{"cells":[{"metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true},"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# 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\n\n\n# You can write up to 5GB 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","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"import tensorflow as tf\nfrom tensorflow.keras.models import Sequential\nfrom tensorflow.keras.layers import Dense, Conv2D, Flatten, Dropout, MaxPooling2D\nfrom tensorflow.keras.preprocessing.image import ImageDataGenerator\nfrom tensorflow.keras import layers\nimport os\nimport numpy as np\nimport matplotlib.pyplot as plt\nimport numpy as np # linear algebra\nimport pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\nimport os\nimport pydicom\nfrom tensorflow import keras\nimport matplotlib.pyplot as plt\nimport matplotlib.animation as animation\nfrom matplotlib.widgets import Slider\n\nfrom IPython.display import HTML\nimport cv2\n\n\nfrom tensorflow.keras.models import Sequential\nfrom tensorflow.keras.layers import Dense, Conv2D, Flatten, Dropout, MaxPooling2D\nfrom tensorflow.keras.preprocessing.image import ImageDataGenerator\nfrom tensorflow.keras import layers\nimport os\nimport numpy as np\nimport matplotlib.pyplot as plt\n\nimport matplotlib.pyplot as plt\nfrom tensorflow.keras.preprocessing.image import ImageDataGenerator\nfrom collections import Counter\nimport tensorflow as tf\n\nfrom tensorflow import keras\nfrom tensorflow.keras.models import Model\nfrom tensorflow.keras.layers import Dense, Embedding, Input, Flatten\nfrom tensorflow.keras.layers import LSTM, Bidirectional, GlobalMaxPool1D, Dropout\nfrom tensorflow.keras.preprocessing.text import Tokenizer\nfrom tensorflow.keras.preprocessing.sequence import pad_sequences\nfrom tensorflow.keras.optimizers import Adam\nfrom sklearn.metrics import roc_auc_score\nimport tensorflow as tf\nfrom tensorflow import keras\nfrom tensorflow.keras import Sequential\nfrom tensorflow.keras.layers import Conv1D, MaxPool1D, BatchNormalization, Flatten, Dropout, Dense, Conv1D, MaxPooling1D, BatchNormalization, GRU\nimport tensorflow_datasets as tfds\nfrom tensorflow import feature_column\nimport pandas as pd\nfrom sklearn.preprocessing import MinMaxScaler,StandardScaler,LabelBinarizer\nimport matplotlib.pyplot as plt","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"#\n# Sample image\n#\n\nimage_string = tf.io.read_file(\"../input/siim-isic-melanoma-classification/jpeg/train/ISIC_0015719.jpg\")\nimage=tf.image.decode_jpeg(image_string,channels=3)\n\n# Image shape \n(image.shape, image.numpy().max())","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"#\n# Sample image\n#\nfig = plt.figure()\nplt.subplot(1,2,1)\nplt.title('Original image')\nplt.imshow(image)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"#\n# Load meta\n#\n\ndata = pd.read_csv(\"../input/siim-isic-melanoma-classification/train.csv\")\nfor ind, row in data.iterrows():\n    data.loc[ind, \"image_path\"] = row.image_name + \".jpg\"\n    \ndata = data[['patient_id', 'sex', 'age_approx', 'anatom_site_general_challenge', 'diagnosis', 'benign_malignant', 'target', 'image_path']]\ndata.target = data.target.astype('str')\n\n","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"pd.set_option('display.max_columns', 500)\npd.set_option('display.width', 1000)\npd.set_option('display.max_colwidth', -1)\npd.set_option('display.max_columns', None)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"# UNDER SAMPLING to remove imbalance and decrese data\ndata_mod = pd.concat([data[data.target == \"1\"][0:584], data[data.target == \"0\"][0:500]]).sample(frac=1, axis=1).sample(frac=1).reset_index(drop=True)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"#\n# image Augmentation\n#\n\ntrain_image_generator = ImageDataGenerator(\n                                            featurewise_center=False, samplewise_center=False,\n                                            featurewise_std_normalization=False, samplewise_std_normalization=False,\n                                            zca_whitening=False, zca_epsilon=1e-06, rotation_range=40, width_shift_range=0.2,\n                                            height_shift_range=0.0, brightness_range=None, shear_range=0.2, zoom_range=0.2,\n                                            channel_shift_range=0.0, fill_mode='nearest', cval=0.0, horizontal_flip=True,\n                                            vertical_flip=True, rescale=1. / 255, preprocessing_function=None,\n                                            data_format=None, validation_split=0.2, dtype=None\n                                        )","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"batch_size = 16\nIMG_SHAPE = (1024, 1024, 3)\nIMG = (IMG_SHAPE[0], IMG_SHAPE[1])","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"train_data_gen = train_image_generator.flow_from_dataframe(\n                                                            data_mod, directory=\"../input/siim-isic-melanoma-classification/jpeg/train\", x_col='image_path', y_col='target', weight_col=None,\n                                                            target_size=IMG, color_mode='rgb', classes=None,\n                                                            class_mode='binary', batch_size=batch_size, shuffle=True, seed=None,\n                                                            save_to_dir=None, save_prefix='', save_format='png', subset=\"training\",\n                                                            interpolation='nearest', validate_filenames=True\n                                                        )","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"Counter(train_data_gen.classes)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"test_data_gen = train_image_generator.flow_from_dataframe(\n                                                            data_mod, directory=\"../input/siim-isic-melanoma-classification/jpeg/train\", x_col='image_path', y_col='target', weight_col=None,\n                                                            target_size=IMG, color_mode='rgb', classes=None,\n                                                            class_mode='binary', batch_size=batch_size, shuffle=True, seed=None,\n                                                            save_to_dir=None, save_prefix='', save_format='png', subset=\"validation\",\n                                                            interpolation='nearest', validate_filenames=True\n                                                        )","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"Counter(test_data_gen.classes)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"\n#\n# RES-NET MODEL Config\n#\ninception_model = tf.keras.applications.ResNet152V2(\n    include_top=True, weights='imagenet', input_tensor=None,\n    pooling=None, classes=1000, classifier_activation='softmax'\n)\n\n# Enable Training of resnet\nfor layer in inception_model.layers:\n    layer.trainable = True\n\n\nMETRICS = [\n      keras.metrics.TruePositives(name='tp'),\n      keras.metrics.FalsePositives(name='fp'),\n      keras.metrics.TrueNegatives(name='tn'),\n      keras.metrics.FalseNegatives(name='fn'), \n      keras.metrics.BinaryAccuracy(name='accuracy'),\n      keras.metrics.Precision(name='precision'),\n      keras.metrics.Recall(name='recall'),\n      keras.metrics.AUC(name='auc')\n]\n","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"model = Sequential()\nmodel.add(inception_model.layers[-2])\nmodel.add(Flatten())\nmodel.add(Dense(512, activation=\"relu\", kernel_regularizer=keras.regularizers.l2(0.001)))\nmodel.add(Dense(1, activation=\"sigmoid\"))\nmodel.compile(optimizer=keras.optimizers.RMSprop(lr=0.001), loss='binary_crossentropy', metrics=[METRICS])","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"keras.backend.set_value(model.optimizer.lr,0.1)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"epochs = 10","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"class_weight = dict(Counter(train_data_gen.classes))\ntotal = len(train_data_gen.classes)\nprint(dict(Counter(train_data_gen.classes)))\nclass_weight = {i:(1/j)*total/len(class_weight) for i,j in class_weight.items()}","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"class_weight","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"import math\n\n# Dynamic learning rate\ndef scheduler(epoch):\n  epoch_limit = 5\n\n  if epoch < epoch_limit:\n    return 0.001\n  else:\n    return  max(0.0001 * math.exp(0.0001 * (epoch_limit - epoch)) , 0.0001)\nlrcallback = tf.keras.callbacks.LearningRateScheduler(scheduler)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"history = model.fit(\n  train_data_gen,\n  steps_per_epoch=len(train_data_gen.filepaths) // batch_size,\n  epochs=20,\n  verbose=1,\n  validation_data=test_data_gen,\n  validation_steps=len(test_data_gen.filepaths) // batch_size,\n  class_weight=class_weight,\n  callbacks=[lrcallback])","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"../input/eff2t/eff2t.h5","execution_count":null},{"metadata":{"trusted":true},"cell_type":"code","source":"model.save(\"res2t.h5\")","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"plt.plot(history.history['accuracy'])\nplt.plot(history.history['val_accuracy'])\nplt.title('model accuracy')\nplt.ylabel('accuracy')\nplt.xlabel('epoch')\nplt.legend(['train', 'val'], loc='upper left')\nplt.show()","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"","execution_count":null,"outputs":[]}],"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":4,"nbformat_minor":4}