cats-and-dogs(image classification).ipynb

{"nbformat":4,"nbformat_minor":0,"metadata":{"colab":{"name":"HW04_20204119_SangwonPark.ipynb","provenance":[],"collapsed_sections":[],"authorship_tag":"ABX9TyMMECX7k7B77MvCEDwuqTwt"},"kernelspec":{"name":"python3","display_name":"Python 3"},"language_info":{"name":"python"},"accelerator":"GPU"},"cells":[{"cell_type":"markdown","metadata":{"id":"msKZ8oWDEPhD"},"source":["# **Catas and dogs dataset**\n","---"]},{"cell_type":"markdown","metadata":{"id":"1EJryjjmE7nV"},"source":["- 강아지와 고양이 이미지 이진분류 task\n","- 사전 훈련된 컨브넷 사용"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":35},"id":"2UqoIcu5Hzvc","executionInfo":{"status":"ok","timestamp":1634534466769,"user_tz":-540,"elapsed":404,"user":{"displayName":"상원박","photoUrl":"https://lh3.googleusercontent.com/a/default-user=s64","userId":"16337076995970243718"}},"outputId":"3f84fc4d-1b2f-4bed-cecb-e747b3c12add"},"source":["import keras\n","keras.__version__"],"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"application/vnd.google.colaboratory.intrinsic+json":{"type":"string"},"text/plain":["'2.6.0'"]},"metadata":{},"execution_count":33}]},{"cell_type":"markdown","metadata":{"id":"pdEK96ivG71q"},"source":["## VGG16 model load"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"WceIZp7yFC5l","executionInfo":{"status":"ok","timestamp":1634541582984,"user_tz":-540,"elapsed":6767,"user":{"displayName":"상원박","photoUrl":"https://lh3.googleusercontent.com/a/default-user=s64","userId":"16337076995970243718"}},"outputId":"35921c92-ff01-4119-9b15-01fd67fd8967"},"source":["import keras\n","from keras.applications.vgg16 import VGG16\n","\n","conv_base = VGG16(weights='imagenet', include_top=False, input_shape=(150,150,3))"],"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/vgg16/vgg16_weights_tf_dim_ordering_tf_kernels_notop.h5\n","58892288/58889256 [==============================] - 1s 0us/step\n","58900480/58889256 [==============================] - 1s 0us/step\n"]}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"Rz4CTJN3HTJ9","executionInfo":{"status":"ok","timestamp":1634522420179,"user_tz":-540,"elapsed":297,"user":{"displayName":"상원박","photoUrl":"https://lh3.googleusercontent.com/a/default-user=s64","userId":"16337076995970243718"}},"outputId":"c46803fe-5260-4d4d-bd03-9c3cf1507c60"},"source":["conv_base.summary()"],"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","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","Total params: 14,714,688\n","Trainable params: 14,714,688\n","Non-trainable params: 0\n","_________________________________________________________________\n"]}]},{"cell_type":"markdown","metadata":{"id":"rdrZGsuAJJlJ"},"source":["## Conv_base 모델 확장해 end-to-end로 학습"]},{"cell_type":"code","metadata":{"id":"WpATqjWLISx3"},"source":["from keras import models, layers\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'))"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"19B4BNSTJfUZ","executionInfo":{"status":"ok","timestamp":1634522762688,"user_tz":-540,"elapsed":299,"user":{"displayName":"상원박","photoUrl":"https://lh3.googleusercontent.com/a/default-user=s64","userId":"16337076995970243718"}},"outputId":"740676e1-75c3-421d-8f7e-3a73b61433fc"},"source":["model.summary()"],"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["Model: \"sequential\"\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 (Dense)                (None, 256)               2097408   \n","_________________________________________________________________\n","dense_1 (Dense)              (None, 1)                 257       \n","=================================================================\n","Total params: 16,812,353\n","Trainable params: 16,812,353\n","Non-trainable params: 0\n","_________________________________________________________________\n"]}]},{"cell_type":"markdown","metadata":{"id":"qyZEUAqeJ1Lj"},"source":["## 모델 커파일 전 합성곱 기반층 동결시키기\n","- top layer의 dense 층은 랜덤하게 초기화되었기 때문에 매우 큰 가중치 업데이트 값이 전파\n","- 이는 사전에 학습된 표현을 크게 훼손\n","- 이를 방지하기 위해 동결"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"d7v_b5rrJzyX","executionInfo":{"status":"ok","timestamp":1634541582985,"user_tz":-540,"elapsed":7,"user":{"displayName":"상원박","photoUrl":"https://lh3.googleusercontent.com/a/default-user=s64","userId":"16337076995970243718"}},"outputId":"61a4f496-7f11-4b96-d20e-d0b1e109d142"},"source":["print(\"conv_base 동결 전 훈련되는 가중치 수 : \", len(model.trainable_weights))\n","conv_base.trainable = False\n","print(\"conv_base 동결 후 훈련되는 가중치 수 : \", len(model.trainable_weights))"],"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["conv_base 동결 전 훈련되는 가중치 수 :  30\n","conv_base 동결 후 훈련되는 가중치 수 :  4\n"]}]},{"cell_type":"markdown","metadata":{"id":"UGDtmOJksTbx"},"source":["## 훈련, 검증, 테스트 폴더로 이미지 복사하기"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"18xe6dpwMVV9","executionInfo":{"status":"ok","timestamp":1634541648511,"user_tz":-540,"elapsed":17561,"user":{"displayName":"상원박","photoUrl":"https://lh3.googleusercontent.com/a/default-user=s64","userId":"16337076995970243718"}},"outputId":"55f35eba-dd2f-4867-de07-edcda2649598"},"source":["# 구글드라이브 Colab에 마운트\n","from google.colab import drive\n","drive.mount('/content/drive')"],"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["Mounted at /content/drive\n"]}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"Qhs7fUtttvmt","executionInfo":{"status":"ok","timestamp":1634541650037,"user_tz":-540,"elapsed":6,"user":{"displayName":"상원박","photoUrl":"https://lh3.googleusercontent.com/a/default-user=s64","userId":"16337076995970243718"}},"outputId":"0dc0f62e-261f-4915-f960-5ebed5eab172"},"source":["cd /content/drive/MyDrive/Colab Notebooks/고급딥러닝HW/datasets"],"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["/content/drive/MyDrive/Colab Notebooks/고급딥러닝HW/datasets\n"]}]},{"cell_type":"code","metadata":{"id":"Bknu1hdOGgf6"},"source":["import os, shutil\n","\n","# 원본 데이터셋을 압축 해제한 경로\n","original_dataset_dir = '/content/drive/MyDrive/Colab Notebooks/고급딥러닝HW/datasets/cats_and_dogs/train'\n","\n","# 소규모 데이터셋을 저장할 디렉토리\n","base_dir = '/content/drive/MyDrive/Colab Notebooks/고급딥러닝HW/datasets/cat_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","# 훈련용 고양이 사진 디렉토리\n","train_cats_dir = os.path.join(train_dir, 'cats')\n","# 훈련용 강아지 사진 디렉토리\n","train_dogs_dir = os.path.join(train_dir, 'dogs')\n","# 검증용 고양이 사진 디렉토리\n","validation_cats_dir = os.path.join(validation_dir, 'cats')\n","# 검증용 강아지 사진 디렉토리\n","validation_dogs_dir = os.path.join(validation_dir, 'dogs')\n","# 테스트용 고양이 사진 디렉토리\n","test_cats_dir = os.path.join(test_dir, 'cats')\n","# 테스트용 강아지 사진 디렉토리\n","test_dogs_dir = os.path.join(test_dir, 'dogs')"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"uzt_9S-XFOWT"},"source":["os.mkdir(base_dir)\n","# 훈련, 검증, 테스트 분할\n","os.mkdir(train_dir)\n","os.mkdir(validation_dir)\n","os.mkdir(test_dir)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"id":"n8dbcxmDsIxP"},"source":["os.mkdir(train_cats_dir)\n","os.mkdir(train_dogs_dir)\n","os.mkdir(validation_cats_dir)\n","os.mkdir(validation_dogs_dir)\n","os.mkdir(test_cats_dir)\n","os.mkdir(test_dogs_dir)\n","# 처음 1000개의 고양이 이미지 train_cats_dir에 복사\n","fnames = ['cat.{}.jpg'.format(i) for i in range(1000)]\n","for fname in fnames:\n","  src = os.path.join(original_dataset_dir, fname)\n","  dst = os.path.join(train_cats_dir, fname)\n","  shutil.copyfile(src, dst)\n","\n","# 다음 500개의 고양이 이미지 validation_cats_dir에 복사\n","fnames = ['cat.{}.jpg'.format(i) for i in range(1000, 1500)]\n","for fname in fnames:\n","  src = os.path.join(original_dataset_dir, fname)\n","  dst = os.path.join(validation_cats_dir, fname)\n","  shutil.copyfile(src, dst)\n","\n","# 다음 500개의 고양이 이미지 test_cats_dir에 복사\n","fnames = ['cat.{}.jpg'.format(i) for i in range(1500, 2000)]\n","for fname in fnames:\n","  src = os.path.join(original_dataset_dir, fname)\n","  dst = os.path.join(test_cats_dir, fname)\n","  shutil.copyfile(src, dst)\n","\n","# 처음 1000개의 강아지 이미지 train_dogs_dir에 복사\n","fnames = ['dog.{}.jpg'.format(i) for i in range(1000)]\n","for fname in fnames:\n","  src = os.path.join(original_dataset_dir, fname)\n","  dst = os.path.join(train_dogs_dir, fname)\n","  shutil.copyfile(src, dst)\n","\n","# 다음 500개의 강아지 이미지 validation_dogs_dir에 복사\n","fnames = ['dog.{}.jpg'.format(i) for i in range(1000, 1500)]\n","for fname in fnames:\n","  src = os.path.join(original_dataset_dir, fname)\n","  dst = os.path.join(validation_dogs_dir, fname)\n","  shutil.copyfile(src, dst)\n","\n","# 다음 500개의 강아지 이미지 test_dogs_dir에 복사\n","fnames = ['dog.{}.jpg'.format(i) for i in range(1500, 2000)]\n","for fname in fnames:\n","  src = os.path.join(original_dataset_dir, fname)\n","  dst = os.path.join(test_dogs_dir, fname)\n","  shutil.copyfile(src, dst)"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"uu3d2e-nwFir","executionInfo":{"status":"ok","timestamp":1634545314891,"user_tz":-540,"elapsed":7,"user":{"displayName":"상원박","photoUrl":"https://lh3.googleusercontent.com/a/default-user=s64","userId":"16337076995970243718"}},"outputId":"d30cb9e6-c73b-4ced-b06a-8825a057523a"},"source":["import os\n","\n","print(len(os.listdir(train_cats_dir)))\n","print(len(os.listdir(train_dogs_dir)))\n","print(len(os.listdir(validation_cats_dir)))\n","print(len(os.listdir(validation_dogs_dir)))\n","print(len(os.listdir(test_cats_dir)))\n","print(len(os.listdir(test_dogs_dir)))"],"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["1000\n","1000\n","500\n","500\n","500\n","500\n"]}]},{"cell_type":"markdown","metadata":{"id":"-7E5BCchyYsv"},"source":["\n","## 데이터 증식을 사용해 모델 훈련"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"5TumruPxytFL","executionInfo":{"status":"ok","timestamp":1634546123810,"user_tz":-540,"elapsed":808922,"user":{"displayName":"상원박","photoUrl":"https://lh3.googleusercontent.com/a/default-user=s64","userId":"16337076995970243718"}},"outputId":"8fc601fb-32ca-4230-d97f-6e8c24d7d92c"},"source":["from keras.preprocessing.image import ImageDataGenerator\n","from tensorflow.keras import optimizers\n","\n","train_datagen = ImageDataGenerator(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","# 검증 데이터는 증식되어선 안됨!!!\n","test_datagen = ImageDataGenerator(rescale=1./255)\n","\n","train_generator = train_datagen.flow_from_directory(\n","      #타깃 티렉토리\n","      train_dir,\n","      # 모든 이미지의 크기를 150X150으로 변경\n","      target_size=(150,150),\n","      batch_size=20,\n","      # binary_crossentropy 손실 사용하므로 이진 레이블 필요\n","      class_mode = 'binary'\n",")\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","\n","model.compile(loss='binary_crossentropy',\n","              optimizer=optimizers.RMSprop(lr=2e-5),\n","              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)"],"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["Found 2000 images belonging to 2 classes.\n","Found 1000 images belonging to 2 classes.\n"]},{"output_type":"stream","name":"stderr","text":["/usr/local/lib/python3.7/dist-packages/keras/optimizer_v2/optimizer_v2.py:356: UserWarning: The `lr` argument is deprecated, use `learning_rate` instead.\n","  \"The `lr` argument is deprecated, use `learning_rate` instead.\")\n","/usr/local/lib/python3.7/dist-packages/keras/engine/training.py:1972: UserWarning: `Model.fit_generator` is deprecated and will be removed in a future version. Please use `Model.fit`, which supports generators.\n","  warnings.warn('`Model.fit_generator` is deprecated and '\n"]},{"output_type":"stream","name":"stdout","text":["Epoch 1/30\n","100/100 - 29s - loss: 0.5930 - acc: 0.7175 - val_loss: 0.4862 - val_acc: 0.8350\n","Epoch 2/30\n","100/100 - 27s - loss: 0.4632 - acc: 0.8200 - val_loss: 0.3969 - val_acc: 0.8550\n","Epoch 3/30\n","100/100 - 27s - loss: 0.3976 - acc: 0.8485 - val_loss: 0.3406 - val_acc: 0.8740\n","Epoch 4/30\n","100/100 - 27s - loss: 0.3632 - acc: 0.8555 - val_loss: 0.3137 - val_acc: 0.8770\n","Epoch 5/30\n","100/100 - 27s - loss: 0.3314 - acc: 0.8670 - val_loss: 0.2968 - val_acc: 0.8800\n","Epoch 6/30\n","100/100 - 27s - loss: 0.3166 - acc: 0.8725 - val_loss: 0.2872 - val_acc: 0.8940\n","Epoch 7/30\n","100/100 - 27s - loss: 0.3074 - acc: 0.8730 - val_loss: 0.2759 - val_acc: 0.8870\n","Epoch 8/30\n","100/100 - 27s - loss: 0.2916 - acc: 0.8810 - val_loss: 0.2651 - val_acc: 0.8940\n","Epoch 9/30\n","100/100 - 27s - loss: 0.2801 - acc: 0.8865 - val_loss: 0.2680 - val_acc: 0.8930\n","Epoch 10/30\n","100/100 - 27s - loss: 0.2700 - acc: 0.8955 - val_loss: 0.2530 - val_acc: 0.8980\n","Epoch 11/30\n","100/100 - 27s - loss: 0.2668 - acc: 0.8905 - val_loss: 0.2492 - val_acc: 0.8980\n","Epoch 12/30\n","100/100 - 27s - loss: 0.2596 - acc: 0.9000 - val_loss: 0.2509 - val_acc: 0.9000\n","Epoch 13/30\n","100/100 - 27s - loss: 0.2623 - acc: 0.8900 - val_loss: 0.2441 - val_acc: 0.8990\n","Epoch 14/30\n","100/100 - 27s - loss: 0.2529 - acc: 0.8980 - val_loss: 0.2415 - val_acc: 0.9080\n","Epoch 15/30\n","100/100 - 27s - loss: 0.2367 - acc: 0.9030 - val_loss: 0.2390 - val_acc: 0.9110\n","Epoch 16/30\n","100/100 - 27s - loss: 0.2404 - acc: 0.9015 - val_loss: 0.2440 - val_acc: 0.8970\n","Epoch 17/30\n","100/100 - 26s - loss: 0.2350 - acc: 0.9065 - val_loss: 0.2376 - val_acc: 0.9030\n","Epoch 18/30\n","100/100 - 26s - loss: 0.2279 - acc: 0.9130 - val_loss: 0.2360 - val_acc: 0.9080\n","Epoch 19/30\n","100/100 - 27s - loss: 0.2215 - acc: 0.9135 - val_loss: 0.2365 - val_acc: 0.9060\n","Epoch 20/30\n","100/100 - 27s - loss: 0.2236 - acc: 0.9110 - val_loss: 0.2432 - val_acc: 0.9000\n","Epoch 21/30\n","100/100 - 27s - loss: 0.2202 - acc: 0.9145 - val_loss: 0.2335 - val_acc: 0.9040\n","Epoch 22/30\n","100/100 - 27s - loss: 0.2165 - acc: 0.9150 - val_loss: 0.2356 - val_acc: 0.9070\n","Epoch 23/30\n","100/100 - 27s - loss: 0.2237 - acc: 0.9040 - val_loss: 0.2356 - val_acc: 0.9070\n","Epoch 24/30\n","100/100 - 27s - loss: 0.2174 - acc: 0.9115 - val_loss: 0.2338 - val_acc: 0.9050\n","Epoch 25/30\n","100/100 - 27s - loss: 0.2060 - acc: 0.9185 - val_loss: 0.2318 - val_acc: 0.9070\n","Epoch 26/30\n","100/100 - 28s - loss: 0.1978 - acc: 0.9240 - val_loss: 0.2389 - val_acc: 0.9050\n","Epoch 27/30\n","100/100 - 27s - loss: 0.2090 - acc: 0.9165 - val_loss: 0.2311 - val_acc: 0.9040\n","Epoch 28/30\n","100/100 - 27s - loss: 0.2021 - acc: 0.9165 - val_loss: 0.2360 - val_acc: 0.9090\n","Epoch 29/30\n","100/100 - 27s - loss: 0.2082 - acc: 0.9135 - val_loss: 0.2398 - val_acc: 0.9040\n","Epoch 30/30\n","100/100 - 27s - loss: 0.2011 - acc: 0.9235 - val_loss: 0.2335 - val_acc: 0.9070\n"]}]},{"cell_type":"markdown","metadata":{"id":"83Tt-yMuInhW"},"source":["모델 저장"]},{"cell_type":"code","metadata":{"id":"lK31wLxEIiUI"},"source":["model.save('cats_and_dogs_small_3.h5')"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"sJ0l3_7EIpnY"},"source":["## 결과 그래프"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":545},"id":"o4j6GiUDIo4E","executionInfo":{"status":"ok","timestamp":1634546124640,"user_tz":-540,"elapsed":15,"user":{"displayName":"상원박","photoUrl":"https://lh3.googleusercontent.com/a/default-user=s64","userId":"16337076995970243718"}},"outputId":"8c72712d-ca7d-444f-c9fe-53f8ff8e3577"},"source":["\n","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(len(acc))\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","plt.legend()\n","\n","plt.figure()\n","\n","plt.plot(epochs, loss, 'bo', label='Training loss')\n","plt.plot(epochs, val_loss, 'b', label='Validation loss')\n","plt.title('Training and validation loss')\n","plt.legend()\n","\n","plt.show()"],"execution_count":null,"outputs":[{"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","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"markdown","metadata":{"id":"xBrk8SNrIwNE"},"source":["## 미세 조정(fine-tuning)\n","1. 사전 훈련된 기반 네트워크 위에 새로운 네트워크 추가\n","2. 기반 네트워크 동결\n","3. 새로 추가한 네트워크 훈련\n","4. 기반 네트워크에서 일부 층의 동결 해제\n","5. 동결 해제한 층과 새로 추가한 층 함께 훈련"]},{"cell_type":"markdown","metadata":{"id":"DPeSCVJoJRD3"},"source":["3번까지의 과정은 위에서 특성 추출 시 이미 완료했으므로 4부터 진행"]},{"cell_type":"code","metadata":{"id":"55UttA4KI-oV"},"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"],"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"YRNLKcofJ4gM"},"source":["- 미세 조정 시작\n","- 학습률을 낮춘 RMSProp 옵티마이저 사용\n","- 기반 네트워크에서 학습된 표현을 해치지 않는 선에서 조금씩 수정하기 위해 학습률 낮춤"]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"AcVMJ00wKJK9","executionInfo":{"status":"ok","timestamp":1634549088430,"user_tz":-540,"elapsed":2963796,"user":{"displayName":"상원박","photoUrl":"https://lh3.googleusercontent.com/a/default-user=s64","userId":"16337076995970243718"}},"outputId":"bd960821-19ba-431e-d353-9b5064ebe0db"},"source":["model.compile(loss='binary_crossentropy',\n","              optimizer=optimizers.RMSprop(lr=1e-5),\n","              metrics=['acc'])\n","\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)"],"execution_count":null,"outputs":[{"output_type":"stream","name":"stderr","text":["/usr/local/lib/python3.7/dist-packages/keras/optimizer_v2/optimizer_v2.py:356: UserWarning: The `lr` argument is deprecated, use `learning_rate` instead.\n","  \"The `lr` argument is deprecated, use `learning_rate` instead.\")\n","/usr/local/lib/python3.7/dist-packages/keras/engine/training.py:1972: UserWarning: `Model.fit_generator` is deprecated and will be removed in a future version. Please use `Model.fit`, which supports generators.\n","  warnings.warn('`Model.fit_generator` is deprecated and '\n"]},{"output_type":"stream","name":"stdout","text":["Epoch 1/100\n","100/100 [==============================] - 31s 289ms/step - loss: 0.2064 - acc: 0.9110 - val_loss: 0.2129 - val_acc: 0.9190\n","Epoch 2/100\n","100/100 [==============================] - 29s 287ms/step - loss: 0.1699 - acc: 0.9340 - val_loss: 0.2207 - val_acc: 0.9110\n","Epoch 3/100\n","100/100 [==============================] - 30s 296ms/step - loss: 0.1548 - acc: 0.9365 - val_loss: 0.1962 - val_acc: 0.9160\n","Epoch 4/100\n","100/100 [==============================] - 30s 295ms/step - loss: 0.1288 - acc: 0.9460 - val_loss: 0.2001 - val_acc: 0.9260\n","Epoch 5/100\n","100/100 [==============================] - 29s 288ms/step - loss: 0.1116 - acc: 0.9605 - val_loss: 0.1997 - val_acc: 0.9270\n","Epoch 6/100\n","100/100 [==============================] - 29s 289ms/step - loss: 0.1070 - acc: 0.9605 - val_loss: 0.2017 - val_acc: 0.9220\n","Epoch 7/100\n","100/100 [==============================] - 29s 289ms/step - loss: 0.0957 - acc: 0.9630 - val_loss: 0.2105 - val_acc: 0.9240\n","Epoch 8/100\n","100/100 [==============================] - 29s 289ms/step - loss: 0.0799 - acc: 0.9710 - val_loss: 0.2579 - val_acc: 0.9100\n","Epoch 9/100\n","100/100 [==============================] - 29s 289ms/step - loss: 0.0789 - acc: 0.9690 - val_loss: 0.2010 - val_acc: 0.9260\n","Epoch 10/100\n","100/100 [==============================] - 29s 287ms/step - loss: 0.0603 - acc: 0.9770 - val_loss: 0.2229 - val_acc: 0.9260\n","Epoch 11/100\n","100/100 [==============================] - 29s 291ms/step - loss: 0.0536 - acc: 0.9815 - val_loss: 0.2293 - val_acc: 0.9250\n","Epoch 12/100\n","100/100 [==============================] - 29s 290ms/step - loss: 0.0485 - acc: 0.9820 - val_loss: 0.2946 - val_acc: 0.9070\n","Epoch 13/100\n","100/100 [==============================] - 29s 292ms/step - loss: 0.0406 - acc: 0.9875 - val_loss: 0.2344 - val_acc: 0.9260\n","Epoch 14/100\n","100/100 [==============================] - 29s 290ms/step - loss: 0.0408 - acc: 0.9865 - val_loss: 0.2451 - val_acc: 0.9240\n","Epoch 15/100\n","100/100 [==============================] - 29s 291ms/step - loss: 0.0380 - acc: 0.9840 - val_loss: 0.2126 - val_acc: 0.9290\n","Epoch 16/100\n","100/100 [==============================] - 29s 291ms/step - loss: 0.0376 - acc: 0.9885 - val_loss: 0.2319 - val_acc: 0.9310\n","Epoch 17/100\n","100/100 [==============================] - 30s 295ms/step - loss: 0.0322 - acc: 0.9890 - val_loss: 0.2253 - val_acc: 0.9310\n","Epoch 18/100\n","100/100 [==============================] - 30s 295ms/step - loss: 0.0298 - acc: 0.9915 - val_loss: 0.2119 - val_acc: 0.9330\n","Epoch 19/100\n","100/100 [==============================] - 29s 293ms/step - loss: 0.0242 - acc: 0.9910 - val_loss: 0.2683 - val_acc: 0.9210\n","Epoch 20/100\n","100/100 [==============================] - 29s 291ms/step - loss: 0.0240 - acc: 0.9920 - val_loss: 0.2212 - val_acc: 0.9370\n","Epoch 21/100\n","100/100 [==============================] - 29s 292ms/step - loss: 0.0282 - acc: 0.9905 - val_loss: 0.2334 - val_acc: 0.9290\n","Epoch 22/100\n","100/100 [==============================] - 29s 295ms/step - loss: 0.0215 - acc: 0.9920 - val_loss: 0.2308 - val_acc: 0.9330\n","Epoch 23/100\n","100/100 [==============================] - 29s 291ms/step - loss: 0.0187 - acc: 0.9940 - val_loss: 0.2418 - val_acc: 0.9310\n","Epoch 24/100\n","100/100 [==============================] - 29s 287ms/step - loss: 0.0216 - acc: 0.9935 - val_loss: 0.2274 - val_acc: 0.9330\n","Epoch 25/100\n","100/100 [==============================] - 29s 287ms/step - loss: 0.0140 - acc: 0.9960 - val_loss: 0.2649 - val_acc: 0.9300\n","Epoch 26/100\n","100/100 [==============================] - 29s 288ms/step - loss: 0.0124 - acc: 0.9955 - val_loss: 0.2609 - val_acc: 0.9340\n","Epoch 27/100\n","100/100 [==============================] - 29s 288ms/step - loss: 0.0154 - acc: 0.9945 - val_loss: 0.2523 - val_acc: 0.9300\n","Epoch 28/100\n","100/100 [==============================] - 29s 290ms/step - loss: 0.0126 - acc: 0.9970 - val_loss: 0.2307 - val_acc: 0.9390\n","Epoch 29/100\n","100/100 [==============================] - 29s 293ms/step - loss: 0.0078 - acc: 0.9980 - val_loss: 0.3001 - val_acc: 0.9200\n","Epoch 30/100\n","100/100 [==============================] - 29s 291ms/step - loss: 0.0093 - acc: 0.9965 - val_loss: 0.2576 - val_acc: 0.9340\n","Epoch 31/100\n","100/100 [==============================] - 29s 290ms/step - loss: 0.0112 - acc: 0.9950 - val_loss: 0.3379 - val_acc: 0.9150\n","Epoch 32/100\n","100/100 [==============================] - 29s 288ms/step - loss: 0.0129 - acc: 0.9955 - val_loss: 0.2557 - val_acc: 0.9330\n","Epoch 33/100\n","100/100 [==============================] - 29s 290ms/step - loss: 0.0074 - acc: 0.9965 - val_loss: 0.3129 - val_acc: 0.9310\n","Epoch 34/100\n","100/100 [==============================] - 29s 287ms/step - loss: 0.0121 - acc: 0.9960 - val_loss: 0.2809 - val_acc: 0.9320\n","Epoch 35/100\n","100/100 [==============================] - 29s 291ms/step - loss: 0.0100 - acc: 0.9975 - val_loss: 0.2497 - val_acc: 0.9340\n","Epoch 36/100\n","100/100 [==============================] - 29s 289ms/step - loss: 0.0093 - acc: 0.9975 - val_loss: 0.2376 - val_acc: 0.9370\n","Epoch 37/100\n","100/100 [==============================] - 29s 288ms/step - loss: 0.0080 - acc: 0.9980 - val_loss: 0.2575 - val_acc: 0.9400\n","Epoch 38/100\n","100/100 [==============================] - 29s 290ms/step - loss: 0.0043 - acc: 0.9990 - val_loss: 0.2655 - val_acc: 0.9350\n","Epoch 39/100\n","100/100 [==============================] - 30s 295ms/step - loss: 0.0071 - acc: 0.9970 - val_loss: 0.2820 - val_acc: 0.9320\n","Epoch 40/100\n","100/100 [==============================] - 29s 292ms/step - loss: 0.0072 - acc: 0.9980 - val_loss: 0.2384 - val_acc: 0.9400\n","Epoch 41/100\n","100/100 [==============================] - 29s 294ms/step - loss: 0.0048 - acc: 0.9985 - val_loss: 0.4970 - val_acc: 0.9010\n","Epoch 42/100\n","100/100 [==============================] - 29s 294ms/step - loss: 0.0092 - acc: 0.9970 - val_loss: 0.2767 - val_acc: 0.9340\n","Epoch 43/100\n","100/100 [==============================] - 29s 292ms/step - loss: 0.0118 - acc: 0.9970 - val_loss: 0.2811 - val_acc: 0.9400\n","Epoch 44/100\n","100/100 [==============================] - 29s 293ms/step - loss: 0.0116 - acc: 0.9960 - val_loss: 0.2549 - val_acc: 0.9340\n","Epoch 45/100\n","100/100 [==============================] - 29s 292ms/step - loss: 0.0136 - acc: 0.9970 - val_loss: 0.4736 - val_acc: 0.9110\n","Epoch 46/100\n","100/100 [==============================] - 29s 292ms/step - loss: 0.0067 - acc: 0.9980 - val_loss: 0.2849 - val_acc: 0.9320\n","Epoch 47/100\n","100/100 [==============================] - 29s 293ms/step - loss: 0.0061 - acc: 0.9975 - val_loss: 0.2688 - val_acc: 0.9360\n","Epoch 48/100\n","100/100 [==============================] - 29s 294ms/step - loss: 0.0077 - acc: 0.9985 - val_loss: 0.3178 - val_acc: 0.9290\n","Epoch 49/100\n","100/100 [==============================] - 29s 294ms/step - loss: 0.0020 - acc: 1.0000 - val_loss: 0.2937 - val_acc: 0.9390\n","Epoch 50/100\n","100/100 [==============================] - 29s 293ms/step - loss: 0.0064 - acc: 0.9990 - val_loss: 0.2915 - val_acc: 0.9360\n","Epoch 51/100\n","100/100 [==============================] - 30s 295ms/step - loss: 0.0070 - acc: 0.9985 - val_loss: 0.2966 - val_acc: 0.9340\n","Epoch 52/100\n","100/100 [==============================] - 30s 297ms/step - loss: 0.0042 - acc: 0.9985 - val_loss: 0.2963 - val_acc: 0.9310\n","Epoch 53/100\n","100/100 [==============================] - 30s 295ms/step - loss: 0.0044 - acc: 0.9985 - val_loss: 0.3205 - val_acc: 0.9340\n","Epoch 54/100\n","100/100 [==============================] - 30s 303ms/step - loss: 0.0077 - acc: 0.9980 - val_loss: 0.3330 - val_acc: 0.9310\n","Epoch 55/100\n","100/100 [==============================] - 30s 303ms/step - loss: 0.0098 - acc: 0.9970 - val_loss: 0.3146 - val_acc: 0.9270\n","Epoch 56/100\n","100/100 [==============================] - 30s 297ms/step - loss: 0.0046 - acc: 0.9985 - val_loss: 0.3860 - val_acc: 0.9240\n","Epoch 57/100\n","100/100 [==============================] - 30s 296ms/step - loss: 0.0043 - acc: 0.9990 - val_loss: 0.2845 - val_acc: 0.9390\n","Epoch 58/100\n","100/100 [==============================] - 29s 294ms/step - loss: 0.0020 - acc: 0.9995 - val_loss: 0.2954 - val_acc: 0.9390\n","Epoch 59/100\n","100/100 [==============================] - 29s 292ms/step - loss: 0.0045 - acc: 0.9985 - val_loss: 0.3151 - val_acc: 0.9400\n","Epoch 60/100\n","100/100 [==============================] - 29s 294ms/step - loss: 0.0041 - acc: 0.9985 - val_loss: 0.3288 - val_acc: 0.9340\n","Epoch 61/100\n","100/100 [==============================] - 29s 295ms/step - loss: 0.0026 - acc: 0.9990 - val_loss: 0.3443 - val_acc: 0.9340\n","Epoch 62/100\n","100/100 [==============================] - 29s 294ms/step - loss: 0.0054 - acc: 0.9985 - val_loss: 0.3920 - val_acc: 0.9250\n","Epoch 63/100\n","100/100 [==============================] - 29s 292ms/step - loss: 0.0055 - acc: 0.9985 - val_loss: 0.3006 - val_acc: 0.9370\n","Epoch 64/100\n","100/100 [==============================] - 29s 292ms/step - loss: 0.0075 - acc: 0.9980 - val_loss: 0.3100 - val_acc: 0.9360\n","Epoch 65/100\n","100/100 [==============================] - 29s 291ms/step - loss: 0.0067 - acc: 0.9980 - val_loss: 0.5391 - val_acc: 0.9070\n","Epoch 66/100\n","100/100 [==============================] - 29s 294ms/step - loss: 0.0113 - acc: 0.9955 - val_loss: 0.2905 - val_acc: 0.9370\n","Epoch 67/100\n","100/100 [==============================] - 29s 290ms/step - loss: 0.0010 - acc: 0.9995 - val_loss: 0.3533 - val_acc: 0.9310\n","Epoch 68/100\n","100/100 [==============================] - 29s 293ms/step - loss: 0.0029 - acc: 0.9985 - val_loss: 0.2968 - val_acc: 0.9340\n","Epoch 69/100\n","100/100 [==============================] - 29s 293ms/step - loss: 0.0015 - acc: 0.9995 - val_loss: 0.3393 - val_acc: 0.9310\n","Epoch 70/100\n","100/100 [==============================] - 29s 294ms/step - loss: 0.0040 - acc: 0.9985 - val_loss: 0.5394 - val_acc: 0.9180\n","Epoch 71/100\n","100/100 [==============================] - 29s 290ms/step - loss: 0.0048 - acc: 0.9985 - val_loss: 0.3843 - val_acc: 0.9250\n","Epoch 72/100\n","100/100 [==============================] - 29s 293ms/step - loss: 0.0054 - acc: 0.9980 - val_loss: 0.3140 - val_acc: 0.9380\n","Epoch 73/100\n","100/100 [==============================] - 29s 292ms/step - loss: 0.0047 - acc: 0.9985 - val_loss: 0.3721 - val_acc: 0.9280\n","Epoch 74/100\n","100/100 [==============================] - 29s 292ms/step - loss: 0.0023 - acc: 0.9990 - val_loss: 0.3522 - val_acc: 0.9370\n","Epoch 75/100\n","100/100 [==============================] - 29s 291ms/step - loss: 0.0029 - acc: 0.9985 - val_loss: 0.3390 - val_acc: 0.9340\n","Epoch 76/100\n","100/100 [==============================] - 30s 295ms/step - loss: 0.0035 - acc: 0.9980 - val_loss: 0.3247 - val_acc: 0.9350\n","Epoch 77/100\n","100/100 [==============================] - 29s 292ms/step - loss: 0.0012 - acc: 1.0000 - val_loss: 0.3841 - val_acc: 0.9270\n","Epoch 78/100\n","100/100 [==============================] - 29s 292ms/step - loss: 0.0012 - acc: 0.9995 - val_loss: 0.3691 - val_acc: 0.9330\n","Epoch 79/100\n","100/100 [==============================] - 29s 290ms/step - loss: 0.0048 - acc: 0.9985 - val_loss: 0.3328 - val_acc: 0.9390\n","Epoch 80/100\n","100/100 [==============================] - 30s 294ms/step - loss: 0.0010 - acc: 0.9995 - val_loss: 0.4043 - val_acc: 0.9250\n","Epoch 81/100\n","100/100 [==============================] - 29s 290ms/step - loss: 0.0032 - acc: 0.9990 - val_loss: 0.4367 - val_acc: 0.9190\n","Epoch 82/100\n","100/100 [==============================] - 30s 295ms/step - loss: 0.0036 - acc: 0.9990 - val_loss: 0.3033 - val_acc: 0.9370\n","Epoch 83/100\n","100/100 [==============================] - 29s 294ms/step - loss: 0.0043 - acc: 0.9995 - val_loss: 0.3564 - val_acc: 0.9360\n","Epoch 84/100\n","100/100 [==============================] - 29s 292ms/step - loss: 0.0034 - acc: 0.9990 - val_loss: 0.4568 - val_acc: 0.9220\n","Epoch 85/100\n","100/100 [==============================] - 29s 295ms/step - loss: 0.0056 - acc: 0.9990 - val_loss: 0.3681 - val_acc: 0.9370\n","Epoch 86/100\n","100/100 [==============================] - 30s 296ms/step - loss: 0.0019 - acc: 0.9995 - val_loss: 0.3682 - val_acc: 0.9340\n","Epoch 87/100\n","100/100 [==============================] - 30s 295ms/step - loss: 0.0041 - acc: 0.9985 - val_loss: 0.3426 - val_acc: 0.9320\n","Epoch 88/100\n","100/100 [==============================] - 30s 296ms/step - loss: 0.0014 - acc: 0.9995 - val_loss: 0.3469 - val_acc: 0.9310\n","Epoch 89/100\n","100/100 [==============================] - 29s 294ms/step - loss: 0.0065 - acc: 0.9995 - val_loss: 0.3865 - val_acc: 0.9350\n","Epoch 90/100\n","100/100 [==============================] - 29s 292ms/step - loss: 0.0028 - acc: 0.9990 - val_loss: 0.3523 - val_acc: 0.9380\n","Epoch 91/100\n","100/100 [==============================] - 29s 293ms/step - loss: 0.0014 - acc: 0.9995 - val_loss: 0.3437 - val_acc: 0.9360\n","Epoch 92/100\n","100/100 [==============================] - 29s 291ms/step - loss: 0.0015 - acc: 0.9995 - val_loss: 0.4448 - val_acc: 0.9330\n","Epoch 93/100\n","100/100 [==============================] - 29s 292ms/step - loss: 9.6054e-04 - acc: 1.0000 - val_loss: 0.3619 - val_acc: 0.9340\n","Epoch 94/100\n","100/100 [==============================] - 29s 289ms/step - loss: 0.0017 - acc: 0.9990 - val_loss: 0.4664 - val_acc: 0.9240\n","Epoch 95/100\n","100/100 [==============================] - 29s 289ms/step - loss: 7.3744e-04 - acc: 1.0000 - val_loss: 0.3957 - val_acc: 0.9340\n","Epoch 96/100\n","100/100 [==============================] - 29s 286ms/step - loss: 0.0045 - acc: 0.9985 - val_loss: 0.3660 - val_acc: 0.9370\n","Epoch 97/100\n","100/100 [==============================] - 29s 288ms/step - loss: 7.1088e-04 - acc: 1.0000 - val_loss: 0.3539 - val_acc: 0.9410\n","Epoch 98/100\n","100/100 [==============================] - 29s 287ms/step - loss: 0.0020 - acc: 0.9990 - val_loss: 0.3796 - val_acc: 0.9290\n","Epoch 99/100\n","100/100 [==============================] - 29s 289ms/step - loss: 0.0025 - acc: 0.9995 - val_loss: 0.3872 - val_acc: 0.9350\n","Epoch 100/100\n","100/100 [==============================] - 29s 289ms/step - loss: 0.0065 - acc: 0.9980 - val_loss: 0.3888 - val_acc: 0.9330\n"]}]},{"cell_type":"code","metadata":{"id":"VvOFedYhbhjU"},"source":["model.save('cats_and_dogs_small_4.h5')"],"execution_count":null,"outputs":[]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":545},"id":"2dzbA-9ibjCM","executionInfo":{"status":"ok","timestamp":1634549089751,"user_tz":-540,"elapsed":1341,"user":{"displayName":"상원박","photoUrl":"https://lh3.googleusercontent.com/a/default-user=s64","userId":"16337076995970243718"}},"outputId":"13ea052e-976a-44d5-8986-1bd06319005f"},"source":["\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(len(acc))\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","plt.legend()\n","\n","plt.figure()\n","\n","plt.plot(epochs, loss, 'bo', label='Training loss')\n","plt.plot(epochs, val_loss, 'b', label='Validation loss')\n","plt.title('Training and validation loss')\n","plt.legend()\n","\n","plt.show()"],"execution_count":null,"outputs":[{"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","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAXQAAAEICAYAAABPgw/pAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO2deXgV5fXHvychEEJAZN83BRQkJBBApSgKtG4FF6xSquJSKnWrS60Wq5TWbtLWn622ohaXomDVWtxKq0BxZxPZN4VgABHCIpAASTi/P859uZPJzL0zd1/O53ny3Dtz5868MwPfOff7nve8xMxQFEVR0p+cZDdAURRFiQ0q6IqiKBmCCrqiKEqGoIKuKIqSIaigK4qiZAgq6IqiKBmCCrriCBG9RUTXxHrbZEJEW4hoZBz2y0R0cuD9X4noZ162jeA444noP5G2M8R+hxNReaz3qySeBslugBI7iOigZbEAwBEAtYHlHzDzTK/7Yubz47FtpsPMN8ZiP0TUDcBmAHnMXBPY90wAnu+hkn2ooGcQzFxo3hPRFgA3MPPb9u2IqIERCUVRMge1XLIA85OaiH5CRF8CmEFEJxLR60S0i4j2Bt53snxnARHdEHg/gYjeI6JpgW03E9H5EW7bnYgWEtEBInqbiB4lor+7tNtLG39BRO8H9vcfImpl+fwqIiojogoimhzi+gwhoi+JKNey7hIiWhF4P5iIPiSifUS0g4j+TEQNXfb1NBH90rL848B3thPRdbZtLySiT4joayL6goimWD5eGHjdR0QHiegMc20t3z+TiBYT0f7A65ler00oiOjUwPf3EdFqIhpt+ewCIloT2Oc2IrorsL5V4P7sI6I9RPQuEam+JBi94NlDOwAtAHQFMBFy72cElrsAqALw5xDfHwJgPYBWAH4H4Ckiogi2fR7AIgAtAUwBcFWIY3pp43cBXAugDYCGAIzA9AHwl8D+OwSO1wkOMPPHAA4BONe23+cD72sB3B44nzMAjADwwxDtRqAN5wXaMwpATwB2//4QgKsBNAdwIYBJRHRx4LOzAq/NmbmQmT+07bsFgDcAPBI4tz8AeIOIWtrOod61CdPmPACvAfhP4Hu3AJhJRL0DmzwFse+aAjgNwLzA+jsBlANoDaAtgJ8C0LoiCUYFPXs4BuABZj7CzFXMXMHMLzNzJTMfAPAggLNDfL+MmZ9g5loAzwBoD/mP63lbIuoCYBCA+5n5KDO/B2CO2wE9tnEGM29g5ioALwIoDqwfC+B1Zl7IzEcA/CxwDdx4AcA4ACCipgAuCKwDMy9l5o+YuYaZtwB43KEdTnwn0L5VzHwI8gCznt8CZl7JzMeYeUXgeF72C8gDYCMzPxdo1wsA1gH4tmUbt2sTitMBFAL4TeAezQPwOgLXBkA1gD5E1IyZ9zLzMsv69gC6MnM1M7/LWigq4aigZw+7mPmwWSCiAiJ6PGBJfA35id/cajvY+NK8YebKwNtCn9t2ALDHsg4AvnBrsMc2fml5X2lpUwfrvgOCWuF2LEg0fikRNQJwKYBlzFwWaEevgJ3wZaAdv4JE6+Go0wYAZbbzG0JE8wOW0n4AN3rcr9l3mW1dGYCOlmW3axO2zcxsffhZ93sZ5GFXRkT/I6IzAusfArAJwH+I6HMiusfbaSixRAU9e7BHS3cC6A1gCDM3Q/AnvpuNEgt2AGhBRAWWdZ1DbB9NG3dY9x04Zku3jZl5DUS4zkdduwUQ62YdgJ6Bdvw0kjZAbCMrz0N+oXRm5hMA/NWy33DR7XaIFWWlC4BtHtoVbr+dbf738f0y82JmHgOxY16FRP5g5gPMfCcz9wAwGsAdRDQiyrYoPlFBz16aQjzpfQE/9oF4HzAQ8S4BMIWIGgaiu2+H+Eo0bXwJwEVE9I1AB+ZUhP/3/jyA2yAPjn/Y2vE1gINEdAqASR7b8CKACUTUJ/BAsbe/KeQXy2EiGgx5kBh2QSyiHi77fhNALyL6LhE1IKIrAPSB2CPR8DEkmr+biPKIaDjkHs0K3LPxRHQCM1dDrskxACCii4jo5EBfyX5Iv0Moi0uJAyro2cvDABoD2A3gIwD/TtBxx0M6FisA/BLAbEi+vBMRt5GZVwO4CSLSOwDshXTahcJ42POYebdl/V0QsT0A4IlAm7204a3AOcyD2BHzbJv8EMBUIjoA4H4Eot3AdyshfQbvBzJHTrftuwLARZBfMRUA7gZwka3dvmHmoxABPx9y3R8DcDUzrwtschWALQHr6UbI/QSk0/dtAAcBfAjgMWaeH01bFP+Q9lsoyYSIZgNYx8xx/4WgKJmORuhKQiGiQUR0EhHlBNL6xkC8WEVRokRHiiqJph2AVyAdlOUAJjHzJ8ltkqJkBmq5KIqiZAhquSiKomQISbNcWrVqxd26dUvW4RVFUdKSpUuX7mbm1k6fJU3Qu3XrhiVLliTr8IqiKGkJEdlHCB9HLRdFUZQMQQVdURQlQ1BBVxRFyRBSKg+9uroa5eXlOHz4cPiNlaSSn5+PTp06IS8vL9lNURQlQEoJenl5OZo2bYpu3brBfe4EJdkwMyoqKlBeXo7u3bsnuzmKogRIKcvl8OHDaNmypYp5ikNEaNmypf6SUpQUI6UEHYCKeZqg90lRUo+UE3RFUfyxeTOwdm1k3z18GHj6aUArgGQGKugWKioqUFxcjOLiYrRr1w4dO3Y8vnz06NGQ312yZAluvfXWsMc488wzw27jhQULFuCiiy6Kyb6U9Oauu4Brr43su6+/Lt9dsSK2bVKSQ0p1ivpl5kxg8mRg61agSxfgwQeB8ePDf8+Nli1bYvny5QCAKVOmoLCwEHfdFZwovaamBg0aOF+y0tJSlJaWhj3GBx98EHkDFcWBvXuBffsi++7XX8trpN9XUou0jdBnzgQmTgTKyuTnYlmZLM+cGdvjTJgwATfeeCOGDBmCu+++G4sWLcIZZ5yBkpISnHnmmVi/fj2AuhHzlClTcN1112H48OHo0aMHHnnkkeP7KywsPL798OHDMXbsWJxyyikYP348TOXLN998E6eccgoGDhyIW2+9NWwkvmfPHlx88cUoKirC6aefjhWBcOt///vf8V8YJSUlOHDgAHbs2IGzzjoLxcXFOO200/Duu+/G9oIpCaeqSv4ioTIwXfeBA7Frj5I80jZCnzw5+I/RUFkp66OJ0p0oLy/HBx98gNzcXHz99dd499130aBBA7z99tv46U9/ipdffrned9atW4f58+fjwIED6N27NyZNmlQvZ/uTTz7B6tWr0aFDBwwdOhTvv/8+SktL8YMf/AALFy5E9+7dMW7cuLDte+CBB1BSUoJXX30V8+bNw9VXX43ly5dj2rRpePTRRzF06FAcPHgQ+fn5mD59Or71rW9h8uTJqK2tRaX9IippR2WleOGRcOiQvJpIXUlv0lbQt271tz4aLr/8cuTm5gIA9u/fj2uuuQYbN24EEaG6utrxOxdeeCEaNWqERo0aoU2bNti5cyc6depUZ5vBgwcfX1dcXIwtW7agsLAQPXr0OJ7fPW7cOEyfPj1k+957773jD5Vzzz0XFRUV+PrrrzF06FDccccdGD9+PC699FJ06tQJgwYNwnXXXYfq6mpcfPHFKC4ujuraKMknmghdBT2zSFvLpUsXf+ujoUmTJsff/+xnP8M555yDVatW4bXXXnPNxW7UqNHx97m5uaipqYlom2i455578OSTT6KqqgpDhw7FunXrcNZZZ2HhwoXo2LEjJkyYgGeffTamx1QST1WVRuiKkLaC/uCDQEFB3XUFBbI+nuzfvx8dO3YEADz99NMx33/v3r3x+eefY8uWLQCA2bPDTzA/bNgwzAx0HixYsACtWrVCs2bN8Nlnn6Ffv374yU9+gkGDBmHdunUoKytD27Zt8f3vfx833HADli1bFvNzUBJLZSVQXQ3U1kb2XUAFPVNIW0EfPx6YPh3o2hUgktfp02Pvn9u5++67ce+996KkpCTmETUANG7cGI899hjOO+88DBw4EE2bNsUJJ5wQ8jtTpkzB0qVLUVRUhHvuuQfPPPMMAODhhx/GaaedhqKiIuTl5eH888/HggUL0L9/f5SUlGD27Nm47bbbYn4OSmIxdksktouJ0LVTNDNI2pyipaWlbJ/gYu3atTj11FOT0p5U4uDBgygsLAQz46abbkLPnj1x++23J7tZ9dD7lXyYgZxAWLZrF9Cqlb/vX3YZ8MorwIQJwIwZMW+eEgeIaCkzO+ZIp22Ensk88cQTKC4uRt++fbF//3784Ac/SHaTlBTF6p1HEqGr5ZJZpG2WSyZz++23p2RErqQeVhGPxnJRQc8MNEJXlDTGOowgkkwXFfTMwpOgE9F5RLSeiDYR0T0On08gol1EtDzwd0Psm6ooip1oI3QdKZpZhLVciCgXwKMARgEoB7CYiOYw8xrbprOZ+eY4tFFRFBesIq4RuuIlQh8MYBMzf87MRwHMAjAmvs1SFMULVstFPXTFi6B3BPCFZbk8sM7OZUS0goheIqLOTjsioolEtISIluzatSuC5saXc845B3Pnzq2z7uGHH8akSZNcvzN8+HCY9MsLLrgA+xzK1k2ZMgXTpk0LeexXX30Va9YEf/Tcf//9ePvtt/0035F0LrP7+uvA2WcDx44luyWpS6w6RQ8c0OucCcSqU/Q1AN2YuQjAfwE847QRM09n5lJmLm3dunWMDh07xo0bh1mzZtVZN2vWLE8FsgCpkti8efOIjm0X9KlTp2LkyJER7StTWLQIWLiwfhE2JUg0lkttLXDkCNCsmSwfPBi7dinJwYugbwNgjbg7BdYdh5krmPlIYPFJAANj07zEMnbsWLzxxhvHJ7PYsmULtm/fjmHDhmHSpEkoLS1F37598cADDzh+v1u3bti9ezcA4MEHH0SvXr3wjW9843iJXUByzAcNGoT+/fvjsssuQ2VlJT744APMmTMHP/7xj1FcXIzPPvsMEyZMwEsvvQQAeOedd1BSUoJ+/frhuuuuw5EjR44f74EHHsCAAQPQr18/rFu3LuT5pVuZXSNQKujuRGO5mO+2ayevarukP17y0BcD6ElE3SFCfiWA71o3IKL2zLwjsDgaQIQTYgX50Y+AwFwTMaO4GHj4YffPW7RogcGDB+Ott97CmDFjMGvWLHznO98BEeHBBx9EixYtUFtbixEjRmDFihUoKipy3M/SpUsxa9YsLF++HDU1NRgwYAAGDpRn3KWXXorvf//7AID77rsPTz31FG655RaMHj0aF110EcaOHVtnX4cPH8aECRPwzjvvoFevXrj66qvxl7/8BT/60Y8AAK1atcKyZcvw2GOPYdq0aXjyySddzy/dyuwaQY+0kmA2EI3lYuyWdu2ADRs00yUTCBuhM3MNgJsBzIUI9YvMvJqIphLR6MBmtxLRaiL6FMCtACbEq8Hxxmq7WO2WF198EQMGDEBJSQlWr15dxx6x8+677+KSSy5BQUEBmjVrhtGjRx//bNWqVRg2bBj69euHmTNnYvXq1SHbs379enTv3h29evUCAFxzzTVYuHDh8c8vvfRSAMDAgQOPF/Ry47333sNVV10FwLnM7iOPPIJ9+/ahQYMGGDRoEGbMmIEpU6Zg5cqVaNq0ach9xwMjUBqhuxON5WKua/v28qoRevrjaaQoM78J4E3buvst7+8FcG8sGxYqko4nY8aMwe23345ly5ahsrISAwcOxObNmzFt2jQsXrwYJ554IiZMmOBaNjccEyZMwKuvvor+/fvj6aefxoIFC6JqrynBG0353XvuuQcXXngh3nzzTQwdOhRz5849Xmb3jTfewIQJE3DHHXfg6quvjqqtflHLJTzRWC7WCB1QQc8EdKSojcLCQpxzzjm47rrrjkfnX3/9NZo0aYITTjgBO3fuxFtvvRVyH2eddRZeffVVVFVV4cCBA3jttdeOf3bgwAG0b98e1dXVx0veAkDTpk1xwOE3b+/evbFlyxZs2rQJAPDcc8/h7LPPjujc0q3Mrlou4YkmQjeCrhF65qC1XBwYN24cLrnkkuPWiyk3e8opp6Bz584YOnRoyO8PGDAAV1xxBfr37482bdpg0KBBxz/7xS9+gSFDhqB169YYMmTIcRG/8sor8f3vfx+PPPLI8c5QAMjPz8eMGTNw+eWXo6amBoMGDcKNN94Y0XmZuU6LiopQUFBQp8zu/PnzkZOTg759++L888/HrFmz8NBDDyEvLw+FhYVJmQhDI/TwVFZK+eiCgsgjdBX0DIKZk/I3cOBAtrNmzZp665TUJd73a8QIZoD5n//0/p3332eurY1fm1KNO+9kLihgbtWKedIkf9+dM0eu79y58vp//xefNiqxBcASdtFVtVyUlMVvhL5uHTB0KGAbG5bRVFUBjRsD+fmRWy7qoWcOarkoKYtfD72iQl5TcBBy3KisFLslPz9yy6V5c/m+Cnr6k3IROidpBiXFH4m4T37TFs12RqiyAROhN24cedpikyYyWlQF3Z333wdcxhOmFCkl6Pn5+aioqFBRT3GYGRUVFcjPz4/rcfxaLuYBkI2CHk2EXlCggh6O558Hpk6NrKJlIkkpy6VTp04oLy9HKhbuUuqSn5+PTp06xfUYfgXdbJdNNUmM5dKwYWSCTiQPAxX00AQqeqC8HDj55OS2JRQpJeh5eXno3r17spuhpAh+PfRstlzy8oC9e/1999AhsVuIRNDTYeh/RQXQsmXij2tizLKy1Bb0lLJcFMVKpB56NkXoVg89kuJcBQXyvmnT1I/QP/4YaNNGspkSjYnQt25N/LH9oIKupCTMkVsu2RShG1GORNBNhA6kh+Wyfr3UbLcUL00YJkJPdUFPKctFUQzV1SLqgHehyuZO0dzcyPLQ00nQv/pKXr/8MrHHZQ5G6GVliT22X1TQlZTEKuJqubhjIvScnMgsl3QSdBMlJ1rQ9+8HTN27VI/Q1XJRUhJrtKmWizvR5KEfOhT00Js1A44elRmMUhUj6Dt2hN4uGp5+Gvjtb52Pm5Ojgq4oEaGCHh7m+nnofoZw2C0XILUzXRJhuTz7LPDXv9ZdZ+yWPn1E0FN57lUVdCUliWQmnmyzXKqrZV5Q0yl67Jis84qToKey7ZIIy2XXLmD79roPRnPcgQPlF0wqD5NRQVdSEhOhN2yoI0XdMOdrInTAn+1iT1sEkifo//0vcO21obdJlKAfPQrs2RNcZyL0wCySKW27qKArKYkRphYt1HJxwyrojRvXXeeFVIrQH3tM/OtQDySr5RKP6iDHjgXF2+rTWyN0ILUzXVTQlZQkGkE/eDA+/+FTDXO+xnIBohf0ZHjotbWAmYlx3z7nbaqqpL2tWontsX9/7Nuxd6+0BRDbxbB7t1zfU0+VZY3QFcUnRphatPDvoR87ltrZGrEiGsulpkasBWuWC5CcCH358qCQuwm6iZKLiuQ1HpkuVm/cKui7dsmDpHlzoLBQBV1RfGOP0L1E3NZIPhs6Rs35RmK5WEvnAskV9Hnzgu/dBN3YLUbQ4+Gjuwn67t1A69ZS86ZLFxV0RfGNVdCty6GoqpL/dEB2+OhGvK2Wi9cI3VyfVBH03Fx5Hy5C79dPXhMp6CZCB4CuXdVDVxTfWC0X63IoKiuD22eToFstF68Rul3QzWjTRAv60aPAu+8Cw4fLslfLJR6Cbn4FNG/uHKEDGqErSkTYI3QvHaOVlcH/eGq5ePuu8dCJklNxcdEiebhceqksh7NcevaUVNZ4RuhFRe4RepcuIvBeO+oTjQq6kpL4FXTmuoKeTRG6mVMUiNxyAZJTE33ePHmYjBkjy6Ei9IYNpY3t2sWvU7RZM6Bbt+D+jxyRa2KN0IHUjdJV0JWUxK+gm+2zUdAjidDdBD3REfq8eUBJCdChgwh2KEFv00bEv127+EXorVtLW3bsqJuXbvXQARV0RfFFVZV4uiecEFwOtz0g/+kBtVy8ftdYLkDiBb2yEvjwQ+Dcc0WomzcPbbmYh3X79vET9DZtRNCrq2V2JCPoGRWhE9F5RLSeiDYR0T0htruMiJiISmPXRCUbOXxYbAQjOOEidPN5NkbosbRcEinoH3wgnaIjRshyKEE3YgvEL0I3D40OHWR5+/agr24i9A4dJNBI1UyXsIJORLkAHgVwPoA+AMYRUR+H7ZoCuA3Ax7FupJJ9RCvo2RChx9pySXSn6IIFQIMGwDe+IcteI/R27SRy9lOIzAtWywUQQbdH6Hl5QMeO6R2hDwawiZk/Z+ajAGYBGOOw3S8A/BaAz6rMilKfqiqN0MNRWQk0aiQRo98I3T6wCEh8p+iGDTLhcmGhLDdv7j7RtRFbQASdObZVD82sROEidCC1Uxe9CHpHAF9YlssD645DRAMAdGbmN0LtiIgmEtESIlqyK5VrUCpJ5/BhiTqNoIeLPI1ANW8ug1SyQdBNLXRAzjkvz3+EnkwPfetWoHPn4LJbhF5ZKe21Wi5AbDNd9u2Tcght2tTd/+7d4u+bznlABD1tLZdwEFEOgD8AuDPctsw8nZlLmbm0tXncKooDxnIxghUuQjdC1qSJ/GWD5VJZGbw+gL+Jog8dksi+UaPgOhOhJ2oCh61bg52MAHDiic6CbmI/a6coEFsf3XqMRo2Ali2DEXqLFsGRrIBkunzxRbCQVyrhRdC3AbA8R9EpsM7QFMBpABYQ0RYApwOYox2jSjRE6qEXFMhP+EyJ0D/6qO4gFytVVXUj7Px8f52iTZoESyUAIujMibl2R45IBGwVdBOh2+v22AXdRNCxFHQzcMkco0OHoIdujz3795dofsmS2B0/VngR9MUAehJRdyJqCOBKAHPMh8y8n5lbMXM3Zu4G4CMAo5k5BU9XSResc2Wa5VBYBb1Jk8wQ9NpaYNQo4MEHnT+3Wi6AvwjdOrmFIZH1XMrL5dUu6EeP1n8oGUE3lkvbtvIarwgdCAq6dZSoYdQoeRD++9/+j3PwoGT3uD2koyWsoDNzDYCbAcwFsBbAi8y8moimEtHo+DRLyXZMhN6ggbdZi6w52cm2XJhjIzaffy7nYcTPjpPl4jdCt5JIQTedinZBB+rbLvboOT9ftvV7jRcvdvfd3QTdKUJv2RIYPNi7oH/5JXDVVVK2oFkzYOhQ4J//9Nd2r3jy0Jn5TWbuxcwnMfODgXX3M/Mch22Ha3SuRIsRdECEKp0slzlzRKiiFfWVK+V1507nz50sFz8eul3QzTR0sch0WbAA+DhEArMfQbeLLeB/+H9tLTByJPDjHzt/7iToX34p194eoQPAeedJHZqKitDH3bRJBPyVV8SqmTpV/n1cfrn3tvuhQXx2qyjRYdIWAREtr52ixnJxS39LBCtXSo50WVnQ7410P4D7g6GqSqJFg1/LxS7oZl8mIo6GH/xAZhXauDH4oLBiBL1Tp+C6UILeqFHd/fgdLbp2rfzymD9ffkFZ+w7MMZo2Df6ba99eHgLWdEkr550H/PznMhfqlVc6H3PZMuD882U/8+YBQ4Z4b2+k6NB/JSUxaYuAiLRXD71xY4nQk2m5GLGKVhitEbrTBB/RWi52Dz1Ww9prasQu2rkT+N3vnLfZulW8cCOgQGjLxUwwYfA7WnTRInndvl3aZsc6cAkI5qIDzoI+aJBk5bjZLps3S0ng/HzgvfcSI+aACrqSokRiuTRqJOllye4UNYIY7VCLVavk9fBhZ1871pZL27bSXxGtoG/dKqJ+4onA73/v3AdgT1kEQkfodlH1K+gffxxMPfzf/+p/bj+GVdCdLJfcXOCb3xRBd0rz/PvfJaiYPx845RTv7YwWFXQlJbEKuhfLxRqtJrtTNBaCXlUldkWPHrLs5KNHk+XiJOg5OTLQx2nQjJ85WjdtkteHHxaxmzy5/jZ+Bd1kuBjatZNz8HqfFy0CzjlHRNtN0K3HCBehA2Kn7NwJrFhR/7OXXwbOPDN4/xKFCrqSkljFyqugm2g1mZ2izLER9LVrRQxHjpRlJ0G3Wy5+8tCd0hYB52Hte/aIv/7ss972bQR95Ejgttvke8uWBT8318hN0O39H3Y7BPCXi15ZKfbVkCHAWWd5i9CtfR9OETogETpQ33bZtAn49FNg7NjwbYs1KuhKysFcP0L3Uj7XCFSTJvL9ZIzk27s3+DCJRtCNfz5qlLw6CZfdcok2Qgec58xcvVq2f+ghb5N1b9okbWnfHvjpT0UQH3gg+PmePSKydkHPzxfbzEuEbkaLTpwosx1dcYWzNw4An3wi/xYGDwbOPlvOz3qOpi6MVdDz8oLHdIvQ27cHiovrC/rLL8urmYUpkaigKynH0aPy6tdDtwo6kJwo3RrdRivojRpJyhtQP0KvrZXrFMs8dEBEdvv2upUMTcS9ahWwcGH4fW/aJEW3iKSe/RVXiJdcUyOfO6UsGuzD/00dF7uolpRIxP3VV2JNvfwy8Je/OLfHdIgOGiQROlD3PPbvl/O1H8PYLm4ROiDZLu+/XzeF8qWX5OHhdH7xRgVdSTlMlOnXQ7daLkByBb1Vq+gFvU8fiRJzc+tH6NY0TYPXTtHqavlzsly6dpWI1dqRuXGjDPA68UTgz38Ov/+NG0XQDUOHyr0wXnMoQbcX6HLKQQfk+n70kTxkVq4EvvUtEVKnXxCLFknfQPv2QL9+ch5W28U+EtXQvr1cI6frZLj2Wrk2118vx96yRUoCXHaZ+3fiiQq6knKYKNOvh27tFAWSK+gDB0Yv6P36iZi3bl0/QrfWQjeYCD2cLeJUOtfglLpoOmevv15GOG7bVv97htpasT569gyuM78y3n+/7r79CLpdbO2MHStiavXqDYsWScQMSMfvsGHOgu5Us+XUU0Mft1cvSc186y1g+nQZQASooCvKcYyg+/HQnSyXZGS6bN0qVkmfPpHnoVdUyE/4fv1k2SlFz5p3bzDvw2WkOE1uYTBzZlo95o0bRaAnTZKO2scfd993eblYQdYIvUsXGUBkFfRGjZy9aa8Rup0xYyRSfumluut375YHjBF0QGyXTZuC9VTcjvGLXwTbHIqbbpK+jjvuEFEvLgZOOin89+KBCrqSctgtF+Ohh4o8U8VyKSsTAWvTRs4jkjaYDlEj6G3bukfodsvF+tmRI86/EkIJuqlPbqJoZhG/nj0lSr/wQhF0t4eG8dutgg5IlG4V9CZejTsAACAASURBVC5d6o/WBOoLuhFdU5DLjRYtJC3RbrssXiyv1oE9Z58tr8ZHt9eKMTRoULe8sBs5OcCMGbLt+vXJi84BFXQlBXGyXGprQ085Zs9yAZJnuXTpEhSHSGwXu6A7Rehulov1s1//Gigqqv8gdJog2pCfL+JpIvQdO+Q6Ggvl5ptFAGfNcm57KEEvL5fr45SyaLAL+rp1IpReOhjHjpXjm+sHiN2SkyMWmKG4WIb5/+MfdWc+imaKho4dJTpv1sy9FEAiUEFXUg4nywUIbbs4RejJslxMhA5ELugtWgRT80yEbhVmJ8vFPg3d8uXyILAXsQoVoQN1c9E3bpRXI9CjRok4/uQnzvVyNm0SAe7Yse56q4/uRdDNua5bJz61dYIJNy6+WMTbart8/LHYX+bfBCCR9+23i999551yjwoL617LSBg7Vtpuf5glEhV0JeVwE/RQHaOp0Cl69Ghw0oZoI/R+/YKWRNu2su/9+4PbOFku9gjd5GVv2FB3/+EE3ZqLbgTdROg5OcATT4g3fffd9b+7aZP4xzk2ZSkqkuMtWFB/YgsrzZvLLzFzDmvXhu+YNLRpI3aKsV3+9S+po2L1zw1TpgC33gr88Y9il8RqAjUnGymRqKArKYeThw64CzpzanSKbtsmbYlG0CsrJRXP2C2A86jIUJ2iJtPFTdBDWS5AMEJnFkFv2LCuAJeUSAfgk0/WH3VpctDtNGgAnH665Iuba+SEdfj/4cNS5MpPLZTLLpOHQFGRROxt24pw2yGS0gSmKmSmzIipgq6kHE4eOuAu6EeOiEgku1PUmo4XqaD/6U/yIPrOd4LrTIegtWPUyUO3doru2hU8/0gi9KoqicJNyqLd8pgyRdZPnBi8X8eOuQs6ILaLqR8eTtD37pV9HTvmPUIHZHRmfr5cw6eeEnHv3995WyLgscekRvp3v+v9GKmMCroSkupqYNo07yMQY4FfD91uP5jXZAp606YS2fpJXdyzRzoyL7pIcqUNThF6OMvFOgzeLuimcqPVV7ZizUU3KYt2CgqAv/5V9v2rX8m67dvl3jltDwR9dOsx7Jx4orzu2ydiDPiL0Nu3Bz77TLJNrrtOfhmEIidH8shvu837MVIZFXQlJAsXSgQzd27ijuk0UhRwj9DtFkJubjBKSyRG0Dt3luivdWt/Efqvfy1i++tf113vFKGHs1yMoPfpU1/QV62SB47pdLVjctG3bAmmLDoxahQwbpwI4ubN7hkuhtNPD3rrnTs7b2O1XNatk+vYq5fztm506CAP02xEBV0JiRkCHq9JbZ2wWy7hPHQncUtExcV//1sGlZh62Fu3ioibdvgR9K1bxW655hrgtNPqftaiRf3h/+Esl88+k/ff/Ka8N3VUABlNWVJSv+PSYKLnDz4IHXEDIua5ucBdd4UX9GbNpG/Aeo3sWAV97Vp5uIQaeq/URaegU0JihnknQ9AjjdCB+E9ysWOH+K5790qBpm9/W0TZRLeAP0E31Qh//vP6n+Xk1B9c5MVy6dBB/OOaGom2Tz5Z3i9fLp2BbrRoIdfvnXdkOZSgd+okFRXvu0/+jeTluUffgFgbmze7f26P0P3454pG6EoYUknQ3Tx0N0GPl+XCLIJYVSX+9rRpst6eX92mjTdBX7ZMaobffLO7t2wX9MpKiYzz8oLrrHnon38unZbGrjC2y/r10m7rQBs7RNKOTz+V5XB51XfeCXTvLsWynDpQrVx7rUyU7MYJJ8jrnj0i6Imc7ScTUEFXQmKEPJGCXlUlUanp0AoXoTtFq/G0XJ5/HnjtNeDBByUXe+FCGZFoF3QvEXptLXDjjbLtffe5b2cfLWqvhQ7Uj9CdBH3pUnkdMCB0u8x5NGoUOuIG5EHyhz/I+2gH1TRqJOexcqWchwq6P9RyUUKSrAi9cePgIA2vHnoiLJcdO4BbbgHOOEPsg8pKsUkmT5ZfBHZBP3iw/lRxVp54QuqNzJwZtBucaNu27pB2p32a5X375L716CEzDZ14Yl1Bb9IE6N079Hka68hpkJATY8bI6MvTTw+/bTiaNwc+/FDeq+XiD43QlZAkS9Cts8FHYrkUFrpbLhUVoevChGLKFDnejBliLTRtKhH222/L53ZBB9yj9K++Au69V4pKjRsX+rjt2tUd/m+ffg4IFpJat062O+mkYJaIEfRly6SWSbih9OY8QvnnVogkSrfmz0dK8+bBzniN0P2hgq64UlsrP/Pz8mSQiZ+JgqOhqqquoOfliQD5yXJxi9CPHQP69gV+9jP/7TpyBHjxReDyy+tGuLfcEvSy/Qj63XdLGx99NPyQ8bZt5SFk6qc4WS5Ect1Wr5ZlM0GxEfTaWpmOLZzdAgQjdK+CHkvML5WWLTNnBGeiUEFXXNm5UwTQjLSzV/ybNy/8xBORYCwXA1Hoaej8dIpu2SLn9cwz/uccnTtX7Ax7NN2xY3CkoT3LBXAW9E8/lTbceac3W8Hkopt74BShA7Ju/Xp5bxX0L76QYx46FLpD1OA3Qo8lRtA1OvePCrriirFbBg2quwyIBTNihBQ3ijV2ywUIPWuRn07RVavk9csvnWd/D8ULL0jUaCZutjJtGjB7dt2ZdUIJ+iOPSHudClw5YUaLmkwXN18+P18i+caNgw8B0zE6e7a8eonQBw+WTJ7Ro721L5YYQVf/3D8q6Iorxjc3gm710detk9f//jf2x3UT9HAeupPlYq8FbgS9SRMRaK8cOgTMmSMlUq2pgoZWrer7x24ldCsqJFPmqquCQ93DYR0tumaNeOFG5K2Ya9CjR9DGMYI+a5ZcVy9CmZ8vQ/udjhFvzDXRCN0/ngSdiM4jovVEtImI7nH4/EYiWklEy4noPSLqE/umKonGROSlpfJqFXRTVvXDD2Nvu9g9dCB0hF5ZKUO9rXU7mjQRS8Xu+69aJbbIJZdI5b+jR7216bXX5DjhOi+tnHCCiL9d0J98Uh5aN9/sfV9GWJctk4FMjRsH89+tmOtm7BYgmEq4davYZ+HqmyQbtVwiJ6ygE1EugEcBnA+gD4BxDoL9PDP3Y+ZiAL8D8IeYt1RJONu2SWfkqaeKMFkF3WRNHD0qNadjid1DB8J76Pbt3Sourl4tQ+vHjZMORq81al54QUZeWotmhYNIIneroNfUSCfoOefUH+IfihNPlHvw0EPS7rfeArp1q7+dNUI3FBYGJ5zw4p8nGxX0yPESoQ8GsImZP2fmowBmARhj3YCZv7YsNgEQZt5xJdX44gupx2FN59u+XQo4NWggYmaP0Hv0EJExQ8RjhV8P3VoL3eA0yUV1tVhFffuKD96ypVgf4TACesUV3nKyrdgHF82ZI9faqUZ3KIjEdmnQQGbaKS523s5J0IGg7ZIOgn7llcDvf1//HJTwePnx1RHAF5blcgBD7BsR0U0A7gDQEMC5TjsiookAJgJAFy+TBCoJ45VX5D/RpZcCZ54p67ZtEyEH6gv6hg0iKh07xl7Q3SwX61yT9u3dBN2a6bJpk/yiOO00eRCNHQs895yIvr02+KJFkuKXny+1T6qr/dkthtat65bQ/dOfxPL59rf97+sXv5CI36lT1mCum33W+V69gPnzvXWIJpvOnWUCDcU/MesUZeZHmfkkAD8B4DiImZmnM3MpM5e2TvME03nzxAvNFMyUY2ZoOCCCbn6qWwW9pkaGlvfsCYwcKb7unj2xa4uT5eI3QneyXEyHqLE6xo2T7z71VLAa4c6dUvFwyBAZMDRhgsxsc+qpwb4EP1gj9IULZQq2H/7Q2xyZdiZMkFrpoXCL0EeOlFl8+vb1f1wlffAi6NsAWKs5dAqsc2MWgIujaVQ68Ic/yES5mYIR9GXLguvcBL2sTCLWXr0kdZFZhCpWOFku4Tx0L5bL6tVimRhvdtgweX/bbWK/XHihLL/wgozgLCuT0rNr1sjkxpHMF2kE/auv5AFy0knyoIgXRtDt/vrYsZKH7pSho2QOXiyXxQB6ElF3iJBfCaDOhE1E1JOZA3kPuBDARmQ4mzdLVLp3r/fUs1TGHqEfOiRzLVoFff9+WW+dOHjwYImG33lH7JpYEEnaolunqNVyWbVKMj7Mtjk5kqXzn/9I+xcskFokf/xj7Drk2rSRSSuuuEL+vXz4odQFjxctW0rlw2hnsFfSk7CCzsw1RHQzgLkAcgH8jZlXE9FUAEuYeQ6Am4loJIBqAHsBXBPPRicbZhlxCEgEF8lP8VTDCPqaNSKcJhq3euiAFKcyGS69eknEd9ZZsfXRI0lbNHnaBqcIfdWq+pZD8+aSPx6LGiROGGdxwQKxdtw6M2PF1KnqP2cznjJSmflNAG/a1t1veZ8hM/J5Y9euoLhs2pT+gn7okNRqKS0FliwBVqwIRsPWCB0Qod+4UYpSmYEzI0YAb75Z16Lxgxn8QyTvjxzx56E7dYqautrmwXv4sLQ7XsLthnnQXHutzHEZb1q1kj8lO9GRohFgnXHFTLuVzpi5MI1lsmxZMEI3Am1et2+XCL1nz6CnPGKEvEYapf/wh8EJhM1AICcP/ehR8e7/8Q/xoU307eShd+ggvxweflg+X7dO6tL4yf2OBaNGyTD/P/85scdVshMV9Agwgk4UnLsxnTF2y7Bh4sEuXRocJeoWoVsn7jXzRJoSsn5Yvx6YPl1muzG1wwFnywWQtLvvfAd4/HHgX/+SdU6CTiQTUHz5pQzkMRUIEy3oTZpINUadF1NJBCroEWAEvaQksyL0rl1FMJctE0EvLBRrBZCOvIICSVcsK6tbhS8nR6L0t9+uXzslHFOnSuTMLFkY9gmiDS1ayGtVleSOt28vufOAe+XBb3xDhsn/5jcy4XFeXnKqBypKolBBj4DNmyUi7d8/MwS9rCw4GnTgQOk83Ly5rh9OJJ+/954IsDVCB8Ra2LEjGAl7Yc0aSRH83vdkedmy+vOJGsaNk0Jga9fK9hdfLKM3KyudI3TDL38p2SWPPy41zDVtT8lkVNAjYPNmSQ07+WT5SR/P2eUTQVmZzN6emysRenW1jCq0d3B26CAdpkD9SHfkSHn1Y7v8/OdiSfzxj/KA/OQTd8ulcWM5hhHkSy8VIX/9dXnAuAn6wIGybW1t4u0WRUk0KugRYATdDK+OlY++b59MLDB/fmz255WysuDEDKbWx4EDQd/c0KFD0FKxC3qXLhK1ey2nu3KldG7eeqtkZZSUiKC7Reh2zj5b8v///ndZDuVRT50qtlA6DHtXlGhQQfdJba14zt26BcuSxkrQV6yQwk1eKwCG4vnngeuvF186HFZB7949WO3OKUIHpOPUeNpWRo2SSSPClaQ9eFAyWwoLZcYeQAR99WoZvASEHxiTlyeTL7z1liyHEvS+fWXffsrVKko6ooLuk+3bxZKwRuix8tHNgB3r7O6R8tBDwN/+JgNZLrwQ+Phj5+2qq6UD1Ag6UTCSdRN0u39uGDVK7CczY7sT+/cD3/qWdFI+/njwwVBSIm0xI1XDReiAWCmmBku4B8App+joSSXzUUEP8NVXwZ/7oTAZLiaSbdky9oJufOpIOXhQ9nHLLdIpuGiRpCS+8Ub9bbdtEw/aOhdmOEF3yxQZPlx8eDcfvaJCsmEWLZLp0KzVC80xzcPAi6B/85vBEaGaFqgoKugAxBcuLpbypOGwCjogtkusLBcj6OXlwdndI2HRIhHpCy4AJk+WvPGiIuCyy6RuiRWTg26tZjwkUBzZXuApXIR+wglS28XNR7/iCsmgefVVKRZl5aSTJEXygw9k2Us0nZ8vvz4AFXRFAVTQAYio7dhRt9KgG5s3iy1hBPCkk2IboZvCTdHYLibKNcLcvLkIee/eku5nrYxoBN0aoV9yiYhySUnd/fbuLUJr6qU7MWoUsHhx/QfSp5/KSNJf/jIowlZyciQN1Mxq7yVCB+QhBWRGgTRFiRYVdATFc6OHGpEmP7tRI1k++WTpyLTPXemXmhp5MJhZ1qMV9FNPrStyLVqIFdK9u4i6qULoFKHn5kqKoL1cbLt24oGfc477sUeNkl8H9kydRx+Vh8H117t/1/oA8SroY8fKw2pIvSlXFCX7UEFHUDw3bw6foWFSFg0nnywCZopARYqpMT58uIhvpD46swi6UxTdurUMs9+/H3jxxeBx27b1LqDhBuYMGSK/MqZPD6Y47t0r6YXf+17oSDoSQc/JkYdIJLXKFSXTUEFHUNCPHZOh7aHYvLmut2zPdPn8c+DZZ/23wfjnvXtLbZRIBX3DBhkZecYZzp+feaYc46mnZNmashgL8vKkL2Lu3OCMTjNmyIChm24K/V2roGtGiqL4RwUdIp7t2sn7ULbL0aOSFWKP0AHpGN2/X2qHXHONdGz6wVpjvKhIOg+PHQv/vW3b6vrVplPRTdCJxPb44AMZRr91a2wFHZB87xEjgNtvl+v52GNSV6V//9Df69MHaNhQ3nuN0BVFCZL1gn7kiFT8u+QSWTbC6sTWrWIjWAW9dWvJztiwQYTcPBD8jvbcsEGyRFq3lgj94MHwNs6WLbLtOecE87E//FA6QUPNuHP11VK75ckn4yPoOTkSlTdoIG377DNvg3oaNpTh+bm58l1FUfyR9YK+bp2M/jzrLMkpDyXo9pRFQCLek04SAfvXv4Df/1488EgEvVcv2V9RkawL1TF65Ahw+eUykOfTT4EnnpD1H34o06jlhLizbdvKrPPTp0vufawFHZCZ2//0J/kF0b598IEZjtLS+E7RpiiZTNYLuhHNfv1EUENZLk6CDojtUlkpeda33y5R6Tvv+CslawQdCE6TZvXRy8uDw+IBmWZsyRIZoDN8OHDffdK+1avd7RYr118fzHSJh6AD0gn6859LhouxUsIxZQowZ0582qMomU7W/7BdsULEplcv+QtVLXDzZrEC7CMoL7pIfOwnn5QI+9xzgZdflu179JBtystlOP5dd0n0aqWqSqwPI+iFhRL1m4fN+vUykvLoUSlKdeqp4kv/+MeSgnjSSTIw6pJL5CESKk/c8K1vyUCh7dvjJ+hEwP33h9/OSvv28qcoin80Ql8pAmkmP9i2zbkcbm2t1Bnp0kU8XivXXCMPAjPT/Lnnyuu8ecFtfvlLmYps4MD6dozJkLGOwCwqkodNTY143vn5Ev1v3y7TmQ0bJjPyAPLrYtIksV6IZLRmOBo0AG64QfLp7SNCFUVJT1TQV4ogAkFBtY/8/OoriWj/+9+69Ufc6N1bokwj6Hv2SCrjhReKTz9yJDBtWtCSsWa4GPr1E/vngQdkKP9f/wr87ncyKcTmzVJl0JoTPnWqePf9+nn3oO+7Tx4a6lkrSmaQ1ZbL3r0SkdsFfcOGYIrdRx/J8PI9eyR328vM7cZ2MVOyPfGE2Cq//rVEw9ddJ3ZJq1bAhAlBQbcWvSoqkrTFX/1KHiKXXx78zCmibtFCcr9DdYbayctzr8uiKEr6kdYRelmZZGp4ydd2wnjUJqvE5JQbga2tlY7ORo0ke8SLmBvOPRfYuVOO8eij0lHar5+kOM6eLbPc33UXsHu3eOTt2wfn7wSCD5kOHbzPGF9aqpM4KEo2k7YR+urVUj51+3bpFBwxwv8+TBaJEc8mTaTD02S6zJsnnZWzZ0unox+Mj37bbVLr5U9/Cn6WkyO1wIuLJVLfsEFsGisnnyyZKNdc4zyZhKIoip20jNAXLZK8cWYR4VmzItvPypVSW8Q61VrPnsEI/W9/EzEdM8b/vrt1k/TGBQvk9aKL6n7ety9w993A009LZ6vd+sjJkayZYcP8H1tRlOwkrQR95kwZFDNkCPD118G0vVdeCV9UywnTIWot7NSrV7Aeyj//CYwfH6ys6BcTpd9yS/3MGEA6JXv0kLarl60oSrSkjaDPnAlMnCgZJ4Ck8913nwj8nj3+ZpsH5DvWDBdDr14ys86f/yyjMf345na+9z0Ztem2j8aNJXslJyc4ObOiKEqkEPsZzhhDSktLecmSJZ6379YtWLvbSpcuEq2PHg0884y3fR04IKmDy5eLJWIdWTlnjlgsTZuKj+1l0oto2bcvODGzoihKKIhoKTOXOn3mKUInovOIaD0RbSKiexw+v4OI1hDRCiJ6h4hiPvZw61bn9V98ISMkX33V25yglZXiZy9dCvzjH/WHyRvr48CB6KJzP6iYK4oSC8IKOhHlAngUwPkA+gAYR0R9bJt9AqCUmYsAvATgd7FuqHVGHfv6K6+UKP3f/w6ut//w2LkTeOklGdzz7rvAc88FZwey0qOHWCANGwLf/W7s2q8oihJvvETogwFsYubPmfkogFkA6uR9MPN8Zq4MLH4EoFNsmynD3O0TARcUyPpzz5VBOrNnS4fmDTfIZ02aSJ3zzp3l9fLLJUPmqafcR3w2bCiDisaN03RBRVHSCy956B0BfGFZLgcQagbH6wG85fQBEU0EMBEAuriF3C6MHy+vkyeL/dKli4i5WT92rKT5zZ4tWSnjx4uVceCAjNLs109SAAcMCF/57913w0+1piiKkmrEdGAREX0PQCmAs50+Z+bpAKYD0inqd//jxwcF3M7EiVKyduxYGczTtq3fvQdp0iTy7yqKoiQLL4K+DYC14GunwLo6ENFIAJMBnM3MR2LTPO+UlISenEJRFCXT8eKhLwbQk4i6E1FDAFcCqDMFARGVAHgcwGhm/ir2zVQURVHCEVbQmbkGwM0A5gJYC+BFZl5NRFOJyOSJPASgEMA/iGg5EemcM4qiKAnGk4fOzG8CeNO27n7L+5ExbpeiKIrik7QZ+q8oiqKERgVdURQlQ0hbQZ85U+q75OTI68yZyW6RoihKcknLCS5M5cXKwNjUsjJZBtzz1BVFUTKdtIzQJ08OirmhslLWK4qiZCtpKehulRfd1iuKomQDaSnooSovKoqiZCtpKehOlReJxEvXDlJFUbKVtBT08eOB6dOBroFpNIiC9c9NB6mKuqIo2UZaCjogor5li4i6fTIL7SBVFCUbSVtBN2gHqaIoipD2gq4dpIqiKELaC3qoqekURVGyibQXdGsHKZG8Tp+uI0YVRck+0nLov51QU9MpiqJkC2kfoSuKoiiCCrqiKEqGkJGCrqV1FUXJRjLCQ7eipXUVRclWMi5C19K6iqJkKxkn6DpyVFGUbCXjBF1HjiqKkq1knKDryFFFUbKVjBN0+8jRli2Bxo2Bq67SjBdFUTKbjBN0IFha97nngKoqoKJCSuxqrXRFUTKZjBR0g1vGy/e+p9G6oiiZR0YLeqjMFo3WFUXJNDwJOhGdR0TriWgTEd3j8PlZRLSMiGqIaGzsmxkZ4TJbND9dUZRMIqygE1EugEcBnA+gD4BxRNTHttlWABMAPB/rBkaDU8aLHc1PVxQlU/ASoQ8GsImZP2fmowBmARhj3YCZtzDzCgDH4tDGiLFPJu0Es/rpiqJkBl4EvSOALyzL5YF1viGiiUS0hIiW7Nq1K5Jd+MZkvPz97+7RuvrpiqJkAgntFGXm6cxcysylrVu3TuShw0br6qcripLueBH0bQA6W5Y7BdalHSZaJ3L+XP10RVHSGS+CvhhATyLqTkQNAVwJYE58mxVftN6LoiiZSFhBZ+YaADcDmAtgLYAXmXk1EU0lotEAQESDiKgcwOUAHiei1fFsdLRovRdFUTIRYuakHLi0tJSXLFmSlGMD0gE6ebLYLC1ayLo9e+q+79JFRF4nxlAUJVUgoqXMXOr0WUaPFA2FW72Xigqt/aIoSnqStYJucKr3YkWzXxRFSReyXtC9ZLZo9ouiKOlA1gu6l8wWzX5RFCUdyHpBD1fvRbNfFEVJF7Je0J1mOGrZUt537SqfaZaLoijpQNYLOhDMeDl2DNi9W/6OHZPIfPJkICdHC3gpipL6NEh2A1KVmTMlZdFkwJgURkAjdkVRUhON0F3Q6esURUk3VNBd0OnrFEVJN1TQXfAyfZ1G64qipBIq6C54mb4O0GhdUZTUQQXdBS/T1xm0PICiKKmACnoIvExfZygrA1q1kj9Nc1QUJRmooHvAa7Rur9R41VUyQEmFXlGURKCC7hE/0brBlJrXkryKoiQCFXSf+PHW3dAMGUVR4oEKegSYaD0aUQckWr/2WrVjFEWJDSroUeA1tTEU1dXh7ZiZM0XsVfQVRQlF1s4pGiuc5iatqJDO0GgubcuW7vsqKNAqkIqSreiconHEqVIjs8xVai/J6wfTkQrUfzA4efAaxSuKohF6ArFXcIwFJnrXKF5RsgON0FMEp8k0GjaMbp9GxJ2ieB29qijZhQp6grFbNH/7W/TZMm6UlcXWlkl1WyfV26cocYeZk/I3cOBAVoL8/e/MBQXMEmvX/yNy/yzcn/mufR9muWtXOb5bu7p2df5+QYF8brYhCr2vWF0nc6yWLeXPvG/Y0Ll92UIi74OSPAAsYRddVUFPIdzEyvznDCf60fwZsQ4nkl4eNm77sp+P3+sSyYOta1f3axyN6IW7V/HE6Ryc/m1k2wMtW1BBzyBCiVs0UXyi/5yie6cHQCzOK9S+wrXDSaDDPVhjJaRObfJ73+0PNC/H6tqVedKk2D+wYvkLwu++UuXXSyzaoYKeoTj94zBiny5/qfIQCtUO+y8Ovw+ScCIZrweZ2YebiET6yyeSX2ChfkGEe6h7sdec2hTq14sXuzGS/39On3kJKPwQStA9pS0S0XkA/g9ALoAnmfk3ts8bAXgWwEAAFQCuYOYtofaZjWmLiSAeqZFKemNPaXVLdY3HMa0D5Ly2Lx7t8LKNaeuePcFBgm7vnQb8Wfdz4ABw9Ki3NnbtKokSXgmVttjAw5dzATwKYBSAcgCLiWgOM6+xbHY9gL3MfDIRXQngtwCu8N5EJVaYvPPJkyXLJdb/meP5ny9a8vKAZs1Ci0c2Yr9fZjme99Hs28u9SEQ7vGxjbauX927X1e+/v1DzF/vFS9riYACbmPlzZj4KYBaAMbZtxgB4JvD+JQAjiIhi10zFDyY1krnuiNWuXWXZuh6Qz9zIy5OIw/p9pxLC1aRRyAAABbNJREFUZh+JuOvmGGYErmnbjBmSChqvNFBFiQfh5i/2hZsXY/4AjIXYLGb5KgB/tm2zCkAny/JnAFo57GsigCUAlnTp0sWfcaTEjUgyNsL5sqG84Wg8bq8dYPFKAQ3llcdyn16vh5Mnm8h26F90f7H20B1X1tkghoJu/dNO0ewkVAdVrNIczXH8poA6dX6GEn/7f8Z4pJWG6uyzn2e0HePWh6ZTB26465HIv7y8yNuU7HOItkM2WkE/A8Bcy/K9AO61bTMXwBmB9w0A7EagTozbnwp69pJOKWR+f72E2t5Plks01ybUwyqSwWVezs/PQ3DSJP8P9VDXxW9GSTTjGtzE2b4f88CJxziFaAW9AYDPAXQH0BDApwD62ra5CcBfA++vBPBiuP2qoCtK/PBiicXrYerlIZjodvh9APt9sCQySAkl6F7TFi8A8DAkbfFvzPwgEU0N7HgOEeUDeA5ACYA9AK5k5s9D7VPTFhVFUfwTVdoiADDzmwDetK273/L+MIDLo2mkoiiKEh1abVFRFCVDUEFXFEXJEFTQFUVRMgQVdEVRlAzBU5ZLXA5MtAtAWYRfbwXJdc82svG8s/Gcgew872w8Z8D/eXdl5tZOHyRN0KOBiJa4pe1kMtl43tl4zkB2nnc2njMQ2/NWy0VRFCVDUEFXFEXJENJV0KcnuwFJIhvPOxvPGcjO887GcwZieN5p6aEriqIo9UnXCF1RFEWxoYKuKIqSIaSdoBPReUS0nog2EdE9yW5PPCCizkQ0n4jWENFqIrotsL4FEf2XiDYGXk9MdltjDRHlEtEnRPR6YLk7EX0cuN+ziahhstsYa4ioORG9RETriGgtEZ2RJff69sC/71VE9AIR5Wfa/SaivxHRV0S0yrLO8d6S8Ejg3FcQ0QC/x0srQbdMWH0+gD4AxhFRn+S2Ki7UALiTmfsAOB3ATYHzvAfAO8zcE8A7geVM4zYAay3LvwXwR2Y+GcBeyITkmcb/Afg3M58CoD/k/DP6XhNRRwC3Aihl5tMgpbnNBPOZdL+fBnCebZ3bvT0fQM/A30QAf/F7sLQSdHibsDrtYeYdzLws8P4A5D94R9SdjPsZABcnp4XxgYg6AbgQwJOBZQJwLmTicSAzz/kEAGcBeAoAmPkoM+9Dht/rAA0ANCaiBgAKAOxAht1vZl4ImSPCitu9HQPg2cA8Fh8BaE5E7f0cL90EvSOALyzL5YF1GQsRdYNMHPIxgLbMvCPw0ZcA2iapWfHiYQB3AzgWWG4JYB8z1wSWM/F+dwewC8CMgNX0JBE1QYbfa2beBmAagK0QId8PYCky/34D7vc2an1LN0HPKoioEMDLAH7EzF9bPwtMRZUxOadEdBGAr5h5abLbkmAaABgA4C/MXALgEGz2SqbdawAI+MZjIA+0DgCaoL41kfHE+t6mm6BvA9DZstwpsC7jIKI8iJjPZOZXAqt3mp9ggdevktW+ODAUwGgi2gKx0s6FeMvNAz/Jgcy83+UAypn548DySxCBz+R7DQAjAWxm5l3MXA3gFci/gUy/34D7vY1a39JN0BcD6BnoCW8I6USZk+Q2xZyAd/wUgLXM/AfLR3MAXBN4fw2AfyW6bfGCme9l5k7M3A1yX+cx83gA8wGMDWyWUecMAMz8JYAviKh3YNUIAGuQwfc6wFYApxNRQeDfuznvjL7fAdzu7RwAVweyXU4HsN9izXjDbfboVP0DcAGADQA+AzA52e2J0zl+A/IzbAWA5YG/CyCe8jsANgJ4G0CLZLc1Tuc/HMDrgfc9ACwCsAnAPwA0Snb74nC+xQCWBO73qwBOzIZ7DeDnANYBWAWZZL5Rpt1vAC9A+giqIb/Grne7twAIksX3GYCVkAwgX8fTof+KoigZQrpZLoqiKIoLKuiKoigZggq6oihKhqCCriiKkiGooCuKomQIKuiKoigZggq6oihKhvD/Qw56POtll0wAAAAASUVORK5CYII=\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/","height":545},"id":"JI7sAHurblLr","executionInfo":{"status":"ok","timestamp":1634549090446,"user_tz":-540,"elapsed":705,"user":{"displayName":"상원박","photoUrl":"https://lh3.googleusercontent.com/a/default-user=s64","userId":"16337076995970243718"}},"outputId":"e8fe7803-0763-4d90-e12c-80ecc16c5b8d"},"source":["def smooth_curve(points, factor=0.8):\n","  smoothed_points = []\n","  for point in points:\n","    if smoothed_points:\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,\n","         smooth_curve(acc), 'bo', label='Smoothed training acc')\n","plt.plot(epochs,\n","         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,\n","         smooth_curve(loss), 'bo', label='Smoothed training loss')\n","plt.plot(epochs,\n","         smooth_curve(val_loss), 'b', label='Smoothed validation loss')\n","plt.title('Training and validation loss')\n","plt.legend()\n","\n","plt.show()"],"execution_count":null,"outputs":[{"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","data":{"image/png":"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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"needs_background":"light"}}]},{"cell_type":"code","metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"cYB5DLa-bmuT","executionInfo":{"status":"ok","timestamp":1634549095852,"user_tz":-540,"elapsed":5413,"user":{"displayName":"상원박","photoUrl":"https://lh3.googleusercontent.com/a/default-user=s64","userId":"16337076995970243718"}},"outputId":"8ba28453-077a-4b40-89f5-11da0b75d437"},"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)"],"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["Found 1000 images belonging to 2 classes.\n"]},{"output_type":"stream","name":"stderr","text":["/usr/local/lib/python3.7/dist-packages/keras/engine/training.py:2006: UserWarning: `Model.evaluate_generator` is deprecated and will be removed in a future version. Please use `Model.evaluate`, which supports generators.\n","  warnings.warn('`Model.evaluate_generator` is deprecated and '\n"]},{"output_type":"stream","name":"stdout","text":["test acc: 0.9290000200271606\n"]}]},{"cell_type":"markdown","metadata":{"id":"UVOG8pC1CvZH"},"source":["## Conclusion\n","\n","- maximum acc = 0.93\n","- feature extraction + data augmentation + fine-tuning"]}]}