brain-mri-transfer-learning-data-aug-97-9.ipynb

{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"name":"python","version":"3.10.12","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"!pip install split-folders","metadata":{"execution":{"iopub.status.busy":"2023-07-16T06:28:21.251423Z","iopub.execute_input":"2023-07-16T06:28:21.251722Z","iopub.status.idle":"2023-07-16T06:28:35.218641Z","shell.execute_reply.started":"2023-07-16T06:28:21.251696Z","shell.execute_reply":"2023-07-16T06:28:35.217445Z"},"trusted":true},"execution_count":1,"outputs":[{"name":"stdout","text":"Collecting split-folders\n  Downloading split_folders-0.5.1-py3-none-any.whl (8.4 kB)\nInstalling collected packages: split-folders\nSuccessfully installed split-folders-0.5.1\n","output_type":"stream"}]},{"cell_type":"code","source":"import tensorflow as tf\ntf.random.set_seed(42)\nimport keras\nimport matplotlib.pyplot as plt\n%matplotlib inline\nimport os\nimport pathlib\nimport splitfolders\nimport shutil\nimport warnings\nimport numpy as np\nnp.random.seed(42)\nwarnings.filterwarnings('ignore')","metadata":{"execution":{"iopub.status.busy":"2023-07-16T06:28:35.221479Z","iopub.execute_input":"2023-07-16T06:28:35.222164Z","iopub.status.idle":"2023-07-16T06:28:43.471437Z","shell.execute_reply.started":"2023-07-16T06:28:35.222123Z","shell.execute_reply":"2023-07-16T06:28:43.470349Z"},"trusted":true},"execution_count":2,"outputs":[{"name":"stderr","text":"/opt/conda/lib/python3.10/site-packages/scipy/__init__.py:146: UserWarning: A NumPy version >=1.16.5 and <1.23.0 is required for this version of SciPy (detected version 1.23.5\n  warnings.warn(f\"A NumPy version >={np_minversion} and <{np_maxversion}\"\n/opt/conda/lib/python3.10/site-packages/tensorflow_io/python/ops/__init__.py:98: UserWarning: unable to load libtensorflow_io_plugins.so: unable to open file: libtensorflow_io_plugins.so, from paths: ['/opt/conda/lib/python3.10/site-packages/tensorflow_io/python/ops/libtensorflow_io_plugins.so']\ncaused by: ['/opt/conda/lib/python3.10/site-packages/tensorflow_io/python/ops/libtensorflow_io_plugins.so: undefined symbol: _ZN3tsl6StatusC1EN10tensorflow5error4CodeESt17basic_string_viewIcSt11char_traitsIcEENS_14SourceLocationE']\n  warnings.warn(f\"unable to load libtensorflow_io_plugins.so: {e}\")\n/opt/conda/lib/python3.10/site-packages/tensorflow_io/python/ops/__init__.py:104: UserWarning: file system plugins are not loaded: unable to open file: libtensorflow_io.so, from paths: ['/opt/conda/lib/python3.10/site-packages/tensorflow_io/python/ops/libtensorflow_io.so']\ncaused by: ['/opt/conda/lib/python3.10/site-packages/tensorflow_io/python/ops/libtensorflow_io.so: undefined symbol: _ZTVN10tensorflow13GcsFileSystemE']\n  warnings.warn(f\"file system plugins are not loaded: {e}\")\n","output_type":"stream"}]},{"cell_type":"markdown","source":"## Lets look at the number of images per class","metadata":{}},{"cell_type":"code","source":"no_path = \"/kaggle/input/brain-mri-images-for-brain-tumor-detection/no\"\nyes_path = \"/kaggle/input/brain-mri-images-for-brain-tumor-detection/yes\"\n\nprint(len(os.listdir(no_path)))\nprint(len(os.listdir(yes_path)))","metadata":{"execution":{"iopub.status.busy":"2023-07-16T06:28:43.474887Z","iopub.execute_input":"2023-07-16T06:28:43.475490Z","iopub.status.idle":"2023-07-16T06:28:43.496117Z","shell.execute_reply.started":"2023-07-16T06:28:43.475461Z","shell.execute_reply":"2023-07-16T06:28:43.495082Z"},"trusted":true},"execution_count":3,"outputs":[{"name":"stdout","text":"98\n155\n","output_type":"stream"}]},{"cell_type":"markdown","source":"This shows that the data is imbalanced, lets balance it using data augmentation","metadata":{}},{"cell_type":"markdown","source":"## Data Augmentation","metadata":{}},{"cell_type":"code","source":"# Making temporary directories that will be deleted later\nmain_dir = 'temp'\nos.mkdir(main_dir)\nno_dir = os.path.join(main_dir, 'no')\nos.mkdir(no_dir)\nyes_dir = os.path.join(main_dir, 'yes')\nos.mkdir(yes_dir)","metadata":{"execution":{"iopub.status.busy":"2023-07-16T06:28:43.499052Z","iopub.execute_input":"2023-07-16T06:28:43.499412Z","iopub.status.idle":"2023-07-16T06:28:43.505597Z","shell.execute_reply.started":"2023-07-16T06:28:43.499378Z","shell.execute_reply":"2023-07-16T06:28:43.504174Z"},"trusted":true},"execution_count":4,"outputs":[]},{"cell_type":"code","source":"# We are creating temporary directories here because we delete these directories later\naug_dir = 'aug_dir'\nos.mkdir(aug_dir)\n\n# create a dir within the base dir to store images of the same class\nimg_dir = os.path.join(aug_dir, 'img_dir')\nos.mkdir(img_dir)\n\n# We are only data augmenting the \"NO\" class\nimg_class = 'no'\n\n# list all images in that directory\nimg_list = os.listdir('/kaggle/input/brain-mri-images-for-brain-tumor-detection/' + img_class)\n\n\nfor fname in img_list:\n    # source path to image\n    src = os.path.join('/kaggle/input/brain-mri-images-for-brain-tumor-detection/' + img_class, fname)\n    # destination path to image\n    dst = os.path.join(img_dir, fname)\n    # copy the image from the source to the destination\n    shutil.copyfile(src, dst)\n\n\n# point to a dir containing the images and not to the images themselves\npath = aug_dir\nsave_path = '/kaggle/working/temp/' + img_class\n\n# Create a data generator\ndatagen = tf.keras.preprocessing.image.ImageDataGenerator(\n    rotation_range=20,\n    zoom_range=0.2,\n    horizontal_flip=True,\n    width_shift_range=0.2,\n    height_shift_range=0.2,\n    shear_range=0.2,\n    fill_mode='nearest')\n\nbatch_size = 1\n\naug_datagen = datagen.flow_from_directory(path,\n                                          save_to_dir=save_path,\n                                          save_format='jpg',\n                                          target_size=(224,224),\n                                          batch_size=batch_size)\n\n\n# Generate the augmented images and add them to the training folders\n    \n# total number of images we want to have in each class\nnum_aug_images_wanted = 250 \n# (Note: We may or may not get the intended num of augmented images, so play will the number. I put 250 here to get 150 images)\n\nnum_files = len(os.listdir(img_dir))\nnum_batches = int(np.ceil((num_aug_images_wanted-num_files)/batch_size))\n\n# run the generator and create our augmented images\nfor i in range(0,num_batches):\n    imgs, labels = next(aug_datagen)\n        \n# delete temporary directory with the raw image files\nshutil.rmtree('aug_dir')","metadata":{"execution":{"iopub.status.busy":"2023-07-16T06:28:43.507852Z","iopub.execute_input":"2023-07-16T06:28:43.508613Z","iopub.status.idle":"2023-07-16T06:28:45.706787Z","shell.execute_reply.started":"2023-07-16T06:28:43.508580Z","shell.execute_reply":"2023-07-16T06:28:45.705736Z"},"trusted":true},"execution_count":5,"outputs":[{"name":"stdout","text":"Found 98 images belonging to 1 classes.\n","output_type":"stream"}]},{"cell_type":"code","source":"# Copying Yes class images in temp directory\norigin_path = '/kaggle/input/brain-mri-images-for-brain-tumor-detection/yes/'\ntarget_path = '/kaggle/working/temp/yes/'\n\n# Fetching the list of all the files\nfiles = os.listdir(origin_path)\n\n# Fetching all the files to directory\nfor file_name in files:\n    shutil.copy(origin_path+file_name, target_path+file_name)\nprint(\"Files are copied successfully\")","metadata":{"execution":{"iopub.status.busy":"2023-07-16T06:28:45.708378Z","iopub.execute_input":"2023-07-16T06:28:45.708950Z","iopub.status.idle":"2023-07-16T06:28:46.260271Z","shell.execute_reply.started":"2023-07-16T06:28:45.708916Z","shell.execute_reply":"2023-07-16T06:28:46.259305Z"},"trusted":true},"execution_count":6,"outputs":[{"name":"stdout","text":"Files are copied successfully\n","output_type":"stream"}]},{"cell_type":"code","source":"no_path = \"/kaggle/working/temp/no\"\nyes_path = \"/kaggle/working/temp/yes\"\n\nprint(len(os.listdir(no_path)))\nprint(len(os.listdir(yes_path)))","metadata":{"execution":{"iopub.status.busy":"2023-07-16T06:28:46.262437Z","iopub.execute_input":"2023-07-16T06:28:46.263154Z","iopub.status.idle":"2023-07-16T06:28:46.270189Z","shell.execute_reply.started":"2023-07-16T06:28:46.263107Z","shell.execute_reply":"2023-07-16T06:28:46.269174Z"},"trusted":true},"execution_count":7,"outputs":[{"name":"stdout","text":"152\n155\n","output_type":"stream"}]},{"cell_type":"markdown","source":"#### The data looks balanced enough.","metadata":{}},{"cell_type":"markdown","source":"## Splitting dataset into train, test and val","metadata":{}},{"cell_type":"code","source":"splitfolders.ratio('/kaggle/working/temp', output='Images', seed=42, ratio=(0.7,0.15,0.15))","metadata":{"execution":{"iopub.status.busy":"2023-07-16T06:28:46.271438Z","iopub.execute_input":"2023-07-16T06:28:46.271992Z","iopub.status.idle":"2023-07-16T06:28:46.335895Z","shell.execute_reply.started":"2023-07-16T06:28:46.271957Z","shell.execute_reply":"2023-07-16T06:28:46.334865Z"},"trusted":true},"execution_count":8,"outputs":[{"name":"stderr","text":"Copying files: 307 files [00:00, 6279.45 files/s]\n","output_type":"stream"}]},{"cell_type":"code","source":"# Deleting the temp directory\nshutil.rmtree('/kaggle/working/temp/yes')\nshutil.rmtree('/kaggle/working/temp/no')","metadata":{"execution":{"iopub.status.busy":"2023-07-16T06:28:46.337427Z","iopub.execute_input":"2023-07-16T06:28:46.338312Z","iopub.status.idle":"2023-07-16T06:28:46.353996Z","shell.execute_reply.started":"2023-07-16T06:28:46.338284Z","shell.execute_reply":"2023-07-16T06:28:46.353058Z"},"trusted":true},"execution_count":9,"outputs":[]},{"cell_type":"markdown","source":"## Preparing the data","metadata":{}},{"cell_type":"code","source":"data_generator = tf.keras.preprocessing.image.ImageDataGenerator(rescale=1/255.)\n\ntrain_generator = data_generator.flow_from_directory(\n    directory = '/kaggle/working/Images/train',\n    class_mode = 'binary',\n    seed = 42,\n    target_size = (224, 224),\n)\nvalid_generator = data_generator.flow_from_directory(\n    directory = '/kaggle/working/Images/val',\n    class_mode = 'binary',\n    seed = 42,\n    target_size = (224, 224),\n)\ntest_generator = data_generator.flow_from_directory(\n    directory = '/kaggle/working/Images/test',\n    class_mode = 'binary',\n    seed = 42,\n    target_size = (224, 224),\n)","metadata":{"execution":{"iopub.status.busy":"2023-07-16T06:28:46.358958Z","iopub.execute_input":"2023-07-16T06:28:46.360886Z","iopub.status.idle":"2023-07-16T06:28:46.391379Z","shell.execute_reply.started":"2023-07-16T06:28:46.360848Z","shell.execute_reply":"2023-07-16T06:28:46.390563Z"},"trusted":true},"execution_count":10,"outputs":[{"name":"stdout","text":"Found 214 images belonging to 2 classes.\nFound 45 images belonging to 2 classes.\nFound 48 images belonging to 2 classes.\n","output_type":"stream"}]},{"cell_type":"markdown","source":"## VGG16","metadata":{}},{"cell_type":"code","source":"basemodel_vgg16 = tf.keras.applications.vgg16.VGG16(include_top=False, weights=\"imagenet\", input_shape=(224,224, 3), classes=2)\nbasemodel_vgg16.trainable = False\nmodel_vgg16 = tf.keras.Sequential([\n    basemodel_vgg16,\n    tf.keras.layers.GlobalAveragePooling2D(),\n    keras.layers.Dense(256, activation='relu', kernel_initializer='he_normal'),\n    keras.layers.Dense(128, activation='relu', kernel_initializer='he_normal'),    \n    keras.layers.Dense(64, activation='relu', kernel_initializer='he_normal'),\n    keras.layers.Dense(32, activation='relu', kernel_initializer='he_normal'),    \n    keras.layers.Dense(16, activation='relu', kernel_initializer='he_normal'),\n    keras.layers.Dense(2, activation='softmax')\n])\n\nmodel_vgg16.summary()","metadata":{"execution":{"iopub.status.busy":"2023-07-16T06:28:46.392329Z","iopub.execute_input":"2023-07-16T06:28:46.392570Z","iopub.status.idle":"2023-07-16T06:28:50.395036Z","shell.execute_reply.started":"2023-07-16T06:28:46.392548Z","shell.execute_reply":"2023-07-16T06:28:50.394284Z"},"trusted":true},"execution_count":11,"outputs":[{"name":"stdout","text":"Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/vgg16/vgg16_weights_tf_dim_ordering_tf_kernels_notop.h5\n58889256/58889256 [==============================] - 0s 0us/step\nModel: \"sequential\"\n_________________________________________________________________\n Layer (type)                Output Shape              Param #   \n=================================================================\n vgg16 (Functional)          (None, 7, 7, 512)         14714688  \n                                                                 \n global_average_pooling2d (G  (None, 512)              0         \n lobalAveragePooling2D)                                          \n                                                                 \n dense (Dense)               (None, 256)               131328    \n                                                                 \n dense_1 (Dense)             (None, 128)               32896     \n                                                                 \n dense_2 (Dense)             (None, 64)                8256      \n                                                                 \n dense_3 (Dense)             (None, 32)                2080      \n                                                                 \n dense_4 (Dense)             (None, 16)                528       \n                                                                 \n dense_5 (Dense)             (None, 2)                 34        \n                                                                 \n=================================================================\nTotal params: 14,889,810\nTrainable params: 175,122\nNon-trainable params: 14,714,688\n_________________________________________________________________\n","output_type":"stream"}]},{"cell_type":"code","source":"cb_earlystopper = tf.keras.callbacks.EarlyStopping(monitor='val_loss', patience=10, verbose=1)\ncb_checkpointer = tf.keras.callbacks.ModelCheckpoint(filepath='../working/model_vgg16.hdf5', monitor='val_loss', save_best_only=True, mode='auto', verbose=1)\n\nCALLBACKS = [cb_earlystopper, cb_checkpointer]","metadata":{"execution":{"iopub.status.busy":"2023-07-16T06:28:50.396046Z","iopub.execute_input":"2023-07-16T06:28:50.396389Z","iopub.status.idle":"2023-07-16T06:28:50.406724Z","shell.execute_reply.started":"2023-07-16T06:28:50.396359Z","shell.execute_reply":"2023-07-16T06:28:50.405865Z"},"trusted":true},"execution_count":12,"outputs":[]},{"cell_type":"code","source":"model_vgg16.compile(optimizer='rmsprop', loss='sparse_categorical_crossentropy', metrics='accuracy')\nhistory_vgg16 = model_vgg16.fit(\n    train_generator,\n    epochs=300,\n    validation_data=valid_generator,\n    callbacks=CALLBACKS\n)","metadata":{"execution":{"iopub.status.busy":"2023-07-16T06:28:50.407977Z","iopub.execute_input":"2023-07-16T06:28:50.408614Z","iopub.status.idle":"2023-07-16T06:30:27.276562Z","shell.execute_reply.started":"2023-07-16T06:28:50.408573Z","shell.execute_reply":"2023-07-16T06:30:27.275448Z"},"trusted":true},"execution_count":13,"outputs":[{"name":"stdout","text":"Epoch 1/300\n7/7 [==============================] - ETA: 0s - loss: 15.2240 - accuracy: 0.4860\nEpoch 1: val_loss improved from inf to 0.66160, saving model to ../working/model_vgg16.hdf5\n7/7 [==============================] - 13s 586ms/step - loss: 15.2240 - accuracy: 0.4860 - val_loss: 0.6616 - val_accuracy: 0.5111\nEpoch 2/300\n7/7 [==============================] - ETA: 0s - loss: 6.8833 - accuracy: 0.6028\nEpoch 2: val_loss improved from 0.66160 to 0.64230, saving model to ../working/model_vgg16.hdf5\n7/7 [==============================] - 1s 194ms/step - loss: 6.8833 - accuracy: 0.6028 - val_loss: 0.6423 - val_accuracy: 0.6444\nEpoch 3/300\n7/7 [==============================] - ETA: 0s - loss: 10.2162 - accuracy: 0.6682\nEpoch 3: val_loss improved from 0.64230 to 0.63027, saving model to ../working/model_vgg16.hdf5\n7/7 [==============================] - 1s 152ms/step - loss: 10.2162 - accuracy: 0.6682 - val_loss: 0.6303 - val_accuracy: 0.5111\nEpoch 4/300\n7/7 [==============================] - ETA: 0s - loss: 1.7720 - accuracy: 0.6495\nEpoch 4: val_loss improved from 0.63027 to 0.52984, saving model to ../working/model_vgg16.hdf5\n7/7 [==============================] - 1s 129ms/step - loss: 1.7720 - accuracy: 0.6495 - val_loss: 0.5298 - val_accuracy: 0.8667\nEpoch 5/300\n7/7 [==============================] - ETA: 0s - loss: 0.9998 - accuracy: 0.7290\nEpoch 5: val_loss improved from 0.52984 to 0.50568, saving model to ../working/model_vgg16.hdf5\n7/7 [==============================] - 1s 159ms/step - loss: 0.9998 - accuracy: 0.7290 - val_loss: 0.5057 - val_accuracy: 0.7778\nEpoch 6/300\n7/7 [==============================] - ETA: 0s - loss: 0.8662 - accuracy: 0.7617\nEpoch 6: val_loss did not improve from 0.50568\n7/7 [==============================] - 1s 109ms/step - loss: 0.8662 - accuracy: 0.7617 - val_loss: 0.6412 - val_accuracy: 0.5111\nEpoch 7/300\n7/7 [==============================] - ETA: 0s - loss: 3.4880 - accuracy: 0.7944\nEpoch 7: val_loss improved from 0.50568 to 0.47458, saving model to ../working/model_vgg16.hdf5\n7/7 [==============================] - 1s 137ms/step - loss: 3.4880 - accuracy: 0.7944 - val_loss: 0.4746 - val_accuracy: 0.7333\nEpoch 8/300\n7/7 [==============================] - ETA: 0s - loss: 0.4747 - accuracy: 0.8364\nEpoch 8: val_loss did not improve from 0.47458\n7/7 [==============================] - 1s 106ms/step - loss: 0.4747 - accuracy: 0.8364 - val_loss: 0.6168 - val_accuracy: 0.5333\nEpoch 9/300\n7/7 [==============================] - ETA: 0s - loss: 0.5130 - accuracy: 0.7617\nEpoch 9: val_loss improved from 0.47458 to 0.29021, saving model to ../working/model_vgg16.hdf5\n7/7 [==============================] - 1s 132ms/step - loss: 0.5130 - accuracy: 0.7617 - val_loss: 0.2902 - val_accuracy: 0.9556\nEpoch 10/300\n7/7 [==============================] - ETA: 0s - loss: 0.5017 - accuracy: 0.7290\nEpoch 10: val_loss did not improve from 0.29021\n7/7 [==============================] - 1s 103ms/step - loss: 0.5017 - accuracy: 0.7290 - val_loss: 0.2999 - val_accuracy: 0.9333\nEpoch 11/300\n7/7 [==============================] - ETA: 0s - loss: 0.3936 - accuracy: 0.8458\nEpoch 11: val_loss did not improve from 0.29021\n7/7 [==============================] - 1s 99ms/step - loss: 0.3936 - accuracy: 0.8458 - val_loss: 0.2968 - val_accuracy: 0.8889\nEpoch 12/300\n7/7 [==============================] - ETA: 0s - loss: 0.3962 - accuracy: 0.8037\nEpoch 12: val_loss improved from 0.29021 to 0.26093, saving model to ../working/model_vgg16.hdf5\n7/7 [==============================] - 1s 126ms/step - loss: 0.3962 - accuracy: 0.8037 - val_loss: 0.2609 - val_accuracy: 0.9111\nEpoch 13/300\n7/7 [==============================] - ETA: 0s - loss: 0.4129 - accuracy: 0.8131\nEpoch 13: val_loss did not improve from 0.26093\n7/7 [==============================] - 1s 132ms/step - loss: 0.4129 - accuracy: 0.8131 - val_loss: 0.2924 - val_accuracy: 0.8889\nEpoch 14/300\n7/7 [==============================] - ETA: 0s - loss: 0.4008 - accuracy: 0.8131\nEpoch 14: val_loss improved from 0.26093 to 0.16930, saving model to ../working/model_vgg16.hdf5\n7/7 [==============================] - 1s 125ms/step - loss: 0.4008 - accuracy: 0.8131 - val_loss: 0.1693 - val_accuracy: 1.0000\nEpoch 15/300\n7/7 [==============================] - ETA: 0s - loss: 0.4052 - accuracy: 0.8178\nEpoch 15: val_loss improved from 0.16930 to 0.16378, saving model to ../working/model_vgg16.hdf5\n7/7 [==============================] - 1s 130ms/step - loss: 0.4052 - accuracy: 0.8178 - val_loss: 0.1638 - val_accuracy: 1.0000\nEpoch 16/300\n7/7 [==============================] - ETA: 0s - loss: 0.3669 - accuracy: 0.8178\nEpoch 16: val_loss did not improve from 0.16378\n7/7 [==============================] - 1s 102ms/step - loss: 0.3669 - accuracy: 0.8178 - val_loss: 0.2132 - val_accuracy: 0.9111\nEpoch 17/300\n7/7 [==============================] - ETA: 0s - loss: 0.3737 - accuracy: 0.8458\nEpoch 17: val_loss improved from 0.16378 to 0.12370, saving model to ../working/model_vgg16.hdf5\n7/7 [==============================] - 1s 157ms/step - loss: 0.3737 - accuracy: 0.8458 - val_loss: 0.1237 - val_accuracy: 1.0000\nEpoch 18/300\n7/7 [==============================] - ETA: 0s - loss: 0.3706 - accuracy: 0.8224\nEpoch 18: val_loss did not improve from 0.12370\n7/7 [==============================] - 1s 105ms/step - loss: 0.3706 - accuracy: 0.8224 - val_loss: 0.2184 - val_accuracy: 0.9111\nEpoch 19/300\n7/7 [==============================] - ETA: 0s - loss: 0.2671 - accuracy: 0.9019\nEpoch 19: val_loss did not improve from 0.12370\n7/7 [==============================] - 1s 101ms/step - loss: 0.2671 - accuracy: 0.9019 - val_loss: 0.2649 - val_accuracy: 0.8889\nEpoch 20/300\n7/7 [==============================] - ETA: 0s - loss: 0.2230 - accuracy: 0.9393\nEpoch 20: val_loss improved from 0.12370 to 0.07400, saving model to ../working/model_vgg16.hdf5\n7/7 [==============================] - 1s 130ms/step - loss: 0.2230 - accuracy: 0.9393 - val_loss: 0.0740 - val_accuracy: 1.0000\nEpoch 21/300\n7/7 [==============================] - ETA: 0s - loss: 0.4898 - accuracy: 0.7757\nEpoch 21: val_loss did not improve from 0.07400\n7/7 [==============================] - 1s 99ms/step - loss: 0.4898 - accuracy: 0.7757 - val_loss: 0.0909 - val_accuracy: 1.0000\nEpoch 22/300\n7/7 [==============================] - ETA: 0s - loss: 0.2976 - accuracy: 0.8551\nEpoch 22: val_loss did not improve from 0.07400\n7/7 [==============================] - 1s 102ms/step - loss: 0.2976 - accuracy: 0.8551 - val_loss: 0.1073 - val_accuracy: 0.9778\nEpoch 23/300\n7/7 [==============================] - ETA: 0s - loss: 0.2169 - accuracy: 0.8925\nEpoch 23: val_loss improved from 0.07400 to 0.06162, saving model to ../working/model_vgg16.hdf5\n7/7 [==============================] - 1s 131ms/step - loss: 0.2169 - accuracy: 0.8925 - val_loss: 0.0616 - val_accuracy: 1.0000\nEpoch 24/300\n7/7 [==============================] - ETA: 0s - loss: 0.4663 - accuracy: 0.7897\nEpoch 24: val_loss did not improve from 0.06162\n7/7 [==============================] - 1s 113ms/step - loss: 0.4663 - accuracy: 0.7897 - val_loss: 0.0887 - val_accuracy: 1.0000\nEpoch 25/300\n7/7 [==============================] - ETA: 0s - loss: 0.2457 - accuracy: 0.8785\nEpoch 25: val_loss did not improve from 0.06162\n7/7 [==============================] - 1s 109ms/step - loss: 0.2457 - accuracy: 0.8785 - val_loss: 0.0907 - val_accuracy: 0.9778\nEpoch 26/300\n7/7 [==============================] - ETA: 0s - loss: 0.2872 - accuracy: 0.8598\nEpoch 26: val_loss did not improve from 0.06162\n7/7 [==============================] - 1s 101ms/step - loss: 0.2872 - accuracy: 0.8598 - val_loss: 0.0795 - val_accuracy: 1.0000\nEpoch 27/300\n7/7 [==============================] - ETA: 0s - loss: 0.1872 - accuracy: 0.9112\nEpoch 27: val_loss improved from 0.06162 to 0.04921, saving model to ../working/model_vgg16.hdf5\n7/7 [==============================] - 1s 131ms/step - loss: 0.1872 - accuracy: 0.9112 - val_loss: 0.0492 - val_accuracy: 1.0000\nEpoch 28/300\n7/7 [==============================] - ETA: 0s - loss: 0.3600 - accuracy: 0.8178\nEpoch 28: val_loss did not improve from 0.04921\n7/7 [==============================] - 1s 102ms/step - loss: 0.3600 - accuracy: 0.8178 - val_loss: 0.0665 - val_accuracy: 1.0000\nEpoch 29/300\n7/7 [==============================] - ETA: 0s - loss: 0.2712 - accuracy: 0.8645\nEpoch 29: val_loss did not improve from 0.04921\n7/7 [==============================] - 1s 129ms/step - loss: 0.2712 - accuracy: 0.8645 - val_loss: 0.1055 - val_accuracy: 0.9778\nEpoch 30/300\n7/7 [==============================] - ETA: 0s - loss: 0.1705 - accuracy: 0.9299\nEpoch 30: val_loss improved from 0.04921 to 0.04567, saving model to ../working/model_vgg16.hdf5\n7/7 [==============================] - 1s 129ms/step - loss: 0.1705 - accuracy: 0.9299 - val_loss: 0.0457 - val_accuracy: 1.0000\nEpoch 31/300\n7/7 [==============================] - ETA: 0s - loss: 0.1851 - accuracy: 0.8925\nEpoch 31: val_loss did not improve from 0.04567\n7/7 [==============================] - 1s 100ms/step - loss: 0.1851 - accuracy: 0.8925 - val_loss: 0.3231 - val_accuracy: 0.8444\nEpoch 32/300\n7/7 [==============================] - ETA: 0s - loss: 0.4085 - accuracy: 0.7850\nEpoch 32: val_loss did not improve from 0.04567\n7/7 [==============================] - 1s 104ms/step - loss: 0.4085 - accuracy: 0.7850 - val_loss: 0.0714 - val_accuracy: 0.9778\nEpoch 33/300\n7/7 [==============================] - ETA: 0s - loss: 0.1553 - accuracy: 0.9299\nEpoch 33: val_loss did not improve from 0.04567\n7/7 [==============================] - 1s 107ms/step - loss: 0.1553 - accuracy: 0.9299 - val_loss: 0.0530 - val_accuracy: 1.0000\nEpoch 34/300\n7/7 [==============================] - ETA: 0s - loss: 0.1709 - accuracy: 0.9252\nEpoch 34: val_loss did not improve from 0.04567\n7/7 [==============================] - 1s 128ms/step - loss: 0.1709 - accuracy: 0.9252 - val_loss: 0.0474 - val_accuracy: 1.0000\nEpoch 35/300\n7/7 [==============================] - ETA: 0s - loss: 0.2671 - accuracy: 0.8551\nEpoch 35: val_loss improved from 0.04567 to 0.04457, saving model to ../working/model_vgg16.hdf5\n7/7 [==============================] - 1s 155ms/step - loss: 0.2671 - accuracy: 0.8551 - val_loss: 0.0446 - val_accuracy: 1.0000\nEpoch 36/300\n7/7 [==============================] - ETA: 0s - loss: 0.1271 - accuracy: 0.9393\nEpoch 36: val_loss improved from 0.04457 to 0.02901, saving model to ../working/model_vgg16.hdf5\n7/7 [==============================] - 1s 126ms/step - loss: 0.1271 - accuracy: 0.9393 - val_loss: 0.0290 - val_accuracy: 1.0000\nEpoch 37/300\n7/7 [==============================] - ETA: 0s - loss: 0.2145 - accuracy: 0.9159\nEpoch 37: val_loss did not improve from 0.02901\n7/7 [==============================] - 1s 99ms/step - loss: 0.2145 - accuracy: 0.9159 - val_loss: 0.5905 - val_accuracy: 0.7556\nEpoch 38/300\n7/7 [==============================] - ETA: 0s - loss: 0.3572 - accuracy: 0.8318\nEpoch 38: val_loss did not improve from 0.02901\n7/7 [==============================] - 1s 104ms/step - loss: 0.3572 - accuracy: 0.8318 - val_loss: 0.0403 - val_accuracy: 1.0000\nEpoch 39/300\n7/7 [==============================] - ETA: 0s - loss: 0.1194 - accuracy: 0.9486\nEpoch 39: val_loss improved from 0.02901 to 0.02614, saving model to ../working/model_vgg16.hdf5\n7/7 [==============================] - 1s 130ms/step - loss: 0.1194 - accuracy: 0.9486 - val_loss: 0.0261 - val_accuracy: 1.0000\nEpoch 40/300\n7/7 [==============================] - ETA: 0s - loss: 0.1432 - accuracy: 0.9299\nEpoch 40: val_loss did not improve from 0.02614\n7/7 [==============================] - 1s 104ms/step - loss: 0.1432 - accuracy: 0.9299 - val_loss: 0.1215 - val_accuracy: 0.9333\nEpoch 41/300\n7/7 [==============================] - ETA: 0s - loss: 0.1207 - accuracy: 0.9299\nEpoch 41: val_loss improved from 0.02614 to 0.01580, saving model to ../working/model_vgg16.hdf5\n7/7 [==============================] - 1s 154ms/step - loss: 0.1207 - accuracy: 0.9299 - val_loss: 0.0158 - val_accuracy: 1.0000\nEpoch 42/300\n7/7 [==============================] - ETA: 0s - loss: 0.5259 - accuracy: 0.8364\nEpoch 42: val_loss did not improve from 0.01580\n7/7 [==============================] - 1s 109ms/step - loss: 0.5259 - accuracy: 0.8364 - val_loss: 0.0332 - val_accuracy: 1.0000\nEpoch 43/300\n7/7 [==============================] - ETA: 0s - loss: 0.0967 - accuracy: 0.9533\nEpoch 43: val_loss did not improve from 0.01580\n7/7 [==============================] - 1s 108ms/step - loss: 0.0967 - accuracy: 0.9533 - val_loss: 0.0234 - val_accuracy: 1.0000\nEpoch 44/300\n7/7 [==============================] - ETA: 0s - loss: 0.0969 - accuracy: 0.9299\nEpoch 44: val_loss improved from 0.01580 to 0.01558, saving model to ../working/model_vgg16.hdf5\n7/7 [==============================] - 1s 127ms/step - loss: 0.0969 - accuracy: 0.9299 - val_loss: 0.0156 - val_accuracy: 1.0000\nEpoch 45/300\n7/7 [==============================] - ETA: 0s - loss: 0.0852 - accuracy: 0.9486\nEpoch 45: val_loss did not improve from 0.01558\n7/7 [==============================] - 1s 108ms/step - loss: 0.0852 - accuracy: 0.9486 - val_loss: 0.0328 - val_accuracy: 0.9778\nEpoch 46/300\n7/7 [==============================] - ETA: 0s - loss: 24.5007 - accuracy: 0.8645\nEpoch 46: val_loss did not improve from 0.01558\n7/7 [==============================] - 1s 97ms/step - loss: 24.5007 - accuracy: 0.8645 - val_loss: 0.0298 - val_accuracy: 1.0000\nEpoch 47/300\n7/7 [==============================] - ETA: 0s - loss: 17.9595 - accuracy: 0.9299\nEpoch 47: val_loss did not improve from 0.01558\n7/7 [==============================] - 1s 102ms/step - loss: 17.9595 - accuracy: 0.9299 - val_loss: 0.0235 - val_accuracy: 1.0000\nEpoch 48/300\n7/7 [==============================] - ETA: 0s - loss: 0.5839 - accuracy: 0.9439\nEpoch 48: val_loss did not improve from 0.01558\n7/7 [==============================] - 1s 99ms/step - loss: 0.5839 - accuracy: 0.9439 - val_loss: 0.0203 - val_accuracy: 1.0000\nEpoch 49/300\n7/7 [==============================] - ETA: 0s - loss: 0.1012 - accuracy: 0.9159\nEpoch 49: val_loss did not improve from 0.01558\n7/7 [==============================] - 1s 98ms/step - loss: 0.1012 - accuracy: 0.9159 - val_loss: 0.0200 - val_accuracy: 1.0000\nEpoch 50/300\n7/7 [==============================] - ETA: 0s - loss: 0.0898 - accuracy: 0.9673\nEpoch 50: val_loss improved from 0.01558 to 0.01539, saving model to ../working/model_vgg16.hdf5\n7/7 [==============================] - 1s 127ms/step - loss: 0.0898 - accuracy: 0.9673 - val_loss: 0.0154 - val_accuracy: 1.0000\nEpoch 51/300\n7/7 [==============================] - ETA: 0s - loss: 0.1183 - accuracy: 0.9159\nEpoch 51: val_loss did not improve from 0.01539\n7/7 [==============================] - 1s 108ms/step - loss: 0.1183 - accuracy: 0.9159 - val_loss: 0.6266 - val_accuracy: 0.8222\nEpoch 52/300\n7/7 [==============================] - ETA: 0s - loss: 0.1970 - accuracy: 0.9252\nEpoch 52: val_loss did not improve from 0.01539\n7/7 [==============================] - 1s 102ms/step - loss: 0.1970 - accuracy: 0.9252 - val_loss: 0.0278 - val_accuracy: 1.0000\nEpoch 53/300\n7/7 [==============================] - ETA: 0s - loss: 0.0847 - accuracy: 0.9533\nEpoch 53: val_loss did not improve from 0.01539\n7/7 [==============================] - 1s 101ms/step - loss: 0.0847 - accuracy: 0.9533 - val_loss: 0.0213 - val_accuracy: 1.0000\nEpoch 54/300\n7/7 [==============================] - ETA: 0s - loss: 0.1512 - accuracy: 0.9346\nEpoch 54: val_loss did not improve from 0.01539\n7/7 [==============================] - 1s 101ms/step - loss: 0.1512 - accuracy: 0.9346 - val_loss: 0.3747 - val_accuracy: 0.8444\nEpoch 55/300\n7/7 [==============================] - ETA: 0s - loss: 0.2165 - accuracy: 0.9019\nEpoch 55: val_loss did not improve from 0.01539\n7/7 [==============================] - 1s 104ms/step - loss: 0.2165 - accuracy: 0.9019 - val_loss: 0.0177 - val_accuracy: 1.0000\nEpoch 56/300\n7/7 [==============================] - ETA: 0s - loss: 0.0818 - accuracy: 0.9393\nEpoch 56: val_loss improved from 0.01539 to 0.01279, saving model to ../working/model_vgg16.hdf5\n7/7 [==============================] - 1s 144ms/step - loss: 0.0818 - accuracy: 0.9393 - val_loss: 0.0128 - val_accuracy: 1.0000\nEpoch 57/300\n7/7 [==============================] - ETA: 0s - loss: 0.0783 - accuracy: 0.9673\nEpoch 57: val_loss improved from 0.01279 to 0.01077, saving model to ../working/model_vgg16.hdf5\n7/7 [==============================] - 1s 128ms/step - loss: 0.0783 - accuracy: 0.9673 - val_loss: 0.0108 - val_accuracy: 1.0000\nEpoch 58/300\n7/7 [==============================] - ETA: 0s - loss: 0.0792 - accuracy: 0.9439\nEpoch 58: val_loss did not improve from 0.01077\n7/7 [==============================] - 1s 99ms/step - loss: 0.0792 - accuracy: 0.9439 - val_loss: 0.0110 - val_accuracy: 1.0000\nEpoch 59/300\n7/7 [==============================] - ETA: 0s - loss: 0.3382 - accuracy: 0.8598\nEpoch 59: val_loss did not improve from 0.01077\n7/7 [==============================] - 1s 101ms/step - loss: 0.3382 - accuracy: 0.8598 - val_loss: 0.5167 - val_accuracy: 0.8444\nEpoch 60/300\n7/7 [==============================] - ETA: 0s - loss: 0.1481 - accuracy: 0.9299\nEpoch 60: val_loss did not improve from 0.01077\n7/7 [==============================] - 1s 99ms/step - loss: 0.1481 - accuracy: 0.9299 - val_loss: 0.0147 - val_accuracy: 1.0000\nEpoch 61/300\n7/7 [==============================] - ETA: 0s - loss: 0.0808 - accuracy: 0.9486\nEpoch 61: val_loss did not improve from 0.01077\n7/7 [==============================] - 1s 99ms/step - loss: 0.0808 - accuracy: 0.9486 - val_loss: 0.0126 - val_accuracy: 1.0000\nEpoch 62/300\n7/7 [==============================] - ETA: 0s - loss: 0.0790 - accuracy: 0.9533\nEpoch 62: val_loss did not improve from 0.01077\n7/7 [==============================] - 1s 101ms/step - loss: 0.0790 - accuracy: 0.9533 - val_loss: 0.0199 - val_accuracy: 1.0000\nEpoch 63/300\n7/7 [==============================] - ETA: 0s - loss: 0.0784 - accuracy: 0.9579\nEpoch 63: val_loss improved from 0.01077 to 0.00994, saving model to ../working/model_vgg16.hdf5\n7/7 [==============================] - 1s 123ms/step - loss: 0.0784 - accuracy: 0.9579 - val_loss: 0.0099 - val_accuracy: 1.0000\nEpoch 64/300\n7/7 [==============================] - ETA: 0s - loss: 0.0759 - accuracy: 0.9579\nEpoch 64: val_loss improved from 0.00994 to 0.00943, saving model to ../working/model_vgg16.hdf5\n7/7 [==============================] - 1s 127ms/step - loss: 0.0759 - accuracy: 0.9579 - val_loss: 0.0094 - val_accuracy: 1.0000\nEpoch 65/300\n7/7 [==============================] - ETA: 0s - loss: 0.0762 - accuracy: 0.9533\nEpoch 65: val_loss did not improve from 0.00943\n7/7 [==============================] - 1s 99ms/step - loss: 0.0762 - accuracy: 0.9533 - val_loss: 0.0099 - val_accuracy: 1.0000\nEpoch 66/300\n7/7 [==============================] - ETA: 0s - loss: 0.5792 - accuracy: 0.8785\nEpoch 66: val_loss improved from 0.00943 to 0.00923, saving model to ../working/model_vgg16.hdf5\n7/7 [==============================] - 1s 126ms/step - loss: 0.5792 - accuracy: 0.8785 - val_loss: 0.0092 - val_accuracy: 1.0000\nEpoch 67/300\n7/7 [==============================] - ETA: 0s - loss: 0.0796 - accuracy: 0.9486\nEpoch 67: val_loss did not improve from 0.00923\n7/7 [==============================] - 1s 107ms/step - loss: 0.0796 - accuracy: 0.9486 - val_loss: 0.0132 - val_accuracy: 1.0000\nEpoch 68/300\n7/7 [==============================] - ETA: 0s - loss: 0.0772 - accuracy: 0.9486\nEpoch 68: val_loss improved from 0.00923 to 0.00831, saving model to ../working/model_vgg16.hdf5\n7/7 [==============================] - 1s 131ms/step - loss: 0.0772 - accuracy: 0.9486 - val_loss: 0.0083 - val_accuracy: 1.0000\nEpoch 69/300\n7/7 [==============================] - ETA: 0s - loss: 0.0739 - accuracy: 0.9579\nEpoch 69: val_loss improved from 0.00831 to 0.00680, saving model to ../working/model_vgg16.hdf5\n7/7 [==============================] - 1s 175ms/step - loss: 0.0739 - accuracy: 0.9579 - val_loss: 0.0068 - val_accuracy: 1.0000\nEpoch 70/300\n7/7 [==============================] - ETA: 0s - loss: 0.0719 - accuracy: 0.9673\nEpoch 70: val_loss improved from 0.00680 to 0.00468, saving model to ../working/model_vgg16.hdf5\n7/7 [==============================] - 1s 131ms/step - loss: 0.0719 - accuracy: 0.9673 - val_loss: 0.0047 - val_accuracy: 1.0000\nEpoch 71/300\n7/7 [==============================] - ETA: 0s - loss: 3.4716 - accuracy: 0.9579\nEpoch 71: val_loss did not improve from 0.00468\n7/7 [==============================] - 1s 104ms/step - loss: 3.4716 - accuracy: 0.9579 - val_loss: 0.0504 - val_accuracy: 0.9556\nEpoch 72/300\n7/7 [==============================] - ETA: 0s - loss: 57.8735 - accuracy: 0.8318\nEpoch 72: val_loss did not improve from 0.00468\n7/7 [==============================] - 1s 103ms/step - loss: 57.8735 - accuracy: 0.8318 - val_loss: 0.0137 - val_accuracy: 1.0000\nEpoch 73/300\n7/7 [==============================] - ETA: 0s - loss: 7.9298 - accuracy: 0.9439\nEpoch 73: val_loss did not improve from 0.00468\n7/7 [==============================] - 1s 100ms/step - loss: 7.9298 - accuracy: 0.9439 - val_loss: 0.0126 - val_accuracy: 1.0000\nEpoch 74/300\n7/7 [==============================] - ETA: 0s - loss: 2.6373 - accuracy: 0.9533\nEpoch 74: val_loss did not improve from 0.00468\n7/7 [==============================] - 1s 101ms/step - loss: 2.6373 - accuracy: 0.9533 - val_loss: 0.0127 - val_accuracy: 1.0000\nEpoch 75/300\n7/7 [==============================] - ETA: 0s - loss: 4.5988 - accuracy: 0.9579\nEpoch 75: val_loss did not improve from 0.00468\n7/7 [==============================] - 1s 99ms/step - loss: 4.5988 - accuracy: 0.9579 - val_loss: 0.0182 - val_accuracy: 1.0000\nEpoch 76/300\n7/7 [==============================] - ETA: 0s - loss: 10.6911 - accuracy: 0.9486\nEpoch 76: val_loss did not improve from 0.00468\n7/7 [==============================] - 1s 103ms/step - loss: 10.6911 - accuracy: 0.9486 - val_loss: 0.0126 - val_accuracy: 1.0000\nEpoch 77/300\n7/7 [==============================] - ETA: 0s - loss: 13.7075 - accuracy: 0.9533\nEpoch 77: val_loss did not improve from 0.00468\n7/7 [==============================] - 1s 106ms/step - loss: 13.7075 - accuracy: 0.9533 - val_loss: 0.0092 - val_accuracy: 1.0000\nEpoch 78/300\n7/7 [==============================] - ETA: 0s - loss: 1.1875 - accuracy: 0.9579\nEpoch 78: val_loss did not improve from 0.00468\n7/7 [==============================] - 1s 100ms/step - loss: 1.1875 - accuracy: 0.9579 - val_loss: 0.0372 - val_accuracy: 0.9778\nEpoch 79/300\n7/7 [==============================] - ETA: 0s - loss: 8.5271 - accuracy: 0.8879 \nEpoch 79: val_loss did not improve from 0.00468\n7/7 [==============================] - 1s 102ms/step - loss: 8.5271 - accuracy: 0.8879 - val_loss: 0.0116 - val_accuracy: 1.0000\nEpoch 80/300\n7/7 [==============================] - ETA: 0s - loss: 4.9828 - accuracy: 0.9252\nEpoch 80: val_loss did not improve from 0.00468\n7/7 [==============================] - 1s 100ms/step - loss: 4.9828 - accuracy: 0.9252 - val_loss: 0.0052 - val_accuracy: 1.0000\nEpoch 80: early stopping\n","output_type":"stream"}]},{"cell_type":"code","source":"plt.figure(1, figsize = (15,8)) \n    \nplt.subplot(221)  \nplt.plot(history_vgg16.history['accuracy'])  \nplt.plot(history_vgg16.history['val_accuracy'])  \nplt.title('model accuracy')  \nplt.ylabel('accuracy')  \nplt.xlabel('epoch')  \nplt.legend(['train', 'valid']) \n    \nplt.subplot(222)  \nplt.plot(history_vgg16.history['loss'])  \nplt.plot(history_vgg16.history['val_loss'])  \nplt.title('model loss')  \nplt.ylabel('loss')  \nplt.xlabel('epoch')  \nplt.legend(['train', 'valid']) \n\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2023-07-16T06:30:27.278212Z","iopub.execute_input":"2023-07-16T06:30:27.278714Z","iopub.status.idle":"2023-07-16T06:30:27.839136Z","shell.execute_reply.started":"2023-07-16T06:30:27.278675Z","shell.execute_reply":"2023-07-16T06:30:27.838181Z"},"trusted":true},"execution_count":14,"outputs":[{"output_type":"display_data","data":{"text/plain":"<Figure size 1500x800 with 2 Axes>","image/png":"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"},"metadata":{}}]},{"cell_type":"code","source":"mymodel = tf.keras.models.load_model('../working/model_vgg16.hdf5')\n\nmymodel.evaluate(test_generator, verbose = 1)","metadata":{"execution":{"iopub.status.busy":"2023-07-16T06:30:27.840142Z","iopub.execute_input":"2023-07-16T06:30:27.840543Z","iopub.status.idle":"2023-07-16T06:30:31.127625Z","shell.execute_reply.started":"2023-07-16T06:30:27.840509Z","shell.execute_reply":"2023-07-16T06:30:31.126536Z"},"trusted":true},"execution_count":15,"outputs":[{"name":"stdout","text":"2/2 [==============================] - 2s 1s/step - loss: 0.0649 - accuracy: 0.9792\n","output_type":"stream"},{"execution_count":15,"output_type":"execute_result","data":{"text/plain":"[0.06493715196847916, 0.9791666865348816]"},"metadata":{}}]},{"cell_type":"markdown","source":"## VGG19","metadata":{}},{"cell_type":"code","source":"basemodel_vgg19 = tf.keras.applications.vgg19.VGG19(include_top=False, weights=\"imagenet\", input_shape=(224,224, 3), classes=2)\nbasemodel_vgg19.trainable = False\nmodel_vgg19 = tf.keras.Sequential([\n    basemodel_vgg19,\n    tf.keras.layers.GlobalAveragePooling2D(),\n    keras.layers.Dense(256, activation='relu', kernel_initializer='he_normal'),\n    keras.layers.Dense(128, activation='relu', kernel_initializer='he_normal'),    \n    keras.layers.Dense(64, activation='relu', kernel_initializer='he_normal'),\n    keras.layers.Dense(32, activation='relu', kernel_initializer='he_normal'),    \n    keras.layers.Dense(16, activation='relu', kernel_initializer='he_normal'),\n    keras.layers.Dense(2, activation='softmax')\n])\n\nmodel_vgg19.summary()","metadata":{"execution":{"iopub.status.busy":"2023-07-16T06:30:31.129130Z","iopub.execute_input":"2023-07-16T06:30:31.129483Z","iopub.status.idle":"2023-07-16T06:30:32.228494Z","shell.execute_reply.started":"2023-07-16T06:30:31.129450Z","shell.execute_reply":"2023-07-16T06:30:32.227711Z"},"trusted":true},"execution_count":16,"outputs":[{"name":"stdout","text":"Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/vgg19/vgg19_weights_tf_dim_ordering_tf_kernels_notop.h5\n80134624/80134624 [==============================] - 0s 0us/step\nModel: \"sequential_1\"\n_________________________________________________________________\n Layer (type)                Output Shape              Param #   \n=================================================================\n vgg19 (Functional)          (None, 7, 7, 512)         20024384  \n                                                                 \n global_average_pooling2d_1   (None, 512)              0         \n (GlobalAveragePooling2D)                                        \n                                                                 \n dense_6 (Dense)             (None, 256)               131328    \n                                                                 \n dense_7 (Dense)             (None, 128)               32896     \n                                                                 \n dense_8 (Dense)             (None, 64)                8256      \n                                                                 \n dense_9 (Dense)             (None, 32)                2080      \n                                                                 \n dense_10 (Dense)            (None, 16)                528       \n                                                                 \n dense_11 (Dense)            (None, 2)                 34        \n                                                                 \n=================================================================\nTotal params: 20,199,506\nTrainable params: 175,122\nNon-trainable params: 20,024,384\n_________________________________________________________________\n","output_type":"stream"}]},{"cell_type":"code","source":"cb_earlystopper = tf.keras.callbacks.EarlyStopping(monitor='val_loss', patience=10, verbose=1)\ncb_checkpointer = tf.keras.callbacks.ModelCheckpoint(filepath='../working/model_vgg19.hdf5', monitor='val_loss', save_best_only=True, mode='auto', verbose=1)\n\nCALLBACKS = [cb_earlystopper, cb_checkpointer]","metadata":{"execution":{"iopub.status.busy":"2023-07-16T06:30:32.229956Z","iopub.execute_input":"2023-07-16T06:30:32.230311Z","iopub.status.idle":"2023-07-16T06:30:32.239779Z","shell.execute_reply.started":"2023-07-16T06:30:32.230277Z","shell.execute_reply":"2023-07-16T06:30:32.238687Z"},"trusted":true},"execution_count":17,"outputs":[]},{"cell_type":"code","source":"model_vgg19.compile(optimizer='rmsprop', loss='sparse_categorical_crossentropy', metrics='accuracy')\nhistory_vgg19 = model_vgg19.fit(\n    train_generator,\n    epochs=300,\n    validation_data=valid_generator,\n    callbacks=CALLBACKS\n)","metadata":{"execution":{"iopub.status.busy":"2023-07-16T06:30:32.241052Z","iopub.execute_input":"2023-07-16T06:30:32.241393Z","iopub.status.idle":"2023-07-16T06:32:16.035001Z","shell.execute_reply.started":"2023-07-16T06:30:32.241361Z","shell.execute_reply":"2023-07-16T06:32:16.033908Z"},"trusted":true},"execution_count":18,"outputs":[{"name":"stdout","text":"Epoch 1/300\n7/7 [==============================] - ETA: 0s - loss: 23.6717 - accuracy: 0.6215\nEpoch 1: val_loss improved from inf to 0.60739, saving model to ../working/model_vgg19.hdf5\n7/7 [==============================] - 3s 199ms/step - loss: 23.6717 - accuracy: 0.6215 - val_loss: 0.6074 - val_accuracy: 0.4889\nEpoch 2/300\n7/7 [==============================] - ETA: 0s - loss: 143.2774 - accuracy: 0.5888\nEpoch 2: val_loss improved from 0.60739 to 0.55187, saving model to ../working/model_vgg19.hdf5\n7/7 [==============================] - 1s 146ms/step - loss: 143.2774 - accuracy: 0.5888 - val_loss: 0.5519 - val_accuracy: 0.8222\nEpoch 3/300\n7/7 [==============================] - ETA: 0s - loss: 111.4110 - accuracy: 0.7056\nEpoch 3: val_loss improved from 0.55187 to 0.45768, saving model to ../working/model_vgg19.hdf5\n7/7 [==============================] - 1s 192ms/step - loss: 111.4110 - accuracy: 0.7056 - val_loss: 0.4577 - val_accuracy: 0.9333\nEpoch 4/300\n7/7 [==============================] - ETA: 0s - loss: 96.9035 - accuracy: 0.7664\nEpoch 4: val_loss did not improve from 0.45768\n7/7 [==============================] - 1s 130ms/step - loss: 96.9035 - accuracy: 0.7664 - val_loss: 0.4752 - val_accuracy: 0.7556\nEpoch 5/300\n7/7 [==============================] - ETA: 0s - loss: 124.0187 - accuracy: 0.7290\nEpoch 5: val_loss improved from 0.45768 to 0.38159, saving model to ../working/model_vgg19.hdf5\n7/7 [==============================] - 1s 148ms/step - loss: 124.0187 - accuracy: 0.7290 - val_loss: 0.3816 - val_accuracy: 0.9556\nEpoch 6/300\n7/7 [==============================] - ETA: 0s - loss: 21.6731 - accuracy: 0.8271\nEpoch 6: val_loss improved from 0.38159 to 0.30665, saving model to ../working/model_vgg19.hdf5\n7/7 [==============================] - 1s 190ms/step - loss: 21.6731 - accuracy: 0.8271 - val_loss: 0.3066 - val_accuracy: 0.9556\nEpoch 7/300\n7/7 [==============================] - ETA: 0s - loss: 5.0229 - accuracy: 0.7570\nEpoch 7: val_loss did not improve from 0.30665\n7/7 [==============================] - 1s 130ms/step - loss: 5.0229 - accuracy: 0.7570 - val_loss: 0.3216 - val_accuracy: 0.9333\nEpoch 8/300\n7/7 [==============================] - ETA: 0s - loss: 336.7788 - accuracy: 0.7850\nEpoch 8: val_loss did not improve from 0.30665\n7/7 [==============================] - 1s 150ms/step - loss: 336.7788 - accuracy: 0.7850 - val_loss: 0.3462 - val_accuracy: 0.8000\nEpoch 9/300\n7/7 [==============================] - ETA: 0s - loss: 236.1078 - accuracy: 0.8178\nEpoch 9: val_loss improved from 0.30665 to 0.22810, saving model to ../working/model_vgg19.hdf5\n7/7 [==============================] - 1s 199ms/step - loss: 236.1078 - accuracy: 0.8178 - val_loss: 0.2281 - val_accuracy: 0.9556\nEpoch 10/300\n7/7 [==============================] - ETA: 0s - loss: 62.5900 - accuracy: 0.8458\nEpoch 10: val_loss did not improve from 0.22810\n7/7 [==============================] - 1s 143ms/step - loss: 62.5900 - accuracy: 0.8458 - val_loss: 0.4786 - val_accuracy: 0.7778\nEpoch 11/300\n7/7 [==============================] - ETA: 0s - loss: 188.8735 - accuracy: 0.8131\nEpoch 11: val_loss improved from 0.22810 to 0.22622, saving model to ../working/model_vgg19.hdf5\n7/7 [==============================] - 2s 250ms/step - loss: 188.8735 - accuracy: 0.8131 - val_loss: 0.2262 - val_accuracy: 0.9333\nEpoch 12/300\n7/7 [==============================] - ETA: 0s - loss: 0.8675 - accuracy: 0.8738\nEpoch 12: val_loss did not improve from 0.22622\n7/7 [==============================] - 1s 134ms/step - loss: 0.8675 - accuracy: 0.8738 - val_loss: 0.2265 - val_accuracy: 0.9333\nEpoch 13/300\n7/7 [==============================] - ETA: 0s - loss: 94.1165 - accuracy: 0.7710 \nEpoch 13: val_loss improved from 0.22622 to 0.19357, saving model to ../working/model_vgg19.hdf5\n7/7 [==============================] - 1s 150ms/step - loss: 94.1165 - accuracy: 0.7710 - val_loss: 0.1936 - val_accuracy: 0.9556\nEpoch 14/300\n7/7 [==============================] - ETA: 0s - loss: 30.0210 - accuracy: 0.8505\nEpoch 14: val_loss did not improve from 0.19357\n7/7 [==============================] - 1s 115ms/step - loss: 30.0210 - accuracy: 0.8505 - val_loss: 0.2562 - val_accuracy: 0.8667\nEpoch 15/300\n7/7 [==============================] - ETA: 0s - loss: 85.8595 - accuracy: 0.9065 \nEpoch 15: val_loss did not improve from 0.19357\n7/7 [==============================] - 1s 120ms/step - loss: 85.8595 - accuracy: 0.9065 - val_loss: 0.2546 - val_accuracy: 0.8889\nEpoch 16/300\n7/7 [==============================] - ETA: 0s - loss: 63.0521 - accuracy: 0.8271\nEpoch 16: val_loss improved from 0.19357 to 0.18858, saving model to ../working/model_vgg19.hdf5\n7/7 [==============================] - 1s 154ms/step - loss: 63.0521 - accuracy: 0.8271 - val_loss: 0.1886 - val_accuracy: 0.9111\nEpoch 17/300\n7/7 [==============================] - ETA: 0s - loss: 59.6237 - accuracy: 0.9159\nEpoch 17: val_loss improved from 0.18858 to 0.16734, saving model to ../working/model_vgg19.hdf5\n7/7 [==============================] - 1s 147ms/step - loss: 59.6237 - accuracy: 0.9159 - val_loss: 0.1673 - val_accuracy: 0.9333\nEpoch 18/300\n7/7 [==============================] - ETA: 0s - loss: 24.6240 - accuracy: 0.8832\nEpoch 18: val_loss improved from 0.16734 to 0.15206, saving model to ../working/model_vgg19.hdf5\n7/7 [==============================] - 1s 144ms/step - loss: 24.6240 - accuracy: 0.8832 - val_loss: 0.1521 - val_accuracy: 0.9556\nEpoch 19/300\n7/7 [==============================] - ETA: 0s - loss: 73.7404 - accuracy: 0.8925\nEpoch 19: val_loss did not improve from 0.15206\n7/7 [==============================] - 1s 116ms/step - loss: 73.7404 - accuracy: 0.8925 - val_loss: 0.3421 - val_accuracy: 0.8444\nEpoch 20/300\n7/7 [==============================] - ETA: 0s - loss: 80.5112 - accuracy: 0.8598\nEpoch 20: val_loss did not improve from 0.15206\n7/7 [==============================] - 1s 113ms/step - loss: 80.5112 - accuracy: 0.8598 - val_loss: 0.3289 - val_accuracy: 0.8444\nEpoch 21/300\n7/7 [==============================] - ETA: 0s - loss: 8.5835 - accuracy: 0.8458 \nEpoch 21: val_loss did not improve from 0.15206\n7/7 [==============================] - 1s 115ms/step - loss: 8.5835 - accuracy: 0.8458 - val_loss: 0.1713 - val_accuracy: 0.9333\nEpoch 22/300\n7/7 [==============================] - ETA: 0s - loss: 190.5753 - accuracy: 0.8598\nEpoch 22: val_loss improved from 0.15206 to 0.09336, saving model to ../working/model_vgg19.hdf5\n7/7 [==============================] - 1s 161ms/step - loss: 190.5753 - accuracy: 0.8598 - val_loss: 0.0934 - val_accuracy: 1.0000\nEpoch 23/300\n7/7 [==============================] - ETA: 0s - loss: 2.8459 - accuracy: 0.8645\nEpoch 23: val_loss did not improve from 0.09336\n7/7 [==============================] - 1s 117ms/step - loss: 2.8459 - accuracy: 0.8645 - val_loss: 0.2188 - val_accuracy: 0.8667\nEpoch 24/300\n7/7 [==============================] - ETA: 0s - loss: 104.8624 - accuracy: 0.8598\nEpoch 24: val_loss did not improve from 0.09336\n7/7 [==============================] - 1s 115ms/step - loss: 104.8624 - accuracy: 0.8598 - val_loss: 0.1286 - val_accuracy: 0.9333\nEpoch 25/300\n7/7 [==============================] - ETA: 0s - loss: 30.7601 - accuracy: 0.8411\nEpoch 25: val_loss did not improve from 0.09336\n7/7 [==============================] - 1s 118ms/step - loss: 30.7601 - accuracy: 0.8411 - val_loss: 0.3748 - val_accuracy: 0.8222\nEpoch 26/300\n7/7 [==============================] - ETA: 0s - loss: 56.7929 - accuracy: 0.8785\nEpoch 26: val_loss improved from 0.09336 to 0.08022, saving model to ../working/model_vgg19.hdf5\n7/7 [==============================] - 1s 150ms/step - loss: 56.7929 - accuracy: 0.8785 - val_loss: 0.0802 - val_accuracy: 0.9778\nEpoch 27/300\n7/7 [==============================] - ETA: 0s - loss: 13.5150 - accuracy: 0.9112\nEpoch 27: val_loss did not improve from 0.08022\n7/7 [==============================] - 1s 114ms/step - loss: 13.5150 - accuracy: 0.9112 - val_loss: 0.0876 - val_accuracy: 0.9778\nEpoch 28/300\n7/7 [==============================] - ETA: 0s - loss: 56.2075 - accuracy: 0.8598\nEpoch 28: val_loss did not improve from 0.08022\n7/7 [==============================] - 1s 116ms/step - loss: 56.2075 - accuracy: 0.8598 - val_loss: 0.2851 - val_accuracy: 0.8667\nEpoch 29/300\n7/7 [==============================] - ETA: 0s - loss: 117.8332 - accuracy: 0.8411\nEpoch 29: val_loss improved from 0.08022 to 0.07021, saving model to ../working/model_vgg19.hdf5\n7/7 [==============================] - 1s 144ms/step - loss: 117.8332 - accuracy: 0.8411 - val_loss: 0.0702 - val_accuracy: 0.9778\nEpoch 30/300\n7/7 [==============================] - ETA: 0s - loss: 31.5863 - accuracy: 0.8785\nEpoch 30: val_loss did not improve from 0.07021\n7/7 [==============================] - 1s 114ms/step - loss: 31.5863 - accuracy: 0.8785 - val_loss: 0.4999 - val_accuracy: 0.7333\nEpoch 31/300\n7/7 [==============================] - ETA: 0s - loss: 16.5217 - accuracy: 0.8925\nEpoch 31: val_loss improved from 0.07021 to 0.06671, saving model to ../working/model_vgg19.hdf5\n7/7 [==============================] - 1s 147ms/step - loss: 16.5217 - accuracy: 0.8925 - val_loss: 0.0667 - val_accuracy: 0.9778\nEpoch 32/300\n7/7 [==============================] - ETA: 0s - loss: 21.3378 - accuracy: 0.9112\nEpoch 32: val_loss improved from 0.06671 to 0.06452, saving model to ../working/model_vgg19.hdf5\n7/7 [==============================] - 1s 148ms/step - loss: 21.3378 - accuracy: 0.9112 - val_loss: 0.0645 - val_accuracy: 0.9778\nEpoch 33/300\n7/7 [==============================] - ETA: 0s - loss: 38.7204 - accuracy: 0.8925\nEpoch 33: val_loss improved from 0.06452 to 0.05908, saving model to ../working/model_vgg19.hdf5\n7/7 [==============================] - 1s 150ms/step - loss: 38.7204 - accuracy: 0.8925 - val_loss: 0.0591 - val_accuracy: 1.0000\nEpoch 34/300\n7/7 [==============================] - ETA: 0s - loss: 0.8732 - accuracy: 0.9206\nEpoch 34: val_loss did not improve from 0.05908\n7/7 [==============================] - 1s 114ms/step - loss: 0.8732 - accuracy: 0.9206 - val_loss: 0.1196 - val_accuracy: 0.9333\nEpoch 35/300\n7/7 [==============================] - ETA: 0s - loss: 24.3787 - accuracy: 0.9159\nEpoch 35: val_loss did not improve from 0.05908\n7/7 [==============================] - 1s 139ms/step - loss: 24.3787 - accuracy: 0.9159 - val_loss: 0.0621 - val_accuracy: 0.9778\nEpoch 36/300\n7/7 [==============================] - ETA: 0s - loss: 17.8380 - accuracy: 0.8785\nEpoch 36: val_loss improved from 0.05908 to 0.04994, saving model to ../working/model_vgg19.hdf5\n7/7 [==============================] - 1s 150ms/step - loss: 17.8380 - accuracy: 0.8785 - val_loss: 0.0499 - val_accuracy: 1.0000\nEpoch 37/300\n7/7 [==============================] - ETA: 0s - loss: 32.4703 - accuracy: 0.9252\nEpoch 37: val_loss did not improve from 0.04994\n7/7 [==============================] - 1s 119ms/step - loss: 32.4703 - accuracy: 0.9252 - val_loss: 0.0634 - val_accuracy: 0.9778\nEpoch 38/300\n7/7 [==============================] - ETA: 0s - loss: 23.8727 - accuracy: 0.9206\nEpoch 38: val_loss did not improve from 0.04994\n7/7 [==============================] - 1s 112ms/step - loss: 23.8727 - accuracy: 0.9206 - val_loss: 0.0552 - val_accuracy: 0.9778\nEpoch 39/300\n7/7 [==============================] - ETA: 0s - loss: 3.2986 - accuracy: 0.9486\nEpoch 39: val_loss did not improve from 0.04994\n7/7 [==============================] - 1s 117ms/step - loss: 3.2986 - accuracy: 0.9486 - val_loss: 0.0537 - val_accuracy: 0.9778\nEpoch 40/300\n7/7 [==============================] - ETA: 0s - loss: 8.8137 - accuracy: 0.8832 \nEpoch 40: val_loss did not improve from 0.04994\n7/7 [==============================] - 1s 151ms/step - loss: 8.8137 - accuracy: 0.8832 - val_loss: 0.0525 - val_accuracy: 0.9778\nEpoch 41/300\n7/7 [==============================] - ETA: 0s - loss: 7.8287 - accuracy: 0.9206\nEpoch 41: val_loss did not improve from 0.04994\n7/7 [==============================] - 1s 123ms/step - loss: 7.8287 - accuracy: 0.9206 - val_loss: 0.0540 - val_accuracy: 0.9778\nEpoch 42/300\n7/7 [==============================] - ETA: 0s - loss: 1.0881 - accuracy: 0.9252\nEpoch 42: val_loss did not improve from 0.04994\n7/7 [==============================] - 1s 121ms/step - loss: 1.0881 - accuracy: 0.9252 - val_loss: 0.0580 - val_accuracy: 0.9778\nEpoch 43/300\n7/7 [==============================] - ETA: 0s - loss: 54.3212 - accuracy: 0.8738\nEpoch 43: val_loss improved from 0.04994 to 0.03751, saving model to ../working/model_vgg19.hdf5\n7/7 [==============================] - 1s 147ms/step - loss: 54.3212 - accuracy: 0.8738 - val_loss: 0.0375 - val_accuracy: 1.0000\nEpoch 44/300\n7/7 [==============================] - ETA: 0s - loss: 3.0125 - accuracy: 0.9252\nEpoch 44: val_loss did not improve from 0.03751\n7/7 [==============================] - 1s 137ms/step - loss: 3.0125 - accuracy: 0.9252 - val_loss: 0.0505 - val_accuracy: 0.9778\nEpoch 45/300\n7/7 [==============================] - ETA: 0s - loss: 23.8675 - accuracy: 0.8598\nEpoch 45: val_loss did not improve from 0.03751\n7/7 [==============================] - 1s 115ms/step - loss: 23.8675 - accuracy: 0.8598 - val_loss: 0.1427 - val_accuracy: 0.9556\nEpoch 46/300\n7/7 [==============================] - ETA: 0s - loss: 23.0388 - accuracy: 0.9112\nEpoch 46: val_loss did not improve from 0.03751\n7/7 [==============================] - 1s 118ms/step - loss: 23.0388 - accuracy: 0.9112 - val_loss: 0.0471 - val_accuracy: 0.9778\nEpoch 47/300\n7/7 [==============================] - ETA: 0s - loss: 11.4666 - accuracy: 0.9299\nEpoch 47: val_loss did not improve from 0.03751\n7/7 [==============================] - 1s 114ms/step - loss: 11.4666 - accuracy: 0.9299 - val_loss: 0.1317 - val_accuracy: 0.9333\nEpoch 48/300\n7/7 [==============================] - ETA: 0s - loss: 27.3855 - accuracy: 0.8178\nEpoch 48: val_loss did not improve from 0.03751\n7/7 [==============================] - 1s 116ms/step - loss: 27.3855 - accuracy: 0.8178 - val_loss: 0.0421 - val_accuracy: 1.0000\nEpoch 49/300\n7/7 [==============================] - ETA: 0s - loss: 1.9589 - accuracy: 0.9439\nEpoch 49: val_loss improved from 0.03751 to 0.03619, saving model to ../working/model_vgg19.hdf5\n7/7 [==============================] - 1s 159ms/step - loss: 1.9589 - accuracy: 0.9439 - val_loss: 0.0362 - val_accuracy: 1.0000\nEpoch 50/300\n7/7 [==============================] - ETA: 0s - loss: 26.8690 - accuracy: 0.9439\nEpoch 50: val_loss improved from 0.03619 to 0.02855, saving model to ../working/model_vgg19.hdf5\n7/7 [==============================] - 1s 151ms/step - loss: 26.8690 - accuracy: 0.9439 - val_loss: 0.0285 - val_accuracy: 1.0000\nEpoch 51/300\n7/7 [==============================] - ETA: 0s - loss: 17.2484 - accuracy: 0.8738\nEpoch 51: val_loss did not improve from 0.02855\n7/7 [==============================] - 1s 114ms/step - loss: 17.2484 - accuracy: 0.8738 - val_loss: 0.1856 - val_accuracy: 0.9333\nEpoch 52/300\n7/7 [==============================] - ETA: 0s - loss: 23.8606 - accuracy: 0.9252\nEpoch 52: val_loss improved from 0.02855 to 0.02645, saving model to ../working/model_vgg19.hdf5\n7/7 [==============================] - 1s 147ms/step - loss: 23.8606 - accuracy: 0.9252 - val_loss: 0.0264 - val_accuracy: 1.0000\nEpoch 53/300\n7/7 [==============================] - ETA: 0s - loss: 8.0407 - accuracy: 0.9393\nEpoch 53: val_loss did not improve from 0.02645\n7/7 [==============================] - 1s 113ms/step - loss: 8.0407 - accuracy: 0.9393 - val_loss: 0.0395 - val_accuracy: 0.9778\nEpoch 54/300\n7/7 [==============================] - ETA: 0s - loss: 7.4825 - accuracy: 0.8925\nEpoch 54: val_loss did not improve from 0.02645\n7/7 [==============================] - 1s 122ms/step - loss: 7.4825 - accuracy: 0.8925 - val_loss: 0.0383 - val_accuracy: 0.9778\nEpoch 55/300\n7/7 [==============================] - ETA: 0s - loss: 4.6694 - accuracy: 0.9486\nEpoch 55: val_loss improved from 0.02645 to 0.02365, saving model to ../working/model_vgg19.hdf5\n7/7 [==============================] - 1s 150ms/step - loss: 4.6694 - accuracy: 0.9486 - val_loss: 0.0236 - val_accuracy: 1.0000\nEpoch 56/300\n7/7 [==============================] - ETA: 0s - loss: 12.0962 - accuracy: 0.9439\nEpoch 56: val_loss did not improve from 0.02365\n7/7 [==============================] - 1s 113ms/step - loss: 12.0962 - accuracy: 0.9439 - val_loss: 0.0295 - val_accuracy: 1.0000\nEpoch 57/300\n7/7 [==============================] - ETA: 0s - loss: 3.9480 - accuracy: 0.8925\nEpoch 57: val_loss did not improve from 0.02365\n7/7 [==============================] - 1s 114ms/step - loss: 3.9480 - accuracy: 0.8925 - val_loss: 0.0265 - val_accuracy: 1.0000\nEpoch 58/300\n7/7 [==============================] - ETA: 0s - loss: 11.1108 - accuracy: 0.9346\nEpoch 58: val_loss improved from 0.02365 to 0.02200, saving model to ../working/model_vgg19.hdf5\n7/7 [==============================] - 1s 149ms/step - loss: 11.1108 - accuracy: 0.9346 - val_loss: 0.0220 - val_accuracy: 1.0000\nEpoch 59/300\n7/7 [==============================] - ETA: 0s - loss: 7.1240 - accuracy: 0.9346\nEpoch 59: val_loss did not improve from 0.02200\n7/7 [==============================] - 1s 114ms/step - loss: 7.1240 - accuracy: 0.9346 - val_loss: 0.0399 - val_accuracy: 0.9778\nEpoch 60/300\n7/7 [==============================] - ETA: 0s - loss: 11.5706 - accuracy: 0.9393\nEpoch 60: val_loss did not improve from 0.02200\n7/7 [==============================] - 1s 115ms/step - loss: 11.5706 - accuracy: 0.9393 - val_loss: 0.0367 - val_accuracy: 0.9778\nEpoch 61/300\n7/7 [==============================] - ETA: 0s - loss: 8.6422 - accuracy: 0.8224 \nEpoch 61: val_loss did not improve from 0.02200\n7/7 [==============================] - 1s 115ms/step - loss: 8.6422 - accuracy: 0.8224 - val_loss: 0.0357 - val_accuracy: 1.0000\nEpoch 62/300\n7/7 [==============================] - ETA: 0s - loss: 21.4148 - accuracy: 0.9439\nEpoch 62: val_loss did not improve from 0.02200\n7/7 [==============================] - 1s 114ms/step - loss: 21.4148 - accuracy: 0.9439 - val_loss: 0.0428 - val_accuracy: 0.9778\nEpoch 63/300\n7/7 [==============================] - ETA: 0s - loss: 3.1027 - accuracy: 0.9486\nEpoch 63: val_loss improved from 0.02200 to 0.02158, saving model to ../working/model_vgg19.hdf5\n7/7 [==============================] - 1s 148ms/step - loss: 3.1027 - accuracy: 0.9486 - val_loss: 0.0216 - val_accuracy: 1.0000\nEpoch 64/300\n7/7 [==============================] - ETA: 0s - loss: 2.8118 - accuracy: 0.9579\nEpoch 64: val_loss improved from 0.02158 to 0.01984, saving model to ../working/model_vgg19.hdf5\n7/7 [==============================] - 1s 149ms/step - loss: 2.8118 - accuracy: 0.9579 - val_loss: 0.0198 - val_accuracy: 1.0000\nEpoch 65/300\n7/7 [==============================] - ETA: 0s - loss: 10.4497 - accuracy: 0.9439\nEpoch 65: val_loss did not improve from 0.01984\n7/7 [==============================] - 1s 114ms/step - loss: 10.4497 - accuracy: 0.9439 - val_loss: 0.0230 - val_accuracy: 1.0000\nEpoch 66/300\n7/7 [==============================] - ETA: 0s - loss: 4.6852 - accuracy: 0.9346\nEpoch 66: val_loss did not improve from 0.01984\n7/7 [==============================] - 1s 115ms/step - loss: 4.6852 - accuracy: 0.9346 - val_loss: 0.0843 - val_accuracy: 0.9556\nEpoch 67/300\n7/7 [==============================] - ETA: 0s - loss: 3.8918 - accuracy: 0.9439\nEpoch 67: val_loss improved from 0.01984 to 0.01634, saving model to ../working/model_vgg19.hdf5\n7/7 [==============================] - 1s 146ms/step - loss: 3.8918 - accuracy: 0.9439 - val_loss: 0.0163 - val_accuracy: 1.0000\nEpoch 68/300\n7/7 [==============================] - ETA: 0s - loss: 16.2221 - accuracy: 0.9112\nEpoch 68: val_loss did not improve from 0.01634\n7/7 [==============================] - 1s 113ms/step - loss: 16.2221 - accuracy: 0.9112 - val_loss: 0.4314 - val_accuracy: 0.8667\nEpoch 69/300\n7/7 [==============================] - ETA: 0s - loss: 11.0949 - accuracy: 0.8832\nEpoch 69: val_loss did not improve from 0.01634\n7/7 [==============================] - 1s 136ms/step - loss: 11.0949 - accuracy: 0.8832 - val_loss: 0.0170 - val_accuracy: 1.0000\nEpoch 70/300\n7/7 [==============================] - ETA: 0s - loss: 5.6687 - accuracy: 0.9533\nEpoch 70: val_loss improved from 0.01634 to 0.01434, saving model to ../working/model_vgg19.hdf5\n7/7 [==============================] - 1s 179ms/step - loss: 5.6687 - accuracy: 0.9533 - val_loss: 0.0143 - val_accuracy: 1.0000\nEpoch 71/300\n7/7 [==============================] - ETA: 0s - loss: 0.8288 - accuracy: 0.9393\nEpoch 71: val_loss did not improve from 0.01434\n7/7 [==============================] - 1s 115ms/step - loss: 0.8288 - accuracy: 0.9393 - val_loss: 0.0156 - val_accuracy: 1.0000\nEpoch 72/300\n7/7 [==============================] - ETA: 0s - loss: 5.8897 - accuracy: 0.9486\nEpoch 72: val_loss did not improve from 0.01434\n7/7 [==============================] - 1s 113ms/step - loss: 5.8897 - accuracy: 0.9486 - val_loss: 0.0169 - val_accuracy: 1.0000\nEpoch 73/300\n7/7 [==============================] - ETA: 0s - loss: 19.0938 - accuracy: 0.9533\nEpoch 73: val_loss improved from 0.01434 to 0.01005, saving model to ../working/model_vgg19.hdf5\n7/7 [==============================] - 1s 156ms/step - loss: 19.0938 - accuracy: 0.9533 - val_loss: 0.0101 - val_accuracy: 1.0000\nEpoch 74/300\n7/7 [==============================] - ETA: 0s - loss: 10.3140 - accuracy: 0.8271\nEpoch 74: val_loss did not improve from 0.01005\n7/7 [==============================] - 1s 114ms/step - loss: 10.3140 - accuracy: 0.8271 - val_loss: 0.0393 - val_accuracy: 0.9778\nEpoch 75/300\n7/7 [==============================] - ETA: 0s - loss: 1.6131 - accuracy: 0.9579\nEpoch 75: val_loss did not improve from 0.01005\n7/7 [==============================] - 1s 114ms/step - loss: 1.6131 - accuracy: 0.9579 - val_loss: 0.0182 - val_accuracy: 1.0000\nEpoch 76/300\n7/7 [==============================] - ETA: 0s - loss: 10.1597 - accuracy: 0.9393\nEpoch 76: val_loss did not improve from 0.01005\n7/7 [==============================] - 1s 113ms/step - loss: 10.1597 - accuracy: 0.9393 - val_loss: 0.0156 - val_accuracy: 1.0000\nEpoch 77/300\n7/7 [==============================] - ETA: 0s - loss: 0.9238 - accuracy: 0.9393\nEpoch 77: val_loss did not improve from 0.01005\n7/7 [==============================] - 1s 116ms/step - loss: 0.9238 - accuracy: 0.9393 - val_loss: 0.0137 - val_accuracy: 1.0000\nEpoch 78/300\n7/7 [==============================] - ETA: 0s - loss: 3.2521 - accuracy: 0.9393\nEpoch 78: val_loss did not improve from 0.01005\n7/7 [==============================] - 1s 116ms/step - loss: 3.2521 - accuracy: 0.9393 - val_loss: 0.0122 - val_accuracy: 1.0000\nEpoch 79/300\n7/7 [==============================] - ETA: 0s - loss: 2.0657 - accuracy: 0.9579\nEpoch 79: val_loss did not improve from 0.01005\n7/7 [==============================] - 1s 113ms/step - loss: 2.0657 - accuracy: 0.9579 - val_loss: 0.0101 - val_accuracy: 1.0000\nEpoch 80/300\n7/7 [==============================] - ETA: 0s - loss: 2.4730 - accuracy: 0.9393\nEpoch 80: val_loss did not improve from 0.01005\n7/7 [==============================] - 1s 117ms/step - loss: 2.4730 - accuracy: 0.9393 - val_loss: 0.0314 - val_accuracy: 0.9778\nEpoch 81/300\n7/7 [==============================] - ETA: 0s - loss: 5.7706 - accuracy: 0.9439\nEpoch 81: val_loss improved from 0.01005 to 0.00812, saving model to ../working/model_vgg19.hdf5\n7/7 [==============================] - 1s 148ms/step - loss: 5.7706 - accuracy: 0.9439 - val_loss: 0.0081 - val_accuracy: 1.0000\nEpoch 82/300\n7/7 [==============================] - ETA: 0s - loss: 1.5464 - accuracy: 0.9673\nEpoch 82: val_loss did not improve from 0.00812\n7/7 [==============================] - 1s 118ms/step - loss: 1.5464 - accuracy: 0.9673 - val_loss: 0.0261 - val_accuracy: 0.9778\nEpoch 83/300\n7/7 [==============================] - ETA: 0s - loss: 4.7907 - accuracy: 0.9299\nEpoch 83: val_loss improved from 0.00812 to 0.00609, saving model to ../working/model_vgg19.hdf5\n7/7 [==============================] - 1s 148ms/step - loss: 4.7907 - accuracy: 0.9299 - val_loss: 0.0061 - val_accuracy: 1.0000\nEpoch 84/300\n7/7 [==============================] - ETA: 0s - loss: 3.4737 - accuracy: 0.9065\nEpoch 84: val_loss did not improve from 0.00609\n7/7 [==============================] - 1s 115ms/step - loss: 3.4737 - accuracy: 0.9065 - val_loss: 1.6109 - val_accuracy: 0.6444\nEpoch 85/300\n7/7 [==============================] - ETA: 0s - loss: 4.0097 - accuracy: 0.9112\nEpoch 85: val_loss did not improve from 0.00609\n7/7 [==============================] - 1s 116ms/step - loss: 4.0097 - accuracy: 0.9112 - val_loss: 0.0271 - val_accuracy: 1.0000\nEpoch 86/300\n7/7 [==============================] - ETA: 0s - loss: 4.5832 - accuracy: 0.9486\nEpoch 86: val_loss did not improve from 0.00609\n7/7 [==============================] - 1s 115ms/step - loss: 4.5832 - accuracy: 0.9486 - val_loss: 0.0233 - val_accuracy: 1.0000\nEpoch 87/300\n7/7 [==============================] - ETA: 0s - loss: 0.5619 - accuracy: 0.9579\nEpoch 87: val_loss did not improve from 0.00609\n7/7 [==============================] - 1s 116ms/step - loss: 0.5619 - accuracy: 0.9579 - val_loss: 0.0378 - val_accuracy: 0.9778\nEpoch 88/300\n7/7 [==============================] - ETA: 0s - loss: 0.8734 - accuracy: 0.9533\nEpoch 88: val_loss did not improve from 0.00609\n7/7 [==============================] - 1s 112ms/step - loss: 0.8734 - accuracy: 0.9533 - val_loss: 0.0216 - val_accuracy: 1.0000\nEpoch 89/300\n7/7 [==============================] - ETA: 0s - loss: 12.7563 - accuracy: 0.9393\nEpoch 89: val_loss did not improve from 0.00609\n7/7 [==============================] - 1s 115ms/step - loss: 12.7563 - accuracy: 0.9393 - val_loss: 0.0178 - val_accuracy: 1.0000\nEpoch 90/300\n7/7 [==============================] - ETA: 0s - loss: 3.3945 - accuracy: 0.9393\nEpoch 90: val_loss did not improve from 0.00609\n7/7 [==============================] - 1s 115ms/step - loss: 3.3945 - accuracy: 0.9393 - val_loss: 0.0128 - val_accuracy: 1.0000\nEpoch 91/300\n7/7 [==============================] - ETA: 0s - loss: 1.7123 - accuracy: 0.9439\nEpoch 91: val_loss did not improve from 0.00609\n7/7 [==============================] - 1s 121ms/step - loss: 1.7123 - accuracy: 0.9439 - val_loss: 0.0142 - val_accuracy: 1.0000\nEpoch 92/300\n7/7 [==============================] - ETA: 0s - loss: 4.1006 - accuracy: 0.9486\nEpoch 92: val_loss did not improve from 0.00609\n7/7 [==============================] - 1s 118ms/step - loss: 4.1006 - accuracy: 0.9486 - val_loss: 0.0090 - val_accuracy: 1.0000\nEpoch 93/300\n7/7 [==============================] - ETA: 0s - loss: 0.3359 - accuracy: 0.9486  \nEpoch 93: val_loss did not improve from 0.00609\n7/7 [==============================] - 1s 116ms/step - loss: 0.3359 - accuracy: 0.9486 - val_loss: 0.0081 - val_accuracy: 1.0000\nEpoch 93: early stopping\n","output_type":"stream"}]},{"cell_type":"code","source":"plt.figure(1, figsize = (15,8)) \n    \nplt.subplot(221)  \nplt.plot(history_vgg19.history['accuracy'])  \nplt.plot(history_vgg19.history['val_accuracy'])  \nplt.title('model accuracy')  \nplt.ylabel('accuracy')  \nplt.xlabel('epoch')  \nplt.legend(['train', 'valid']) \n    \nplt.subplot(222)  \nplt.plot(history_vgg19.history['loss'])  \nplt.plot(history_vgg19.history['val_loss'])  \nplt.title('model loss')  \nplt.ylabel('loss')  \nplt.xlabel('epoch')  \nplt.legend(['train', 'valid']) \n\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2023-07-16T06:32:16.036391Z","iopub.execute_input":"2023-07-16T06:32:16.036779Z","iopub.status.idle":"2023-07-16T06:32:16.533963Z","shell.execute_reply.started":"2023-07-16T06:32:16.036745Z","shell.execute_reply":"2023-07-16T06:32:16.530810Z"},"trusted":true},"execution_count":19,"outputs":[{"output_type":"display_data","data":{"text/plain":"<Figure size 1500x800 with 2 Axes>","image/png":"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"},"metadata":{}}]},{"cell_type":"code","source":"mymodel = tf.keras.models.load_model('../working/model_vgg19.hdf5')\n\nmymodel.evaluate(test_generator, verbose = 1)","metadata":{"execution":{"iopub.status.busy":"2023-07-16T06:32:16.535520Z","iopub.execute_input":"2023-07-16T06:32:16.535914Z","iopub.status.idle":"2023-07-16T06:32:17.637273Z","shell.execute_reply.started":"2023-07-16T06:32:16.535880Z","shell.execute_reply":"2023-07-16T06:32:17.636341Z"},"trusted":true},"execution_count":20,"outputs":[{"name":"stdout","text":"2/2 [==============================] - 0s 45ms/step - loss: 0.4005 - accuracy: 0.9167\n","output_type":"stream"},{"execution_count":20,"output_type":"execute_result","data":{"text/plain":"[0.40054166316986084, 0.9166666865348816]"},"metadata":{}}]},{"cell_type":"markdown","source":"## ResNet-50","metadata":{}},{"cell_type":"code","source":"basemodel_resnet50 = tf.keras.applications.resnet_v2.ResNet50V2(include_top=False, weights=\"imagenet\", input_shape=(224,224, 3), classes=2)\nbasemodel_resnet50.trainable = False\nmodel_resnet50 = tf.keras.Sequential([\n    basemodel_resnet50,\n    tf.keras.layers.GlobalAveragePooling2D(),\n    keras.layers.Dense(256, activation='relu', kernel_initializer='he_normal'),\n    keras.layers.Dense(128, activation='relu', kernel_initializer='he_normal'),    \n    keras.layers.Dense(64, activation='relu', kernel_initializer='he_normal'),\n    keras.layers.Dense(32, activation='relu', kernel_initializer='he_normal'),    \n    keras.layers.Dense(16, activation='relu', kernel_initializer='he_normal'),\n    keras.layers.Dense(2, activation='softmax')\n])\n\nmodel_resnet50.summary()","metadata":{"execution":{"iopub.status.busy":"2023-07-16T06:32:17.638641Z","iopub.execute_input":"2023-07-16T06:32:17.639106Z","iopub.status.idle":"2023-07-16T06:32:20.363658Z","shell.execute_reply.started":"2023-07-16T06:32:17.639069Z","shell.execute_reply":"2023-07-16T06:32:20.362692Z"},"trusted":true},"execution_count":21,"outputs":[{"name":"stdout","text":"Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/resnet/resnet50v2_weights_tf_dim_ordering_tf_kernels_notop.h5\n94668760/94668760 [==============================] - 0s 0us/step\nModel: \"sequential_2\"\n_________________________________________________________________\n Layer (type)                Output Shape              Param #   \n=================================================================\n resnet50v2 (Functional)     (None, 7, 7, 2048)        23564800  \n                                                                 \n global_average_pooling2d_2   (None, 2048)             0         \n (GlobalAveragePooling2D)                                        \n                                                                 \n dense_12 (Dense)            (None, 256)               524544    \n                                                                 \n dense_13 (Dense)            (None, 128)               32896     \n                                                                 \n dense_14 (Dense)            (None, 64)                8256      \n                                                                 \n dense_15 (Dense)            (None, 32)                2080      \n                                                                 \n dense_16 (Dense)            (None, 16)                528       \n                                                                 \n dense_17 (Dense)            (None, 2)                 34        \n                                                                 \n=================================================================\nTotal params: 24,133,138\nTrainable params: 568,338\nNon-trainable params: 23,564,800\n_________________________________________________________________\n","output_type":"stream"}]},{"cell_type":"code","source":"cb_earlystopper = tf.keras.callbacks.EarlyStopping(monitor='val_loss', patience=10, verbose=1)\ncb_checkpointer = tf.keras.callbacks.ModelCheckpoint(filepath='../working/model_resnet50.hdf5', monitor='val_loss', save_best_only=True, mode='auto', verbose=1)\n\nCALLBACKS = [cb_earlystopper, cb_checkpointer]","metadata":{"execution":{"iopub.status.busy":"2023-07-16T06:32:20.365348Z","iopub.execute_input":"2023-07-16T06:32:20.365723Z","iopub.status.idle":"2023-07-16T06:32:20.374689Z","shell.execute_reply.started":"2023-07-16T06:32:20.365688Z","shell.execute_reply":"2023-07-16T06:32:20.373614Z"},"trusted":true},"execution_count":22,"outputs":[]},{"cell_type":"code","source":"model_resnet50.compile(optimizer='rmsprop', loss='sparse_categorical_crossentropy', metrics='accuracy')\nhistory_resnet50 = model_resnet50.fit(\n    train_generator,\n    epochs=300,\n    validation_data=valid_generator,\n    callbacks=CALLBACKS\n)","metadata":{"execution":{"iopub.status.busy":"2023-07-16T06:32:20.376065Z","iopub.execute_input":"2023-07-16T06:32:20.376558Z","iopub.status.idle":"2023-07-16T06:34:22.670521Z","shell.execute_reply.started":"2023-07-16T06:32:20.376525Z","shell.execute_reply":"2023-07-16T06:34:22.669481Z"},"trusted":true},"execution_count":23,"outputs":[{"name":"stdout","text":"Epoch 1/300\n7/7 [==============================] - ETA: 0s - loss: 0.9516 - accuracy: 0.5561\nEpoch 1: val_loss improved from inf to 0.57555, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 9s 516ms/step - loss: 0.9516 - accuracy: 0.5561 - val_loss: 0.5756 - val_accuracy: 0.6444\nEpoch 2/300\n7/7 [==============================] - ETA: 0s - loss: 0.6211 - accuracy: 0.6776\nEpoch 2: val_loss improved from 0.57555 to 0.33749, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 170ms/step - loss: 0.6211 - accuracy: 0.6776 - val_loss: 0.3375 - val_accuracy: 0.9333\nEpoch 3/300\n7/7 [==============================] - ETA: 0s - loss: 0.5582 - accuracy: 0.7991\nEpoch 3: val_loss did not improve from 0.33749\n7/7 [==============================] - 1s 92ms/step - loss: 0.5582 - accuracy: 0.7991 - val_loss: 0.3714 - val_accuracy: 0.8444\nEpoch 4/300\n7/7 [==============================] - ETA: 0s - loss: 0.4025 - accuracy: 0.8084\nEpoch 4: val_loss improved from 0.33749 to 0.25629, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 175ms/step - loss: 0.4025 - accuracy: 0.8084 - val_loss: 0.2563 - val_accuracy: 0.9556\nEpoch 5/300\n7/7 [==============================] - ETA: 0s - loss: 0.3323 - accuracy: 0.8692\nEpoch 5: val_loss did not improve from 0.25629\n7/7 [==============================] - 1s 102ms/step - loss: 0.3323 - accuracy: 0.8692 - val_loss: 0.3328 - val_accuracy: 0.8889\nEpoch 6/300\n7/7 [==============================] - ETA: 0s - loss: 0.1914 - accuracy: 0.9439\nEpoch 6: val_loss improved from 0.25629 to 0.13923, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 212ms/step - loss: 0.1914 - accuracy: 0.9439 - val_loss: 0.1392 - val_accuracy: 0.9556\nEpoch 7/300\n7/7 [==============================] - ETA: 0s - loss: 0.1771 - accuracy: 0.9206\nEpoch 7: val_loss did not improve from 0.13923\n7/7 [==============================] - 1s 94ms/step - loss: 0.1771 - accuracy: 0.9206 - val_loss: 0.8768 - val_accuracy: 0.7778\nEpoch 8/300\n7/7 [==============================] - ETA: 0s - loss: 0.2309 - accuracy: 0.9112\nEpoch 8: val_loss improved from 0.13923 to 0.05759, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 168ms/step - loss: 0.2309 - accuracy: 0.9112 - val_loss: 0.0576 - val_accuracy: 0.9778\nEpoch 9/300\n7/7 [==============================] - ETA: 0s - loss: 0.1366 - accuracy: 0.9346\nEpoch 9: val_loss did not improve from 0.05759\n7/7 [==============================] - 1s 92ms/step - loss: 0.1366 - accuracy: 0.9346 - val_loss: 0.1315 - val_accuracy: 0.9111\nEpoch 10/300\n6/7 [========================>.....] - ETA: 0s - loss: 0.0626 - accuracy: 0.9792\nEpoch 10: val_loss improved from 0.05759 to 0.05442, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 163ms/step - loss: 0.0595 - accuracy: 0.9813 - val_loss: 0.0544 - val_accuracy: 0.9778\nEpoch 11/300\n7/7 [==============================] - ETA: 0s - loss: 0.0616 - accuracy: 0.9766\nEpoch 11: val_loss improved from 0.05442 to 0.04375, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 225ms/step - loss: 0.0616 - accuracy: 0.9766 - val_loss: 0.0437 - val_accuracy: 0.9778\nEpoch 12/300\n7/7 [==============================] - ETA: 0s - loss: 0.0079 - accuracy: 1.0000\nEpoch 12: val_loss did not improve from 0.04375\n7/7 [==============================] - 1s 101ms/step - loss: 0.0079 - accuracy: 1.0000 - val_loss: 0.0462 - val_accuracy: 0.9778\nEpoch 13/300\n7/7 [==============================] - ETA: 0s - loss: 0.0029 - accuracy: 1.0000\nEpoch 13: val_loss did not improve from 0.04375\n7/7 [==============================] - 1s 93ms/step - loss: 0.0029 - accuracy: 1.0000 - val_loss: 0.0475 - val_accuracy: 0.9778\nEpoch 14/300\n7/7 [==============================] - ETA: 0s - loss: 0.0012 - accuracy: 1.0000  \nEpoch 14: val_loss did not improve from 0.04375\n7/7 [==============================] - 1s 93ms/step - loss: 0.0012 - accuracy: 1.0000 - val_loss: 0.0441 - val_accuracy: 0.9556\nEpoch 15/300\n7/7 [==============================] - ETA: 0s - loss: 6.0195e-04 - accuracy: 1.0000\nEpoch 15: val_loss did not improve from 0.04375\n7/7 [==============================] - 1s 90ms/step - loss: 6.0195e-04 - accuracy: 1.0000 - val_loss: 0.0460 - val_accuracy: 0.9778\nEpoch 16/300\n7/7 [==============================] - ETA: 0s - loss: 4.0715e-04 - accuracy: 1.0000\nEpoch 16: val_loss improved from 0.04375 to 0.03962, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 176ms/step - loss: 4.0715e-04 - accuracy: 1.0000 - val_loss: 0.0396 - val_accuracy: 0.9556\nEpoch 17/300\n7/7 [==============================] - ETA: 0s - loss: 2.5693e-04 - accuracy: 1.0000\nEpoch 17: val_loss improved from 0.03962 to 0.03871, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 165ms/step - loss: 2.5693e-04 - accuracy: 1.0000 - val_loss: 0.0387 - val_accuracy: 0.9778\nEpoch 18/300\n7/7 [==============================] - ETA: 0s - loss: 1.8519e-04 - accuracy: 1.0000\nEpoch 18: val_loss improved from 0.03871 to 0.03642, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 173ms/step - loss: 1.8519e-04 - accuracy: 1.0000 - val_loss: 0.0364 - val_accuracy: 0.9778\nEpoch 19/300\n7/7 [==============================] - ETA: 0s - loss: 1.2769e-04 - accuracy: 1.0000\nEpoch 19: val_loss improved from 0.03642 to 0.03482, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 166ms/step - loss: 1.2769e-04 - accuracy: 1.0000 - val_loss: 0.0348 - val_accuracy: 0.9556\nEpoch 20/300\n7/7 [==============================] - ETA: 0s - loss: 9.7573e-05 - accuracy: 1.0000\nEpoch 20: val_loss did not improve from 0.03482\n7/7 [==============================] - 1s 97ms/step - loss: 9.7573e-05 - accuracy: 1.0000 - val_loss: 0.0382 - val_accuracy: 0.9778\nEpoch 21/300\n7/7 [==============================] - ETA: 0s - loss: 7.1271e-05 - accuracy: 1.0000\nEpoch 21: val_loss improved from 0.03482 to 0.03289, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 2s 259ms/step - loss: 7.1271e-05 - accuracy: 1.0000 - val_loss: 0.0329 - val_accuracy: 0.9778\nEpoch 22/300\n7/7 [==============================] - ETA: 0s - loss: 5.0214e-05 - accuracy: 1.0000\nEpoch 22: val_loss improved from 0.03289 to 0.03087, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 170ms/step - loss: 5.0214e-05 - accuracy: 1.0000 - val_loss: 0.0309 - val_accuracy: 0.9778\nEpoch 23/300\n7/7 [==============================] - ETA: 0s - loss: 3.9480e-05 - accuracy: 1.0000\nEpoch 23: val_loss did not improve from 0.03087\n7/7 [==============================] - 1s 96ms/step - loss: 3.9480e-05 - accuracy: 1.0000 - val_loss: 0.0311 - val_accuracy: 0.9778\nEpoch 24/300\n7/7 [==============================] - ETA: 0s - loss: 2.9904e-05 - accuracy: 1.0000\nEpoch 24: val_loss improved from 0.03087 to 0.02943, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 168ms/step - loss: 2.9904e-05 - accuracy: 1.0000 - val_loss: 0.0294 - val_accuracy: 0.9778\nEpoch 25/300\n7/7 [==============================] - ETA: 0s - loss: 2.3467e-05 - accuracy: 1.0000\nEpoch 25: val_loss improved from 0.02943 to 0.02820, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 2s 235ms/step - loss: 2.3467e-05 - accuracy: 1.0000 - val_loss: 0.0282 - val_accuracy: 1.0000\nEpoch 26/300\n7/7 [==============================] - ETA: 0s - loss: 1.9239e-05 - accuracy: 1.0000\nEpoch 26: val_loss did not improve from 0.02820\n7/7 [==============================] - 1s 92ms/step - loss: 1.9239e-05 - accuracy: 1.0000 - val_loss: 0.0282 - val_accuracy: 0.9778\nEpoch 27/300\n7/7 [==============================] - ETA: 0s - loss: 1.6295e-05 - accuracy: 1.0000\nEpoch 27: val_loss improved from 0.02820 to 0.02702, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 164ms/step - loss: 1.6295e-05 - accuracy: 1.0000 - val_loss: 0.0270 - val_accuracy: 0.9778\nEpoch 28/300\n7/7 [==============================] - ETA: 0s - loss: 1.3798e-05 - accuracy: 1.0000\nEpoch 28: val_loss improved from 0.02702 to 0.02644, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 167ms/step - loss: 1.3798e-05 - accuracy: 1.0000 - val_loss: 0.0264 - val_accuracy: 0.9778\nEpoch 29/300\n7/7 [==============================] - ETA: 0s - loss: 1.1768e-05 - accuracy: 1.0000\nEpoch 29: val_loss improved from 0.02644 to 0.02576, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 219ms/step - loss: 1.1768e-05 - accuracy: 1.0000 - val_loss: 0.0258 - val_accuracy: 0.9778\nEpoch 30/300\n7/7 [==============================] - ETA: 0s - loss: 1.0277e-05 - accuracy: 1.0000\nEpoch 30: val_loss improved from 0.02576 to 0.02536, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 166ms/step - loss: 1.0277e-05 - accuracy: 1.0000 - val_loss: 0.0254 - val_accuracy: 0.9778\nEpoch 31/300\n7/7 [==============================] - ETA: 0s - loss: 9.1410e-06 - accuracy: 1.0000\nEpoch 31: val_loss improved from 0.02536 to 0.02486, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 166ms/step - loss: 9.1410e-06 - accuracy: 1.0000 - val_loss: 0.0249 - val_accuracy: 0.9778\nEpoch 32/300\n7/7 [==============================] - ETA: 0s - loss: 8.1283e-06 - accuracy: 1.0000\nEpoch 32: val_loss improved from 0.02486 to 0.02477, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 176ms/step - loss: 8.1283e-06 - accuracy: 1.0000 - val_loss: 0.0248 - val_accuracy: 0.9778\nEpoch 33/300\n6/7 [========================>.....] - ETA: 0s - loss: 7.6168e-06 - accuracy: 1.0000\nEpoch 33: val_loss improved from 0.02477 to 0.02431, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 2s 307ms/step - loss: 7.4092e-06 - accuracy: 1.0000 - val_loss: 0.0243 - val_accuracy: 1.0000\nEpoch 34/300\n7/7 [==============================] - ETA: 0s - loss: 6.7346e-06 - accuracy: 1.0000\nEpoch 34: val_loss did not improve from 0.02431\n7/7 [==============================] - 1s 97ms/step - loss: 6.7346e-06 - accuracy: 1.0000 - val_loss: 0.0245 - val_accuracy: 0.9778\nEpoch 35/300\n7/7 [==============================] - ETA: 0s - loss: 6.2684e-06 - accuracy: 1.0000\nEpoch 35: val_loss improved from 0.02431 to 0.02410, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 167ms/step - loss: 6.2684e-06 - accuracy: 1.0000 - val_loss: 0.0241 - val_accuracy: 0.9778\nEpoch 36/300\n7/7 [==============================] - ETA: 0s - loss: 5.7793e-06 - accuracy: 1.0000\nEpoch 36: val_loss improved from 0.02410 to 0.02375, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 170ms/step - loss: 5.7793e-06 - accuracy: 1.0000 - val_loss: 0.0238 - val_accuracy: 1.0000\nEpoch 37/300\n7/7 [==============================] - ETA: 0s - loss: 5.3682e-06 - accuracy: 1.0000\nEpoch 37: val_loss improved from 0.02375 to 0.02351, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 178ms/step - loss: 5.3682e-06 - accuracy: 1.0000 - val_loss: 0.0235 - val_accuracy: 1.0000\nEpoch 38/300\n7/7 [==============================] - ETA: 0s - loss: 5.0145e-06 - accuracy: 1.0000\nEpoch 38: val_loss did not improve from 0.02351\n7/7 [==============================] - 1s 96ms/step - loss: 5.0145e-06 - accuracy: 1.0000 - val_loss: 0.0236 - val_accuracy: 1.0000\nEpoch 39/300\n7/7 [==============================] - ETA: 0s - loss: 4.6781e-06 - accuracy: 1.0000\nEpoch 39: val_loss improved from 0.02351 to 0.02342, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 172ms/step - loss: 4.6781e-06 - accuracy: 1.0000 - val_loss: 0.0234 - val_accuracy: 1.0000\nEpoch 40/300\n7/7 [==============================] - ETA: 0s - loss: 4.4146e-06 - accuracy: 1.0000\nEpoch 40: val_loss improved from 0.02342 to 0.02336, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 168ms/step - loss: 4.4146e-06 - accuracy: 1.0000 - val_loss: 0.0234 - val_accuracy: 1.0000\nEpoch 41/300\n7/7 [==============================] - ETA: 0s - loss: 4.1656e-06 - accuracy: 1.0000\nEpoch 41: val_loss improved from 0.02336 to 0.02325, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 182ms/step - loss: 4.1656e-06 - accuracy: 1.0000 - val_loss: 0.0233 - val_accuracy: 1.0000\nEpoch 42/300\n7/7 [==============================] - ETA: 0s - loss: 3.9394e-06 - accuracy: 1.0000\nEpoch 42: val_loss improved from 0.02325 to 0.02323, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 2s 294ms/step - loss: 3.9394e-06 - accuracy: 1.0000 - val_loss: 0.0232 - val_accuracy: 1.0000\nEpoch 43/300\n7/7 [==============================] - ETA: 0s - loss: 3.7656e-06 - accuracy: 1.0000\nEpoch 43: val_loss improved from 0.02323 to 0.02307, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 166ms/step - loss: 3.7656e-06 - accuracy: 1.0000 - val_loss: 0.0231 - val_accuracy: 1.0000\nEpoch 44/300\n6/7 [========================>.....] - ETA: 0s - loss: 3.6321e-06 - accuracy: 1.0000\nEpoch 44: val_loss improved from 0.02307 to 0.02306, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 166ms/step - loss: 3.6030e-06 - accuracy: 1.0000 - val_loss: 0.0231 - val_accuracy: 1.0000\nEpoch 45/300\n7/7 [==============================] - ETA: 0s - loss: 3.4231e-06 - accuracy: 1.0000\nEpoch 45: val_loss improved from 0.02306 to 0.02305, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 173ms/step - loss: 3.4231e-06 - accuracy: 1.0000 - val_loss: 0.0230 - val_accuracy: 1.0000\nEpoch 46/300\n7/7 [==============================] - ETA: 0s - loss: 3.2537e-06 - accuracy: 1.0000\nEpoch 46: val_loss improved from 0.02305 to 0.02286, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 2s 320ms/step - loss: 3.2537e-06 - accuracy: 1.0000 - val_loss: 0.0229 - val_accuracy: 1.0000\nEpoch 47/300\n7/7 [==============================] - ETA: 0s - loss: 3.0922e-06 - accuracy: 1.0000\nEpoch 47: val_loss improved from 0.02286 to 0.02258, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 169ms/step - loss: 3.0922e-06 - accuracy: 1.0000 - val_loss: 0.0226 - val_accuracy: 1.0000\nEpoch 48/300\n7/7 [==============================] - ETA: 0s - loss: 2.9886e-06 - accuracy: 1.0000\nEpoch 48: val_loss did not improve from 0.02258\n7/7 [==============================] - 1s 95ms/step - loss: 2.9886e-06 - accuracy: 1.0000 - val_loss: 0.0227 - val_accuracy: 1.0000\nEpoch 49/300\n7/7 [==============================] - ETA: 0s - loss: 2.8476e-06 - accuracy: 1.0000\nEpoch 49: val_loss improved from 0.02258 to 0.02257, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 2s 231ms/step - loss: 2.8476e-06 - accuracy: 1.0000 - val_loss: 0.0226 - val_accuracy: 1.0000\nEpoch 50/300\n7/7 [==============================] - ETA: 0s - loss: 2.7635e-06 - accuracy: 1.0000\nEpoch 50: val_loss did not improve from 0.02257\n7/7 [==============================] - 1s 92ms/step - loss: 2.7635e-06 - accuracy: 1.0000 - val_loss: 0.0226 - val_accuracy: 1.0000\nEpoch 51/300\n7/7 [==============================] - ETA: 0s - loss: 2.6449e-06 - accuracy: 1.0000\nEpoch 51: val_loss improved from 0.02257 to 0.02252, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 168ms/step - loss: 2.6449e-06 - accuracy: 1.0000 - val_loss: 0.0225 - val_accuracy: 1.0000\nEpoch 52/300\n7/7 [==============================] - ETA: 0s - loss: 2.5446e-06 - accuracy: 1.0000\nEpoch 52: val_loss improved from 0.02252 to 0.02251, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 163ms/step - loss: 2.5446e-06 - accuracy: 1.0000 - val_loss: 0.0225 - val_accuracy: 1.0000\nEpoch 53/300\n7/7 [==============================] - ETA: 0s - loss: 2.4544e-06 - accuracy: 1.0000\nEpoch 53: val_loss did not improve from 0.02251\n7/7 [==============================] - 1s 92ms/step - loss: 2.4544e-06 - accuracy: 1.0000 - val_loss: 0.0225 - val_accuracy: 1.0000\nEpoch 54/300\n6/7 [========================>.....] - ETA: 0s - loss: 2.2631e-06 - accuracy: 1.0000\nEpoch 54: val_loss improved from 0.02251 to 0.02241, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 171ms/step - loss: 2.3686e-06 - accuracy: 1.0000 - val_loss: 0.0224 - val_accuracy: 1.0000\nEpoch 55/300\n7/7 [==============================] - ETA: 0s - loss: 2.3101e-06 - accuracy: 1.0000\nEpoch 55: val_loss improved from 0.02241 to 0.02240, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 168ms/step - loss: 2.3101e-06 - accuracy: 1.0000 - val_loss: 0.0224 - val_accuracy: 1.0000\nEpoch 56/300\n7/7 [==============================] - ETA: 0s - loss: 2.2159e-06 - accuracy: 1.0000\nEpoch 56: val_loss improved from 0.02240 to 0.02229, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 172ms/step - loss: 2.2159e-06 - accuracy: 1.0000 - val_loss: 0.0223 - val_accuracy: 1.0000\nEpoch 57/300\n7/7 [==============================] - ETA: 0s - loss: 2.1435e-06 - accuracy: 1.0000\nEpoch 57: val_loss did not improve from 0.02229\n7/7 [==============================] - 1s 92ms/step - loss: 2.1435e-06 - accuracy: 1.0000 - val_loss: 0.0223 - val_accuracy: 1.0000\nEpoch 58/300\n7/7 [==============================] - ETA: 0s - loss: 2.0739e-06 - accuracy: 1.0000\nEpoch 58: val_loss improved from 0.02229 to 0.02227, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 163ms/step - loss: 2.0739e-06 - accuracy: 1.0000 - val_loss: 0.0223 - val_accuracy: 1.0000\nEpoch 59/300\n7/7 [==============================] - ETA: 0s - loss: 2.0238e-06 - accuracy: 1.0000\nEpoch 59: val_loss improved from 0.02227 to 0.02216, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 2s 289ms/step - loss: 2.0238e-06 - accuracy: 1.0000 - val_loss: 0.0222 - val_accuracy: 1.0000\nEpoch 60/300\n7/7 [==============================] - ETA: 0s - loss: 1.9647e-06 - accuracy: 1.0000\nEpoch 60: val_loss did not improve from 0.02216\n7/7 [==============================] - 1s 90ms/step - loss: 1.9647e-06 - accuracy: 1.0000 - val_loss: 0.0222 - val_accuracy: 1.0000\nEpoch 61/300\n7/7 [==============================] - ETA: 0s - loss: 1.9124e-06 - accuracy: 1.0000\nEpoch 61: val_loss improved from 0.02216 to 0.02213, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 165ms/step - loss: 1.9124e-06 - accuracy: 1.0000 - val_loss: 0.0221 - val_accuracy: 1.0000\nEpoch 62/300\n7/7 [==============================] - ETA: 0s - loss: 1.8544e-06 - accuracy: 1.0000\nEpoch 62: val_loss improved from 0.02213 to 0.02210, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 168ms/step - loss: 1.8544e-06 - accuracy: 1.0000 - val_loss: 0.0221 - val_accuracy: 1.0000\nEpoch 63/300\n6/7 [========================>.....] - ETA: 0s - loss: 1.8310e-06 - accuracy: 1.0000\nEpoch 63: val_loss improved from 0.02210 to 0.02198, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 221ms/step - loss: 1.8099e-06 - accuracy: 1.0000 - val_loss: 0.0220 - val_accuracy: 1.0000\nEpoch 64/300\n7/7 [==============================] - ETA: 0s - loss: 1.7586e-06 - accuracy: 1.0000\nEpoch 64: val_loss did not improve from 0.02198\n7/7 [==============================] - 1s 101ms/step - loss: 1.7586e-06 - accuracy: 1.0000 - val_loss: 0.0220 - val_accuracy: 1.0000\nEpoch 65/300\n7/7 [==============================] - ETA: 0s - loss: 1.7235e-06 - accuracy: 1.0000\nEpoch 65: val_loss improved from 0.02198 to 0.02186, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 173ms/step - loss: 1.7235e-06 - accuracy: 1.0000 - val_loss: 0.0219 - val_accuracy: 1.0000\nEpoch 66/300\n6/7 [========================>.....] - ETA: 0s - loss: 1.6801e-06 - accuracy: 1.0000\nEpoch 66: val_loss improved from 0.02186 to 0.02180, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 167ms/step - loss: 1.6778e-06 - accuracy: 1.0000 - val_loss: 0.0218 - val_accuracy: 1.0000\nEpoch 67/300\n7/7 [==============================] - ETA: 0s - loss: 1.6349e-06 - accuracy: 1.0000\nEpoch 67: val_loss did not improve from 0.02180\n7/7 [==============================] - 1s 92ms/step - loss: 1.6349e-06 - accuracy: 1.0000 - val_loss: 0.0219 - val_accuracy: 1.0000\nEpoch 68/300\n7/7 [==============================] - ETA: 0s - loss: 1.5920e-06 - accuracy: 1.0000\nEpoch 68: val_loss did not improve from 0.02180\n7/7 [==============================] - 1s 93ms/step - loss: 1.5920e-06 - accuracy: 1.0000 - val_loss: 0.0218 - val_accuracy: 1.0000\nEpoch 69/300\n6/7 [========================>.....] - ETA: 0s - loss: 1.5274e-06 - accuracy: 1.0000\nEpoch 69: val_loss improved from 0.02180 to 0.02180, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 228ms/step - loss: 1.5547e-06 - accuracy: 1.0000 - val_loss: 0.0218 - val_accuracy: 1.0000\nEpoch 70/300\n6/7 [========================>.....] - ETA: 0s - loss: 1.3970e-06 - accuracy: 1.0000\nEpoch 70: val_loss did not improve from 0.02180\n7/7 [==============================] - 1s 94ms/step - loss: 1.5191e-06 - accuracy: 1.0000 - val_loss: 0.0219 - val_accuracy: 1.0000\nEpoch 71/300\n7/7 [==============================] - ETA: 0s - loss: 1.4868e-06 - accuracy: 1.0000\nEpoch 71: val_loss improved from 0.02180 to 0.02178, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 167ms/step - loss: 1.4868e-06 - accuracy: 1.0000 - val_loss: 0.0218 - val_accuracy: 1.0000\nEpoch 72/300\n6/7 [========================>.....] - ETA: 0s - loss: 1.4423e-06 - accuracy: 1.0000\nEpoch 72: val_loss improved from 0.02178 to 0.02162, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 162ms/step - loss: 1.4533e-06 - accuracy: 1.0000 - val_loss: 0.0216 - val_accuracy: 1.0000\nEpoch 73/300\n7/7 [==============================] - ETA: 0s - loss: 1.4182e-06 - accuracy: 1.0000\nEpoch 73: val_loss did not improve from 0.02162\n7/7 [==============================] - 1s 105ms/step - loss: 1.4182e-06 - accuracy: 1.0000 - val_loss: 0.0216 - val_accuracy: 1.0000\nEpoch 74/300\n7/7 [==============================] - ETA: 0s - loss: 1.3859e-06 - accuracy: 1.0000\nEpoch 74: val_loss improved from 0.02162 to 0.02158, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 172ms/step - loss: 1.3859e-06 - accuracy: 1.0000 - val_loss: 0.0216 - val_accuracy: 1.0000\nEpoch 75/300\n7/7 [==============================] - ETA: 0s - loss: 1.3598e-06 - accuracy: 1.0000\nEpoch 75: val_loss did not improve from 0.02158\n7/7 [==============================] - 1s 99ms/step - loss: 1.3598e-06 - accuracy: 1.0000 - val_loss: 0.0216 - val_accuracy: 1.0000\nEpoch 76/300\n7/7 [==============================] - ETA: 0s - loss: 1.3319e-06 - accuracy: 1.0000\nEpoch 76: val_loss improved from 0.02158 to 0.02156, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 213ms/step - loss: 1.3319e-06 - accuracy: 1.0000 - val_loss: 0.0216 - val_accuracy: 1.0000\nEpoch 77/300\n7/7 [==============================] - ETA: 0s - loss: 1.3057e-06 - accuracy: 1.0000\nEpoch 77: val_loss improved from 0.02156 to 0.02150, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 166ms/step - loss: 1.3057e-06 - accuracy: 1.0000 - val_loss: 0.0215 - val_accuracy: 1.0000\nEpoch 78/300\n7/7 [==============================] - ETA: 0s - loss: 1.2823e-06 - accuracy: 1.0000\nEpoch 78: val_loss improved from 0.02150 to 0.02146, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 167ms/step - loss: 1.2823e-06 - accuracy: 1.0000 - val_loss: 0.0215 - val_accuracy: 1.0000\nEpoch 79/300\n7/7 [==============================] - ETA: 0s - loss: 1.2511e-06 - accuracy: 1.0000\nEpoch 79: val_loss improved from 0.02146 to 0.02140, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 177ms/step - loss: 1.2511e-06 - accuracy: 1.0000 - val_loss: 0.0214 - val_accuracy: 1.0000\nEpoch 80/300\n6/7 [========================>.....] - ETA: 0s - loss: 1.2889e-06 - accuracy: 1.0000\nEpoch 80: val_loss did not improve from 0.02140\n7/7 [==============================] - 1s 93ms/step - loss: 1.2294e-06 - accuracy: 1.0000 - val_loss: 0.0214 - val_accuracy: 1.0000\nEpoch 81/300\n7/7 [==============================] - ETA: 0s - loss: 1.2088e-06 - accuracy: 1.0000\nEpoch 81: val_loss did not improve from 0.02140\n7/7 [==============================] - 1s 88ms/step - loss: 1.2088e-06 - accuracy: 1.0000 - val_loss: 0.0215 - val_accuracy: 1.0000\nEpoch 82/300\n7/7 [==============================] - ETA: 0s - loss: 1.1832e-06 - accuracy: 1.0000\nEpoch 82: val_loss did not improve from 0.02140\n7/7 [==============================] - 1s 93ms/step - loss: 1.1832e-06 - accuracy: 1.0000 - val_loss: 0.0215 - val_accuracy: 1.0000\nEpoch 83/300\n7/7 [==============================] - ETA: 0s - loss: 1.1592e-06 - accuracy: 1.0000\nEpoch 83: val_loss did not improve from 0.02140\n7/7 [==============================] - 1s 96ms/step - loss: 1.1592e-06 - accuracy: 1.0000 - val_loss: 0.0214 - val_accuracy: 1.0000\nEpoch 84/300\n7/7 [==============================] - ETA: 0s - loss: 1.1414e-06 - accuracy: 1.0000\nEpoch 84: val_loss did not improve from 0.02140\n7/7 [==============================] - 1s 98ms/step - loss: 1.1414e-06 - accuracy: 1.0000 - val_loss: 0.0215 - val_accuracy: 1.0000\nEpoch 85/300\n7/7 [==============================] - ETA: 0s - loss: 1.1213e-06 - accuracy: 1.0000\nEpoch 85: val_loss did not improve from 0.02140\n7/7 [==============================] - 1s 92ms/step - loss: 1.1213e-06 - accuracy: 1.0000 - val_loss: 0.0215 - val_accuracy: 1.0000\nEpoch 86/300\n6/7 [========================>.....] - ETA: 0s - loss: 1.1350e-06 - accuracy: 1.0000\nEpoch 86: val_loss did not improve from 0.02140\n7/7 [==============================] - 1s 87ms/step - loss: 1.1063e-06 - accuracy: 1.0000 - val_loss: 0.0215 - val_accuracy: 1.0000\nEpoch 87/300\n7/7 [==============================] - ETA: 0s - loss: 1.0846e-06 - accuracy: 1.0000\nEpoch 87: val_loss did not improve from 0.02140\n7/7 [==============================] - 1s 92ms/step - loss: 1.0846e-06 - accuracy: 1.0000 - val_loss: 0.0214 - val_accuracy: 1.0000\nEpoch 88/300\n7/7 [==============================] - ETA: 0s - loss: 1.0701e-06 - accuracy: 1.0000\nEpoch 88: val_loss did not improve from 0.02140\n7/7 [==============================] - 1s 91ms/step - loss: 1.0701e-06 - accuracy: 1.0000 - val_loss: 0.0214 - val_accuracy: 1.0000\nEpoch 89/300\n7/7 [==============================] - ETA: 0s - loss: 1.0523e-06 - accuracy: 1.0000\nEpoch 89: val_loss improved from 0.02140 to 0.02136, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 167ms/step - loss: 1.0523e-06 - accuracy: 1.0000 - val_loss: 0.0214 - val_accuracy: 1.0000\nEpoch 90/300\n7/7 [==============================] - ETA: 0s - loss: 1.0328e-06 - accuracy: 1.0000\nEpoch 90: val_loss did not improve from 0.02136\n7/7 [==============================] - 1s 90ms/step - loss: 1.0328e-06 - accuracy: 1.0000 - val_loss: 0.0214 - val_accuracy: 1.0000\nEpoch 91/300\n7/7 [==============================] - ETA: 0s - loss: 1.0194e-06 - accuracy: 1.0000\nEpoch 91: val_loss did not improve from 0.02136\n7/7 [==============================] - 1s 92ms/step - loss: 1.0194e-06 - accuracy: 1.0000 - val_loss: 0.0214 - val_accuracy: 1.0000\nEpoch 92/300\n6/7 [========================>.....] - ETA: 0s - loss: 1.0201e-06 - accuracy: 1.0000\nEpoch 92: val_loss did not improve from 0.02136\n7/7 [==============================] - 1s 90ms/step - loss: 1.0027e-06 - accuracy: 1.0000 - val_loss: 0.0214 - val_accuracy: 1.0000\nEpoch 93/300\n7/7 [==============================] - ETA: 0s - loss: 9.8709e-07 - accuracy: 1.0000\nEpoch 93: val_loss did not improve from 0.02136\n7/7 [==============================] - 1s 92ms/step - loss: 9.8709e-07 - accuracy: 1.0000 - val_loss: 0.0214 - val_accuracy: 1.0000\nEpoch 94/300\n7/7 [==============================] - ETA: 0s - loss: 9.7094e-07 - accuracy: 1.0000\nEpoch 94: val_loss did not improve from 0.02136\n7/7 [==============================] - 1s 97ms/step - loss: 9.7094e-07 - accuracy: 1.0000 - val_loss: 0.0214 - val_accuracy: 1.0000\nEpoch 95/300\n6/7 [========================>.....] - ETA: 0s - loss: 9.2759e-07 - accuracy: 1.0000\nEpoch 95: val_loss improved from 0.02136 to 0.02133, saving model to ../working/model_resnet50.hdf5\n7/7 [==============================] - 1s 168ms/step - loss: 9.5479e-07 - accuracy: 1.0000 - val_loss: 0.0213 - val_accuracy: 1.0000\nEpoch 96/300\n7/7 [==============================] - ETA: 0s - loss: 9.4030e-07 - accuracy: 1.0000\nEpoch 96: val_loss did not improve from 0.02133\n7/7 [==============================] - 1s 98ms/step - loss: 9.4030e-07 - accuracy: 1.0000 - val_loss: 0.0214 - val_accuracy: 1.0000\nEpoch 97/300\n7/7 [==============================] - ETA: 0s - loss: 9.2860e-07 - accuracy: 1.0000\nEpoch 97: val_loss did not improve from 0.02133\n7/7 [==============================] - 1s 95ms/step - loss: 9.2860e-07 - accuracy: 1.0000 - val_loss: 0.0215 - val_accuracy: 1.0000\nEpoch 98/300\n7/7 [==============================] - ETA: 0s - loss: 9.1134e-07 - accuracy: 1.0000\nEpoch 98: val_loss did not improve from 0.02133\n7/7 [==============================] - 1s 94ms/step - loss: 9.1134e-07 - accuracy: 1.0000 - val_loss: 0.0214 - val_accuracy: 1.0000\nEpoch 99/300\n7/7 [==============================] - ETA: 0s - loss: 8.9852e-07 - accuracy: 1.0000\nEpoch 99: val_loss did not improve from 0.02133\n7/7 [==============================] - 1s 99ms/step - loss: 8.9852e-07 - accuracy: 1.0000 - val_loss: 0.0214 - val_accuracy: 1.0000\nEpoch 100/300\n7/7 [==============================] - ETA: 0s - loss: 8.8850e-07 - accuracy: 1.0000\nEpoch 100: val_loss did not improve from 0.02133\n7/7 [==============================] - 1s 92ms/step - loss: 8.8850e-07 - accuracy: 1.0000 - val_loss: 0.0214 - val_accuracy: 1.0000\nEpoch 101/300\n7/7 [==============================] - ETA: 0s - loss: 8.7401e-07 - accuracy: 1.0000\nEpoch 101: val_loss did not improve from 0.02133\n7/7 [==============================] - 1s 93ms/step - loss: 8.7401e-07 - accuracy: 1.0000 - val_loss: 0.0214 - val_accuracy: 1.0000\nEpoch 102/300\n7/7 [==============================] - ETA: 0s - loss: 8.6232e-07 - accuracy: 1.0000\nEpoch 102: val_loss did not improve from 0.02133\n7/7 [==============================] - 1s 91ms/step - loss: 8.6232e-07 - accuracy: 1.0000 - val_loss: 0.0214 - val_accuracy: 1.0000\nEpoch 103/300\n7/7 [==============================] - ETA: 0s - loss: 8.4839e-07 - accuracy: 1.0000\nEpoch 103: val_loss did not improve from 0.02133\n7/7 [==============================] - 1s 92ms/step - loss: 8.4839e-07 - accuracy: 1.0000 - val_loss: 0.0214 - val_accuracy: 1.0000\nEpoch 104/300\n6/7 [========================>.....] - ETA: 0s - loss: 7.4381e-07 - accuracy: 1.0000\nEpoch 104: val_loss did not improve from 0.02133\n7/7 [==============================] - 1s 90ms/step - loss: 8.3502e-07 - accuracy: 1.0000 - val_loss: 0.0214 - val_accuracy: 1.0000\nEpoch 105/300\n7/7 [==============================] - ETA: 0s - loss: 8.2666e-07 - accuracy: 1.0000\nEpoch 105: val_loss did not improve from 0.02133\n7/7 [==============================] - 1s 94ms/step - loss: 8.2666e-07 - accuracy: 1.0000 - val_loss: 0.0214 - val_accuracy: 1.0000\nEpoch 105: early stopping\n","output_type":"stream"}]},{"cell_type":"code","source":"plt.figure(1, figsize = (15,8)) \n    \nplt.subplot(221)  \nplt.plot(history_resnet50.history['accuracy'])  \nplt.plot(history_resnet50.history['val_accuracy'])  \nplt.title('model accuracy')  \nplt.ylabel('accuracy')  \nplt.xlabel('epoch')  \nplt.legend(['train', 'valid']) \n    \nplt.subplot(222)  \nplt.plot(history_resnet50.history['loss'])  \nplt.plot(history_resnet50.history['val_loss'])  \nplt.title('model loss')  \nplt.ylabel('loss')  \nplt.xlabel('epoch')  \nplt.legend(['train', 'valid']) \n\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2023-07-16T06:34:22.671986Z","iopub.execute_input":"2023-07-16T06:34:22.672380Z","iopub.status.idle":"2023-07-16T06:34:23.156106Z","shell.execute_reply.started":"2023-07-16T06:34:22.672343Z","shell.execute_reply":"2023-07-16T06:34:23.155190Z"},"trusted":true},"execution_count":24,"outputs":[{"output_type":"display_data","data":{"text/plain":"<Figure size 1500x800 with 2 Axes>","image/png":"iVBORw0KGgoAAAANSUhEUgAABNEAAAFrCAYAAAD/xR/KAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8pXeV/AAAACXBIWXMAAA9hAAAPYQGoP6dpAACKvklEQVR4nO3deXxU1d3H8e9kkkz2hARIAoR9F1kEpYDUHUTFpS64IoqtiMpWtSqtC/UR64LUKmirSG1RqXUpKpXGDVFEEQEVEESWsCSENfs6c58/7sxkJpkshFmyfN6v1zwzc+65d85c6OPhN7/fORbDMAwBAAAAAAAAqFVYqAcAAAAAAAAANHUE0QAAAAAAAIB6EEQDAAAAAAAA6kEQDQAAAAAAAKgHQTQAAAAAAACgHgTRAAAAAAAAgHoQRAMAAAAAAADqQRANAAAAAAAAqAdBNAAAAAAAAKAeBNEANBm7du2SxWLR4sWLj/vcTz/9VBaLRZ9++qnfxwUAAICmKRjzR+aZAFwIogEAAAAAAAD1IIgGAE1YSUmJDMMI9TAAAAAAoNUjiAbA7aGHHpLFYtF3332nK6+8UomJiUpOTtasWbNUWVmprVu36vzzz1d8fLy6du2qxx9/vMY1srKydP3116t9+/ay2Wzq16+fnnrqKTkcDq9++/fv11VXXaX4+HglJiZqwoQJysnJ8Tmub775RhdffLGSk5MVFRWlIUOG6F//+lejvuPBgwc1depU9e/fX3FxcWrfvr3OPvtsrVq1qkbfsrIyzZkzR/369VNUVJRSUlJ01llnafXq1e4+DodDf/nLXzR48GBFR0crKSlJv/jFL7Rs2TJ3H4vFooceeqjG9bt27apJkya53y9evFgWi0X/+9//dPPNN6tdu3aKiYlRWVmZtm/frptuukm9evVSTEyMOnbsqPHjx+v777+vcd1jx47pt7/9rbp37y6bzab27dvrggsu0I8//ijDMNSrVy+NHTu2xnmFhYVKTEzU7bfffpx3FQAAtFatYf5Ym2XLlmnEiBGKiYlRfHy8zjvvPH355ZdefQ4ePKjf/OY3ysjIkM1mU7t27TRq1Ch9+OGH7j7r16/XRRdd5P7+HTp00IUXXqi9e/f6dbwATlx4qAcAoOm56qqrdP311+vWW29VZmamHn/8cVVUVOjDDz/U1KlTddddd+nVV1/V7373O/Xs2VO/+tWvJJmThJEjR6q8vFx//OMf1bVrV7333nu666679PPPP2vBggWSzOyqc889V/v379fcuXPVu3dvvf/++5owYUKNsXzyySc6//zzNXz4cD3//PNKTEzU66+/rgkTJqi4uNgrCNUQR44ckSQ9+OCDSktLU2Fhod5++22deeaZ+uijj3TmmWdKkiorKzVu3DitWrVKM2bM0Nlnn63KykqtWbNGWVlZGjlypCRp0qRJ+uc//6nJkydrzpw5ioyM1Lfffqtdu3Y17uZLuvnmm3XhhRfqH//4h4qKihQREaH9+/crJSVFjz32mNq1a6cjR47o73//u4YPH67169erT58+kqSCggKdfvrp2rVrl373u99p+PDhKiws1Geffabs7Gz17dtXd955p2bMmKGffvpJvXr1cn/uK6+8ovz8fIJoAADguLXk+aMvr776qq677jqNGTNGr732msrKyvT444+755Snn366JOmGG27Qt99+q//7v/9T7969dezYMX377bc6fPiwJKmoqEjnnXeeunXrpueee06pqanKycnRJ598ooKCghMeJwA/MwDA6cEHHzQkGU899ZRX++DBgw1JxltvveVuq6ioMNq1a2f86le/crfde++9hiTjq6++8jr/tttuMywWi7F161bDMAxj4cKFhiTjP//5j1e/X//614Yk4+WXX3a39e3b1xgyZIhRUVHh1feiiy4y0tPTDbvdbhiGYXzyySeGJOOTTz45ru9cWVlpVFRUGOecc45x2WWXudtfeeUVQ5Lxt7/9rdZzP/vsM0OSMXv27Do/Q5Lx4IMP1mjv0qWLceONN7rfv/zyy4YkY+LEiQ0ad3l5udGrVy9j5syZ7vY5c+YYkozMzMxaz83Pzzfi4+ON6dOne7X379/fOOuss+r9bAAAAJfWMH+s3s9utxsdOnQwTj75ZPe1DMMwCgoKjPbt2xsjR450t8XFxRkzZsyo9drffPONIcl455136hwDgKaBck4ANVx00UVe7/v16yeLxaJx48a528LDw9WzZ0/t3r3b3fbxxx+rf//+Ou2007zOnzRpkgzD0McffyzJ/HUwPj5eF198sVe/a6+91uv99u3b9eOPP+q6666TZGaHuR4XXHCBsrOztXXr1uP+fs8//7xOOeUURUVFKTw8XBEREfroo4+0ZcsWd5///ve/ioqK0s0331zrdf773/9Kkt8zty6//PIabZWVlXr00UfVv39/RUZGKjw8XJGRkfrpp59qjLt3794699xza71+fHy8brrpJi1evFhFRUWSzD+7zZs364477vDrdwEAAK1DS58/etq6dav279+vG264QWFhVf+kjouL0+WXX641a9aouLhYknTaaadp8eLFeuSRR7RmzRpVVFR4Xatnz55q06aNfve73+n555/X5s2bT2hsAAKLIBqAGpKTk73eR0ZGKiYmRlFRUTXaS0tL3e8PHz6s9PT0Gtfr0KGD+7jrOTU1tUa/tLQ0r/cHDhyQJN11112KiIjwekydOlWSdOjQoeP6bvPmzdNtt92m4cOH680339SaNWu0du1anX/++SopKXH3O3jwoDp06OA1Maru4MGDslqtNcZ9onzdw1mzZukPf/iDLr30Ur377rv66quvtHbtWg0aNKjGuDt16lTvZ9x5550qKCjQkiVLJEnPPvusOnXqpEsuucR/XwQAALQaLXn+WJ1rTLWN2+Fw6OjRo5KkpUuX6sYbb9SLL76oESNGKDk5WRMnTnSv5ZaYmKiVK1dq8ODBuv/++3XSSSepQ4cOevDBB2sE3ACEHmuiAfCblJQUZWdn12jfv3+/JKlt27bufl9//XWNftUXhnX1v++++9zrZlTnWgusof75z3/qzDPP1MKFC73aq6850a5dO33++edyOBy1BtLatWsnu92unJwcn5MoF5vNprKyshrtrglYdRaLxee4J06cqEcffdSr/dChQ0pKSvIaU0MWoe3Zs6fGjRun5557TuPGjdOyZcv08MMPy2q11nsuAACAvzSH+aOvMUuqddxhYWFq06aNezzz58/X/PnzlZWVpWXLlunee+9Vbm6uPvjgA0nSySefrNdff12GYei7777T4sWLNWfOHEVHR+vee+89obEC8C8y0QD4zTnnnKPNmzfr22+/9Wp/5ZVXZLFYdNZZZ0mSzjrrLBUUFHjtYCmZC7R66tOnj3r16qWNGzdq2LBhPh/x8fHHNUaLxSKbzebV9t1339XYSWncuHEqLS3V4sWLa72WqzyhekCuuq5du+q7777zavv4449VWFh4QuN+//33tW/fvhpj2rZtm7v0oS7Tp0/Xd999pxtvvFFWq1W//vWvGzweAAAAf2gO88fq+vTpo44dO+rVV1+VYRju9qKiIr355pvuHTur69y5s+644w6dd955Nb6vZM73Bg0apKefflpJSUk++wAILTLRAPjNzJkz9corr+jCCy/UnDlz1KVLF73//vtasGCBbrvtNvXu3VuSNHHiRD399NOaOHGi/u///k+9evXS8uXLtWLFihrXfOGFFzRu3DiNHTtWkyZNUseOHXXkyBFt2bJF3377rd54443jGuNFF12kP/7xj3rwwQd1xhlnaOvWrZozZ466deumyspKd79rrrlGL7/8sqZMmaKtW7fqrLPOksPh0FdffaV+/frp6quv1ujRo3XDDTfokUce0YEDB3TRRRfJZrNp/fr1iomJ0Z133inJ3JXpD3/4gx544AGdccYZ2rx5s5599lklJiYe17gXL16svn37auDAgVq3bp2eeOKJGqWbM2bM0NKlS3XJJZfo3nvv1WmnnaaSkhKtXLlSF110kXsiKknnnXee+vfvr08++cS9rTwAAEAwNYf5Y3VhYWF6/PHHdd111+miiy7SrbfeqrKyMj3xxBM6duyYHnvsMUlSXl6ezjrrLF177bXq27ev4uPjtXbtWn3wwQfuLLn33ntPCxYs0KWXXqru3bvLMAy99dZbOnbsmM4777wTGieAAAjptgYAmhTX7koHDx70ar/xxhuN2NjYGv3POOMM46STTvJq2717t3HttdcaKSkpRkREhNGnTx/jiSee8Nq5yDAMY+/evcbll19uxMXFGfHx8cbll19urF69usbuSoZhGBs3bjSuuuoqo3379kZERISRlpZmnH322cbzzz/v7tPQ3ZXKysqMu+66y+jYsaMRFRVlnHLKKcY777xj3HjjjUaXLl28+paUlBgPPPCA0atXLyMyMtJISUkxzj77bGP16tXuPna73Xj66aeNAQMGGJGRkUZiYqIxYsQI49133/X6zHvuucfIyMgwoqOjjTPOOMPYsGFDrbtzrl27tsa4jx49akyePNlo3769ERMTY5x++unGqlWrjDPOOMM444wzavSdPn260blzZyMiIsJo3769ceGFFxo//vhjjes+9NBDhiRjzZo1dd43AAAAX1rD/LG2fu+8844xfPhwIyoqyoiNjTXOOecc44svvnAfLy0tNaZMmWIMHDjQSEhIMKKjo40+ffoYDz74oFFUVGQYhmH8+OOPxjXXXGP06NHDiI6ONhITE43TTjvNWLx4cZ1jAhAaFsPwyD8FALQqw4YNk8Vi0dq1a0M9FAAAAABo0ijnBIBWJj8/Xz/88IPee+89rVu3Tm+//XaohwQAAAAATR5BNABoZb799ludddZZSklJ0YMPPqhLL7001EMCAAAAgCaPck4AAAAAAACgHmGhHgAAAAAAAADQ1BFEAwAAAAAAAOpBEA0AAAAAAACoR6vbWMDhcGj//v2Kj4+XxWIJ9XAAAEAzYRiGCgoK1KFDB4WF8TtkU8Q8DwAANEZD53mtLoi2f/9+ZWRkhHoYAACgmdqzZ486deoU6mHAB+Z5AADgRNQ3z2t1QbT4+HhJ5o1JSEgI8WgAAEBzkZ+fr4yMDPdcAk0P8zwAANAYDZ3ntbogmiu1PyEhgckVAAA4bpQJNl3M8wAAwImob57Hgh4AAAAAAABAPQiiAQAAAAAAAPUgiAYAAAAAAADUo9WtiQYAAAAAANBcGIahyspK2e32UA+l2bJarQoPDz/htW0JogEAAAAAADRB5eXlys7OVnFxcaiH0uzFxMQoPT1dkZGRjb4GQTQAAAAAAIAmxuFwaOfOnbJarerQoYMiIyPZJbwRDMNQeXm5Dh48qJ07d6pXr14KC2vc6mYhDaJ99tlneuKJJ7Ru3TplZ2fr7bff1qWXXlrnOStXrtSsWbO0adMmdejQQffcc4+mTJkSnAEDAAAAAAAEQXl5uRwOhzIyMhQTExPq4TRr0dHRioiI0O7du1VeXq6oqKhGXSekGwsUFRVp0KBBevbZZxvUf+fOnbrgggs0evRorV+/Xvfff7+mTZumN998M8AjBQAAAAAACL7GZk3Bmz/uY0gz0caNG6dx48Y1uP/zzz+vzp07a/78+ZKkfv366ZtvvtGTTz6pyy+/3Oc5ZWVlKisrc7/Pz88/oTG3Jmt3HdF/NuyTwwj1SIKjQ/FW9Shcpy/aXSWHpWlUOmcU/aDOxZu0uu2VMiwN/x98v7zP1LvgqwCODACaJkdMe42Y/GSoh4EW6oMfcvTZTwd1es+2uuDk9FAPBwAABFnTiBQ00JdffqkxY8Z4tY0dO1YvvfSSKioqFBERUeOcuXPn6uGHHw7WEFsMh8PQ9NfWa39eaaiHEiSGPo6cre5hOfosq0yv2c8J9YAUJoe+sN2vdMsRrdjl0DLHqAadl6QCPWh7QDZLZYBHCABNz+6jGZIIoiEwNuw5ple/ylJUuJUgGgAArVCzCqLl5OQoNTXVqy01NVWVlZU6dOiQ0tNrTmbuu+8+zZo1y/0+Pz9fGRkZAR9rc7cu66j255UqzhauX4/uHurhBFxqwQ/qvjFHknRHyjqlDbwtxCOSMo59rfQfjkiSZqWuV4+TbmrQeSdn/1u2nyt1LKqTtrS/MJBDBIAmJyw2RV1CPQi0WHE2qySpqIwfqgAACJauXbtqxowZmjFjRqiH0ryCaJJq7ERhGIbPdhebzSabzRbwcbU0yzbslySNPSlN08/tFeLRBMH7L7hfdszfoOlDI6Q2XUM3Hkl6e577ZddjazR9eIIUn1rHCU4vfSJJSvrlFI0YeWegRgcAQKsTZzOnzoUE0QAAqNOZZ56pwYMHu5fjOhFr165VbGzsiQ/KD5rV6nRpaWnKycnxasvNzVV4eLhSUlJCNKqWp9Lu0PLvsyVJFw/uEOLRBEFlufSDc3OKmLbm83f/Ct14JKm8WNqyzHwd01YyHNIP/67/vCM7pD1fSZYwacAVgR0jAACtTCxBNAAA/MIwDFVWNuy/p+3atWsyu5M2qyDaiBEjlJmZ6dX2v//9T8OGDfO5Hhoa58sdh3W4qFzJsZEa2aMVBCe3fyiVHJFi20vnOdfP2/i6ZIRwR4Uf35fKC81suDN+VzWm+riCf93OkBJYqwUAAH+KjyKIBgAILcMwVFxeGZKH0cB/I0+aNEkrV67Un//8Z1ksFlksFi1evFgWi0UrVqzQsGHDZLPZtGrVKv3888+65JJLlJqaqri4OJ166qn68MMPva7XtWtXr4w2i8WiF198UZdddpliYmLUq1cvLVu2zJ+3uVYhLecsLCzU9u3b3e937typDRs2KDk5WZ07d9Z9992nffv26ZVXXpEkTZkyRc8++6xmzZqlX//61/ryyy/10ksv6bXXXgvVV2iRXKWc4wakKcLarOKsjfOdMzh18pVS/0ul5XdLR36W9q2TOg0L7ZgGTpAGXC6tuE/K+U7K3SK17+f7HMOoCrQNujo44wQAoBVxZaKxJhoAIFRKKuzq/8CKkHz25jljFRNZfxjpz3/+s7Zt26YBAwZozpw5kqRNmzZJku655x49+eST6t69u5KSkrR3715dcMEFeuSRRxQVFaW///3vGj9+vLZu3arOnTvX+hkPP/ywHn/8cT3xxBP6y1/+ouuuu067d+9WcnKyf75sLUIaIfnmm280ZMgQDRkyRJI0a9YsDRkyRA888IAkKTs7W1lZWe7+3bp10/Lly/Xpp59q8ODB+uMf/6hnnnlGl19+eUjG3xKVVdr1wSazZHb8oFZQyllyTNr6gfl60ATJFif1vch835DMr0AoOCD9/LH5euAEKTZF6jWm/jHtXSsd3SlFxFR9BwAA4DeUcwIAUL/ExERFRkYqJiZGaWlpSktLk9Vqbs4zZ84cnXfeeerRo4dSUlI0aNAg3XrrrTr55JPVq1cvPfLII+revXu9mWWTJk3SNddco549e+rRRx9VUVGRvv7664B/t5Bmop155pl1pgMuXry4RtsZZ5yhb7/9NoCjat0+23ZIBaWVSk2w6dSugY3gNgmb/yPZy6R2/aS0gWbboAnS9/8y10k7f65kDXKp8A9vmmugdRwmpfQw2wZOkLYul75/QzrnQSnMR/z7u6Xmc7/xZjAQAAD4VTxBNABAiEVHWLV5ztiQffaJGjbMu9qrqKhIDz/8sN577z3t379flZWVKikp8Uqo8mXgwIHu17GxsYqPj1dubu4Jj68+zW53TgTWuxvNUs6LBnaQNcz3jqctiivwNGiC5NrhtduZUlyqVHjAXC+tz7ggj8lHSWbv8yVbopS/T9r9udTtl97neG6OMHBCcMYJAEArQzknACDULBZLg0oqm6rqu2zefffdWrFihZ588kn17NlT0dHRuuKKK1ReXl7ndaqvi2+xWORwOPw+3upawYJXaKji8kplbj4gqZWUch7dLe3+QpLFXA/NxRpetbNlsEs6c3+UsjdKYeHSSb+qao+Ikk66xDmmpTXP254plRw1g3/dzgjOWAEAaGVcQbQKu6GySnuIRwMAQNMVGRkpu73+/1auWrVKkyZN0mWXXaaTTz5ZaWlp2rVrV+AH2EgE0VohwzA0c+kGTV68Vluy893tH23JVUmFXRnJ0RrUKTGEIwyS7507WXY9XUrs5H1skDOba+t/zXXTgsWVhdZrjLkWmqeBzsy0zf+Ryou9j2302BzB2nx/lQAAoCmLs1X9N7awlGw0AABq07VrV3311VfatWuXDh06VGuWWM+ePfXWW29pw4YN2rhxo6699tqgZJQ1FkG0Vmjv0RK9vX6fPvoxVxc+s0q/f+d7HS0qd5dyjh/YQRZLCy/lNIyqjC5fO1mmDTTXSbOXmUGrYHA4pO/eMF8PvKrm8c4jpMTOUnmBuT6aS8lRaZtzcwRKOQEACBhrmMW9HkxRGZloAADU5q677pLValX//v3Vrl27Wtc4e/rpp9WmTRuNHDlS48eP19ixY3XKKacEebQNZzHqWtm/BcrPz1diYqLy8vKUkJAQ6uH4X2me9K8bpQG/kk6Z6LPLR5tzdODV2zTCulmuP31rmEUOw5BhSJ3aRCvSWkd8deAE6czf1X78ozlS3l7pkudqX5R/41Jp1VOSo9qvuD3PlS54vPZrr5onbVginehfW8Nh7mQZHiXd9ZMU5ePvwudPSx8+JNkSpNh2J/Z5DeGolI7tNtc+u2ubWcJZ3Ud/lFY9KUUlSjFtzbaKEqlgv9S+v3Tb6qq13QAAftXi5xAtQDD+jIY98qEOFZZp+bTR6t+BvwcAgMApLS3Vzp071a1bN0VF+fj3IY5LXfezoXMI6r5amt1fSjs+MRegryWItmfPTk0K/9h84xlvsTgfefV8xso/ScNukuLa1zx2+GczOCZJ/S+R+l5Ys4/DbganCvbXPPb1z9Kwm6X2fWseKz4iffKo5KioZ4DH4eQrfAfQJDNYuPJxqSzffATL4Gt8B9AkafC10upnzGBpabU/qKE3EUADACDA4qPCdaiwjB06AQBohQiitTTlhebzsT1mtpaPoMrR/T9Lkooi2yr2un/Kbhj6aMsBfbL1oH41pKNO7Zpc+/X/e4+U8530w1vSL6bUPP79G1Wvv1vqO4i2a5UZQItKlK5+TbI4s95WPibt+NQ879wHa563+R0zgNa+v3ThvNrH2FBh4VL6oNqPJ3SQ7vhGOlb31rp+ZY2U0gfWfjylh3TnOilvn3d7ZIyUenJgxwYAABRrc5VzEkQDAKC1IYjW0lSUmM+VJVLxYSm2bY0uJQd3S5LKE7ootssIWSWN6SqNGdeA6w+53gykffd6zSCaYXjvZrn1A3NR/ugk736utchO+pXUdVRV+9CbzCDa929IZ/9BCgvzfd7ga6UuIxowWD9I7Gg+mpKkzuYDAAAEnWtzgQKCaAAAtDpsLNDSuIJoks8Mqgq7Q+EFeyVJkSldjv/6J/1Kslil/eulg9u8j+1da64zFhErpfR0Lsr/jnef8mJpyzLzdfVF8Hufb64HlrdH2v2F97EjO6U9a8ystQFXHP+4AQAA/MAVRCMTDQCA1ocgWktTUVT1Om9PjcO7DxcpzTgoSYpp1+34rx/Xzlz8XzKz0Ty5stD6jTcz1qSq7DGXrcvNktOkLlLnX3gfi4iSTrrE97W/+5f53O0MKSH9+McNAADgB7EE0QAAaLUIorU0XploNYNoW3MK1dFySJJkScpo3GcMcmaQffcvyeEwX1eWS5veqjp+8lWSLFLWauno7qpzXYG2gRN8L4I/8GrzefOyqu9iGFVBtUFXN27MAAAAfuAu5ywliAYAQGtDEK2lqSiueu0jE23rgQJ3EE2NDaL1uUCyJZjXz/rSbNueKZUcleLTzWyxxI5St9Hmse+dWWSFudLPzl1BawuGdR4hJWaYu2Fu/a/Ztm+ddGSHFBEj9b2ocWMGAADwg6CWc+75Wvriz+bO5gAAIOQIorU09WSibcvOrwqiJTZycfqIaKn/xeZrV4aYK8Ps5CukMHPXKndW2calZjbZD29Khl3qOMzcZdKXsDBp4FXOay/1vnbfiyRbXOPGDAAA4Afucs7yIATRPrhXynzAXHcWAACEHEG0lqbcMxOt5sYC+w/kKN7iDLQldmr857gCZJv+IxXkSNs+cLZ7bBbQb7wUHiUd/kna/21VMKy+kkzXtbd/KOVnm8E3qaqMFAAAIESCWs5Zmu/9DAAAQoogWkvjWc5ZLROttMIuu3N9Mkd0ihQZ0/jP6TJKSugkleVJ79wm2cul9idJaSdX9YlKkPpeaL7++BEpe4MUFm7u8FmXdr2lDkMkR6X0n6lSyREpLlXqdmbjxwsAAOAHQS3ntJebz46KwH8WAABNSNeuXTV//nz3e4vFonfeeafW/rt27ZLFYtGGDRsCOi6CaC2NZzln6TGprMD9dntuodJ1WJJkadPIUk6XsDBp4JXma/c6Zz4yxVxZZa4+Pc+TYlPqv74ro8113oArJGt448cLAADgB3FR5nykMChBtArvZwAAWqns7GyNGzcu1MMgiNZkGIZ3AKw2FaX1HC/2fu+RjbY1p2pTAUtiIzcV8DTQsyzTIp18Zc0+Pc6WYttVvW9oSeaAKySL9fjPAwAACCDXmmiFZUFY7N+ViUYQDQDQyqWlpclms4V6GATRmowV90t/6iod3FZ7nx/ekh7tYC7UX5vqQTSPHTq3ee7M6Y8gWvu+Uvog83X3M6SEDjX7WMPNgJgk2RKl3g2MHMe1k3qeY75u109KG3ji4wUAADhBcTbzR77glHM6g2eUcwIAXAxDKi8KzcMwGjTEF154QR07dpTD4fBqv/jii3XjjTfq559/1iWXXKLU1FTFxcXp1FNP1YcffljnNauXc3799dcaMmSIoqKiNGzYMK1fv/64b2VjUB/XVOxcJVWWSlmrzTXBfNnxibm75Z6vas/McmWzRcZL5QXSsarNBbYeKNCVloPmmyQ/BNEk6Yx7peV3S6Pvqr3P8N9I2zOlUyZKEVENv/bo30q5W6Qz75UslhMfKwAAwAmKs0VIClY5Z7n3MwAAFcVmck0o3L9fioytt9uVV16padOm6ZNPPtE555jJMUePHtWKFSv07rvvqrCwUBdccIEeeeQRRUVF6e9//7vGjx+vrVu3qnPn+peeKioq0kUXXaSzzz5b//znP7Vz505Nnz79hL9eQxBEaypKjprP1TYD8OI6Vj3bzFN5kfnctpe5I6ZnJlqOnzPRJKnvBeajLsndpTvXHf+1O/9CmvlD48YFAAAQALHOTLTgBtHIRAMANB/Jyck6//zz9eqrr7qDaG+88YaSk5N1zjnnyGq1atCgQe7+jzzyiN5++20tW7ZMd9xxR73XX7Jkiex2uxYtWqSYmBiddNJJ2rt3r2677baAfScXgmhNRckR8zmvjiCa61h5Ye19XJlo7fqaQTRn4C2/tEL780rV0eYMovkrEw0AAKAVce3OWV7pUHmlQ5HhAVodxWE3KxAkc8dyAAAkKSLGzAgL1Wc30HXXXaff/OY3WrBggWw2m5YsWaKrr75aVqtVRUVFevjhh/Xee+9p//79qqysVElJibKysuq/sKQtW7Zo0KBBiompGs+IESOO++s0BkG0pqCitCq7rLZMNMOQ8vaar13ZZj6v5QqiOUtCnef8dKBANpWrnSXfbPdXJhoAAEAr4tpYQDLXRYsMjwzMB3lmn1HOCQBwsVgaVFIZauPHj5fD4dD777+vU089VatWrdK8efMkSXfffbdWrFihJ598Uj179lR0dLSuuOIKlZc37L93RgPXZgsEgmhNgauUU6o9E63ooLlmmiSV11LOaRhShaucs4/X9X70LOWMjJOi25zgoAEAAFqfCGuYbOFhKqt0qLCsUm1iAxVE8/iHBOWcAIBmJjo6Wr/61a+0ZMkSbd++Xb1799bQoUMlSatWrdKkSZN02WWXSZIKCwu1a9euBl+7f//++sc//qGSkhJFR0dLktasWeP37+ALu3M2BZ5BtPz9kt1Hyr5HhpqjrJZyTnu5ZDh3v2jnDKIV5EiV5TXXQ2OhfgAAgEaJjzJ/hy4qD2CZpVcmGkE0AEDzc9111+n999/XokWLdP3117vbe/bsqbfeeksbNmzQxo0bde2119bYybMu1157rcLCwjR58mRt3rxZy5cv15NPPhmIr1ADQbSmwLUemmSufVHgo745r6o2+GjeMd/X8dxwIKmzFB4tyZDy92rrAY8gGuuhAQAANJqrpLOwNJBBNI9MNAdBNABA83P22WcrOTlZW7du1bXXXutuf/rpp9WmTRuNHDlS48eP19ixY3XKKac0+LpxcXF69913tXnzZg0ZMkSzZ8/Wn/70p0B8hRoo52wKio94vz+2xwyCVW9zMsoKVWl3KNxaLQbqKvMMi5CsEVJiJ+nwTzKOZWlrTrlG+XtnTgAAgFYoNtIZRAvkDp2UcwIAmjmr1ar9+2smCXXt2lUff/yxV9vtt9/u9b56eWf1ddB+8YtfaMOGDXX2CQQy0ZoCz3JOyfe6aB5tUUapvtxxuGYf16YCrh0znBlnBQd26WhxBZloAAAAfhAXFYwgGuWcAAA0NQTRmoISH5lo1Xm0xahM727YV7OPq5wzwlxYz5Vxdmz/z5KkHhFHvdoBAABw/OKc5ZxFwcpEo5wTAIAmgSBaU1AjEy2rRhfDIxMtzGLok01ZKqu0e3dyBdEivTPRSg/tkiR1DHNlolUrFQUAAECDuddEK7PX0/MEeJVzltfeDwAABA1BtGA4tF06sKn246410dp0M5/ryUSTJHtpkT7bdsi7jzsTzRlEc2acheXvkVV2JVeyJhoAAAiuBQsWqFu3boqKitLQoUO1atWqOvsvWbJEgwYNUkxMjNLT03XTTTfp8GEfy1iEUFxQNhbwLOcM4OcAAIAGI4gWaIYhLRorvXiuVFbou48rEy19kPlcfU200nxZyvIkSZXOP7IYS6ne3VhtgT73mmje5ZxRRfuVqqMKk12yRkpxqSf0lQAAABpi6dKlmjFjhmbPnq3169dr9OjRGjdunLKyambdS9Lnn3+uiRMnavLkydq0aZPeeOMNrV27VrfcckuQR163OJtVklRUTjknACDwgrFgfmvgj/tIEC3Q7OVS8SEzSyzfxzpmko8g2l4z+ObiDKodNeJUaImXJMWqVJmbD6jYc/JWy8YC7Y1D6hftzHZL6CiF8ccOAAACb968eZo8ebJuueUW9evXT/Pnz1dGRoYWLlzos/+aNWvUtWtXTZs2Td26ddPpp5+uW2+9Vd98802QR163qnLOYO3OSTknALRGERERkqTi4uIQj6RlcN1H131tjHB/DQa1qCyrel2YK7XrU7OPq5wzdYAki1RZKhUdlOLam+3OUs59RlulhpVK9jx1TZC25tn10ZZcjR/UwexXXmQ+u4Jo8R1kl1WRFruuSD0g5UhK7OT3rwgAAFBdeXm51q1bp3vvvderfcyYMVq9erXPc0aOHKnZs2dr+fLlGjdunHJzc/Xvf/9bF154oc/+ZWVlKiurmmvl5+f77wvUgXJOAEAwWK1WJSUlKTc3V5IUExMji8US4lE1P4ZhqLi4WLm5uUpKSpLVam30tQiiBZrnL4eFB3z3cWWixbWX4tOlgv1m4MwVRMurCqK1jTgi2aWzu8dqxXrp3Y37q4Jo1co5y40wHVIbddAhDbf+aB5jUwEAABAEhw4dkt1uV2qq9zISqampysnJ8XnOyJEjtWTJEk2YMEGlpaWqrKzUxRdfrL/85S8++8+dO1cPP/yw38den6DvzkkmGgC0WmlpaZLkDqSh8ZKSktz3s7EIogWaZyZa0cGaxw1DKnFmosUkm5liBfvNwFmnoWb7MXPdkH1GWw2MKJVKpVGdo6X10qdbDyq/tEIJURE1dudc9dNBxTnaqkPYIbU5tM48xqYCAAAgiKr/Ym4YRq2/om/evFnTpk3TAw88oLFjxyo7O1t33323pkyZopdeeqlG//vuu0+zZs1yv8/Pz1dGRuDnOnFRQS7nZE00AGi1LBaL0tPT1b59e1VU8N+DxoqIiDihDDQXgmiBVlla9brQR+S4orhqkhTdxlzHbO/X3psLeGSiWSLNrLWOMQ71ah+nn3ILteKHHF05LKPG7pzvbtyvXxptJUmWMmd5QxJBNAAAEHht27aV1WqtkXWWm5tbIzvNZe7cuRo1apTuvvtuSdLAgQMVGxur0aNH65FHHlF6erpXf5vNJpvNFpgvUIfgrIlGOScAoIrVavVLEAgnhhXmA83zV8QiH0E013poYRFSZFxVptgxjyCa8/Veo63Co8yNBSwVRe4yzne/yzb7eZRzlpTb9b/NB7TPGURzIxMNAAAEQWRkpIYOHarMzEyv9szMTI0cOdLnOcXFxQqrtgGS6x8MTWlnMso5AQBonQiiBZrXxgI+yjld66FFt5EslqpMsVoy0SKj48y28mJ3EO2L7Ye0PbfQIxMtVh//mKvicruKorx/sSUTDQAABMusWbP04osvatGiRdqyZYtmzpyprKwsTZkyRZJZjjlx4kR3//Hjx+utt97SwoULtWPHDn3xxReaNm2aTjvtNHXo0CFUX6OGuGDvzkk5JwAATQLlnIFW38YCnuuhSVKic+F/VyZaZZn7vH1GW9liE8z28iJ1axur03u21efbD+nmxWuV2bVANkmKiNa7G/dLkrr26CttdX2YRUpgd04AABAcEyZM0OHDhzVnzhxlZ2drwIABWr58ubp06SJJys7OVlZWlrv/pEmTVFBQoGeffVa//e1vlZSUpLPPPlt/+tOfQvUVfApOEI1yTgAAmhqCaIFW38YC7kw0ZxDNnYnmnFDm7ZUkFRs2lUYkVWWiVRRJkuZfPViXLfhCWUeK9a19v0ZIKrXY9PFWs3R02KBBVUG0+DQpPNJPXwwAAKB+U6dO1dSpU30eW7x4cY22O++8U3feeWeAR3ViXGuilVY4VGl3KNwagOIOyjkBAGhyKOcMNHu1IFr19Txca6JFtzGfXWuWleZJpflepZzpidGyRLrKOc0gWts4m16edJoSosJVWlwoSfo+t0LllQ71aBerHj37VH0W66EBAACcsFhb1cLORWX2wHwI5ZwAADQ5BNECrbLar4ilx7yPu8s5nUE0W1xVQC1vj7usc5/RVqkJUVJkrHmsvNh9iZ7t4/TCDcMUazEDdq+tPyRJGj+ogyyRMVJsO7Mj66EBAACcMFu4VZHO7LPC8gCVWnqVcxJEAwCgKSCIFmiVpd7vq28uUHLMfHYFziTvHTq9MtE8g2iFXpcZ0SNFPduYv4oerTBLDFwbD7ivRyYaAACAX8RFBXiHTq9yToJoAAA0BQTRAq36GhZFud7v3eWcyVVtSc7NBbwy0VKU5hVEK6rxUcmRZjlBqSI1KCNJPdo5Sz/b9XU+96lxDgAAAI6fq6SzoDQIQTTKOQEAaBLYWCDQPDcWkGru0OneWMBXJlqWVyba0MQoKSLGPFZRrBqcJZ53X3SK0k86par93Iek7mdKJ13WuO8AAAAAL7GRgc5Eo5wTAICmhky0QKueiVajnNO1JppnJpprh849ZiBNZhAtLSFKcm8s4F3OKckdWDulRwelJ0ZXtcenSoMmsDMnAACAn8RTzgkAQKsT8iDaggUL1K1bN0VFRWno0KFatWpVnf2fe+459evXT9HR0erTp49eeeWVII20kapnolUv5/SZidbJfD66W8rfJ0naZ7SrVs7pIxOtosR8joiueQwAAAB+E2szg2gFQclEK6+9HwAACJqQBtGWLl2qGTNmaPbs2Vq/fr1Gjx6tcePGKSsry2f/hQsX6r777tNDDz2kTZs26eGHH9btt9+ud999N8gjPw726uWcDVgTzVXOmfO95KhUhWHVAbVxBtGc5ZzV10QzjKoST1egDQAAAAHhCqIFJRPNsJtzPQAAEFIhDaLNmzdPkydP1i233KJ+/fpp/vz5ysjI0MKFC332/8c//qFbb71VEyZMUPfu3XX11Vdr8uTJ+tOf/hTkkR+HSucEyOospSzyKOc0jKpMtBgfGws4F5HNMZIVFmZV21ibRzlntSBaZakk5+SKTDQAAICAincG0QqDsbGAREknAABNQMiCaOXl5Vq3bp3GjBnj1T5mzBitXr3a5zllZWWKioryaouOjtbXX3+tigrfE4uysjLl5+d7PYKqstR8dpVoem4sUJZv/rIoeZdzxqRI4VWBsH1qq9SEKIWFWTw2Fijy/kXSVcopVfUBAABAQLgy0QrLg1DOKVHSCQBAExCyINqhQ4dkt9uVmprq1Z6amqqcnByf54wdO1Yvvvii1q1bJ8Mw9M0332jRokWqqKjQoUOHfJ4zd+5cJSYmuh8ZGRl+/y51ck143EE0j0w0VylneLR39pjFUrW5gJybCiQ6g4euUk3DURWgk6pKOa02Kczqxy8AAACA6uKCWc4puSsUAABA6IR8YwGLxeL13jCMGm0uf/jDHzRu3Dj94he/UEREhC655BJNmjRJkmS1+g4c3XfffcrLy3M/9uzZ49fx18u1sYBrnbOi3KoMMl+bCrgkVgXR9rp25pS81zvz3FzA9ZpSTgAAgICLo5wTAIBWJ2RBtLZt28pqtdbIOsvNza2RneYSHR2tRYsWqbi4WLt27VJWVpa6du2q+Ph4tW3b1uc5NptNCQkJXo+gcm0s4MpEs5dLpXnm6xJnJprnemgutWWihVmlcOfr8sKq/q5MNEo5AQAAAs5dzllmD8wH1CjnJIgGAECohSyIFhkZqaFDhyozM9OrPTMzUyNHjqzz3IiICHXq1ElWq1Wvv/66LrroIoWFhTypzjfXxgK2BPMhVW0uUHLMfK4nE22fZyaaVJWN5rm5gGtNtEiCaAAAAIEWF0U5JwAArU14KD981qxZuuGGGzRs2DCNGDFCf/3rX5WVlaUpU6ZIMksx9+3bp1deeUWStG3bNn399dcaPny4jh49qnnz5umHH37Q3//+91B+jbq5MtHCbVJsO3MzgcIDUtteVWui+QqiuXboVLVMNEmKiJV0uCr7TDI3GpAo5wQAAAiCOJu5lEhhsIJoZKIBABByIQ2iTZgwQYcPH9acOXOUnZ2tAQMGaPny5erSpYskKTs7W1lZWe7+drtdTz31lLZu3aqIiAidddZZWr16tbp27Rqib9AArkw0a6QU11468rNUmGu2udZE81XO6ZGJlm2kKD3RVyaaZzmnMxONck4AAICAi40MdCYa5ZwAADQ1IQ2iSdLUqVM1depUn8cWL17s9b5fv35av359EEblR+5MtCgzE03yKOesIxOtfT8ZkfHaVNZWZYpUqs9yTs9MNFcQjUw0AACAQHOVcwYvE63cdz8AABA0TXQhsRbEtTtneKQU59wwwZWJ5i7n9JGJFp2kY7/5RleUPSBJ1YJozmwzzzXRXK8jPHbvBAAAQEC4d+cMViaaI0CfAwAAGowgWqC5gmhWm1nOKUlF1co5fWWiSdpfHq1S2dQ2LlKR4R5/VJFx5rPPck4y0QAAAALNtTtncblddofh/w9gTTQAAJocgmiBZvfIRHOVc7rXRHNmovlaE03SgfxSSfLeVECqWvfMa2MB52uCaAAAAAHnykSTpKLyAGSJuXfjtJhPlHMCABByBNECzb2xgEcmWvWNBWrJRMvOcwbREqoFxtxronmUc7qCaJGUcwIAAASaLTxM4WFmgCsgmwu4Ms9cFQgOMtEAAAg1gmiB5s5Es0mxrnJO58YCda2JJumAK4iWaPM+4C7n9AyiUc4JAAAQLBaLxb25QGCCaM4fYl1r4VLOCQBAyBFECzR3Jlqkdyaawy6V5pnv68lES0+snonmY2MBdzlnjD9GDQAAgHrERppBtILSQAbRnFUGBNEAAAg5gmiB5s5Ei6oKotnLpGNZkpyL0NYSRMtxronmtTOn5Lucs5wgGgAAQDC51kUrKrP798IOu2Q4zNeundcp5wQAIOQIogVapcfGAhHRUmS8+f7gVvM5Mt485kOOOxOt+sYCzslUha9MNMo5AQAAgsFVzlno73JOz00EKOcEAKDJIIgWaK4gmtW5rlmcc4fOQ84gWi1ZaFJVEK1BmWjuNdHIRAMAAAiGWFswgmiUcwIA0FQQRAskh6Mq9T7cFURLNZ9dmWgxvoNohWWVKnBOyNKqZ6K5g2jFVW2uIFokQTQAAIBgiLNZJQVgYwHPgJnrB1LPwBoAAAgJgmiB5DnZsTpLNmOdmWgHfzSfa1sPLc8MisVHhbvX23BzB9EKq9pcpZ1kogEAAARFXKAz0cIiquaQjgBsXgAAAI4LQbRAcm0qIHlkojk3Fzi4zXyOTvZ5atYRM8sso42PoFid5ZysiQYAABAMAS/ntEZK1ghnG+WcAACEGkG0QKr0lYnmDKKVF5jPtWSiZR02g2idk30E0VzZZhU+yjnJRAMAAAiKePfunAEq57RGmNloEuWcAAA0AQTRAsnusamAxWK+dm0s4BJTWyaaGRTrnOIrEy3OfPbMRCunnBMAACCY3JlopUHIRKOcEwCAkCOIFkiunTldpZxS1cYCLvWVc/rKRPMs5zQM8zXlnAAAAEEV3HJOMtEAAAg1gmiB5AqiuUo5papyTpdayjn3Hq2jnNO1A6dhNz/D4ZAqKecEAAAIpvgoZzlneTDKOVkTDQCAUAuvvwsaze4rE63+ck7DMNyZaL7XRIutel1eJEXYq95HEkQDAAAIhthIyjkBAGhNCKIFUqXHBMilAZloh4vKVVxul8UidUzyUZ5pDTfXWbOXSRVFkoyqY+GUcwIAAAQD5ZwAALQulHMGkq9MtMiYqo0BJJ9rormy0DokRisyvJY/Is910Vy7dIZHSWH8kQIAAASDu5yzzF5Pz+PkWc7p+jGWck4AAEKOiEsg+cpEk6Q4j2w0H5loe5xBtE5t6sgqc+/QWWw+JNZDAwAACKKgZKKFOQtHCKIBABByBNECyVcmmuRd0hmdVOO0rMN1rIfm4lr7rLywKhONIBoAAEDQxNmqNhYwDKOe3sfBHUSL8FgTjSAaAAChRhAtkCpLzefwKO921+YCUYlSmLXGaXVuKuDiVc7p2pmT9dAAAACCxRVEMwypuNyPJZ3ucs5IyjkBAGhCCKIFUm3lnK5MNB/roUkeQbSUOoJorqyziuKqTDR25gQAAAiaqIgwhVnM134t6fRZzsnGAgAAhBpBtECqrZzTtSZajO8gmmtNtIw6M9Fca6JRzgkAABAKFovFnY2WX+LHTDGvck7nj7EOP6+7BgAAjhtBtECqLRMtPt18jm1X45SySruy880y0IaVcxZTzgkAABAiPdqbP2yuzzrmv4t6lXM610QjEw0AgJAjiBZItWWi9b9EOvUWafRdNU7Zd7REhiHFRFqVEhtZ47ibe2OBIjLRAAAAQmR0L/NH0ZU/HfTfRdmdEwCAJokgWiBVOoNo1TPRopOkC5+SMk6tcYrnpgIWi6X2a7vKOSuKzGw0iSAaAABochYsWKBu3bopKipKQ4cO1apVq+rsX1ZWptmzZ6tLly6y2Wzq0aOHFi1aFKTRHr9f9morSfpi+yHZHX7aoZNyTgAAmqTwUA+gRXNNgKpnotWhQeuhSVUBs/IiKcJZ2kk5JwAAaEKWLl2qGTNmaMGCBRo1apReeOEFjRs3Tps3b1bnzp19nnPVVVfpwIEDeumll9SzZ0/l5uaqsrLpBpAGZyQp3hauY8UV+mFfngZlJJ34RSnnBACgSSKIFkiV5tpmCo9q8CmemWh1cq+JViRFFnu3AQAANAHz5s3T5MmTdcstt0iS5s+frxUrVmjhwoWaO3dujf4ffPCBVq5cqR07dig52dyAqWvXrrVev6ysTGVlZe73+fn5/v0CDRBuDdPInilasemAVv100E9BNI9yTncQjXJOAABCjXLOQKptY4E6NDyI5tqd03NNNDLRAABA01BeXq5169ZpzJgxXu1jxozR6tWrfZ6zbNkyDRs2TI8//rg6duyo3r1766677lJJSYnP/nPnzlViYqL7kZGR4ffv0RCuddE+23bIPxd0Z6JFSGEE0QAAaCrIRAuk2jYWqEPWEXOSWH8QzdfGAgTRAABA03Do0CHZ7XalpqZ6taempionJ8fnOTt27NDnn3+uqKgovf322zp06JCmTp2qI0eO+FwX7b777tOsWbPc7/Pz80MSSDujtxlE+zbrqApKKxQfFXFiF/SVieYgiAYAQKgRRAuk48xEMwyj4WuiuUo3K4qlCuevs2wsAAAAmpjqGyUZhlHr5kkOh0MWi0VLlixRYmKiJLMk9IorrtBzzz2n6GjvHwxtNptstob/WBkoGckx6poSo12Hi7VmxxGd1z+1/pPq4hVEc84jyUQDACDkKOcMpOPMRDtaXKHCMnPh3E5t6skqc20mUF7I7pwAAKDJadu2raxWa42ss9zc3BrZaS7p6enq2LGjO4AmSf369ZNhGNq7d29Ax3uiqko6D574xbzKOcO92wAAQMgQRAukSmcQrYGZaK710FITbIqKsNbd2XNjgQqCaAAAoGmJjIzU0KFDlZmZ6dWemZmpkSNH+jxn1KhR2r9/vwoLC91t27ZtU1hYmDp16hTQ8Z6oXzpLOlf95I8gmo9MNMo5AQAIuUYF0T799FM/D6OFck2AGpiJ1uBNBSSPIJpnOSdrogEAgKZj1qxZevHFF7Vo0SJt2bJFM2fOVFZWlqZMmSLJXNNs4sSJ7v7XXnutUlJSdNNNN2nz5s367LPPdPfdd+vmm2+uUcoZMvZKqaK0RvMvuicrPMyiXYeLlXW4+AQ/wxVEi/DYnbP8xK4JAABOWKOCaOeff7569OihRx55RHv27PH3mFoOVyZaeFSDujd4PTSpWiZakbONTDQAANB0TJgwQfPnz9ecOXM0ePBgffbZZ1q+fLm6dOkiScrOzlZWVpa7f1xcnDIzM3Xs2DENGzZM1113ncaPH69nnnkmVF/B2/9+L/0xRfrs8RqH4qMidErnNpKkz040G81dzhnpUc5ZeWLXBAAAJ6xRQbT9+/dr+vTpeuutt9StWzeNHTtW//rXv1Rezi9kXo63nPNwIzLRKopYEw0AADRZU6dO1a5du1RWVqZ169bpl7/8pfvY4sWLa1Q49O3bV5mZmSouLtaePXv01FNPNZ0sNNeatCVHfR7+Ze+2kvxQ0kk5JwAATVKjgmjJycmaNm2avv32W33zzTfq06ePbr/9dqWnp2vatGnauHGjv8fZPB3nxgLHVc7pCpg5KqXSPGdbE5lgAgAAtETRZqaZio/4POzaXGD19sOqtDsa/zm1lXMaRuOvCQAATtgJbywwePBg3Xvvvbr99ttVVFSkRYsWaejQoRo9erQ2bdrkjzE2X5UevyI2QKPWRJOk4sPmc0Ss774AAAA4cTHJ5nMtmWgDOiYqKSZCBWWV2rDnWOM/x7Oc0xVEkySHvfHXBAAAJ6zRQbSKigr9+9//1gUXXKAuXbpoxYoVevbZZ3XgwAHt3LlTGRkZuvLKK/051ubnODLRKuwOZeeZGwQ0KIhmjZCsrus6f5UkEw0AACBwXJloJb4z0axhFo3qaZZ0fvbTocZ/jmc5Z5hnEI2STgAAQqlRQbQ777xT6enpmjJlinr37q3169fryy+/1C233KLY2FhlZGToscce048//ujv8TYv7jXR6g+i7T9WIoch2cLD1C6+YeWfNTYSYE00AACAwIl2ZaIdq7XLL3uZQbQ1Px9u/Of4Kuf0bAcAACER3piTNm/erL/85S+6/PLLFRnpu1SxQ4cO+uSTT05ocM2ea6ITXn85p2cpp8Viadj1I+O8ywnYnRMAACBwYupeE02S+qQlSJL2HC1u/Od47c7pGURjh04AAEKpUUG0jz76qP4Lh4frjDPOaMzlW47jyEQ7rvXQXKpnnoVHNfxcAAAAHB9XOWdFkTnP87FkR3qiOR87kF+qSrtD4dZGFH54lXOGSRarZNjJRAMAIMQaVc45d+5cLVq0qEb7okWL9Kc//emEB9ViVDZ8TTRXEC3jeIJonpsLRMRIDc1gAwAAwPGzJUoW5/S5ls0F2sbZFB5mkcOQDhaWNe5z3Jloziw01yZVrIkGAEBINSqI9sILL6hv37412k866SQ9//zzJzyoFuM4NhbYfagRmWheQTQ2FQAAAAiosDApKsl8XUtJpzXMotQEMxstO6+0cZ9jr7bDuyuYZieIBgBAKDUqiJaTk6P09PQa7e3atVN2dvYJD6pFcDgkh3PdinrKOT/5MVf/25wjSeqbHt/wz/AKosXW3g8AAAD+EePaXMB3JpokpTlLOrOP+SmIFuZcgYUgGgAAIdWoIFpGRoa++OKLGu1ffPGFOnTocFzXWrBggbp166aoqCgNHTpUq1atqrP/kiVLNGjQIMXExCg9PV033XSTDh8+gd2PAsXukb5fx8YCm/fn645Xv5XDkK4Y2kkjuqc0/DPIRAMAAAgu17poJbVvLuBaFy07r6Rxn+Eu53QGzyjnBACgSWhUEO2WW27RjBkz9PLLL2v37t3avXu3Fi1apJkzZ+rXv/51g6+zdOlSzZgxQ7Nnz9b69es1evRojRs3TllZWT77f/7555o4caImT56sTZs26Y033tDatWt1yy23NOZrBFalRxCtlky0nLxS3bx4rYrK7RrZI0WPXnZyw3fmlLw3FiCIBgAAEHjR9WeiVQXR/F3OycYCAACEUqN257znnnt05MgRTZ06VeXl5n/Mo6Ki9Lvf/U733Xdfg68zb948TZ482R0Emz9/vlasWKGFCxdq7ty5NfqvWbNGXbt21bRp0yRJ3bp106233qrHH3+8MV8jsDwnOdaIGoeLyip18+K1yskvVc/2cVp4/VBFhh9nTDMyzuM15ZwAAAAB5yrnrGVNNElKTzR/3MxpTBDNMOoIolUe//UAAIDfNCoTzWKx6E9/+pMOHjyoNWvWaOPGjTpy5IgeeOCBBl+jvLxc69at05gxY7zax4wZo9WrV/s8Z+TIkdq7d6+WL18uwzB04MAB/fvf/9aFF15Y6+eUlZUpPz/f6xEUrkw0q63GrpmVdofufG29Nmfnq21cpF6edKoSo2sG2upFOScAAEBwucs5G5KJ1ohyTodHoMwVPAtzPlPOCQBASDUqiOYSFxenU089VQMGDJDNVv8OlJ4OHToku92u1NRUr/bU1FTl5OT4PGfkyJFasmSJJkyYoMjISKWlpSkpKUl/+ctfav2cuXPnKjEx0f3IyMg4rnE2miuIFh5V49BHP+bq4x9zZQsP098mDlPG8ezI6SmSck4AAICgcpdz1p6JlnYi5Zxe1QyUcwIA0JQ0qpxTktauXas33nhDWVlZ7pJOl7feeqvB16m+BphhGLWuC7Z582ZNmzZNDzzwgMaOHavs7GzdfffdmjJlil566SWf59x3332aNWuW+31+fn5wAmmujQV8bCqw50ixJGnMSWka0rlN4z/Ds5yT3TkBAAACLzrJfK4jE61DkvnjZm5BmSrtDoVbj+N36zqDaJRzAgAQSo3KRHv99dc1atQobd68WW+//bYqKiq0efNmffzxx0pMTGzQNdq2bSur1Voj6yw3N7dGdprL3LlzNWrUKN19990aOHCgxo4dqwULFmjRokXKzs72eY7NZlNCQoLXIyg8yzmryS8xU/GTGlPC6YmNBQAAAILLvSZa7UG0tnE2hYdZZHcYOlhYVms/n+weJZthzt+7w8hEAwCgKWhUEO3RRx/V008/rffee0+RkZH685//rC1btuiqq65S586dG3SNyMhIDR06VJmZmV7tmZmZGjlypM9ziouLFRbmPWSr1SrJzGBrUlyTHB+ZaHnOIFqj1kHz5LUmWiNLQgEAANBw7jXRai/ntIZZlJrQyJJOz00FXNUZrow01kQDACCkGhVE+/nnn92L+dtsNhUVFclisWjmzJn661//2uDrzJo1Sy+++KIWLVqkLVu2aObMmcrKytKUKVMkmaWYEydOdPcfP3683nrrLS1cuFA7duzQF198oWnTpum0005Thw4dGvNVAqeOTDT/BdE8d+ckiAYAAE7c3//+d73//vvu9/fcc4+SkpI0cuRI7d69O4QjayLca6LVnokmVa2Ldtw7dFbfmVOSrM6MNMo5AQAIqUYF0ZKTk1VQUCBJ6tixo3744QdJ0rFjx1RcXNzg60yYMEHz58/XnDlzNHjwYH322Wdavny5unTpIknKzs5WVlaWu/+kSZM0b948PfvssxowYICuvPJK9enT57jWYAuaoGSiUc4JAAD869FHH1V0tDmv+PLLL/Xss8/q8ccfV9u2bTVz5swQj64JcJdzHpHqqIRwBdH2HzvOHTpd5ZxWj3ki5ZwAADQJjdpYYPTo0crMzNTJJ5+sq666StOnT9fHH3+szMxMnXPOOcd1ralTp2rq1Kk+jy1evLhG25133qk777yzMcMOrgZkoiVQzgkAAJqYPXv2qGfPnpKkd955R1dccYV+85vfaNSoUTrzzDNDO7imwFXOaS+TKkpqrQbo4NdMNMo5AQBoChoVRHv22WdVWmpOCO677z5FRETo888/169+9Sv94Q9/8OsAm61K54QpPIDlnBEE0QAAgH/FxcXp8OHD6ty5s/73v/+5s8+ioqJUUnKcWVUtUWScmRnmqDDXRasliJaWaGbzndCaaC7uck6CaAAAhNJxB9EqKyv17rvvauzYsZKksLAw3XPPPbrnnnv8PrhmzV3O6SuIZq5nwcYCAACgqTnvvPN0yy23aMiQIdq2bZt7HdxNmzapa9euoR1cU2CxmNloRbnmumiJnXx2c2WiZef5oZzTFVAjiAYAQEgd95po4eHhuu2221RWdpzbdbc2tZRzGoahfFcmWow/g2isiQYAAE7cc889pxEjRujgwYN68803lZKSIklat26drrnmmhCPronwXBetFn7dWMC1JhrlnAAAhFSjyjmHDx+u9evXuzcAgA+1bCxQWuFQud0hyd+ZaATRAADAiUtKStKzzz5bo/3hhx8OwWiaKNe6aCW1B9HSneWcBwrKZHcYsoZZGnZtdxDNMxMt3PsYAAAIiUYF0aZOnarf/va32rt3r4YOHarY2Fiv4wMHDvTL4Jq1WjLRXOuhWcMsio20nthnWCPMXynt5d4BNQAAgEb64IMPFBcXp9NPP12SmZn2t7/9Tf3799dzzz2nNm3ahHiETUC0MxOt5GitXdrF22QNs8juMHSwoMydmVYvdzmnj40F7JWNGCwAAPCX4y7nlKQJEyZo586dmjZtmkaNGqXBgwdryJAh7mfI3LFJqpGJ5rmpgMXSwF8k6+IKnpGJBgAA/ODuu+9Wfn6+JOn777/Xb3/7W11wwQXasWOHZs2aFeLRNRExzkBiHeWc1jCLUuPNH1P3H8+6aHWVc5KJBgBASDUqE23nzp3+HkfLU+maAPnORDvhUk6XoTdJe9dK7fr553oAAKBV27lzp/r37y9JevPNN3XRRRfp0Ucf1bfffqsLLrggxKNrItzlnLVnoknmumj780qPb100nxsLsCYaAABNQaOCaKyF1gCVzslSuO8gWoK/gmjnPuif6wAAAEiKjIxUcXGxJOnDDz/UxIkTJUnJycnuDLVWrwHlnJKUnhQtZR1T9nEF0XxkormCaJRzAgAQUo0Kor3yyit1HndNtlo198YCAc5EAwAA8KPTTz9ds2bN0qhRo/T1119r6dKlkqRt27apU6dOIR5dE9HATLT0BHMdtOxjlHMCANASNCqINn36dK/3FRUVKi4uVmRkpGJiYgiiSfVuLEAQDQAANEXPPvuspk6dqn//+99auHChOnbsKEn673//q/PPPz/Eo2siYpyZaHWsiSY5M9EkZedTzgkAQEvQqCDa0aM1f3X76aefdNttt+nuu+8+4UG1CO5MtNo2FmjUrQcAAAiozp0767333qvR/vTTT4dgNE1UQzPREv2UieYu5ySIBgBAKPktktOrVy899thjuv766/Xjjz/667LNVy2ZaPlkogEAgCbObrfrnXfe0ZYtW2SxWNSvXz9dcsklslqtoR5a0+BeE63uTLQ0ZxDt+DYW8BVEc74miAYAQEj5NR3KarVq//79/rxk82V3BtFqzUQjiAYAAJqe7du364ILLtC+ffvUp08fGYahbdu2KSMjQ++//7569OgR6iGGXozHxgKGIVksPrt1SDTLOQ8UlMnuMGQN893Pi69yzjDnlJ1yTgAAQqpRQbRly5Z5vTcMQ9nZ2Xr22Wc1atQovwys2at0/YrImmgAAKD5mDZtmnr06KE1a9YoOdkMFh0+fFjXX3+9pk2bpvfffz/EI2wCXOWcjkqprECKSvDZrV28TdYwi+wOQwcLytyZaXWinBMAgCarUUG0Sy+91Ou9xWJRu3btdPbZZ+upp57yx7iav0pn2j67cwIAgGZk5cqVXgE0SUpJSdFjjz3Gj6UuEdFSeLRUWWKWdNYSRLOGWZQab9P+vFJl55WcQBCNck4AAJqCRgXRHA6Hv8fR8rg3FvAdREsgiAYAAJogm82mgoKCGu2FhYWKjIz0cUYrFd1GKigxSzrbdK21W1pilDOIVqohDbmuz3JOVyZaeWNHCwAA/CAs1ANosWrZWIBMNAAA0JRddNFF+s1vfqOvvvpKhmHIMAytWbNGU6ZM0cUXXxzq4TUdrnXRiuveXCDduS5adkM3F6irnNNReTwjBAAAftaoINoVV1yhxx57rEb7E088oSuvvPKEB9UiuDPR2FgAAAA0H88884x69OihESNGKCoqSlFRURo5cqR69uyp+fPnh3p4TYdrXbSSo3V2S3fv0FnSsOu6g2gec0XWRAMAoEloVDnnypUr9eCDD9ZoP//88/Xkk0+e8KBaBB+ZaKUVdpVXmqWwBNEAAEBTlJSUpP/85z/avn27tmzZIsMw1L9/f/Xs2TPUQ2taGhhEc62Dtr/BmWiuck6PH2Ip5wQAoEloVBCttjUxIiIilJ+ff8KDahHsziCax5poriw0a5hFcbZG3XoAAAC/mzVrVp3HP/30U/frefPmBXg0zUQDyzk7JJnlnDmUcwIA0Ow1KpIzYMAALV26VA888IBX++uvv67+/fv7ZWDNXmXNCZB7U4GocFksllCMCgAAoIb169c3qB/zFw/HmYmWfcwf5ZxkogEAEEqNCqL94Q9/0OWXX66ff/5ZZ599tiTpo48+0muvvaY33njDrwNsturIRKOUEwAANCWffPJJqIfQ/EQ7M9FK6ttYwAyiHSgok91hyBpWTyCyznJO1kQDACCUGrWxwMUXX6x33nlH27dv19SpU/Xb3/5We/fu1YcffqhLL73Uz0Nspip9BNGKCaIBAIDWZcGCBerWrZuioqI0dOhQrVq1qkHnffHFFwoPD9fgwYMDO8DGamAmWvv4KFnDLLI7DB0qLKv/upRzAgDQZDV6Ya4LL7xQF154oT/H0rL42FjAXc5JEA0AALQCS5cu1YwZM7RgwQKNGjVKL7zwgsaNG6fNmzerc+fOtZ6Xl5eniRMn6pxzztGBAweCOOLj0MA10axhFrWPtyk7r1T7j5UoNSGq7uu6M9Eo5wQAoKlpVCba2rVr9dVXX9Vo/+qrr/TNN9+c8KCaPYddMuzma8o5AQBAKzVv3jxNnjxZt9xyi/r166f58+crIyNDCxcurPO8W2+9Vddee61GjBgRpJE2QgMz0aSqddEatLmAz0w052vKOQEACKlGBdFuv/127dmzp0b7vn37dPvtt5/woJq9So9UfV8bCxBEAwAALVx5ebnWrVunMWPGeLWPGTNGq1evrvW8l19+WT///LMefPDBej+jrKxM+fn5Xo+gaeCaaJKU0SZGkvRTbmH91/UVRAtzFo8QRAMAIKQaFUTbvHmzTjnllBrtQ4YM0ebNm094UM2e3SOIRiYaAABohQ4dOiS73a7U1FSv9tTUVOXk5Pg856efftK9996rJUuWKDy8/lVH5s6dq8TERPcjIyPDL2NvEFc5Z8kxyeGos+uwrmbW2tc76w+41VnO6SCIBgBAKDUqiGaz2XyuT5Gdnd2gCU+LV+lar8JS9cuhpHyCaAAAoJWxWLx3ozQMo0abJNntdl177bV6+OGH1bt37wZd+7777lNeXp774atSImCikpwvDKn0WJ1dh3dLkSSt231UFfa6A26UcwIA0HQ1Koh23nnnuSctLseOHdP999+v8847z2+Da7bsHjtzekwSyUQDAACtRdu2bWW1WmtkneXm5tbITpOkgoICffPNN7rjjjsUHh6u8PBwzZkzRxs3blR4eLg+/vjjGufYbDYlJCR4PYImPFKKjDdf17MuWq/2cWoTE6GSCru+25tXZ1/f5ZzOuaNhrzfrDQAABE6jgmhPPfWU9uzZoy5duuiss87SWWedpW7duiknJ0dPPfWUv8fY/PjYmVMiiAYAAFqPyMhIDR06VJmZmV7tmZmZGjlyZI3+CQkJ+v7777Vhwwb3Y8qUKerTp482bNig4cOHB2voDdfAzQXCwiw6rZtZ/vnVzsN1X7Ouck6Jkk4AAEKoUbWXHTt21HfffaclS5Zo48aNio6O1k033aRrrrlGEREEiNxBtHCCaAAAoPWaNWuWbrjhBg0bNkwjRozQX//6V2VlZWnKlCmSzHLMffv26ZVXXlFYWJgGDBjgdX779u0VFRVVo73JiGkj5WVJxfWvdTa8W4pWbDqgr3Yc0dQz6+jos5wzwvt4tTkmAAAIjkYvYBYbG6vTTz9dnTt3Vnm5+R/7//73v5Kkiy++2D+ja67sBNEAAAAmTJigw4cPa86cOcrOztaAAQO0fPlydenSRZK5nm5WVlaIR3kCGpiJJknDu5uZaN/sOqJKu0Ph1loKQuoq55RYFw0AgBBqVBBtx44duuyyy/T999/LYrHUWCDWbrf7bYDNUqWPyY8IogEAgNZn6tSpmjp1qs9jixcvrvPchx56SA899JD/B+Uv0a4dOqtlou1cJaX0lBLS3U190xKUEBWu/NJKbdqfr0EZSb6v6aucM8wqySLJkByV/ho9AAA4To1aE2369Onq1q2bDhw4oJiYGP3www9auXKlhg0bpk8//dTPQ2yGfGSilVbYVVZpLgSbQBANAACg+YtxBdE8MtF+eFP6+0XSO7d5dbU2dF00X5loFktVUM11HAAABF2jgmhffvml5syZo3bt2iksLExWq1Wnn3665s6dq2nTpvl7jM2Pj0y0fGcWmsUixdsaXUULAACApsJVzulaE81eKX3yqPl6/7eSYXh1H94tRZL01Y5a1lBzOKoyzapVNLjfU84JAEDINCqIZrfbFRcXJ8ncvnz//v2SpC5dumjr1q3+G11z5SMTzVXKmRAVobAwi6+zAAAA0JxEV8tE+/4N6fB283VpnlTsnXHmWhft611HZHd4B9gkee+8aa1WuRDm/BGWIBoAACHTqJSoAQMG6LvvvlP37t01fPhwPf7444qMjNRf//pXde/e3d9jbH5cu3N6/ILIemgAAAAtjHtjgSNmcGvlY97HD2+XYtu63/ZPT1CcLVwFpZXakp2vAR0Tvft7lmrWyERzziEdBNEAAAiVRmWi/f73v5fDYa7v9cgjj2j37t0aPXq0li9frmeeecavA2yWXEG08Ch3E0E0AACAFsa1JlrxEWnja9LRXVJMWyljuNl+6Cev7uHWMA3ragbevtrpo6TTM8uMck4AAJqcRmWijR071v26e/fu2rx5s44cOaI2bdp47dLZavko58wvJYgGAADQorgy0QpzpZVPmK9Pn2kG0/Z8VVXa6eG0bsn6dOtBfbXjsCaf3s37oCsTzRLm3JHTA+WcAACEXKMy0XxJTk4mgObiY2OBvGKCaAAAAC2Ka020gv1SXpYUlyoNu1lq28ts9xFEc20u8PWuI3JUXxfN186cLq42yjkBAAgZvwXR4MHnxgLmTksJBNEAAABaBlc5p8vo30qRMVJKD/O9jyDawE6Jio6w6lhxhbblFngfdGWZ+QyiRXj3AQAAQUcQLRB8ZaKxJhoAAEDLEpUoyVmJkdBROuVG83WKMxPtyA7JYfc6JcIapqFdnOui7ai2Lpo7E83HfJFyTgAAQo4gWiD4zEQjiAYAANCihFmrstFG/1aKcG4qldhJstrMoNixrBqnDe+WLIsc+npHrvcByjkBAGjSGrWxAOpRWWo+k4kGAADQsp03R8r5QRpyQ1VbmFVK7i4d3CId/llK9t5A4BddE7Q88j4Z2yNkOL6RxbWJgLuc08d80V3OWR6ALwEAABqCIFoguMo5w6PcTfkE0QAAAFqeIdf7bm/b0xlE2y71Otfr0MDwPbKF7ZEkHT6Uo5T2Hc0DdWaisSYaAAChRjlnIFDOCQAA0Lql9DSfD/9U45AtZ5379cHsvVUH6gqihTnnkI5Kf40QAAAcJ4JogcDGAgAAAK2ba3MBHzt0as9X7pfHDu6raqecEwCAJi3kQbQFCxaoW7duioqK0tChQ7Vq1apa+06aNEkWi6XG46STTgriiBuATDQAAIDWzZ2J9nPNY3u+dr8sOrK/qp1yTgAAmrSQBtGWLl2qGTNmaPbs2Vq/fr1Gjx6tcePGKSur5i5GkvTnP/9Z2dnZ7seePXuUnJysK6+8Msgjr0elM4jmnACVVzpUUmFub04QDQAAoBVwBdHy9kjlxVXtefvMNqfyvANVxxpSzkkQDQCAkAlpEG3evHmaPHmybrnlFvXr10/z589XRkaGFi5c6LN/YmKi0tLS3I9vvvlGR48e1U033VTrZ5SVlSk/P9/rEXCuCZAzE82VhWaxSPFR7OUAAADQ4sWmSNFtzNdHdlS17/3aq5ujMLfqTZ3lnM7AmoMgGgAAoRKyIFp5ebnWrVunMWPGeLWPGTNGq1evbtA1XnrpJZ177rnq0qVLrX3mzp2rxMRE9yMjI+OExt0glaXms9U7iBZvC1dYmCXwnw8AAIDQ87W5wJ61kiTDYpUkhRcfqjrmDqL5KucM9+4DAACCLmRBtEOHDslutys1NdWrPTU1VTk5OfWen52drf/+97+65ZZb6ux33333KS8vz/3Ys2dPnf39wrWxQLg5AXKvhxZDKScAAECr4WtzAeemAiUdhkuSossPyzAM8xjlnAAANGkhry20WLwzswzDqNHmy+LFi5WUlKRLL720zn42m002m63OPn7n3lggSpKUz6YCAAAArU9KD/PZtblARYmUvVGSFN7/YmnfaiUZx5RfUmn+2Eo5JwAATVrIMtHatm0rq9VaI+ssNze3RnZadYZhaNGiRbrhhhsUGenjl7pQq/T+FZGdOQEAAFohVznnIWc55/4NZhAstr0iu5wmSWprydOeo86NBxq0O2d54MYLAADqFLIgWmRkpIYOHarMzEyv9szMTI0cObLOc1euXKnt27dr8uTJgRxi47kz0bzXRCOIBgAA0Iq0dZVz/iQZhruUUxmnSXHtJEkpyte+GkE0X5loriBaZQAHDAAA6hLScs5Zs2bphhtu0LBhwzRixAj99a9/VVZWlqZMmSLJXM9s3759euWVV7zOe+mllzR8+HANGDAgFMOunzsTjSAaAABAq5XcXZJFKs2Tio9Ie81NBZQxXIptL0myWSp18GCupPS6NxZwrYlGOScAACET0iDahAkTdPjwYc2ZM0fZ2dkaMGCAli9f7t5tMzs7W1lZWV7n5OXl6c0339Sf//znUAy5YdyZaN7lnAkE0QAAAFqPiGgpMUPKyzKz0dyZaMOliCiVWmMVZS9S3qH9kgZRzgkAQBMX8o0Fpk6dqqlTp/o8tnjx4hptiYmJKi4uDvCoTlClM4hGJhoAAEDrltLDDKL9lCkVHTQzytIHSZLKbCmKKi5SydFssy/lnAAANGkhWxOtRav0nYlGEA0AAKCVcW0usPE187nDYCnC3MHdEWOui1aRf8A81pByTjLRAAAIGYJogeAu5zQnSATRAAAAWinX5gL5+8znjOHuQ+Hx5rpoKso1n+ss53S2sSYaAAAhQxDN3+yVkuEwXzsnO/kE0QAAAFqnlB7e7zNOc7+0JaVJkmIqjqqwrLKeck7nKix2gmgAAIQKQTR/c2WhSVI4a6IBAAC0aim9vN93qgqiRSaaQbR2ytO+oyUNLOckiAYAQKgQRPO3So8gGhsLAAAAtG6JndxzQiV2lhLSq47FmmuitbXkad+xYso5AQBo4gii+Ztr8mMJk6zhqrA7VFxul0QQDQAAoNUJs0rJ3c3XHqWckqQ4c020thZXJhrlnAAANGUE0fzNlYlWLQtNkuKjCKIBAAC0Oh0Gm8/dRnu3xzqDaMrT3vrKOV1tBNEAAAiZ8FAPoMVxBdHCzYmOK4gWHxUua5glVKMCAABAqJw3R+pxjjTgcu/22LaSpLaWfO09VlJ3OadrTTTKOQEACBky0fzNtbFAeJQk1kMDAABo9eLaSwOvlMLCarZLirGU6dCRox6ZaHWVc5YHcKAAAKAuBNH8rdL1CyKbCgAAAKAOkXFyWM0fXkuPZjdsYwF7ZZAGBwAAqiOI5m9273LOfIJoAAAA8MVikeHMRrMWH5SjsgHlnGSiAQAQMgTR/K2WjQUIogEAAKC6MI8dOisqXPNIX+WczsAaa6IBABAyBNH8zfXroGtjgWKCaAAAAPDN4g6i5cvuDqL5Kuc010Q7Vliif6zZHazhAQAADwTR/I1MNAAAADSUa4dO5clRVxDNWc5ZWVGmx5ZvkcNhBGuEAADAiSCav1V6r4nmCqIlEEQDAABAdbFmJlqKJU+Ge2OB2ss5I1SponK79hwtDtYIAQCAE0E0f7P7zkQjiAYAAIAaPNZEk9253pmPTLQyw5y2R8guSdqSnR+c8QEAADeCaP7mzkSjnBMAAAD1iG0nyVwTLcxRexDt690FkqRwVUqSNmcXBGd8AADAjSCav7k3FiCIBgAAgHq4MtGUJ6vhCqLVnDd+9NNRSVKkxS7JIBMNAIAQIIjmb9U2FsgniAYAAFqxBQsWqFu3boqKitLQoUO1atWqWvu+9dZbOu+889SuXTslJCRoxIgRWrFiRRBHGwLONdHaWfIUbphZZtUz0SrsDn207Yj7fbjsBNEAAAgBgmj+5s5E895YgCAaAABobZYuXaoZM2Zo9uzZWr9+vUaPHq1x48YpKyvLZ//PPvtM5513npYvX65169bprLPO0vjx47V+/fogjzyInLtzJliKFW5xmG3Vgmhf7zyiQyVVu3GGy669R0uUX1oRtGECAACCaP7nkYlWYXeoqNxc/JUgGgAAaG3mzZunyZMn65ZbblG/fv00f/58ZWRkaOHChT77z58/X/fcc49OPfVU9erVS48++qh69eqld99912f/srIy5efnez2aneg2Uli1eWK1cs4PfshRpcLd7zMSzNdbc1gXDQCAYCKI5m+VpeZzeKS7lFOSEqLCazkBAACg5SkvL9e6des0ZswYr/YxY8Zo9erVDbqGw+FQQUGBkpOTfR6fO3euEhMT3Y+MjIwTHnfQWSzuzQXcPDLRHA5DKzblqEJWd1v/1GhJ7NAJAECwEUTzN1c5p9XmLuWMs4Ur3MqtBgAArcehQ4dkt9uVmprq1Z6amqqcnJwGXeOpp55SUVGRrrrqKp/H77vvPuXl5bkfe/bsOeFxh0RctSCaR2ba+j3HlFtQpnhbhIww80fZPu2jJBFEAwAg2EiP8jdXOWd4FOuhAQCAVs9isXi9NwyjRpsvr732mh566CH95z//Ufv27X32sdlsstlsfhlnSMVWfb9KWXWooFxpiWag7IMfsiVJZ/drL8v2CMlRqT5toyTla3M25ZwAAAQT6VH+5rGxgCuIlkAQDQAAtDJt27aV1WqtkXWWm5tbIzutuqVLl2ry5Mn617/+pXPPPTeQw2waPMo5y41wXfLc5/pu7zEZhqEPNpn37/yT0txlnr3amoHDrTn5sjuMmtcDAAABQRDN34bdLF36vNTzPOWXmtuUJ0aT8AcAAFqXyMhIDR06VJmZmV7tmZmZGjlyZK3nvfbaa5o0aZJeffVVXXjhhYEeZtPgUc5pt4TrQH6Zrnz+Sz394U/ac6REtvAwndGnnWQ155Qd48MVFRGm0gqHdh0uCtWoAQBodQii+VvGadLga6TU/pRzAgCAVm3WrFl68cUXtWjRIm3ZskUzZ85UVlaWpkyZIslc02zixInu/q+99pomTpyop556Sr/4xS+Uk5OjnJwc5eXlheorBIdHOWdsTIzO6tNOZZUOPfPRT5KkM3q3U0xkuHutNKtRqT6p8ZKkHynpBAAgaAiiBVA+QTQAANCKTZgwQfPnz9ecOXM0ePBgffbZZ1q+fLm6dOkiScrOzlZWVpa7/wsvvKDKykrdfvvtSk9Pdz+mT58eqq8QHHFVQbSw8Ei9eOOp+vXobu628wekmS9cu3Y6KtQvPUESmwsAABBM1BkGEJloAACgtZs6daqmTp3q89jixYu93n/66aeBH1BT5LEmmqwRsoZZNPvC/hrYKUkb9xzTRQM7OI85p+72CvVLT5FEEA0AgGAiiBZAecUE0QAAAFAPj0w0d7aZpPGDOmj8oA41j9nJRAMAIBQo5wwgMtEAAABQL69MtMja+znXRJO9XH3TzTXR9ueV6lhxeQAHBwAAXAiiBZAriJZAEA0AAAC1iUmRLM5pubWOeaOrnNNRqYSoCHVqEy1J2sLmAgAABAVBtAAiEw0AAAD1CrOagTSp7kw0j3JOSeqbRkknAADBRBAtgAiiAQAAoEFineuiNbCcU5L6O0s6f8whiAYAQDCwsUAA5RNEAwAEkd1uV0VFRaiH0WxFRETIarWGehhoreLaSbmqp5zTecxRKUkemwtQzgkAQDAQRAsQu8NQQZk5wSGIBgAIJMMwlJOTo2PHjoV6KM1eUlKS0tLSZLFYQj0UtDYNyUSzemeiuYJoWw8UqNLuULiVIhMAAAKJIFqAuLLQJDYWAAAEliuA1r59e8XExBAAagTDMFRcXKzc3FxJUnp6eohHhFbHtUNnXZlo7nJOc57ZOTlGsZFWFZXbtfNQkXqlxgd4kAAAtG4E0QLEtR5abKRVEfwqCAAIELvd7g6gpaSkhHo4zVp0tLnTYW5urtq3b09pJ4IrPtV8Do+qvU+1cs6wMIv6pMXr26xj2pydTxANAIAAI7oTIGwqAAAIBtcaaDExMSEeScvguo+sLYegO+lXUp8LpKGTau9TrZxTkvqlxStVR7RpX15gxwcAAAiiBYoriEYpJwAgGCjh9A/uI0ImKUO65jWp6+m193Gtl2avCvJeYX9fX0XdoditbwV4gAAAgCBagJCJBgAAAL8Kc67E4qgKovU78pH5fOxTlVc6QjEqAABaDYJoAUIQDQCA4Onatavmz58f6mEAgWX13lhAFaWy5W6UJA2xbNPm/ZR0AgAQSATRAoQgGgAAdTvzzDM1Y8YMv1xr7dq1+s1vfuOXawFNVvVyzv3rZXGuj9bOkqetWzeFaGAAALQOBNECJJ8gGgAAJ8QwDFVWVjaob7t27dhcAS2fq5zTtbHAnjVeh0t+/iLIAwIAoHUhiBYgZKIBAELFMAwVl1eG5GEYRoPGOGnSJK1cuVJ//vOfZbFYZLFYtHjxYlksFq1YsULDhg2TzWbTqlWr9PPPP+uSSy5Ramqq4uLidOqpp+rDDz/0ul71ck6LxaIXX3xRl112mWJiYtSrVy8tW7bMn7cZCD5XJprDGVzOMoNolRFxkqS43PWhGBUAAK1GeKgH0FK5g2gxBNEAAMFVUmFX/wdWhOSzN88Zq5jI+qcXf/7zn7Vt2zYNGDBAc+bMkSRt2mSWot1zzz168skn1b17dyUlJWnv3r264IIL9MgjjygqKkp///vfNX78eG3dulWdO3eu9TMefvhhPf7443riiSf0l7/8Rdddd512796t5ORk/3xZINg810RzONxBNMcpk6SvnlWfyi3KzitRemJ06MYIAEALFvJMtAULFqhbt26KiorS0KFDtWrVqjr7l5WVafbs2erSpYtsNpt69OihRYsWBWm0DUcmGgAAtUtMTFRkZKRiYmKUlpamtLQ0Wa1WSdKcOXN03nnnqUePHkpJSdGgQYN066236uSTT1avXr30yCOPqHv37vVmlk2aNEnXXHONevbsqUcffVRFRUX6+uuvg/H1gMDwLOc8tFUqPSZFxCjyF7+WJPWzZGnjz/tDNz4AAFq4kGaiLV26VDNmzNCCBQs0atQovfDCCxo3bpw2b95c6y/LV111lQ4cOKCXXnpJPXv2VG5uboPXSwkmVxAtgSAaACDIoiOs2jxnbMg++0QNGzbM631RUZEefvhhvffee9q/f78qKytVUlKirKysOq8zcOBA9+vY2FjFx8crNzf3hMcHhIxnOWfWl+brTsOkNl2VF9FOiRUHdeDHL6VTeoRujAAAtGAhDaLNmzdPkydP1i233CJJmj9/vlasWKGFCxdq7ty5Nfp/8MEHWrlypXbs2OEuxejatWswh9xgZKIBAELFYrE0qKSyqYqNjfV6f/fdd2vFihV68skn1bNnT0VHR+uKK65QeXl5ndeJiPD+b7DFYpHD4fD7eIGgcZdzlrtLOdV5hCSpsN0pSty/QmH7vpZ0fWjGBwBACxeycs7y8nKtW7dOY8aM8WofM2aMVq9e7fOcZcuWadiwYXr88cfVsWNH9e7dW3fddZdKSkpq/ZyysjLl5+d7PYKBIBoAAHWLjIyU3W6vt9+qVas0adIkXXbZZTr55JOVlpamXbt2BX6AQFPjuSaaKxOt8y8kSTHdzWBax4LvVVpR//+uAADA8QtZEO3QoUOy2+1KTU31ak9NTVVOTo7Pc3bs2KHPP/9cP/zwg95++23Nnz9f//73v3X77bfX+jlz585VYmKi+5GRkeHX7+GL3WGooNQsMU2IIogGAIAvXbt21VdffaVdu3bp0KFDtWaJ9ezZU2+99ZY2bNigjRs36tprryWjDK1TmHNeeSzLfFjCpE6nSpKS+pwuSRps2aYf9h4L0QABAGjZQr6xgMVi8XpvGEaNNheHwyGLxaIlS5botNNO0wUXXKB58+Zp8eLFtWaj3XfffcrLy3M/9uzZ4/fvUF1BaYX7NZloAAD4dtddd8lqtap///5q165drWucPf3002rTpo1Gjhyp8ePHa+zYsTrllFOCPFqgCXBlouV8Zz6nDpBs8ZIkS/ogVVgilGwp1M9bN4ZogAAAtGwhWzClbdu2slqtNbLOcnNza2SnuaSnp6tjx45KTEx0t/Xr10+GYWjv3r3q1atXjXNsNptsNpt/B18PVylndIRVkeEhj1MCANAk9e7dW19++aVX26RJk2r069q1qz7++GOvtupZ6NXLOw3DqHGdY8eONWqcQJPh2ljAcGZiOtdDkySFR+pwfH+l5W9Uyc9rJJ0d9OEBANDShSzCExkZqaFDhyozM9OrPTMzUyNHjvR5zqhRo7R//34VFha627Zt26awsDB16tQpoOM9HqyHBgAAAL8Lq/b7t3M9NBcj4zRJUvzh9T4DyQAA4MSENE1q1qxZevHFF7Vo0SJt2bJFM2fOVFZWlqZMmSLJLMWcOHGiu/+1116rlJQU3XTTTdq8ebM+++wz3X333br55psVHR0dqq9RA0E0AAAA+J0rE82lWhAtpa+5Llr/yi3ae7T2jbcAAEDjhKycU5ImTJigw4cPa86cOcrOztaAAQO0fPlydenSRZKUnZ3ttT5KXFycMjMzdeedd2rYsGFKSUnRVVddpUceeSRUX8EngmgAAADwO6vH3DKps5TQwetwZFczqNbHslf//XmPMpL7BHN0AAC0eCENoknS1KlTNXXqVJ/HFi9eXKOtb9++NUpAmxpXEC2BIBoAAAD8xbOc03M9NJf4NB2NTFeb8mwd+vEL6VSCaAAA+FPIg2gtEZloAAAA8DvPcs5qpZwuRe2Hqs3e9xS27xvtOHiVsvNKte9YiQ4Vlun8k9LUvV1ckAYLAEDLQxAtAAiiAQAAwO88yzl9ZaJJius5Qtr7njKKftDZT630Opa5+YDenjoqkCMEAKBFC+nGAi1VPkE0AAAA+FukM4ssuo3U1nepZmIvM0h2SthPSomsUM/2cRrdq60sFml91jHtP8aGAwAANBZBtACoykQj0Q8AAAB+0r6fdPYfpMv+KoX5nsZb0gbIiEpSgqVY37T5vT68qFT/mDxcw7q0kSSt2JQTzBEDANCiEOUJgPySSklSYgyZaAAAAPATi0X65V1197FGyHLVK9J/bpclb4/06lVSv/G6rMftWrtL+uj7LN3Uu1w6/LOUt0cqPuzxOCJFJUjpg6T0IeZzXDvJMKSSo1JBtpSfLVnDpS6jvMtLAQBoBQiiBQBrogEAEHhdu3bVjBkzNGPGDEmSxWLR22+/rUsvvdRn/127dqlbt25av369Bg8eHLRxAkHX/Qxp6hpp5Z+kL5+TtryrqyM+0i9tMeqQfVh6zqj7/C3vVr2ObSeVFUqV1cpAY9pKJ18pDb5GShtoBvgAAGjhCKIFAEE0AACCLzs7W23atAn1MICmwRYnjfmjNHCC9N5Mhe39Wp0sxZKkivBYRbTrKbXpagbJopOlmBQpJlkqzJWyN0j7N0iHt0tFB6uuGZ0sJXSQCg+Y7V8tNB/t+0sdT5Ei4yWb8xGVKCVlmJ+RmFF/1pphSEd3SaXHpNSTzWw3AACaGP7rFAAE0QAACL60tLRQDwFoetIGSDevkLJW680NOXp0TbkG9Oqhv08eXv+5ZQXSoZ+k6CQpPl2KiDbb7RXSzx9LG1+Tflwu5W42H7WxhEmJnaTEzmZ5aFyqM3jXRjqyQ8reKGV/J5Xlmf2jk6XeY6U+F0g9zjYDggAANAEE0fzM4TCUX2oG0RIIogEAQsEwpIri0Hx2REyDyrpeeOEFzZkzR3v27FGYxwLpF198sdq0aaMHHnhAs2bN0po1a1RUVKR+/fpp7ty5Ovfcc2u9ZvVyzq+//lq33nqrtmzZogEDBmj27Nkn/PWAZiksTOp6ugbHFurwmpVaveOw8koq6v/B1xZvZphVZ40wg1y9x0olx6St/5Xy95lBt/JC87n4iHQsSzq2W6osdb7OqvvzrJFSeLRUcsQM0G18zWxLO9kZhMswH/FpZmDOUSk57JJhN8cUlyrFpUnxqebYAQDwM4JoflZQVinDucwEmWgAgJCoKJYe7RCaz75/vxQZW2+3K6+8UtOmTdMnn3yic845R5J09OhRrVixQu+++64KCwt1wQUX6JFHHlFUVJT+/ve/a/z48dq6das6d+5c7/WLiop00UUX6eyzz9Y///lP7dy5U9OnTz/hrwc0Zz3axalX+zj9lFuoj388oMuGdDrxi0Ynmeui1cbhMMs/j+6SCvab5aKFB8zn4sNSUmfnRgaDpHZ9JVmkPWvMwNyP70tHd0r71pmP4xERK8WmmNlu0W3M7LaoBPP6nsKjzO/g6heVZGbcRUSbxzyfI2KkcBvrvwFAK0YQzc/ynaWcURFhsoVbQzwaAACapuTkZJ1//vl69dVX3UG0N954Q8nJyTrnnHNktVo1aNAgd/9HHnlEb7/9tpYtW6Y77rij3usvWbJEdrtdixYtUkxMjE466STt3btXt912W8C+E9AcnD8gTT99vF0f/JDjnyBafcLCpIR089FQXU83H2MeMctJD22T8vaau4nm7ZUKcsxAlsUqhTkfFaXO4NwBMxuuokg6VlR/9ttxsziDaZFmlpw10syCc70OjzIDbeE2KSzCGXCzOMdrMdvCbc6+NiksvCqbzpVZZwmr6uPZzxphnm+NcH7v8KqHJcy8vlFt0wiLxXnM+XCNxes5rGp8nu1e13GeH2b1vlaN21P9PI8/J/efV7j3+GWRZEiGwzl+w+OeeY65lj8Pr3E39I/R4574+r7uPg0Zg6rOtxznOAA0OwTR/Iz10AAAIRcRY2aEheqzG+i6667Tb37zGy1YsEA2m01LlizR1VdfLavVqqKiIj388MN67733tH//flVWVqqkpERZWQ37B/GWLVs0aNAgxcRUjWfEiBHH/XWAlub8AWn6y8fbtXLbQRWXVyomsgn/c8Bikdr1Nh/Ho6zQDKYVH5FKjlY9yvJr9q0o8e5Tesxsqyg1y1ArS83sXkel8wTDDNBVFJ3ot0NLViPQ6GtHXF8BQI9gnDvAGVZ7YK4hQVLX51cPsLo/x+Nza/Bxnmfg0fX5XsHV2nh+N3kETKv18Qxcuk+r9n287kctgVT3uI1q72sbXm33oq7zazlWPUBd2/31HLfrfng9e3xGzZOrjbu+Yw24T65AtuEwHzXGWpdqf3d9vvZQ/fs2lC1BuuGthvcPgCb8X83miSAaACDkLJYGlVSG2vjx4+VwOPT+++/r1FNP1apVqzRv3jxJ0t13360VK1boySefVM+ePRUdHa0rrrhC5eXlDbq2Ud9kGWil+qcnKCM5WnuOlOizbQd1/oDjyBBrLmxx5iOlh/+uaa8wg2uVpVJ5kWQvN9vcz2VSZbnzudT5ulze/yh2mH0ry6rOsVdUZWVZrGbmniHndar1c1RWfWb17DV3kE8e/3B1frbnP4oNh2r+Q93zH88e/6D1+se1YX6e6xoOewNumvM891hdzw7nuCslR4XZp3rWlzzHXP0f2XUFppoIX0EIACcuOvS7sBNE8zOCaAAANEx0dLR+9atfacmSJdq+fbt69+6toUOHSpJWrVqlSZMm6bLLLpMkFRYWateuXQ2+dv/+/fWPf/xDJSUlio42dxRcs2aN378D0NxYLBadf1Ka/rZqp/77Q07LDKIFgtVZRqmEUI8E1VUP/lU/ViPDqFpQsXqwyzPA6BV0bMA4fGXz+MrgqR6gdL12H/e4huQs863lOjW+Sy2lsa574JUJVv1zPdt8ZH95fUdnP4dHcNWoI7ha/dou1TPTagRyjarP8mzzdU71P2+fGX71jc2jrUHn+2qr/mdQ22d6fl9fpdb1ZH81ODvOR+Dc15+Br4zG6tdsyD2s8bra53l9bi0l5LWxhj7OQhDNzwiiAQDQcNddd53Gjx+vTZs26frrr3e39+zZU2+99ZbGjx8vi8WiP/zhD3I4Gv6r/rXXXqvZs2dr8uTJ+v3vf69du3bpySefDMRXQD0WLFigJ554QtnZ2TrppJM0f/58jR49utb+K1eu1KxZs7Rp0yZ16NBB99xzj6ZMmRLEEbd85w9I199W7dTHW3JVVmlnHV80b41ah4y/8wAap66iZTSCK4iWQBANAIB6nX322UpOTtbWrVt17bXXutuffvpptWnTRiNHjtT48eM1duxYnXLKKQ2+blxcnN59911t3rxZQ4YM0ezZs/WnP/0pEF8BdVi6dKlmzJih2bNna/369Ro9erTGjRtX69p2O3fu1AUXXKDRo0dr/fr1uv/++zVt2jS9+eabQR55yzYkI0nt420qKKvUmKc/00PLNumTrbkqrWhIiR4AAK2XxWhli4bk5+crMTFReXl5Skjwfzr2nz74UQs//Vk3jeqqB8ef5PfrAwDgqbS0VDt37lS3bt0UFRUV6uE0e3Xdz0DPIVqi4cOH65RTTtHChQvdbf369dOll16quXPn1uj/u9/9TsuWLdOWLVvcbVOmTNHGjRv15Zdf1vt5/Bk13OIvdur/lm9Rhb3qnwK28DD1S09QaoJNqQlRah9vU/v4KMXYrLKFWxUVEaaoCKsirWGyhlkUZrEoLEyyWiwKc763WizmJpRhFlnkTBKSpapayPl/XG3uYimPTKKqNtd7z7Ij72N1aUhukuW4M5hODPs2AsCJibUFpqCyoXMIyjn9jHJOAAAAqby8XOvWrdO9997r1T5mzBitXr3a5zlffvmlxowZ49U2duxYvfTSS6qoqFBEhPf8qqysTGVlZe73+fk+dl+ET5NGddMVwzL0xfZD+nTrQX26NVfZeaXasOdYqIcGAIBPSTER2vDAmPo7BhBBND8rLDV3xiGIBgAAWrNDhw7JbrcrNTXVqz01NVU5OTk+z8nJyfHZv7KyUocOHVJ6uvci+HPnztXDDz/s34G3InG2cI09KU1jT0qTYRj6KbdQOw4WKregTAfyS5WbX6aDhWUqKbertNKhsgq7yiodKq90yO4wZDcMOTyeDUNyGM737vXAzXZDkmEYzucQf3EAABqJIJqf/fnqwXrs8pMVFuTUcAAAgKaoermcYRh1ltD56u+rXZLuu+8+zZo1y/0+Pz9fGRkZJzLcVstisah3arx6p8YH7TNdf7aeQTWj+jGf5/loq20HvHrOAwDgeBBE8zOLxaKYSG4rAABo3dq2bSur1Voj6yw3N7dGtplLWlqaz/7h4eFKSUmp0d9ms8lms/lv0AgqV2DUd0yVH6QBAE0Pu3MCANACtLJ9ggKG++g/kZGRGjp0qDIzM73aMzMzNXLkSJ/njBgxokb///3vfxo2bFiN9dAAAACCjSAaAADNmCuwUFxcHOKRtAyu+0jAxj9mzZqlF198UYsWLdKWLVs0c+ZMZWVlacqUKZLMcsyJEye6+0+ZMkW7d+/WrFmztGXLFi1atEgvvfSS7rrrrlB9BQAAADfqDgEAaMasVquSkpKUm5srSYqJialzvSn4ZhiGiouLlZubq6SkJFmt1lAPqUWYMGGCDh8+rDlz5ig7O1sDBgzQ8uXL1aVLF0lSdna2srKy3P27deum5cuXa+bMmXruuefUoUMHPfPMM7r88stD9RUAAADcLEYrq1vIz89XYmKi8vLylJCQEOrhAABwwgzDUE5Ojo4dOxbqoTR7SUlJSktL8xmIZA7R9PFnBAAAGqOhcwgy0QAAaOYsFovS09PVvn17VVRUhHo4zVZERAQZaAAAAKgVQTQAAFoIq9VKEAgAAAAIEDYWAAAAAAAAAOpBEA0AAAAAAACoB0E0AAAAAAAAoB6tbk0012ak+fn5IR4JAABoTlxzh1a2sXmzwjwPAAA0RkPnea0uiFZQUCBJysjICPFIAABAc1RQUKDExMRQDwM+MM8DAAAnor55nsVoZT+nOhwO7d+/X/Hx8bJYLH6/fn5+vjIyMrRnzx4lJCT4/fqoiXseXNzv4OOeBxf3O7ia0/02DEMFBQXq0KGDwsJYEaMpCvQ8T2pef2dbAu53cHG/g497Hlzc7+BqTve7ofO8VpeJFhYWpk6dOgX8cxISEpr8X5KWhnseXNzv4OOeBxf3O7iay/0mA61pC9Y8T2o+f2dbCu53cHG/g497Hlzc7+BqLve7IfM8fkYFAAAAAAAA6kEQDQAAAAAAAKgHQTQ/s9lsevDBB2Wz2UI9lFaDex5c3O/g454HF/c7uLjfaG74Oxtc3O/g4n4HH/c8uLjfwdUS73er21gAAAAAAAAAOF5kogEAAAAAAAD1IIgGAAAAAAAA1IMgGgAAAAAAAFAPgmgAAAAAAABAPQii+dmCBQvUrVs3RUVFaejQoVq1alWoh9QizJ07V6eeeqri4+PVvn17XXrppdq6datXH8Mw9NBDD6lDhw6Kjo7WmWeeqU2bNoVoxC3L3LlzZbFYNGPGDHcb99v/9u3bp+uvv14pKSmKiYnR4MGDtW7dOvdx7rn/VFZW6ve//726deum6Ohode/eXXPmzJHD4XD34X6fmM8++0zjx49Xhw4dZLFY9M4773gdb8j9LSsr05133qm2bdsqNjZWF198sfbu3RvEbwF4Y54XGMzzQot5XnAwzwse5nmB16rneQb85vXXXzciIiKMv/3tb8bmzZuN6dOnG7Gxscbu3btDPbRmb+zYscbLL79s/PDDD8aGDRuMCy+80OjcubNRWFjo7vPYY48Z8fHxxptvvml8//33xoQJE4z09HQjPz8/hCNv/r7++muja9euxsCBA43p06e727nf/nXkyBGjS5cuxqRJk4yvvvrK2Llzp/Hhhx8a27dvd/fhnvvPI488YqSkpBjvvfeesXPnTuONN94w4uLijPnz57v7cL9PzPLly43Zs2cbb775piHJePvtt72ON+T+TpkyxejYsaORmZlpfPvtt8ZZZ51lDBo0yKisrAzytwGY5wUS87zQYZ4XHMzzgot5XuC15nkeQTQ/Ou2004wpU6Z4tfXt29e49957QzSilis3N9eQZKxcudIwDMNwOBxGWlqa8dhjj7n7lJaWGomJicbzzz8fqmE2ewUFBUavXr2MzMxM44wzznBPrrjf/ve73/3OOP3002s9zj33rwsvvNC4+eabvdp+9atfGddff71hGNxvf6s+uWrI/T127JgRERFhvP766+4++/btM8LCwowPPvggaGMHXJjnBQ/zvOBgnhc8zPOCi3lecLW2eR7lnH5SXl6udevWacyYMV7tY8aM0erVq0M0qpYrLy9PkpScnCxJ2rlzp3Jycrzuv81m0xlnnMH9PwG33367LrzwQp177rle7dxv/1u2bJmGDRumK6+8Uu3bt9eQIUP0t7/9zX2ce+5fp59+uj766CNt27ZNkrRx40Z9/vnnuuCCCyRxvwOtIfd33bp1qqio8OrToUMHDRgwgD8DBB3zvOBinhcczPOCh3lecDHPC62WPs8LD/UAWopDhw7JbrcrNTXVqz01NVU5OTkhGlXLZBiGZs2apdNPP10DBgyQJPc99nX/d+/eHfQxtgSvv/66vv32W61du7bGMe63/+3YsUMLFy7UrFmzdP/99+vrr7/WtGnTZLPZNHHiRO65n/3ud79TXl6e+vbtK6vVKrvdrv/7v//TNddcI4m/44HWkPubk5OjyMhItWnTpkYf/ruKYGOeFzzM84KDeV5wMc8LLuZ5odXS53kE0fzMYrF4vTcMo0YbTswdd9yh7777Tp9//nmNY9x//9izZ4+mT5+u//3vf4qKiqq1H/fbfxwOh4YNG6ZHH31UkjRkyBBt2rRJCxcu1MSJE939uOf+sXTpUv3zn//Uq6++qpNOOkkbNmzQjBkz1KFDB914443uftzvwGrM/eXPAKHE/08IPOZ5gcc8L/iY5wUX87ymoaXO8yjn9JO2bdvKarXWiJrm5ubWiMCi8e68804tW7ZMn3zyiTp16uRuT0tLkyTuv5+sW7dOubm5Gjp0qMLDwxUeHq6VK1fqmWeeUXh4uPuecr/9Jz09Xf379/dq69evn7KysiTxd9zf7r77bt177726+uqrdfLJJ+uGG27QzJkzNXfuXEnc70BryP1NS0tTeXm5jh49WmsfIFiY5wUH87zgYJ4XfMzzgot5Xmi19HkeQTQ/iYyM1NChQ5WZmenVnpmZqZEjR4ZoVC2HYRi644479NZbb+njjz9Wt27dvI5369ZNaWlpXve/vLxcK1eu5P43wjnnnKPvv/9eGzZscD+GDRum6667Ths2bFD37t253342atQobd261att27Zt6tKliyT+jvtbcXGxwsK8/xNotVrdW59zvwOrIfd36NChioiI8OqTnZ2tH374gT8DBB3zvMBinhdczPOCj3lecDHPC60WP88L9k4GLZlr6/OXXnrJ2Lx5szFjxgwjNjbW2LVrV6iH1uzddtttRmJiovHpp58a2dnZ7kdxcbG7z2OPPWYkJiYab731lvH9998b11xzDdsU+5Hnrk2Gwf32t6+//toIDw83/u///s/46aefjCVLlhgxMTHGP//5T3cf7rn/3HjjjUbHjh3dW5+/9dZbRtu2bY177rnH3Yf7fWIKCgqM9evXG+vXrzckGfPmzTPWr19v7N692zCMht3fKVOmGJ06dTI+/PBD49tvvzXOPvvsZrH1OVom5nmBwzwv9JjnBRbzvOBinhd4rXmeRxDNz5577jmjS5cuRmRkpHHKKae4t+bGiZHk8/Hyyy+7+zgcDuPBBx800tLSDJvNZvzyl780vv/++9ANuoWpPrnifvvfu+++awwYMMCw2WxG3759jb/+9a9ex7nn/pOfn29Mnz7d6Ny5sxEVFWV0797dmD17tlFWVubuw/0+MZ988onP/7994403GobRsPtbUlJi3HHHHUZycrIRHR1tXHTRRUZWVlYIvg1gYp4XGMzzQo95XuAxzwse5nmB15rneRbDMIzg5b0BAAAAAAAAzQ9rogEAAAAAAAD1IIgGAAAAAAAA1IMgGgAAAAAAAFAPgmgAAAAAAABAPQiiAQAAAAAAAPUgiAYAAAAAAADUgyAaAAAAAAAAUA+CaAAAAAAAAEA9CKIBwAn69NNPZbFYdOzYsVAPBQAAAH7EPA+AJ4JoAAAAAAAAQD0IogEAAAAAAAD1IIgGoNkzDEOPP/64unfvrujoaA0aNEj//ve/JVWl4L///vsaNGiQoqKiNHz4cH3//fde13jzzTd10kknyWazqWvXrnrqqae8jpeVlemee+5RRkaGbDabevXqpZdeesmrz7p16zRs2DDFxMRo5MiR2rp1a2C/OAAAQAvHPA9AU0IQDUCz9/vf/14vv/yyFi5cqE2bNmnmzJm6/vrrtXLlSnefu+++W08++aTWrl2r9u3b6+KLL1ZFRYUkc1J01VVX6eqrr9b333+vhx56SH/4wx+0ePFi9/kTJ07U66+/rmeeeUZbtmzR888/r7i4OK9xzJ49W0899ZS++eYbhYeH6+abbw7K9wcAAGipmOcBaEoshmEYoR4EADRWUVGR2rZtq48//lgjRoxwt99yyy0qLi7Wb37zG5111ll6/fXXNWHCBEnSkSNH1KlTJy1evFhXXXWVrrvuOh08eFD/+9//3Offc889ev/997Vp0yZt27ZNffr0UWZmps4999waY/j000911lln6cMPP9Q555wjSVq+fLkuvPBClZSUKCoqKsB3AQAAoOVhngegqSETDUCztnnzZpWWluq8885TXFyc+/HKK6/o559/dvfznHglJyerT58+2rJliyRpy5YtGjVqlNd1R40apZ9++kl2u10bNmyQ1WrVGWecUedYBg4c6H6dnp4uScrNzT3h7wgAANAaMc8D0NSEh3oAAHAiHA6HJOn9999Xx44dvY7ZbDavCVZ1FotFkrnWhuu1i2eSbnR0dIPGEhERUeParvEBAADg+DDPA9DUkIkGoFnr37+/bDabsrKy1LNnT69HRkaGu9+aNWvcr48ePapt27apb9++7mt8/vnnXtddvXq1evfuLavVqpNPPlkOh8Nr7Q0AAAAEFvM8AE0NmWgAmrX4+HjdddddmjlzphwOh04//XTl5+dr9erViouLU5cuXSRJc+bMUUpKilJTUzV79my1bdtWl156qSTpt7/9rU499VT98Y9/1IQJE/Tll1/q2Wef1YIFCyRJXbt21Y033qibb75ZzzzzjAYNGqTdu3crNzdXV111Vai+OgAAQIvGPA9AU0MQDUCz98c//lHt27fX3LlztWPHDiUlJemUU07R/fff706zf+yxxzR9+nT99NNPGjRokJYtW6bIyEhJ0imnnKJ//etfeuCBB/THP/5R6enpmjNnjiZNmuT+jIULF+r+++/X1KlTdfjwYXXu3Fn3339/KL4uAABAq8E8D0BTwu6cAFo0145KR48eVVJSUqiHAwAAAD9hngcg2FgTDQAAAAAAAKgHQTQAAAAAAACgHpRzAgAAAAAAAPUgEw0AAAAAAACoB0E0AAAAAAAAoB4E0QAAAAAAAIB6EEQDAAAAAAAA6kEQDQAAAAAAAKgHQTQAAAAAAACgHgTRAAAAAAAAgHoQRAMAAAAAAADq8f9QA6/fiKE26gAAAABJRU5ErkJggg=="},"metadata":{}}]},{"cell_type":"code","source":"mymodel = tf.keras.models.load_model('../working/model_resnet50.hdf5')\n\nmymodel.evaluate(test_generator, verbose = 1)","metadata":{"execution":{"iopub.status.busy":"2023-07-16T06:34:23.157635Z","iopub.execute_input":"2023-07-16T06:34:23.157994Z","iopub.status.idle":"2023-07-16T06:34:26.999746Z","shell.execute_reply.started":"2023-07-16T06:34:23.157961Z","shell.execute_reply":"2023-07-16T06:34:26.998692Z"},"trusted":true},"execution_count":25,"outputs":[{"name":"stdout","text":"2/2 [==============================] - 2s 601ms/step - loss: 0.5916 - accuracy: 0.9375\n","output_type":"stream"},{"execution_count":25,"output_type":"execute_result","data":{"text/plain":"[0.5915923714637756, 0.9375]"},"metadata":{}}]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Conclusion","metadata":{}},{"cell_type":"markdown","source":"VGG16 gives the highest testing accuracy with 97.9%","metadata":{}}]}