{"cells":[{"metadata":{},"cell_type":"markdown","source":"![](https://raw.githubusercontent.com/visipedia/imet-fgvcx/master/assets/banner.png)\n

Using the bottleneck features of a pre-trained network

\n\n#### I'm sharing this code since it can get a little tricky to beginners do this, especially finding the right methods, APIs and models, I hope this can help someone.\n\n#### Using the bottleneck features of a pre-trained model is basically using the models as a preprocessing step on your data, so you get a model (VGG16 in this case) and pass your data through it, the output will be the representation of your data according to the model. Then take these features and use on any other model (another deep learning model or even SVM).\n\n#### What you need to know:\n- Similar to fine-tuning you need to remove the top of the pre-trained model.\n- Essentially it works as a pipeline with two models.\n- To have good results with this approach your data need to be similar to the same data used to train the model (ImageNet).\n- This will make your training a lot faster since you only need to pass all data once through the big model (VGG 16) than just train a smaller one on a less complex data."},{"metadata":{},"cell_type":"markdown","source":"### Dependencies"},{"metadata":{"trusted":true,"_kg_hide-input":true,"_kg_hide-output":true},"cell_type":"code","source":"import os\nimport cv2\nimport math\nimport warnings\nimport numpy as np\nimport pandas as pd\nimport seaborn as sns\nimport tensorflow as tf\nimport matplotlib.pyplot as plt\nfrom sklearn.model_selection import train_test_split\nfrom sklearn.metrics import confusion_matrix, fbeta_score\nfrom keras import optimizers\nfrom keras import backend as K\nfrom keras.models import Sequential\nfrom keras.applications.vgg16 import VGG16\nfrom keras.preprocessing.image import ImageDataGenerator\nfrom keras.callbacks import LearningRateScheduler, EarlyStopping\nfrom keras.layers import Dense, Dropout, Flatten, Conv2D, MaxPool2D, Activation, BatchNormalization\n\n# Set seeds to make the experiment more reproducible.\nfrom tensorflow import set_random_seed\nfrom numpy.random import seed\nset_random_seed(0)\nseed(0)\n\n%matplotlib inline\nsns.set(style=\"whitegrid\")\nwarnings.filterwarnings(\"ignore\")","execution_count":1,"outputs":[{"output_type":"stream","text":"Using TensorFlow backend.\n","name":"stderr"}]},{"metadata":{},"cell_type":"markdown","source":"### Load data"},{"metadata":{"_kg_hide-input":true,"trusted":true},"cell_type":"code","source":"train = pd.read_csv('../input/imet-2019-fgvc6/train.csv')\nlabels = pd.read_csv('../input/imet-2019-fgvc6/labels.csv')\ntest = pd.read_csv('../input/imet-2019-fgvc6/sample_submission.csv')\n\ntrain[\"attribute_ids\"] = train[\"attribute_ids\"].apply(lambda x:x.split(\" \"))\ntrain[\"id\"] = train[\"id\"].apply(lambda x: x + \".png\")\ntest[\"id\"] = test[\"id\"].apply(lambda x: x + \".png\")\n\nprint('Number of train samples: ', train.shape[0])\nprint('Number of test samples: ', test.shape[0])\nprint('Number of labels: ', labels.shape[0])\ndisplay(train.head())\ndisplay(labels.head())","execution_count":2,"outputs":[{"output_type":"stream","text":"Number of train samples: 109237\nNumber of test samples: 7443\nNumber of labels: 1103\n","name":"stdout"},{"output_type":"display_data","data":{"text/plain":" id attribute_ids\n0 1000483014d91860.png [147, 616, 813]\n1 1000fe2e667721fe.png [51, 616, 734, 813]\n2 1001614cb89646ee.png [776]\n3 10041eb49b297c08.png [51, 671, 698, 813, 1092]\n4 100501c227f8beea.png [13, 404, 492, 903, 1093]","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 \n \n \n \n \n \n \n \n \n \n
idattribute_ids
01000483014d91860.png[147, 616, 813]
11000fe2e667721fe.png[51, 616, 734, 813]
21001614cb89646ee.png[776]
310041eb49b297c08.png[51, 671, 698, 813, 1092]
4100501c227f8beea.png[13, 404, 492, 903, 1093]
\n
"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":" attribute_id attribute_name\n0 0 culture::abruzzi\n1 1 culture::achaemenid\n2 2 culture::aegean\n3 3 culture::afghan\n4 4 culture::after british","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 \n \n \n \n \n \n \n \n \n \n
attribute_idattribute_name
00culture::abruzzi
11culture::achaemenid
22culture::aegean
33culture::afghan
44culture::after british
\n
"},"metadata":{}}]},{"metadata":{"trusted":true},"cell_type":"code","source":"# Parameters\nBATCH_SIZE = 64\nEPOCHS = 200\nLEARNING_RATE = 0.0001\nHEIGHT = 128\nWIDTH = 128\nCANAL = 3\nN_CLASSES = labels.shape[0]\nES_PATIENCE = 5\nDECAY_DROP = 0.5\nDECAY_EPOCHS = 10\nclasses = list(map(str, range(N_CLASSES)))","execution_count":3,"outputs":[]},{"metadata":{"_kg_hide-input":true,"_kg_hide-output":true,"trusted":true},"cell_type":"code","source":"def f2_score_thr(threshold=0.5):\n def f2_score(y_true, y_pred):\n beta = 2\n y_pred = K.cast(K.greater(K.clip(y_pred, 0, 1), threshold), K.floatx())\n\n true_positives = K.sum(K.clip(y_true * y_pred, 0, 1), axis=1)\n predicted_positives = K.sum(K.clip(y_pred, 0, 1), axis=1)\n possible_positives = K.sum(K.clip(y_true, 0, 1), axis=1)\n\n precision = true_positives / (predicted_positives + K.epsilon())\n recall = true_positives / (possible_positives + K.epsilon())\n\n return K.mean(((1+beta**2)*precision*recall) / ((beta**2)*precision+recall+K.epsilon()))\n return f2_score\n\ndef step_decay(epoch):\n initial_lrate = LEARNING_RATE\n drop = DECAY_DROP\n epochs_drop = DECAY_EPOCHS\n lrate = initial_lrate * math.pow(drop, math.floor((1+epoch)/epochs_drop))\n \n return lrate","execution_count":4,"outputs":[]},{"metadata":{"trusted":true,"_kg_hide-input":true},"cell_type":"code","source":"train_datagen = ImageDataGenerator(rescale=1./255)\n\ntrain_generator = train_datagen.flow_from_dataframe(\n dataframe=train,\n directory=\"../input/imet-2019-fgvc6/train\",\n x_col=\"id\",\n y_col=\"attribute_ids\",\n batch_size=BATCH_SIZE,\n shuffle=False,\n class_mode=None,\n target_size=(HEIGHT, WIDTH))\n\ntest_datagen = ImageDataGenerator(rescale=1./255)\n\ntest_generator = test_datagen.flow_from_dataframe( \n dataframe=test,\n directory = \"../input/imet-2019-fgvc6/test\", \n x_col=\"id\",\n target_size=(HEIGHT, WIDTH),\n batch_size=1,\n shuffle=False,\n class_mode=None)","execution_count":5,"outputs":[{"output_type":"stream","text":"Found 109237 images.\nFound 7443 images.\n","name":"stdout"}]},{"metadata":{},"cell_type":"markdown","source":"### Bottleneck model - VGG16 (feature extractor)"},{"metadata":{"trusted":true,"_kg_hide-output":true},"cell_type":"code","source":"model_vgg = VGG16(weights=None, include_top=False)\nmodel_vgg.load_weights('../input/vgg16/vgg16_weights_tf_dim_ordering_tf_kernels_notop.h5')","execution_count":6,"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.\n","name":"stdout"}]},{"metadata":{},"cell_type":"markdown","source":"#### Pass all the train data through the pre-trained model to extract features"},{"metadata":{"_kg_hide-output":true,"trusted":true},"cell_type":"code","source":"STEP_SIZE_TRAIN = train_generator.n // train_generator.batch_size\ntrain_data = model_vgg.predict_generator(train_generator, STEP_SIZE_TRAIN)","execution_count":7,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"#### Recreate the train labels"},{"metadata":{"trusted":true},"cell_type":"code","source":"train_labels = []\nfor label in train['attribute_ids'][:train_data.shape[0]].values:\n zeros = np.zeros(N_CLASSES)\n for label_i in label:\n zeros[int(label_i)] = 1\n train_labels.append(zeros)\n \ntrain_labels = np.asarray(train_labels)","execution_count":8,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"#### Split the new train data into train and validation for the next model"},{"metadata":{"trusted":true},"cell_type":"code","source":"X_train, X_val, Y_train, Y_val = train_test_split(train_data, train_labels, test_size=0.2, random_state=0)","execution_count":9,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"### Second model - Deep Learning MLP"},{"metadata":{"trusted":true,"_kg_hide-output":true},"cell_type":"code","source":"model = Sequential()\nmodel.add(Flatten(input_shape=train_data.shape[1:]))\nmodel.add(Dense(2048, activation='relu'))\nmodel.add(Dropout(0.5))\nmodel.add(Dense(1024, activation='relu'))\nmodel.add(Dropout(0.5))\nmodel.add(Dense(N_CLASSES, activation=\"sigmoid\"))\n\noptimizer = optimizers.Adam(lr=LEARNING_RATE)\nthresholds = [0.1, 0.15, 0.2, 0.25, 0.28, 0.3, 0.4, 0.5]\nmetrics = [\"accuracy\", \"categorical_accuracy\", f2_score_thr(0.1), f2_score_thr(0.15), f2_score_thr(0.2), \n f2_score_thr(0.25), f2_score_thr(0.28), f2_score_thr(0.3), f2_score_thr(0.4), f2_score_thr(0.5)]\nlrate = LearningRateScheduler(step_decay)\nes = EarlyStopping(monitor='val_loss', mode='min', verbose=1, patience=ES_PATIENCE)\ncallbacks = [es]\nmodel.compile(optimizer=optimizer, loss=\"binary_crossentropy\", metrics=metrics)","execution_count":10,"outputs":[{"output_type":"stream","text":"WARNING: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","name":"stdout"}]},{"metadata":{"trusted":true,"_kg_hide-output":true,"_kg_hide-input":true},"cell_type":"code","source":"history = model.fit(x=X_train, y=Y_train,\n validation_data=(X_val, Y_val),\n epochs=EPOCHS,\n batch_size=BATCH_SIZE,\n callbacks=callbacks,\n verbose=2)","execution_count":11,"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 87347 samples, validate on 21837 samples\nEpoch 1/200\n - 18s - loss: 0.0253 - acc: 0.9925 - categorical_accuracy: 0.0686 - f2_score: 0.1695 - f2_score_1: 0.1414 - f2_score_2: 0.1162 - f2_score_3: 0.0953 - f2_score_4: 0.0845 - f2_score_5: 0.0781 - f2_score_6: 0.0519 - f2_score_7: 0.0337 - val_loss: 0.0128 - val_acc: 0.9972 - val_categorical_accuracy: 0.1243 - val_f2_score: 0.2746 - val_f2_score_1: 0.2165 - val_f2_score_2: 0.1681 - val_f2_score_3: 0.1316 - val_f2_score_4: 0.1134 - val_f2_score_5: 0.1031 - val_f2_score_6: 0.0615 - val_f2_score_7: 0.0353\nEpoch 2/200\n - 17s - loss: 0.0135 - acc: 0.9971 - categorical_accuracy: 0.1220 - f2_score: 0.2752 - f2_score_1: 0.2369 - f2_score_2: 0.1994 - f2_score_3: 0.1672 - f2_score_4: 0.1498 - f2_score_5: 0.1390 - f2_score_6: 0.0950 - f2_score_7: 0.0644 - val_loss: 0.0120 - val_acc: 0.9972 - val_categorical_accuracy: 0.1556 - val_f2_score: 0.3363 - val_f2_score_1: 0.2918 - val_f2_score_2: 0.2458 - val_f2_score_3: 0.2041 - val_f2_score_4: 0.1824 - val_f2_score_5: 0.1696 - val_f2_score_6: 0.1181 - val_f2_score_7: 0.0810\nEpoch 3/200\n - 17s - loss: 0.0125 - acc: 0.9972 - categorical_accuracy: 0.1467 - f2_score: 0.3189 - f2_score_1: 0.2828 - f2_score_2: 0.2443 - f2_score_3: 0.2086 - f2_score_4: 0.1892 - f2_score_5: 0.1769 - f2_score_6: 0.1263 - f2_score_7: 0.0898 - val_loss: 0.0115 - val_acc: 0.9973 - val_categorical_accuracy: 0.1723 - val_f2_score: 0.3667 - val_f2_score_1: 0.3330 - val_f2_score_2: 0.2931 - val_f2_score_3: 0.2522 - val_f2_score_4: 0.2280 - val_f2_score_5: 0.2130 - val_f2_score_6: 0.1500 - val_f2_score_7: 0.1070\nEpoch 4/200\n - 17s - loss: 0.0120 - acc: 0.9972 - categorical_accuracy: 0.1584 - f2_score: 0.3464 - f2_score_1: 0.3128 - f2_score_2: 0.2747 - f2_score_3: 0.2379 - f2_score_4: 0.2176 - f2_score_5: 0.2048 - f2_score_6: 0.1506 - f2_score_7: 0.1097 - val_loss: 0.0111 - val_acc: 0.9973 - val_categorical_accuracy: 0.1669 - val_f2_score: 0.3882 - val_f2_score_1: 0.3608 - val_f2_score_2: 0.3245 - val_f2_score_3: 0.2883 - val_f2_score_4: 0.2683 - val_f2_score_5: 0.2540 - val_f2_score_6: 0.1934 - val_f2_score_7: 0.1418\nEpoch 5/200\n - 17s - loss: 0.0115 - acc: 0.9973 - categorical_accuracy: 0.1705 - f2_score: 0.3663 - f2_score_1: 0.3339 - f2_score_2: 0.2967 - f2_score_3: 0.2591 - f2_score_4: 0.2383 - f2_score_5: 0.2254 - f2_score_6: 0.1701 - f2_score_7: 0.1263 - val_loss: 0.0110 - val_acc: 0.9973 - val_categorical_accuracy: 0.1583 - val_f2_score: 0.4003 - val_f2_score_1: 0.3840 - val_f2_score_2: 0.3526 - val_f2_score_3: 0.3206 - val_f2_score_4: 0.3007 - val_f2_score_5: 0.2875 - val_f2_score_6: 0.2241 - val_f2_score_7: 0.1721\nEpoch 6/200\n - 17s - loss: 0.0112 - acc: 0.9973 - categorical_accuracy: 0.1812 - f2_score: 0.3809 - f2_score_1: 0.3510 - f2_score_2: 0.3145 - f2_score_3: 0.2777 - f2_score_4: 0.2565 - f2_score_5: 0.2433 - f2_score_6: 0.1850 - f2_score_7: 0.1402 - val_loss: 0.0108 - val_acc: 0.9974 - val_categorical_accuracy: 0.1715 - val_f2_score: 0.4131 - val_f2_score_1: 0.4002 - val_f2_score_2: 0.3743 - val_f2_score_3: 0.3411 - val_f2_score_4: 0.3208 - val_f2_score_5: 0.3073 - val_f2_score_6: 0.2423 - val_f2_score_7: 0.1861\nEpoch 7/200\n - 17s - loss: 0.0110 - acc: 0.9973 - categorical_accuracy: 0.1925 - f2_score: 0.3943 - f2_score_1: 0.3657 - f2_score_2: 0.3295 - f2_score_3: 0.2931 - f2_score_4: 0.2723 - f2_score_5: 0.2589 - f2_score_6: 0.1997 - f2_score_7: 0.1520 - val_loss: 0.0108 - val_acc: 0.9974 - val_categorical_accuracy: 0.1848 - val_f2_score: 0.4200 - val_f2_score_1: 0.4106 - val_f2_score_2: 0.3839 - val_f2_score_3: 0.3511 - val_f2_score_4: 0.3319 - val_f2_score_5: 0.3189 - val_f2_score_6: 0.2555 - val_f2_score_7: 0.1998\nEpoch 8/200\n - 17s - loss: 0.0107 - acc: 0.9974 - categorical_accuracy: 0.2018 - f2_score: 0.4059 - f2_score_1: 0.3782 - f2_score_2: 0.3432 - f2_score_3: 0.3065 - f2_score_4: 0.2857 - f2_score_5: 0.2723 - f2_score_6: 0.2121 - f2_score_7: 0.1640 - val_loss: 0.0107 - val_acc: 0.9974 - val_categorical_accuracy: 0.1937 - val_f2_score: 0.4242 - val_f2_score_1: 0.4147 - val_f2_score_2: 0.3909 - val_f2_score_3: 0.3619 - val_f2_score_4: 0.3425 - val_f2_score_5: 0.3298 - val_f2_score_6: 0.2672 - val_f2_score_7: 0.2118\nEpoch 9/200\n - 17s - loss: 0.0105 - acc: 0.9974 - categorical_accuracy: 0.2117 - f2_score: 0.4161 - f2_score_1: 0.3899 - f2_score_2: 0.3555 - f2_score_3: 0.3193 - f2_score_4: 0.2981 - f2_score_5: 0.2845 - f2_score_6: 0.2231 - f2_score_7: 0.1741 - val_loss: 0.0105 - val_acc: 0.9974 - val_categorical_accuracy: 0.2069 - val_f2_score: 0.4335 - val_f2_score_1: 0.4219 - val_f2_score_2: 0.3980 - val_f2_score_3: 0.3671 - val_f2_score_4: 0.3488 - val_f2_score_5: 0.3364 - val_f2_score_6: 0.2738 - val_f2_score_7: 0.2171\nEpoch 10/200\n - 17s - loss: 0.0103 - acc: 0.9974 - categorical_accuracy: 0.2200 - f2_score: 0.4251 - f2_score_1: 0.3999 - f2_score_2: 0.3656 - f2_score_3: 0.3308 - f2_score_4: 0.3105 - f2_score_5: 0.2972 - f2_score_6: 0.2361 - f2_score_7: 0.1855 - val_loss: 0.0104 - val_acc: 0.9974 - val_categorical_accuracy: 0.2063 - val_f2_score: 0.4357 - val_f2_score_1: 0.4267 - val_f2_score_2: 0.4026 - val_f2_score_3: 0.3732 - val_f2_score_4: 0.3550 - val_f2_score_5: 0.3420 - val_f2_score_6: 0.2808 - val_f2_score_7: 0.2236\nEpoch 11/200\n - 17s - loss: 0.0102 - acc: 0.9974 - categorical_accuracy: 0.2272 - f2_score: 0.4334 - f2_score_1: 0.4086 - f2_score_2: 0.3759 - f2_score_3: 0.3412 - f2_score_4: 0.3208 - f2_score_5: 0.3073 - f2_score_6: 0.2451 - f2_score_7: 0.1934 - val_loss: 0.0103 - val_acc: 0.9974 - val_categorical_accuracy: 0.2105 - val_f2_score: 0.4404 - val_f2_score_1: 0.4334 - val_f2_score_2: 0.4095 - val_f2_score_3: 0.3805 - val_f2_score_4: 0.3627 - val_f2_score_5: 0.3504 - val_f2_score_6: 0.2907 - val_f2_score_7: 0.2344\nEpoch 12/200\n - 16s - loss: 0.0100 - acc: 0.9974 - categorical_accuracy: 0.2348 - f2_score: 0.4414 - f2_score_1: 0.4176 - f2_score_2: 0.3851 - f2_score_3: 0.3509 - f2_score_4: 0.3304 - f2_score_5: 0.3173 - f2_score_6: 0.2549 - f2_score_7: 0.2022 - val_loss: 0.0104 - val_acc: 0.9974 - val_categorical_accuracy: 0.2266 - val_f2_score: 0.4441 - val_f2_score_1: 0.4402 - val_f2_score_2: 0.4212 - val_f2_score_3: 0.3935 - val_f2_score_4: 0.3758 - val_f2_score_5: 0.3636 - val_f2_score_6: 0.3039 - val_f2_score_7: 0.2472\nEpoch 13/200\n - 17s - loss: 0.0099 - acc: 0.9975 - categorical_accuracy: 0.2429 - f2_score: 0.4493 - f2_score_1: 0.4265 - f2_score_2: 0.3943 - f2_score_3: 0.3604 - f2_score_4: 0.3403 - f2_score_5: 0.3272 - f2_score_6: 0.2651 - f2_score_7: 0.2118 - val_loss: 0.0103 - val_acc: 0.9974 - val_categorical_accuracy: 0.2230 - val_f2_score: 0.4456 - val_f2_score_1: 0.4415 - val_f2_score_2: 0.4214 - val_f2_score_3: 0.3935 - val_f2_score_4: 0.3755 - val_f2_score_5: 0.3640 - val_f2_score_6: 0.2995 - val_f2_score_7: 0.2426\nEpoch 14/200\n - 17s - loss: 0.0097 - acc: 0.9975 - categorical_accuracy: 0.2490 - f2_score: 0.4563 - f2_score_1: 0.4337 - f2_score_2: 0.4026 - f2_score_3: 0.3691 - f2_score_4: 0.3492 - f2_score_5: 0.3363 - f2_score_6: 0.2740 - f2_score_7: 0.2205 - val_loss: 0.0101 - val_acc: 0.9975 - val_categorical_accuracy: 0.2335 - val_f2_score: 0.4496 - val_f2_score_1: 0.4453 - val_f2_score_2: 0.4262 - val_f2_score_3: 0.4006 - val_f2_score_4: 0.3835 - val_f2_score_5: 0.3723 - val_f2_score_6: 0.3120 - val_f2_score_7: 0.2571\nEpoch 15/200\n - 17s - loss: 0.0096 - acc: 0.9975 - categorical_accuracy: 0.2582 - f2_score: 0.4635 - f2_score_1: 0.4421 - f2_score_2: 0.4110 - f2_score_3: 0.3776 - f2_score_4: 0.3577 - f2_score_5: 0.3443 - f2_score_6: 0.2830 - f2_score_7: 0.2293 - val_loss: 0.0101 - val_acc: 0.9975 - val_categorical_accuracy: 0.2300 - val_f2_score: 0.4532 - val_f2_score_1: 0.4490 - val_f2_score_2: 0.4269 - val_f2_score_3: 0.4007 - val_f2_score_4: 0.3826 - val_f2_score_5: 0.3709 - val_f2_score_6: 0.3132 - val_f2_score_7: 0.2578\nEpoch 16/200\n - 17s - loss: 0.0095 - acc: 0.9975 - categorical_accuracy: 0.2661 - f2_score: 0.4710 - f2_score_1: 0.4502 - f2_score_2: 0.4198 - f2_score_3: 0.3868 - f2_score_4: 0.3675 - f2_score_5: 0.3545 - f2_score_6: 0.2924 - f2_score_7: 0.2385 - val_loss: 0.0102 - val_acc: 0.9975 - val_categorical_accuracy: 0.2378 - val_f2_score: 0.4534 - val_f2_score_1: 0.4517 - val_f2_score_2: 0.4328 - val_f2_score_3: 0.4057 - val_f2_score_4: 0.3879 - val_f2_score_5: 0.3766 - val_f2_score_6: 0.3202 - val_f2_score_7: 0.2675\n","name":"stdout"},{"output_type":"stream","text":"Epoch 17/200\n - 16s - loss: 0.0094 - acc: 0.9975 - categorical_accuracy: 0.2723 - f2_score: 0.4775 - f2_score_1: 0.4568 - f2_score_2: 0.4275 - f2_score_3: 0.3947 - f2_score_4: 0.3752 - f2_score_5: 0.3624 - f2_score_6: 0.3008 - f2_score_7: 0.2463 - val_loss: 0.0101 - val_acc: 0.9975 - val_categorical_accuracy: 0.2466 - val_f2_score: 0.4551 - val_f2_score_1: 0.4541 - val_f2_score_2: 0.4380 - val_f2_score_3: 0.4144 - val_f2_score_4: 0.3989 - val_f2_score_5: 0.3875 - val_f2_score_6: 0.3326 - val_f2_score_7: 0.2778\nEpoch 18/200\n - 17s - loss: 0.0093 - acc: 0.9975 - categorical_accuracy: 0.2780 - f2_score: 0.4830 - f2_score_1: 0.4630 - f2_score_2: 0.4341 - f2_score_3: 0.4018 - f2_score_4: 0.3822 - f2_score_5: 0.3691 - f2_score_6: 0.3078 - f2_score_7: 0.2528 - val_loss: 0.0101 - val_acc: 0.9975 - val_categorical_accuracy: 0.2310 - val_f2_score: 0.4563 - val_f2_score_1: 0.4546 - val_f2_score_2: 0.4383 - val_f2_score_3: 0.4145 - val_f2_score_4: 0.3983 - val_f2_score_5: 0.3883 - val_f2_score_6: 0.3326 - val_f2_score_7: 0.2776\nEpoch 19/200\n - 17s - loss: 0.0091 - acc: 0.9976 - categorical_accuracy: 0.2829 - f2_score: 0.4890 - f2_score_1: 0.4699 - f2_score_2: 0.4413 - f2_score_3: 0.4093 - f2_score_4: 0.3899 - f2_score_5: 0.3765 - f2_score_6: 0.3165 - f2_score_7: 0.2607 - val_loss: 0.0101 - val_acc: 0.9975 - val_categorical_accuracy: 0.2475 - val_f2_score: 0.4574 - val_f2_score_1: 0.4561 - val_f2_score_2: 0.4354 - val_f2_score_3: 0.4093 - val_f2_score_4: 0.3926 - val_f2_score_5: 0.3815 - val_f2_score_6: 0.3224 - val_f2_score_7: 0.2702\nEpoch 20/200\n - 17s - loss: 0.0090 - acc: 0.9976 - categorical_accuracy: 0.2920 - f2_score: 0.4952 - f2_score_1: 0.4766 - f2_score_2: 0.4479 - f2_score_3: 0.4174 - f2_score_4: 0.3984 - f2_score_5: 0.3857 - f2_score_6: 0.3239 - f2_score_7: 0.2687 - val_loss: 0.0100 - val_acc: 0.9975 - val_categorical_accuracy: 0.2439 - val_f2_score: 0.4612 - val_f2_score_1: 0.4573 - val_f2_score_2: 0.4405 - val_f2_score_3: 0.4163 - val_f2_score_4: 0.4012 - val_f2_score_5: 0.3896 - val_f2_score_6: 0.3348 - val_f2_score_7: 0.2826\nEpoch 21/200\n - 17s - loss: 0.0089 - acc: 0.9976 - categorical_accuracy: 0.2973 - f2_score: 0.5002 - f2_score_1: 0.4822 - f2_score_2: 0.4548 - f2_score_3: 0.4243 - f2_score_4: 0.4055 - f2_score_5: 0.3930 - f2_score_6: 0.3326 - f2_score_7: 0.2766 - val_loss: 0.0101 - val_acc: 0.9975 - val_categorical_accuracy: 0.2424 - val_f2_score: 0.4594 - val_f2_score_1: 0.4613 - val_f2_score_2: 0.4468 - val_f2_score_3: 0.4250 - val_f2_score_4: 0.4109 - val_f2_score_5: 0.4006 - val_f2_score_6: 0.3469 - val_f2_score_7: 0.2945\nEpoch 22/200\n - 17s - loss: 0.0088 - acc: 0.9976 - categorical_accuracy: 0.3040 - f2_score: 0.5066 - f2_score_1: 0.4899 - f2_score_2: 0.4625 - f2_score_3: 0.4321 - f2_score_4: 0.4137 - f2_score_5: 0.4014 - f2_score_6: 0.3405 - f2_score_7: 0.2847 - val_loss: 0.0099 - val_acc: 0.9975 - val_categorical_accuracy: 0.2460 - val_f2_score: 0.4629 - val_f2_score_1: 0.4616 - val_f2_score_2: 0.4459 - val_f2_score_3: 0.4230 - val_f2_score_4: 0.4070 - val_f2_score_5: 0.3971 - val_f2_score_6: 0.3442 - val_f2_score_7: 0.2920\nEpoch 23/200\n - 16s - loss: 0.0087 - acc: 0.9976 - categorical_accuracy: 0.3117 - f2_score: 0.5126 - f2_score_1: 0.4956 - f2_score_2: 0.4691 - f2_score_3: 0.4391 - f2_score_4: 0.4210 - f2_score_5: 0.4087 - f2_score_6: 0.3491 - f2_score_7: 0.2930 - val_loss: 0.0100 - val_acc: 0.9975 - val_categorical_accuracy: 0.2508 - val_f2_score: 0.4640 - val_f2_score_1: 0.4641 - val_f2_score_2: 0.4477 - val_f2_score_3: 0.4246 - val_f2_score_4: 0.4093 - val_f2_score_5: 0.3992 - val_f2_score_6: 0.3467 - val_f2_score_7: 0.2958\nEpoch 24/200\n - 16s - loss: 0.0086 - acc: 0.9976 - categorical_accuracy: 0.3161 - f2_score: 0.5183 - f2_score_1: 0.5029 - f2_score_2: 0.4763 - f2_score_3: 0.4464 - f2_score_4: 0.4275 - f2_score_5: 0.4152 - f2_score_6: 0.3555 - f2_score_7: 0.2997 - val_loss: 0.0100 - val_acc: 0.9975 - val_categorical_accuracy: 0.2521 - val_f2_score: 0.4653 - val_f2_score_1: 0.4656 - val_f2_score_2: 0.4495 - val_f2_score_3: 0.4303 - val_f2_score_4: 0.4165 - val_f2_score_5: 0.4073 - val_f2_score_6: 0.3561 - val_f2_score_7: 0.3048\nEpoch 25/200\n - 17s - loss: 0.0085 - acc: 0.9977 - categorical_accuracy: 0.3219 - f2_score: 0.5227 - f2_score_1: 0.5073 - f2_score_2: 0.4813 - f2_score_3: 0.4520 - f2_score_4: 0.4340 - f2_score_5: 0.4216 - f2_score_6: 0.3627 - f2_score_7: 0.3069 - val_loss: 0.0099 - val_acc: 0.9975 - val_categorical_accuracy: 0.2553 - val_f2_score: 0.4662 - val_f2_score_1: 0.4648 - val_f2_score_2: 0.4481 - val_f2_score_3: 0.4253 - val_f2_score_4: 0.4107 - val_f2_score_5: 0.4010 - val_f2_score_6: 0.3496 - val_f2_score_7: 0.3003\nEpoch 26/200\n - 17s - loss: 0.0084 - acc: 0.9977 - categorical_accuracy: 0.3297 - f2_score: 0.5288 - f2_score_1: 0.5132 - f2_score_2: 0.4880 - f2_score_3: 0.4594 - f2_score_4: 0.4417 - f2_score_5: 0.4296 - f2_score_6: 0.3710 - f2_score_7: 0.3151 - val_loss: 0.0100 - val_acc: 0.9975 - val_categorical_accuracy: 0.2532 - val_f2_score: 0.4658 - val_f2_score_1: 0.4669 - val_f2_score_2: 0.4537 - val_f2_score_3: 0.4336 - val_f2_score_4: 0.4191 - val_f2_score_5: 0.4094 - val_f2_score_6: 0.3592 - val_f2_score_7: 0.3074\nEpoch 27/200\n - 16s - loss: 0.0083 - acc: 0.9977 - categorical_accuracy: 0.3356 - f2_score: 0.5332 - f2_score_1: 0.5184 - f2_score_2: 0.4931 - f2_score_3: 0.4654 - f2_score_4: 0.4481 - f2_score_5: 0.4358 - f2_score_6: 0.3773 - f2_score_7: 0.3210 - val_loss: 0.0099 - val_acc: 0.9975 - val_categorical_accuracy: 0.2603 - val_f2_score: 0.4666 - val_f2_score_1: 0.4666 - val_f2_score_2: 0.4523 - val_f2_score_3: 0.4304 - val_f2_score_4: 0.4169 - val_f2_score_5: 0.4068 - val_f2_score_6: 0.3578 - val_f2_score_7: 0.3101\nEpoch 28/200\n - 16s - loss: 0.0082 - acc: 0.9977 - categorical_accuracy: 0.3397 - f2_score: 0.5384 - f2_score_1: 0.5242 - f2_score_2: 0.4994 - f2_score_3: 0.4713 - f2_score_4: 0.4541 - f2_score_5: 0.4426 - f2_score_6: 0.3845 - f2_score_7: 0.3276 - val_loss: 0.0099 - val_acc: 0.9975 - val_categorical_accuracy: 0.2701 - val_f2_score: 0.4667 - val_f2_score_1: 0.4654 - val_f2_score_2: 0.4484 - val_f2_score_3: 0.4267 - val_f2_score_4: 0.4123 - val_f2_score_5: 0.4025 - val_f2_score_6: 0.3524 - val_f2_score_7: 0.3027\nEpoch 29/200\n - 16s - loss: 0.0081 - acc: 0.9977 - categorical_accuracy: 0.3462 - f2_score: 0.5435 - f2_score_1: 0.5302 - f2_score_2: 0.5060 - f2_score_3: 0.4778 - f2_score_4: 0.4609 - f2_score_5: 0.4493 - f2_score_6: 0.3916 - f2_score_7: 0.3356 - val_loss: 0.0099 - val_acc: 0.9975 - val_categorical_accuracy: 0.2732 - val_f2_score: 0.4689 - val_f2_score_1: 0.4682 - val_f2_score_2: 0.4531 - val_f2_score_3: 0.4327 - val_f2_score_4: 0.4185 - val_f2_score_5: 0.4083 - val_f2_score_6: 0.3575 - val_f2_score_7: 0.3095\nEpoch 30/200\n - 17s - loss: 0.0080 - acc: 0.9977 - categorical_accuracy: 0.3497 - f2_score: 0.5491 - f2_score_1: 0.5359 - f2_score_2: 0.5120 - f2_score_3: 0.4841 - f2_score_4: 0.4670 - f2_score_5: 0.4556 - f2_score_6: 0.3984 - f2_score_7: 0.3432 - val_loss: 0.0099 - val_acc: 0.9975 - val_categorical_accuracy: 0.2630 - val_f2_score: 0.4684 - val_f2_score_1: 0.4696 - val_f2_score_2: 0.4544 - val_f2_score_3: 0.4345 - val_f2_score_4: 0.4215 - val_f2_score_5: 0.4116 - val_f2_score_6: 0.3633 - val_f2_score_7: 0.3159\nEpoch 31/200\n - 16s - loss: 0.0080 - acc: 0.9978 - categorical_accuracy: 0.3573 - f2_score: 0.5536 - f2_score_1: 0.5413 - f2_score_2: 0.5177 - f2_score_3: 0.4913 - f2_score_4: 0.4738 - f2_score_5: 0.4626 - f2_score_6: 0.4058 - f2_score_7: 0.3509 - val_loss: 0.0099 - val_acc: 0.9975 - val_categorical_accuracy: 0.2708 - val_f2_score: 0.4693 - val_f2_score_1: 0.4695 - val_f2_score_2: 0.4584 - val_f2_score_3: 0.4397 - val_f2_score_4: 0.4268 - val_f2_score_5: 0.4178 - val_f2_score_6: 0.3711 - val_f2_score_7: 0.3219\nEpoch 32/200\n - 16s - loss: 0.0079 - acc: 0.9978 - categorical_accuracy: 0.3646 - f2_score: 0.5584 - f2_score_1: 0.5453 - f2_score_2: 0.5225 - f2_score_3: 0.4959 - f2_score_4: 0.4791 - f2_score_5: 0.4677 - f2_score_6: 0.4119 - f2_score_7: 0.3571 - val_loss: 0.0099 - val_acc: 0.9975 - val_categorical_accuracy: 0.2650 - val_f2_score: 0.4687 - val_f2_score_1: 0.4678 - val_f2_score_2: 0.4551 - val_f2_score_3: 0.4355 - val_f2_score_4: 0.4227 - val_f2_score_5: 0.4127 - val_f2_score_6: 0.3646 - val_f2_score_7: 0.3181\n","name":"stdout"},{"output_type":"stream","text":"Epoch 33/200\n - 16s - loss: 0.0078 - acc: 0.9978 - categorical_accuracy: 0.3687 - f2_score: 0.5636 - f2_score_1: 0.5512 - f2_score_2: 0.5296 - f2_score_3: 0.5030 - f2_score_4: 0.4870 - f2_score_5: 0.4759 - f2_score_6: 0.4193 - f2_score_7: 0.3647 - val_loss: 0.0099 - val_acc: 0.9975 - val_categorical_accuracy: 0.2722 - val_f2_score: 0.4706 - val_f2_score_1: 0.4700 - val_f2_score_2: 0.4555 - val_f2_score_3: 0.4354 - val_f2_score_4: 0.4218 - val_f2_score_5: 0.4135 - val_f2_score_6: 0.3668 - val_f2_score_7: 0.3209\nEpoch 34/200\n - 16s - loss: 0.0077 - acc: 0.9978 - categorical_accuracy: 0.3730 - f2_score: 0.5681 - f2_score_1: 0.5562 - f2_score_2: 0.5338 - f2_score_3: 0.5080 - f2_score_4: 0.4915 - f2_score_5: 0.4808 - f2_score_6: 0.4246 - f2_score_7: 0.3707 - val_loss: 0.0100 - val_acc: 0.9975 - val_categorical_accuracy: 0.2659 - val_f2_score: 0.4681 - val_f2_score_1: 0.4707 - val_f2_score_2: 0.4588 - val_f2_score_3: 0.4402 - val_f2_score_4: 0.4279 - val_f2_score_5: 0.4187 - val_f2_score_6: 0.3724 - val_f2_score_7: 0.3262\nEpoch 35/200\n - 17s - loss: 0.0076 - acc: 0.9978 - categorical_accuracy: 0.3774 - f2_score: 0.5721 - f2_score_1: 0.5603 - f2_score_2: 0.5390 - f2_score_3: 0.5130 - f2_score_4: 0.4970 - f2_score_5: 0.4858 - f2_score_6: 0.4313 - f2_score_7: 0.3767 - val_loss: 0.0100 - val_acc: 0.9975 - val_categorical_accuracy: 0.2797 - val_f2_score: 0.4691 - val_f2_score_1: 0.4711 - val_f2_score_2: 0.4574 - val_f2_score_3: 0.4367 - val_f2_score_4: 0.4242 - val_f2_score_5: 0.4149 - val_f2_score_6: 0.3661 - val_f2_score_7: 0.3215\nEpoch 00035: early stopping\n","name":"stdout"}]},{"metadata":{},"cell_type":"markdown","source":"### Model graph loss"},{"metadata":{"_kg_hide-input":true,"trusted":true},"cell_type":"code","source":"sns.set_style(\"whitegrid\")\nfig, (ax1, ax2, ax3) = plt.subplots(1, 3, sharex='col', figsize=(20,7))\n\n\nax1.plot(history.history['loss'], label='Train loss')\nax1.plot(history.history['val_loss'], label='Validation loss')\nax1.legend(loc='best')\nax1.set_title('Loss')\n\nax2.plot(history.history['acc'], label='Train Accuracy')\nax2.plot(history.history['val_acc'], label='Validation accuracy')\nax2.legend(loc='best')\nax2.set_title('Accuracy')\n\nax3.plot(history.history['categorical_accuracy'], label='Train Cat Accuracy')\nax3.plot(history.history['val_categorical_accuracy'], label='Validation Cat Accuracy')\nax3.legend(loc='best')\nax3.set_title('Cat Accuracy')\n\nplt.xlabel('Epochs')\nsns.despine()\nplt.show()","execution_count":12,"outputs":[{"output_type":"display_data","data":{"text/plain":"
","image/png":"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\n"},"metadata":{}}]},{"metadata":{"_kg_hide-input":true,"_kg_hide-output":false,"trusted":true},"cell_type":"code","source":"fig, axes = plt.subplots(4, 2, sharex='col', figsize=(20,7))\n\naxes[0][0].plot(history.history['f2_score'], label='Train F2 Score')\naxes[0][0].plot(history.history['val_f2_score'], label='Validation F2 Score')\naxes[0][0].legend(loc='best')\naxes[0][0].set_title('F2 Score threshold 0.1')\n\naxes[0][1].plot(history.history['f2_score_1'], label='Train F2 Score')\naxes[0][1].plot(history.history['val_f2_score_1'], label='Validation F2 Score')\naxes[0][1].legend(loc='best')\naxes[0][1].set_title('F2 Score threshold 0.15')\n\naxes[1][0].plot(history.history['f2_score_2'], label='Train F2 Score')\naxes[1][0].plot(history.history['val_f2_score_2'], label='Validation F2 Score')\naxes[1][0].legend(loc='best')\naxes[1][0].set_title('F2 Score threshold 0.2')\n\naxes[1][1].plot(history.history['f2_score_3'], label='Train F2 Score')\naxes[1][1].plot(history.history['val_f2_score_3'], label='Validation F2 Score')\naxes[1][1].legend(loc='best')\naxes[1][1].set_title('F2 Score threshold 0.25')\n\naxes[2][0].plot(history.history['f2_score_4'], label='Train F2 Score')\naxes[2][0].plot(history.history['val_f2_score_4'], label='Validation F2 Score')\naxes[2][0].legend(loc='best')\naxes[2][0].set_title('F2 Score threshold 0.28')\n\naxes[2][1].plot(history.history['f2_score_5'], label='Train F2 Score')\naxes[2][1].plot(history.history['val_f2_score_5'], label='Validation F2 Score')\naxes[2][1].legend(loc='best')\naxes[2][1].set_title('F2 Score threshold 0.3')\n\naxes[3][0].plot(history.history['f2_score_6'], label='Train F2 Score')\naxes[3][0].plot(history.history['val_f2_score_6'], label='Validation F2 Score')\naxes[3][0].legend(loc='best')\naxes[3][0].set_title('F2 Score threshold 0.4')\n\naxes[3][1].plot(history.history['f2_score_7'], label='Train F2 Score')\naxes[3][1].plot(history.history['val_f2_score_7'], label='Validation F2 Score')\naxes[3][1].legend(loc='best')\naxes[3][1].set_title('F2 Score threshold 0.5')\n\nplt.xlabel('Epochs')\nsns.despine()\nplt.show()","execution_count":13,"outputs":[{"output_type":"display_data","data":{"text/plain":"
","image/png":"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\n"},"metadata":{}}]},{"metadata":{},"cell_type":"markdown","source":"### Find best threshold value"},{"metadata":{"trusted":true,"_kg_hide-input":true},"cell_type":"code","source":"best_thr = 0\nbest_thr_val = history.history['val_f2_score'][-1]\nfor i in range(1, len(metrics)-2):\n if best_thr_val < history.history['val_f2_score_%s' % i][-1]:\n best_thr_val = history.history['val_f2_score_%s' % i][-1]\n best_thr = i\n\nthreshold = thresholds[best_thr]\nprint('Best threshold is: %s' % threshold)","execution_count":14,"outputs":[{"output_type":"stream","text":"Best threshold is: 0.15\n","name":"stdout"}]},{"metadata":{},"cell_type":"markdown","source":"### Apply model to test set and output predictions"},{"metadata":{"trusted":true,"_kg_hide-input":false},"cell_type":"code","source":"test_generator.reset()\nSTEP_SIZE_TEST = test_generator.n//test_generator.batch_size\n# Pass the test data through the pre-trained model to extract features\nbottleneck_preds = model_vgg.predict_generator(test_generator, steps=STEP_SIZE_TEST)\n# Make prediction using the second model\npreds = model.predict(bottleneck_preds)","execution_count":15,"outputs":[]},{"metadata":{"trusted":true,"_kg_hide-input":true},"cell_type":"code","source":"predictions = []\nfor pred_ar in preds:\n valid = ''\n for idx, pred in enumerate(pred_ar):\n if pred > threshold:\n if len(valid) == 0:\n valid += str(idx)\n else:\n valid += (' %s' % idx)\n if len(valid) == 0:\n valid = str(np.argmax(pred_ar))\n predictions.append(valid)","execution_count":16,"outputs":[]},{"metadata":{"trusted":true,"_kg_hide-input":true},"cell_type":"code","source":"filenames = test_generator.filenames\nresults = pd.DataFrame({'id':filenames, 'attribute_ids':predictions})\nresults['id'] = results['id'].map(lambda x: str(x)[:-4])\nresults.to_csv('submission.csv',index=False)\nresults.head(10)","execution_count":17,"outputs":[{"output_type":"execute_result","execution_count":17,"data":{"text/plain":" id attribute_ids\n0 10023b2cc4ed5f68 121 223 343 344 369 766 1039 1059\n1 100fbe75ed8fd887 121 1039 1059 1085\n2 101b627524a04f19 79 304 482 498 703 718 813 961\n3 10234480c41284c6 13 51 480 483 725 738 776 830 923 963 1046\n4 1023b0e2636dcea8 147 156 283 322 584 813 903 954 1046 1092\n5 1039cd6cf85845c 13 405 896 903 1092\n6 103a5b3f83fbe88 194 744 813 1092\n7 10413aaae8d6a9a2 51 147 813 1092\n8 10423822b93a65ab 51 147\n9 1052bf702cb099f7 188 597 612 671 723 780","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 \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 \n \n \n \n \n \n \n \n
idattribute_ids
010023b2cc4ed5f68121 223 343 344 369 766 1039 1059
1100fbe75ed8fd887121 1039 1059 1085
2101b627524a04f1979 304 482 498 703 718 813 961
310234480c41284c613 51 480 483 725 738 776 830 923 963 1046
41023b0e2636dcea8147 156 283 322 584 813 903 954 1046 1092
51039cd6cf85845c13 405 896 903 1092
6103a5b3f83fbe88194 744 813 1092
710413aaae8d6a9a251 147 813 1092
810423822b93a65ab51 147
91052bf702cb099f7188 597 612 671 723 780
\n
"},"metadata":{}}]}],"metadata":{"kernelspec":{"display_name":"Python 3","language":"python","name":"python3"},"language_info":{"name":"python","version":"3.6.4","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"}},"nbformat":4,"nbformat_minor":1}