{"cells":[{"metadata":{"_uuid":"b7da188542295605973b40ac1b1ce0cfc4ce7773"},"cell_type":"markdown","source":""},{"metadata":{"trusted":true,"_uuid":"f2e5e5157318e7dbde5e9280015583a851b5915b","_kg_hide-input":true},"cell_type":"code","source":"import numpy as np\nimport pandas as pd\nimport pydicom\nimport matplotlib.pyplot as plt\nfrom PIL import Image\nimport tensorflow as tf\nfrom random import shuffle\nfrom sklearn.model_selection import train_test_split\n\nimport keras\nfrom keras.models import Sequential\nfrom keras.utils import plot_model\nfrom keras.layers import Dense, Dropout, Flatten\nfrom keras.layers import Conv2D, MaxPooling2D, ZeroPadding2D, Convolution2D\nfrom keras.optimizers import SGD\nfrom keras.layers. normalization import BatchNormalization\nfrom keras.preprocessing.image import ImageDataGenerator\nimport numpy as np","execution_count":1,"outputs":[{"output_type":"stream","text":"Using TensorFlow backend.\n","name":"stderr"}]},{"metadata":{"trusted":true,"_uuid":"587457c23462591dfddb80156e6e5d02a4d06d52","scrolled":false},"cell_type":"code","source":"def data_loader_csv(path,columns=None):\n    data = pd.read_csv(path)\n    if not columns==None:\n        data = data.filter(columns)\n    return data\n\ndef image_reader(image_path, show='False'):\n    image_arr = pydicom.read_file(image_path)\n    image_arr = image_arr.pixel_array\n    if show:\n        plt.imshow(image_arr,cmap='gray')\n        plt.show()\n    return image_arr\n\ndef label_one_shot(label_value):\n    if label_value == 0:\n        return np.array([0, 1])\n    elif label_value == 1:\n        return np.array([1, 0])\n    \n\ndef load_data(dataset, resize_size):\n    data = []\n    for i in range(len(dataset)):\n        array_img = image_reader('../input/rsna-pneumonia-detection-challenge/stage_2_train_images/'+dataset[i][0]+'.dcm', show=False)\n        img = Image.fromarray(array_img)\n        img = img.resize(resize_size)\n        array_img = np.array(img) / 255\n        data.append([array_img, label_one_shot(dataset[i][1])])\n        \n    shuffle(data)\n    \n    return data\n \ndata_rsna = data_loader_csv('../input/rsna-pneumonia-detection-challenge/stage_2_train_labels.csv', ['patientId','Target'])\n#data_rsna.head(10)\ndataset = data_rsna.values\n\narr = image_reader('../input/rsna-pneumonia-detection-challenge/stage_2_train_images/'+dataset[0][0]+'.dcm', show=True) #/ test\n\nimage_height = int(arr.shape[0] / 8)  # Image size checking -> 128 x 128\nimage_width = int(arr.shape[1] / 8)\n\ndata_r = load_data(dataset, (image_width, image_height)) # Image size 128x128\nprint('Complete load_data!')","execution_count":2,"outputs":[{"output_type":"display_data","data":{"text/plain":"<Figure size 432x288 with 1 Axes>","image/png":"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\n"},"metadata":{}},{"output_type":"stream","text":"Complete load_data!\n","name":"stdout"}]},{"metadata":{"trusted":true,"_uuid":"eeb34e0fff49995cf671289f4001957b742f9452","_kg_hide-input":false},"cell_type":"code","source":"Images = np.array([i[0] for i in data_r]).reshape(-1, image_height, image_width, 1).astype('float32') # trainImages = Image size 128x128\nLabels = np.array([i[1] for i in data_r]).astype('float32') # trainLabels\nX_train, X_test, y_train, y_test = train_test_split(Images, Labels , test_size=0.1)","execution_count":3,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"keras.optimizers.SGD(lr=0.001, momentum=0.0, decay=0.0, nesterov=False)","execution_count":12,"outputs":[{"output_type":"execute_result","execution_count":12,"data":{"text/plain":"<keras.optimizers.SGD at 0x7f6e205e7b70>"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"model = Sequential()\nmodel.add(Conv2D(8, kernel_size=(4, 4), strides=(1, 1), padding='same',activation='relu', input_shape= X_train.shape[1:]))\nmodel.add(MaxPooling2D(pool_size=(2, 2), strides=(2, 2)))\nmodel.add(BatchNormalization())\nmodel.add(Conv2D(16, (2, 2), activation='relu', padding='same'))\nmodel.add(MaxPooling2D(pool_size=(2, 2)))\nmodel.add(BatchNormalization())\nmodel.add(Dropout(0.25))\nmodel.add(Flatten())\nmodel.add(Dense(1000, activation='relu'))\nmodel.add(Dropout(0.5))\nmodel.add(Dense(2, activation='softmax'))\nmodel.compile(loss='binary_crossentropy', optimizer='sgd', metrics=['accuracy'])\nmodel.summary()","execution_count":46,"outputs":[{"output_type":"stream","text":"_________________________________________________________________\nLayer (type)                 Output Shape              Param #   \n=================================================================\nconv2d_21 (Conv2D)           (None, 128, 128, 8)       136       \n_________________________________________________________________\nmax_pooling2d_21 (MaxPooling (None, 64, 64, 8)         0         \n_________________________________________________________________\nbatch_normalization_21 (Batc (None, 64, 64, 8)         32        \n_________________________________________________________________\nconv2d_22 (Conv2D)           (None, 64, 64, 16)        528       \n_________________________________________________________________\nmax_pooling2d_22 (MaxPooling (None, 32, 32, 16)        0         \n_________________________________________________________________\nbatch_normalization_22 (Batc (None, 32, 32, 16)        64        \n_________________________________________________________________\ndropout_21 (Dropout)         (None, 32, 32, 16)        0         \n_________________________________________________________________\nflatten_11 (Flatten)         (None, 16384)             0         \n_________________________________________________________________\ndense_21 (Dense)             (None, 1000)              16385000  \n_________________________________________________________________\ndropout_22 (Dropout)         (None, 1000)              0         \n_________________________________________________________________\ndense_22 (Dense)             (None, 2)                 2002      \n=================================================================\nTotal params: 16,387,762\nTrainable params: 16,387,714\nNon-trainable params: 48\n_________________________________________________________________\n","name":"stdout"}]},{"metadata":{"trusted":true,"_uuid":"7f6c689490a43e27449800a024b980b70ac35858"},"cell_type":"code","source":"from IPython.display import SVG\nfrom keras.utils.vis_utils import model_to_dot\n%matplotlib inline\nSVG(model_to_dot(model, show_shapes=True).create(prog='dot', format='svg'))","execution_count":47,"outputs":[{"output_type":"execute_result","execution_count":47,"data":{"text/plain":"<IPython.core.display.SVG object>","image/svg+xml":"<svg height=\"958pt\" viewBox=\"0.00 0.00 447.00 958.00\" width=\"447pt\" xmlns=\"http://www.w3.org/2000/svg\" xmlns:xlink=\"http://www.w3.org/1999/xlink\">\n<g class=\"graph\" id=\"graph0\" transform=\"scale(1 1) rotate(0) translate(4 954)\">\n<title>G</title>\n<polygon fill=\"white\" points=\"-4,4 -4,-954 443,-954 443,4 -4,4\" stroke=\"none\"/>\n<!-- 140110097278792 -->\n<g class=\"node\" id=\"node1\"><title>140110097278792</title>\n<polygon fill=\"none\" points=\"64.5,-830.5 64.5,-876.5 374.5,-876.5 374.5,-830.5 64.5,-830.5\" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"129.5\" y=\"-849.8\">conv2d_21: Conv2D</text>\n<polyline fill=\"none\" points=\"194.5,-830.5 194.5,-876.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"222\" y=\"-861.3\">input:</text>\n<polyline fill=\"none\" points=\"194.5,-853.5 249.5,-853.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"222\" y=\"-838.3\">output:</text>\n<polyline fill=\"none\" points=\"249.5,-830.5 249.5,-876.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"312\" y=\"-861.3\">(None, 128, 128, 1)</text>\n<polyline fill=\"none\" points=\"249.5,-853.5 374.5,-853.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"312\" y=\"-838.3\">(None, 128, 128, 8)</text>\n</g>\n<!-- 140110097280136 -->\n<g class=\"node\" id=\"node2\"><title>140110097280136</title>\n<polygon fill=\"none\" points=\"22.5,-747.5 22.5,-793.5 416.5,-793.5 416.5,-747.5 22.5,-747.5\" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"129.5\" y=\"-766.8\">max_pooling2d_21: MaxPooling2D</text>\n<polyline fill=\"none\" points=\"236.5,-747.5 236.5,-793.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"264\" y=\"-778.3\">input:</text>\n<polyline fill=\"none\" points=\"236.5,-770.5 291.5,-770.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"264\" y=\"-755.3\">output:</text>\n<polyline fill=\"none\" points=\"291.5,-747.5 291.5,-793.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"354\" y=\"-778.3\">(None, 128, 128, 8)</text>\n<polyline fill=\"none\" points=\"291.5,-770.5 416.5,-770.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"354\" y=\"-755.3\">(None, 64, 64, 8)</text>\n</g>\n<!-- 140110097278792&#45;&gt;140110097280136 -->\n<g class=\"edge\" id=\"edge2\"><title>140110097278792-&gt;140110097280136</title>\n<path d=\"M219.5,-830.366C219.5,-822.152 219.5,-812.658 219.5,-803.725\" fill=\"none\" stroke=\"black\"/>\n<polygon fill=\"black\" points=\"223,-803.607 219.5,-793.607 216,-803.607 223,-803.607\" stroke=\"black\"/>\n</g>\n<!-- 140110097279072 -->\n<g class=\"node\" id=\"node3\"><title>140110097279072</title>\n<polygon fill=\"none\" points=\"3.5,-664.5 3.5,-710.5 435.5,-710.5 435.5,-664.5 3.5,-664.5\" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"136.5\" y=\"-683.8\">batch_normalization_21: BatchNormalization</text>\n<polyline fill=\"none\" points=\"269.5,-664.5 269.5,-710.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"297\" y=\"-695.3\">input:</text>\n<polyline fill=\"none\" points=\"269.5,-687.5 324.5,-687.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"297\" y=\"-672.3\">output:</text>\n<polyline fill=\"none\" points=\"324.5,-664.5 324.5,-710.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"380\" y=\"-695.3\">(None, 64, 64, 8)</text>\n<polyline fill=\"none\" points=\"324.5,-687.5 435.5,-687.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"380\" y=\"-672.3\">(None, 64, 64, 8)</text>\n</g>\n<!-- 140110097280136&#45;&gt;140110097279072 -->\n<g class=\"edge\" id=\"edge3\"><title>140110097280136-&gt;140110097279072</title>\n<path d=\"M219.5,-747.366C219.5,-739.152 219.5,-729.658 219.5,-720.725\" fill=\"none\" stroke=\"black\"/>\n<polygon fill=\"black\" points=\"223,-720.607 219.5,-710.607 216,-720.607 223,-720.607\" stroke=\"black\"/>\n</g>\n<!-- 140110097405712 -->\n<g class=\"node\" id=\"node4\"><title>140110097405712</title>\n<polygon fill=\"none\" points=\"68,-581.5 68,-627.5 371,-627.5 371,-581.5 68,-581.5\" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"133\" y=\"-600.8\">conv2d_22: Conv2D</text>\n<polyline fill=\"none\" points=\"198,-581.5 198,-627.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"225.5\" y=\"-612.3\">input:</text>\n<polyline fill=\"none\" points=\"198,-604.5 253,-604.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"225.5\" y=\"-589.3\">output:</text>\n<polyline fill=\"none\" points=\"253,-581.5 253,-627.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"312\" y=\"-612.3\">(None, 64, 64, 8)</text>\n<polyline fill=\"none\" points=\"253,-604.5 371,-604.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"312\" y=\"-589.3\">(None, 64, 64, 16)</text>\n</g>\n<!-- 140110097279072&#45;&gt;140110097405712 -->\n<g class=\"edge\" id=\"edge4\"><title>140110097279072-&gt;140110097405712</title>\n<path d=\"M219.5,-664.366C219.5,-656.152 219.5,-646.658 219.5,-637.725\" fill=\"none\" stroke=\"black\"/>\n<polygon fill=\"black\" points=\"223,-637.607 219.5,-627.607 216,-637.607 223,-637.607\" stroke=\"black\"/>\n</g>\n<!-- 140110097057888 -->\n<g class=\"node\" id=\"node5\"><title>140110097057888</title>\n<polygon fill=\"none\" points=\"26,-498.5 26,-544.5 413,-544.5 413,-498.5 26,-498.5\" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"133\" y=\"-517.8\">max_pooling2d_22: MaxPooling2D</text>\n<polyline fill=\"none\" points=\"240,-498.5 240,-544.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"267.5\" y=\"-529.3\">input:</text>\n<polyline fill=\"none\" points=\"240,-521.5 295,-521.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"267.5\" y=\"-506.3\">output:</text>\n<polyline fill=\"none\" points=\"295,-498.5 295,-544.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"354\" y=\"-529.3\">(None, 64, 64, 16)</text>\n<polyline fill=\"none\" points=\"295,-521.5 413,-521.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"354\" y=\"-506.3\">(None, 32, 32, 16)</text>\n</g>\n<!-- 140110097405712&#45;&gt;140110097057888 -->\n<g class=\"edge\" id=\"edge5\"><title>140110097405712-&gt;140110097057888</title>\n<path d=\"M219.5,-581.366C219.5,-573.152 219.5,-563.658 219.5,-554.725\" fill=\"none\" stroke=\"black\"/>\n<polygon fill=\"black\" points=\"223,-554.607 219.5,-544.607 216,-554.607 223,-554.607\" stroke=\"black\"/>\n</g>\n<!-- 140110097278232 -->\n<g class=\"node\" id=\"node6\"><title>140110097278232</title>\n<polygon fill=\"none\" points=\"0,-415.5 0,-461.5 439,-461.5 439,-415.5 0,-415.5\" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"133\" y=\"-434.8\">batch_normalization_22: BatchNormalization</text>\n<polyline fill=\"none\" points=\"266,-415.5 266,-461.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"293.5\" y=\"-446.3\">input:</text>\n<polyline fill=\"none\" points=\"266,-438.5 321,-438.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"293.5\" y=\"-423.3\">output:</text>\n<polyline fill=\"none\" points=\"321,-415.5 321,-461.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"380\" y=\"-446.3\">(None, 32, 32, 16)</text>\n<polyline fill=\"none\" points=\"321,-438.5 439,-438.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"380\" y=\"-423.3\">(None, 32, 32, 16)</text>\n</g>\n<!-- 140110097057888&#45;&gt;140110097278232 -->\n<g class=\"edge\" id=\"edge6\"><title>140110097057888-&gt;140110097278232</title>\n<path d=\"M219.5,-498.366C219.5,-490.152 219.5,-480.658 219.5,-471.725\" fill=\"none\" stroke=\"black\"/>\n<polygon fill=\"black\" points=\"223,-471.607 219.5,-461.607 216,-471.607 223,-471.607\" stroke=\"black\"/>\n</g>\n<!-- 140110096238464 -->\n<g class=\"node\" id=\"node7\"><title>140110096238464</title>\n<polygon fill=\"none\" points=\"67.5,-332.5 67.5,-378.5 371.5,-378.5 371.5,-332.5 67.5,-332.5\" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"133\" y=\"-351.8\">dropout_21: Dropout</text>\n<polyline fill=\"none\" points=\"198.5,-332.5 198.5,-378.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"226\" y=\"-363.3\">input:</text>\n<polyline fill=\"none\" points=\"198.5,-355.5 253.5,-355.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"226\" y=\"-340.3\">output:</text>\n<polyline fill=\"none\" points=\"253.5,-332.5 253.5,-378.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"312.5\" y=\"-363.3\">(None, 32, 32, 16)</text>\n<polyline fill=\"none\" points=\"253.5,-355.5 371.5,-355.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"312.5\" y=\"-340.3\">(None, 32, 32, 16)</text>\n</g>\n<!-- 140110097278232&#45;&gt;140110096238464 -->\n<g class=\"edge\" id=\"edge7\"><title>140110097278232-&gt;140110096238464</title>\n<path d=\"M219.5,-415.366C219.5,-407.152 219.5,-397.658 219.5,-388.725\" fill=\"none\" stroke=\"black\"/>\n<polygon fill=\"black\" points=\"223,-388.607 219.5,-378.607 216,-388.607 223,-388.607\" stroke=\"black\"/>\n</g>\n<!-- 140110097567192 -->\n<g class=\"node\" id=\"node8\"><title>140110097567192</title>\n<polygon fill=\"none\" points=\"75,-249.5 75,-295.5 364,-295.5 364,-249.5 75,-249.5\" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"133\" y=\"-268.8\">flatten_11: Flatten</text>\n<polyline fill=\"none\" points=\"191,-249.5 191,-295.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"218.5\" y=\"-280.3\">input:</text>\n<polyline fill=\"none\" points=\"191,-272.5 246,-272.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"218.5\" y=\"-257.3\">output:</text>\n<polyline fill=\"none\" points=\"246,-249.5 246,-295.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"305\" y=\"-280.3\">(None, 32, 32, 16)</text>\n<polyline fill=\"none\" points=\"246,-272.5 364,-272.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"305\" y=\"-257.3\">(None, 16384)</text>\n</g>\n<!-- 140110096238464&#45;&gt;140110097567192 -->\n<g class=\"edge\" id=\"edge8\"><title>140110096238464-&gt;140110097567192</title>\n<path d=\"M219.5,-332.366C219.5,-324.152 219.5,-314.658 219.5,-305.725\" fill=\"none\" stroke=\"black\"/>\n<polygon fill=\"black\" points=\"223,-305.607 219.5,-295.607 216,-305.607 223,-305.607\" stroke=\"black\"/>\n</g>\n<!-- 140110097059512 -->\n<g class=\"node\" id=\"node9\"><title>140110097059512</title>\n<polygon fill=\"none\" points=\"89.5,-166.5 89.5,-212.5 349.5,-212.5 349.5,-166.5 89.5,-166.5\" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"144\" y=\"-185.8\">dense_21: Dense</text>\n<polyline fill=\"none\" points=\"198.5,-166.5 198.5,-212.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"226\" y=\"-197.3\">input:</text>\n<polyline fill=\"none\" points=\"198.5,-189.5 253.5,-189.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"226\" y=\"-174.3\">output:</text>\n<polyline fill=\"none\" points=\"253.5,-166.5 253.5,-212.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"301.5\" y=\"-197.3\">(None, 16384)</text>\n<polyline fill=\"none\" points=\"253.5,-189.5 349.5,-189.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"301.5\" y=\"-174.3\">(None, 1000)</text>\n</g>\n<!-- 140110097567192&#45;&gt;140110097059512 -->\n<g class=\"edge\" id=\"edge9\"><title>140110097567192-&gt;140110097059512</title>\n<path d=\"M219.5,-249.366C219.5,-241.152 219.5,-231.658 219.5,-222.725\" fill=\"none\" stroke=\"black\"/>\n<polygon fill=\"black\" points=\"223,-222.607 219.5,-212.607 216,-222.607 223,-222.607\" stroke=\"black\"/>\n</g>\n<!-- 140110096004544 -->\n<g class=\"node\" id=\"node10\"><title>140110096004544</title>\n<polygon fill=\"none\" points=\"82,-83.5 82,-129.5 357,-129.5 357,-83.5 82,-83.5\" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"147.5\" y=\"-102.8\">dropout_22: Dropout</text>\n<polyline fill=\"none\" points=\"213,-83.5 213,-129.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"240.5\" y=\"-114.3\">input:</text>\n<polyline fill=\"none\" points=\"213,-106.5 268,-106.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"240.5\" y=\"-91.3\">output:</text>\n<polyline fill=\"none\" points=\"268,-83.5 268,-129.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"312.5\" y=\"-114.3\">(None, 1000)</text>\n<polyline fill=\"none\" points=\"268,-106.5 357,-106.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"312.5\" y=\"-91.3\">(None, 1000)</text>\n</g>\n<!-- 140110097059512&#45;&gt;140110096004544 -->\n<g class=\"edge\" id=\"edge10\"><title>140110097059512-&gt;140110096004544</title>\n<path d=\"M219.5,-166.366C219.5,-158.152 219.5,-148.658 219.5,-139.725\" fill=\"none\" stroke=\"black\"/>\n<polygon fill=\"black\" points=\"223,-139.607 219.5,-129.607 216,-139.607 223,-139.607\" stroke=\"black\"/>\n</g>\n<!-- 140110094867648 -->\n<g class=\"node\" id=\"node11\"><title>140110094867648</title>\n<polygon fill=\"none\" points=\"93,-0.5 93,-46.5 346,-46.5 346,-0.5 93,-0.5\" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"147.5\" y=\"-19.8\">dense_22: Dense</text>\n<polyline fill=\"none\" points=\"202,-0.5 202,-46.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"229.5\" y=\"-31.3\">input:</text>\n<polyline fill=\"none\" points=\"202,-23.5 257,-23.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"229.5\" y=\"-8.3\">output:</text>\n<polyline fill=\"none\" points=\"257,-0.5 257,-46.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"301.5\" y=\"-31.3\">(None, 1000)</text>\n<polyline fill=\"none\" points=\"257,-23.5 346,-23.5 \" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"301.5\" y=\"-8.3\">(None, 2)</text>\n</g>\n<!-- 140110096004544&#45;&gt;140110094867648 -->\n<g class=\"edge\" id=\"edge11\"><title>140110096004544-&gt;140110094867648</title>\n<path d=\"M219.5,-83.3664C219.5,-75.1516 219.5,-65.6579 219.5,-56.7252\" fill=\"none\" stroke=\"black\"/>\n<polygon fill=\"black\" points=\"223,-56.6068 219.5,-46.6068 216,-56.6069 223,-56.6068\" stroke=\"black\"/>\n</g>\n<!-- 140110097277280 -->\n<g class=\"node\" id=\"node12\"><title>140110097277280</title>\n<polygon fill=\"none\" points=\"160.5,-913.5 160.5,-949.5 278.5,-949.5 278.5,-913.5 160.5,-913.5\" stroke=\"black\"/>\n<text font-family=\"Times,serif\" font-size=\"14.00\" text-anchor=\"middle\" x=\"219.5\" y=\"-927.8\">140110097277280</text>\n</g>\n<!-- 140110097277280&#45;&gt;140110097278792 -->\n<g class=\"edge\" id=\"edge1\"><title>140110097277280-&gt;140110097278792</title>\n<path d=\"M219.5,-913.254C219.5,-905.363 219.5,-895.749 219.5,-886.602\" fill=\"none\" stroke=\"black\"/>\n<polygon fill=\"black\" points=\"223,-886.591 219.5,-876.591 216,-886.591 223,-886.591\" stroke=\"black\"/>\n</g>\n</g>\n</svg>"},"metadata":{}}]},{"metadata":{"trusted":true,"_uuid":"743b6aff6cdde97fbd7f177237d2f2d95016e4e2","scrolled":true},"cell_type":"code","source":"history = model.fit(X_train, y_train, validation_split = 0.1, batch_size = 50, epochs = 20, verbose = 1)","execution_count":48,"outputs":[{"output_type":"stream","text":"Train on 24483 samples, validate on 2721 samples\nEpoch 1/20\n24483/24483 [==============================] - 7s 299us/step - loss: 0.6622 - acc: 0.7486 - val_loss: 0.4688 - val_acc: 0.7685\nEpoch 2/20\n24483/24483 [==============================] - 6s 233us/step - loss: 0.4707 - acc: 0.7784 - val_loss: 0.4674 - val_acc: 0.7762\nEpoch 3/20\n24483/24483 [==============================] - 6s 232us/step - loss: 0.4568 - acc: 0.7854 - val_loss: 0.4564 - val_acc: 0.7832\nEpoch 4/20\n24483/24483 [==============================] - 6s 233us/step - loss: 0.4468 - acc: 0.7918 - val_loss: 0.4388 - val_acc: 0.7924\nEpoch 5/20\n24483/24483 [==============================] - 6s 233us/step - loss: 0.4367 - acc: 0.7962 - val_loss: 0.4410 - val_acc: 0.7898\nEpoch 6/20\n24483/24483 [==============================] - 6s 235us/step - loss: 0.4273 - acc: 0.8034 - val_loss: 0.4653 - val_acc: 0.7832\nEpoch 7/20\n24483/24483 [==============================] - 6s 233us/step - loss: 0.4158 - acc: 0.8104 - val_loss: 0.4358 - val_acc: 0.7946\nEpoch 8/20\n24483/24483 [==============================] - 6s 232us/step - loss: 0.4053 - acc: 0.8140 - val_loss: 0.4248 - val_acc: 0.8001\nEpoch 9/20\n24483/24483 [==============================] - 6s 231us/step - loss: 0.3899 - acc: 0.8226 - val_loss: 0.4309 - val_acc: 0.7975\nEpoch 10/20\n24483/24483 [==============================] - 6s 237us/step - loss: 0.3785 - acc: 0.8292 - val_loss: 0.4318 - val_acc: 0.7964\nEpoch 11/20\n24483/24483 [==============================] - 6s 240us/step - loss: 0.3712 - acc: 0.8367 - val_loss: 0.4479 - val_acc: 0.8056\nEpoch 12/20\n24483/24483 [==============================] - 6s 233us/step - loss: 0.3512 - acc: 0.8437 - val_loss: 0.4255 - val_acc: 0.8118\nEpoch 13/20\n24483/24483 [==============================] - 6s 231us/step - loss: 0.3320 - acc: 0.8576 - val_loss: 0.4311 - val_acc: 0.8067\nEpoch 14/20\n24483/24483 [==============================] - 6s 231us/step - loss: 0.3157 - acc: 0.8608 - val_loss: 0.4353 - val_acc: 0.8093\nEpoch 15/20\n24483/24483 [==============================] - 6s 232us/step - loss: 0.2964 - acc: 0.8736 - val_loss: 0.4198 - val_acc: 0.8181\nEpoch 16/20\n24483/24483 [==============================] - 6s 231us/step - loss: 0.2780 - acc: 0.8814 - val_loss: 0.4237 - val_acc: 0.8291\nEpoch 17/20\n24483/24483 [==============================] - 6s 231us/step - loss: 0.2577 - acc: 0.8920 - val_loss: 0.4604 - val_acc: 0.8166\nEpoch 18/20\n24483/24483 [==============================] - 6s 231us/step - loss: 0.2383 - acc: 0.9029 - val_loss: 0.4448 - val_acc: 0.8284\nEpoch 19/20\n24483/24483 [==============================] - 6s 231us/step - loss: 0.2247 - acc: 0.9077 - val_loss: 0.4795 - val_acc: 0.8196\nEpoch 20/20\n24483/24483 [==============================] - 6s 231us/step - loss: 0.2043 - acc: 0.9171 - val_loss: 0.4700 - val_acc: 0.8181\n","name":"stdout"}]},{"metadata":{"trusted":true},"cell_type":"code","source":"# Plot training & validation accuracy values\nplt.plot(history.history['acc'])\nplt.plot(history.history['val_acc'])\nplt.title('Model accuracy')\nplt.ylabel('Accuracy')\nplt.xlabel('Epoch')\nplt.legend(['Train', 'Test'], loc='upper left')\nplt.show()","execution_count":49,"outputs":[{"output_type":"display_data","data":{"text/plain":"<Figure size 432x288 with 1 Axes>","image/png":"iVBORw0KGgoAAAANSUhEUgAAAZIAAAEWCAYAAABMoxE0AAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzs3Xd4VGX2wPHvSSOUQCAJvYUqTVooomADRSzYFcWKIr9de1lx13UV1111V10LKwsKKAqIWEAFARFFVJAAoUsIPSFAKCFACGnn98e90SGkDCSTSTmf58nDzL3vvXNmgDl5u6gqxhhjzJkK8HcAxhhjKjZLJMYYY0rEEokxxpgSsURijDGmRCyRGGOMKRFLJMYYY0rEEokxhRCRliKiIhLkRdk7RWRJWcRlTHljicRUCiKyXUQyRSQy3/E4Nxm09E9kxlR+lkhMZbINGJb3RES6ANX9F0754E2NypiSsERiKpMpwO0ez+8A3vcsICJ1ROR9EUkRkR0i8rSIBLjnAkXk3yKyX0S2ApcXcO27IpIsIkki8ncRCfQmMBH5WET2iMhhEVksIp08zlUXkVfceA6LyBIRqe6eO09EfhKRVBHZJSJ3use/E5F7PO5xUtOaWwv7o4hsBja7x15375EmIitEpL9H+UAR+bOIbBGRI+75ZiIyVkReyfdevhCRh71536ZqsERiKpOlQG0R6eB+wd8EfJCvzJtAHaAVcD5O4rnLPXcvcAXQHYgBrs937XtANtDGLXMJcA/emQu0BeoDK4EPPc79G+gJ9APqAX8CckWkuXvdm0AU0A2I8/L1AK4G+gAd3efL3XvUA6YCH4tIqHvuUZza3BCgNnA3kO6+52EeyTYSuBiYdhpxmMpOVe3Hfir8D7AdGAg8DfwTGAwsAIIABVoCgcAJoKPHdfcB37mPvwVGeZy7xL02CGjgXlvd4/wwYJH7+E5giZexhrv3rYPzy9xxoGsB5Z4CPivkHt8B93g8P+n13ftfVEwch/JeF9gEDC2k3EZgkPv4fmCOv/++7ad8/VjbqalspgCLgWjyNWsBkUAIsMPj2A6gifu4MbAr37k8LYBgIFlE8o4F5CtfILd29AJwA07NItcjnmpAKLClgEubFXLcWyfFJiKP4dSgGuMkmtpuDMW91nvAcJzEPBx4vQQxmUrImrZMpaKqO3A63YcAn+Y7vR/IwkkKeZoDSe7jZJwvVM9zeXbh1EgiVTXc/amtqp0o3i3AUJwaUx2c2hGAuDFlAK0LuG5XIccBjgE1PJ43LKDMb0t7u/0hTwI3AnVVNRw47MZQ3Gt9AAwVka5AB+DzQsqZKsoSiamMRuA06xzzPKiqOcAM4AURCRORFjh9A3n9KDOAB0WkqYjUBUZ7XJsMzAdeEZHaIhIgIq1F5Hwv4gnDSUIHcL78/+Fx31xgIvCqiDR2O73PEZFqOP0oA0XkRhEJEpEIEenmXhoHXCsiNUSkjfuei4shG0gBgkTkGZwaSZ53gOdFpK04zhaRCDfGRJz+lSnAJ6p63Iv3bKoQSySm0lHVLaoaW8jpB3B+m98KLMHpdJ7onpsAzANW43SI56/R3I7TNLYBp39hJtDIi5Dex2kmS3KvXZrv/OPAWpwv64PAS0CAqu7EqVk95h6PA7q617wGZAJ7cZqePqRo83A67uPdWDI4uenrVZxEOh9IA97l5KHT7wFdcJKJMScRVdvYyhhTNBEZgFNza+nWooz5jdVIjDFFEpFg4CHgHUsipiCWSIwxhRKRDkAqThPef/wcjimnfJpIRGSwiGwSkQQRGV3A+RYislBE1rgzdZu6x7uJyM8ist49d5PHNZNFZJu7hlKcR+ejMaaUqepGVa2pqv1UNc3f8ZjyyWd9JO7Y+XhgEJA36mOYqm7wKPMx8KWqviciFwF3qeptItIOUFXdLCKNgRVAB1VNFZHJ7jUzfRK4McaY0+LLCYm9gQRV3QogItNxxtJv8CjTEXjEfbwId3y6qsbnFVDV3SKyD2ciV+qZBBIZGaktW7Y8k0uNMabKWrFixX5VjSqunC8TSRNOHl6YiLPuj6fVwHU4M2WvAcJEJEJVD+QVEJHeOEMuPWfdvuCOg18IjFbVE/lfXERGAiMBmjdvTmxsYaNBjTHGFEREdhRfyrd9JFLAsfztaI8D54vIKpwF9JJwJk05NxBphDNu/S6P0SJPAWcBvXAWn3uyoBdX1fGqGqOqMVFRxSZUY4wxZ8iXNZJETl5uoimw27OAqu4GrgUQkVrAdap62H1eG/gKeFpVl3pck+w+PCEik3CSkTHGGD/xZY1kOdBWRKJFJAS4GZjtWUBEIvOWp8apaUx0j4cAnwHvq+rH+a5p5P4pOMtkr/PhezDGGFMMn9VIVDVbRO7HWZohEJioqutFZAwQq6qzgQuAf4qI4qzY+kf38huBAUBE3kY+wJ2qGgd8KCJROE1nccCoM4kvKyuLxMREMjIyzuwNVjChoaE0bdqU4OBgf4dijKlkqsQSKTExMZq/s33btm2EhYURERGBx7LglZKqcuDAAY4cOUJ0dLS/wzHGVBAiskJVY4orV2VntmdkZFSJJAIgIkRERFSZ2pcxpmxV2UQCVIkkkqcqvVdjTNmq0onEGGMqq92px3nui/Vk5fh+nU1LJH5y4MABunXrRrdu3WjYsCFNmjT57XlmZqZX97jrrrvYtGmTjyM1xlQkqsrnq5K49D+L+Wj5Ln5NPuLz17Q92/0kIiKCuLg4AJ599llq1arF44+fPCVGVVFVAgIKzveTJk3yeZzGmIrj0LFMnv58HV+tTaZni7q8emNXWkTU9PnrWo2knElISKBz586MGjWKHj16kJyczMiRI4mJiaFTp06MGTPmt7LnnXcecXFxZGdnEx4ezujRo+natSvnnHMO+/bt8+O7MMaUtUWb9nHJfxYzf8Me/jS4PTPuO6dMkghYjQSA575Yz4bdpbtCdsfGtfnblZ3O6NoNGzYwadIkxo0bB8CLL75IvXr1yM7O5sILL+T666+nY8eOJ11z+PBhzj//fF588UUeffRRJk6cyOjRp6zcb4ypZI6dyOaFORuZumwn7RuEMfmuXnRqXKdMY7BEUg61bt2aXr16/fZ82rRpvPvuu2RnZ7N79242bNhwSiKpXr06l112GQA9e/bkhx9+KNOYjTFlb8WOQzw6I46dB9MZOaAVjw5qR2hwYJnHYYkEzrjm4Cs1a/5eHd28eTOvv/46v/zyC+Hh4QwfPrzA+SAhISG/PQ4MDCQ7O/uUMsaYyiEzO5fXF8bz9ndbaFSnOtPv7UufVhF+i8f6SMq5tLQ0wsLCqF27NsnJycybN8/fIRlj/Ch+7xGu+e+PjF20het7NuXrh/v7NYmA1UjKvR49etCxY0c6d+5Mq1atOPfcc/0dkjHGD3JzlXeXbONf8zcRVi2I8bf15JJODf0dFlCF19rauHEjHTp08FNE/lEV37MxlUHioXQem7GaZdsOMqhjA/55bRcia1Xz+et6u9aW1UiMMaacUlVmrkjkuS+cHcr/df3ZXN+zablb8sgSiTHGlEO/7knj1fnxzN+wl97R9Xjlhq40q1fD32EVyBKJMcaUE4mH0pm9ejez43bz654jhAQF8JchHRhxXjQBAeWrFuLJEokxxvjRwWOZfLU2mdlxSSzffgiAni3qMmZoJ4Z0aVQmfSElZYnEGGPKWHpmNgs27GVW3G4Wx6eQnau0rV+LJy5tz1VdG5fbJqzCWCIxxpgykJWTyw+bU5gVt5v56/dyPCuHxnVCGdE/mqFdm9ChUVi560T3liUSPzlw4AAXX3wxAHv27CEwMJCoqCgAfvnll5Nmqhdl4sSJDBkyhIYNy8d4cmPM73JzlRU7DzErLomv1iRzKD2L8BrBXNOjCUO7NqZXy3rluu/DW5ZI/MSbZeS9MXHiRHr06GGJxJhyZv76PTz3xQaSUo8TGhzAoI4NGdq1MQPaRRESVLkWFfFpIhGRwcDrQCDwjqq+mO98C2AiEAUcBIaraqJ77g7gabfo31X1Pfd4T2AyUB2YAzyklWxW5XvvvcfYsWPJzMykX79+vPXWW+Tm5nLXXXcRFxeHqjJy5EgaNGhAXFwcN910E9WrVz+tmowxxjdyc5X/fBPPG98m0KlxbR6/tB2XdGxIzWqV9/d2n70zEQkExgKDgERguYjMVtUNHsX+Dbyvqu+JyEXAP4HbRKQe8DcgBlBghXvtIeBtYCSwFCeRDAbmlijYuaNhz9oS3eIUDbvAZS8WXy6fdevW8dlnn/HTTz8RFBTEyJEjmT59Oq1bt2b//v2sXevEmZqaSnh4OG+++SZvvfUW3bp1K934jTGn7fDxLB6evopFm1K4MaYpY4Z29stqvGXNlymyN5CgqlsBRGQ6MBTwTCQdgUfcx4uAz93HlwILVPWge+0CYLCIfAfUVtWf3ePvA1dT0kRSjnzzzTcsX76cmBhnVYLjx4/TrFkzLr30UjZt2sRDDz3EkCFDuOSSS/wcqTHG06Y9R7hvSixJqcf5+9WdubVP8wrbeX66fJlImgC7PJ4nAn3ylVkNXIfT/HUNECYiEYVc28T9SSzg+ClEZCROzYXmzZsXHekZ1Bx8RVW5++67ef755085t2bNGubOncsbb7zBJ598wvjx4/0QoTEmv6/WJPPEzNXUrBbE9JF96dminr9DKlO+7PEpKBXn78t4HDhfRFYB5wNJQHYR13pzT+eg6nhVjVHVmLzRUBXBwIEDmTFjBvv37wec0V07d+4kJSUFVeWGG27gueeeY+XKlQCEhYVx5MgRf4ZsTJWVk6v8c+5G/jh1JWc1DOPLB86rckkEfFsjSQSaeTxvCuz2LKCqu4FrAUSkFnCdqh4WkUTggnzXfufes2lR96zounTpwt/+9jcGDhxIbm4uwcHBjBs3jsDAQEaMGIGqIiK89NJLANx1113cc8891tluTBk7dCyTB6ev4ofN+7m1T3P+dmWnSjcay1s+W0ZeRIKAeOBinJrGcuAWVV3vUSYSOKiquSLyApCjqs+4ne0rgB5u0ZVAT1U9KCLLgQeAZTid7W+q6pyiYrFl5B1V8T0b4wvrdx/mvikr2Jd2guev7sRNvYppPq+gvF1G3mfpU1WzgfuBecBGYIaqrheRMSJylVvsAmCTiMQDDYAX3GsPAs/jJJ/lwJi8jnfg/4B3gARgC5Woo90YU/7Nikviurd/IjtHmTHqnEqbRE6HTwc2uzWFOfmOPePxeCYws5BrJ+LMMcl/PBboXLqRGmNM0bJzcvnn3F95d8k2ekfXY+wtPYgKK/8LKpaFyjtDxgt5/Q1VQSWbs2lMmdp/9AT3T13J0q0HubNfS/5yeQeCA6tmf0hBqmwiCQ0N5cCBA0RERFT6ZKKqHDhwgNDQUH+HYkyFsyYxlVFTVnDgWCav3tiVa3s0Lf6iKqbKJpKmTZuSmJhISkqKv0MpE6GhoTRtav8BjDkdH8fu4i+fryOqVjU++b9+dG5Sx98hlUtVNpEEBwcTHR3t7zCMMeXQiewcnv9yAx8s3cm5bSJ4c1gP6tW0ofWFqbKJxBhjCpKUepw/fLCC1YmHuW9AK564tD1B1h9SJEskxhjjWhyfwkPTV5GVo4wb3pPBnW17Bm9YIjHGVHm5ucpbixJ47Zt42tUP4+3hPWgVVcvfYVUYlkiMMVVaanomj3wUx6JNKVzTvQkvXNOZGiH21Xg67NMyxlRZ65IOM+qDFexNy+D5qzszvAot/V6aLJEYY6qkj5bv5K+z1hNZM4SPR/WjW7Nwf4dUYVkiMcZUKRlZOTwzax0zYhPp3zaS12/ubkN7S8gSiTGmyth5IJ1RH6xgQ3IaD17UhocGtiMwwJqySsoSiTGmSli4cS+PfBQHwMQ7Y7jorAZ+jqjysERijKnUcnKV1xbE89aiBDo1rs244T1pVq+Gv8OqVCyRGGMqrQNHT/DQ9DiWJOznpphmPDe0E6HBgf4Oq9KxRGKMqZQWx6fw5CdrOHAsk5eu62IbUPmQJRJjTKWyJeUoL3y1kW9/3Ud0ZE0+tVV7fc4SiTGmUkhNz+T1hZuZ8vMOqgcH8uchZ3FHv5ZUC7KmLF+zRGKMqdCyc3KZ+stOXl0QT9rxLG7u3ZxHB7UjspZtg1tWLJEYYyqs7+NT+PuXG9i87yjntIrgmSs70qFRbX+HVeX4NJGIyGDgdSAQeEdVX8x3vjnwHhDulhmtqnNE5FbgCY+iZwM9VDVORL4DGgHH3XOXqOo+X74PY0z5krDvKC98tYFFm1JoEVGD8bf1ZFDHBrZOlp/4LJGISCAwFhgEJALLRWS2qm7wKPY0MENV3xaRjsAcoKWqfgh86N6nCzBLVeM8rrtVVWN9FbsxpnyyfpDyyZc1kt5AgqpuBRCR6cBQwDORKJBXD60D7C7gPsOAaT6M0xhTzmXl5DJ12U5e+8b6QcojXyaSJsAuj+eJQJ98ZZ4F5ovIA0BNYGAB97kJJwF5miQiOcAnwN9VVUslYmNMufN9fArPf7mBBOsHKbd8mUgKaqzM/4U/DJisqq+IyDnAFBHprKq5ACLSB0hX1XUe19yqqkkiEoaTSG4D3j/lxUVGAiMBmje3iUjGVDRJqcd5+rO11g9SAfgykSQCzTyeN+XUpqsRwGAAVf1ZREKBSCCv8/xm8jVrqWqS++cREZmK04R2SiJR1fHAeICYmBirsRhTgSTsO8pt7y4j7XiW9YNUAAE+vPdyoK2IRItICE5SmJ2vzE7gYgAR6QCEAinu8wDgBmB6XmERCRKRSPdxMHAFsA5jTKWxLukwN/3vZ7Jycvl4VD9GDmhtSaSc81mNRFWzReR+YB7O0N6JqrpeRMYAsao6G3gMmCAij+A0e93p0d8xAEjM66x3VQPmuUkkEPgGmOCr92CMKVux2w9y1+TlhFUL4oN7+tAqqpa/QzJekKrQTx0TE6OxsTZa2Jjy7Pv4FO6bEkvjOtWZck8fmoRX93dIVZ6IrFDVmOLK2cx2Y4zfzV2bzIPTV9Gmfhjv392bqDAb1luRWCIxxvjVjNhdjP5kDd2ahTPpzt7UqRHs75DMabJEYozxm4lLtjHmyw30bxvJ/27rSY0Q+0qqiOxvzRhT5lSVNxYm8No38VzaqQFvDOtuI7MqMEskxpgypar8/auNvLtkG9f1aMpL13UhKNCXMxGMr1kiMcaUmZxc5c+fruWj2F3c2a8lz1zRkYAAm6le0VkiMcaUiczsXB75KI6v1ibz4EVteGRQO1vupJKwRGKM8bnjmTmM+mAF38en8JchHbh3QCt/h2RKkSUSY4xPpWVkMWLycmJ3HOLFa7twc29bRLWysURijPGZA0dPcMekX9i05whvDuvOFWc39ndIxgcskRhjSl1aRhYrdxzi+S83kHjoOONvj+HC9vX9HZbxEUskxpgS2516nNgdh4jdfpDl2w/x6540VCEsNIj37+5Nn1YR/g7R+JAlEmPMacnJVeL3HiF2+0E3eRwiKfU4ADVDAunRoi4PXdyWXi3r0a1ZODWr2ddMZWd/w8aYImVk5RC3K/W3xLFixyGOZGQDUD+sGr2i63FP/2h6tazHWQ3DbHJhFWSJxBhToB8T9vPv+ZtYl3SYrBxnu4l2DWpxZdfGxLSoS6+W9What7rNBTGWSIwxp1qyeT8j3ltOozqh3NO/Fb1a1qVH87qE1wjxd2imHLJEYow5yU9b9nPP+8uJjqzJ1Hv7Uq+mJQ9TNGvMNMb8ZtnWA4yYHEvzejX48J4+lkSMVyyRGGOA3/dLbxweyof39CWilu1SaLxjicQYw8qdh7hz0nIa1g5l2r19batbc1oskRhTxa3elcod7/5CRK0Qpt7bl/q1Q/0dkqlgfJpIRGSwiGwSkQQRGV3A+eYiskhEVonIGhEZ4h5vKSLHRSTO/RnncU1PEVnr3vMNsbGHxpyxtYmHue3dZYTXDGbavX1pWMeSiDl9PkskIhIIjAUuAzoCw0SkY75iTwMzVLU7cDPwX49zW1S1m/szyuP428BIoK37M9hX78GYymz97sMMf3cZYaFOEmkcXt3fIZkKypc1kt5AgqpuVdVMYDowNF8ZBWq7j+sAu4u6oYg0Amqr6s+qqsD7wNWlG7Yxld+ve9IY/s4yaoYEMu3evjStW8PfIZkKrNhEIiL3i0jdM7h3E2CXx/NE95inZ4HhIpIIzAEe8DgX7TZ5fS8i/T3umVjMPfPiHikisSISm5KScgbhG1M5xe89wq0TllEtKJCp9/aleYQlEVMy3tRIGgLLRWSG2+fhbZ9EQeU03/NhwGRVbQoMAaaISACQDDR3m7weBaaKSG0v7+kcVB2vqjGqGhMVFeVlyMZUbgn7jnDLhKUEBghT7+1Dy8ia/g7JVALFJhJVfRqnL+Jd4E5gs4j8Q0RaF3NpItDM43lTTm26GgHMcF/nZyAUiFTVE6p6wD2+AtgCtHPv2bSYexpjCrA15SjDJiwDhKn39qVVVC1/h2QqCa/6SNz+iD3uTzZQF5gpIi8XcdlyoK2IRItICE5n+ux8ZXYCFwOISAecRJIiIlFuZz0i0gonkW1V1WTgiIj0dWtGtwOzvHurxlRd2/cfY9iEpeTmKlPv7UOb+pZETOkpdq0tEXkQuAPYD7wDPKGqWW4T1GbgTwVdp6rZInI/MA8IBCaq6noRGQPEqups4DFggog8gtNEdaeqqogMAMaISDaQA4xS1YPurf8PmAxUB+a6P8aYQuw8kM6wCUvJzM5l2si+tGsQ5u+QTCUjTmWjiALOF/+7qrqjgHMdVHWjr4IrLTExMRobG+vvMIwpc7sOpnPz+KUcPZHN1Hv70KlxHX+HZCoQEVmhqjHFlfOmaWsOkFcbQETCRKQPQEVIIsZUVVtSjnLLO0s5kpHFh/dYEjG+480y8m8DPTyeHyvgmDGmnEg8lM6bCxOYuTKRGiGBfDCiD52bWBIxvuNNIhH1aP9S1VwRsX1MjCln9qZl8Na3CUxfvhNBuK1vC/5wYWvqh9myJ8a3vEkIW90O97fd538AtvouJGPM6dh/9ATjvtvClKU7yMlVbuzVjPsvbGNLnpgy400iGQW8gbMulgILcda6Msb4UWp6JuMXb2XyT9vJyMrhmu5NeejitjZT3ZS5YhOJqu7DmQNijCkHjmRkMXHJdt75YStHM7O54uzGPDywLa1tgqHxE2/mkYTizEDvhDNhEABVvduHcRlj8knPzOa9n3bwv8VbSE3P4tJODXhkUDvOali7+IuN8SFvmramAL8ClwJjgFsBG/ZrTBnJyMrhw2U7efu7BPYfzeTC9lE8Oqg9XZraSCxTPniTSNqo6g0iMlRV3xORqTiz1Y0xPpSZncuM2F289W0Ce9Iy6Nc6gv/d1o6eLer5OzRjTuJNIsly/0wVkc4462219FlExlRxqsrcdXt4+etf2X4gnZgWdXn1pq70ax3p79CMKZA3iWS8ux/J0ziLLtYC/urTqIypomK3H+SFORtZtTOV9g3CmHRnLy5oH4XtKG3KsyITibswY5qqHgIWA63KJCpjqpitKUd56etfmbd+L/XDqvHSdV24vmczAgMsgZjyr8hE4s5ivx93zxBjTOk6cPQEry/czNRlO6kWFMBjg9oxon80NUJs8QhTcXjzr3WBiDwOfISzzhYAHsu6G2NO0/HMHCb+uI23v9vC8awchvVuxkMXtyMqrJq/QzPmtHmTSPLmi/zR45hizVzGnLacXOXTlYm8Mj+ePWkZDOrYgCcHn2UbTZkKzZuZ7dFlEYgxld3i+BT+MWcjv+45Qtdm4bwxrDu9o20or6n4vJnZfntBx1X1/dIPx5jKZ8PuNP45dyM/bN5P83o1eOuW7lzepZGNxDKVhjdNW708Hofi7LG+ErBEYkwREg+l89qCzXy6KpE61YP56xUdGd63OdWCAv0dWtWQnQmHtkNUO39HUul507T1gOdzEamDs2yKMaYAW1OO8vZ3W/hsVRIBAcLI/q34wwVtqFMj2N+hVR1pu+Gj2yApFgY+B+c9XPYxZByGbT9Au8EQWLlH4Z3Ju0sH2pZ2IMZUdBuT0xi7KIE5a5MJDgxgeN8WjBzQyvYFKWs7l8GM2+DEUWh1AXzzN0jfDwPHQIA3u4uXgkPb4cMbYf8maNQVrnoLGp1dNq/tB970kXyBM0oLnD3eO+LlvBIRGQy8DgQC76jqi/nONwfeA8LdMqNVdY6IDAJeBEKATOAJVf3WveY7oBFw3L3NJe5S98b4RdyuVN76NoFvNu6lVrUgRg5ozYjzom0orz+seA++egzqNIHbPoeos+DrJ+GnN+HYAbjqDQj0cc0wMRam3Qw5mTDwWfj5vzD+Ajj3ITj/SQiufDtWelMj+bfH42xgh6omFneRiAQCY4FBQCKwXERmq+oGj2JPAzNU9W0R6QjMwVnHaz9wparudtf3mgc08bjuVlWN9SJ2Y3xCVVm27SBvfZvAkoT9hNcI5pGB7bizX0trwvKH7EyY9xQsfwdaXwTXvQs13BFxl70MNaNg0Qtw/CBcPwlCfLT514bZ8Om9ENYQbpnj9M/0uAPm/xWWvAobZ8NVb0KLfr55fT/xJpHsBJJVNQNARKqLSEtV3V7Mdb2BBFXd6l43HRgKeCYSBfI2U6gD7AZQ1VUeZdYDoSJSTVVPeBGvMT6jqnwXn8LYbxOI3XGIyFrV+POQs7ilTwtqVavc7eDl1tEUmHE77PwJ+j0AFz97cp+ECJz/J6gR4dRWPrgWhk2D6nVLLwZVp9az4Blo2su5f013kc0a9eDqsdDlevjiIZh0GcSMcGoroZVjLxlv/uV/DHimzxz3WK+Ci/+mCbDL43ki0CdfmWeB+SLyAFATGFjAfa4DVuVLIpNEJAf4BPi7qmr+i0RkJO6WwM2bNy8mVGOKlpurzN+wh7cWJbAuKY0m4dUZM7QTN8Y0IzTYRmH5ze5VMH240wdy7QQ4+8bCy/Ya4Xypf3IvTLochn8CtRuVPIacbJjzOKyYBB2vhmvGQXAB/WKtL4Q//AzfvgDL3ob4r+GK16DdpSWPwc+86XkKUtXMvCfu4xAvritokHz+L/xhwGRVbQoMAaa4C0U6NxDpBLwE3Odxza2q2gXo7/7cVtCLq+p4VY1R1ZioqCgvwjXmVNl6wlj8AAAgAElEQVQ5uXy2KpFL/7OYUR+s5NiJHF6+7mwWPX4Bt5/T0pKIP62ZARMHO4/vnld0EsnT6Rq49WNI3QETL4EDW0oWQ0YaTLvJSSLnPeI0mxWURPKE1ITB/4ARCyC0Dky9EWaOgGP7SxaHn3mTSFJE5Kq8JyIyFKcPoziJQDOP501xm648jMDtuFfVn3HmqUS6r9MU+Ay4XVV/+9tW1ST3zyPAVJwmNGNKlaryxerdXPTK9zzy0WoCRHhjWHe+efR8buzVjJCgMhr9Y06Vkw3z/uL0RTTpCSO/g8bdvL++9YVwxxeQeQzevQR2x51ZHIcTnWaqLYvgytedpipvR4U1jYGR38MFf4YNs+CtXk5iPLVxpULw5l2PAv4sIjtFZCfwJCfXEAqzHGgrItEiEgLcjLOfiaedOBMcEZEOOIkkRUTCga+Ap1T1x7zCIhIkInmJJhi4AljnRSzGeG3XwXTumrycB6atIiw0iPG39WTuQ/25qmtjW9bd39IPwofXw89vQa974fZZUOsMWhya9IC750NwDZh8BWxbfHrX746DCRdD6k4YPhN63nn6MQSFwAVPwqgfIKK1kxg/vAFSdxV/bTkjBXQvFFxQpJZb/ojXNxcZAvwHZ2jvRFV9QUTGALGqOtsdqTUBZ7MsBf6kqvNF5GngKWCzx+0uwVl9eDEQ7N7zG+BRVc0pKo6YmBiNjbVBXqZo2Tm5TPxxG68t2EyAwOOXtuf2c1pa8igv9q6H6bc4kw0vfwV6FLh60+lJ2w1TroWDW5yRXh2vKv6aTV/DzLud/pZbZkCDjiWPIzcHfpkAC58DCXBqNzEjym7eSyFEZIWqxhRbrrhEIiL/AF5W1VT3eV3gMVV9ulQiLQOWSExx1iSm8tSna1m/O42BHRowZmgnm0hYnmyYBZ/9H1QLg5umQLNSbNFOP+jM+0hcDpe/CjF3FV522f/g69HQ8Gy45SNnmG9pOrQDvnwYtnwLzfo6816i2pfua5yG0kwkq1S1e75jK1W1RwljLDOWSExhjp3I5pX58Uz+aRuRtaoxZmgnLu3U0BZULC9yc+G7f8Dif0GTGLjpg9IZaZVfZjp8fAdsng8XPQ39H3eGDf8WRw7M+zMsGwftL4frJjgd576gCqunOwkrIxWq1YawRlC7MdRu4rz/2o0hrLF7rLEztNkH/2a9TSTeDP8N9JzDISLVAZuyayq8hRv38sys9ew+fJzhfVrwxOD21A61yYTlxvFU+GwUxM+FbsOd5ixfzQoPqQE3T4VZf4Rv/+6Morr0n07TUuYx+OQe2DQH+v4RLnkeAnw4Wk8Eug2DNhc7CeVwIhzZ7TTDbdkER/eA5p58TWA1p3bkmWhqN3ESUOsLnRFiPuRNIvkAWCgik9znd+Esa2JMhbQvLYPnvtjAV2uTadegFjNH9aNni1KcnGZKbuOXzuTBYylw2b+g970++Y37JIHBcPU4qBEJS8dC+gG4+Bn4aDjsWQtD/u3EUVZq1YdzHzz1eE42HNvnJJa0JEhLdv48kuwcS1rpfH457tS7+2P9n0hU9WURWYMzWVCAr4EWPo3KGB/IzVWmLd/Ji3N/5UR2Lk9c2p57+7eyobzlydF9MOcJ2PA5NOjszBBvUoat6AEBcOkLzkiwb56F9Z85v+0Pm15+Jg4GBv3epEUhrU6qTt/Pkd0Q7vuva2/XdNgD5AI3AttwZpQbU2HE7z3Cnz9dS+yOQ/RrHcEL13QhOtJHbdzm9KnC6mnw9VOQlQ4X/dVZ5NDXCywWRMSZXFgzCmInOrPPG3Ut+zhKQgRqRjg/ZaDQRCIi7XDmfgwDDgAf4XTOX1gmkRlTCjKychi7KIFx32+hVrUg/n1DV67r0cQ608uTU0YqvVk+NqPqPtz5McUqqkbyK/ADziq8CQAi8kiZRGVMCSWlHmfJ5hTGfb+VbfuPcW33Jvzl8g5E1LJxIuVGbg78Mh4WjnHmTgz5d7mYO2FOX1GJ5DqcGskiEfkamE7B62cZ43eH07P4eet+liTs58eEA2zbfwyA6MiafDCiD+e1jfRzhOYk+zbCrPudHQzbXuLM3whvVvx1plwqNJGo6mfAZyJSE7gaeARoICJvA5+p6vwyitGYU5zIzmHFjkP8mLCfJQkHWJuYSq5CzZBA+raK4La+LTivbSRt69eyZqzyJDvT2Zdj8b+dyYXXvuMsr25/RxWaN6O2jgEfAh+KSD3gBmA0YInElJncXGXjnjSWbHZqHcu3HyQjK5fAAKF7s3AeuKgt57WNpFuzcIIDq3jTiCpsWQg/vuH85t/hSuh2qzP6yZ9f2ImxTi0kZSN0uQEGv/j7nh2mQvN6ra2KzGa2V0wHj2Uyf/0eliTs56ctBzh4zNnNoG39WpzbJpL+bSPpHV2PMJtE6MjJgnWfOBss7V3nTEZrGgObF0B2hrPtbLdb4OybIaxB2cWVecyZ5Lf0bWfIaiXZg6MqKM2Z7caUqdT0TMYv3srkn7aTnplD/bBqXNA+ivPaRHJum0ga1K58e16XSEYarHzP+aJOS4KoDnD129D5emeF2YzDznyIuKnODn7fPAdtBkL3W6HdYAjy4QCELd86uwKm7qx0uwKa31mNxJQbaRlZvPvDNiYu2caRE9lcfnYj/nBBazo2qm39HAVJS3Z22oudDCcOQ8v+0O9BaDuo8Cas/ZudhLJ6ujNZrXpd6HKjU1Np1PXMm75U4eheZwZ48mrnzz1rnVV1I9pUyn3Kq4JSW7SxMrBEUr4dO5HN5J+2M37xVg4fz+LSTg14ZFA7zmpov7kWaN+vTvPVmo9Ac6DjUCeBnM4M8Nwc2LoIVn0Iv37lLKfRoLOTULrcWPQeH7k5zs6Ce9a4CcP981jK72XqtnRWyG3WB3rd47s1soxPWSLxYImkfDqemcOUpdsZ9/1WDh7L5KKz6vPooHZ0buLbdYEqJFXY8aPTgb55HgRVhx63Qd8/QL3okt37+CGnbyVuKiStgIAgaHupk1Si+zu1mLxkkbzG2Rck+7hzbUAw1O/gJI1GZ0PDLtCgk8/XdjJlwxKJB0sk5UtGVg7TftnJf7/bQsqRE/RvG8kjg9rRo7ktnHiK3BzYONtJILtXOgsK9rnP+S2/Rr3Sf719G52EsuYjp6nKU2gdJ2E07OL+nA2R7Zx+GFMpWSLxYImkfMjMzmVG7C7e+jaBPWkZ9Imux2OXtKd3tA++ECuDdZ86O+Yd2g71WkO/+6HrMAgugw23crKdIcR71jid9w27QHhzm+9RxdioLVNuZOXk8unKRN5YmEBS6nF6tqjLqzd25ZzWEdaJXpBj+50l1Dd87vzWf9MH0H6Ib/fAyC8wyBmia8N0jRcskRifyclVZsUl8frCzew4kM7ZTevwwjWdOb9dlCWQwmyYDV8+AifS4OK/OZ3ogfbf1JRv9i/UlDpV5et1e3hlQTwJ+47SoVFtJtwew8AO9S2BFCb9oLMPx7qZzjDcq7+ABh39HZUxXvFpIhGRwcDrQCDwjqq+mO98c5zdFsPdMqNVdY577ilgBJADPKiq87y5p/EfVWXx5v38e94m1iYdpnVUTf57aw8Gd2pIQIAlkEL9OsdZRj39IFz4NJz3sH/24TDmDPkskYhIIDAWGAQkAstFZLaqbvAo9jQwQ1XfFpGOwBygpfv4ZqAT0Bj4xt0fBS/uafwgdvtBXp63iV+2HaRJeHX+fUNXru7WmKCqvu5VUY4fcjZyWj0NGnSB4Z84ndrGVDC+rJH0BhJUdSuAiEwHhgKeX/oK5M06qwPsdh8PBaar6glgm4gkuPfDi3uaMrQu6TCvzN/Eok0pRNaqxpihnbipVzOqBZVhx3BhThyFI3uceRZl2VHtjc0LYPYDztay5z8J/R+3YbSmwvJlImkC7PJ4ngj0yVfmWWC+iDwA1MTZFz7v2qX5rm3iPi7unqYMbEk5yqsL4vlqTTJ1qgfz5OCzuKNfC2qElINut+TVEDsJ1n4MmUedyXsNOnrMf+jqPA/xw1a7GYdh3l9g1RRnWO2wadC4e9nHYUwp8uX/+oIaxfNPWhkGTFbVV0TkHGCKiHQu4tqC2kkKnAgjIiOBkQDNmzf3OmhTtKTU47z+TTwzVyQSGhzIAxe14Z7+rahT3c9t+pnHnHkXKyY5s7ODQqHTtdDiHGdJkT1rYP3nsGKye4E4a0A17PL7jOyGZ0Ot+r6Lccu3MOsBZ42r8x6FC0b7dsFEY8qILxNJIuC55VlTfm+6yjMCGAygqj+LSCgQWcy1xd0T937jgfHgTEg8s7dg8qQcOcHYRQlMXbYTgDv7RfOHC1sT6e+ta/ducJLH6unOkNmos2DwS9D1JmdBQk+qcDjRY42otc4Ofes//b1MrQa/J5WGXSCqPdSMgur1znwY7okjMP+vTpyR7WDEN9C055m/Z2PKGV8mkuVAWxGJBpJwOs9vyVdmJ3AxMFlEOgChQAowG5gqIq/idLa3BX7BqakUd09Tig6nZzH+hy1MXLKdzJxcbujZlAcubkuT8DKYXV2YrOOwYRbEToRdyyAwBDpeDTF3QfNzCp99LeJs5xreDM66/Pfjxw/BnnW/J5c9a2HrG5CbffL1oeHORkw1IpylSmpGeDzOO+7+1Ix0ms62LYZZf4TUXdDvAWdUli1gaCoZnyUSVc0WkfuBeThDdSeq6noRGQPEqups4DFggog8gtNEdac6a7asF5EZOJ3o2cAfVTUHoKB7+uo9VFWqyurEw8xdm8y0X3aSlpHNlV0b88jAtrSKquW/wFLind/q46ZCRqqzbMglf4eutzhf6meqel1nccLo/r8fyz4BKb86q9ymH3B+ju13H+93li1JinWe5084eYKqO4sb1msNd8+D5tadZyonW2vLAM5Wtit2HmLu2j18vS6Z3YczCAoQLjqrPg8NbEunxn5azTX7BGz8wuk837HEWW22wxXQ8y6IHuD/tZ9UnQ70k5LN/t+TTvW6zgq9ITX8G6cxZ8DW2jLFys7J5ZftB5m7dg/z1u9h35EThAQFMKBtJI9d0p6BHRpQp4afOtFVIfZdWPQP5ws5vIWzu1634UXvlVHWRKB6uPMT0drf0RjjF5ZIqpisnFx+2nKAr9clM2/9Xg4eyyQ0OIAL29dncOeGXHRWff/vgZ6T5SwXsmKSU+s492FodSEE2ORGY8ojSySVWXYmZBwmI6QuSxIOMHfdHr7ZuJfDx7OoGRLIRR0aMKRzQ85vH1U+5n+As0zIx3c4ndTnPuwsXGgJxJhyrZx8e5jSptuXkDFzFNWP7uKY1iE4twUdA6Pp2qQrrbucQ8/uMYRWK2czqVPiYdpNzhDdq8dBt2H+jsgY4wVLJJVMdsZRdswYTeutU9ib24DPA4dzbvhBesh2BhyegyTNcgZOL6zhbInqOdu7fgf/dQonLISP73IWK7zjC2je1z9xGGNOmyWSSuJ4Zg6LFsymS+xTtNZkPgu+nNyL/8b/xbT5fd2r7EzYv+n3uRLJa2DtJ858DAAJgIi2J8/2btzD6Uj2pWXj4evRzmTCYdOgbgvfvp4xplRZIqngDh3L5IMlv1Ln55cZrl+SElif5QPeY+j5Q09duj0o5PcaSB5VSN3pMRlvjTPJb91M53xwDWeobb8HoHaj0g0+JwvmPumMzmp3GVw3AaqFle5rGGN8zhJJBZV4KJ13ftjGxuXf8oL8lzYBu9l31q00uPYlGpzOl7GIUwOo28KZn5En/aCTWFZPh2XjYPk70H24s1dGeCmsXZZ+ED6+E7Z9D+c+5Haql7MVeo0xXrEJiRXMxuQ0/vf9Fuat2clDgTMZGfglObUaEnzNWGh9kW9e9OA2+PF1WPUBoHD2zdD/0TOfN7F/M0y9yakJXfk6dL+1VMM1xpQObyckWiKpAFSVpVsPMu77LXwfn0KvkB2MrTGB+hlbocftzjIhoWUw8/xwEvz0hrOCbk6ms7pu/8dOb0vYLd86NZGAILjpQ2d1XmNMuWSJxENFTSQ5ucqCDXt4+/utrN6VSsOawptNFhKzaxJSqz5c9Sa0HVT2gR3dBz+/Bcvfdfb7OOsKGPB48ftq/DLB6ROJag/DplunujHlnCUSDxUtkeTmKnPX7eG1b+JJ2HeUFhE1+FPXTC5LeI6AfeudRQoH/+PUZdLLWvpBp/9k2Thnvak2g2DAE6cuTpiT7YzKWj4B2g2G696xTnVjKgBLJB4qSiJRVRZu3McrC+LZmJxGm/q1ePjClgxJnUbAD/9ylie/4j9w1hB/h3qyjMNOZ/zPY511sVr2dxJK9ABnld6P74St30G/B531sqxT3ZgKwRKJh/KeSFSVHzbv55UF8azelUrLiBo8PLAdV9ZPIfDLB52tY7vcAJe9DDXq+TvcwmUec/pPfnwDju6Bpr3h+EE4tAOu/I8z6ssYU2HY6r8VxLKtB3hlfjy/bD9Ik/DqvHzd2VzbNI2gH56CWbOcTZNunAIdr/J3qMULqQnn/BFiRkDcB7DkP5CVDnfMhhb9/B2dMcZHLJH4ycqdh3h1fjxLEvZTP6wazw/txE2tThCy5Dn4aiaE1IIBf4Jz/uD/vpDTFRwKve6BHnc6o7tsLw5jKjVLJGVsXdJhXlsQz8Jf9xFRM4SnL+/Abe2Vaj+9DPOnQVCoM0Hv3IfKdzOWNwKDznyfc2NMhWH/y8tI/N4jvLYgnrnr9lCnejBPXNqeuzoFUWPpqzDuQ2deRZ//c2aO16rv73CNMcZrlkh8bNv+Y/znm3hmr95NzZAgHrq4Lfd0CyXslzfgf+85a131vMuZ2Ffaa1kZY0wZsETiQ+/8sJV/zv2VkMAA7hvQmlE9wwhfORb+9y7kZkO3W51hsuHN/B2qMcacMUskPpKanskr8+Pp1zqC165sTuTqcTBhPGRnQNdhTgKpF+3vMI0xpsR8mkhEZDDwOhAIvKOqL+Y7/xpwofu0BlBfVcNF5ELgNY+iZwE3q+rnIjIZOB847J67U1XjfPg2zsgHS3cQnJXGq1E/EfnORGcpkS7Xw/mjIbKNv8MzxphS47NEIiKBwFhgEJAILBeR2aq6Ia+Mqj7iUf4BoLt7fBHQzT1eD0gA5nvc/glVnemr2EsqIyOD9CVv82ONjwlbmQYdh8IFTzk7EBpjTCXjyxpJbyBBVbcCiMh0YCiwoZDyw4C/FXD8emCuqqb7JMrSpAqbF3Bi1p/4U+42DjfoC1e/DI26+jsyY4zxmQAf3rsJsMvjeaJ77BQi0gKIBr4t4PTNwLR8x14QkTUi8pqIVCvkniNFJFZEYlNSUk4/+tO1dwNMuQam3kBa+gnGhP2V2vfNtSRijKn0fJlIpIBjhS3sdTMwU1VzTrqBSCOgCzDP4/BTOH0mvYB6wJMF3VBVx6tqjKrGREVFnW7s3ju6D754GMadC7tXsbHbn7no+It0G3gLEuDLj9cYY8oHX37TJQKe41qbArsLKVtQrQPgRuAzVc3KO6Cqyeo4AUzCaUIre1kZ8MOr8EYPWDUFeo+EB1fxdHJ/GtQNY0jnhn4Jyxhjypov+0iWA21FJBpIwkkWt+QvJCLtgbrAzwXcYxhODcSzfCNVTRYRAa4G1pV24EVShfWfwTd/c7aKbXcZXPI8RLYldvtBVuw4xLNXdiQo0GojxpiqwWeJRFWzReR+nGapQGCiqq4XkTFArKrOdosOA6ZrvvXsRaQlTo3m+3y3/lBEonCazuKAUb56D6dIXAHznoJdy6BBZ7h9FrS64LfT/1u8lfAawdzYyyYYGmOqDp/OI1HVOcCcfMeeyff82UKu3U4BnfOqelHpReil1F2wcAysnQE168OVbzh7a3hs0LQl5SjfbNzL/Re2oUaIzfM0xlQd9o1XlBNH4cf/wE9vOk1a/R+D8x4pcJvYd37YSnBgAHf0a1n2cRpjjB9ZIinKh9fDzp+h83XOFrHhzQsstu9IBp+sSOL6mKZE1ipwNLIxxlRalkiKcuGfnf1BmhU9MOy9n7aTlZvLvf1blVFgxhhTflgiKUr0gGKLHDuRzQdLd3JJxwZER9Ysg6CMMaZ8sTGqJfTR8l0cPp7Ffee39ncoxhjjF5ZISiArJ5d3l2yjV8u69GhewfZVN8aYUmKJpATmrE0mKfU4IwdYbcQYU3VZIjlDqsr/vt9K66iaXHyW7bFujKm6LJGcoR8TDrAhOY2RA1oREFDQ+pTGGFM1WCI5Q/9bvIWosGpc3b3AlfGNMabKsERyBtbvPswPm/dzZ7+WVAsKLP4CY4ypxCyRnIEJi7dSIySQ4X1a+DsUY4zxO0skpykp9ThfrElmWO/m1KkR7O9wjDHG7yyRnKaJS7YBcPd50X6OxBhjygdLJKfhcHoW037ZyZVnN6JJeHV/h2OMMeWCJZLT8MGyHaRn5tgERGOM8WCJxEsnsnOY/NN2+reNpGPj2v4Oxxhjyg1LJF76fFUSKUdOcJ/VRowx5iSWSLyQm6uMX7yVjo1qc26bCH+HY4wx5YolEi8s/HUfW1KOcd/5rRCx5VCMMcaTTxOJiAwWkU0ikiAiows4/5qIxLk/8SKS6nEux+PcbI/j0SKyTEQ2i8hHIhLiy/cAMH7xFpqEV2dIl0a+filjjKlwfJZIRCQQGAtcBnQEholIR88yqvqIqnZT1W7Am8CnHqeP551T1as8jr8EvKaqbYFDwAhfvQeAlTsPsXz7IUacF01woFXgjDEmP19+M/YGElR1q6pmAtOBoUWUHwZMK+qG4rQrXQTMdA+9B1xdCrEWavz3W6lTPZibejXz5csYY0yF5ctE0gTY5fE80T12ChFpAUQD33ocDhWRWBFZKiJ5ySICSFXVbC/uOdK9PjYlJeWM3sDWlKPM27CH2/q2oGY1297eGGMK4stvx4J6pbWQsjcDM1U1x+NYc1XdLSKtgG9FZC2Q5u09VXU8MB4gJiamsNct0jtLthEcGMAd/VqeyeXGGFMl+LJGkgh4tgc1BXYXUvZm8jVrqepu98+twHdAd2A/EC4ieQmwqHuWWPN6NRhxXjRRYdV89RLGGFPh+bJGshxoKyLRQBJOsrglfyERaQ/UBX72OFYXSFfVEyISCZwLvKyqKiKLgOtx+lzuAGb56g2MOt8mHxpjTHF8ViNx+zHuB+YBG4EZqrpeRMaIiOcorGHAdFX1bH7qAMSKyGpgEfCiqm5wzz0JPCoiCTh9Ju/66j0YY4wpnpz8/V05xcTEaGxsrL/DMMaYCkVEVqhqTHHlbGKEMcaYErFEYowxpkQskRhjjCkRSyTGGGNKxBKJMcaYErFEYowxpkSqxPBfEUkBdpzh5ZE4M+rLK4uvZCy+krH4Sqa8x9dCVaOKK1QlEklJiEisN+Oo/cXiKxmLr2QsvpIp7/F5y5q2jDHGlIglEmOMMSViiaR44/0dQDEsvpKx+ErG4iuZ8h6fV6yPxBhjTIlYjcQYY0yJWCIxxhhTIpZIXCIyWEQ2iUiCiIwu4Hw1EfnIPb9MRFqWYWzNRGSRiGwUkfUi8lABZS4QkcMiEuf+PFNW8bmvv11E1rqvfcqa/eJ4w/381ohIjzKMrb3H5xInImki8nC+MmX6+YnIRBHZJyLrPI7VE5EFIrLZ/bNuIdfe4ZbZLCJ3lGF8/xKRX92/v89EJLyQa4v8t+DD+J4VkSSPv8MhhVxb5P91H8b3kUds20UkrpBrff75lTpVrfI/QCCwBWgFhACrgY75yvwBGOc+vhn4qAzjawT0cB+HAfEFxHcB8KUfP8PtQGQR54cAcwEB+gLL/Ph3vQdnopXfPj9gANADWOdx7GVgtPt4NPBSAdfVA7a6f9Z1H9cto/guAYLcxy8VFJ83/xZ8GN+zwONe/P0X+X/dV/HlO/8K8Iy/Pr/S/rEaiaM3kKCqW1U1E2cb36H5ygwF3nMfzwQuFhEpi+BUNVlVV7qPj+DsONmkLF67FA0F3lfHUiBcRBr5IY6LgS2qeqYrHZQKVV0MHMx32PPf2HvA1QVceimwQFUPquohYAEwuCziU9X56ux8CrAUaFrar+utQj4/b3jzf73EiorP/d64EZhW2q/rL5ZIHE2AXR7PEzn1i/q3Mu5/psM4W/2WKbdJrTuwrIDT54jIahGZKyKdyjQwUGC+iKwQkZEFnPfmMy4LN1P4f2B/fn4ADVQ1GZxfHoD6BZQpL5/j3Tg1zIIU92/Bl+53m94mFtI0WB4+v/7AXlXdXMh5f35+Z8QSiaOgmkX+cdHelPEpEakFfAI8rKpp+U6vxGmu6Qq8CXxelrEB56pqD+Ay4I8iMiDf+fLw+YUAVwEfF3Da35+ft8rD5/gXIBv4sJAixf1b8JW3gdZANyAZp/koP79/fsAwiq6N+OvzO2OWSByJQDOP502B3YWVEZEgoA5nVrU+IyISjJNEPlTVT/OfV9U0VT3qPp4DBItIZFnFp6q73T/3AZ/hNCF48uYz9rXLgJWqujf/CX9/fq69ec197p/7Cijj18/R7dy/ArhV3Qb9/Lz4t+ATqrpXVXNUNReYUMjr+vvzCwKuBT4qrIy/Pr+SsETiWA60FZFo97fWm4HZ+crMBvJGyFzP/7d3Py82hXEcx9/fEONnfpQfCYmVkjRNmqykSRZKFpMUMZtRYiULO9nYSDIbv1L+AhZCXaXkx6wYREyyUJRRSCTpa/F8T52ue5mce86Z9HnVac597jP3fu+5z+l7zvOc+xy41W5H6rToU70APHP3k23qLMrGbMysh/TdfqgovhlmNitbJw3KPmmqdhXYHVdvbQA+Zd04FWp7JFjn9svJt7E9wJUWdW4AfWY2N7pu+qKsdGa2BTgCbHP3r23qjKctlBVffsxte5v3Hc++XqbNwHN3f9PqyTq3XyF1j/ZPlIV0VdEL0hUdR6PsGGmnAZhG6hIZBYaBlRXGtpF0+j0CPIxlKzAIDEadA8BT0lUo94HeCuNbGe/7KGLItl8+PgOGYvs+Bror/n6nkxLDnFxZbduPlNDeAj9IR8kDpDG3BvAy/s6LuvVyPggAAAHcSURBVN3A+dz/7ot2OArsrTC+UdL4QtYGs6sYlwDX/tQWKorvcrStEVJyWNwcXzz+bV+vIr4ov5S1uVzdyrdfpxdNkSIiIoWoa0tERApRIhERkUKUSEREpBAlEhERKUSJREREClEiEekAM/vZNMNwx2aVNbMV+VlkRSaayXUHIPKf+Obu6+oOQqQOOiMRKVHcW+KEmQ3HsirKl5tZIyYYbJjZsihfGPf6eBRLb7zUJDM7Z+l+NDfNrKu2DyXSRIlEpDO6mrq2+nPPfXb3HuAMcCrKzpCm1V9LmvzwdJSfBm57mjxyPenXzQCrgSF3XwN8BHaU/HlExk2/bBfpADP74u4zW5S/Bja5+6uYePOdu883szHSFB4/ovytuy8ws/fAUnf/nnuNFaR7kKyOx0eAKe5+vPxPJvJ3OiMRKZ+3WW9Xp5XvufWfaHxTJhAlEpHy9ef+3ov1u6SZZwF2AXdivQHsBzCzSWY2u6ogRf6VjmpEOqPLzB7mHl939+wS4Klm9oB04LYzyg4CF83sMPAe2Bvlh4CzZjZAOvPYT5pFVmTC0hiJSIlijKTb3cfqjkWkLOraEhGRQnRGIiIiheiMREREClEiERGRQpRIRESkECUSEREpRIlEREQK+QVDYkWbtnFUWQAAAABJRU5ErkJggg==\n"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"# Plot training & validation loss values\nplt.plot(history.history['loss'])\nplt.plot(history.history['val_loss'])\nplt.title('Model loss')\nplt.ylabel('Loss')\nplt.xlabel('Epoch')\nplt.legend(['Train', 'Test'], loc='upper left')\nplt.show()","execution_count":50,"outputs":[{"output_type":"display_data","data":{"text/plain":"<Figure size 432x288 with 1 Axes>","image/png":"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\n"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"test_loss, test_acc = model.evaluate(X_test, y_test)\nprint('Test Accuracy:', test_acc)\nprint('Test Loss:', test_loss)","execution_count":51,"outputs":[{"output_type":"stream","text":"3023/3023 [==============================] - 0s 135us/step\nTest Accuracy: 0.8144227588488256\nTest Loss: 0.47433689975537413\n","name":"stdout"}]}],"metadata":{"kernelspec":{"display_name":"Python 3","language":"python","name":"python3"},"language_info":{"name":"python","version":"3.6.6","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"}},"nbformat":4,"nbformat_minor":1}