuQc3ydS2INz6AiMoCGADgO2be67KjiSuccNBlx78g3nogCtgAYK9H/ToS0SpXSOk4gFE2z2vOvddj314AdSzbvu5NQJuZ2dr4Wc97G6Sx20tEnxNRZ9f+pwHsAPAxEe0ionH2qqGEEhX0koOnt/QXAE0BdGTmS+F+xPcVRgkFBwBUIaLyln11/ZQvjo0HrOd2XbOqr8LMvBkiXDeiYLgFkNDNVgCNXXY8WhQbIGEjK29CnlDqMnNFAC9ZzhvIu90PCUVZqQfgFxt2BTpvXY/4d/55mXkdM98CCccsgXj+YOaTzPwXZr4c8pTwEBH1KqYtSpCooJdcKkBi0r+64rFPOH1Bl8ebDmACESW6vLvf+DmkODYuBHAzEXVzdWA+icB/728CuB/ScLzjYccJAKeI6EoAo23a8DaAYUTU3NWgeNpfAfLEcpaIOkAaEsMRSIjoch/nXg6gCRH9lohKE9EdAJpDwiPF4RtIbH8sEZUhoh6Q32i+6zcbSkQVmfkC5J5cBAAiupmIrnD1lZj9F71eQXEMFfSSy1QA5QAcBfA/AB+G6bpDIR2LWQD+CWABJF/eG1NRRBuZeROA+yAifQBANqTTzh9vAegB4DNmPmrZ/zBEbE8CeMVlsx0bVrjq8BkkHPGZR5ExAJ4kopMA/g6Xt+s6NgfSZ/CVK3Okk8e5swDcDHmKyQIwFsDNHnYHDTOfB9AP8qRyFMB0AHcz81ZXkbsA7HGFnkYBuNO1vzGAlQBOAfgawHRmXl0cW5TgIe23UCIJES0AsJWZHX9CUJR4Rz10JawQ0VVE1IiISrnS+m6BxGIVRSkmOlJUCTeXAVgE6aDMBDCamb+PrEmKEh9oyEVRFCVO0JCLoihKnBCxkEu1atW4QYMGkbq8oihKTLJ+/fqjzFzd23cRE/QGDRogPT09UpdXFEWJSYjIc4RwPhpyURRFiRNU0BVFUeIEFXRFUZQ4QfPQFaUEcuHCBWRmZuLs2bOBCysRISkpCSkpKShTpoztY1TQFaUEkpmZiQoVKqBBgwbwvU6JEimYGVlZWcjMzETDhg1tH6chF0UpgZw9exZVq1ZVMY9SiAhVq1YN+glKBV1RSigq5tFNUX4fFXRFcRBmYPZs4MyZSFuilARU0BXFQbZsAYYNA5Yvj7Ql0UVWVhZSU1ORmpqKyy67DHXq1MnfPn/+vN9j09PTcf/99we8RpcuXUJlbsygnaKK4iA5OQXfY5V584DHHgN+/hmoVw+YNAkYOrTo56tatSo2bNgAAJgwYQKSk5Px8MMP53+fm5uL0qW9y1NaWhrS0tICXmPt2rVFNzBGUQ9dURzknGstpgBOZ1Qzbx4wciSwd6+EkPbule1580J7nWHDhuGhhx5Cz5498cgjj+Dbb79Fly5d0LZtW3Tp0gXbtm0DAKxevRo333wzAGkMRowYgR49euDyyy/HtGnT8s+XnJycX75Hjx4YOHAgrrzySgwdOhRmltnly5fjyiuvRLdu3XD//ffnn9fKnj170L17d7Rr1w7t2rUr0FBMnjwZrVq1Qps2bTBunKyLvWPHDvTu3Rtt2rRBu3btsHPnztDeKD+oh64oDmKE/JyvRfZigMceK/yEkZMj+4vjpXsjIyMDK1euREJCAk6cOIE1a9agdOnSWLlyJR599FG8++67hY7ZunUrVq1ahZMnT6Jp06YYPXp0odzt77//Hps2bULt2rXRtWtXfPXVV0hLS8O9996LNWvWoGHDhhgyZIhXm2rUqIFPPvkESUlJ2L59O4YMGYL09HSsWLECS5YswTfffIPy5cvj2LFjAIChQ4di3Lhx6N+/P86ePYu8vLzQ3iQ/qKArioPEg4f+88/B7S8OgwYNQkJCAgDg+PHjuOeee7B9+3YQES5cuOD1mJtuuglly5ZF2bJlUaNGDRw6dAgpKSkFynTo0CF/X2pqKvbs2YPk5GRcfvnl+XneQ4YMwYwZMwqd/8KFC/jjH/+IDRs2ICEhARkZGQCAlStXYvjw4ShfvjwAoEqVKjh58iR++eUX9O/fH4AMDgonGnJRFAcxgh7LHnq9esHtLw6XXHJJ/ufHH38cPXv2xE8//YSlS5f6zMkuW7Zs/ueEhATk5ubaKmN3cZ/nnnsONWvWxA8//ID09PT8TltmLpRaGOkFg1TQFcVBjGceyx76pEmAywnNp3x52e8kx48fR506dQAAr7/+esjPf+WVV2LXrl3Ys2cPAGDBggU+7ahVqxZKlSqFOXPm4OLFiwCA66+/HrNmzUKOKx517NgxXHrppUhJScGSJUsAAOfOncv/PhyooCuKg8SDhz50KDBjBlC/PkAk7zNmhD5+7snYsWMxfvx4dO3aNV9EQ0m5cuUwffp09OnTB926dUPNmjVRsWLFQuXGjBmD2bNno1OnTsjIyMh/iujTpw/69euHtLQ0pKam4plnngEAzJkzB9OmTUPr1q3RpUsXHDx4MOS2+yJia4qmpaWxLnChxDuzZgG/+x3w178CkydH2ho3W7ZsQbNmzSJtRsQ5deoUkpOTwcy477770LhxYzz44IORNisfb78TEa1nZq95m+qhK4qDxIOHHs+88sorSE1NRYsWLXD8+HHce++9kTapWGiWi6I4SDzE0OOZBx98MKo88uKiHrqiOEg8pC0qsYMKuqI4iIZclHCigq4oDqIhFyWcqKArioOoh66EExV0RXEQ9dBDh5lsa//+/Rg4cKDXMj169ECgdOipU6cWGOzTt29f/PrrryGzM5KooCuKg6iHHnpq166NhQsXFvl4T0Ffvnw5KlWqFALLIo8KuqI4iGa5eOeRRx7B9OnT87cnTJiAZ599FqdOnUKvXr3Qrl07tGrVCu+9916hY/fs2YOWLVsCAM6cOYPBgwejdevWuOOOO3DGsjTU6NGjkZaWhhYtWuCJJ54AAEybNg379+9Hz5490bNnTwBAgwYNcPToUQDAlClT0LJlS7Rs2RJTp07Nv16zZs3whz/8AS1atMD1119f4DqGpUuXomPHjmjbti169+6NQ4cOAZDBS8OHD0erVq3QunXr/BkjP/zwQ7Rr1w5t2rRBr169intLAWgeuqI4SixMn/vAA4BrrYmQkZoKuPTQK4MHD8YDDzyAMWPGAADefvttfPjhh0hKSsLixYtx6aWX4ujRo+jUqRP69evnc33NF198EeXLl8fGjRuxceNGtGvXLv+7SZMmoUqVKrh48SJ69eqFjRs34v7778eUKVOwatUqVKtWrcC51q9fj9deew3ffPMNmBkdO3bENddcg8qVK2P79u1466238Morr+D222/Hu+++izvvvLPA8d26dcP//vc/EBFmzpyJyZMn49lnn8XEiRNRsWJF/PjjjwCA7OxsHDlyBH/4wx/yp+41U+8WFxV0RXEQ9dC907ZtWxw+fBj79+/HkSNHULlyZdSrVw8XLlzAo48+ijVr1qBUqVL45ZdfcOjQIVx22WVez7NmzZr85ehat26N1q1b53/39ttvY8aMGcjNzcWBAwewefPmAt978uWXX6J///75c7UMGDAAX3zxBfr164eGDRsiNTUVANC+ffv8Cb2sZGZm4o477sCBAwdw/vz5/Gl5V65cifnz5+eXq1y5MpYuXYqrr746v0yVKlXs3zw/qKArioPEgofuz5N2koEDB2LhwoU4ePAgBg8eDACYN28ejhw5gvXr16NMmTJo0KCBz2lzDd689927d+OZZ57BunXrULlyZQwbNizgefzNa+U5/a63kMuf/vQnPPTQQ+jXrx9Wr16NCRMm5J/X2zS7vp46ioOtGDoR9SGibUS0g4jGefm+IhEtJaIfiGgTEQ0PuaWKEoOoh+6bwYMHY/78+Vi4cGF+1srx48dRo0YNlClTBqtWrcLevXv9nuPqq6/GPNdaeD/99BM2btwIADhx4gQuueQSVKxYEYcOHcKKFSvyj6lQoQJOnjzp9VxLlixBTk4OTp8+jcWLF6N79+6262Od7nf27Nn5+6+//no8//zz+dvZ2dno3LkzPv/8c+zevRsAQhZyCSjoRJQA4AUANwJoDmAIETX3KHYfgM3M3AZADwDPElFiSCxUlBhGBd03LVq0wMmTJ1GnTh3UqlULgCzflp6ejrS0NMybNw9XXnml33OMHj0ap06dQuvWrTF58mR06NABANCmTRu0bdsWLVq0wIgRI9C1a9f8Y0aOHIkbb7wxv1PU0K5dOwwbNgwdOnRAx44d8fvf/x5t27a1XZ8JEyZg0KBB6N69e4H4/N/+9jdkZ2ejZcuWaNOmDVatWoXq1atjxowZGDBgANq0aYM77rjD9nX8EXD6XCLqDGACM9/g2h4PAMz8b0uZ8QDqQoS9AYBPADRhZp+L6en0uUpJoGtXYO1aoFo14MiRSFvjRqfPjQ2cmD63DoB9lu1M1z4rzwNoBmA/gB8B/NmfmCtKSUE9dCWc2BF0b5F7T7f+BgAbANQGkArgeSK6tNCJiEYSUToRpR+JJndFURwiFjpFlfjBjqBnQsIphhSIJ25lOIBFLOwAsBtAoeAXM89g5jRmTqtevXpRbVaUmMHqoUd4/eBCRHpBY8U/Rfl97Aj6OgCNiaihq6NzMID3Pcr8DKAXABBRTQBNAewK2hpFiTOMh84MeFmMPmIkJSUhKytLRT1KYWZkZWUhKSkpqOMC5qEzcy4R/RHARwASAMxi5k1ENMr1/UsAJgJ4nYh+hIRoHmHmo8FWQlHiDWuo5fx5oEyZyNliJSUlBZmZmdDQZ/SSlJSElJSUoI6xNbCImZcDWO6x7yXL5/0Arg/qyopSArAK+rlzgGsQYsQpU6ZM/ihFJX7QybkUxUHOnwfKlXN/VhQnUUFXFAc5dw6oUEE+q6ArTqOCrigOwQxcuOAWdE1dVJxGBV1RHMJ45OqhK+FCBV1RHMJ45OqhK+FCBV1RHEI9dCXcqKArikOoh66EGxV0RXEI9dCVcKOCrigOoR66Em5U0BXFITwFXT10xWlU0BXFITxDLuqhK06jgq4oDqEeuhJuVNAVxSG0U1QJNyroiuIQxkO/9NKC24riFCroiuIQ6qEr4UYFXVEcwnjkyckFtxXFKVTQFcUhtFNUCTcq6IriEEbAk5KA0qXVQ1ecRwVdURzCCHjZskBionroivOooCuKQxgBL1tWXuqhK06jgq4oDmEEPDFRPXQlPKigK4pDWEMu6qEr4UAFXVEc4vx5oFQpICFBPXQlPKigK4pDnDsnnjmggq6EBxV0RXGI8+fdgq4hFyUcqKArikOcOyeeOaAeuhIeVNAVxSHUQ1fCjQq6ojiEeuhKuFFBVxSHsHaKqoeuhAMVdEVxiPPn1UNXwosKuqI4hHroSrhRQVcUh7B2iqqHroQDFXRFcQjtFFXCjQq6ojiEhlyUcKOCrigOoZ2iSrixJehE1IeIthHRDiIa56NMDyLaQESbiOjz0JqpKLGHeuhKuCkdqAARJQB4AcB1ADIBrCOi95l5s6VMJQDTAfRh5p+JqIZD9ipKzODZKXrhAsAMEEXWLiV+seOhdwCwg5l3MfN5APMB3OJR5rcAFjHzzwDAzIdDa6aixB7WTlEj7Bp2UZzEjqDXAbDPsp3p2melCYDKRLSaiNYT0d3eTkREI4konYjSjxw5UjSLFSVG8Jw+F1BBV5zFjqB7e0Bkj+3SANoDuAnADQAeJ6ImhQ5insHMacycVr169aCNVZRYwtopaoRd4+iKkwSMoUM88rqW7RQA+72UOcrMpwGcJqI1ANoAyAiJlYoSg6iHroQbOx76OgCNiaghESUCGAzgfY8y7wHoTkSliag8gI4AtoTWVEWJHfLygNzcwoKuHrriJAE9dGbOJaI/AvgIQAKAWcy8iYhGub5/iZm3ENGHADYCyAMwk5l/ctJwRYlmjCeunaJKOLETcgEzLwew3GPfSx7bTwN4OnSmKUrsYoRbQy5KONGRooriACa0op2iSjhRQVcUBzDCrR66Ek5U0BXFAXzF0NVDV5xEBV1RHEA9dCUSqKArigN4doqqh66EAxV0RXEAz05R9dCVcKCCrsQtL70EbN0amWv7Crmoh644iQq6EpecPw+MHg28+mrkrg/owCIlvKigK3FJVpa8Hz0ametrp6gSCVTQlbjECHmkZmnWTlElEqigK3FJtHjo2imqhBMVdCUuMUIeaUFXD10JJyroSlwSaUH37BQtU6bgfkVxAhV0JS4xQn78eGRE1NNDJxJxVw9dcRIVdCUuMTF0z8/hwrNTFBBBVw9dcRIVdCUkvPcesGFDpK1wYw21RCLs4tkpaj6rh644iQq6EhJGjQKeeirSVriJtKB7xtAB8dbVQ1ecRAVdKTbMEtY4dCjSlrg5ehRISXF/DjfnzgEJCfIyqIeuOI0KulJsTp8GLlyIPkFv1kw+R2Jw0blzBePngHroivOooCvFxnQ6RpOgZ2UBTZrI50iFXKzhFkA7RRXnUUFXis2xY/KelQXk5kbWFkC845MngVq1gEqVIhdy8eaha8hFcRIVdKXYGEEHIjd3ihXzxFC1KlCtWuQ8dE9BVw9dcRoVdKXYWPO8oyHsYgS8WjWgevXIxdC9hVzUQ1ecRAU9CDIygOXLI21F9GH10A8fjpwdBtPAVKsWOQ+9OJ2iq1cD27Y5YpbiMMyRvb4KehA88wxw552RtiL6sAp6tHnokQy5FNVDHzIEeOghZ+wqadx6KzBhgnPnz8sD3n4bGDMGaNsWSE4G0tOdu14gVNCD4OhRIDtbUvQUN1lZQCnXX1K0Cnq4PaeieuinTwMHDwJr14pYxDPMzv4vZWTICOanny7odISSp58G7rgDmDsXqFIFyMkB1qxx5lp2UEEPguxseXfqjyNWOXYMqF1bBCsaQi5G0KtWlRj62bPyjxZOfHWKBvLQ9+yR919/BTZvdsKy6OGNNyQT6cQJZ86/YIG85+QAM2aE/vzbt4v337+/aMOnnwI1akT2d1NBDwIj5JGakjVaOXZMxLNmzejx0C+9VKasrVZN9oW7Y9Rbp6gdD333bvfnr74KvV3RxJo18nS3erUz51+wAOjWDbjuOuC//w1thlFeHvCHP8hv+vzz7hHBLVqooMcMxkOPxOx90UxWljxuRougZ2W5hdy8h7sR9hZyseOhG0EvXz7+BX3TJnn/5BNnzr1pk4RDHnoI2L9fYt2hYtYs4PPPpV+tdm33/ubN5bqR6hxVQQ8C9dC9c+yYCHqNGtETcom0oHvrFLXroZcrB1x/fWgF/dQpEbRIZ2EYmN2e7Mcfh/78CxZIv87AgcANN8g0EFOmeK//qlUSA7fLgQPAww8DPXoAv/tdwe+aN5cQ0v79xTK/yKig2+T8eemwAlTQPYnGkIsR8urV3fvCiS8P3Y6gN2gAdO8O7Nol4hEKxowRb/Wbb0JzvuKyb5+M5r3ySum83Lu3eOezjlBmFkG/5hrgsstkcZEHHwS+/75wh+W2bUC/fsC999oPyTz/vNj+8stybistWsi7r7ALMzB2LPD11/auFSwq6DYx4RZABd2KmWnR6qFHOjsjWjz0onaKNmwIdO0q26Hw0pctA+bMkc+RzMCwYsItf/6zvBcn7HLmDFCvnjzV7N8P/PCDNBKDB7vL3Hmn/C386U/SmADSWTpwoLzn5ADffhv4WhcuSLjlppvccwVZad68YP08mTNHMmOcCDMBKui2UUH3zqlT4h0ZDz03VzI0IklWltgDABUrSodVtHSKXrwoL1/s3i2C3rYtkJRUfEHPzhbvs2VLoFEj4Isvine+UGEEb9AgiUEXR+DWrpUnmZUrgdatgUcekd98wAB3mXLlREz37AHS0uQ+jBkjdsybJ572Z58FvtbSpZJWOnKk9++rV5eGw5uHnpEh17z6auCxx4pU1YCooNvEmqqonaJuzH0xnaJAZMMuZ89KI2M8c6LIDC7y5aGb77yRnS1roDZsKGU7dgS+/LJ4djz0kPwer78uMd+vvgrPE9SePcBPP/n+ftMm+XupWlU865Ur/Td0/li5EihdGli3DqhfX2LyvXu7/wYMffpIyKliRbkXs2cDjz8unnzbtvYEfcYMmWe/Tx/fZVq0KOyhnzsn1ylbVhoQ6zz5ocSWoBNRHyLaRkQ7iGicn3JXEdFFIhoYOhOjA+OhlyqlHroV07iZkAsQWUG3Dvs3RELQfXnogG9BNxkuDRvKe9euEvc1fTfB8tlnIuSPPAK0by9x+exs3+GAUDJ6tAi1L5HetMkdb77uOnEMvv++aNf69FOgUyep49dfA//3f8Dkyd7LNmsmoZUBA4Dbbwf+/nfZf+21cqx1vAJzwRz53bulsfj976UB8UXz5uKhWztgx4+X+s2a5V54xQkCCjoRJQB4AcCNAJoDGEJEzX2UewrAR6E2MhownmiDBiroVsx9MSEXILKZLtZRoobq1aOnU9R85w1vgn7xogjQiRPAtGnA4sX2bZg9G6hc2S1a3bvLu9NhF2bxhA8c8J5jnpcngmcEvXdveS9Ktkt2tgy179VLthMTgfvvl9CLLypVAt55RzpOjad87bXS0K5d6y73j3/I38706VKnmTPliW/ECP82NW8uYceDB2V7xw7gueekkbvlluDrGAx2PPQOAHYw8y5mPg9gPgBvZv0JwLsAoiBxLfQYD71xYxV0K1YPPRpCLt4EvVq18MbQTZy8uB56584iIGPHAnXrSgfiyJH25py/eFEmkrvxRvd1GzaUeLXTgr5zp/v/5c03C3//88/y1GEEvUYNIDW1aIK+erWIrWkUikq3buJ1r1ol27/+KmmOZcsC990n3vysWUDfvvJb+MPUyzwJzZolT/ZOxc2t2BH0OgD2WbYzXfvyIaI6APoDeMnfiYhoJBGlE1H6kWiYODsIjCd6xRUaQ7dijaFXqSJ/uNEQcjGdooD9kMvf/gbcfXfxbTCCXRQPvVIleQHiXbdrJ4/qffsCTz4p9bATV//mGyn7m9+49xGJl/7FF87mo69bJ++pqcC77xaurxE6I3yAhGfWri3coZ6dDTz6qNvb9WTlSuCSS4AOHYpnc4UKcg4TR58+XVITV6+Wxc8XLxYb7r038LlMpsvmzdL4vv66/H516vg9LCTYEXTyss/zz2EqgEeY2W+3BjPPYOY0Zk6rbhKEY4TsbOlMqVlTOq50gi7BKugJCfKIGm0hl2rVROgDdQYuXy4CVNTOOYMvQbfjoRvv3LB0qXQwvvWWdHCWKyc2BmLpUvE4PTvvuncHfvnFPWeME6SnS4bOk0/K/8qHHxb83pug//a3In7/+lfBso8+Cvz739Kh6O3J5NNPJd/c82moKFx7rTRGhw5JiOTGG6VBHTtWGsEJE2RfIGrUkP+HzZuBFSsk9OQ5AMkp7Ah6JgDrQ0YKAM9xUGkA5hPRHgADAUwnoltDYWC0kJ0tHpPx/NRLF7KyxEMyYmV3cNFPP7knTwolRtCrVHHvq15dxNxfOiWzpJXl5MikS8XBeKTeps8FghP0WrXcnWiXXCICvXhx4MZp6VIRb+PtG4oSR2cGpk4FfvzRXvl168Q779NH7r1n2GXTJqlX5crufW3aAMOHy3V27JB9GzZIVklamgyz95wGNzNTBgaZ+HlxufZaacyHDZO/o/Hj3d917gw88YS97BQid6bLzJnyP3HTTaGxMRB2BH0dgMZE1JCIEgEMBvC+tQAzN2TmBszcAMBCAGOYeUmojY0kZnh7pAaqRCtmlKihRg17gj55svzjhDqF7uhREbEyZdz77PxmBw+6s0mKmm1hMIIeTMiF2T2oyB8DBoiHbcIa3ti9W8TEGm4xtGwp9ycYQZ81S0Za/va3gZ9eLl4EvvsOuOoq+Q0GDZLG5dQpdxlrhouVf/5T7tnYsXI/7r9f/uc++UQ83EmTCnr7n34q78WNnxs6d5brf/ihxNRN41cUmjeXv6MPPpC/c+vfo5MEFHRmzgXwR0j2yhYAbzPzJiIaRUSjnDYwWjAeugp6QUxDZ6hZ017IJSNDcsZDNbTdYJ2Yy2BnxsWMDPfn4gq68cCD6RQ9eFDuRyBBv/lmEQd/YZelS+Xdm6CXKiXZM3YFfe9eEfOUFHmqev11/+W3bpWG8aqrZHvIEBnJ+d57sp2XB2zZ4l3Qa9WSEMvixdIR+cUXEm6pVElmS2zVSkZ8LlsmIc9PP5UngJYt7dUlEElJ7hG6jz5avHM1by734eLFwFkxocRWHjozL2fmJszciJknufa9xMyFOkGZeRgzLwy1oZHG00PXkItghv0b7IZcTFhj167Q2mMd9m+w0wgbe6pXl0f94lAUD90zw8UXlSpJiGHRIt8dm0uXyhwpV1zh/fvu3SVUEajhzcsTMWKWKQM6dZKBOP7y4s2TQ1qavHfpIsPyZ8yQcNaePfLuTdABaTzq1wdefFHyyocPl/3lykmqYWKiNFQpKcCSJXIvStlSMXvcd59c09/AITuY+l19tfcpApxCR4raRD1073gLueTkFHzE9iQry92ZunNnaO05erSgPYC9CboyMkQsbrpJPHSrWK5eLR2S3sIN3rztonSK2hV0QMIuO3d6j2mfOCHx5ptv9n28iTkbT94X06dL1seUKWLXM8/IE9WUKb6PWbdOMkaaNpXtUqVEpNescc94CPgW9KQkiaNXrCheuTVm3bSpNAhLlognfe4ccNtt/usQLAMGSIjJc9KtYElNlcb3gQdCYFQQqKDbgNntoRuxUEEXvIVcAP/en7XT0QlB9/TQ7fxm27fLXCft20u5X35xf/fvf0vWg6eQ/fOfMpuftSwQuFPUn4feoIFvGw233CKCs2hR4e+WL5dwhLdwi6F9exlPYSbs8sbHH0ssu08fGRkJiIgOGCBpfL7SCNPT5fxWr/mBB6RRrFwZeOEF2WdS+7xx660SHuvcufB3iYlS/0WLJEQ1MErHpFetKv8b/fuH97oq6DbIyZF/ksqVxctKTlZBBwo2dAY7g4uMoJcpE56QS/ny8vIUXisZGfJo3LatbJs4+vHjMtikfHkZGGLCMYsWSfghO1tCAVZMNk1SUsH9Vg9961Z5HJ84UTz/3bvl3pUrF7iONWpI2GTGjILzpXz7LTBqlIRaunTxfTwRcNdd4sl7m7b2nXfEw2/cWGLmVm/1P/+RBmnSpMLHnT8v98fEz61ccw2wfj3w0kuSzuiZfeOJnU7E4nrRThMJ+1TQbWDNtQYit5J8IC5etLeqfKg4edI906LBznwu27eLB9epU2g9dDMNqmfIBQB69pT5q73NXnjxotjRuLGkzxG5hXvFCmnM58+X3/3OOyWscPfdMnlWq1aFV8J5911pADyFzXjoS5fKsenpMiz/uuukAbETbjE895y8d+okjcvXX8t5qlaVzkJ/c40AUg9AJoqyMnOm5Hx36CCCbxpoQ+PGUvdXXy3cyfzjjyLqJn7uSUKCDMx5/HF7dVSCRwXdBmYYs8mbNQNVoo3x4/17ZqHGs6ED7IVcMjIktNCsWWg99I0b5d3Eb6288YZ0tt16a+Fr7tsnDWGTJvL01bix20NfskTq1Lcv8NprknLXpYvEeBcvliyOr7+W4eyAdBjOny+hgAoVCl7HeOhz54oXvXWrnPObb4IX9HbtpEFo2VLiyNdeK43p559LJ2QgGjaU1Lw5c9z9BUuXyjqZN9wgIRdfXvTDD0vmyvPPF9xvOkS9eehKeFBBt0GseOjffis5wEWdnS9YrPO4GOx66I0bA5dfLl7eyZOhscd4394atSpVJN3t4kWJLx8/XtAeQGwCpEPr++9F5JcvlxVtEhJE6B54QMIBixZJmt2gQXLMQlde16JFUh+TnWGlcmUJqQwZIil59epJjnJ6uuRS33prcPWtXVti0yNHypPC558HN5PfXXdJo5KeLg3SPfdIQ7F4sTxh+KJZM7knzz9f8G9t3Tp5QrDTD6A4gwq6Dbx56NEo6GaEXbhWHbfOtGhITBTPzpegM7sFvVEj2RcqL33tWmkkLrvM+/dNmojwZmRI1orBCLpJL2vbVrIpFi8WcbbOkDdlinQIduwo21dcISJowi6vvSbe79VXF75+hQryd/PmmwUFs1kzGTxz++3B1zkpSUJJ335bcLFiOwwaJL/Xa6+5h9YvWFA4O8cbjzwiv/+sWbL93nvi7ffuHf2x7XhGBd0GRriMoFetGn2CfuaMu9MvHPNdA95DLoCEKHwNGDp8WETSeOhAaOLozCLoZmCIL669VjzRt9+WewaIwF9yiXjcgLtjdOJE2W8dWk4EXHppwXMOGiRhk88/lw7U4cN950b783zDTeXK8rTy4osSNnrlFd+565506SL3+tln5alk0CBp2F5+2VmbFf+ooNvAeOjWkMvJk/YXlQ0HJu0NCJ+ge5vZEBBBXLPGe962GZHZpInbQ7cK+unT7tBIMOzZI56znT6EO+6QPPkVK2TbPDEYz9II+ubNMhmTZ7aKJybsctddco577gnO9khibL33XrkvwTB2rGTJ3HabiPlHH0nfghI5VNBtcOyYxFCTk2U7GkeLGlEsUyb8Hrp1kiVAcpV9TfNqjVdXrCiNpDXk8tRT4jX26hXcSvBmYQI7gt6zpww2MpODZWS44+eA9AOY8IWdBQlM/vq+fWK3nU7JaOHmmyUOP21a0Y5NS5N8cRXz6EAF3QbZ2SI8xoOLxtGiRtB79gyvoCcnFx5AYxZV8La6zvbtklJXv75sN2pU0ENftkwEcf16WXVm1qyCsyQySwOwfHlBL/6rryRG7WsEopXSpcWrXLZMzr17d+Hh2amp0ojbnSXPxL+9dYZGM0RFn362VCm57199pWIeLaig2+DYsYJeaLQKeoUK8s/5888F10J8802Z2yPU2S9ZWd5zvpOTZcGCxYsLzzeyfbvEzk2e9OWXuz30/fslu2T0aOCHHyRz43e/k3vfpIl7kYBGjURo//tf93nXrpWcbLuL795xh+Ssv/CCNAxWDx2QcMJ//1v46cMX994rQ+OjdeSiUyQmaidoNKGCbgPjoRuicfj/zp0idGbmOWumy9y5MhmTv/Uo8/IKe72B8BwlaqV/f2lYvvuu4H4zItPQqJGEVnJz3VOj9u0rQv/555L9MWmSeN779skTyPTpkkP9739LI3XihAxqCSYHv3t3yYYxA3Q8PfRrrpGGxS4VKwJ/+UtoFlpQlKKigm4DMzGXIVo99EaNCq9nePase6Fef1OfvvWWeL2vvGL/mv4E/Te/kUdyayOSlyeplVZvuFEjEfN9+6RBSUkRzxwQb7t3b/eUqj/+KCMbR4+WIeiHD4uH/e23cu5AGS5WEhLEmzb9IJ4euqLEIiroNvAUrmhbtcjMBdKokeRAlyvnFvQvv5T0vE6dZOY8Xx2NJp/4P/+xv7yer5ALII3eNdcUFPT9+8UWq3ia1MWtW2V0Yt++9h7hu3aVgT6TJ4tnT+TODbeLyeqwrkSlKLGMCroNPD30xETJRY4WDz0zU0S4USPxips1cwv6Rx9J5svMmRLP9jbD3u7dIvZXXy2C728WPgNz4bnQPenfX0I/27bJtueITMCdujh7tqSC9u0b+NqGiRPFhqlTxav3zA8PRJcuEpNv0kTjwEp8oIIeALMWpadwRdNoUZMlYsTRrGcIiNfbrZvs69FDwi6eHZWzZ4ugzZkj+cT/+pf3BXmtrFol9Tc5294wQ9nN6jomB90q6LVrSwP57rvS8ASzPuRVV8kQ9IsXizaHTalSMsCoKCl7ihKNqKAH4PhxEUDPbIdoFHQzyq9FCxk1umWLTFh1ww2yf9gwKWudcTAvT0S+d29JF3z8cSnz1lvy/ZEjMuWp58x6EyfKyMq77/ZtV926kqP82GOSZTNtmqQz1rUsOZ6QIGGi3FwJ0Zhcf7s8+aQ0BNddF9xxhi5dZGZBRYkHVNAD4GvwTDQN/9+5U0TNTMxkOkanTpV3I+i33SZD2a2do6tWSZjFrHvYr5/kf0+cKKl49epJJ+TAgW6v/csvpaN17NjAoygXLpQ49xVXSCPTpUvhYfEmjl6UldHbtJHO0XAvJKAo0YgKegA8h/0bomkK3Z07xcs1OdgmdfGNN2ReldatZTs5WYapL1ggixhcvCidoZUqucMjpUoBf/ubxLtnz5bh7E89JUP5n3hCykycKKMpR44MbFvt2sBf/yqDeI4dc6/UbsWEioKJn1upVElj4IoCAAGmwVd8eejVqolnmJcX2kVqi4JJWTTUqyee+OnT4llb7Rs3Tgbh3H67zBtuvHOrpz1woAjwVVe5p8PNyJDYelKSxOWfeir4iaZ83acRI2QovqYOKkrxUA89AL489I4dZaThsmXBn/PMGZmVLhQhG+bCgl6qlHvNRhNuMTRtKpknCxaIOF+44F4z0kAk4Q8j5oCMmmzVSlbYqVIluEE3gWjbVs6rXraiFA8V9AD48tBvu03mI3n66eDOt2OHdBSOGiXD460LLRSFo0dlpKRV0AF3HP366wsfk5AgHvr338s0t/4yVQzlykmYplo16Tj1XI1HUZTIo4IeAM/FLQylS8siCV9+KXNJB4JZBtmkpcmQ+H/8Q0Y+9uvnnpe7KHimLBoefFCyU6xetidEEuqwS9OmMjjogQeCNlNRlDAQc4L+2WciikZonebYMfFOvWVzjBghQu/LSz97VhYA6NdPhHXAAMn2+O47CTG88YYsRWZWiykKvgS9dWvJUgk1dlZjVxQlMsScoFepIlOrvvSSvfLMkq3xzjuyAEKweE7MZSU5GbjvPllI2AyasTJxoiyom5Ehc0fPnCkevVlzccgQWZfx/fflPJ4DfuxgBD2YBYYVRYlTmDkir/bt23NRuf565po1mc+cCVx22TJmkUp5NWnC/Nln9q5z6hRz3brM3bv7LnPwIHPZsswjRxbcv2uX7B86NPB1xo0T2557zp5dGzYwP/ss89ixzM2bM9epY+84RVFiHwDp7ENXY1LQP/1ULH/5Zf/lzp0TAW/alHntWubJk5lTUpg7drR3HSO0X3zhv9zIkcyJiWKX4bbbmMuXZ963L/B1Ll5k7t+fmYh56VLZ9+uvzB98wLxnT8GyH37InJQkdiUmSn3Gj7dXH0VRYp+4E/S8POa0NOYrrmDOzfVdbsoUqeEHH7j3Pfus7Nu0yf81tmxhLlOG+Z57Attz+DBzy5YitCtWMK9eLdd48klb1WFmeRpo1445OZm5WzfmhAQ5R7lyUo/cXHnaSExkTk1lzsyU+6AoSski7gSdmfntt8X6d97x/v3hw8wVKzLfcENB4Tt0iLl0aeaHH/Z97rw85muvZa5UScrb4cgREdrEROb69Znr1WPOybFbGyEzU54m2rdnfvRRaRxuvlnqmZoqDUz79sxZWcGdV1GU+CEuBT03Vzz0q67y7qmOGiVerjdP/JZbJAZ//rz3c7/+utyZF14IzqZjx8QegHnBguCO9UVeHvObbzJXq8bcqRNzdnZozqsoSmziT9BjLsvFkJAgGSTr1kn2iJUPPgBmzADGjHGPmLQyfDhw6JB7yTPDpk0yp8mwYTIDX7Bpf5Ury1wlH38sc6aEAiLJhtm3T1IcK1UKzXkVRYk/YlbQAckDv+EGGXVpVsb56isR07ZtZS1Kb/TtK3nhr70m2ydOyERTrVrJ7IMTJ4ow211w2EqFCjKVa6iHsScluRdWVhRF8UZMC3qZMrIwQocOMjjnxRcl37tuXVmf0tfw9DJlZBbBpUslP71VK+DVV2V05a5dMttgsPNyK4qiRJqYFnRAZhX84AOZqW/MGJkB8KOP/A95ByTskpsrc5qULSue/bPP6tqSiqLELrYEnYj6ENE2ItpBROO8fD+UiDa6XmuJqE3oTfVNlSoStx4xQt7NSEx/tGghMfIHHwQ2bJBFlBVFUWIZkk5TPwWIEgBkALgOQCaAdQCGMPNmS5kuALYwczYR3QhgAjP7XYM9LS2N09PTi2u/oihKiYKI1jNzmrfv7HjoHQDsYOZdzHwewHwAt1gLMPNaZjbTZf0PQEpxDPbFvHnifZcqJe/z5jlxFUVRlNjEjqDXAbDPsp3p2ueL3wFY4e0LIhpJROlElH7Ec9XhAMybJ5koe/fKrCx798q2irqiKIpgR9C9JeB5jdMQUU+IoD/i7XtmnsHMacycVj2YibghK8fn5BTcl5Mj+xVFURR7a4pmAqhr2U4BsN+zEBG1BjATwI3MHPLlk3/+Obj9iqIoJQ07Hvo6AI2JqCERJQIYDOB9awEiqgdgEYC7mNnLzODFp1694PYriqKUNAIKOjPnAvgjgI8AbAHwNjNvIqJRRDTKVezvAKoCmE5EG4go5OkrkyYVXmW+fHnfo0EVRVFKGgHTFp2iKGmL8+ZJzPznn92rCB07Jl76pEnA0KEOGKooihJFFDdtMWoYOhTYsweYM0cWVs7K0owXRVEUQ0wJukEzXhRFUQoTk4KuGS+KoiiFiUlB14wXRVGUwsSkoGvGi6IoSmFiUtCHDpUVierXl4Uk6teXbc1yURSlJBOza+AMHaoCriiKYiUmPXRvWGdirFZNXjoro6IoJYmY9dCtmJkYTSpjlmUmGZOjDqhHryhKfBMXHrq3vHQrmqOuKEpJIC4E3U7+ueaoK4oS78SFoNvJP9ccdUVR4p24EHRveelWiCSWrh2kiqLEM3Eh6J556VWryguQbTOhpE7ipShKPBMXgg64Z2LMywOOHpVX/fpuMTfk5AB33qneuqIo8UfcCLo3/HWE7t0L3HWXePAq7oqixANxLeiBOkI1FKMoSjwR14IeqLPUiuaqK4oS68S1oFs7S+2gmTCKosQycS3ogLuzdO5ce966hl8URYlV4l7QDZ7eOpHvshp+URQlFikxgg64vXVmWWjaXyhGwy+KosQaJUrQrRhxDyTqmtqoKEqsUGIF3RAoE0ZTGxVFiRVKvKAHkwkTaJSpdZEN9egVRQk3xJ5j48NEWloap6enR+TavmjQQDxxO5g5YurXdy9ObV1kAxDPX9c6VRQllBDRemZO8/ZdiffQrQQzEMkzFPPnPxdeZMNXtox68oqiOIEKuoVgUhut5OQUXPbOiplPxog4kXS07t0rjYLG5hVFCRUq6B4Ek9poB2ZZsHrECHc4x84MkOrFK4oSLBpDt4HnItROYmLz1nncrfvNPO/HjsnkY5MmaYxeUUoSGkMvJkUNxRQFI+Ke7azZzsqSlwnXmDz5atXk5enR+/L09QlAUeIQZo7Iq3379hyrzJ3LXL8+s8hqdL7Kl2cePVre7e6fO7dg/Yjkvaj7FUUJPQDS2YeuqqAXg7lzvQtj1aqBBZco8qLv+UpI8G6b2ba73zQOVqGvWlVe/hoDX2UC/QaBGhk71w53QxSpa0d74xvt9kUDKugO4u0P0JvQlylTWFS8lYunl79Gy1dj4K2MpyCPHu1+QrLbyATTQNlpGHx99hQhb8f7u7adRi1Ym/xd19eTmZ36hbrR9OUgFbUBdqKx92dDsPevqBRb0AH0AbANwA4A47x8TwCmub7fCKBdoHPGi6D7wu4fnzV8UxRx0pczr+Lce9N4h+K38xR6J/4efD2ZFefeBGo0fTU4wdgX6FxONPb+bEhMDO7+FVXciyXoABIA7ARwOYBEAD8AaO5Rpi+AFS5h7wTgm0DnjXdBLwrBeBRO/GPrS1/6Cu/L+gRiF3+CbifLpQOAHcy8i5nPA5gP4BaPMrcAeMN1vf8BqEREtYLvoi3ZmBz4vDx5N+mI1v1Hj8qL2Z0nTyTpjCalMVgSEkJUAUVRgiLUay/YEfQ6APZZtjNd+4ItAyIaSUTpRJR+5MiRYG1VPPAm9N5WZipfHhg92vv+uXOB2bMLf2dSMz1TNAPtVxQlOMxo8lBgR9C9/atyEcqAmWcwcxozp1WvXt2OfUqQWHPmieR9xgxg+nTv+4cO9X7MnDmFnwIC7ffWmBihtz5B+GoM/JXxR6BGxs61I9EgRera5nrR/mQW7faFinr1QngyX7EY8wLQGcBHlu3xAMZ7lHkZwBDL9jYAtfydV2Po8YmdzuBgy/jKcilKRkWgfgoguM43b589O8cCZbAUpY8kWJt8Xdczq6SoHY3F7Yz0Flf2Z19xO3CLY7ev/dZMtqJkGdkFxewULQ1gF4CGcHeKtvAocxMKdop+G+i8KuhKtFLcXOhQ5VIXN0ffCVtD0WgWNe2zOOcKld3FHccQir8Nf4JO8r1/iKgvgKmQjJdZzDyJiEa5PPyXiIgAPA9Jb8wBMJyZ/U7UEktzuSiKokQL/uZyKW3nBMy8HMByj30vWT4zgPuKY6SiKIpSPHRyLkVRlDhBBV1RFCVOUEFXFEWJE1TQFUVR4gRbWS6OXJjoCIC9QRxSDcBRh8yJZkpivUtinYGSWe+SWGegePWuz8xeR2ZGTNCDhYjSfaXqxDMlsd4lsc5Ayax3Sawz4Fy9NeSiKIoSJ6igK4qixAmxJOgzIm1AhCiJ9S6JdQZKZr1LYp0Bh+odMzF0RVEUxT+x5KEriqIoflBBVxRFiRNiQtCJqA8RbSOiHUQ0LtL2OAER1SWiVUS0hYg2EdGfXfurENEnRLTd9V450raGGiJKIKLviWiZa7sk1LkSES0koq2u37xzCan3g66/75+I6C0iSoq3ehPRLCI6TEQ/Wfb5rCMRjXdp2zYiuqE41456QSeiBAAvALgRQHMAQ4ioeWStcoRcAH9h5maQOeXvc9VzHIBPmbkxgE9d2/HGnwFssWyXhDr/H4APmflKAG0g9Y/rehNRHQD3A0hj5paQ6bgHI/7q/TpkKnErXuvo+h8fDKCF65jpLs0rElEv6LC3SHXMw8wHmPk71+eTkH/wOpC6znYVmw3g1ogY6BBElAJZIGWmZXe81/lSAFcDeBUAmPk8M/+KOK+3i9IAyhFRaQDlAexHnNWbmdcAOOax21cdbwEwn5nPMfNuADsgmlckYkHQbS1AHU8QUQMAbQF8A6AmMx8ARPQB1IigaU4wFcBYAHmWffFe58sBHAHwmivUNJOILkGc15uZfwHwDICfARwAcJyZP0ac19uFrzqGVN9iQdBtLUAdLxBRMoB3ATzAzCcibY+TENHNAA4z8/pI2xJmSgNoB+BFZm4L4DRiP8wQEFfc+BbIcpa1AVxCRHdG1qqIE1J9iwVBzwRQ17KdAnlMizuIqAxEzOcx8yLX7kNEVMv1fS0AhyNlnwN0BdCPiPZAQmnXEtFcxHedAfmbzmTmb1zbCyECH+/17g1gNzMfYeYLABYB6IL4rzfgu44h1bdYEPR1ABoTUUMiSoR0ILwfYZtCjmtd1lcBbGHmKZav3gdwj+vzPQDeC7dtTsHM45k5hZkbQH7Xz5j5TsRxnQGAmQ8C2EdETV27egHYjDivNyTU0omIyrv+3ntB+orivd6A7zq+D2AwEZUlooYAGgP4tshX8bV6dDS9APQFkAFgJ4DHIm2PQ3XsBnnU2ghgg+vVF0BVSK/4dtd7lUjb6lD9ewBY5voc93UGkAog3fV7LwFQuYTU+x8AtgL4CcAcAGXjrd4A3oL0EVyAeOC/81dHAI+5tG0bgBuLc20d+q8oihInxELIRVEURbGBCrqiKEqcoIKuKIoSJ6igK4qixAkq6IqiKHGCCrqiKEqcoIKuKIoSJ/w/ol0vT71Q/2IAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "acc = history.history['acc']\n",
    "val_acc = history.history['val_acc']\n",
    "loss = history.history['loss']\n",
    "val_loss = history.history['val_loss']\n",
    "\n",
    "epochs = range(1,len(acc) + 1)\n",
    "\n",
    "plt.plot(epochs,acc,'bo',label='Training acc')\n",
    "plt.plot(epochs,val_acc,'b',label='validation acc')\n",
    "plt.title('Training and validation accuracy')\n",
    "\n",
    "plt.legend()\n",
    "\n",
    "plt.figure()\n",
    "\n",
    "plt.plot(epochs,loss,'bo',label='Training acc')\n",
    "plt.plot(epochs,val_loss,'b',label='validation acc')\n",
    "plt.title('Training and validation loss')\n",
    "plt.legend()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "3476fbfa",
   "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": "iVBORw0KGgoAAAANSUhEUgAAAXQAAAEICAYAAABPgw/pAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAAsTAAALEwEAmpwYAAA0cElEQVR4nO3deZhT5fXA8e9hk1VEQEV2FEU2EVDApUBFZHFXlK2KVqk7tW7g1qq1orUq/sSFKmIFAXepImrrgrhUQAFZFBGGRVCGZVCEEZg5vz9OAmGYZDIzyWRycz7Pk2eSe29u3puZOXnzLucVVcU551z6q5DqAjjnnEsMD+jOORcQHtCdcy4gPKA751xAeEB3zrmA8IDunHMB4QHdASAib4nIRQk4TxMR2SoiFRNRrmQRkR4isiYJ5x0mIrMiHm8VkRbxHFuC10rI76yQ804Qkb8m+rwu+SqlugCu5ERka8TD6sCvQF7o8R9UdVK851LVvokok6quAmom4lxBoKoJeS9E5C/A4ao6NOLcCfmdueDwgJ7GIoOFiGQBl6rqfwoeJyKVVHVXWZbNOVf2vMklgMLNCSJys4j8ADwjInVE5A0RyRaRzaH7jSKe84GIXBq6P0xEZonIA6FjV4hI34hjm4vITBH5WUT+IyJjRWRiaF8zEVERqRRx3rtF5OPQ8e+ISL2Ic10oIitFZKOI3C4iWSLSK8p19ReRL0XkJxFZHaq1hveFX/ciEVklIhtE5NaI/dVCTQmbRWQxcGyM9+8JEXmgwLbXReRPofsjReS70PUsFpGzY5xLReTw0P26IjItVP7PgcMKHDsmdF0/ichcETkptL0PcAtwQagJZ34hv7MKInJb6L1cLyL/EpHa8bw3RRGRy0RkmYhsCpX/0NB2EZGHQq+3RUQWiEjb0L5+offmZxH5XkRuiPf1XMl5QA+uQ4ADgabAcOx3/UzocRNgO/BojOd3Ab4B6gH3A0+LiIT2PQ98DtQF/gL8roiyDAYuBg4CqgA3AIhIa+AxYAjQAKgNNIxxnl+AC4EDgP7AFSJyVoFjTgSOBE4G7hCRo0Lb/4wF0MOAU4FYbc/PY8FTQuWsA/QGpoT2fwecFCrvncBEEWkQ43xhY4Fc7FovCd0izQY6YL+354EXRaSqqs4A/gZMVdWaqnp0IeceFrr1BFpgzV4Ff7/R3puoROS3wL3A+aFyr2TP+9Ab+A1wBPY7uQDYGNr3NNbsVwtoC7xX1Gu50vOAHlz5wJ9V9VdV3a6qG1X1ZVXdpqo/A/cA3WM8f6Wq/lNV84BnsX/mg0WkCVa7vUNVd6jqLGBaEWV5RlWXqup24AUsaAGcB/xbVWep6g7gDiBqciFV/UBVv1LVfFVdAEwu5BruDF3vfGA+EA5+5wP3qOomVV0NPBKjvB+FynFSRDk/VdW1oXK8qKprQ+WYCnwLHBfrDRDrJD4Xe99+UdWF2PsaeX0TQ7+nXar6D2A/LADHYwjwoKouV9WtwChgYPibUki096ao845X1S9U9dfQebuJSDNgJ1ALaAWIqi5R1XWh5+0EWovI/qq6WVW/iPM6XCl4QA+ubFXNDT8Qkeoi8mToK/lPwEzgAIk+GuWH8B1V3Ra6WxM4FNgUsQ1gdRFl+SHi/jb2dJoeGvnc0Dk3EoWIdBGR98WajbYAl2PfIIr9WlhNs1BqGeumAINCmwYDuzuYQ81E80QkR0RysBpowXIUVB/rs4paBhG5XkSWhJovcrBvAEWdN+zQAudbGXq9gyO2RXtv4j5v6MNiI9BQVd/DvgWMBX4UkXEisn/o0HOBfsBKEflQRLrFeR2uFDygB1fBmu71WG2vi6ruj31VBhCKZx1woIhUj9jWuGRFZB0Q2Y5fDWvGieZ57NtAY1WtDTxB/OVfV6CcTYo4fjJwnog0xZqfXg6VsSnwT+BqoK6qHgAsjKMc2cCuaGUItZffjH2TqBM675aI8xaVFnUt1pwWee5dwI9FPK8oe51XRGpgv6PvAVT1EVXtBLTBml5uDG2frapnYs1sr2HfzFySeUDPHLWwdvMcETkQa1MuNlVdCcwB/iIiVUI1r9NLWKaXgNNF5HgRqYK1R8cKjLWwbwe5InIcVnOO1wvAKLHO4UbANbEOVtUvsSD8FPC2quaEdtXAgms2gIhcjNXQYwo1Xb2CvW/VQ/0Hke34tbAAnA1UEpE7gP0j9v8INBORaP+zk4HrxDqsa7Knzb20o5ueBy4WkQ4isl/ovP9T1SwROTb0raky1r+RC+SF/i6GiEhtVd0J/MSe4bQuiTygZ46HgWrABuAzYEYpzjUE6IZ99f4rMBUbA18sqroIC6xTsBr0z8D6GOe6ErhLRH7G2tuLU+u7E2s6WAG8AzwXx3MmA72woBYu82LgH8CnWJBtB3wcZxmuxpo5fgAmYJ3UYW8DbwFLQ+XMZe/mmRdDPzeKSGHt0eOxa5qJXWMuRXxoxUNV/wvcjn1DWYd1Kg8M7d4f+7ayOVTmjUB4dNDvgKxQ897lwFBc0okvcOFKS0SmAl+raolq/RHnqQnkAC1VdUUiyuZcJvEauiu20Fftw0Jjn/sAZ2LtpCU51+mhJogaWO3uKyArYYV1LoN4QHclcQjwAbAVG/53RajNuSTOxDre1gItgYHqXxudKxFvcnHOuYDwGrpzzgVEypJz1atXT5s1a5aql3fOubQ0d+7cDapav7B9KQvozZo1Y86cOal6eeecS0siEnWWsze5OOdcQHhAd865gPCA7pxzAVGuVizauXMna9asITc3t+iDnUuwqlWr0qhRIypXrpzqojhXIuUqoK9Zs4ZatWrRrFkz9qyl4FzyqSobN25kzZo1NG/ePNXFca5EylWTS25uLnXr1vVg7sqciFC3bl3/dujSWrkK6IAHc5cy/rfn0l25C+jOBYkqjB8Pv/yS6pK4TOABPUWysrJ4/vndabaZMGECV199dYnP98EHH3Daaafts33evHlMnz692Odbu3Yt5513XpHH9evXj5ycnGKfv6CsrCzati1ynYi0s3Ah/P738PjjqS6JywRpHdAnTYJmzaBCBfs5aVJRzyg/Cgb0ZIkV0Hftir6YzaGHHspLL71U5PmnT5/OAQccUNLiBd7G0Aqpr7+e2nK4zJC2AX3SJBg+HFautK+1K1fa49IE9V9++YX+/ftz9NFH07ZtW6ZOnQpYmoJbbrmFbt260blzZ7744gtOPfVUDjvsMJ544gnARknceOONtG3blnbt2u1+brTtI0eO5KOPPqJDhw489NBDgNWK+/TpQ8uWLbnpppt2l+udd96hW7dudOzYkQEDBrB161YAZsyYQatWrTjxxBN55ZVX9rmeHTt2cMcddzB16lQ6dOjA1KlT+ctf/sLw4cPp3bs3F154IVlZWZx00kl07NiRjh078sknnwB715gnTJjAOeecU2jZmjVrxoYNG8jKyuKoo47isssuo02bNvTu3Zvt27cDMHv2bNq3b0+3bt12vxex5ObmcvHFF9OuXTuOOeYY3n//fQAWLVrEcccdR4cOHWjfvj3ffvtt1N9ZebF5s/385BNYvz61ZXEZQFVTcuvUqZMWtHjx4n22RdO0qaqF8r1vTZvGfYp9vPTSS3rppZfufpyTkxN6rab62GOPqarqH//4R23Xrp3+9NNPun79eq1fv/7u5/bq1Ut37dqlP/zwgzZu3FjXrl0bdfv777+v/fv33/1azzzzjDZv3lxzcnJ0+/bt2qRJE121apVmZ2frSSedpFu3blVV1dGjR+udd96p27dv10aNGunSpUs1Pz9fBwwYsNf5Is971VVX7X785z//WTt27Kjbtm1TVdVffvlFt2/frqqqS5cu1fDvZcWKFdqmTZuYZQu/N9nZ2bpixQqtWLGifvnll6qqOmDAAH3uuedUVbVNmzb68ccfq6rqzTffvPu8kSJf74EHHtBhw4apquqSJUu0cePGun37dr366qt14sSJqqr666+/6rZt26L+zkqqOH+D8Rg/fs/f5tNPJ/TULkMBczRKXE3bGvqqVcXbHo927drxn//8h5tvvpmPPvqI2rVr7953xhln7D6mS5cu1KpVi/r161O1alVycnKYNWsWgwYNomLFihx88MF0796d2bNnR91emJNPPpnatWtTtWpVWrduzcqVK/nss89YvHgxJ5xwAh06dODZZ59l5cqVfP311zRv3pyWLVsiIgwdGv+SjWeccQbVqlUDbDLXZZddRrt27RgwYACLFy+Ou2wFNW/enA4dOgDQqVMnsrKyyMnJ4eeff+b4448HYPDgotd1njVrFr/73e8AaNWqFU2bNmXp0qV069aNv/3tb9x3332sXLmSatWqxfydlQfh7oV69eC111JZEpcJ0jagN2lSvO3xOOKII5g7dy7t2rVj1KhR3HXXXbv37bfffgBUqFBh9/3w4127dqFRFgqJtr0wkeetWLHi7vOecsopzJs3j3nz5rF48WKefvppoOTD7GrUqLH7/kMPPcTBBx/M/PnzmTNnDjt27Ii7bPGWv7iiPWfw4MFMmzaNatWqceqpp/Lee+/F/J2VB5s3gwgMGgTvvAOh1jLnkiJtA/o990D16ntvq17dtpfU2rVrqV69OkOHDuWGG27giy8KW1y9cL/5zW+YOnUqeXl5ZGdnM3PmTI477rio22vVqsXPP/9c5Hm7du3Kxx9/zLJlywDYtm0bS5cupVWrVqxYsYLvvvsOgMmTJxf6/KJeZ8uWLTRo0IAKFSrw3HPPkZeXF/c1x6NOnTrUqlWLzz77DIApU6YU+Zzf/OY3TAp1hixdupRVq1Zx5JFHsnz5clq0aMG1117LGWecwYIFC0r1OysLOTlQuzacfTb8+qsFdeeSpVxN/S+OIUPs5623WjNLkyYWzMPbS+Krr77ixhtvpEKFClSuXJnHizHW7Oyzz+bTTz/l6KOPRkS4//77OeSQQ6Jur1u3LpUqVeLoo49m2LBh1KlTp9Dz1q9fnwkTJjBo0CB+/fVXAP76179yxBFHMG7cOPr370+9evU48cQTWbhw4T7P79mzJ6NHj6ZDhw6MGjVqn/1XXnkl5557Li+++CI9e/bcq/aeKE8//TSXXXYZNWrUoEePHkU2i1x55ZVcfvnltGvXjkqVKjFhwgT2228/pk6dysSJE6lcuTKHHHIId9xxB7Nnzy7x76wsbN4MBxwAJ50EderYaJdzzkl1qVxQpWxN0c6dO2vBBS6WLFnCUUcdlZLyuOTZunUrNWvWBGD06NGsW7eOMWPGpLhUhUv03+Dpp8P338MXX8BFF8G//22jXSqlbVXKpZqIzFXVzoXtS9smF5c+3nzzTTp06EDbtm356KOPuO2221JdpDITrqEDnHWWPZ45M5UlckHm9QSXdBdccAEXXHBBqouREjk5cOSRdr93b2tP/+c/4be/TWmxXEB5Dd25JMrJ2VNDr1EDLrsMXnyxdMNrnYvGA7pzSRTZ5AJwzTX285FHUlIcF3Ae0J1Lkh07YNs2G90S1qQJDBhgzS4//ZS6srlg8oDuXJKEZ4kWzF12/fUWzMePL+sSuaDzgJ4iZZU+tzTnmTZtGqNHjy70uPAwxGhycnJ47LHHdj+ONx1vPHr06EHBIa/lUTigF5xi0LmzjUsfMwZiJLx0aea99+Cww+C//01dGTygp0hZpc8tjTPOOIORI0eW6LkFA3q86XiDJJxpsbDswn/6E2RlQSFJMl0a+vRTOOMMWL4cLrkkdSkePKBHCFr6XIAuXbqwaNGi3Y979OjB3Llz+fzzzzn++OM55phjOP744/nmm2/2eW7kt4YVK1bQrVs3jj32WG6//fbdx2zdupWTTz6Zjh070q5dO14PJf4eOXIk3333HR06dODGG2/cKx1vtPS4sdL0RjN58mTatWtH27ZtufnmmwHIy8tj2LBhu9/z8Pv7yCOP0Lp1a9q3b8/AgQOLPHdpRauhg004atkSRo+2XIwufX35JfTtCw0awMsvw+rVNoM9JaKlYUz2raj0uSNGqHbvntjbiBGx01IGMX3ugw8+qHfccYeqqq5du1ZbtmypqqpbtmzRnTt3qqrqu+++q+ecc46q6l7liky9e/rpp+uzzz6rqqqPPvqo1qhRQ1VVd+7cqVu2bFFV1ezsbD3ssMM0Pz9/r3S4qvGlx42VpjdS9+7ddfbs2fr9999r48aNdf369bpz507t2bOnvvrqqzpnzhzt1avX7uM3b96sqqoNGjTQ3NzcvbYVlMj0uVOmWNrcRYsK3//007b/rbf23v7YY6oTJqjm5SWsKC5JvvlGtV491SZNVFeutG1XX60qohrKGJ1wBDF9bjIEMX3u+eefz4svvgjACy+8wIABAwBLyjVgwADatm3Lddddt1ctvjAff/wxgwYNAtid2hasQnDLLbfQvn17evXqxffff8+PP/4Y81zR0uNGew+imT17Nj169KB+/fpUqlSJIUOGMHPmTFq0aMHy5cu55pprmDFjBvvvvz8A7du3Z8iQIUycOJFKZTD3PlaTC8DQodC48d4J5V59Fa68EoYNgy5d4H//S3YpXW4ubNlS/Odt3mzftAD+8589mV7/9jf7vf7+93v+BgDy8+Gjj+APf7CafDKU25miDz9c9q8ZTsU6ffp0Ro0aRe/evbnjjjuA1KfPLZhNcd68eXGlz23YsCF169ZlwYIFTJ06lSeffBKA22+/nZ49e/Lqq6+SlZVFjx49ijxXYa83adIksrOzmTt3LpUrV6ZZs2bk5ubGPE+s9ySeNL1FnadOnTrMnz+ft99+m7Fjx/LCCy8wfvx43nzzTWbOnMm0adO4++67WbRoUVIDe6wmF4AqVeDGG+Haa+0fvUULuPRS6NTJto0cCV27Qr9+Fjj69Stdemi3t/x8eO45GDXKOqc/+mjPrN6i7NwJ558PK1ZYZ2jLlnv21aoFTz5pzTB160Lr1nDMMTBrlvWbVK++9/GJ5DX0CEFMnwswcOBA7r//frZs2UK7du0Aq6E3bNgQsLbropxwwgm7U99Oiljnb8uWLRx00EFUrlyZ999/f3eNOtb1RUuPW1xdunThww8/ZMOGDeTl5TF58mS6d+/Ohg0byM/P59xzz+Xuu+/miy++ID8/n9WrV9OzZ0/uv/9+cnJydvdFJMvmzRa0q1aNfszvfw/168Pdd8OFF1pt8fnn7f4338Btt8GSJXDFFdC0qeVV9zb3kvnhBxuB8vLLMG6cfQMaNgwaNbKc9aecEv8M3uuus1r5uHFw4on77u/TxzpK77zTPoTffReOOMI+QH78EW64IaGXtlu5raGnQhDT5wKcd955jBgxYq/OzJtuuomLLrqIBx98kN/GkVhkzJgxDB48mDFjxnDuuefu3j5kyBBOP/10OnfuTIcOHWjVqhUAdevW5YQTTqBt27b07duXq666avdzoqXHLa4GDRpw77330rNnT1SVfv36ceaZZzJ//nwuvvhi8vPzAbj33nvJy8tj6NChbNmyBVXluuuuS/ri1jk5VjuP9UWqenUb8RLObPzUU/aPD1bTu/tuuOsuC+5PPmnfXHv0sK/tLj7LlsF998Gzz1rNOqxhQ5g40T4kFyyA7t0tqH/0ERx0UOHnys21YDx2rP0cNiz663btarey5OlznYuQyL/B88+3QPH117GP++kn+6rfvTtMnhz9AyA/32p+n3wC8+fbmGcX25/+ZOP9K1e25qzzzoMDD7QP2kMOse1hs2ZZArXq1S2g77ef/TztNMuU+csvMHCgvffXXQd//ztUrFj21xQrfa7X0J1LknANvSj77w/ffmvJu2LV5itUsNmlbdtabvUPP0xNQEkXb78NDz1k79Xo0RbAYznxRHjrLfuWlJtrK0wtW2b9Gddea8G/dm144w3o379srqG44groItIHGANUBJ5S1UKnD4rIscBnwAWqmlmzSJwrICfHOsXiUcTE290aNYL/+z9rY//HPyCOofoZaccOGDECDj/cmqribdHr3t1ukZYutdFHq1bBLbdYU015VWRAF5GKwFjgFGANMFtEpqnq4kKOuw94uzQFUtUSL37sXGkkuvlx82YLKIk2dCi89hrcfrs1B7RunfjXSHePPmr9Dm+8EX8wj+aIIyA0Z63ci2eUy3HAMlVdrqo7gCnAmYUcdw3wMrC+pIWpWrUqGzduTPg/lnNFUVU2btxI1VhDUoopMhd6IonA449bU81FF8WXD+bFF62d/tVXE1+e8ubHH210Sb9+5bdpJFniaXJpCKyOeLwG6BJ5gIg0BM4GfgscG+1EIjIcGA7QpJABtY0aNWLNmjVkZ2fHUSznEqtq1ao0atQoIedS3TcXeiIddBA89ph1vN5/vzUFRPPhh1arr1DBFqgeMcKeU6VKcsqWSvn5ls1y+3ZrP8808QT0wto/ClahHwZuVtW8WM0lqjoOGAc2yqXg/sqVK9O8efM4iuRc+fbLL5CXF1+naEkNGGAB/S9/sYlHoSkGe1m0yEZotGgB779vnYNjxsDHH8M77yS3fGVtxQq4+GL7ALvttj3DPzNJPE0ua4DGEY8bAWsLHNMZmCIiWcB5wGMiclYiCuhcOipq2n+ijB1rQXnQIFizZu99WVk2W7FqVRu9ccghNo795Zfhiy8s6AXFM89A+/aWKOuZZ2zsfiaKJ6DPBlqKSHMRqQIMBKZFHqCqzVW1mao2A14CrlTV1xJdWOfSRVHT/hOlXj2YNAlWrrSUAbNmWXPPU09ZgMvJgTffhGbN9jznnHPgqqvgiScsAJYXeXkle97EiZay9thj4auvbLJPpo6rKDKgq+ou4Gps9MoS4AVVXSQil4vI5ckuoHPpKNpqRcnQq5cl8apdG3r2hOOPt8WoO3e2iU0dO+77nLvusiGVV1+dulQCqvYBdNNNcNRRUK2adfIuWGD7N22yqfJXXGEfPsuX73uOzz6zCUM9eti484zPdRMtDWOyb4Wlz3UuKF5/3VLjzplTdq+5ebNq//6q1aurPvpo0el3w+l7//WvMinePh54wF6/cmXVXr1UL71UtUYN29amjWrFina/WjX7CaqHHaY6apTq4sWqq1apHnywaosWqhs2pOYaUoEY6XPL1dR/54LiX/+y2uayZWU7RV/VZjlWq1b0sfn50K2bNdcsXmxT4stKbq41A7VpY6s2hTNVb95sE4Heesu+aZx9tn3T+PZb68R9801LipWXZ7luwGrpmTQW36f+O1fGyqpTtCCR+II52DDGsWMtcHbtCtOmQSi3WtKFsw4+//yeYA7W5zBypN0iHXmk3a65xp43ebKVd+TIzArmRfH0uc4lQbgNPTJYlUedO1tK2ZwcSyc7fXrhx23atHemwtLIz7e0BR07Wpt/cR18MPzxj5aHvHfvxJQpKDygO5cEmzdbk0AZLIxUaiedBHPm2Fj1006zTtYHH7QRMOPGWW6TunWheXNbjWfDhtK93htv2LT8G2/M3NEoyeIB3bkkiDfTYnnRpImNOBk1yhaCuP56q0H/4Q/WxHHrrTYS5dZbbXm1//u/+M8d7tIM+/vfbbGO885L/HVkujSoPziXfpKVxyWZatSw9U3vuccyC374oaXq7dBhT0160SKrWY8YYR8CZxaW1QnLpfL447Btm92qVbNl2A4/3D44Hn44Pb69pBsf5eJcEnTvbkHwgw9SXZLE277drm/xYltso337vfe//bYtxNGrl6UjqF7dmqC+/BLmzbN+hW+/jT9lsNubj3JxGUfVao8//2wzKDt1smnwZVVrzsmxNucgqlbN0vced5zlkJk9e8+SbZs326zNNm3g3//edz3VXbusczXekTiueLwN3QXSunUWUL791nJjDx5syZomTSqbmZHJzLRYHhx6KLz+OmRn25T7yZPtfb3mGli/3tbvLCwTcaVKHsyTyQO6C6TFoeVXnnvOaumzZlmNeehQOPVUm0yTTOnWKVoSnTrZJJ+6de0Ds00b+8C87Tbb58qeB3QXSIsW2c/WrW0tyBNOsPbeRx+1mYWnnbbvuOrvvktMsqpdu+xDJMg19LDjj7cml/Hj7VtJ166xc7O75PKA7gJp0SKrOYbbdsEWVL7qKmsOWLjQZkmGrV9viwR36WLtw/FYscI6PQs24WzZYj+DXkMPq1jR8pCvXm0jYypXTnWJMpcHdBdIixZZE0BhE1fOOstGYfz5z9bWnp9vKVc3b7bnDBhgOcPBcrEMH25D7gYPhnvvtVmOXbvaRJyePS2zYeQycGWZabE8qVQpmKsgpRMf5eICR9Xa0AcOLHy/CDzyiI2xvukmG2f91ltWYx861EbDXHCBBf233rIa54knWpPN5Ml2jmOOgfvusynx991nnYOTJ1t+lHfftWMypYbuyg8P6C5w1q2zWnKbNtGPadnSJsjccw9MmWJZ/a64woL9jBk2HO+DD+BPf7Jbgwb2vC1b4KefbLZkWJMmlle8fXubVbl1q2Uu9KRRrqx5QHeBE+4QjRXQwTrvJk60JpenntrTPFOrliWs+vVXmxQTqXbtfRNuXXmltdXfe69Npjn7bGuK8eYHV9Y8oLvACQ9ZLCqgV68On39unXoFc4FXrLhvMI/lvPM8N4lLPQ/oLnDCI1zq1y/62MhRMM6lOx/l4gIn1ggX54LMA7oLFNU9Ad25TOMB3QXKunU2EsUDustEHtBdoERO+Xcu03hAdwmlCkuX2ljsWHbt2rOQciLFO2TRuSDygO4S4rXX4NxzbdTIkUfawgarV0c//pJLbOr88uWJLcfixVCvno9ecZnJA7orNVUL0LNmQf/+tsDwpk02yebHH/c9fuFCm9CTk2NT7HfsSFxZvEPUZTIP6K7U1q615pPbb4cJE+C662D6dFizxoL6xo17H3/nnbb82D//aavNjxyZmHKER7h4+7nLVD6xyJVauN26bds92044AaZNsxp7r162zuRBB8H8+fDSSxb8L73UHj/0kDXTbN1qeVTmz7dZmjVr2jT7xo1tlfhGjSxRVoUKtv+cc/aezXn33TbCpVu3sr1+58oLXyTaldqDD8L111tO8YKzM2fMsMDbpIllIbz2Wnj/fcslXqcO5ObaIgnhhSXatLGAvGOHBfjNm20F+lWr9l2Q4uijLc3tYYdZLpbLLoMLL7RvCT6pyAWVLxLtkmrRIqt9FzbVvk8fq52fdpotKvzDD9bkEk4tW7UqvPGGJcPq0WPvLIaR8vNhwwbIy7P7c+bYogqdOtk6lvfea68VmWTLuUzjNXRXal26QI0a8N570Y/58ktbyzMvz2rn++9f+tfNyrKEWHPnQufOVvOvWbP053WuPItVQ/dOUVcq+fk2VDCy/bwwxxxjbeP/+19igjlAs2Y2suaf/7SFKDyYu0znTS6uVFatsrbueIYKhheJSKSqVa1z1TnnNXRXSgsX2s+iaujOueSLK6CLSB8R+UZElonIPqOGReRMEVkgIvNEZI6InJj4orryyKfaO1d+FNnkIiIVgbHAKcAaYLaITFPVxRGH/ReYpqoqIu2BF4BWySiwK18WLoSGDTNvhXvnyqN4aujHActUdbmq7gCmAGdGHqCqW3XPcJkaQGqGzrgyt2iRN7c4V17EE9AbApFpltaEtu1FRM4Wka+BN4FLCjuRiAwPNcnMyc7OLkl5XTmSlwdLlnhzi3PlRTwBvbBpGvvUwFX1VVVtBZwF3F3YiVR1nKp2VtXO9eNZ8NGVa8uX20xPr6E7Vz7EE9DXAJHz9xoBa6MdrKozgcNEpF4py+bKOR/h4lz5Ek9Anw20FJHmIlIFGAhMizxARA4XsQnXItIRqAJs3OdMLlDCI1yOOiq15XDOmSJHuajqLhG5GngbqAiMV9VFInJ5aP8TwLnAhSKyE9gOXKCpyingyszChdC8uc/QdK68iGumqKpOB6YX2PZExP37gPsSWzRX1lassNS0Bx5oQXrDBpsJ+sMPlgHxwAP3HLtjB8yb5x2izpUnPvXfkZ8PN98MDzywZ5uILRgRVq0aDBli0+xnzbIc5t9/D5dfXvbldc4VzgN6MSxYYIszNG0Khx9u62ZG1lrT0c6dlkf82WctWHfrZisMbdkCBx9secxr14bJk+G55yw9LUDPnvD009C7d2rL75zbwwN6MYwebYEtrGZNy+XdvXvqyhSPOXMsX/jmzfDzzzZ+vEULOOIIS2s7Y4blKL/99ui5xHv0sHO88gq0b2+5zZ1z5YvnQy+GDh2s1vr44/Dtt7Z2ZlaWLbXWq1eqS1e4jRttZZ/cXBuNUquWbf/uOxtHnp8Pjz4KV1yR2nI65+LjKxYlQF4efPONBe4WLez2wQdwyim2Gs8rr0C/fqku5d5UrRll/Xr47DPo2HHv/Tt3wvbtictP7pxLLU+fG6eVK/fUcsMOOshW6WnbFs4+G2bOTF35CvPEE/Daa9ZUVDCYgy247MHcueDwGnqcFodySxacRFO3ri1+fPzxcNZZ8OmntoJ9Knz4obXx778/VK8O991n62z+8Y+pKY9zrmx5QI/TkiX2s7BZkXXqwJtvQteu1uzy2WeFL5icbNdfD199ZWPJc3Nt0s+ECfbYORd8/q8epyVLrEM0vFp9QS1awL//DWvXQt++MGWKjdMuzM6ddr7t2xNXvqVLbbHk0aPtvL/+atsOPjhxr+GcK9+8hh6nJUugdevYx3TpYk0eF10EgwbZtqZNraZ86KH2YTB/vg0jzM2FKlXsOT17wvDhtlBESU2ebEMOL7jAHlepUvJzOefSkw9bjIOqrcgzdCiMHVv08bt2WeCeORM+/xxWr7aa+8aNNlW+a1cby714sY2UmTvXhhM+/LB9GEQbCx6rfK1a2QfCe++V4AKdc2kjUMMWN22yzr4777QV38vCunXw00/xZxWsVAk6dbJbPJYtg0sugYsvtpmoDz9sM1Hj9eWX1rxy443xP8c5Fzxp14Y+Ywbcf7+NBy+rRY9idYgmwuGHW0394Yethn3EETa2fcYMm/hTlOeftyGI556bnPI559JD2gX0wYNh6lRrpujaFb7+OvmvGQ7oRbWhl0aFCjBihNXWb7/d2tn79rVr/Pzz6M/Ly7P28759o3fYOucyQ9oFdIDzz4f334etWy2Z1Ny5yX29xYstQdUhhyT3dcA6T++80yYyjR9v7e9du1oCrS++sOafXbv2HP/RR9Y+P3hw8svmnCvf0rpTNCvLRoj88gt88knx2p2Lo2dPG5Xy6afJOX8sP/0Ed90FY8bsCeQi1olarZoNT9y506b3V69e9uVzzpWtQHWKRmrWzNqZTzgBTj0VPv44ObXoJUtSl6dl//0tT/mVV9qCEj/8AD/+aOltt2+HbdssE6IHc+dcWgd0sGn2b74Jv/2tBd3XXrMc3pFycy2x1uLFlmXw9NMtA2E8Nm2yAJrM9vN4hBOCOedcNGkf0MEm57z8Mpx5pgW9Cy6Aq66yDsaXX4a337amibDRoy07YjyLMyR7hItzziVKWnaKFqZPH8tRPmKE5Sc/4QSbpPPll/CHP9jImAULrN398MOhf38b7lcUD+jOuXQRiBp6WJMm8I9/2LC/11+3IHzssfvOvPzwQ6vNDxliwwNvuMFGlxSkavurVrUp/M45V54FpoYe6YADrHZ+3HGFT6OvXds6Uy+5xEaPNGtmszTffRfWrLFA/tln1i7/5JP2s2LFsr4K55wrnkAG9HhUrWqLHH/77Z4mmd69oXFjGxLYrZt1oj7yiLW3O+dceZfW49ATKSfH2tu//tpujRrZOps1a6a6ZM45t0dgx6En0gEH2ASinj1TXRLnnCuZjG1ycc65oPGA7pxzAeEB3TnnAsIDunPOBYQHdOecCwgP6M45FxAe0J1zLiA8oDvnXEB4QHfOuYCIK6CLSB8R+UZElonIyEL2DxGRBaHbJyIS5/IRzjnnEqXIgC4iFYGxQF+gNTBIRAqu37MC6K6q7YG7gXGJLqhzzrnY4qmhHwcsU9XlqroDmAKcGXmAqn6iqptDDz8DGiW2mM4554oST0BvCKyOeLwmtC2a3wNvFbZDRIaLyBwRmZOdnR1/KZ1zzhUpnoBeyBIRFJpzV0R6YgH95sL2q+o4Ve2sqp3r168ffymdc84VKZ70uWuAxhGPGwFrCx4kIu2Bp4C+qroxMcVzzjkXr3hq6LOBliLSXESqAAOBaZEHiEgT4BXgd6q6NPHFNJMm2XJxFSrYz0mTkvVKzjmXfoqsoavqLhG5GngbqAiMV9VFInJ5aP8TwB1AXeAxsUU8d0VbUaOkJk2C4cNh2zZ7vHKlPQZb7Nk55zJd2ixB16yZBfHCNG0K99zjgd05F3yxlqBLm5miq1ZF3xeurXsTjHMuk6VNQG/SJPb+bdvg1lvLpizOOVcepU1Av+ceqF499jGxavHOORd0aRPQhwyBceOsvTyaomrxzjkXZGkT0MGCelYWTJy4b229cmXYutWHNDrnMldaBfSwyNq6CNStaz83bgRV7yR1zmWmtAzosKe2np8PNWvCjh177/dOUudcpknbgB4pWmeod5I65zJJIAJ6tM5QVW9Pd85ljkAE9FhDGr093TmXKQIR0Isa0ujt6c65TBCIgA57OkmlsOzteHu6cy74AhPQw6K1p/ukI+dc0AUuoBfWni5ibeneQeqcC7LABfSC7ekiNtoFvIPUORdsgQvosKc9vWnTPcE8zDtInXNBFciAHuYTjpxzmSTQAd07SJ1zmSTQAd07SJ1zmSTQAd07SJ1zmSTQAR28g9Q5lzkCH9DDvIPUORd0GRPQPSOjcy7oMiage0ZG51zQZUxA94yMzrmgy5iADp6R0TkXbBkV0MN8wpFzLogyMqD7hCPnXBBlZED3CUfOuSDKyIAOPuHIORc8GRvQw3zCkXMuKDI+oPuEI+dcUGR8QPcJR865oMj4gO4TjpxzQRFXQBeRPiLyjYgsE5GRhexvJSKfisivInJD4ouZXD7hyDkXBEUGdBGpCIwF+gKtgUEi0rrAYZuAa4EHEl7CMuQTjpxz6SyeGvpxwDJVXa6qO4ApwJmRB6jqelWdDexMQhnLjE84cs6ls3gCekNgdcTjNaFtxSYiw0VkjojMyc7OLskpksonHDnn0lk8Ab2wlmUtZFuRVHWcqnZW1c7169cvySmSziccOefSVTwBfQ3QOOJxI2BtcopTfviEI+dcuoknoM8GWopIcxGpAgwEpiW3WKnnE46cc+mmyICuqruAq4G3gSXAC6q6SEQuF5HLAUTkEBFZA/wJuE1E1ojI/skseLL5hCPnXLoRLdhQXEY6d+6sc+bMSclrx2vSJGszX7my8P1Nm1p7u3POlRURmauqnQvbl/EzRWPxCUfOuXTiAT0O3p7unEsHHtDj4O3pzrl04AE9Dp7AyzmXDjygx6mo9nRPD+CcSzUP6MUUK1GXN78451LJA3oxxWpPB2t+GTrUa+vOubLnAb2YimpPD/PaunOurHlAL4HIBF6xeGepc64seUAvhaKaX8A7S51zZccDeil484tzrjzxgF5K4eaXiROL7iy96CKoUMFr7M655PCAniDx1Nbz8ixdgNfYnXPJ4AE9geLtLAUf3uicSzwP6EkQT2dp2MqV8Lvf2QxUD+7OudLwgJ4Ekc0vIlCxYuzjfSFq51wieEBPknDzS34+PPts/DV2b4pxzpWUB/QyEO/wxkjeFOOcKy4P6GUk3uGNkbwpxjlXHB7Qy1jB2nq0dLwFeVOMc64oHtBTIFxbV4XnnvOmGOdcYnhAT7HSNsV4cHfOhXlALydK2hTjwd05F+YBvRwpTVMMeHB3LtN5QC+nStIUE8mDu3OZxwN6OVfSpphIhQX3evXs5tkfnQsOD+hpIFpTTGmC+8aNdgtnf7z44j0B3oO9c+nJA3qaSWRwj7Rz554AXzDYe63eufTgAT2NJSu4FxStVu+B3rnyxQN6QJRVcI9UnEAfz33/MHCudDygB1AqgnukwgJ9PPfj+TDwoO9cdB7QA66w4C4CdevaLXy/SpVUl9QU9WEQqwO3LO+n6wfLpElW9nS+BheDqqbk1qlTJ3Xlx8SJqk2bqoqo1q1rN7DHFkr9VvAWfm/C75eIvYdXXLHve1ke7hf2+6xcOTVlatrU/uZi/R3G+15GO1dx/+5Lc56yBMzRKHFVNFwlKmOdO3fWOXPmpOS1XfwmTYJbb4VVq+DAA23bxo1Ws0/Rn44LiPDfUN269rg0f1cFz7Vp056/12j3C3u9kpxn0yZo0gT69YPp0/f+XynqOffcY9+ii3etMldVOxe6L56ALiJ9gDFAReApVR1dYL+E9vcDtgHDVPWLWOf0gJ7ePNA7V3rVq9vEweIE9VgBvcg2dBGpCIwF+gKtgUEi0rrAYX2BlqHbcODx+Ivn0lHkEnsbNtitqHb6WPeh7DptnSsvtm2zilGixNMpehywTFWXq+oOYApwZoFjzgT+FWri+Qw4QEQaJK6YLl0UFujjuV+cD4Py0oHrXCKsWpW4c8UT0BsCqyMerwltK+4xiMhwEZkjInOys7OLW1YXcPF+GIwfX7JvAYm6D+n9bSJcdv9wLB+aNEncueIJ6IX96RZsJY3nGFR1nKp2VtXO9evXj6d8zu2jpN8CEnU/1reJpk3hiitS+4ET637TplZ21dR9OELsD8Twvnjey6LOFY/w81PxIV29unWMJky04S/hG9ANeDvi8ShgVIFjngQGRTz+BmgQ67w+bNG5zFXYMNmSDh2Mdq7iDp8s7XmKO1y1pMMkKc2wRRGpBCwFTga+B2YDg1V1UcQx/YGrsVEuXYBHVPW4WOf1US7OOVd8sUa5VCrqyaq6S0SuBt7Ghi2OV9VFInJ5aP8TwHQsmC/Dhi1enKjCO+eci0+RAR1AVadjQTty2xMR9xW4KrFFc845Vxyey8U55wLCA7pzzgWEB3TnnAuIlCXnEpFsYGUxnlIP2JCk4pRnmXjdmXjNkJnXnYnXDKW77qaqWuhEnpQF9OISkTnRhuoEWSZedyZeM2TmdWfiNUPyrtubXJxzLiA8oDvnXECkU0Afl+oCpEgmXncmXjNk5nVn4jVDkq47bdrQnXPOxZZONXTnnHMxeEB3zrmASIuALiJ9ROQbEVkmIiNTXZ5kEJHGIvK+iCwRkUUiMiK0/UAReVdEvg39rJPqsiaaiFQUkS9F5I3Q40y45gNE5CUR+Tr0O++WIdd9Xejve6GITBaRqkG7bhEZLyLrRWRhxLao1ygio0Kx7RsRObU0r13uA3qca5oGwS7gelU9CugKXBW6zpHAf1W1JfDf0OOgGQEsiXicCdc8Bpihqq2Ao7HrD/R1i0hD4Fqgs6q2xbK3DiR41z0B6FNgW6HXGPofHwi0CT3nsVDMK5FyH9CJb03TtKeq61T1i9D9n7F/8IbYtT4bOuxZ4KyUFDBJRKQR0B94KmJz0K95f+A3wNMAqrpDVXMI+HWHVAKqhdZZqA6sJWDXraozgU0FNke7xjOBKar6q6quwFKQx1xLIpZ0COhxrVcaJCLSDDgG+B9wsKquAwv6wEEpLFoyPAzcBORHbAv6NbcAsoFnQk1NT4lIDQJ+3ar6PfAAsApYB2xR1XcI+HWHRLvGhMa3dAjoca1XGhQiUhN4Gfijqv6U6vIkk4icBqxX1bmpLksZqwR0BB5X1WOAX0j/ZoYihdqNzwSaA4cCNURkaGpLlXIJjW/pENDXAI0jHjfCvqYFjohUxoL5JFV9JbT5RxFpENrfAFifqvIlwQnAGSKShTWl/VZEJhLsawb7m16jqv8LPX4JC/BBv+5ewApVzVbVncArwPEE/7oh+jUmNL6lQ0CfDbQUkeYiUgXrQJiW4jIlnIgI1qa6RFUfjNg1DbgodP8i4PWyLluyqOooVW2kqs2w3+t7qjqUAF8zgKr+AKwWkSNDm04GFhPw68aaWrqKSPXQ3/vJWF9R0K8bol/jNGCgiOwnIs2BlsDnJX6VaKtHl6cbtl7pUuA74NZUlydJ13gi9lVrATAvdOsH1MV6xb8N/Tww1WVN0vX3AN4I3Q/8NQMdgDmh3/drQJ0Mue47ga+BhcBzwH5Bu25gMtZHsBOrgf8+1jUCt4Zi2zdA39K8tk/9d865gEiHJhfnnHNx8IDunHMB4QHdOecCwgO6c84FhAd055wLCA/ozjkXEB7QnXMuIP4f0aA6o5gyGXUAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "def smooth_curve(points, factor=0.8): \n",
    "    smoothed_points = []\n",
    "    for point in points: \n",
    "        if smoothed_points: # not empty\n",
    "            previous = smoothed_points[-1]\n",
    "            smoothed_points.append(previous * factor + point * (1 - factor)) \n",
    "        else:\n",
    "            smoothed_points.append(point) \n",
    "    return smoothed_points\n",
    "\n",
    "plt.plot(epochs,smooth_curve(acc),'bo',label='smoothed training acc')\n",
    "plt.plot(epochs,smooth_curve(val_acc),'b',label='smoothed validation acc')\n",
    "plt.title('Training and validation accuracy')\n",
    "plt.legend()\n",
    "\n",
    "plt.figure()\n",
    "\n",
    "plt.plot(epochs,smooth_curve(loss),'bo',label='smoothed training loss')\n",
    "plt.plot(epochs,smooth_curve(val_loss),'b',label='smoothed validation loss')\n",
    "plt.title('Trainging and validation loss')\n",
    "plt.legend()\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "87b223a4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 1000 images belonging to 2 classes.\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/var/folders/pn/m2bqkk014sv8108lk7tl4xy80000gn/T/ipykernel_19487/2020560468.py:7: UserWarning: `Model.evaluate_generator` is deprecated and will be removed in a future version. Please use `Model.evaluate`, which supports generators.\n",
      "  test_loss, test_acc = model.evaluate_generator(test_generator,steps=50)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "test acc: 0.9300000667572021\n"
     ]
    }
   ],
   "source": [
    "test_generator = test_datagen.flow_from_directory(\n",
    "        test_dir,\n",
    "        target_size=(150,150),\n",
    "        batch_size=20,\n",
    "        class_mode='binary')\n",
    "\n",
    "test_loss, test_acc = model.evaluate_generator(test_generator,steps=50)\n",
    "print('test acc:', test_acc)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "499a6d8b",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}