{"cells":[{"metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true},"cell_type":"code","source":"import numpy as np # linear algebra\nimport pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\nimport tensorflow as tf\nimport tensorflow.keras as keras\nimport matplotlib.pyplot as plt\nimport seaborn as sns\nfrom sklearn.model_selection import train_test_split\nimport tensorflow\nfrom tensorflow.keras.callbacks import EarlyStopping,ModelCheckpoint\nfrom tensorflow.keras.preprocessing.image import ImageDataGenerator\nimport PIL\nfrom PIL import Image\nfrom tensorflow.keras.applications import DenseNet121\n\nfrom tensorflow.keras.applications.resnet50 import ResNet50\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# Any results you write to the current directory are saved as output.","execution_count":null,"outputs":[]},{"metadata":{"_uuid":"d629ff2d2480ee46fbb7e2d37f6b5fab8052498a","_cell_guid":"79c7e3d0-c299-4dcb-8224-4455121ee9b0","trusted":true},"cell_type":"code","source":"base_path = '/kaggle/input/Kannada-MNIST/'\ntrain = pd.read_csv(base_path + 'train.csv')\ntest = pd.read_csv(base_path + 'test.csv')\ndig_mnist = pd.read_csv(base_path + 'Dig-MNIST.csv')\n\n\ntrain_label = train['label']\ntrain_data_raw = train.drop('label', axis=1, inplace=False)\ntrain_data = []\n\ntrain_data_raw = np.array(train_data_raw)\n\nfor i in train_data_raw:\n    image = i.reshape(28,28,1)\n    train_data.append(image)\n\n#test...\ntest.drop('id', axis=1, inplace=True)\n\ntest_data_raw = np.array(test)\n\ntest_data = []\nfor i in test_data_raw:\n    image = i.reshape(28,28,1)\n    test_data.append(image)\n\n","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"plt.figure( figsize = (10,10))\nfor i in range(1,10):\n    plt.subplot(3,4,i)\n    plt.title(train_label[i])\n    plt.imshow(train_data[i].reshape([28,28]))\n    \nplt.plot()\ntrain_label.value_counts().plot(kind='barh')\n\n","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"#create a np.array of pixel\ntrain_data = np.array(train_data)/255.0\ntest_data = np.array(test_data)/255.0\n\ntrain_label  = np.array(train_label).reshape(60000,-1)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"es = EarlyStopping(monitor='accuracy', mode='min', verbose=1, patience=5,baseline=0.99)\n\ndef define_model():\n    model2 = keras.models.Sequential([\n        keras.layers.Conv2D(16, (3,3), input_shape=(28,28,1), activation='relu'),#32\n        keras.layers.BatchNormalization(),\n        #keras.layers.Conv2D(16,(3,3), activation='relu'),\n        #keras.layers.BatchNormalization(),\n        keras.layers.Conv2D(16, (5,5), activation='relu', padding='same'),\n        keras.layers.BatchNormalization(),\n        keras.layers.MaxPooling2D(2,2),\n        keras.layers.Dropout(0.4),\n        keras.layers.Conv2D(32, (3,3), activation='relu'),\n        keras.layers.BatchNormalization(),\n        #keras.layers.Conv2D(32,(3,3), activation='relu'),\n        #keras.layers.BatchNormalization(),\n        keras.layers.Conv2D(32, (5,5), activation='relu', padding='same'),\n        keras.layers.BatchNormalization(),\n        keras.layers.MaxPooling2D(2,2),\n        keras.layers.BatchNormalization(),\n        keras.layers.Dropout(0.5),\n        keras.layers.Conv2D(64, (3,3), activation='relu'),\n        keras.layers.BatchNormalization(),\n        #keras.layers.Conv2D(64, (3,3), activation='relu', padding='same'),\n        #keras.layers.BatchNormalization(),\n        keras.layers.Conv2D(64, (5,5), activation='relu', padding='same'),\n        keras.layers.BatchNormalization(),\n        keras.layers.MaxPooling2D(2,2),\n        keras.layers.BatchNormalization(),\n        keras.layers.Dropout(0.5),\n        keras.layers.Flatten(),\n        keras.layers.Dense(128), #256 prima!\n        keras.layers.BatchNormalization(),\n        keras.layers.Activation('relu'), \n        keras.layers.Dropout(0.6),\n        keras.layers.Dense(64),#32 prima\n        keras.layers.BatchNormalization(),\n        keras.layers.Activation('relu'),\n        keras.layers.Dropout(0.5),\n        keras.layers.Dense(10, activation='softmax')\n    ])\n    return model2\n\n\nmodel = define_model()\n\nmodel.compile(optimizer='adam', loss='sparse_categorical_crossentropy', metrics=['accuracy'])\n\nmodel.summary()","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"def define_model_ResNet50():\n    model = keras.models.Sequential()\n    input_layer = keras.layers.Input(shape=(224, 224, 3), name='image_input')\n    model.add(DenseNet121(weights=None, include_top=False, input_tensor=input_layer))\n    model.add(keras.layers.Flatten())\n    model.add(keras.layers.Dense(128))\n    model.add(keras.layers.BatchNormalization())\n    model.add(keras.layers.Activation('relu'))\n    model.add(keras.layers.Dropout(0.3))\n    model.add(keras.layers.Dense(32))\n    model.add(keras.layers.BatchNormalization())\n    model.add(keras.layers.Activation('relu'))\n    model.add(keras.layers.Dropout(0.1))\n    model.add(keras.layers.Dense(10, activation='softmax'))\n    \n    model.layers[0].trainable = True\n    model.summary()\n    return model\n","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"EPOCHS =100\nX_train ,X_test,Y_train,Y_test= train_test_split(train_data,train_label,test_size=0.1)\nprint(np.array(Y_train).shape)\n\nes = keras.callbacks.EarlyStopping(monitor='accuracy', mode='min', patience=3, baseline=0.99)\ncheckpoint = ModelCheckpoint('best_weights.h5', monitor='val_loss', sava_best_only=True, mode='auto', period=1)\n\n#prova resnet\nmodel = define_model()\nmodel.compile(optimizer='adam', loss='sparse_categorical_crossentropy', metrics=['accuracy'])\n","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"#per il resnet..!!\n\n#base_path = '/kaggle/input/Kannada-MNIST/'\n#train_res = pd.read_csv(base_path + 'train.csv')\n#test_res = pd.read_csv(base_path + 'test.csv')\n#dig_mnist_res = pd.read_csv(base_path + 'Dig-MNIST.csv')\n\n","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"#X_train_images = []#\n#x_list = []\n\n#count = 0\n#mappa = []\n#len = train_res.shape[0]\n\n\n#for i in range(0,len):\n#    save_dir = '/kaggle/tmp/'\n    \n##    record = train_res.loc[i]\n#    tmp_img_label = record['label']\n #   record.drop('label',inplace=True)\n #   record = np.array(record).reshape(28,28,1)\n #   if not os.path.exists(save_dir):\n #       os.makedirs(save_dir)\n #   tmpImg = np.dstack([record,record,record])\n #   new_im = PIL.Image.fromarray(np.array(tmpImg).astype('uint8'))\n  #  img_path = save_dir + \"img\" + str(count) + \".png\"\n  #  new_im.save(img_path)\n   # mappa.append([img_path, tmp_img_label])\n    #count = count + 1\n","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"#df = pd.DataFrame(df,columns = ['filename','label'])\n#df = pd.DataFrame(mappa)\n#df.head()\n\n\n#df.to_csv('/kaggle/tmp/train.csv',index=False)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"\n\n#mappaDUE = pd.read_csv('/kaggle/tmp/train.csv', names =['filename','label'])\n#df = pd.DataFrame(mappaDUE)\n#df.head()\n#df.to_csv('/kaggle/tmp/train.csv',index=False)\n\n    \n#for i in os.listdir('/kaggle/tmp/'):\n    #X_train_images.append(new_img)\n    #img = tensorflow.keras.preprocessing.image.load_img('/kaggle/tmp/' + i, target_size=(224, 224))\n    #x = tensorflow.keras.preprocessing.image.img_to_array(img)\n    #del img\n    #x_list.append(x)\n    \n\n#datagen_res = ImageDataGenerator(featurewise_center=False,\n#        rescale=1./255,# set input mean to 0 over the dataset\n#        samplewise_center=False,  # set each sample mean to 0\n#        featurewise_std_normalization=False,  # divide inputs by std of the dataset\n#        samplewise_std_normalization=False,  # divide each input by its std\n#        zca_whitening=False,  # apply ZCA whitening\n##        rotation_range=10,  # randomly rotate images in the range (degrees, 0 to 180)\n#        zoom_range = 0.15, # Randomly zoom image  era 0.05 messo a 0.15\n#       width_shift_range=0.15,  # randomly shift images horizontally (fraction of total width)\n #       height_shift_range=0.15,  # randomly shift images vertically (fraction of total height)\n #       horizontal_flip=False,  # randomly flip images\n #       vertical_flip=False,\n #       validation_split=0.1)  # randomly flip images\n","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"#mappaDUE['label'] = mappaDUE['label'].astype(str)\n\n#def create_flow(datagen, subset):\n#    return datagen_res.flow_from_dataframe(\n#        mappaDUE, #dataframe contenente il mapping file - label\n#        directory='/kaggle/tmp/', # cartella contente le immagini puntate!!\n#        x_col='filename', \n#        y_col=['label'],\n#        class_mode='raw',\n #       target_size=(128, 128),\n#        batch_size=128,\n#        subset=subset\n#    )\n\n#train_gen = create_flow(datagen_res, 'training')\n#val_gen = create_flow(datagen_res, 'validation')\nprint(np.array(X_train).shape)\nprint(np.array(Y_train).shape)\n#history = model.fit_generator(datagen.flow(X_train,Y_train),\n#                              #train_gen,\n#                              epochs = EPOCHS, validation_data =(X_test,Y_test), #val_gen,\n#                              verbose = 1, callbacks=[es,checkpoint])\n\n\ndatagen = ImageDataGenerator(featurewise_center=False,  # set input mean to 0 over the dataset\n        samplewise_center=False,  # set each sample mean to 0\n        featurewise_std_normalization=False,  # divide inputs by std of the dataset\n        samplewise_std_normalization=False,  # divide each input by its std\n        zca_whitening=False,  # apply ZCA whitening\n        rotation_range=20,  # randomly rotate images in the range (degrees, 0 to 180)\n        zoom_range = 0.2, # Randomly zoom image \n        width_shift_range=0.20,  # randomly shift images horizontally (fraction of total width)\n        height_shift_range=0.20,  # randomly shift images vertically (fraction of total height)\n        horizontal_flip=False,  # randomly flip images\n        vertical_flip=False)  # randomly flip images\n\n\nhistory = model.fit_generator(datagen.flow(X_train,Y_train),\n                              epochs = EPOCHS, validation_data = (X_test,Y_test),\n                              verbose = 1, callbacks=[es,checkpoint])","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"model_prediction = define_model()\nmodel_prediction.load_weights('best_weights.h5')","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"\n#len_test = np.array(test_res).shape[0]\n#test_res.drop('id', axis=1, inplace=True)\n#mappaT = []\n\n#count = 0\n#for i in range(0,len_test):\n#    save_dir = '/kaggle/tmpTst/'\n    \n#    record = test_res.loc[i]\n    #tmp_img_label = record['label']\n    #record.drop('label',inplace=True)\n#    record = np.array(record).reshape(28,28,1)\n#    if not os.path.exists(save_dir):\n#        os.makedirs(save_dir)\n#    tmpImg = np.dstack([record,record,record])\n#    new_im = PIL.Image.fromarray(np.array(tmpImg).astype('uint8'))\n#    img_path = save_dir + \"img\" + str(count) + \".png\"\n#     new_im.save(img_path)\n#    mappaT.append([img_path, ''])\n#    count = count + 1\n    \n#df = pd.DataFrame(mappaT)\n#df.head()\n#df.to_csv('/kaggle/tmpTst/test.csv',index=False)\n#mappaDUETST = pd.read_csv('/kaggle/tmpTst/test.csv', names =['filename','label'])\n#df = pd.DataFrame(mappaDUETST)\n#df.head()\n#df.to_csv('/kaggle/tmpTst/test.csv',index=False)\n\n\n#def create_test_gen(datagen):\n#    return datagen.flow_from_dataframe(\n#        mappaDUETST,\n#        directory='/kaggle/tmpTst/',\n#        x_col='filename',\n#        class_mode=None,\n#        target_size=(224, 224),\n#        batch_size=128,\n#        shuffle=False\n#    )\n\n#test_gen = create_test_gen(datagen_res)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"#results = model_prediction.predict_generator(test_gen)\n\nresults = model_prediction.predict(test_data)\n# select the indix with the maximum probability\nresults = np.argmax(results,axis = 1)\nresults = pd.Series(results,name=\"Label\")\n\nsubmission_file = pd.read_csv('/kaggle/input/Kannada-MNIST/sample_submission.csv')\nsubmission_file['label'] = results\n#submission = pd.concat([pd.Series(range(1,28001),name = \"ImageId\"),results],axis = 1)\nprint(results)\nsubmission_file.to_csv(\"submission.csv\",index=False)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"acc = history.history['acc']\n\nepochs_ = range(0, EPOCHS)\n\nplt.plot(epochs_ , acc, label='accuracy')\nplt.xlabel('epochs')\nplt.ylabel('accuracy')\n\nplt.title('accuracy vs epochs')\nplt.legend()","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":1}