{"cells":[{"metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true},"cell_type":"code","source":"import numpy as np \nimport pandas as pd \nimport cv2\nfrom matplotlib import pyplot as plt\nimport os\nfrom subprocess import check_output\nimport cv2\nfrom PIL import Image\nimport glob","execution_count":1,"outputs":[]},{"metadata":{"_cell_guid":"79c7e3d0-c299-4dcb-8224-4455121ee9b0","_uuid":"d629ff2d2480ee46fbb7e2d37f6b5fab8052498a","trusted":true},"cell_type":"code","source":"trainLabels = pd.read_csv(\"../input/labels/trainLabels.csv\")\nprint(trainLabels.head())\n\nfilelist = glob.glob('../input/bloodvessel/bloodvesselextraction/BloodVesselExtraction/*.jpeg')\n##filelist = glob.glob('../input/diabetic-retinopathy-detection/*.jpeg')\nnp.size(filelist)","execution_count":3,"outputs":[{"output_type":"stream","text":" image level\n0 10_left 0\n1 10_right 0\n2 13_left 0\n3 13_right 0\n4 15_left 1\n","name":"stdout"},{"output_type":"execute_result","execution_count":3,"data":{"text/plain":"500"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"img_data = []\nimg_label = []\nimg_r = 512\nimg_c = 512\nfor file in filelist:\n tmp = cv2.imread(file)\n tmp = cv2.resize(tmp,(img_r, img_c), interpolation = cv2.INTER_CUBIC)\n tmp = cv2.cvtColor(tmp, cv2.COLOR_BGR2GRAY)\n img_data.append(np.array(tmp).flatten())\n tmpfn = file\n tmpfn = tmpfn.replace(\"../input/bloodvessel/bloodvesselextraction/BloodVesselExtraction/\",\"\")\n ##tmpfn = tmpfn.replace(\"../input/diabetic-retinopathy-detection/\",\"\")\n tmpfn = tmpfn.replace(\".jpeg\",\"\")\n img_label.append(trainLabels.loc[trainLabels.image==tmpfn, 'level'].values[0])","execution_count":4,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"data = pd.DataFrame({'img_data':img_data,'label':img_label})\ndata.sample(3)","execution_count":5,"outputs":[{"output_type":"execute_result","execution_count":5,"data":{"text/plain":" img_data label\n489 [255, 255, 255, 255, 255, 255, 255, 255, 255, ... 0\n267 [255, 255, 255, 255, 255, 255, 255, 255, 255, ... 0\n10 [255, 255, 255, 255, 255, 255, 255, 255, 255, ... 0","text/html":"
\n\n\n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n \n
img_datalabel
489[255, 255, 255, 255, 255, 255, 255, 255, 255, ...0
267[255, 255, 255, 255, 255, 255, 255, 255, 255, ...0
10[255, 255, 255, 255, 255, 255, 255, 255, 255, ...0
\n
"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"data[['label']].hist(figsize = (10, 5))","execution_count":6,"outputs":[{"output_type":"execute_result","execution_count":6,"data":{"text/plain":"array([[]],\n dtype=object)"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"
","image/png":"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\n"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"from sklearn.model_selection import train_test_split\nX = data['img_data']\ny = data['label']\nX_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)\n\n\nfrom sklearn.utils import shuffle\n\ndata,label = shuffle(X_train,y_train, random_state=2)\ntrain_data = pd.DataFrame({'data': data, 'label':label})\ntrain_df = train_data.groupby(['label']).apply(lambda x: x.sample(160, replace = True)\n ).reset_index(drop = True)\nprint('New Data Size:', train_df.shape[0], 'Old Size:', train_data.shape[0])\ntrain_df[['label']].hist(figsize = (10, 5))","execution_count":7,"outputs":[{"output_type":"stream","text":"New Data Size: 800 Old Size: 400\n","name":"stdout"},{"output_type":"execute_result","execution_count":7,"data":{"text/plain":"array([[]],\n dtype=object)"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"
","image/png":"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\n"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"X_train = train_df['data']\ny_train = train_df['label']\n\nX_train = np.asarray(X_train)\ny_train = np.asarray(y_train)\nX_test = np.asarray(X_test)\ny_test = np.asarray(y_test)","execution_count":8,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"X_train_resh = np.zeros([X_train.shape[0],img_r, img_c, 1])\nfor i in range (X_train.shape[0]-1):\n X_train_resh[i] = np.reshape(X_train[i], (img_r, img_c, 1))\n \nX_test_resh = np.zeros([X_test.shape[0],img_r, img_c, 1])\nfor i in range (X_test.shape[0]-1):\n X_test_resh[i] = np.reshape(X_test[i], (img_r, img_c, 1))\nprint(X_test_resh.shape)","execution_count":9,"outputs":[{"output_type":"stream","text":"(100, 512, 512, 1)\n","name":"stdout"}]},{"metadata":{"trusted":true},"cell_type":"code","source":"from keras.utils import np_utils\nnb_classes = 5\nY_train = np_utils.to_categorical(y_train, nb_classes)\nY_test = np_utils.to_categorical(y_test, nb_classes)","execution_count":10,"outputs":[{"output_type":"stream","text":"Using TensorFlow backend.\n","name":"stderr"}]},{"metadata":{"trusted":true},"cell_type":"code","source":"import matplotlib.pyplot as plt\nimport matplotlib\n\nimg=X_train_resh[100].reshape(img_r,img_c)\nplt.imshow(img)\nplt.imshow(img,cmap='gray')","execution_count":11,"outputs":[{"output_type":"execute_result","execution_count":11,"data":{"text/plain":""},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"
","image/png":"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\n"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"import keras\nfrom keras.layers.core import Layer\nimport keras.backend as K\nimport tensorflow as tf\n\nfrom keras.models import Model\nfrom keras.layers import Conv2D, MaxPool2D, \\\n Dropout, Dense, Input, concatenate, \\\n GlobalAveragePooling2D, AveragePooling2D,\\\n Flatten\n\nimport cv2 \nimport numpy as np \nfrom keras.datasets import cifar10 \nfrom keras import backend as K \nfrom keras.utils import np_utils\n\nimport math \nfrom keras.optimizers import SGD \nfrom keras.callbacks import LearningRateScheduler\n\ndef inception_module(x,\n filters_1x1,\n filters_3x3_reduce,\n filters_3x3,\n filters_5x5_reduce,\n filters_5x5,\n filters_pool_proj,\n name=None):\n \n conv_1x1 = Conv2D(filters_1x1, (1, 1), padding='same', activation='relu', kernel_initializer=kernel_init, bias_initializer=bias_init)(x)\n \n conv_3x3 = Conv2D(filters_3x3_reduce, (1, 1), padding='same', activation='relu', kernel_initializer=kernel_init, bias_initializer=bias_init)(x)\n conv_3x3 = Conv2D(filters_3x3, (3, 3), padding='same', activation='relu', kernel_initializer=kernel_init, bias_initializer=bias_init)(conv_3x3)\n\n conv_5x5 = Conv2D(filters_5x5_reduce, (1, 1), padding='same', activation='relu', kernel_initializer=kernel_init, bias_initializer=bias_init)(x)\n conv_5x5 = Conv2D(filters_5x5, (5, 5), padding='same', activation='relu', kernel_initializer=kernel_init, bias_initializer=bias_init)(conv_5x5)\n\n pool_proj = MaxPool2D((3, 3), strides=(1, 1), padding='same')(x)\n pool_proj = Conv2D(filters_pool_proj, (1, 1), padding='same', activation='relu', kernel_initializer=kernel_init, bias_initializer=bias_init)(pool_proj)\n\n output = concatenate([conv_1x1, conv_3x3, conv_5x5, pool_proj], axis=3, name=name)\n \n return output\n\nkernel_init = keras.initializers.glorot_uniform()\nbias_init = keras.initializers.Constant(value=0.2)\n\ninput_layer = Input(shape=(512, 512, 1))\n\nx = Conv2D(64, (7, 7), padding='same', strides=(2, 2), activation='relu', name='conv_1_7x7/2', kernel_initializer=kernel_init, bias_initializer=bias_init)(input_layer)\nx = MaxPool2D((3, 3), padding='same', strides=(2, 2), name='max_pool_1_3x3/2')(x)\nx = Conv2D(64, (1, 1), padding='same', strides=(1, 1), activation='relu', name='conv_2a_3x3/1')(x)\nx = Conv2D(192, (3, 3), padding='same', strides=(1, 1), activation='relu', name='conv_2b_3x3/1')(x)\nx = MaxPool2D((3, 3), padding='same', strides=(2, 2), name='max_pool_2_3x3/2')(x)\n\nx = inception_module(x,\n filters_1x1=64,\n filters_3x3_reduce=96,\n filters_3x3=128,\n filters_5x5_reduce=16,\n filters_5x5=32,\n filters_pool_proj=32,\n name='inception_3a')\n\nx = inception_module(x,\n filters_1x1=128,\n filters_3x3_reduce=128,\n filters_3x3=192,\n filters_5x5_reduce=32,\n filters_5x5=96,\n filters_pool_proj=64,\n name='inception_3b')\n\nx = MaxPool2D((3, 3), padding='same', strides=(2, 2), name='max_pool_3_3x3/2')(x)\n\nx = inception_module(x,\n filters_1x1=192,\n filters_3x3_reduce=96,\n filters_3x3=208,\n filters_5x5_reduce=16,\n filters_5x5=48,\n filters_pool_proj=64,\n name='inception_4a')\n\n\nx1 = AveragePooling2D((5, 5), strides=3)(x)\nx1 = Conv2D(128, (1, 1), padding='same', activation='relu')(x1)\nx1 = Flatten()(x1)\nx1 = Dense(6400, activation='relu')(x1)\nx1 = Dense(3200, activation='relu')(x1)\nx1 = Dense(1600, activation='relu')(x1)\nx1 = Dense(1024, activation='relu')(x1)\nx1 = Dense(800, activation='relu')(x1)\nx1 = Dense(400, activation='relu')(x1)\nx1 = Dense(200, activation='relu')(x1)\nx1 = Dropout(0.7)(x1)\nx1 = Dense(5, activation='softmax', name='auxilliary_output_1')(x1)\n\nx = inception_module(x,\n filters_1x1=160,\n filters_3x3_reduce=112,\n filters_3x3=224,\n filters_5x5_reduce=24,\n filters_5x5=64,\n filters_pool_proj=64,\n name='inception_4b')\n\nx = inception_module(x,\n filters_1x1=128,\n filters_3x3_reduce=128,\n filters_3x3=256,\n filters_5x5_reduce=24,\n filters_5x5=64,\n filters_pool_proj=64,\n name='inception_4c')\n\nx = inception_module(x,\n filters_1x1=112,\n filters_3x3_reduce=144,\n filters_3x3=288,\n filters_5x5_reduce=32,\n filters_5x5=64,\n filters_pool_proj=64,\n name='inception_4d')\n\n\nx2 = AveragePooling2D((5, 5), strides=3)(x)\nx2 = Conv2D(128, (1, 1), padding='same', activation='relu')(x2)\nx2 = Flatten()(x2)\nx2 = Dense(6400, activation='relu')(x2)\nx2 = Dense(3200, activation='relu')(x2)\nx2 = Dense(1600, activation='relu')(x2)\nx2 = Dense(1024, activation='relu')(x2)\nx2 = Dense(800, activation='relu')(x2)\nx2 = Dense(400, activation='relu')(x2)\nx2 = Dense(200, activation='relu')(x2)\nx2 = Dropout(0.7)(x2)\nx2 = Dense(5, activation='softmax', name='auxilliary_output_2')(x2)\n\nx = inception_module(x,\n filters_1x1=256,\n filters_3x3_reduce=160,\n filters_3x3=320,\n filters_5x5_reduce=32,\n filters_5x5=128,\n filters_pool_proj=128,\n name='inception_4e')\n\nx = MaxPool2D((3, 3), padding='same', strides=(2, 2), name='max_pool_4_3x3/2')(x)\n\nx = inception_module(x,\n filters_1x1=256,\n filters_3x3_reduce=160,\n filters_3x3=320,\n filters_5x5_reduce=32,\n filters_5x5=128,\n filters_pool_proj=128,\n name='inception_5a')\n\nx = inception_module(x,\n filters_1x1=384,\n filters_3x3_reduce=192,\n filters_3x3=384,\n filters_5x5_reduce=48,\n filters_5x5=128,\n filters_pool_proj=128,\n name='inception_5b')\n\nx = GlobalAveragePooling2D(name='avg_pool_5_3x3/1')(x)\nx = Dense(6400, activation='relu')(x)\nx = Dense(3200, activation='relu')(x)\nx = Dense(1600, activation='relu')(x)\nx = Dense(1024, activation='relu')(x)\nx = Dense(800, activation='relu')(x)\nx = Dense(400, activation='relu')(x)\nx = Dense(200, activation='relu')(x)\nx = Dropout(0.4)(x)\n\nx = Dense(5, activation='softmax', name='output')(x)\n\nmodel = Model(input_layer, [x, x1, x2], name='inception_v1')\nprint(model.summary())","execution_count":12,"outputs":[{"output_type":"stream","text":"WARNING:tensorflow:From /opt/conda/lib/python3.6/site-packages/tensorflow/python/framework/op_def_library.py:263: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version.\nInstructions for updating:\nColocations handled automatically by placer.\nWARNING:tensorflow:From /opt/conda/lib/python3.6/site-packages/keras/backend/tensorflow_backend.py:3445: calling dropout (from tensorflow.python.ops.nn_ops) with keep_prob is deprecated and will be removed in a future version.\nInstructions for updating:\nPlease use `rate` instead of `keep_prob`. Rate should be set to `rate = 1 - keep_prob`.\n__________________________________________________________________________________________________\nLayer (type) Output Shape Param # Connected to \n==================================================================================================\ninput_1 (InputLayer) (None, 512, 512, 1) 0 \n__________________________________________________________________________________________________\nconv_1_7x7/2 (Conv2D) (None, 256, 256, 64) 3200 input_1[0][0] \n__________________________________________________________________________________________________\nmax_pool_1_3x3/2 (MaxPooling2D) (None, 128, 128, 64) 0 conv_1_7x7/2[0][0] \n__________________________________________________________________________________________________\nconv_2a_3x3/1 (Conv2D) (None, 128, 128, 64) 4160 max_pool_1_3x3/2[0][0] \n__________________________________________________________________________________________________\nconv_2b_3x3/1 (Conv2D) (None, 128, 128, 192 110784 conv_2a_3x3/1[0][0] \n__________________________________________________________________________________________________\nmax_pool_2_3x3/2 (MaxPooling2D) (None, 64, 64, 192) 0 conv_2b_3x3/1[0][0] \n__________________________________________________________________________________________________\nconv2d_2 (Conv2D) (None, 64, 64, 96) 18528 max_pool_2_3x3/2[0][0] \n__________________________________________________________________________________________________\nconv2d_4 (Conv2D) (None, 64, 64, 16) 3088 max_pool_2_3x3/2[0][0] \n__________________________________________________________________________________________________\nmax_pooling2d_1 (MaxPooling2D) (None, 64, 64, 192) 0 max_pool_2_3x3/2[0][0] \n__________________________________________________________________________________________________\nconv2d_1 (Conv2D) (None, 64, 64, 64) 12352 max_pool_2_3x3/2[0][0] \n__________________________________________________________________________________________________\nconv2d_3 (Conv2D) (None, 64, 64, 128) 110720 conv2d_2[0][0] \n__________________________________________________________________________________________________\nconv2d_5 (Conv2D) (None, 64, 64, 32) 12832 conv2d_4[0][0] \n__________________________________________________________________________________________________\nconv2d_6 (Conv2D) (None, 64, 64, 32) 6176 max_pooling2d_1[0][0] \n__________________________________________________________________________________________________\ninception_3a (Concatenate) (None, 64, 64, 256) 0 conv2d_1[0][0] \n conv2d_3[0][0] \n conv2d_5[0][0] \n conv2d_6[0][0] \n__________________________________________________________________________________________________\nconv2d_8 (Conv2D) (None, 64, 64, 128) 32896 inception_3a[0][0] \n__________________________________________________________________________________________________\nconv2d_10 (Conv2D) (None, 64, 64, 32) 8224 inception_3a[0][0] \n__________________________________________________________________________________________________\nmax_pooling2d_2 (MaxPooling2D) (None, 64, 64, 256) 0 inception_3a[0][0] \n__________________________________________________________________________________________________\nconv2d_7 (Conv2D) (None, 64, 64, 128) 32896 inception_3a[0][0] \n__________________________________________________________________________________________________\nconv2d_9 (Conv2D) (None, 64, 64, 192) 221376 conv2d_8[0][0] \n__________________________________________________________________________________________________\nconv2d_11 (Conv2D) (None, 64, 64, 96) 76896 conv2d_10[0][0] \n__________________________________________________________________________________________________\nconv2d_12 (Conv2D) (None, 64, 64, 64) 16448 max_pooling2d_2[0][0] \n__________________________________________________________________________________________________\ninception_3b (Concatenate) (None, 64, 64, 480) 0 conv2d_7[0][0] \n conv2d_9[0][0] \n conv2d_11[0][0] \n conv2d_12[0][0] \n__________________________________________________________________________________________________\nmax_pool_3_3x3/2 (MaxPooling2D) (None, 32, 32, 480) 0 inception_3b[0][0] \n__________________________________________________________________________________________________\nconv2d_14 (Conv2D) (None, 32, 32, 96) 46176 max_pool_3_3x3/2[0][0] \n__________________________________________________________________________________________________\nconv2d_16 (Conv2D) (None, 32, 32, 16) 7696 max_pool_3_3x3/2[0][0] \n__________________________________________________________________________________________________\nmax_pooling2d_3 (MaxPooling2D) (None, 32, 32, 480) 0 max_pool_3_3x3/2[0][0] \n__________________________________________________________________________________________________\nconv2d_13 (Conv2D) (None, 32, 32, 192) 92352 max_pool_3_3x3/2[0][0] \n__________________________________________________________________________________________________\nconv2d_15 (Conv2D) (None, 32, 32, 208) 179920 conv2d_14[0][0] \n__________________________________________________________________________________________________\nconv2d_17 (Conv2D) (None, 32, 32, 48) 19248 conv2d_16[0][0] \n__________________________________________________________________________________________________\nconv2d_18 (Conv2D) (None, 32, 32, 64) 30784 max_pooling2d_3[0][0] \n__________________________________________________________________________________________________\ninception_4a (Concatenate) (None, 32, 32, 512) 0 conv2d_13[0][0] \n conv2d_15[0][0] \n conv2d_17[0][0] \n conv2d_18[0][0] \n__________________________________________________________________________________________________\nconv2d_21 (Conv2D) (None, 32, 32, 112) 57456 inception_4a[0][0] \n__________________________________________________________________________________________________\nconv2d_23 (Conv2D) (None, 32, 32, 24) 12312 inception_4a[0][0] \n__________________________________________________________________________________________________\nmax_pooling2d_4 (MaxPooling2D) (None, 32, 32, 512) 0 inception_4a[0][0] \n__________________________________________________________________________________________________\nconv2d_20 (Conv2D) (None, 32, 32, 160) 82080 inception_4a[0][0] \n__________________________________________________________________________________________________\nconv2d_22 (Conv2D) (None, 32, 32, 224) 226016 conv2d_21[0][0] \n__________________________________________________________________________________________________\nconv2d_24 (Conv2D) (None, 32, 32, 64) 38464 conv2d_23[0][0] \n__________________________________________________________________________________________________\nconv2d_25 (Conv2D) (None, 32, 32, 64) 32832 max_pooling2d_4[0][0] \n__________________________________________________________________________________________________\ninception_4b (Concatenate) (None, 32, 32, 512) 0 conv2d_20[0][0] \n conv2d_22[0][0] \n conv2d_24[0][0] \n conv2d_25[0][0] \n__________________________________________________________________________________________________\nconv2d_27 (Conv2D) (None, 32, 32, 128) 65664 inception_4b[0][0] \n__________________________________________________________________________________________________\nconv2d_29 (Conv2D) (None, 32, 32, 24) 12312 inception_4b[0][0] \n__________________________________________________________________________________________________\nmax_pooling2d_5 (MaxPooling2D) (None, 32, 32, 512) 0 inception_4b[0][0] \n__________________________________________________________________________________________________\nconv2d_26 (Conv2D) (None, 32, 32, 128) 65664 inception_4b[0][0] \n__________________________________________________________________________________________________\nconv2d_28 (Conv2D) (None, 32, 32, 256) 295168 conv2d_27[0][0] \n__________________________________________________________________________________________________\nconv2d_30 (Conv2D) (None, 32, 32, 64) 38464 conv2d_29[0][0] \n__________________________________________________________________________________________________\nconv2d_31 (Conv2D) (None, 32, 32, 64) 32832 max_pooling2d_5[0][0] \n__________________________________________________________________________________________________\ninception_4c (Concatenate) (None, 32, 32, 512) 0 conv2d_26[0][0] \n conv2d_28[0][0] \n conv2d_30[0][0] \n conv2d_31[0][0] \n__________________________________________________________________________________________________\nconv2d_33 (Conv2D) (None, 32, 32, 144) 73872 inception_4c[0][0] \n__________________________________________________________________________________________________\nconv2d_35 (Conv2D) (None, 32, 32, 32) 16416 inception_4c[0][0] \n__________________________________________________________________________________________________\nmax_pooling2d_6 (MaxPooling2D) (None, 32, 32, 512) 0 inception_4c[0][0] \n__________________________________________________________________________________________________\nconv2d_32 (Conv2D) (None, 32, 32, 112) 57456 inception_4c[0][0] \n__________________________________________________________________________________________________\nconv2d_34 (Conv2D) (None, 32, 32, 288) 373536 conv2d_33[0][0] \n__________________________________________________________________________________________________\nconv2d_36 (Conv2D) (None, 32, 32, 64) 51264 conv2d_35[0][0] \n__________________________________________________________________________________________________\nconv2d_37 (Conv2D) (None, 32, 32, 64) 32832 max_pooling2d_6[0][0] \n__________________________________________________________________________________________________\ninception_4d (Concatenate) (None, 32, 32, 528) 0 conv2d_32[0][0] \n conv2d_34[0][0] \n conv2d_36[0][0] \n conv2d_37[0][0] \n__________________________________________________________________________________________________\nconv2d_40 (Conv2D) (None, 32, 32, 160) 84640 inception_4d[0][0] \n__________________________________________________________________________________________________\nconv2d_42 (Conv2D) (None, 32, 32, 32) 16928 inception_4d[0][0] \n__________________________________________________________________________________________________\nmax_pooling2d_7 (MaxPooling2D) (None, 32, 32, 528) 0 inception_4d[0][0] \n__________________________________________________________________________________________________\nconv2d_39 (Conv2D) (None, 32, 32, 256) 135424 inception_4d[0][0] \n__________________________________________________________________________________________________\nconv2d_41 (Conv2D) (None, 32, 32, 320) 461120 conv2d_40[0][0] \n__________________________________________________________________________________________________\nconv2d_43 (Conv2D) (None, 32, 32, 128) 102528 conv2d_42[0][0] \n__________________________________________________________________________________________________\nconv2d_44 (Conv2D) (None, 32, 32, 128) 67712 max_pooling2d_7[0][0] \n__________________________________________________________________________________________________\ninception_4e (Concatenate) (None, 32, 32, 832) 0 conv2d_39[0][0] \n conv2d_41[0][0] \n conv2d_43[0][0] \n conv2d_44[0][0] \n__________________________________________________________________________________________________\nmax_pool_4_3x3/2 (MaxPooling2D) (None, 16, 16, 832) 0 inception_4e[0][0] \n__________________________________________________________________________________________________\nconv2d_46 (Conv2D) (None, 16, 16, 160) 133280 max_pool_4_3x3/2[0][0] \n__________________________________________________________________________________________________\nconv2d_48 (Conv2D) (None, 16, 16, 32) 26656 max_pool_4_3x3/2[0][0] \n__________________________________________________________________________________________________\nmax_pooling2d_8 (MaxPooling2D) (None, 16, 16, 832) 0 max_pool_4_3x3/2[0][0] \n__________________________________________________________________________________________________\nconv2d_45 (Conv2D) (None, 16, 16, 256) 213248 max_pool_4_3x3/2[0][0] \n__________________________________________________________________________________________________\nconv2d_47 (Conv2D) (None, 16, 16, 320) 461120 conv2d_46[0][0] \n__________________________________________________________________________________________________\nconv2d_49 (Conv2D) (None, 16, 16, 128) 102528 conv2d_48[0][0] \n__________________________________________________________________________________________________\nconv2d_50 (Conv2D) (None, 16, 16, 128) 106624 max_pooling2d_8[0][0] \n__________________________________________________________________________________________________\ninception_5a (Concatenate) (None, 16, 16, 832) 0 conv2d_45[0][0] \n conv2d_47[0][0] \n conv2d_49[0][0] \n conv2d_50[0][0] \n__________________________________________________________________________________________________\nconv2d_52 (Conv2D) (None, 16, 16, 192) 159936 inception_5a[0][0] \n__________________________________________________________________________________________________\nconv2d_54 (Conv2D) (None, 16, 16, 48) 39984 inception_5a[0][0] \n__________________________________________________________________________________________________\nmax_pooling2d_9 (MaxPooling2D) (None, 16, 16, 832) 0 inception_5a[0][0] \n__________________________________________________________________________________________________\nconv2d_51 (Conv2D) (None, 16, 16, 384) 319872 inception_5a[0][0] \n__________________________________________________________________________________________________\nconv2d_53 (Conv2D) (None, 16, 16, 384) 663936 conv2d_52[0][0] \n__________________________________________________________________________________________________\nconv2d_55 (Conv2D) (None, 16, 16, 128) 153728 conv2d_54[0][0] \n__________________________________________________________________________________________________\nconv2d_56 (Conv2D) (None, 16, 16, 128) 106624 max_pooling2d_9[0][0] \n__________________________________________________________________________________________________\naverage_pooling2d_1 (AveragePoo (None, 10, 10, 512) 0 inception_4a[0][0] \n__________________________________________________________________________________________________\naverage_pooling2d_2 (AveragePoo (None, 10, 10, 528) 0 inception_4d[0][0] \n__________________________________________________________________________________________________\ninception_5b (Concatenate) (None, 16, 16, 1024) 0 conv2d_51[0][0] \n conv2d_53[0][0] \n conv2d_55[0][0] \n conv2d_56[0][0] \n__________________________________________________________________________________________________\nconv2d_19 (Conv2D) (None, 10, 10, 128) 65664 average_pooling2d_1[0][0] \n__________________________________________________________________________________________________\nconv2d_38 (Conv2D) (None, 10, 10, 128) 67712 average_pooling2d_2[0][0] \n__________________________________________________________________________________________________\navg_pool_5_3x3/1 (GlobalAverage (None, 1024) 0 inception_5b[0][0] \n__________________________________________________________________________________________________\nflatten_1 (Flatten) (None, 12800) 0 conv2d_19[0][0] \n__________________________________________________________________________________________________\nflatten_2 (Flatten) (None, 12800) 0 conv2d_38[0][0] \n__________________________________________________________________________________________________\ndense_15 (Dense) (None, 6400) 6560000 avg_pool_5_3x3/1[0][0] \n__________________________________________________________________________________________________\ndense_1 (Dense) (None, 6400) 81926400 flatten_1[0][0] \n__________________________________________________________________________________________________\ndense_8 (Dense) (None, 6400) 81926400 flatten_2[0][0] \n__________________________________________________________________________________________________\ndense_16 (Dense) (None, 3200) 20483200 dense_15[0][0] \n__________________________________________________________________________________________________\ndense_2 (Dense) (None, 3200) 20483200 dense_1[0][0] \n__________________________________________________________________________________________________\ndense_9 (Dense) (None, 3200) 20483200 dense_8[0][0] \n__________________________________________________________________________________________________\ndense_17 (Dense) (None, 1600) 5121600 dense_16[0][0] \n__________________________________________________________________________________________________\ndense_3 (Dense) (None, 1600) 5121600 dense_2[0][0] \n__________________________________________________________________________________________________\ndense_10 (Dense) (None, 1600) 5121600 dense_9[0][0] \n__________________________________________________________________________________________________\ndense_18 (Dense) (None, 1024) 1639424 dense_17[0][0] \n__________________________________________________________________________________________________\ndense_4 (Dense) (None, 1024) 1639424 dense_3[0][0] \n__________________________________________________________________________________________________\ndense_11 (Dense) (None, 1024) 1639424 dense_10[0][0] \n__________________________________________________________________________________________________\ndense_19 (Dense) (None, 800) 820000 dense_18[0][0] \n__________________________________________________________________________________________________\ndense_5 (Dense) (None, 800) 820000 dense_4[0][0] \n__________________________________________________________________________________________________\ndense_12 (Dense) (None, 800) 820000 dense_11[0][0] \n__________________________________________________________________________________________________\ndense_20 (Dense) (None, 400) 320400 dense_19[0][0] \n__________________________________________________________________________________________________\ndense_6 (Dense) (None, 400) 320400 dense_5[0][0] \n__________________________________________________________________________________________________\ndense_13 (Dense) (None, 400) 320400 dense_12[0][0] \n__________________________________________________________________________________________________\ndense_21 (Dense) (None, 200) 80200 dense_20[0][0] \n__________________________________________________________________________________________________\ndense_7 (Dense) (None, 200) 80200 dense_6[0][0] \n__________________________________________________________________________________________________\ndense_14 (Dense) (None, 200) 80200 dense_13[0][0] \n__________________________________________________________________________________________________\ndropout_3 (Dropout) (None, 200) 0 dense_21[0][0] \n__________________________________________________________________________________________________\ndropout_1 (Dropout) (None, 200) 0 dense_7[0][0] \n__________________________________________________________________________________________________\ndropout_2 (Dropout) (None, 200) 0 dense_14[0][0] \n__________________________________________________________________________________________________\noutput (Dense) (None, 5) 1005 dropout_3[0][0] \n__________________________________________________________________________________________________\nauxilliary_output_1 (Dense) (None, 5) 1005 dropout_1[0][0] \n__________________________________________________________________________________________________\nauxilliary_output_2 (Dense) (None, 5) 1005 dropout_2[0][0] \n==================================================================================================\nTotal params: 261,910,943\nTrainable params: 261,910,943\nNon-trainable params: 0\n__________________________________________________________________________________________________\nNone\n","name":"stdout"}]},{"metadata":{"trusted":true},"cell_type":"code","source":"sgd = SGD(lr = 0.01, momentum=0.9, nesterov=False)\nmodel.compile(loss=['categorical_crossentropy', 'categorical_crossentropy', 'categorical_crossentropy'], loss_weights=[1, 0.3, 0.3], optimizer=sgd, metrics=['accuracy'])\nfrom keras.callbacks import ReduceLROnPlateau\nreduceLROnPlat = ReduceLROnPlateau(monitor='val_loss', factor=0.8, patience=3, verbose=1, mode='auto', epsilon=0.0001, cooldown=5, min_lr=0.0001)\n","execution_count":13,"outputs":[{"output_type":"stream","text":"/opt/conda/lib/python3.6/site-packages/keras/callbacks.py:1065: UserWarning: `epsilon` argument is deprecated and will be removed, use `min_delta` instead.\n warnings.warn('`epsilon` argument is deprecated and '\n","name":"stderr"}]},{"metadata":{"trusted":true},"cell_type":"code","source":"history = model.fit(X_train_resh, [Y_train, Y_train, Y_train], validation_data=(X_test_resh, [Y_test, Y_test, Y_test]), epochs=200, batch_size=32, callbacks=[reduceLROnPlat])","execution_count":14,"outputs":[{"output_type":"stream","text":"WARNING:tensorflow:From /opt/conda/lib/python3.6/site-packages/tensorflow/python/ops/math_ops.py:3066: to_int32 (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.\nInstructions for updating:\nUse tf.cast instead.\nTrain on 800 samples, validate on 100 samples\nEpoch 1/200\n800/800 [==============================] - 25s 31ms/step - loss: 2.6911 - output_loss: 1.6307 - auxilliary_output_1_loss: 1.8527 - auxilliary_output_2_loss: 1.6820 - output_acc: 0.1762 - auxilliary_output_1_acc: 0.1825 - auxilliary_output_2_acc: 0.1775 - val_loss: 2.6455 - val_output_loss: 1.6709 - val_auxilliary_output_1_loss: 1.6388 - val_auxilliary_output_2_loss: 1.6098 - val_output_acc: 0.0100 - val_auxilliary_output_1_acc: 0.0400 - val_auxilliary_output_2_acc: 0.0400\nEpoch 2/200\n800/800 [==============================] - 11s 13ms/step - loss: 2.5838 - output_loss: 1.6170 - auxilliary_output_1_loss: 1.6098 - auxilliary_output_2_loss: 1.6130 - output_acc: 0.2100 - auxilliary_output_1_acc: 0.2125 - auxilliary_output_2_acc: 0.1787 - val_loss: 2.5872 - val_output_loss: 1.6211 - val_auxilliary_output_1_loss: 1.6187 - val_auxilliary_output_2_loss: 1.6014 - val_output_acc: 0.0400 - val_auxilliary_output_1_acc: 0.0100 - val_auxilliary_output_2_acc: 0.7500\nEpoch 3/200\n800/800 [==============================] - 11s 13ms/step - loss: 2.5811 - output_loss: 1.6133 - auxilliary_output_1_loss: 1.6116 - auxilliary_output_2_loss: 1.6144 - output_acc: 0.2100 - auxilliary_output_1_acc: 0.2038 - auxilliary_output_2_acc: 0.1800 - val_loss: 2.6026 - val_output_loss: 1.6350 - val_auxilliary_output_1_loss: 1.6102 - val_auxilliary_output_2_loss: 1.6151 - val_output_acc: 0.1600 - val_auxilliary_output_1_acc: 0.0400 - val_auxilliary_output_2_acc: 0.0300\nEpoch 4/200\n800/800 [==============================] - 11s 13ms/step - loss: 2.5783 - output_loss: 1.6139 - auxilliary_output_1_loss: 1.6071 - auxilliary_output_2_loss: 1.6076 - output_acc: 0.2062 - auxilliary_output_1_acc: 0.2238 - auxilliary_output_2_acc: 0.2162 - val_loss: 2.5754 - val_output_loss: 1.6027 - val_auxilliary_output_1_loss: 1.6185 - val_auxilliary_output_2_loss: 1.6238 - val_output_acc: 0.0400 - val_auxilliary_output_1_acc: 0.0500 - val_auxilliary_output_2_acc: 0.0400\nEpoch 5/200\n800/800 [==============================] - 11s 14ms/step - loss: 2.5762 - output_loss: 1.6121 - auxilliary_output_1_loss: 1.6062 - auxilliary_output_2_loss: 1.6074 - output_acc: 0.1950 - auxilliary_output_1_acc: 0.2300 - auxilliary_output_2_acc: 0.2025 - val_loss: 2.5622 - val_output_loss: 1.6003 - val_auxilliary_output_1_loss: 1.5990 - val_auxilliary_output_2_loss: 1.6074 - val_output_acc: 0.7200 - val_auxilliary_output_1_acc: 0.4200 - val_auxilliary_output_2_acc: 0.0100\nEpoch 6/200\n800/800 [==============================] - 11s 13ms/step - loss: 2.5756 - output_loss: 1.6115 - auxilliary_output_1_loss: 1.6050 - auxilliary_output_2_loss: 1.6089 - output_acc: 0.2062 - auxilliary_output_1_acc: 0.2400 - auxilliary_output_2_acc: 0.2062 - val_loss: 2.5628 - val_output_loss: 1.6021 - val_auxilliary_output_1_loss: 1.5961 - val_auxilliary_output_2_loss: 1.6063 - val_output_acc: 0.7500 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.1000\nEpoch 7/200\n800/800 [==============================] - 11s 13ms/step - loss: 2.5756 - output_loss: 1.6127 - auxilliary_output_1_loss: 1.6044 - auxilliary_output_2_loss: 1.6054 - output_acc: 0.1900 - auxilliary_output_1_acc: 0.2213 - auxilliary_output_2_acc: 0.2363 - val_loss: 2.5947 - val_output_loss: 1.6310 - val_auxilliary_output_1_loss: 1.6135 - val_auxilliary_output_2_loss: 1.5988 - val_output_acc: 0.0400 - val_auxilliary_output_1_acc: 0.0400 - val_auxilliary_output_2_acc: 0.5600\nEpoch 8/200\n800/800 [==============================] - 11s 13ms/step - loss: 2.5690 - output_loss: 1.6113 - auxilliary_output_1_loss: 1.5972 - auxilliary_output_2_loss: 1.5953 - output_acc: 0.1900 - auxilliary_output_1_acc: 0.2375 - auxilliary_output_2_acc: 0.2662 - val_loss: 2.5923 - val_output_loss: 1.6284 - val_auxilliary_output_1_loss: 1.6003 - val_auxilliary_output_2_loss: 1.6128 - val_output_acc: 0.0400 - val_auxilliary_output_1_acc: 0.1200 - val_auxilliary_output_2_acc: 0.1000\n\nEpoch 00008: ReduceLROnPlateau reducing learning rate to 0.007999999821186066.\nEpoch 9/200\n800/800 [==============================] - 11s 13ms/step - loss: 2.5598 - output_loss: 1.6095 - auxilliary_output_1_loss: 1.5806 - auxilliary_output_2_loss: 1.5869 - output_acc: 0.2100 - auxilliary_output_1_acc: 0.2850 - auxilliary_output_2_acc: 0.2850 - val_loss: 2.5697 - val_output_loss: 1.6158 - val_auxilliary_output_1_loss: 1.5901 - val_auxilliary_output_2_loss: 1.5896 - val_output_acc: 0.1600 - val_auxilliary_output_1_acc: 0.1500 - val_auxilliary_output_2_acc: 0.0600\nEpoch 10/200\n800/800 [==============================] - 11s 13ms/step - loss: 2.5465 - output_loss: 1.6096 - auxilliary_output_1_loss: 1.5612 - auxilliary_output_2_loss: 1.5618 - output_acc: 0.2112 - auxilliary_output_1_acc: 0.2888 - auxilliary_output_2_acc: 0.3150 - val_loss: 2.5680 - val_output_loss: 1.6213 - val_auxilliary_output_1_loss: 1.5595 - val_auxilliary_output_2_loss: 1.5963 - val_output_acc: 0.0200 - val_auxilliary_output_1_acc: 0.0400 - val_auxilliary_output_2_acc: 0.0500\nEpoch 11/200\n800/800 [==============================] - 11s 13ms/step - loss: 2.5561 - output_loss: 1.6102 - auxilliary_output_1_loss: 1.5839 - auxilliary_output_2_loss: 1.5689 - output_acc: 0.2100 - auxilliary_output_1_acc: 0.2775 - auxilliary_output_2_acc: 0.2575 - val_loss: 2.6158 - val_output_loss: 1.5935 - val_auxilliary_output_1_loss: 1.7316 - val_auxilliary_output_2_loss: 1.6761 - val_output_acc: 0.7400 - val_auxilliary_output_1_acc: 0.0400 - val_auxilliary_output_2_acc: 0.0400\nEpoch 12/200\n800/800 [==============================] - 11s 13ms/step - loss: 2.5782 - output_loss: 1.6103 - auxilliary_output_1_loss: 1.6170 - auxilliary_output_2_loss: 1.6092 - output_acc: 0.2125 - auxilliary_output_1_acc: 0.2162 - auxilliary_output_2_acc: 0.2238 - val_loss: 2.5878 - val_output_loss: 1.6251 - val_auxilliary_output_1_loss: 1.5973 - val_auxilliary_output_2_loss: 1.6116 - val_output_acc: 0.1300 - val_auxilliary_output_1_acc: 0.4000 - val_auxilliary_output_2_acc: 0.1000\nEpoch 13/200\n800/800 [==============================] - 11s 14ms/step - loss: 2.5422 - output_loss: 1.6085 - auxilliary_output_1_loss: 1.5671 - auxilliary_output_2_loss: 1.5453 - output_acc: 0.2137 - auxilliary_output_1_acc: 0.2538 - auxilliary_output_2_acc: 0.3175 - val_loss: 2.5448 - val_output_loss: 1.6085 - val_auxilliary_output_1_loss: 1.5152 - val_auxilliary_output_2_loss: 1.6058 - val_output_acc: 0.0400 - val_auxilliary_output_1_acc: 0.4200 - val_auxilliary_output_2_acc: 0.0700\nEpoch 14/200\n800/800 [==============================] - 11s 14ms/step - loss: 2.5026 - output_loss: 1.6076 - auxilliary_output_1_loss: 1.5302 - auxilliary_output_2_loss: 1.4530 - output_acc: 0.2213 - auxilliary_output_1_acc: 0.2988 - auxilliary_output_2_acc: 0.3600 - val_loss: 2.5106 - val_output_loss: 1.6073 - val_auxilliary_output_1_loss: 1.5060 - val_auxilliary_output_2_loss: 1.5051 - val_output_acc: 0.0400 - val_auxilliary_output_1_acc: 0.2400 - val_auxilliary_output_2_acc: 0.1700\nEpoch 15/200\n800/800 [==============================] - 11s 13ms/step - loss: 2.4659 - output_loss: 1.6007 - auxilliary_output_1_loss: 1.4784 - auxilliary_output_2_loss: 1.4053 - output_acc: 0.2175 - auxilliary_output_1_acc: 0.3300 - auxilliary_output_2_acc: 0.4000 - val_loss: 2.6592 - val_output_loss: 1.5702 - val_auxilliary_output_1_loss: 1.8178 - val_auxilliary_output_2_loss: 1.8124 - val_output_acc: 0.0700 - val_auxilliary_output_1_acc: 0.0300 - val_auxilliary_output_2_acc: 0.0500\nEpoch 16/200\n800/800 [==============================] - 11s 13ms/step - loss: 2.3801 - output_loss: 1.5977 - auxilliary_output_1_loss: 1.3383 - auxilliary_output_2_loss: 1.2695 - output_acc: 0.1975 - auxilliary_output_1_acc: 0.4413 - auxilliary_output_2_acc: 0.4537 - val_loss: 2.5602 - val_output_loss: 1.5933 - val_auxilliary_output_1_loss: 1.6629 - val_auxilliary_output_2_loss: 1.5602 - val_output_acc: 0.0800 - val_auxilliary_output_1_acc: 0.1200 - val_auxilliary_output_2_acc: 0.2000\n","name":"stdout"},{"output_type":"stream","text":"Epoch 17/200\n800/800 [==============================] - 11s 13ms/step - loss: 2.4653 - output_loss: 1.6082 - auxilliary_output_1_loss: 1.3893 - auxilliary_output_2_loss: 1.4677 - output_acc: 0.2125 - auxilliary_output_1_acc: 0.4037 - auxilliary_output_2_acc: 0.3775 - val_loss: 2.5887 - val_output_loss: 1.6209 - val_auxilliary_output_1_loss: 1.6373 - val_auxilliary_output_2_loss: 1.5889 - val_output_acc: 0.0600 - val_auxilliary_output_1_acc: 0.1500 - val_auxilliary_output_2_acc: 0.0200\n\nEpoch 00017: ReduceLROnPlateau reducing learning rate to 0.006399999558925629.\nEpoch 18/200\n800/800 [==============================] - 11s 13ms/step - loss: 2.3220 - output_loss: 1.5845 - auxilliary_output_1_loss: 1.1772 - auxilliary_output_2_loss: 1.2810 - output_acc: 0.2387 - auxilliary_output_1_acc: 0.5225 - auxilliary_output_2_acc: 0.4713 - val_loss: 2.4283 - val_output_loss: 1.6162 - val_auxilliary_output_1_loss: 1.3162 - val_auxilliary_output_2_loss: 1.3908 - val_output_acc: 0.0500 - val_auxilliary_output_1_acc: 0.2500 - val_auxilliary_output_2_acc: 0.3000\nEpoch 19/200\n800/800 [==============================] - 11s 13ms/step - loss: 2.1889 - output_loss: 1.5690 - auxilliary_output_1_loss: 0.9935 - auxilliary_output_2_loss: 1.0727 - output_acc: 0.2387 - auxilliary_output_1_acc: 0.6112 - auxilliary_output_2_acc: 0.5875 - val_loss: 2.4643 - val_output_loss: 1.5808 - val_auxilliary_output_1_loss: 1.4595 - val_auxilliary_output_2_loss: 1.4855 - val_output_acc: 0.1000 - val_auxilliary_output_1_acc: 0.2700 - val_auxilliary_output_2_acc: 0.2900\nEpoch 20/200\n800/800 [==============================] - 11s 14ms/step - loss: 2.0888 - output_loss: 1.5629 - auxilliary_output_1_loss: 0.8263 - auxilliary_output_2_loss: 0.9266 - output_acc: 0.2575 - auxilliary_output_1_acc: 0.6713 - auxilliary_output_2_acc: 0.6512 - val_loss: 2.6390 - val_output_loss: 1.6397 - val_auxilliary_output_1_loss: 1.8087 - val_auxilliary_output_2_loss: 1.5222 - val_output_acc: 0.0500 - val_auxilliary_output_1_acc: 0.2400 - val_auxilliary_output_2_acc: 0.2500\nEpoch 21/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.9395 - output_loss: 1.5507 - auxilliary_output_1_loss: 0.6642 - auxilliary_output_2_loss: 0.6317 - output_acc: 0.2412 - auxilliary_output_1_acc: 0.7437 - auxilliary_output_2_acc: 0.7612 - val_loss: 2.4839 - val_output_loss: 1.6331 - val_auxilliary_output_1_loss: 1.3207 - val_auxilliary_output_2_loss: 1.5155 - val_output_acc: 0.0700 - val_auxilliary_output_1_acc: 0.4400 - val_auxilliary_output_2_acc: 0.4800\nEpoch 22/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.8618 - output_loss: 1.5349 - auxilliary_output_1_loss: 0.5466 - auxilliary_output_2_loss: 0.5433 - output_acc: 0.2625 - auxilliary_output_1_acc: 0.7838 - auxilliary_output_2_acc: 0.7913 - val_loss: 2.6034 - val_output_loss: 1.6229 - val_auxilliary_output_1_loss: 1.5085 - val_auxilliary_output_2_loss: 1.7598 - val_output_acc: 0.0100 - val_auxilliary_output_1_acc: 0.4600 - val_auxilliary_output_2_acc: 0.2800\nEpoch 23/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.8132 - output_loss: 1.5497 - auxilliary_output_1_loss: 0.4514 - auxilliary_output_2_loss: 0.4269 - output_acc: 0.2662 - auxilliary_output_1_acc: 0.8337 - auxilliary_output_2_acc: 0.8488 - val_loss: 2.5335 - val_output_loss: 1.6143 - val_auxilliary_output_1_loss: 1.4726 - val_auxilliary_output_2_loss: 1.5912 - val_output_acc: 0.1400 - val_auxilliary_output_1_acc: 0.4900 - val_auxilliary_output_2_acc: 0.5400\nEpoch 24/200\n800/800 [==============================] - 11s 14ms/step - loss: 1.8517 - output_loss: 1.5358 - auxilliary_output_1_loss: 0.5596 - auxilliary_output_2_loss: 0.4934 - output_acc: 0.2725 - auxilliary_output_1_acc: 0.7900 - auxilliary_output_2_acc: 0.8163 - val_loss: 2.6251 - val_output_loss: 1.6632 - val_auxilliary_output_1_loss: 1.9674 - val_auxilliary_output_2_loss: 1.2389 - val_output_acc: 0.0600 - val_auxilliary_output_1_acc: 0.3400 - val_auxilliary_output_2_acc: 0.5100\n\nEpoch 00024: ReduceLROnPlateau reducing learning rate to 0.0051199994981288915.\nEpoch 25/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.8243 - output_loss: 1.5444 - auxilliary_output_1_loss: 0.6351 - auxilliary_output_2_loss: 0.2977 - output_acc: 0.2575 - auxilliary_output_1_acc: 0.8062 - auxilliary_output_2_acc: 0.8987 - val_loss: 2.4310 - val_output_loss: 1.5610 - val_auxilliary_output_1_loss: 1.3677 - val_auxilliary_output_2_loss: 1.5325 - val_output_acc: 0.0400 - val_auxilliary_output_1_acc: 0.4700 - val_auxilliary_output_2_acc: 0.5200\nEpoch 26/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.7025 - output_loss: 1.5390 - auxilliary_output_1_loss: 0.3282 - auxilliary_output_2_loss: 0.2168 - output_acc: 0.2637 - auxilliary_output_1_acc: 0.8813 - auxilliary_output_2_acc: 0.9288 - val_loss: 2.5634 - val_output_loss: 1.5599 - val_auxilliary_output_1_loss: 1.5419 - val_auxilliary_output_2_loss: 1.8029 - val_output_acc: 0.6400 - val_auxilliary_output_1_acc: 0.5500 - val_auxilliary_output_2_acc: 0.4000\nEpoch 27/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.6133 - output_loss: 1.5216 - auxilliary_output_1_loss: 0.1727 - auxilliary_output_2_loss: 0.1329 - output_acc: 0.2963 - auxilliary_output_1_acc: 0.9487 - auxilliary_output_2_acc: 0.9513 - val_loss: 2.7923 - val_output_loss: 1.5637 - val_auxilliary_output_1_loss: 2.0401 - val_auxilliary_output_2_loss: 2.0554 - val_output_acc: 0.3500 - val_auxilliary_output_1_acc: 0.4400 - val_auxilliary_output_2_acc: 0.5400\nEpoch 28/200\n800/800 [==============================] - 11s 14ms/step - loss: 1.6106 - output_loss: 1.5181 - auxilliary_output_1_loss: 0.2034 - auxilliary_output_2_loss: 0.1048 - output_acc: 0.2963 - auxilliary_output_1_acc: 0.9350 - auxilliary_output_2_acc: 0.9650 - val_loss: 2.6514 - val_output_loss: 1.7128 - val_auxilliary_output_1_loss: 1.4159 - val_auxilliary_output_2_loss: 1.7128 - val_output_acc: 0.0500 - val_auxilliary_output_1_acc: 0.4700 - val_auxilliary_output_2_acc: 0.4700\nEpoch 29/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.6094 - output_loss: 1.5259 - auxilliary_output_1_loss: 0.1609 - auxilliary_output_2_loss: 0.1173 - output_acc: 0.2825 - auxilliary_output_1_acc: 0.9550 - auxilliary_output_2_acc: 0.9662 - val_loss: 2.7557 - val_output_loss: 1.5739 - val_auxilliary_output_1_loss: 1.6290 - val_auxilliary_output_2_loss: 2.3102 - val_output_acc: 0.4300 - val_auxilliary_output_1_acc: 0.6300 - val_auxilliary_output_2_acc: 0.5700\nEpoch 30/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.5721 - output_loss: 1.5045 - auxilliary_output_1_loss: 0.1176 - auxilliary_output_2_loss: 0.1074 - output_acc: 0.3050 - auxilliary_output_1_acc: 0.9688 - auxilliary_output_2_acc: 0.9725 - val_loss: 2.8674 - val_output_loss: 1.7101 - val_auxilliary_output_1_loss: 1.5365 - val_auxilliary_output_2_loss: 2.3214 - val_output_acc: 0.0400 - val_auxilliary_output_1_acc: 0.4400 - val_auxilliary_output_2_acc: 0.4000\nEpoch 31/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.5593 - output_loss: 1.4963 - auxilliary_output_1_loss: 0.0929 - auxilliary_output_2_loss: 0.1172 - output_acc: 0.3012 - auxilliary_output_1_acc: 0.9725 - auxilliary_output_2_acc: 0.9713 - val_loss: 2.8112 - val_output_loss: 1.6520 - val_auxilliary_output_1_loss: 1.7932 - val_auxilliary_output_2_loss: 2.0709 - val_output_acc: 0.2200 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.6000\n\nEpoch 00031: ReduceLROnPlateau reducing learning rate to 0.004095999523997307.\nEpoch 32/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.5225 - output_loss: 1.4813 - auxilliary_output_1_loss: 0.0679 - auxilliary_output_2_loss: 0.0695 - output_acc: 0.3087 - auxilliary_output_1_acc: 0.9800 - auxilliary_output_2_acc: 0.9862 - val_loss: 3.1913 - val_output_loss: 1.7077 - val_auxilliary_output_1_loss: 2.0671 - val_auxilliary_output_2_loss: 2.8783 - val_output_acc: 0.0600 - val_auxilliary_output_1_acc: 0.6200 - val_auxilliary_output_2_acc: 0.4900\nEpoch 33/200\n","name":"stdout"},{"output_type":"stream","text":"800/800 [==============================] - 11s 13ms/step - loss: 1.5120 - output_loss: 1.4888 - auxilliary_output_1_loss: 0.0305 - auxilliary_output_2_loss: 0.0468 - output_acc: 0.3100 - auxilliary_output_1_acc: 0.9962 - auxilliary_output_2_acc: 0.9888 - val_loss: 2.9181 - val_output_loss: 1.6608 - val_auxilliary_output_1_loss: 2.2888 - val_auxilliary_output_2_loss: 1.9020 - val_output_acc: 0.0200 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.6400\nEpoch 34/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.4712 - output_loss: 1.4468 - auxilliary_output_1_loss: 0.0369 - auxilliary_output_2_loss: 0.0443 - output_acc: 0.3463 - auxilliary_output_1_acc: 0.9862 - auxilliary_output_2_acc: 0.9888 - val_loss: 3.1990 - val_output_loss: 1.6390 - val_auxilliary_output_1_loss: 2.2523 - val_auxilliary_output_2_loss: 2.9479 - val_output_acc: 0.4000 - val_auxilliary_output_1_acc: 0.5600 - val_auxilliary_output_2_acc: 0.5000\nEpoch 35/200\n800/800 [==============================] - 11s 14ms/step - loss: 1.4787 - output_loss: 1.4625 - auxilliary_output_1_loss: 0.0167 - auxilliary_output_2_loss: 0.0375 - output_acc: 0.3150 - auxilliary_output_1_acc: 0.9962 - auxilliary_output_2_acc: 0.9937 - val_loss: 3.6845 - val_output_loss: 1.5432 - val_auxilliary_output_1_loss: 3.6926 - val_auxilliary_output_2_loss: 3.4451 - val_output_acc: 0.7100 - val_auxilliary_output_1_acc: 0.4300 - val_auxilliary_output_2_acc: 0.6000\nEpoch 36/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.6031 - output_loss: 1.5409 - auxilliary_output_1_loss: 0.0312 - auxilliary_output_2_loss: 0.1761 - output_acc: 0.2737 - auxilliary_output_1_acc: 0.9937 - auxilliary_output_2_acc: 0.9500 - val_loss: 2.9056 - val_output_loss: 1.6534 - val_auxilliary_output_1_loss: 2.3756 - val_auxilliary_output_2_loss: 1.7985 - val_output_acc: 0.0200 - val_auxilliary_output_1_acc: 0.6600 - val_auxilliary_output_2_acc: 0.5900\nEpoch 37/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.4878 - output_loss: 1.4629 - auxilliary_output_1_loss: 0.0321 - auxilliary_output_2_loss: 0.0508 - output_acc: 0.3337 - auxilliary_output_1_acc: 0.9925 - auxilliary_output_2_acc: 0.9888 - val_loss: 3.0593 - val_output_loss: 1.6761 - val_auxilliary_output_1_loss: 2.6605 - val_auxilliary_output_2_loss: 1.9504 - val_output_acc: 0.0500 - val_auxilliary_output_1_acc: 0.6000 - val_auxilliary_output_2_acc: 0.6700\nEpoch 38/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.4955 - output_loss: 1.4688 - auxilliary_output_1_loss: 0.0527 - auxilliary_output_2_loss: 0.0362 - output_acc: 0.3012 - auxilliary_output_1_acc: 0.9925 - auxilliary_output_2_acc: 0.9888 - val_loss: 3.1048 - val_output_loss: 1.6414 - val_auxilliary_output_1_loss: 2.5127 - val_auxilliary_output_2_loss: 2.3653 - val_output_acc: 0.3300 - val_auxilliary_output_1_acc: 0.5600 - val_auxilliary_output_2_acc: 0.6200\n\nEpoch 00038: ReduceLROnPlateau reducing learning rate to 0.0032767996191978457.\nEpoch 39/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.4489 - output_loss: 1.4369 - auxilliary_output_1_loss: 0.0227 - auxilliary_output_2_loss: 0.0174 - output_acc: 0.3488 - auxilliary_output_1_acc: 0.9937 - auxilliary_output_2_acc: 0.9975 - val_loss: 3.2697 - val_output_loss: 1.6067 - val_auxilliary_output_1_loss: 2.5071 - val_auxilliary_output_2_loss: 3.0362 - val_output_acc: 0.3700 - val_auxilliary_output_1_acc: 0.6400 - val_auxilliary_output_2_acc: 0.5300\nEpoch 40/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.4677 - output_loss: 1.4555 - auxilliary_output_1_loss: 0.0284 - auxilliary_output_2_loss: 0.0122 - output_acc: 0.3488 - auxilliary_output_1_acc: 0.9913 - auxilliary_output_2_acc: 0.9962 - val_loss: 3.2465 - val_output_loss: 1.5308 - val_auxilliary_output_1_loss: 3.3425 - val_auxilliary_output_2_loss: 2.3767 - val_output_acc: 0.4500 - val_auxilliary_output_1_acc: 0.5100 - val_auxilliary_output_2_acc: 0.6500\nEpoch 41/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.4743 - output_loss: 1.4542 - auxilliary_output_1_loss: 0.0560 - auxilliary_output_2_loss: 0.0109 - output_acc: 0.2950 - auxilliary_output_1_acc: 0.9862 - auxilliary_output_2_acc: 0.9975 - val_loss: 3.2529 - val_output_loss: 1.6840 - val_auxilliary_output_1_loss: 2.6767 - val_auxilliary_output_2_loss: 2.5530 - val_output_acc: 0.0800 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5900\nEpoch 42/200\n800/800 [==============================] - 11s 14ms/step - loss: 1.5707 - output_loss: 1.5335 - auxilliary_output_1_loss: 0.0687 - auxilliary_output_2_loss: 0.0552 - output_acc: 0.2913 - auxilliary_output_1_acc: 0.9825 - auxilliary_output_2_acc: 0.9838 - val_loss: 2.7978 - val_output_loss: 1.5216 - val_auxilliary_output_1_loss: 2.1146 - val_auxilliary_output_2_loss: 2.1393 - val_output_acc: 0.6100 - val_auxilliary_output_1_acc: 0.4500 - val_auxilliary_output_2_acc: 0.4200\nEpoch 43/200\n800/800 [==============================] - 11s 14ms/step - loss: 1.6262 - output_loss: 1.5980 - auxilliary_output_1_loss: 0.0541 - auxilliary_output_2_loss: 0.0401 - output_acc: 0.2337 - auxilliary_output_1_acc: 0.9862 - auxilliary_output_2_acc: 0.9900 - val_loss: 3.3244 - val_output_loss: 1.5958 - val_auxilliary_output_1_loss: 3.4505 - val_auxilliary_output_2_loss: 2.3114 - val_output_acc: 0.0400 - val_auxilliary_output_1_acc: 0.4200 - val_auxilliary_output_2_acc: 0.5900\nEpoch 44/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.5825 - output_loss: 1.5667 - auxilliary_output_1_loss: 0.0388 - auxilliary_output_2_loss: 0.0139 - output_acc: 0.2250 - auxilliary_output_1_acc: 0.9875 - auxilliary_output_2_acc: 0.9988 - val_loss: 2.9481 - val_output_loss: 1.6367 - val_auxilliary_output_1_loss: 2.1067 - val_auxilliary_output_2_loss: 2.2645 - val_output_acc: 0.0400 - val_auxilliary_output_1_acc: 0.6100 - val_auxilliary_output_2_acc: 0.5700\nEpoch 45/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.5513 - output_loss: 1.5418 - auxilliary_output_1_loss: 0.0196 - auxilliary_output_2_loss: 0.0122 - output_acc: 0.2538 - auxilliary_output_1_acc: 0.9950 - auxilliary_output_2_acc: 0.9962 - val_loss: 3.1420 - val_output_loss: 1.6456 - val_auxilliary_output_1_loss: 2.3378 - val_auxilliary_output_2_loss: 2.6502 - val_output_acc: 0.4200 - val_auxilliary_output_1_acc: 0.6400 - val_auxilliary_output_2_acc: 0.5600\n\nEpoch 00045: ReduceLROnPlateau reducing learning rate to 0.0026214396581053737.\nEpoch 46/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.5088 - output_loss: 1.5010 - auxilliary_output_1_loss: 0.0068 - auxilliary_output_2_loss: 0.0195 - output_acc: 0.2762 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9975 - val_loss: 3.3703 - val_output_loss: 1.6086 - val_auxilliary_output_1_loss: 3.0104 - val_auxilliary_output_2_loss: 2.8620 - val_output_acc: 0.3700 - val_auxilliary_output_1_acc: 0.6100 - val_auxilliary_output_2_acc: 0.6400\nEpoch 47/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.4683 - output_loss: 1.4635 - auxilliary_output_1_loss: 0.0090 - auxilliary_output_2_loss: 0.0070 - output_acc: 0.3287 - auxilliary_output_1_acc: 0.9975 - auxilliary_output_2_acc: 0.9988 - val_loss: 3.5082 - val_output_loss: 1.7093 - val_auxilliary_output_1_loss: 2.9959 - val_auxilliary_output_2_loss: 3.0001 - val_output_acc: 0.0200 - val_auxilliary_output_1_acc: 0.6300 - val_auxilliary_output_2_acc: 0.6300\nEpoch 48/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.4158 - output_loss: 1.4094 - auxilliary_output_1_loss: 0.0041 - auxilliary_output_2_loss: 0.0172 - output_acc: 0.3575 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9950 - val_loss: 3.5616 - val_output_loss: 1.6884 - val_auxilliary_output_1_loss: 3.3118 - val_auxilliary_output_2_loss: 2.9320 - val_output_acc: 0.1500 - val_auxilliary_output_1_acc: 0.6000 - val_auxilliary_output_2_acc: 0.6500\nEpoch 49/200\n","name":"stdout"},{"output_type":"stream","text":"800/800 [==============================] - 11s 13ms/step - loss: 1.4600 - output_loss: 1.4553 - auxilliary_output_1_loss: 0.0091 - auxilliary_output_2_loss: 0.0064 - output_acc: 0.3400 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 3.4613 - val_output_loss: 1.5201 - val_auxilliary_output_1_loss: 3.3277 - val_auxilliary_output_2_loss: 3.1433 - val_output_acc: 0.6100 - val_auxilliary_output_1_acc: 0.6100 - val_auxilliary_output_2_acc: 0.6100\nEpoch 50/200\n800/800 [==============================] - 11s 14ms/step - loss: 1.5203 - output_loss: 1.5149 - auxilliary_output_1_loss: 0.0078 - auxilliary_output_2_loss: 0.0103 - output_acc: 0.3050 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 3.3333 - val_output_loss: 1.6256 - val_auxilliary_output_1_loss: 3.0208 - val_auxilliary_output_2_loss: 2.6714 - val_output_acc: 0.0600 - val_auxilliary_output_1_acc: 0.6100 - val_auxilliary_output_2_acc: 0.5800\nEpoch 51/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.4530 - output_loss: 1.4494 - auxilliary_output_1_loss: 0.0060 - auxilliary_output_2_loss: 0.0058 - output_acc: 0.3525 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 3.6519 - val_output_loss: 1.7012 - val_auxilliary_output_1_loss: 3.3781 - val_auxilliary_output_2_loss: 3.1244 - val_output_acc: 0.0500 - val_auxilliary_output_1_acc: 0.6100 - val_auxilliary_output_2_acc: 0.6200\nEpoch 52/200\n800/800 [==============================] - 11s 14ms/step - loss: 1.3656 - output_loss: 1.3618 - auxilliary_output_1_loss: 0.0041 - auxilliary_output_2_loss: 0.0087 - output_acc: 0.4000 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9975 - val_loss: 3.7048 - val_output_loss: 1.6341 - val_auxilliary_output_1_loss: 3.5471 - val_auxilliary_output_2_loss: 3.3553 - val_output_acc: 0.1600 - val_auxilliary_output_1_acc: 0.6100 - val_auxilliary_output_2_acc: 0.6500\n\nEpoch 00052: ReduceLROnPlateau reducing learning rate to 0.0020971518009901048.\nEpoch 53/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.3250 - output_loss: 1.3220 - auxilliary_output_1_loss: 0.0057 - auxilliary_output_2_loss: 0.0042 - output_acc: 0.3850 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 3.9546 - val_output_loss: 1.7032 - val_auxilliary_output_1_loss: 3.8264 - val_auxilliary_output_2_loss: 3.6783 - val_output_acc: 0.0900 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.6100\nEpoch 54/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.3084 - output_loss: 1.3051 - auxilliary_output_1_loss: 0.0045 - auxilliary_output_2_loss: 0.0065 - output_acc: 0.4225 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 4.0401 - val_output_loss: 1.6428 - val_auxilliary_output_1_loss: 4.1095 - val_auxilliary_output_2_loss: 3.8816 - val_output_acc: 0.3000 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.6300\nEpoch 55/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.3608 - output_loss: 1.3575 - auxilliary_output_1_loss: 0.0050 - auxilliary_output_2_loss: 0.0060 - output_acc: 0.3675 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 3.9018 - val_output_loss: 1.5822 - val_auxilliary_output_1_loss: 4.0857 - val_auxilliary_output_2_loss: 3.6465 - val_output_acc: 0.4200 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.6100\nEpoch 56/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.3483 - output_loss: 1.3452 - auxilliary_output_1_loss: 0.0056 - auxilliary_output_2_loss: 0.0047 - output_acc: 0.3775 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 4.1823 - val_output_loss: 1.7622 - val_auxilliary_output_1_loss: 4.1509 - val_auxilliary_output_2_loss: 3.9161 - val_output_acc: 0.0400 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5800\nEpoch 57/200\n800/800 [==============================] - 11s 14ms/step - loss: 1.2623 - output_loss: 1.2599 - auxilliary_output_1_loss: 0.0052 - auxilliary_output_2_loss: 0.0028 - output_acc: 0.4313 - auxilliary_output_1_acc: 0.9975 - auxilliary_output_2_acc: 0.9988 - val_loss: 4.0120 - val_output_loss: 1.6065 - val_auxilliary_output_1_loss: 4.1544 - val_auxilliary_output_2_loss: 3.8637 - val_output_acc: 0.3600 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5800\nEpoch 58/200\n800/800 [==============================] - 11s 14ms/step - loss: 1.5164 - output_loss: 1.5129 - auxilliary_output_1_loss: 0.0033 - auxilliary_output_2_loss: 0.0083 - output_acc: 0.2963 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 3.3799 - val_output_loss: 1.5424 - val_auxilliary_output_1_loss: 3.4229 - val_auxilliary_output_2_loss: 2.7020 - val_output_acc: 0.2800 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5100\nEpoch 59/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.4165 - output_loss: 1.4129 - auxilliary_output_1_loss: 0.0057 - auxilliary_output_2_loss: 0.0065 - output_acc: 0.3562 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9975 - val_loss: 3.9742 - val_output_loss: 1.6767 - val_auxilliary_output_1_loss: 3.9240 - val_auxilliary_output_2_loss: 3.7342 - val_output_acc: 0.0800 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5400\n\nEpoch 00059: ReduceLROnPlateau reducing learning rate to 0.0016777213662862779.\nEpoch 60/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.2765 - output_loss: 1.2733 - auxilliary_output_1_loss: 0.0060 - auxilliary_output_2_loss: 0.0047 - output_acc: 0.4275 - auxilliary_output_1_acc: 0.9975 - auxilliary_output_2_acc: 0.9975 - val_loss: 4.2106 - val_output_loss: 1.7144 - val_auxilliary_output_1_loss: 3.9281 - val_auxilliary_output_2_loss: 4.3923 - val_output_acc: 0.2100 - val_auxilliary_output_1_acc: 0.6100 - val_auxilliary_output_2_acc: 0.5400\nEpoch 61/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.1696 - output_loss: 1.1676 - auxilliary_output_1_loss: 0.0038 - auxilliary_output_2_loss: 0.0029 - output_acc: 0.5025 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 4.2137 - val_output_loss: 1.7143 - val_auxilliary_output_1_loss: 4.1673 - val_auxilliary_output_2_loss: 4.1641 - val_output_acc: 0.1400 - val_auxilliary_output_1_acc: 0.6000 - val_auxilliary_output_2_acc: 0.6400\nEpoch 62/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.1302 - output_loss: 1.1273 - auxilliary_output_1_loss: 0.0058 - auxilliary_output_2_loss: 0.0037 - output_acc: 0.5212 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9975 - val_loss: 4.1596 - val_output_loss: 1.6230 - val_auxilliary_output_1_loss: 4.2324 - val_auxilliary_output_2_loss: 4.2230 - val_output_acc: 0.2300 - val_auxilliary_output_1_acc: 0.6000 - val_auxilliary_output_2_acc: 0.6200\nEpoch 63/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.0634 - output_loss: 1.0615 - auxilliary_output_1_loss: 0.0042 - auxilliary_output_2_loss: 0.0021 - output_acc: 0.5463 - auxilliary_output_1_acc: 0.9975 - auxilliary_output_2_acc: 0.9988 - val_loss: 4.2048 - val_output_loss: 1.6342 - val_auxilliary_output_1_loss: 4.2487 - val_auxilliary_output_2_loss: 4.3201 - val_output_acc: 0.1700 - val_auxilliary_output_1_acc: 0.6100 - val_auxilliary_output_2_acc: 0.6400\nEpoch 64/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.2714 - output_loss: 1.2698 - auxilliary_output_1_loss: 0.0025 - auxilliary_output_2_loss: 0.0027 - output_acc: 0.4612 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 3.7696 - val_output_loss: 1.6613 - val_auxilliary_output_1_loss: 3.8503 - val_auxilliary_output_2_loss: 3.1771 - val_output_acc: 0.1400 - val_auxilliary_output_1_acc: 0.6100 - val_auxilliary_output_2_acc: 0.6100\nEpoch 65/200\n","name":"stdout"},{"output_type":"stream","text":"800/800 [==============================] - 11s 14ms/step - loss: 1.0750 - output_loss: 1.0730 - auxilliary_output_1_loss: 0.0030 - auxilliary_output_2_loss: 0.0036 - output_acc: 0.5463 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9975 - val_loss: 4.2680 - val_output_loss: 1.6896 - val_auxilliary_output_1_loss: 4.1930 - val_auxilliary_output_2_loss: 4.4017 - val_output_acc: 0.2100 - val_auxilliary_output_1_acc: 0.6100 - val_auxilliary_output_2_acc: 0.6300\nEpoch 66/200\n800/800 [==============================] - 11s 14ms/step - loss: 1.0180 - output_loss: 1.0120 - auxilliary_output_1_loss: 0.0107 - auxilliary_output_2_loss: 0.0092 - output_acc: 0.5637 - auxilliary_output_1_acc: 0.9975 - auxilliary_output_2_acc: 0.9975 - val_loss: 4.4757 - val_output_loss: 1.6979 - val_auxilliary_output_1_loss: 4.1858 - val_auxilliary_output_2_loss: 5.0734 - val_output_acc: 0.1100 - val_auxilliary_output_1_acc: 0.6600 - val_auxilliary_output_2_acc: 0.5100\n\nEpoch 00066: ReduceLROnPlateau reducing learning rate to 0.0013421771116554739.\nEpoch 67/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.0345 - output_loss: 1.0237 - auxilliary_output_1_loss: 0.0318 - auxilliary_output_2_loss: 0.0043 - output_acc: 0.5637 - auxilliary_output_1_acc: 0.9925 - auxilliary_output_2_acc: 0.9975 - val_loss: 3.9543 - val_output_loss: 1.6761 - val_auxilliary_output_1_loss: 3.6205 - val_auxilliary_output_2_loss: 3.9734 - val_output_acc: 0.2800 - val_auxilliary_output_1_acc: 0.6500 - val_auxilliary_output_2_acc: 0.6000\nEpoch 68/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.0849 - output_loss: 1.0805 - auxilliary_output_1_loss: 0.0096 - auxilliary_output_2_loss: 0.0050 - output_acc: 0.5425 - auxilliary_output_1_acc: 0.9975 - auxilliary_output_2_acc: 0.9988 - val_loss: 4.1154 - val_output_loss: 1.6500 - val_auxilliary_output_1_loss: 4.2217 - val_auxilliary_output_2_loss: 3.9965 - val_output_acc: 0.1700 - val_auxilliary_output_1_acc: 0.6000 - val_auxilliary_output_2_acc: 0.6000\nEpoch 69/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.9202 - output_loss: 0.9028 - auxilliary_output_1_loss: 0.0551 - auxilliary_output_2_loss: 0.0030 - output_acc: 0.6050 - auxilliary_output_1_acc: 0.9888 - auxilliary_output_2_acc: 0.9988 - val_loss: 4.2887 - val_output_loss: 1.4134 - val_auxilliary_output_1_loss: 5.1411 - val_auxilliary_output_2_loss: 4.4431 - val_output_acc: 0.2200 - val_auxilliary_output_1_acc: 0.4900 - val_auxilliary_output_2_acc: 0.5900\nEpoch 70/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.0345 - output_loss: 1.0243 - auxilliary_output_1_loss: 0.0312 - auxilliary_output_2_loss: 0.0030 - output_acc: 0.5575 - auxilliary_output_1_acc: 0.9900 - auxilliary_output_2_acc: 0.9988 - val_loss: 3.7736 - val_output_loss: 1.5161 - val_auxilliary_output_1_loss: 3.8515 - val_auxilliary_output_2_loss: 3.6736 - val_output_acc: 0.3000 - val_auxilliary_output_1_acc: 0.6100 - val_auxilliary_output_2_acc: 0.6100\nEpoch 71/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.8012 - output_loss: 0.7989 - auxilliary_output_1_loss: 0.0049 - auxilliary_output_2_loss: 0.0025 - output_acc: 0.6475 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 4.1404 - val_output_loss: 1.5724 - val_auxilliary_output_1_loss: 4.1506 - val_auxilliary_output_2_loss: 4.4094 - val_output_acc: 0.1800 - val_auxilliary_output_1_acc: 0.6200 - val_auxilliary_output_2_acc: 0.5900\nEpoch 72/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.7333 - output_loss: 0.7317 - auxilliary_output_1_loss: 0.0033 - auxilliary_output_2_loss: 0.0024 - output_acc: 0.6638 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 4.0637 - val_output_loss: 1.5124 - val_auxilliary_output_1_loss: 3.9853 - val_auxilliary_output_2_loss: 4.5189 - val_output_acc: 0.3500 - val_auxilliary_output_1_acc: 0.6300 - val_auxilliary_output_2_acc: 0.5900\nEpoch 73/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.7506 - output_loss: 0.7484 - auxilliary_output_1_loss: 0.0052 - auxilliary_output_2_loss: 0.0022 - output_acc: 0.6737 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 3.9775 - val_output_loss: 1.5148 - val_auxilliary_output_1_loss: 3.7450 - val_auxilliary_output_2_loss: 4.4639 - val_output_acc: 0.2300 - val_auxilliary_output_1_acc: 0.6600 - val_auxilliary_output_2_acc: 0.5900\n\nEpoch 00073: ReduceLROnPlateau reducing learning rate to 0.001073741726577282.\nEpoch 74/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.7902 - output_loss: 0.7881 - auxilliary_output_1_loss: 0.0043 - auxilliary_output_2_loss: 0.0026 - output_acc: 0.6825 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 4.0767 - val_output_loss: 1.5287 - val_auxilliary_output_1_loss: 3.7934 - val_auxilliary_output_2_loss: 4.6998 - val_output_acc: 0.3200 - val_auxilliary_output_1_acc: 0.6600 - val_auxilliary_output_2_acc: 0.5700\nEpoch 75/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.6857 - output_loss: 0.6839 - auxilliary_output_1_loss: 0.0032 - auxilliary_output_2_loss: 0.0028 - output_acc: 0.7150 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 4.2000 - val_output_loss: 1.5874 - val_auxilliary_output_1_loss: 3.8916 - val_auxilliary_output_2_loss: 4.8171 - val_output_acc: 0.3500 - val_auxilliary_output_1_acc: 0.6700 - val_auxilliary_output_2_acc: 0.5700\nEpoch 76/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.5600 - output_loss: 0.5587 - auxilliary_output_1_loss: 0.0025 - auxilliary_output_2_loss: 0.0020 - output_acc: 0.7612 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 4.2567 - val_output_loss: 1.5268 - val_auxilliary_output_1_loss: 4.1263 - val_auxilliary_output_2_loss: 4.9731 - val_output_acc: 0.4800 - val_auxilliary_output_1_acc: 0.6500 - val_auxilliary_output_2_acc: 0.5800\nEpoch 77/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.9739 - output_loss: 0.9718 - auxilliary_output_1_loss: 0.0034 - auxilliary_output_2_loss: 0.0035 - output_acc: 0.6288 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9975 - val_loss: 3.7107 - val_output_loss: 1.2337 - val_auxilliary_output_1_loss: 3.6632 - val_auxilliary_output_2_loss: 4.5934 - val_output_acc: 0.2700 - val_auxilliary_output_1_acc: 0.6500 - val_auxilliary_output_2_acc: 0.5700\nEpoch 78/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.3957 - output_loss: 1.3934 - auxilliary_output_1_loss: 0.0028 - auxilliary_output_2_loss: 0.0049 - output_acc: 0.3737 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 3.2675 - val_output_loss: 1.5755 - val_auxilliary_output_1_loss: 3.1310 - val_auxilliary_output_2_loss: 2.5088 - val_output_acc: 0.1500 - val_auxilliary_output_1_acc: 0.6400 - val_auxilliary_output_2_acc: 0.5900\nEpoch 79/200\n800/800 [==============================] - 11s 13ms/step - loss: 1.1487 - output_loss: 1.1462 - auxilliary_output_1_loss: 0.0051 - auxilliary_output_2_loss: 0.0033 - output_acc: 0.4913 - auxilliary_output_1_acc: 0.9975 - auxilliary_output_2_acc: 0.9988 - val_loss: 3.7501 - val_output_loss: 1.6057 - val_auxilliary_output_1_loss: 3.5502 - val_auxilliary_output_2_loss: 3.5976 - val_output_acc: 0.2400 - val_auxilliary_output_1_acc: 0.6600 - val_auxilliary_output_2_acc: 0.5700\nEpoch 80/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.9187 - output_loss: 0.9171 - auxilliary_output_1_loss: 0.0031 - auxilliary_output_2_loss: 0.0020 - output_acc: 0.6187 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 4.3715 - val_output_loss: 2.0599 - val_auxilliary_output_1_loss: 3.7097 - val_auxilliary_output_2_loss: 3.9956 - val_output_acc: 0.2200 - val_auxilliary_output_1_acc: 0.6300 - val_auxilliary_output_2_acc: 0.6100\n\nEpoch 00080: ReduceLROnPlateau reducing learning rate to 0.0008589933626353742.\nEpoch 81/200\n","name":"stdout"},{"output_type":"stream","text":"800/800 [==============================] - 11s 14ms/step - loss: 0.7662 - output_loss: 0.7646 - auxilliary_output_1_loss: 0.0028 - auxilliary_output_2_loss: 0.0027 - output_acc: 0.6775 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 4.0275 - val_output_loss: 1.5477 - val_auxilliary_output_1_loss: 3.8804 - val_auxilliary_output_2_loss: 4.3856 - val_output_acc: 0.3000 - val_auxilliary_output_1_acc: 0.6500 - val_auxilliary_output_2_acc: 0.6100\nEpoch 82/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.5330 - output_loss: 0.5317 - auxilliary_output_1_loss: 0.0027 - auxilliary_output_2_loss: 0.0017 - output_acc: 0.7762 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 4.3419 - val_output_loss: 1.6273 - val_auxilliary_output_1_loss: 4.1465 - val_auxilliary_output_2_loss: 4.9020 - val_output_acc: 0.3600 - val_auxilliary_output_1_acc: 0.6500 - val_auxilliary_output_2_acc: 0.5800\nEpoch 83/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.7542 - output_loss: 0.7503 - auxilliary_output_1_loss: 0.0058 - auxilliary_output_2_loss: 0.0071 - output_acc: 0.6737 - auxilliary_output_1_acc: 0.9975 - auxilliary_output_2_acc: 0.9988 - val_loss: 4.0665 - val_output_loss: 1.6148 - val_auxilliary_output_1_loss: 3.5186 - val_auxilliary_output_2_loss: 4.6535 - val_output_acc: 0.3200 - val_auxilliary_output_1_acc: 0.7100 - val_auxilliary_output_2_acc: 0.5900\nEpoch 84/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.5029 - output_loss: 0.5000 - auxilliary_output_1_loss: 0.0060 - auxilliary_output_2_loss: 0.0037 - output_acc: 0.8050 - auxilliary_output_1_acc: 0.9975 - auxilliary_output_2_acc: 0.9988 - val_loss: 4.0418 - val_output_loss: 1.3465 - val_auxilliary_output_1_loss: 4.0482 - val_auxilliary_output_2_loss: 4.9359 - val_output_acc: 0.5100 - val_auxilliary_output_1_acc: 0.6900 - val_auxilliary_output_2_acc: 0.5600\nEpoch 85/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.4430 - output_loss: 0.4402 - auxilliary_output_1_loss: 0.0062 - auxilliary_output_2_loss: 0.0034 - output_acc: 0.8225 - auxilliary_output_1_acc: 0.9975 - auxilliary_output_2_acc: 0.9975 - val_loss: 4.2263 - val_output_loss: 1.5292 - val_auxilliary_output_1_loss: 4.1644 - val_auxilliary_output_2_loss: 4.8262 - val_output_acc: 0.3900 - val_auxilliary_output_1_acc: 0.6500 - val_auxilliary_output_2_acc: 0.6100\nEpoch 86/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.4012 - output_loss: 0.3999 - auxilliary_output_1_loss: 0.0025 - auxilliary_output_2_loss: 0.0020 - output_acc: 0.8250 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 4.0238 - val_output_loss: 1.3958 - val_auxilliary_output_1_loss: 4.1907 - val_auxilliary_output_2_loss: 4.5695 - val_output_acc: 0.4400 - val_auxilliary_output_1_acc: 0.6600 - val_auxilliary_output_2_acc: 0.6300\nEpoch 87/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.4101 - output_loss: 0.4086 - auxilliary_output_1_loss: 0.0029 - auxilliary_output_2_loss: 0.0019 - output_acc: 0.8150 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 4.2895 - val_output_loss: 1.5519 - val_auxilliary_output_1_loss: 4.2527 - val_auxilliary_output_2_loss: 4.8728 - val_output_acc: 0.4800 - val_auxilliary_output_1_acc: 0.6700 - val_auxilliary_output_2_acc: 0.6100\n\nEpoch 00087: ReduceLROnPlateau reducing learning rate to 0.0006871947087347508.\nEpoch 88/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.3696 - output_loss: 0.3680 - auxilliary_output_1_loss: 0.0032 - auxilliary_output_2_loss: 0.0021 - output_acc: 0.8375 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 4.2145 - val_output_loss: 1.5787 - val_auxilliary_output_1_loss: 4.3386 - val_auxilliary_output_2_loss: 4.4474 - val_output_acc: 0.3500 - val_auxilliary_output_1_acc: 0.6400 - val_auxilliary_output_2_acc: 0.6600\nEpoch 89/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.2963 - output_loss: 0.2950 - auxilliary_output_1_loss: 0.0025 - auxilliary_output_2_loss: 0.0018 - output_acc: 0.8738 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 4.7605 - val_output_loss: 1.9414 - val_auxilliary_output_1_loss: 4.3447 - val_auxilliary_output_2_loss: 5.0523 - val_output_acc: 0.4300 - val_auxilliary_output_1_acc: 0.6700 - val_auxilliary_output_2_acc: 0.6000\nEpoch 90/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.4245 - output_loss: 0.4226 - auxilliary_output_1_loss: 0.0044 - auxilliary_output_2_loss: 0.0021 - output_acc: 0.8362 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 4.8403 - val_output_loss: 2.0526 - val_auxilliary_output_1_loss: 4.4060 - val_auxilliary_output_2_loss: 4.8866 - val_output_acc: 0.3200 - val_auxilliary_output_1_acc: 0.6600 - val_auxilliary_output_2_acc: 0.6100\nEpoch 91/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.5177 - output_loss: 0.5163 - auxilliary_output_1_loss: 0.0022 - auxilliary_output_2_loss: 0.0024 - output_acc: 0.7913 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 4.0930 - val_output_loss: 1.3986 - val_auxilliary_output_1_loss: 4.3773 - val_auxilliary_output_2_loss: 4.6043 - val_output_acc: 0.4100 - val_auxilliary_output_1_acc: 0.6400 - val_auxilliary_output_2_acc: 0.6200\nEpoch 92/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.2680 - output_loss: 0.2667 - auxilliary_output_1_loss: 0.0025 - auxilliary_output_2_loss: 0.0018 - output_acc: 0.9062 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 4.5209 - val_output_loss: 1.7387 - val_auxilliary_output_1_loss: 4.4427 - val_auxilliary_output_2_loss: 4.8313 - val_output_acc: 0.4400 - val_auxilliary_output_1_acc: 0.6400 - val_auxilliary_output_2_acc: 0.6100\nEpoch 93/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.2444 - output_loss: 0.2416 - auxilliary_output_1_loss: 0.0027 - auxilliary_output_2_loss: 0.0066 - output_acc: 0.9050 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9975 - val_loss: 4.7083 - val_output_loss: 1.6437 - val_auxilliary_output_1_loss: 4.5417 - val_auxilliary_output_2_loss: 5.6736 - val_output_acc: 0.5600 - val_auxilliary_output_1_acc: 0.6300 - val_auxilliary_output_2_acc: 0.5400\nEpoch 94/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.2206 - output_loss: 0.2191 - auxilliary_output_1_loss: 0.0034 - auxilliary_output_2_loss: 0.0017 - output_acc: 0.9100 - auxilliary_output_1_acc: 0.9975 - auxilliary_output_2_acc: 0.9988 - val_loss: 5.1875 - val_output_loss: 2.0187 - val_auxilliary_output_1_loss: 4.5716 - val_auxilliary_output_2_loss: 5.9909 - val_output_acc: 0.4400 - val_auxilliary_output_1_acc: 0.6200 - val_auxilliary_output_2_acc: 0.5200\n\nEpoch 00094: ReduceLROnPlateau reducing learning rate to 0.0005497557576745749.\nEpoch 95/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.2035 - output_loss: 0.2021 - auxilliary_output_1_loss: 0.0024 - auxilliary_output_2_loss: 0.0021 - output_acc: 0.9175 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 4.7802 - val_output_loss: 1.6268 - val_auxilliary_output_1_loss: 4.6347 - val_auxilliary_output_2_loss: 5.8765 - val_output_acc: 0.5500 - val_auxilliary_output_1_acc: 0.6100 - val_auxilliary_output_2_acc: 0.5300\nEpoch 96/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.2293 - output_loss: 0.2273 - auxilliary_output_1_loss: 0.0022 - auxilliary_output_2_loss: 0.0045 - output_acc: 0.9175 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 5.0667 - val_output_loss: 2.0054 - val_auxilliary_output_1_loss: 4.5169 - val_auxilliary_output_2_loss: 5.6874 - val_output_acc: 0.5700 - val_auxilliary_output_1_acc: 0.6300 - val_auxilliary_output_2_acc: 0.5300\nEpoch 97/200\n","name":"stdout"},{"output_type":"stream","text":"800/800 [==============================] - 11s 14ms/step - loss: 0.3381 - output_loss: 0.3363 - auxilliary_output_1_loss: 0.0031 - auxilliary_output_2_loss: 0.0028 - output_acc: 0.8750 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9975 - val_loss: 4.5937 - val_output_loss: 1.5326 - val_auxilliary_output_1_loss: 4.5885 - val_auxilliary_output_2_loss: 5.6153 - val_output_acc: 0.3800 - val_auxilliary_output_1_acc: 0.6200 - val_auxilliary_output_2_acc: 0.5400\nEpoch 98/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.2443 - output_loss: 0.2430 - auxilliary_output_1_loss: 0.0025 - auxilliary_output_2_loss: 0.0020 - output_acc: 0.8987 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 5.1156 - val_output_loss: 2.0833 - val_auxilliary_output_1_loss: 4.5796 - val_auxilliary_output_2_loss: 5.5283 - val_output_acc: 0.5200 - val_auxilliary_output_1_acc: 0.6400 - val_auxilliary_output_2_acc: 0.5600\nEpoch 99/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.2044 - output_loss: 0.2032 - auxilliary_output_1_loss: 0.0023 - auxilliary_output_2_loss: 0.0018 - output_acc: 0.9225 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 4.7675 - val_output_loss: 1.7059 - val_auxilliary_output_1_loss: 4.5568 - val_auxilliary_output_2_loss: 5.6486 - val_output_acc: 0.5300 - val_auxilliary_output_1_acc: 0.6400 - val_auxilliary_output_2_acc: 0.5400\nEpoch 100/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.1303 - output_loss: 0.1276 - auxilliary_output_1_loss: 0.0022 - auxilliary_output_2_loss: 0.0068 - output_acc: 0.9575 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9975 - val_loss: 4.6832 - val_output_loss: 1.6023 - val_auxilliary_output_1_loss: 4.6930 - val_auxilliary_output_2_loss: 5.5767 - val_output_acc: 0.5500 - val_auxilliary_output_1_acc: 0.6300 - val_auxilliary_output_2_acc: 0.5600\nEpoch 101/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.1647 - output_loss: 0.1635 - auxilliary_output_1_loss: 0.0024 - auxilliary_output_2_loss: 0.0016 - output_acc: 0.9325 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 4.7678 - val_output_loss: 1.6575 - val_auxilliary_output_1_loss: 4.7113 - val_auxilliary_output_2_loss: 5.6563 - val_output_acc: 0.6100 - val_auxilliary_output_1_acc: 0.6200 - val_auxilliary_output_2_acc: 0.5500\n\nEpoch 00101: ReduceLROnPlateau reducing learning rate to 0.0004398046061396599.\nEpoch 102/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.2175 - output_loss: 0.2163 - auxilliary_output_1_loss: 0.0025 - auxilliary_output_2_loss: 0.0016 - output_acc: 0.9288 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 4.7337 - val_output_loss: 1.7265 - val_auxilliary_output_1_loss: 4.7327 - val_auxilliary_output_2_loss: 5.2916 - val_output_acc: 0.4300 - val_auxilliary_output_1_acc: 0.6000 - val_auxilliary_output_2_acc: 0.5900\nEpoch 103/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.1425 - output_loss: 0.1406 - auxilliary_output_1_loss: 0.0020 - auxilliary_output_2_loss: 0.0044 - output_acc: 0.9487 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9975 - val_loss: 5.4386 - val_output_loss: 1.7142 - val_auxilliary_output_1_loss: 4.7418 - val_auxilliary_output_2_loss: 7.6728 - val_output_acc: 0.6100 - val_auxilliary_output_1_acc: 0.6200 - val_auxilliary_output_2_acc: 0.4500\nEpoch 104/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.2262 - output_loss: 0.2232 - auxilliary_output_1_loss: 0.0033 - auxilliary_output_2_loss: 0.0066 - output_acc: 0.9075 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 4.6646 - val_output_loss: 1.7106 - val_auxilliary_output_1_loss: 4.7079 - val_auxilliary_output_2_loss: 5.1386 - val_output_acc: 0.3900 - val_auxilliary_output_1_acc: 0.6000 - val_auxilliary_output_2_acc: 0.6100\nEpoch 105/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.1334 - output_loss: 0.1317 - auxilliary_output_1_loss: 0.0025 - auxilliary_output_2_loss: 0.0032 - output_acc: 0.9550 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 4.5725 - val_output_loss: 1.6270 - val_auxilliary_output_1_loss: 4.7688 - val_auxilliary_output_2_loss: 5.0496 - val_output_acc: 0.5800 - val_auxilliary_output_1_acc: 0.6000 - val_auxilliary_output_2_acc: 0.6200\nEpoch 106/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.1115 - output_loss: 0.1100 - auxilliary_output_1_loss: 0.0029 - auxilliary_output_2_loss: 0.0022 - output_acc: 0.9637 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 4.8117 - val_output_loss: 1.7849 - val_auxilliary_output_1_loss: 4.9340 - val_auxilliary_output_2_loss: 5.1555 - val_output_acc: 0.5500 - val_auxilliary_output_1_acc: 0.6100 - val_auxilliary_output_2_acc: 0.6100\nEpoch 107/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0771 - output_loss: 0.0758 - auxilliary_output_1_loss: 0.0025 - auxilliary_output_2_loss: 0.0016 - output_acc: 0.9775 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 5.0521 - val_output_loss: 2.0678 - val_auxilliary_output_1_loss: 4.9627 - val_auxilliary_output_2_loss: 4.9848 - val_output_acc: 0.4700 - val_auxilliary_output_1_acc: 0.6100 - val_auxilliary_output_2_acc: 0.6200\nEpoch 108/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0635 - output_loss: 0.0612 - auxilliary_output_1_loss: 0.0030 - auxilliary_output_2_loss: 0.0043 - output_acc: 0.9875 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9975 - val_loss: 5.1316 - val_output_loss: 2.1144 - val_auxilliary_output_1_loss: 4.9066 - val_auxilliary_output_2_loss: 5.1508 - val_output_acc: 0.6000 - val_auxilliary_output_1_acc: 0.6200 - val_auxilliary_output_2_acc: 0.6000\n\nEpoch 00108: ReduceLROnPlateau reducing learning rate to 0.00035184368025511505.\nEpoch 109/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0591 - output_loss: 0.0579 - auxilliary_output_1_loss: 0.0024 - auxilliary_output_2_loss: 0.0015 - output_acc: 0.9825 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 5.1004 - val_output_loss: 2.0763 - val_auxilliary_output_1_loss: 4.8447 - val_auxilliary_output_2_loss: 5.2355 - val_output_acc: 0.5600 - val_auxilliary_output_1_acc: 0.6300 - val_auxilliary_output_2_acc: 0.6000\nEpoch 110/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0550 - output_loss: 0.0516 - auxilliary_output_1_loss: 0.0082 - auxilliary_output_2_loss: 0.0031 - output_acc: 0.9825 - auxilliary_output_1_acc: 0.9975 - auxilliary_output_2_acc: 0.9988 - val_loss: 5.7454 - val_output_loss: 2.6142 - val_auxilliary_output_1_loss: 4.9606 - val_auxilliary_output_2_loss: 5.4766 - val_output_acc: 0.4100 - val_auxilliary_output_1_acc: 0.6100 - val_auxilliary_output_2_acc: 0.5800\nEpoch 111/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.2440 - output_loss: 0.2428 - auxilliary_output_1_loss: 0.0024 - auxilliary_output_2_loss: 0.0014 - output_acc: 0.9125 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 5.7368 - val_output_loss: 2.7198 - val_auxilliary_output_1_loss: 4.7463 - val_auxilliary_output_2_loss: 5.3105 - val_output_acc: 0.4300 - val_auxilliary_output_1_acc: 0.6100 - val_auxilliary_output_2_acc: 0.5700\nEpoch 112/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.2607 - output_loss: 0.2597 - auxilliary_output_1_loss: 0.0020 - auxilliary_output_2_loss: 0.0014 - output_acc: 0.9075 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 4.7680 - val_output_loss: 1.6986 - val_auxilliary_output_1_loss: 4.9654 - val_auxilliary_output_2_loss: 5.2658 - val_output_acc: 0.4900 - val_auxilliary_output_1_acc: 0.6000 - val_auxilliary_output_2_acc: 0.5800\nEpoch 113/200\n","name":"stdout"},{"output_type":"stream","text":"800/800 [==============================] - 11s 13ms/step - loss: 0.1146 - output_loss: 0.1133 - auxilliary_output_1_loss: 0.0026 - auxilliary_output_2_loss: 0.0018 - output_acc: 0.9625 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 4.8465 - val_output_loss: 1.7499 - val_auxilliary_output_1_loss: 5.0340 - val_auxilliary_output_2_loss: 5.2882 - val_output_acc: 0.6500 - val_auxilliary_output_1_acc: 0.6100 - val_auxilliary_output_2_acc: 0.5900\nEpoch 114/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0591 - output_loss: 0.0561 - auxilliary_output_1_loss: 0.0023 - auxilliary_output_2_loss: 0.0078 - output_acc: 0.9888 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 5.0670 - val_output_loss: 1.8352 - val_auxilliary_output_1_loss: 5.0075 - val_auxilliary_output_2_loss: 5.7652 - val_output_acc: 0.5800 - val_auxilliary_output_1_acc: 0.6100 - val_auxilliary_output_2_acc: 0.5600\nEpoch 115/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0548 - output_loss: 0.0537 - auxilliary_output_1_loss: 0.0020 - auxilliary_output_2_loss: 0.0016 - output_acc: 0.9850 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 5.2121 - val_output_loss: 1.9136 - val_auxilliary_output_1_loss: 5.0899 - val_auxilliary_output_2_loss: 5.9051 - val_output_acc: 0.5700 - val_auxilliary_output_1_acc: 0.6100 - val_auxilliary_output_2_acc: 0.5500\n\nEpoch 00115: ReduceLROnPlateau reducing learning rate to 0.0002814749488607049.\nEpoch 116/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0326 - output_loss: 0.0316 - auxilliary_output_1_loss: 0.0020 - auxilliary_output_2_loss: 0.0016 - output_acc: 0.9937 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 5.2159 - val_output_loss: 1.8867 - val_auxilliary_output_1_loss: 5.1510 - val_auxilliary_output_2_loss: 5.9464 - val_output_acc: 0.5800 - val_auxilliary_output_1_acc: 0.6000 - val_auxilliary_output_2_acc: 0.5500\nEpoch 117/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0283 - output_loss: 0.0273 - auxilliary_output_1_loss: 0.0022 - auxilliary_output_2_loss: 0.0013 - output_acc: 0.9937 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 5.5580 - val_output_loss: 2.2372 - val_auxilliary_output_1_loss: 5.1094 - val_auxilliary_output_2_loss: 5.9601 - val_output_acc: 0.5200 - val_auxilliary_output_1_acc: 0.6100 - val_auxilliary_output_2_acc: 0.5500\nEpoch 118/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0252 - output_loss: 0.0238 - auxilliary_output_1_loss: 0.0027 - auxilliary_output_2_loss: 0.0019 - output_acc: 0.9950 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 5.4924 - val_output_loss: 2.1703 - val_auxilliary_output_1_loss: 5.0883 - val_auxilliary_output_2_loss: 5.9853 - val_output_acc: 0.6100 - val_auxilliary_output_1_acc: 0.6100 - val_auxilliary_output_2_acc: 0.5500\nEpoch 119/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0181 - output_loss: 0.0170 - auxilliary_output_1_loss: 0.0022 - auxilliary_output_2_loss: 0.0015 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 5.5228 - val_output_loss: 2.1915 - val_auxilliary_output_1_loss: 5.1032 - val_auxilliary_output_2_loss: 6.0011 - val_output_acc: 0.6200 - val_auxilliary_output_1_acc: 0.6000 - val_auxilliary_output_2_acc: 0.5400\nEpoch 120/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0170 - output_loss: 0.0157 - auxilliary_output_1_loss: 0.0026 - auxilliary_output_2_loss: 0.0017 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 5.6364 - val_output_loss: 2.2904 - val_auxilliary_output_1_loss: 5.1487 - val_auxilliary_output_2_loss: 6.0044 - val_output_acc: 0.5700 - val_auxilliary_output_1_acc: 0.6000 - val_auxilliary_output_2_acc: 0.5500\nEpoch 121/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0216 - output_loss: 0.0202 - auxilliary_output_1_loss: 0.0023 - auxilliary_output_2_loss: 0.0023 - output_acc: 0.9962 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 5.8213 - val_output_loss: 2.4554 - val_auxilliary_output_1_loss: 5.1469 - val_auxilliary_output_2_loss: 6.0730 - val_output_acc: 0.5500 - val_auxilliary_output_1_acc: 0.6000 - val_auxilliary_output_2_acc: 0.5500\nEpoch 122/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0212 - output_loss: 0.0199 - auxilliary_output_1_loss: 0.0026 - auxilliary_output_2_loss: 0.0017 - output_acc: 0.9962 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 5.7495 - val_output_loss: 2.3752 - val_auxilliary_output_1_loss: 5.1382 - val_auxilliary_output_2_loss: 6.1095 - val_output_acc: 0.5500 - val_auxilliary_output_1_acc: 0.6000 - val_auxilliary_output_2_acc: 0.5500\n\nEpoch 00122: ReduceLROnPlateau reducing learning rate to 0.0002251799684017897.\nEpoch 123/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0280 - output_loss: 0.0267 - auxilliary_output_1_loss: 0.0027 - auxilliary_output_2_loss: 0.0015 - output_acc: 0.9950 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 5.7476 - val_output_loss: 2.3634 - val_auxilliary_output_1_loss: 5.1607 - val_auxilliary_output_2_loss: 6.1200 - val_output_acc: 0.5800 - val_auxilliary_output_1_acc: 0.6000 - val_auxilliary_output_2_acc: 0.5500\nEpoch 124/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0293 - output_loss: 0.0279 - auxilliary_output_1_loss: 0.0026 - auxilliary_output_2_loss: 0.0020 - output_acc: 0.9900 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 6.2482 - val_output_loss: 2.8453 - val_auxilliary_output_1_loss: 5.2112 - val_auxilliary_output_2_loss: 6.1319 - val_output_acc: 0.4600 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5500\nEpoch 125/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0205 - output_loss: 0.0189 - auxilliary_output_1_loss: 0.0034 - auxilliary_output_2_loss: 0.0018 - output_acc: 0.9975 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 5.8721 - val_output_loss: 2.4641 - val_auxilliary_output_1_loss: 5.1945 - val_auxilliary_output_2_loss: 6.1656 - val_output_acc: 0.5400 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5500\nEpoch 126/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0143 - output_loss: 0.0132 - auxilliary_output_1_loss: 0.0022 - auxilliary_output_2_loss: 0.0015 - output_acc: 0.9975 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 5.9021 - val_output_loss: 2.4914 - val_auxilliary_output_1_loss: 5.2106 - val_auxilliary_output_2_loss: 6.1584 - val_output_acc: 0.5500 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5500\nEpoch 127/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0106 - output_loss: 0.0095 - auxilliary_output_1_loss: 0.0020 - auxilliary_output_2_loss: 0.0016 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 5.9640 - val_output_loss: 2.5502 - val_auxilliary_output_1_loss: 5.2218 - val_auxilliary_output_2_loss: 6.1578 - val_output_acc: 0.5600 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5500\nEpoch 128/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0122 - output_loss: 0.0111 - auxilliary_output_1_loss: 0.0024 - auxilliary_output_2_loss: 0.0015 - output_acc: 0.9975 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 5.9240 - val_output_loss: 2.5066 - val_auxilliary_output_1_loss: 5.2322 - val_auxilliary_output_2_loss: 6.1591 - val_output_acc: 0.5900 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5500\nEpoch 129/200\n","name":"stdout"},{"output_type":"stream","text":"800/800 [==============================] - 11s 14ms/step - loss: 0.0097 - output_loss: 0.0083 - auxilliary_output_1_loss: 0.0028 - auxilliary_output_2_loss: 0.0018 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 5.9418 - val_output_loss: 2.5307 - val_auxilliary_output_1_loss: 5.2174 - val_auxilliary_output_2_loss: 6.1530 - val_output_acc: 0.6000 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5500\n\nEpoch 00129: ReduceLROnPlateau reducing learning rate to 0.0001801439793780446.\nEpoch 130/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0127 - output_loss: 0.0096 - auxilliary_output_1_loss: 0.0025 - auxilliary_output_2_loss: 0.0081 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 6.0096 - val_output_loss: 2.6450 - val_auxilliary_output_1_loss: 5.2114 - val_auxilliary_output_2_loss: 6.0040 - val_output_acc: 0.5700 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5600\nEpoch 131/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0115 - output_loss: 0.0096 - auxilliary_output_1_loss: 0.0046 - auxilliary_output_2_loss: 0.0015 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9975 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.0692 - val_output_loss: 2.6983 - val_auxilliary_output_1_loss: 5.2533 - val_auxilliary_output_2_loss: 5.9829 - val_output_acc: 0.5600 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5600\nEpoch 132/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0105 - output_loss: 0.0094 - auxilliary_output_1_loss: 0.0022 - auxilliary_output_2_loss: 0.0017 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.1585 - val_output_loss: 2.7541 - val_auxilliary_output_1_loss: 5.3559 - val_auxilliary_output_2_loss: 5.9922 - val_output_acc: 0.5600 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5600\nEpoch 133/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0104 - output_loss: 0.0084 - auxilliary_output_1_loss: 0.0049 - auxilliary_output_2_loss: 0.0017 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9975 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.1459 - val_output_loss: 2.7566 - val_auxilliary_output_1_loss: 5.3053 - val_auxilliary_output_2_loss: 5.9927 - val_output_acc: 0.5600 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5600\nEpoch 134/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0107 - output_loss: 0.0083 - auxilliary_output_1_loss: 0.0044 - auxilliary_output_2_loss: 0.0036 - output_acc: 0.9975 - auxilliary_output_1_acc: 0.9975 - auxilliary_output_2_acc: 0.9988 - val_loss: 6.2601 - val_output_loss: 2.8632 - val_auxilliary_output_1_loss: 5.2919 - val_auxilliary_output_2_loss: 6.0312 - val_output_acc: 0.5300 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5500\nEpoch 135/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0099 - output_loss: 0.0088 - auxilliary_output_1_loss: 0.0017 - auxilliary_output_2_loss: 0.0022 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.0922 - val_output_loss: 2.7463 - val_auxilliary_output_1_loss: 5.3303 - val_auxilliary_output_2_loss: 5.8229 - val_output_acc: 0.5600 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5600\nEpoch 136/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0089 - output_loss: 0.0078 - auxilliary_output_1_loss: 0.0019 - auxilliary_output_2_loss: 0.0018 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.2344 - val_output_loss: 2.8997 - val_auxilliary_output_1_loss: 5.3381 - val_auxilliary_output_2_loss: 5.7776 - val_output_acc: 0.5200 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5600\n\nEpoch 00136: ReduceLROnPlateau reducing learning rate to 0.00014411518350243568.\nEpoch 137/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0078 - output_loss: 0.0067 - auxilliary_output_1_loss: 0.0024 - auxilliary_output_2_loss: 0.0014 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.1221 - val_output_loss: 2.7888 - val_auxilliary_output_1_loss: 5.3433 - val_auxilliary_output_2_loss: 5.7676 - val_output_acc: 0.5500 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5500\nEpoch 138/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0082 - output_loss: 0.0070 - auxilliary_output_1_loss: 0.0024 - auxilliary_output_2_loss: 0.0017 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.1325 - val_output_loss: 2.7982 - val_auxilliary_output_1_loss: 5.3422 - val_auxilliary_output_2_loss: 5.7721 - val_output_acc: 0.5600 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5600\nEpoch 139/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0082 - output_loss: 0.0070 - auxilliary_output_1_loss: 0.0022 - auxilliary_output_2_loss: 0.0018 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.1802 - val_output_loss: 2.8443 - val_auxilliary_output_1_loss: 5.3383 - val_auxilliary_output_2_loss: 5.7814 - val_output_acc: 0.5600 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5600\nEpoch 140/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0081 - output_loss: 0.0068 - auxilliary_output_1_loss: 0.0030 - auxilliary_output_2_loss: 0.0015 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.1970 - val_output_loss: 2.8578 - val_auxilliary_output_1_loss: 5.3489 - val_auxilliary_output_2_loss: 5.7820 - val_output_acc: 0.5600 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5600\nEpoch 141/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0077 - output_loss: 0.0066 - auxilliary_output_1_loss: 0.0020 - auxilliary_output_2_loss: 0.0015 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.1765 - val_output_loss: 2.8363 - val_auxilliary_output_1_loss: 5.3621 - val_auxilliary_output_2_loss: 5.7718 - val_output_acc: 0.5700 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5500\nEpoch 142/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0081 - output_loss: 0.0070 - auxilliary_output_1_loss: 0.0022 - auxilliary_output_2_loss: 0.0016 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.1202 - val_output_loss: 2.7802 - val_auxilliary_output_1_loss: 5.3423 - val_auxilliary_output_2_loss: 5.7909 - val_output_acc: 0.6100 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5600\nEpoch 143/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0079 - output_loss: 0.0066 - auxilliary_output_1_loss: 0.0024 - auxilliary_output_2_loss: 0.0020 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.3778 - val_output_loss: 3.0353 - val_auxilliary_output_1_loss: 5.3603 - val_auxilliary_output_2_loss: 5.7815 - val_output_acc: 0.5200 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5500\n\nEpoch 00143: ReduceLROnPlateau reducing learning rate to 0.00011529214680194855.\nEpoch 144/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0076 - output_loss: 0.0062 - auxilliary_output_1_loss: 0.0024 - auxilliary_output_2_loss: 0.0024 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.2063 - val_output_loss: 2.8568 - val_auxilliary_output_1_loss: 5.3630 - val_auxilliary_output_2_loss: 5.8020 - val_output_acc: 0.5700 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5500\nEpoch 145/200\n","name":"stdout"},{"output_type":"stream","text":"800/800 [==============================] - 11s 13ms/step - loss: 0.0069 - output_loss: 0.0059 - auxilliary_output_1_loss: 0.0019 - auxilliary_output_2_loss: 0.0017 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.2678 - val_output_loss: 2.9178 - val_auxilliary_output_1_loss: 5.3571 - val_auxilliary_output_2_loss: 5.8094 - val_output_acc: 0.5600 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5600\nEpoch 146/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0070 - output_loss: 0.0056 - auxilliary_output_1_loss: 0.0027 - auxilliary_output_2_loss: 0.0019 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 6.2448 - val_output_loss: 2.9014 - val_auxilliary_output_1_loss: 5.3442 - val_auxilliary_output_2_loss: 5.8004 - val_output_acc: 0.5700 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5500\nEpoch 147/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0069 - output_loss: 0.0055 - auxilliary_output_1_loss: 0.0026 - auxilliary_output_2_loss: 0.0019 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.2978 - val_output_loss: 2.9543 - val_auxilliary_output_1_loss: 5.3363 - val_auxilliary_output_2_loss: 5.8084 - val_output_acc: 0.5600 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5600\nEpoch 148/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0092 - output_loss: 0.0081 - auxilliary_output_1_loss: 0.0021 - auxilliary_output_2_loss: 0.0015 - output_acc: 0.9975 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.4443 - val_output_loss: 3.1014 - val_auxilliary_output_1_loss: 5.3259 - val_auxilliary_output_2_loss: 5.8172 - val_output_acc: 0.4900 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5600\nEpoch 149/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0076 - output_loss: 0.0065 - auxilliary_output_1_loss: 0.0020 - auxilliary_output_2_loss: 0.0017 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.1359 - val_output_loss: 2.7887 - val_auxilliary_output_1_loss: 5.3293 - val_auxilliary_output_2_loss: 5.8280 - val_output_acc: 0.5900 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5600\nEpoch 150/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0103 - output_loss: 0.0082 - auxilliary_output_1_loss: 0.0033 - auxilliary_output_2_loss: 0.0037 - output_acc: 0.9975 - auxilliary_output_1_acc: 0.9975 - auxilliary_output_2_acc: 0.9988 - val_loss: 6.2486 - val_output_loss: 2.9239 - val_auxilliary_output_1_loss: 5.2734 - val_auxilliary_output_2_loss: 5.8087 - val_output_acc: 0.5700 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5600\n\nEpoch 00150: ReduceLROnPlateau reducing learning rate to 0.0001.\nEpoch 151/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0085 - output_loss: 0.0074 - auxilliary_output_1_loss: 0.0021 - auxilliary_output_2_loss: 0.0015 - output_acc: 0.9975 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.2355 - val_output_loss: 2.9250 - val_auxilliary_output_1_loss: 5.2251 - val_auxilliary_output_2_loss: 5.8100 - val_output_acc: 0.5700 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5600\nEpoch 152/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0072 - output_loss: 0.0060 - auxilliary_output_1_loss: 0.0024 - auxilliary_output_2_loss: 0.0016 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.2187 - val_output_loss: 2.9073 - val_auxilliary_output_1_loss: 5.2319 - val_auxilliary_output_2_loss: 5.8063 - val_output_acc: 0.5800 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5500\nEpoch 153/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0070 - output_loss: 0.0059 - auxilliary_output_1_loss: 0.0022 - auxilliary_output_2_loss: 0.0017 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.2812 - val_output_loss: 2.9691 - val_auxilliary_output_1_loss: 5.2246 - val_auxilliary_output_2_loss: 5.8160 - val_output_acc: 0.5600 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5500\nEpoch 154/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0098 - output_loss: 0.0083 - auxilliary_output_1_loss: 0.0035 - auxilliary_output_2_loss: 0.0018 - output_acc: 0.9975 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.2648 - val_output_loss: 2.9518 - val_auxilliary_output_1_loss: 5.2446 - val_auxilliary_output_2_loss: 5.7987 - val_output_acc: 0.5700 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5500\nEpoch 155/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0065 - output_loss: 0.0052 - auxilliary_output_1_loss: 0.0030 - auxilliary_output_2_loss: 0.0015 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.1815 - val_output_loss: 2.8681 - val_auxilliary_output_1_loss: 5.2458 - val_auxilliary_output_2_loss: 5.7990 - val_output_acc: 0.5800 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5500\nEpoch 156/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0070 - output_loss: 0.0059 - auxilliary_output_1_loss: 0.0024 - auxilliary_output_2_loss: 0.0014 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.2409 - val_output_loss: 2.9282 - val_auxilliary_output_1_loss: 5.2480 - val_auxilliary_output_2_loss: 5.7944 - val_output_acc: 0.5700 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5500\nEpoch 157/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0071 - output_loss: 0.0058 - auxilliary_output_1_loss: 0.0029 - auxilliary_output_2_loss: 0.0015 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.2320 - val_output_loss: 2.9092 - val_auxilliary_output_1_loss: 5.2741 - val_auxilliary_output_2_loss: 5.8019 - val_output_acc: 0.5700 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5500\nEpoch 158/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0068 - output_loss: 0.0055 - auxilliary_output_1_loss: 0.0027 - auxilliary_output_2_loss: 0.0017 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.2429 - val_output_loss: 2.9163 - val_auxilliary_output_1_loss: 5.2855 - val_auxilliary_output_2_loss: 5.8032 - val_output_acc: 0.5600 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5500\nEpoch 159/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0136 - output_loss: 0.0050 - auxilliary_output_1_loss: 0.0220 - auxilliary_output_2_loss: 0.0066 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9975 - auxilliary_output_2_acc: 0.9988 - val_loss: 6.3854 - val_output_loss: 3.0420 - val_auxilliary_output_1_loss: 5.2780 - val_auxilliary_output_2_loss: 5.8666 - val_output_acc: 0.5400 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5600\nEpoch 160/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0065 - output_loss: 0.0054 - auxilliary_output_1_loss: 0.0023 - auxilliary_output_2_loss: 0.0016 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.2288 - val_output_loss: 2.8716 - val_auxilliary_output_1_loss: 5.2865 - val_auxilliary_output_2_loss: 5.9042 - val_output_acc: 0.6000 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5600\nEpoch 161/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0076 - output_loss: 0.0049 - auxilliary_output_1_loss: 0.0073 - auxilliary_output_2_loss: 0.0017 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9975 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.3375 - val_output_loss: 2.9750 - val_auxilliary_output_1_loss: 5.3086 - val_auxilliary_output_2_loss: 5.8999 - val_output_acc: 0.5600 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5600\n","name":"stdout"},{"output_type":"stream","text":"Epoch 162/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0072 - output_loss: 0.0054 - auxilliary_output_1_loss: 0.0028 - auxilliary_output_2_loss: 0.0032 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 6.3656 - val_output_loss: 2.9907 - val_auxilliary_output_1_loss: 5.3176 - val_auxilliary_output_2_loss: 5.9322 - val_output_acc: 0.5600 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5600\nEpoch 163/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0081 - output_loss: 0.0057 - auxilliary_output_1_loss: 0.0018 - auxilliary_output_2_loss: 0.0064 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 6.2506 - val_output_loss: 2.8526 - val_auxilliary_output_1_loss: 5.3225 - val_auxilliary_output_2_loss: 6.0039 - val_output_acc: 0.5900 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5500\nEpoch 164/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0073 - output_loss: 0.0055 - auxilliary_output_1_loss: 0.0020 - auxilliary_output_2_loss: 0.0041 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 6.5662 - val_output_loss: 3.1628 - val_auxilliary_output_1_loss: 5.3215 - val_auxilliary_output_2_loss: 6.0232 - val_output_acc: 0.5400 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5500\nEpoch 165/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0066 - output_loss: 0.0054 - auxilliary_output_1_loss: 0.0022 - auxilliary_output_2_loss: 0.0016 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.4707 - val_output_loss: 3.0518 - val_auxilliary_output_1_loss: 5.3197 - val_auxilliary_output_2_loss: 6.0767 - val_output_acc: 0.5600 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5500\nEpoch 166/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0081 - output_loss: 0.0063 - auxilliary_output_1_loss: 0.0044 - auxilliary_output_2_loss: 0.0017 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9975 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.4369 - val_output_loss: 3.0227 - val_auxilliary_output_1_loss: 5.2975 - val_auxilliary_output_2_loss: 6.0831 - val_output_acc: 0.5600 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5500\nEpoch 167/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0057 - output_loss: 0.0044 - auxilliary_output_1_loss: 0.0026 - auxilliary_output_2_loss: 0.0016 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.3707 - val_output_loss: 2.9579 - val_auxilliary_output_1_loss: 5.2931 - val_auxilliary_output_2_loss: 6.0832 - val_output_acc: 0.5800 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5500\nEpoch 168/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0064 - output_loss: 0.0048 - auxilliary_output_1_loss: 0.0038 - auxilliary_output_2_loss: 0.0016 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9975 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.4287 - val_output_loss: 3.0128 - val_auxilliary_output_1_loss: 5.3018 - val_auxilliary_output_2_loss: 6.0845 - val_output_acc: 0.5600 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5500\nEpoch 169/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0066 - output_loss: 0.0052 - auxilliary_output_1_loss: 0.0032 - auxilliary_output_2_loss: 0.0015 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.4395 - val_output_loss: 3.0249 - val_auxilliary_output_1_loss: 5.2972 - val_auxilliary_output_2_loss: 6.0849 - val_output_acc: 0.5700 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5500\nEpoch 170/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0082 - output_loss: 0.0071 - auxilliary_output_1_loss: 0.0023 - auxilliary_output_2_loss: 0.0015 - output_acc: 0.9975 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.5281 - val_output_loss: 3.1347 - val_auxilliary_output_1_loss: 5.2286 - val_auxilliary_output_2_loss: 6.0826 - val_output_acc: 0.5500 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5500\nEpoch 171/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0071 - output_loss: 0.0060 - auxilliary_output_1_loss: 0.0020 - auxilliary_output_2_loss: 0.0016 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.4124 - val_output_loss: 3.0198 - val_auxilliary_output_1_loss: 5.2312 - val_auxilliary_output_2_loss: 6.0775 - val_output_acc: 0.5700 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5500\nEpoch 172/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0066 - output_loss: 0.0053 - auxilliary_output_1_loss: 0.0028 - auxilliary_output_2_loss: 0.0015 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.4626 - val_output_loss: 3.0667 - val_auxilliary_output_1_loss: 5.2370 - val_auxilliary_output_2_loss: 6.0826 - val_output_acc: 0.5600 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5500\nEpoch 173/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0069 - output_loss: 0.0058 - auxilliary_output_1_loss: 0.0021 - auxilliary_output_2_loss: 0.0015 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.4633 - val_output_loss: 3.0626 - val_auxilliary_output_1_loss: 5.2484 - val_auxilliary_output_2_loss: 6.0872 - val_output_acc: 0.5600 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5500\nEpoch 174/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0080 - output_loss: 0.0055 - auxilliary_output_1_loss: 0.0068 - auxilliary_output_2_loss: 0.0015 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9975 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.4381 - val_output_loss: 3.0266 - val_auxilliary_output_1_loss: 5.2845 - val_auxilliary_output_2_loss: 6.0871 - val_output_acc: 0.5800 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5500\nEpoch 175/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0067 - output_loss: 0.0056 - auxilliary_output_1_loss: 0.0022 - auxilliary_output_2_loss: 0.0016 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.4451 - val_output_loss: 3.0218 - val_auxilliary_output_1_loss: 5.3243 - val_auxilliary_output_2_loss: 6.0867 - val_output_acc: 0.5800 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5500\nEpoch 176/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0058 - output_loss: 0.0043 - auxilliary_output_1_loss: 0.0019 - auxilliary_output_2_loss: 0.0031 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 6.5701 - val_output_loss: 3.1130 - val_auxilliary_output_1_loss: 5.3234 - val_auxilliary_output_2_loss: 6.2002 - val_output_acc: 0.5600 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5500\nEpoch 177/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0062 - output_loss: 0.0048 - auxilliary_output_1_loss: 0.0029 - auxilliary_output_2_loss: 0.0016 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.5906 - val_output_loss: 3.1293 - val_auxilliary_output_1_loss: 5.3257 - val_auxilliary_output_2_loss: 6.2118 - val_output_acc: 0.5600 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5500\nEpoch 178/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0069 - output_loss: 0.0059 - auxilliary_output_1_loss: 0.0018 - auxilliary_output_2_loss: 0.0014 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.4886 - val_output_loss: 3.0277 - val_auxilliary_output_1_loss: 5.3238 - val_auxilliary_output_2_loss: 6.2126 - val_output_acc: 0.5900 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5500\n","name":"stdout"},{"output_type":"stream","text":"Epoch 179/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0076 - output_loss: 0.0062 - auxilliary_output_1_loss: 0.0028 - auxilliary_output_2_loss: 0.0017 - output_acc: 0.9975 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.6155 - val_output_loss: 3.1573 - val_auxilliary_output_1_loss: 5.3105 - val_auxilliary_output_2_loss: 6.2168 - val_output_acc: 0.5400 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5500\nEpoch 180/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0065 - output_loss: 0.0050 - auxilliary_output_1_loss: 0.0022 - auxilliary_output_2_loss: 0.0028 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 6.4344 - val_output_loss: 2.9735 - val_auxilliary_output_1_loss: 5.3245 - val_auxilliary_output_2_loss: 6.2121 - val_output_acc: 0.6000 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5500\nEpoch 181/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0059 - output_loss: 0.0047 - auxilliary_output_1_loss: 0.0022 - auxilliary_output_2_loss: 0.0019 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.5869 - val_output_loss: 3.1260 - val_auxilliary_output_1_loss: 5.3297 - val_auxilliary_output_2_loss: 6.2069 - val_output_acc: 0.5700 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5500\nEpoch 182/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0059 - output_loss: 0.0047 - auxilliary_output_1_loss: 0.0025 - auxilliary_output_2_loss: 0.0017 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.5005 - val_output_loss: 3.0388 - val_auxilliary_output_1_loss: 5.3358 - val_auxilliary_output_2_loss: 6.2032 - val_output_acc: 0.5800 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5500\nEpoch 183/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0060 - output_loss: 0.0046 - auxilliary_output_1_loss: 0.0031 - auxilliary_output_2_loss: 0.0016 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9975 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.5043 - val_output_loss: 3.0413 - val_auxilliary_output_1_loss: 5.3424 - val_auxilliary_output_2_loss: 6.2010 - val_output_acc: 0.5900 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5500\nEpoch 184/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0060 - output_loss: 0.0039 - auxilliary_output_1_loss: 0.0028 - auxilliary_output_2_loss: 0.0042 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 6.6113 - val_output_loss: 3.1474 - val_auxilliary_output_1_loss: 5.3494 - val_auxilliary_output_2_loss: 6.1967 - val_output_acc: 0.5600 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5600\nEpoch 185/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0071 - output_loss: 0.0058 - auxilliary_output_1_loss: 0.0027 - auxilliary_output_2_loss: 0.0017 - output_acc: 0.9975 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.6340 - val_output_loss: 3.1663 - val_auxilliary_output_1_loss: 5.3570 - val_auxilliary_output_2_loss: 6.2019 - val_output_acc: 0.5700 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5600\nEpoch 186/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0057 - output_loss: 0.0042 - auxilliary_output_1_loss: 0.0032 - auxilliary_output_2_loss: 0.0017 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9975 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.5098 - val_output_loss: 3.0744 - val_auxilliary_output_1_loss: 5.2501 - val_auxilliary_output_2_loss: 6.2013 - val_output_acc: 0.5800 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5600\nEpoch 187/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0063 - output_loss: 0.0051 - auxilliary_output_1_loss: 0.0024 - auxilliary_output_2_loss: 0.0017 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.4667 - val_output_loss: 3.0352 - val_auxilliary_output_1_loss: 5.2402 - val_auxilliary_output_2_loss: 6.1980 - val_output_acc: 0.6100 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5600\nEpoch 188/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0060 - output_loss: 0.0049 - auxilliary_output_1_loss: 0.0023 - auxilliary_output_2_loss: 0.0014 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.6526 - val_output_loss: 3.2220 - val_auxilliary_output_1_loss: 5.2428 - val_auxilliary_output_2_loss: 6.1927 - val_output_acc: 0.5500 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5600\nEpoch 189/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0061 - output_loss: 0.0051 - auxilliary_output_1_loss: 0.0020 - auxilliary_output_2_loss: 0.0012 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.4670 - val_output_loss: 3.0363 - val_auxilliary_output_1_loss: 5.2474 - val_auxilliary_output_2_loss: 6.1881 - val_output_acc: 0.6000 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5600\nEpoch 190/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0049 - output_loss: 0.0038 - auxilliary_output_1_loss: 0.0021 - auxilliary_output_2_loss: 0.0016 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.5406 - val_output_loss: 3.1105 - val_auxilliary_output_1_loss: 5.2485 - val_auxilliary_output_2_loss: 6.1854 - val_output_acc: 0.5700 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5600\nEpoch 191/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0060 - output_loss: 0.0046 - auxilliary_output_1_loss: 0.0023 - auxilliary_output_2_loss: 0.0022 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.5497 - val_output_loss: 3.1240 - val_auxilliary_output_1_loss: 5.2466 - val_auxilliary_output_2_loss: 6.1724 - val_output_acc: 0.5700 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5600\nEpoch 192/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0052 - output_loss: 0.0041 - auxilliary_output_1_loss: 0.0022 - auxilliary_output_2_loss: 0.0017 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.4687 - val_output_loss: 3.0425 - val_auxilliary_output_1_loss: 5.2462 - val_auxilliary_output_2_loss: 6.1745 - val_output_acc: 0.6100 - val_auxilliary_output_1_acc: 0.5900 - val_auxilliary_output_2_acc: 0.5600\nEpoch 193/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0057 - output_loss: 0.0046 - auxilliary_output_1_loss: 0.0021 - auxilliary_output_2_loss: 0.0017 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.5741 - val_output_loss: 3.1432 - val_auxilliary_output_1_loss: 5.2673 - val_auxilliary_output_2_loss: 6.1691 - val_output_acc: 0.5600 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5600\nEpoch 194/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0061 - output_loss: 0.0049 - auxilliary_output_1_loss: 0.0025 - auxilliary_output_2_loss: 0.0014 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.5604 - val_output_loss: 3.1302 - val_auxilliary_output_1_loss: 5.2720 - val_auxilliary_output_2_loss: 6.1620 - val_output_acc: 0.5700 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5600\nEpoch 195/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0062 - output_loss: 0.0051 - auxilliary_output_1_loss: 0.0020 - auxilliary_output_2_loss: 0.0018 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.4991 - val_output_loss: 3.0676 - val_auxilliary_output_1_loss: 5.2700 - val_auxilliary_output_2_loss: 6.1684 - val_output_acc: 0.6000 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5600\n","name":"stdout"},{"output_type":"stream","text":"Epoch 196/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0058 - output_loss: 0.0046 - auxilliary_output_1_loss: 0.0027 - auxilliary_output_2_loss: 0.0015 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.5208 - val_output_loss: 3.0873 - val_auxilliary_output_1_loss: 5.2726 - val_auxilliary_output_2_loss: 6.1725 - val_output_acc: 0.5800 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5600\nEpoch 197/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0051 - output_loss: 0.0040 - auxilliary_output_1_loss: 0.0021 - auxilliary_output_2_loss: 0.0015 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.6241 - val_output_loss: 3.1883 - val_auxilliary_output_1_loss: 5.2756 - val_auxilliary_output_2_loss: 6.1770 - val_output_acc: 0.5700 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5600\nEpoch 198/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0064 - output_loss: 0.0044 - auxilliary_output_1_loss: 0.0022 - auxilliary_output_2_loss: 0.0045 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 0.9988 - val_loss: 6.5142 - val_output_loss: 3.0677 - val_auxilliary_output_1_loss: 5.2684 - val_auxilliary_output_2_loss: 6.2200 - val_output_acc: 0.6000 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5600\nEpoch 199/200\n800/800 [==============================] - 11s 14ms/step - loss: 0.0058 - output_loss: 0.0047 - auxilliary_output_1_loss: 0.0021 - auxilliary_output_2_loss: 0.0016 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.5507 - val_output_loss: 3.1045 - val_auxilliary_output_1_loss: 5.2679 - val_auxilliary_output_2_loss: 6.2195 - val_output_acc: 0.5900 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5600\nEpoch 200/200\n800/800 [==============================] - 11s 13ms/step - loss: 0.0061 - output_loss: 0.0050 - auxilliary_output_1_loss: 0.0023 - auxilliary_output_2_loss: 0.0015 - output_acc: 0.9988 - auxilliary_output_1_acc: 0.9988 - auxilliary_output_2_acc: 1.0000 - val_loss: 6.6698 - val_output_loss: 3.2219 - val_auxilliary_output_1_loss: 5.2714 - val_auxilliary_output_2_loss: 6.2217 - val_output_acc: 0.5600 - val_auxilliary_output_1_acc: 0.5800 - val_auxilliary_output_2_acc: 0.5600\n","name":"stdout"}]},{"metadata":{"trusted":true},"cell_type":"code","source":"import matplotlib.pyplot as pyplot\npyplot.subplot(211)\npyplot.title('Cross-Entropy Loss', pad=-40)\npyplot.plot(history.history['loss'], label='train')\npyplot.plot(history.history['val_loss'], label='test')\npyplot.legend()\n# plot accuracy learning curves\npyplot.subplot(212)\npyplot.title('Accuracy', pad=-40)\npyplot.plot(history.history['output_acc'], label='train')\npyplot.plot(history.history['val_output_acc'], label='test')\npyplot.legend()\npyplot.show()","execution_count":15,"outputs":[{"output_type":"display_data","data":{"text/plain":"
","image/png":"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\n"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"pyplot.savefig(\"trainmetrics.png\")","execution_count":23,"outputs":[{"output_type":"display_data","data":{"text/plain":"
"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"from sklearn.metrics import accuracy_score, classification_report\npred_Y = model.predict(X_test_resh, batch_size = 32, verbose = True)\n\npred_Y_cat = np.argmax(pred_Y, -1)\n#print(pred_Y_cat[0])\n\ntest_Y_cat = np.argmax(Y_test, -1)\nprint('Accuracy on Test Data: %2.2f%%' % (accuracy_score(test_Y_cat, pred_Y_cat[0])))\nprint(classification_report(test_Y_cat, pred_Y_cat[0]))","execution_count":16,"outputs":[{"output_type":"stream","text":"100/100 [==============================] - 1s 9ms/step\nAccuracy on Test Data: 0.56%\n precision recall f1-score support\n\n 0 0.79 0.64 0.71 75\n 1 0.00 0.00 0.00 4\n 2 0.21 0.44 0.29 16\n 3 0.00 0.00 0.00 1\n 4 0.25 0.25 0.25 4\n\n micro avg 0.56 0.56 0.56 100\n macro avg 0.25 0.27 0.25 100\nweighted avg 0.63 0.56 0.59 100\n\n","name":"stdout"},{"output_type":"stream","text":"/opt/conda/lib/python3.6/site-packages/sklearn/metrics/classification.py:1143: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples.\n 'precision', 'predicted', average, warn_for)\n","name":"stderr"}]},{"metadata":{"trusted":true},"cell_type":"code","source":"import seaborn as sns\nfrom sklearn.metrics import confusion_matrix\nsns.heatmap(confusion_matrix(test_Y_cat, pred_Y_cat[0]), \n annot=True, fmt=\"d\", cbar = False, cmap = plt.cm.Blues, vmax = X_test_resh.shape[0]//16)","execution_count":17,"outputs":[{"output_type":"execute_result","execution_count":17,"data":{"text/plain":""},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"
","image/png":"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\n"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"pre = pred_Y[0]\n\nfrom sklearn.metrics import roc_curve, roc_auc_score\nsick_vec = test_Y_cat>0\nsick_score = np.sum(pre[:,1:],1)\nfpr, tpr, _ = roc_curve(sick_vec, sick_score)\nfig, ax1 = plt.subplots(1,1, figsize = (6, 6), dpi = 150)\nax1.plot(fpr, tpr, 'b.-', label = 'Model Prediction (AUC: %2.2f)' % roc_auc_score(sick_vec, sick_score))\nax1.plot(fpr, fpr, 'g-', label = 'Random Guessing')\nax1.legend()\nax1.set_xlabel('False Positive Rate')\nax1.set_ylabel('True Positive Rate');","execution_count":18,"outputs":[{"output_type":"display_data","data":{"text/plain":"
","image/png":"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\n"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"fig.savefig(\"roc.png\")","execution_count":26,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"from keras.models import model_from_json\n# Model to JSON\nmodel_json = model.to_json()\nwith open(\"model_project_work.json\", \"w\") as json_file:\n json_file.write(model_json)\n# Weights to HDF5\nmodel.save(\"model_project_work.h5\")","execution_count":19,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"os.listdir(\"../working\")","execution_count":27,"outputs":[{"output_type":"execute_result","execution_count":27,"data":{"text/plain":"['.ipynb_checkpoints',\n 'model_project_work.json',\n '__notebook_source__.ipynb',\n 'trainmetrics.png',\n 'roc.png',\n 'model_project_work.h5']"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"from keras.models import load_model\nnew_model = load_model('model_project_work.h5')\n","execution_count":21,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"pred_Y = new_model.predict(X_test_resh, batch_size = 32, verbose = True)\n\npred_Y_cat = np.argmax(pred_Y, -1)\n#print(pred_Y_cat[0])\n\ntest_Y_cat = np.argmax(Y_test, -1)\nprint('Accuracy on Test Data: %2.2f%%' % (accuracy_score(test_Y_cat, pred_Y_cat[0])))\nprint(classification_report(test_Y_cat, pred_Y_cat[0]))","execution_count":22,"outputs":[{"output_type":"stream","text":"100/100 [==============================] - 1s 11ms/step\nAccuracy on Test Data: 0.56%\n precision recall f1-score support\n\n 0 0.79 0.64 0.71 75\n 1 0.00 0.00 0.00 4\n 2 0.21 0.44 0.29 16\n 3 0.00 0.00 0.00 1\n 4 0.25 0.25 0.25 4\n\n micro avg 0.56 0.56 0.56 100\n macro avg 0.25 0.27 0.25 100\nweighted avg 0.63 0.56 0.59 100\n\n","name":"stdout"},{"output_type":"stream","text":"/opt/conda/lib/python3.6/site-packages/sklearn/metrics/classification.py:1143: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples.\n 'precision', 'predicted', average, warn_for)\n/opt/conda/lib/python3.6/site-packages/sklearn/metrics/classification.py:1143: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples.\n 'precision', 'predicted', average, warn_for)\n/opt/conda/lib/python3.6/site-packages/sklearn/metrics/classification.py:1143: UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples.\n 'precision', 'predicted', average, warn_for)\n","name":"stderr"}]}],"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}