{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"# This Python 3 environment comes with many helpful analytics libraries installed\n# It is defined by the kaggle/python Docker image: https://github.com/kaggle/docker-python\n# For example, here's several helpful packages to load\n\nimport numpy as np # linear algebra\nimport pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\n\n\n# Input data files are available in the read-only \"../input/\" directory\n# For example, running this (by clicking run or pressing Shift+Enter) will list all files under the input directory\n\nimport os\nfor dirname, _, filenames in os.walk('/kaggle/input'):\n    for filename in filenames:\n        print(os.path.join(dirname, filename))\n\n# You can write up to 20GB to the current directory (/kaggle/working/) that gets preserved as output when you create a version using \"Save & Run All\" \n# You can also write temporary files to /kaggle/temp/, but they won't be saved outside of the current session","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# 이미지 데이터를 보고 c0 ~ c9인지 판단하기(Ex) c0(안전운전), c6(드링킹), c9(통화 중))\nimport pandas as pd\ntrain = pd.read_csv('/kaggle/input/state-farm-distracted-driver-detection/driver_imgs_list.csv')\ntrain","metadata":{"_uuid":"d629ff2d2480ee46fbb7e2d37f6b5fab8052498a","_cell_guid":"79c7e3d0-c299-4dcb-8224-4455121ee9b0","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import glob\nglob.glob('/kaggle/input/state-farm-distracted-driver-detection/imgs/train/*/*')","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"from PIL import Image\nImage.open('/kaggle/input/state-farm-distracted-driver-detection/imgs/train/c5/img_68208.jpg')","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"Image.open('/kaggle/input/state-farm-distracted-driver-detection/imgs/train/c0/img_44733.jpg')","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"pd.options.display.max_colwidth = 999","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# classname과 path가 일치하지 않는다\ntrain['path'] = glob.glob('/kaggle/input/state-farm-distracted-driver-detection/imgs/train/*/*')\ntrain","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train['subject'].unique()","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# test['subject'].unique()","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# classname과 path를 일치하도록!\ntrain['classname'] = train['path'].apply(lambda x: x.split('/')[-2])\ntrain","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# K-fold CV(교차검정)\nfrom sklearn.model_selection import train_test_split\n\n# random_state\n# stratify: 비율을 동일하게 추출\nx_train, x_valid = train_test_split(train, test_size = 0.2, random_state = 42, stratify = train['classname'])\nx_valid['classname'].value_counts()","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Data augmentation(데이터 증가)\n# horizontal_flip\nfrom keras.preprocessing.image import ImageDataGenerator\nidg = ImageDataGenerator(horizontal_flip = True)\nidg2 = ImageDataGenerator()\n\n# target_size, batch_size\ntrain_generator = idg.flow_from_dataframe(x_train, x_col = 'path', y_col = 'classname', target_size = (100,100), batch_size = 256) \nvalid_generator = idg.flow_from_dataframe(x_valid, x_col = 'path', y_col = 'classname', target_size = (100,100), batch_size = 256)\n\ntrain_generator\nvalid_generator","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# callback(호출)\n# ReduceLROnPlateau\nfrom keras.callbacks import *\nrl = ReduceLROnPlateau(patience = 2, factor = 0.1, verbose = 1)\n\n# patience: 가장 좋았을 때와 대비하여 '연속' 3번 실행, EarlyStopping(조기종료)\nes = EarlyStopping(patience = 3, verbose = 1)\n# 최적의 순간만 저장하겠다\nmc = ModelCheckpoint('best.h5', save_best_only = True, verbose = 1)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# EfficientNetB0\nfrom keras import *\nfrom keras.layers import *\nfrom tensorflow.keras.applications.efficientnet import EfficientNetB0\n\n# loss, val_acc\n# acc: 0.9948, val_acc: 0.9944\nmodel = Sequential()\nmodel.add(EfficientNetB0(weights = 'imagenet', include_top = False, pooling = 'avg'))\nmodel.add(Dense(10, activation = 'softmax'))\nmodel.compile(metrics = ['acc'], optimizer = 'adam', loss = 'categorical_crossentropy')\nmodel.fit(train_generator, epochs = 10, validation_data = valid_generator, callbacks = [es,mc,rl])\n\nmodel.summary()","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"test = pd.DataFrame({'path': glob.glob('/kaggle/input/state-farm-distracted-driver-detection/imgs/test/*')})\npd.options.display.max_colwidth = 999","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# 결과값이 나오지 않음\ntest","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"### KeyError: 'path'?\n# class_mode, shuffle\ntest_generator = idg.flow_from_dataframe(test, x_col = 'path', y_col = None, class_mode = None, shuffle = False)\ntest_generator","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# h5는 가중치를 저장할 수 있는 확장자의 한 예\nmodel.load_weights('best.h5')\n\n# predict_classes\nresult = model.predict_classes(test_generator, verbose = 1)\nresult","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"result.shape","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"### Question\nclass_list = []\nlabel_list = sorted(train['label'].unique())\n\nfor j in result:\n     target = label_list[j]\n     class_list.append(target)","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"class_list","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# class_indices\ntrain_generator.class_indices","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"sub['classname']","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"sub = pd.read_csv('/kaggle/input/state-farm-distracted-driver-detection/sample_submission.csv')\nsub","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"sub['classname'] = class_list","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# 순서 바꾸기\n# 모델 90점이상 만들자\nsub.to_csv('Driver.csv', index = 0)","metadata":{"trusted":true},"execution_count":null,"outputs":[]}]}