Model_03-Adadelta-k=10.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Mandatory imports\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.metrics import log_loss\n",
    "from sklearn.model_selection import StratifiedKFold, StratifiedShuffleSplit\n",
    "from os.path import join as opj\n",
    "from matplotlib import pyplot as plt\n",
    "from mpl_toolkits.mplot3d import Axes3D\n",
    "import pylab\n",
    "plt.rcParams['figure.figsize'] = 10, 10\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "train = pd.read_json(\"input/train.json\")\n",
    "target_train=train['is_iceberg']\n",
    "test = pd.read_json(\"input/test.json\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "target_train=train['is_iceberg']\n",
    "test['inc_angle']=pd.to_numeric(test['inc_angle'], errors='coerce')\n",
    "train['inc_angle']=pd.to_numeric(train['inc_angle'], errors='coerce')#We have only 133 NAs.\n",
    "train['inc_angle']=train['inc_angle'].fillna(method='pad')\n",
    "X_angle=train['inc_angle']\n",
    "test['inc_angle']=pd.to_numeric(test['inc_angle'], errors='coerce')\n",
    "X_test_angle=test['inc_angle']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Generate the training data\n",
    "X_band_1=np.array([np.array(band).astype(np.float32).reshape(75, 75) for band in train[\"band_1\"]])\n",
    "X_band_2=np.array([np.array(band).astype(np.float32).reshape(75, 75) for band in train[\"band_2\"]])\n",
    "X_band_3=(X_band_1+X_band_2)/2\n",
    "#X_band_3=np.array([np.full((75, 75), angel).astype(np.float32) for angel in train[\"inc_angle\"]])\n",
    "X_train = np.concatenate([X_band_1[:, :, :, np.newaxis]\n",
    "                          , X_band_2[:, :, :, np.newaxis]\n",
    "                         , X_band_3[:, :, :, np.newaxis]], axis=-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "X_band_test_1=np.array([np.array(band).astype(np.float32).reshape(75, 75) for band in test[\"band_1\"]])\n",
    "X_band_test_2=np.array([np.array(band).astype(np.float32).reshape(75, 75) for band in test[\"band_2\"]])\n",
    "X_band_test_3=(X_band_test_1+X_band_test_2)/2\n",
    "#X_band_test_3=np.array([np.full((75, 75), angel).astype(np.float32) for angel in test[\"inc_angle\"]])\n",
    "X_test = np.concatenate([X_band_test_1[:, :, :, np.newaxis]\n",
    "                          , X_band_test_2[:, :, :, np.newaxis]\n",
    "                         , X_band_test_3[:, :, :, np.newaxis]], axis=-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    }
   ],
   "source": [
    "#Import Keras.\n",
    "from matplotlib import pyplot\n",
    "from keras.optimizers import RMSprop\n",
    "# from keras.preprocessing.image import ImageDataGenerator\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Conv2D, MaxPooling2D, Dense, Dropout, Input, Flatten, Activation\n",
    "from keras.layers import GlobalMaxPooling2D\n",
    "from keras.layers.normalization import BatchNormalization\n",
    "from keras.layers.merge import Concatenate\n",
    "from keras.models import Model\n",
    "from keras import initializers\n",
    "from keras.optimizers import Adam\n",
    "from keras.optimizers import rmsprop\n",
    "from keras.layers.advanced_activations import LeakyReLU, PReLU\n",
    "from keras.optimizers import SGD\n",
    "from keras.optimizers import Adadelta\n",
    "from keras.callbacks import ModelCheckpoint, Callback, EarlyStopping\n",
    "\n",
    "from keras.datasets import cifar10\n",
    "from keras.applications.inception_v3 import InceptionV3\n",
    "from keras.applications.vgg16 import VGG16\n",
    "from keras.applications.xception import Xception\n",
    "from keras.applications.mobilenet import MobileNet\n",
    "from keras.applications.vgg19 import VGG19\n",
    "from keras.layers import Concatenate, Dense, LSTM, Input, concatenate\n",
    "from keras.preprocessing import image\n",
    "from keras.applications.vgg16 import preprocess_input"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "#Data Aug for multi-input\n",
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "batch_size=64\n",
    "# Define the image transformations here\n",
    "gen = ImageDataGenerator(horizontal_flip = True,\n",
    "                         vertical_flip = True,\n",
    "                         width_shift_range = 0.,\n",
    "                         height_shift_range = 0.,\n",
    "                         channel_shift_range=0,\n",
    "                         zoom_range = 0.2,\n",
    "                         rotation_range = 10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Here is the function that merges our two generators\n",
    "# We use the exact same generator with the same random seed for both the y and angle arrays\n",
    "def gen_flow_for_two_inputs(X1, X2, y):\n",
    "    genX1 = gen.flow(X1,y,  batch_size=batch_size,seed=55)\n",
    "    genX2 = gen.flow(X1,X2, batch_size=batch_size,seed=55)\n",
    "    while True:\n",
    "            X1i = genX1.next()\n",
    "            X2i = genX2.next()\n",
    "            #Assert arrays are equal - this was for peace of mind, but slows down training\n",
    "            #np.testing.assert_array_equal(X1i[0],X2i[0])\n",
    "            yield [X1i[0], X2i[1]], X1i[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Finally create generator\n",
    "def get_callbacks(filepath, patience=2):\n",
    "    es = EarlyStopping('val_loss', patience=10, mode=\"min\")\n",
    "    msave = ModelCheckpoint(filepath, save_best_only=True)\n",
    "    return [es, msave]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "def getVggAngleModel():\n",
    "    input_2 = Input(shape=[1], name=\"angle\")\n",
    "    angle_layer = Dense(1, )(input_2)\n",
    "    base_model = VGG16(weights='imagenet', include_top=False, \n",
    "                 input_shape=X_train.shape[1:], classes=1)\n",
    "    x = base_model.get_layer('block5_pool').output\n",
    "    \n",
    "\n",
    "    x = GlobalMaxPooling2D()(x)\n",
    "    merge_one = concatenate([x, angle_layer])\n",
    "    merge_one = Dense(512, activation='relu', name='fc2')(merge_one)\n",
    "    merge_one = Dropout(0.3)(merge_one)\n",
    "    merge_one = Dense(512, activation='relu', name='fc3')(merge_one)\n",
    "    merge_one = Dropout(0.3)(merge_one)\n",
    "    \n",
    "    predictions = Dense(1, activation='sigmoid')(merge_one)\n",
    "    \n",
    "    model = Model(inputs=[base_model.input, input_2], outputs=predictions)\n",
    "    \n",
    "    adadelta = Adadelta(lr=1e-3, rho=0.95, epsilon=1e-6, decay=1e-6)\n",
    "    model.compile(loss='binary_crossentropy',\n",
    "                    optimizer=adadelta,\n",
    "                    metrics=['accuracy'])\n",
    "    \n",
    "#     sgd = SGD(lr=1e-3, decay=1e-6, momentum=0.9, nesterov=True)\n",
    "#     model.compile(loss='binary_crossentropy',\n",
    "#                   optimizer=sgd,\n",
    "#                   metrics=['accuracy'])\n",
    "    return model\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "===================FOLD= 0\n",
      "Epoch 1/50\n",
      "24/24 [==============================] - 18s 760ms/step - loss: 0.9204 - acc: 0.5175 - val_loss: 0.6633 - val_acc: 0.6296\n",
      "Epoch 2/50\n",
      "24/24 [==============================] - 13s 552ms/step - loss: 0.7795 - acc: 0.5804 - val_loss: 0.5711 - val_acc: 0.7469\n",
      "Epoch 3/50\n",
      "24/24 [==============================] - 13s 554ms/step - loss: 0.7338 - acc: 0.6038 - val_loss: 0.5015 - val_acc: 0.7469\n",
      "Epoch 4/50\n",
      "24/24 [==============================] - 13s 553ms/step - loss: 0.6466 - acc: 0.6431 - val_loss: 0.4436 - val_acc: 0.8025\n",
      "Epoch 5/50\n",
      "24/24 [==============================] - 13s 555ms/step - loss: 0.6073 - acc: 0.6658 - val_loss: 0.3965 - val_acc: 0.8457\n",
      "Epoch 6/50\n",
      "24/24 [==============================] - 13s 558ms/step - loss: 0.5586 - acc: 0.7019 - val_loss: 0.3657 - val_acc: 0.8395\n",
      "Epoch 7/50\n",
      "24/24 [==============================] - 12s 509ms/step - loss: 0.5111 - acc: 0.7358 - val_loss: 0.3818 - val_acc: 0.8086\n",
      "Epoch 8/50\n",
      "24/24 [==============================] - 13s 559ms/step - loss: 0.4936 - acc: 0.7618 - val_loss: 0.3223 - val_acc: 0.8519\n",
      "Epoch 9/50\n",
      "24/24 [==============================] - 12s 512ms/step - loss: 0.4574 - acc: 0.7649 - val_loss: 0.3269 - val_acc: 0.8580\n",
      "Epoch 10/50\n",
      "24/24 [==============================] - 13s 561ms/step - loss: 0.4242 - acc: 0.7975 - val_loss: 0.3106 - val_acc: 0.8457\n",
      "Epoch 11/50\n",
      "24/24 [==============================] - 13s 560ms/step - loss: 0.4248 - acc: 0.7934 - val_loss: 0.3086 - val_acc: 0.8642\n",
      "Epoch 12/50\n",
      "24/24 [==============================] - 13s 562ms/step - loss: 0.3856 - acc: 0.8190 - val_loss: 0.3008 - val_acc: 0.8642\n",
      "Epoch 13/50\n",
      "24/24 [==============================] - 13s 560ms/step - loss: 0.3613 - acc: 0.8281 - val_loss: 0.2888 - val_acc: 0.8827\n",
      "Epoch 14/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3608 - acc: 0.8342 - val_loss: 0.2891 - val_acc: 0.8765\n",
      "Epoch 15/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.3591 - acc: 0.8396 - val_loss: 0.2808 - val_acc: 0.8765\n",
      "Epoch 16/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.3478 - acc: 0.8364 - val_loss: 0.2781 - val_acc: 0.8580\n",
      "Epoch 17/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.3201 - acc: 0.8527 - val_loss: 0.2890 - val_acc: 0.8580\n",
      "Epoch 18/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.3514 - acc: 0.8457 - val_loss: 0.2667 - val_acc: 0.8827\n",
      "Epoch 19/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3385 - acc: 0.8387 - val_loss: 0.2678 - val_acc: 0.8580\n",
      "Epoch 20/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3318 - acc: 0.8520 - val_loss: 0.3028 - val_acc: 0.8519\n",
      "Epoch 21/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.3307 - acc: 0.8536 - val_loss: 0.2825 - val_acc: 0.8765\n",
      "Epoch 22/50\n",
      "24/24 [==============================] - 13s 562ms/step - loss: 0.3036 - acc: 0.8584 - val_loss: 0.2587 - val_acc: 0.8889\n",
      "Epoch 23/50\n",
      "24/24 [==============================] - 14s 578ms/step - loss: 0.3140 - acc: 0.8592 - val_loss: 0.2571 - val_acc: 0.8827\n",
      "Epoch 24/50\n",
      "24/24 [==============================] - 12s 514ms/step - loss: 0.3063 - acc: 0.8610 - val_loss: 0.2600 - val_acc: 0.8827\n",
      "Epoch 25/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.3028 - acc: 0.8656 - val_loss: 0.2473 - val_acc: 0.8765\n",
      "Epoch 26/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.3052 - acc: 0.8683 - val_loss: 0.2435 - val_acc: 0.8765\n",
      "Epoch 27/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2881 - acc: 0.8705 - val_loss: 0.2438 - val_acc: 0.8765\n",
      "Epoch 28/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2993 - acc: 0.8662 - val_loss: 0.2795 - val_acc: 0.8765\n",
      "Epoch 29/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2700 - acc: 0.8826 - val_loss: 0.2496 - val_acc: 0.8889\n",
      "Epoch 30/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2908 - acc: 0.8676 - val_loss: 0.2611 - val_acc: 0.8951\n",
      "Epoch 31/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2761 - acc: 0.8775 - val_loss: 0.2442 - val_acc: 0.8889\n",
      "Epoch 32/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2782 - acc: 0.8766 - val_loss: 0.2525 - val_acc: 0.8951\n",
      "Epoch 33/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.2852 - acc: 0.8713 - val_loss: 0.2406 - val_acc: 0.8889\n",
      "Epoch 34/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.2629 - acc: 0.8839 - val_loss: 0.2386 - val_acc: 0.9012\n",
      "Epoch 35/50\n",
      "24/24 [==============================] - 12s 513ms/step - loss: 0.2830 - acc: 0.8791 - val_loss: 0.2521 - val_acc: 0.9074\n",
      "Epoch 36/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.2474 - acc: 0.8917 - val_loss: 0.2343 - val_acc: 0.8951\n",
      "Epoch 37/50\n",
      "24/24 [==============================] - 12s 513ms/step - loss: 0.2703 - acc: 0.8811 - val_loss: 0.2349 - val_acc: 0.9012\n",
      "Epoch 38/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.2535 - acc: 0.8858 - val_loss: 0.2233 - val_acc: 0.9012\n",
      "Epoch 39/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2872 - acc: 0.8890 - val_loss: 0.2442 - val_acc: 0.9074\n",
      "Epoch 40/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2642 - acc: 0.8849 - val_loss: 0.2363 - val_acc: 0.9198\n",
      "Epoch 41/50\n",
      "24/24 [==============================] - 12s 514ms/step - loss: 0.2725 - acc: 0.8848 - val_loss: 0.2343 - val_acc: 0.9074\n",
      "Epoch 42/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2679 - acc: 0.8866 - val_loss: 0.2293 - val_acc: 0.8951\n",
      "Epoch 43/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2652 - acc: 0.8790 - val_loss: 0.2465 - val_acc: 0.9074\n",
      "Epoch 44/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2649 - acc: 0.8871 - val_loss: 0.2354 - val_acc: 0.9198\n",
      "Epoch 45/50\n",
      "24/24 [==============================] - 12s 514ms/step - loss: 0.2557 - acc: 0.8890 - val_loss: 0.2290 - val_acc: 0.9012\n",
      "Epoch 46/50\n",
      "24/24 [==============================] - 12s 509ms/step - loss: 0.2501 - acc: 0.8910 - val_loss: 0.2509 - val_acc: 0.9074\n",
      "Epoch 47/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2630 - acc: 0.8869 - val_loss: 0.2343 - val_acc: 0.9259\n",
      "Epoch 48/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2471 - acc: 0.8909 - val_loss: 0.2332 - val_acc: 0.9259\n",
      "Train loss: 0.193217491641\n",
      "Train accuracy: 0.921636615811\n",
      "Test loss: 0.223294598373\n",
      "Test accuracy: 0.901234567901\n",
      "\n",
      "===================FOLD= 1\n",
      "Epoch 1/50\n",
      "24/24 [==============================] - 17s 695ms/step - loss: 0.7977 - acc: 0.5313 - val_loss: 0.6148 - val_acc: 0.6522\n",
      "Epoch 2/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.6803 - acc: 0.6047 - val_loss: 0.5627 - val_acc: 0.7019\n",
      "Epoch 3/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.6558 - acc: 0.6180 - val_loss: 0.5260 - val_acc: 0.7702\n",
      "Epoch 4/50\n",
      "24/24 [==============================] - 13s 562ms/step - loss: 0.6160 - acc: 0.6440 - val_loss: 0.4994 - val_acc: 0.7578\n",
      "Epoch 5/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.5980 - acc: 0.6543 - val_loss: 0.4609 - val_acc: 0.8012\n",
      "Epoch 6/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.5463 - acc: 0.6982 - val_loss: 0.4316 - val_acc: 0.8261\n",
      "Epoch 7/50\n",
      "24/24 [==============================] - 14s 566ms/step - loss: 0.5124 - acc: 0.7375 - val_loss: 0.4156 - val_acc: 0.8075\n",
      "Epoch 8/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.4748 - acc: 0.7518 - val_loss: 0.3921 - val_acc: 0.8385\n",
      "Epoch 9/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.4361 - acc: 0.7864 - val_loss: 0.3708 - val_acc: 0.8447\n",
      "Epoch 10/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.4283 - acc: 0.7919 - val_loss: 0.3624 - val_acc: 0.8447\n",
      "Epoch 11/50\n",
      "24/24 [==============================] - 12s 514ms/step - loss: 0.4128 - acc: 0.7950 - val_loss: 0.3680 - val_acc: 0.8075\n",
      "Epoch 12/50\n",
      "24/24 [==============================] - 14s 566ms/step - loss: 0.3848 - acc: 0.8321 - val_loss: 0.3249 - val_acc: 0.8634\n",
      "Epoch 13/50\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "24/24 [==============================] - 14s 564ms/step - loss: 0.3785 - acc: 0.8278 - val_loss: 0.3103 - val_acc: 0.8385\n",
      "Epoch 14/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3663 - acc: 0.8330 - val_loss: 0.3346 - val_acc: 0.8075\n",
      "Epoch 15/50\n",
      "24/24 [==============================] - 14s 567ms/step - loss: 0.3501 - acc: 0.8412 - val_loss: 0.2887 - val_acc: 0.8696\n",
      "Epoch 16/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3487 - acc: 0.8447 - val_loss: 0.3457 - val_acc: 0.8261\n",
      "Epoch 17/50\n",
      "24/24 [==============================] - 12s 517ms/step - loss: 0.3192 - acc: 0.8545 - val_loss: 0.2958 - val_acc: 0.8323\n",
      "Epoch 18/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3210 - acc: 0.8585 - val_loss: 0.2898 - val_acc: 0.8509\n",
      "Epoch 19/50\n",
      "24/24 [==============================] - 14s 566ms/step - loss: 0.3149 - acc: 0.8627 - val_loss: 0.2695 - val_acc: 0.8944\n",
      "Epoch 20/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3046 - acc: 0.8664 - val_loss: 0.2810 - val_acc: 0.8634\n",
      "Epoch 21/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2905 - acc: 0.8713 - val_loss: 0.2719 - val_acc: 0.8634\n",
      "Epoch 22/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.3043 - acc: 0.8628 - val_loss: 0.2630 - val_acc: 0.8820\n",
      "Epoch 23/50\n",
      "24/24 [==============================] - 14s 578ms/step - loss: 0.3103 - acc: 0.8661 - val_loss: 0.2610 - val_acc: 0.8634\n",
      "Epoch 24/50\n",
      "24/24 [==============================] - 12s 514ms/step - loss: 0.2945 - acc: 0.8701 - val_loss: 0.2867 - val_acc: 0.8571\n",
      "Epoch 25/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2840 - acc: 0.8666 - val_loss: 0.2671 - val_acc: 0.8571\n",
      "Epoch 26/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.2821 - acc: 0.8723 - val_loss: 0.2552 - val_acc: 0.8820\n",
      "Epoch 27/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2952 - acc: 0.8692 - val_loss: 0.2626 - val_acc: 0.8634\n",
      "Epoch 28/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.2743 - acc: 0.8719 - val_loss: 0.2524 - val_acc: 0.8944\n",
      "Epoch 29/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2740 - acc: 0.8821 - val_loss: 0.2525 - val_acc: 0.8758\n",
      "Epoch 30/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.2763 - acc: 0.8804 - val_loss: 0.2514 - val_acc: 0.8758\n",
      "Epoch 31/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2640 - acc: 0.8896 - val_loss: 0.2659 - val_acc: 0.8634\n",
      "Epoch 32/50\n",
      "24/24 [==============================] - 12s 513ms/step - loss: 0.2628 - acc: 0.8890 - val_loss: 0.2539 - val_acc: 0.8820\n",
      "Epoch 33/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2619 - acc: 0.8883 - val_loss: 0.2890 - val_acc: 0.8696\n",
      "Epoch 34/50\n",
      "24/24 [==============================] - 12s 517ms/step - loss: 0.2587 - acc: 0.8924 - val_loss: 0.2626 - val_acc: 0.8571\n",
      "Epoch 35/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2912 - acc: 0.8678 - val_loss: 0.2936 - val_acc: 0.8696\n",
      "Epoch 36/50\n",
      "24/24 [==============================] - 12s 517ms/step - loss: 0.2544 - acc: 0.8878 - val_loss: 0.2653 - val_acc: 0.8758\n",
      "Epoch 37/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.2711 - acc: 0.8842 - val_loss: 0.2509 - val_acc: 0.8944\n",
      "Epoch 38/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2394 - acc: 0.9012 - val_loss: 0.2550 - val_acc: 0.8634\n",
      "Epoch 39/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2537 - acc: 0.8956 - val_loss: 0.2892 - val_acc: 0.8634\n",
      "Epoch 40/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2313 - acc: 0.9010 - val_loss: 0.3299 - val_acc: 0.8447\n",
      "Epoch 41/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.2603 - acc: 0.8915 - val_loss: 0.2411 - val_acc: 0.8820\n",
      "Epoch 42/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2632 - acc: 0.8857 - val_loss: 0.2433 - val_acc: 0.8882\n",
      "Epoch 43/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2394 - acc: 0.9054 - val_loss: 0.2455 - val_acc: 0.8882\n",
      "Epoch 44/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.2506 - acc: 0.8870 - val_loss: 0.2411 - val_acc: 0.8944\n",
      "Epoch 45/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2356 - acc: 0.9066 - val_loss: 0.2534 - val_acc: 0.8696\n",
      "Epoch 46/50\n",
      "24/24 [==============================] - 13s 558ms/step - loss: 0.2302 - acc: 0.8992 - val_loss: 0.2316 - val_acc: 0.8882\n",
      "Epoch 47/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2363 - acc: 0.8946 - val_loss: 0.2367 - val_acc: 0.8944\n",
      "Epoch 48/50\n",
      "24/24 [==============================] - 12s 517ms/step - loss: 0.2460 - acc: 0.8902 - val_loss: 0.2453 - val_acc: 0.8944\n",
      "Epoch 49/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2302 - acc: 0.9046 - val_loss: 0.2393 - val_acc: 0.8944\n",
      "Epoch 50/50\n",
      "24/24 [==============================] - 12s 517ms/step - loss: 0.2461 - acc: 0.8997 - val_loss: 0.2338 - val_acc: 0.8882\n",
      "Train loss: 0.192233548983\n",
      "Train accuracy: 0.92376992377\n",
      "Test loss: 0.231610448801\n",
      "Test accuracy: 0.888198757764\n",
      "\n",
      "===================FOLD= 2\n",
      "Epoch 1/50\n",
      "24/24 [==============================] - 15s 609ms/step - loss: 0.9308 - acc: 0.5330 - val_loss: 0.6268 - val_acc: 0.6273\n",
      "Epoch 2/50\n",
      "24/24 [==============================] - 14s 566ms/step - loss: 0.8200 - acc: 0.5734 - val_loss: 0.5728 - val_acc: 0.6335\n",
      "Epoch 3/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.8143 - acc: 0.5724 - val_loss: 0.5188 - val_acc: 0.7578\n",
      "Epoch 4/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.7440 - acc: 0.6073 - val_loss: 0.4808 - val_acc: 0.7516\n",
      "Epoch 5/50\n",
      "24/24 [==============================] - 14s 566ms/step - loss: 0.6263 - acc: 0.6844 - val_loss: 0.4359 - val_acc: 0.7888\n",
      "Epoch 6/50\n",
      "24/24 [==============================] - 12s 517ms/step - loss: 0.6064 - acc: 0.7005 - val_loss: 0.4450 - val_acc: 0.7329\n",
      "Epoch 7/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.5581 - acc: 0.7256 - val_loss: 0.3812 - val_acc: 0.8447\n",
      "Epoch 8/50\n",
      "24/24 [==============================] - 14s 566ms/step - loss: 0.5031 - acc: 0.7409 - val_loss: 0.3540 - val_acc: 0.8634\n",
      "Epoch 9/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.4718 - acc: 0.7763 - val_loss: 0.3252 - val_acc: 0.8634\n",
      "Epoch 10/50\n",
      "24/24 [==============================] - 14s 566ms/step - loss: 0.4555 - acc: 0.7815 - val_loss: 0.3190 - val_acc: 0.8509\n",
      "Epoch 11/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.4564 - acc: 0.7913 - val_loss: 0.3111 - val_acc: 0.8696\n",
      "Epoch 12/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.4133 - acc: 0.8149 - val_loss: 0.3090 - val_acc: 0.8509\n",
      "Epoch 13/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.4109 - acc: 0.8063 - val_loss: 0.3079 - val_acc: 0.8447\n",
      "Epoch 14/50\n",
      "24/24 [==============================] - 12s 514ms/step - loss: 0.3746 - acc: 0.8354 - val_loss: 0.3525 - val_acc: 0.8261\n",
      "Epoch 15/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.3688 - acc: 0.8336 - val_loss: 0.2825 - val_acc: 0.8634\n",
      "Epoch 16/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.3939 - acc: 0.8341 - val_loss: 0.2854 - val_acc: 0.8571\n",
      "Epoch 17/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3955 - acc: 0.8255 - val_loss: 0.2838 - val_acc: 0.8696\n",
      "Epoch 18/50\n",
      "24/24 [==============================] - 12s 517ms/step - loss: 0.4073 - acc: 0.8171 - val_loss: 0.2914 - val_acc: 0.8571\n",
      "Epoch 19/50\n",
      "24/24 [==============================] - 14s 566ms/step - loss: 0.3609 - acc: 0.8376 - val_loss: 0.2631 - val_acc: 0.8820\n",
      "Epoch 20/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3802 - acc: 0.8268 - val_loss: 0.2636 - val_acc: 0.8758\n",
      "Epoch 21/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3241 - acc: 0.8614 - val_loss: 0.2864 - val_acc: 0.8634\n",
      "Epoch 22/50\n",
      "24/24 [==============================] - 12s 517ms/step - loss: 0.3580 - acc: 0.8504 - val_loss: 0.2684 - val_acc: 0.8571\n",
      "Epoch 23/50\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "24/24 [==============================] - 13s 558ms/step - loss: 0.3360 - acc: 0.8593 - val_loss: 0.2578 - val_acc: 0.8696\n",
      "Epoch 24/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.3313 - acc: 0.8521 - val_loss: 0.2817 - val_acc: 0.8571\n",
      "Epoch 25/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.3190 - acc: 0.8581 - val_loss: 0.2685 - val_acc: 0.8509\n",
      "Epoch 26/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.3429 - acc: 0.8547 - val_loss: 0.2676 - val_acc: 0.8758\n",
      "Epoch 27/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.3352 - acc: 0.8538 - val_loss: 0.2453 - val_acc: 0.8820\n",
      "Epoch 28/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3321 - acc: 0.8420 - val_loss: 0.2673 - val_acc: 0.9006\n",
      "Epoch 29/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.2978 - acc: 0.8676 - val_loss: 0.2394 - val_acc: 0.8696\n",
      "Epoch 30/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3084 - acc: 0.8666 - val_loss: 0.2492 - val_acc: 0.8820\n",
      "Epoch 31/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3245 - acc: 0.8640 - val_loss: 0.2460 - val_acc: 0.8758\n",
      "Epoch 32/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.3082 - acc: 0.8614 - val_loss: 0.2577 - val_acc: 0.8882\n",
      "Epoch 33/50\n",
      "24/24 [==============================] - 12s 514ms/step - loss: 0.2805 - acc: 0.8752 - val_loss: 0.2823 - val_acc: 0.9006\n",
      "Epoch 34/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3188 - acc: 0.8774 - val_loss: 0.2405 - val_acc: 0.8696\n",
      "Epoch 35/50\n",
      "24/24 [==============================] - 12s 518ms/step - loss: 0.3115 - acc: 0.8695 - val_loss: 0.2466 - val_acc: 0.8696\n",
      "Epoch 36/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2942 - acc: 0.8621 - val_loss: 0.2494 - val_acc: 0.8944\n",
      "Epoch 37/50\n",
      "24/24 [==============================] - 14s 567ms/step - loss: 0.2958 - acc: 0.8847 - val_loss: 0.2369 - val_acc: 0.8882\n",
      "Epoch 38/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.3385 - acc: 0.8521 - val_loss: 0.2323 - val_acc: 0.8882\n",
      "Epoch 39/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2726 - acc: 0.8825 - val_loss: 0.2628 - val_acc: 0.9068\n",
      "Epoch 40/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2878 - acc: 0.8870 - val_loss: 0.2493 - val_acc: 0.8944\n",
      "Epoch 41/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2865 - acc: 0.8791 - val_loss: 0.2406 - val_acc: 0.8882\n",
      "Epoch 42/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.2866 - acc: 0.8761 - val_loss: 0.2306 - val_acc: 0.8820\n",
      "Epoch 43/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2726 - acc: 0.8891 - val_loss: 0.2338 - val_acc: 0.8944\n",
      "Epoch 44/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2866 - acc: 0.8780 - val_loss: 0.2324 - val_acc: 0.8882\n",
      "Epoch 45/50\n",
      "24/24 [==============================] - 14s 566ms/step - loss: 0.2601 - acc: 0.8967 - val_loss: 0.2306 - val_acc: 0.9006\n",
      "Epoch 46/50\n",
      "24/24 [==============================] - 12s 510ms/step - loss: 0.2664 - acc: 0.8863 - val_loss: 0.2577 - val_acc: 0.9006\n",
      "Epoch 47/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2570 - acc: 0.8917 - val_loss: 0.2345 - val_acc: 0.9068\n",
      "Epoch 48/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2745 - acc: 0.8790 - val_loss: 0.2310 - val_acc: 0.9130\n",
      "Epoch 49/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.2676 - acc: 0.8967 - val_loss: 0.2230 - val_acc: 0.9006\n",
      "Epoch 50/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2798 - acc: 0.8775 - val_loss: 0.2318 - val_acc: 0.9130\n",
      "Train loss: 0.192361263774\n",
      "Train accuracy: 0.92376992377\n",
      "Test loss: 0.223038307536\n",
      "Test accuracy: 0.900621118012\n",
      "\n",
      "===================FOLD= 3\n",
      "Epoch 1/50\n",
      "24/24 [==============================] - 16s 681ms/step - loss: 0.9396 - acc: 0.5267 - val_loss: 0.6423 - val_acc: 0.5750\n",
      "Epoch 2/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.8261 - acc: 0.5536 - val_loss: 0.5839 - val_acc: 0.6562\n",
      "Epoch 3/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.7854 - acc: 0.5697 - val_loss: 0.5285 - val_acc: 0.7375\n",
      "Epoch 4/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.7774 - acc: 0.5751 - val_loss: 0.4787 - val_acc: 0.7812\n",
      "Epoch 5/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.6621 - acc: 0.6603 - val_loss: 0.4340 - val_acc: 0.7688\n",
      "Epoch 6/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.6134 - acc: 0.6883 - val_loss: 0.3956 - val_acc: 0.7875\n",
      "Epoch 7/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.5551 - acc: 0.7150 - val_loss: 0.3531 - val_acc: 0.8375\n",
      "Epoch 8/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.5243 - acc: 0.7510 - val_loss: 0.3348 - val_acc: 0.8313\n",
      "Epoch 9/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.4534 - acc: 0.7814 - val_loss: 0.3429 - val_acc: 0.8438\n",
      "Epoch 10/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.4814 - acc: 0.7718 - val_loss: 0.2991 - val_acc: 0.8562\n",
      "Epoch 11/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.4502 - acc: 0.8024 - val_loss: 0.2998 - val_acc: 0.8625\n",
      "Epoch 12/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.4158 - acc: 0.8187 - val_loss: 0.2813 - val_acc: 0.8562\n",
      "Epoch 13/50\n",
      "24/24 [==============================] - 12s 514ms/step - loss: 0.3955 - acc: 0.8137 - val_loss: 0.2837 - val_acc: 0.8750\n",
      "Epoch 14/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.4158 - acc: 0.8064 - val_loss: 0.2922 - val_acc: 0.8625\n",
      "Epoch 15/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.4047 - acc: 0.8212 - val_loss: 0.2695 - val_acc: 0.8812\n",
      "Epoch 16/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.3694 - acc: 0.8347 - val_loss: 0.2627 - val_acc: 0.8938\n",
      "Epoch 17/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.3572 - acc: 0.8376 - val_loss: 0.2608 - val_acc: 0.8812\n",
      "Epoch 18/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.4014 - acc: 0.8228 - val_loss: 0.2622 - val_acc: 0.8750\n",
      "Epoch 19/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.3463 - acc: 0.8497 - val_loss: 0.2607 - val_acc: 0.8750\n",
      "Epoch 20/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.3529 - acc: 0.8479 - val_loss: 0.2521 - val_acc: 0.9000\n",
      "Epoch 21/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.3527 - acc: 0.8397 - val_loss: 0.2517 - val_acc: 0.8812\n",
      "Epoch 22/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.3535 - acc: 0.8318 - val_loss: 0.2413 - val_acc: 0.8938\n",
      "Epoch 23/50\n",
      "24/24 [==============================] - 13s 522ms/step - loss: 0.3472 - acc: 0.8401 - val_loss: 0.2598 - val_acc: 0.8750\n",
      "Epoch 24/50\n",
      "24/24 [==============================] - 12s 514ms/step - loss: 0.3408 - acc: 0.8454 - val_loss: 0.2508 - val_acc: 0.9062\n",
      "Epoch 25/50\n",
      "24/24 [==============================] - 12s 517ms/step - loss: 0.3059 - acc: 0.8629 - val_loss: 0.2775 - val_acc: 0.8688\n",
      "Epoch 26/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3114 - acc: 0.8731 - val_loss: 0.2677 - val_acc: 0.8812\n",
      "Epoch 27/50\n",
      "24/24 [==============================] - 14s 566ms/step - loss: 0.3226 - acc: 0.8625 - val_loss: 0.2396 - val_acc: 0.8875\n",
      "Epoch 28/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3040 - acc: 0.8643 - val_loss: 0.2519 - val_acc: 0.9000\n",
      "Epoch 29/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3308 - acc: 0.8661 - val_loss: 0.2408 - val_acc: 0.9062\n",
      "Epoch 30/50\n",
      "24/24 [==============================] - 12s 517ms/step - loss: 0.3245 - acc: 0.8611 - val_loss: 0.2722 - val_acc: 0.8812\n",
      "Epoch 31/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3111 - acc: 0.8697 - val_loss: 0.2417 - val_acc: 0.9125\n",
      "Epoch 32/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2934 - acc: 0.8707 - val_loss: 0.2442 - val_acc: 0.9062\n",
      "Epoch 33/50\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2947 - acc: 0.8653 - val_loss: 0.2801 - val_acc: 0.8812\n",
      "Epoch 34/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3001 - acc: 0.8772 - val_loss: 0.2784 - val_acc: 0.8812\n",
      "Epoch 35/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.2807 - acc: 0.8831 - val_loss: 0.2375 - val_acc: 0.9062\n",
      "Epoch 36/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3021 - acc: 0.8658 - val_loss: 0.3534 - val_acc: 0.8688\n",
      "Epoch 37/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2773 - acc: 0.8748 - val_loss: 0.2641 - val_acc: 0.8750\n",
      "Epoch 38/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.2875 - acc: 0.8721 - val_loss: 0.2309 - val_acc: 0.9187\n",
      "Epoch 39/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2885 - acc: 0.8686 - val_loss: 0.3000 - val_acc: 0.8688\n",
      "Epoch 40/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2917 - acc: 0.8687 - val_loss: 0.2688 - val_acc: 0.9000\n",
      "Epoch 41/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2569 - acc: 0.8920 - val_loss: 0.2397 - val_acc: 0.9062\n",
      "Epoch 42/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2832 - acc: 0.8868 - val_loss: 0.2911 - val_acc: 0.8875\n",
      "Epoch 43/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.3037 - acc: 0.8671 - val_loss: 0.2648 - val_acc: 0.8938\n",
      "Epoch 44/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2679 - acc: 0.8853 - val_loss: 0.2699 - val_acc: 0.8938\n",
      "Epoch 45/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2623 - acc: 0.8938 - val_loss: 0.3041 - val_acc: 0.8750\n",
      "Epoch 46/50\n",
      "24/24 [==============================] - 12s 508ms/step - loss: 0.2838 - acc: 0.8726 - val_loss: 0.2808 - val_acc: 0.8625\n",
      "Epoch 47/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2825 - acc: 0.8731 - val_loss: 0.2529 - val_acc: 0.9000\n",
      "Epoch 48/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2772 - acc: 0.8831 - val_loss: 0.2418 - val_acc: 0.9125\n",
      "Train loss: 0.203669224626\n",
      "Train accuracy: 0.918282548476\n",
      "Test loss: 0.230892792344\n",
      "Test accuracy: 0.91875\n",
      "\n",
      "===================FOLD= 4\n",
      "Epoch 1/50\n",
      "24/24 [==============================] - 15s 631ms/step - loss: 0.7930 - acc: 0.5249 - val_loss: 0.6433 - val_acc: 0.6562\n",
      "Epoch 2/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.6924 - acc: 0.5915 - val_loss: 0.5997 - val_acc: 0.7063\n",
      "Epoch 3/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.6502 - acc: 0.6248 - val_loss: 0.5603 - val_acc: 0.7812\n",
      "Epoch 4/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.6467 - acc: 0.6343 - val_loss: 0.5282 - val_acc: 0.8000\n",
      "Epoch 5/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.5896 - acc: 0.6708 - val_loss: 0.4846 - val_acc: 0.8000\n",
      "Epoch 6/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.5425 - acc: 0.7104 - val_loss: 0.4350 - val_acc: 0.8625\n",
      "Epoch 7/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.5280 - acc: 0.7278 - val_loss: 0.4108 - val_acc: 0.8562\n",
      "Epoch 8/50\n",
      "24/24 [==============================] - 13s 562ms/step - loss: 0.4982 - acc: 0.7383 - val_loss: 0.3671 - val_acc: 0.9062\n",
      "Epoch 9/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.4645 - acc: 0.7684 - val_loss: 0.3394 - val_acc: 0.9062\n",
      "Epoch 10/50\n",
      "24/24 [==============================] - 14s 566ms/step - loss: 0.4455 - acc: 0.7927 - val_loss: 0.3258 - val_acc: 0.8875\n",
      "Epoch 11/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.4249 - acc: 0.7972 - val_loss: 0.3219 - val_acc: 0.8875\n",
      "Epoch 12/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.4120 - acc: 0.8057 - val_loss: 0.3015 - val_acc: 0.9000\n",
      "Epoch 13/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3934 - acc: 0.8163 - val_loss: 0.3078 - val_acc: 0.8938\n",
      "Epoch 14/50\n",
      "24/24 [==============================] - 14s 566ms/step - loss: 0.3536 - acc: 0.8423 - val_loss: 0.2785 - val_acc: 0.9125\n",
      "Epoch 15/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.3565 - acc: 0.8353 - val_loss: 0.2565 - val_acc: 0.9187\n",
      "Epoch 16/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.3502 - acc: 0.8391 - val_loss: 0.2197 - val_acc: 0.9313\n",
      "Epoch 17/50\n",
      "24/24 [==============================] - 13s 562ms/step - loss: 0.3269 - acc: 0.8586 - val_loss: 0.2189 - val_acc: 0.9437\n",
      "Epoch 18/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.3159 - acc: 0.8568 - val_loss: 0.2252 - val_acc: 0.9375\n",
      "Epoch 19/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.3388 - acc: 0.8342 - val_loss: 0.2169 - val_acc: 0.9313\n",
      "Epoch 20/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.3219 - acc: 0.8533 - val_loss: 0.1910 - val_acc: 0.9563\n",
      "Epoch 21/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3023 - acc: 0.8666 - val_loss: 0.2257 - val_acc: 0.9125\n",
      "Epoch 22/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.2991 - acc: 0.8601 - val_loss: 0.1896 - val_acc: 0.9563\n",
      "Epoch 23/50\n",
      "24/24 [==============================] - 12s 510ms/step - loss: 0.3253 - acc: 0.8520 - val_loss: 0.1906 - val_acc: 0.9625\n",
      "Epoch 24/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2958 - acc: 0.8710 - val_loss: 0.2227 - val_acc: 0.9125\n",
      "Epoch 25/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2819 - acc: 0.8718 - val_loss: 0.1910 - val_acc: 0.9625\n",
      "Epoch 26/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.2877 - acc: 0.8790 - val_loss: 0.1702 - val_acc: 0.9500\n",
      "Epoch 27/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3113 - acc: 0.8676 - val_loss: 0.1943 - val_acc: 0.9375\n",
      "Epoch 28/50\n",
      "24/24 [==============================] - 12s 517ms/step - loss: 0.2712 - acc: 0.8769 - val_loss: 0.1801 - val_acc: 0.9500\n",
      "Epoch 29/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.2841 - acc: 0.8728 - val_loss: 0.1648 - val_acc: 0.9750\n",
      "Epoch 30/50\n",
      "24/24 [==============================] - 12s 513ms/step - loss: 0.2651 - acc: 0.8861 - val_loss: 0.1695 - val_acc: 0.9563\n",
      "Epoch 31/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2636 - acc: 0.8839 - val_loss: 0.2002 - val_acc: 0.9187\n",
      "Epoch 32/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2602 - acc: 0.8779 - val_loss: 0.1767 - val_acc: 0.9437\n",
      "Epoch 33/50\n",
      "24/24 [==============================] - 12s 517ms/step - loss: 0.2687 - acc: 0.8816 - val_loss: 0.1857 - val_acc: 0.9375\n",
      "Epoch 34/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2714 - acc: 0.8769 - val_loss: 0.1907 - val_acc: 0.9375\n",
      "Epoch 35/50\n",
      "24/24 [==============================] - 12s 517ms/step - loss: 0.2703 - acc: 0.8717 - val_loss: 0.1691 - val_acc: 0.9500\n",
      "Epoch 36/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.2702 - acc: 0.8905 - val_loss: 0.1622 - val_acc: 0.9500\n",
      "Epoch 37/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.2679 - acc: 0.8814 - val_loss: 0.1583 - val_acc: 0.9688\n",
      "Epoch 38/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2480 - acc: 0.8865 - val_loss: 0.1659 - val_acc: 0.9625\n",
      "Epoch 39/50\n",
      "24/24 [==============================] - 12s 517ms/step - loss: 0.2513 - acc: 0.8922 - val_loss: 0.1870 - val_acc: 0.9375\n",
      "Epoch 40/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2683 - acc: 0.8801 - val_loss: 0.1597 - val_acc: 0.9625\n",
      "Epoch 41/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.2448 - acc: 0.8923 - val_loss: 0.1452 - val_acc: 0.9750\n",
      "Epoch 42/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2518 - acc: 0.8860 - val_loss: 0.1577 - val_acc: 0.9625\n",
      "Epoch 43/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2554 - acc: 0.8813 - val_loss: 0.1769 - val_acc: 0.9437\n",
      "Epoch 44/50\n",
      "24/24 [==============================] - 12s 514ms/step - loss: 0.2498 - acc: 0.8936 - val_loss: 0.1742 - val_acc: 0.9250\n",
      "Epoch 45/50\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2434 - acc: 0.8976 - val_loss: 0.1800 - val_acc: 0.9500\n",
      "Epoch 46/50\n",
      "24/24 [==============================] - 12s 510ms/step - loss: 0.2428 - acc: 0.9006 - val_loss: 0.1674 - val_acc: 0.9438\n",
      "Epoch 47/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2341 - acc: 0.9033 - val_loss: 0.1679 - val_acc: 0.9500\n",
      "Epoch 48/50\n",
      "24/24 [==============================] - 12s 512ms/step - loss: 0.2270 - acc: 0.9065 - val_loss: 0.1455 - val_acc: 0.9625\n",
      "Epoch 49/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2425 - acc: 0.8902 - val_loss: 0.1723 - val_acc: 0.9500\n",
      "Epoch 50/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2448 - acc: 0.8981 - val_loss: 0.1673 - val_acc: 0.9625\n",
      "Train loss: 0.202101165403\n",
      "Train accuracy: 0.918975069252\n",
      "Test loss: 0.145169643313\n",
      "Test accuracy: 0.975\n",
      "\n",
      "===================FOLD= 5\n",
      "Epoch 1/50\n",
      "24/24 [==============================] - 15s 639ms/step - loss: 0.9396 - acc: 0.5277 - val_loss: 0.6526 - val_acc: 0.6000\n",
      "Epoch 2/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.8431 - acc: 0.5632 - val_loss: 0.5775 - val_acc: 0.6562\n",
      "Epoch 3/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.7653 - acc: 0.5980 - val_loss: 0.5305 - val_acc: 0.7125\n",
      "Epoch 4/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.7116 - acc: 0.6300 - val_loss: 0.4777 - val_acc: 0.7562\n",
      "Epoch 5/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.6323 - acc: 0.6767 - val_loss: 0.4432 - val_acc: 0.7750\n",
      "Epoch 6/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.6341 - acc: 0.6730 - val_loss: 0.3918 - val_acc: 0.8187\n",
      "Epoch 7/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.5735 - acc: 0.7236 - val_loss: 0.3541 - val_acc: 0.8438\n",
      "Epoch 8/50\n",
      "24/24 [==============================] - 14s 566ms/step - loss: 0.5393 - acc: 0.7189 - val_loss: 0.3250 - val_acc: 0.8500\n",
      "Epoch 9/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.4527 - acc: 0.7818 - val_loss: 0.3245 - val_acc: 0.8500\n",
      "Epoch 10/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.5032 - acc: 0.7599 - val_loss: 0.2749 - val_acc: 0.8812\n",
      "Epoch 11/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.4739 - acc: 0.7836 - val_loss: 0.2673 - val_acc: 0.8750\n",
      "Epoch 12/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.4588 - acc: 0.7849 - val_loss: 0.2577 - val_acc: 0.8938\n",
      "Epoch 13/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.4276 - acc: 0.8028 - val_loss: 0.2577 - val_acc: 0.8938\n",
      "Epoch 14/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.4223 - acc: 0.8018 - val_loss: 0.2486 - val_acc: 0.8875\n",
      "Epoch 15/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3964 - acc: 0.8124 - val_loss: 0.2821 - val_acc: 0.8875\n",
      "Epoch 16/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.4070 - acc: 0.8233 - val_loss: 0.2345 - val_acc: 0.9062\n",
      "Epoch 17/50\n",
      "24/24 [==============================] - 12s 517ms/step - loss: 0.3808 - acc: 0.8272 - val_loss: 0.2892 - val_acc: 0.8875\n",
      "Epoch 18/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3649 - acc: 0.8376 - val_loss: 0.2577 - val_acc: 0.9000\n",
      "Epoch 19/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.3746 - acc: 0.8305 - val_loss: 0.2304 - val_acc: 0.9062\n",
      "Epoch 20/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3359 - acc: 0.8518 - val_loss: 0.2470 - val_acc: 0.9187\n",
      "Epoch 21/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.3770 - acc: 0.8360 - val_loss: 0.2415 - val_acc: 0.9250\n",
      "Epoch 22/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.3561 - acc: 0.8366 - val_loss: 0.2164 - val_acc: 0.9125\n",
      "Epoch 23/50\n",
      "24/24 [==============================] - 12s 509ms/step - loss: 0.3293 - acc: 0.8475 - val_loss: 0.2361 - val_acc: 0.8938\n",
      "Epoch 24/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.3416 - acc: 0.8436 - val_loss: 0.2099 - val_acc: 0.9250\n",
      "Epoch 25/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.3429 - acc: 0.8406 - val_loss: 0.2230 - val_acc: 0.9250\n",
      "Epoch 26/50\n",
      "24/24 [==============================] - 14s 566ms/step - loss: 0.3230 - acc: 0.8606 - val_loss: 0.2064 - val_acc: 0.9313\n",
      "Epoch 27/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.3065 - acc: 0.8653 - val_loss: 0.2041 - val_acc: 0.9313\n",
      "Epoch 28/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.3201 - acc: 0.8564 - val_loss: 0.2112 - val_acc: 0.9187\n",
      "Epoch 29/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.3260 - acc: 0.8614 - val_loss: 0.2414 - val_acc: 0.9313\n",
      "Epoch 30/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.3264 - acc: 0.8564 - val_loss: 0.2411 - val_acc: 0.9250\n",
      "Epoch 31/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3130 - acc: 0.8681 - val_loss: 0.2078 - val_acc: 0.9062\n",
      "Epoch 32/50\n",
      "24/24 [==============================] - 12s 517ms/step - loss: 0.2914 - acc: 0.8653 - val_loss: 0.2314 - val_acc: 0.9187\n",
      "Epoch 33/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.3443 - acc: 0.8526 - val_loss: 0.2483 - val_acc: 0.9125\n",
      "Epoch 34/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2957 - acc: 0.8733 - val_loss: 0.2360 - val_acc: 0.9250\n",
      "Epoch 35/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.3090 - acc: 0.8536 - val_loss: 0.2236 - val_acc: 0.9313\n",
      "Epoch 36/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.3208 - acc: 0.8606 - val_loss: 0.2021 - val_acc: 0.9375\n",
      "Epoch 37/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2799 - acc: 0.8730 - val_loss: 0.2891 - val_acc: 0.8875\n",
      "Epoch 38/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.3168 - acc: 0.8663 - val_loss: 0.1993 - val_acc: 0.9250\n",
      "Epoch 39/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.3028 - acc: 0.8668 - val_loss: 0.2312 - val_acc: 0.9250\n",
      "Epoch 40/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.3036 - acc: 0.8671 - val_loss: 0.2051 - val_acc: 0.9187\n",
      "Epoch 41/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2497 - acc: 0.8958 - val_loss: 0.2016 - val_acc: 0.9187\n",
      "Epoch 42/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2927 - acc: 0.8759 - val_loss: 0.2004 - val_acc: 0.9187\n",
      "Epoch 43/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2914 - acc: 0.8705 - val_loss: 0.2115 - val_acc: 0.9250\n",
      "Epoch 44/50\n",
      "24/24 [==============================] - 12s 517ms/step - loss: 0.2608 - acc: 0.8845 - val_loss: 0.2045 - val_acc: 0.9187\n",
      "Epoch 45/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2576 - acc: 0.8933 - val_loss: 0.2174 - val_acc: 0.9187\n",
      "Epoch 46/50\n",
      "24/24 [==============================] - 12s 509ms/step - loss: 0.2899 - acc: 0.8739 - val_loss: 0.2187 - val_acc: 0.9250\n",
      "Epoch 47/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.2577 - acc: 0.8907 - val_loss: 0.1951 - val_acc: 0.9187\n",
      "Epoch 48/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2754 - acc: 0.8809 - val_loss: 0.2021 - val_acc: 0.9250\n",
      "Epoch 49/50\n",
      "24/24 [==============================] - 12s 514ms/step - loss: 0.2562 - acc: 0.8881 - val_loss: 0.2447 - val_acc: 0.9125\n",
      "Epoch 50/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2608 - acc: 0.8853 - val_loss: 0.2491 - val_acc: 0.9125\n",
      "Train loss: 0.196741641517\n",
      "Train accuracy: 0.923130193906\n",
      "Test loss: 0.195147423446\n",
      "Test accuracy: 0.91875\n",
      "\n",
      "===================FOLD= 6\n",
      "Epoch 1/50\n",
      "24/24 [==============================] - 16s 650ms/step - loss: 0.9060 - acc: 0.5226 - val_loss: 0.6657 - val_acc: 0.5437\n",
      "Epoch 2/50\n",
      "24/24 [==============================] - 14s 566ms/step - loss: 0.8404 - acc: 0.5541 - val_loss: 0.6123 - val_acc: 0.6125\n",
      "Epoch 3/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.7718 - acc: 0.5925 - val_loss: 0.5752 - val_acc: 0.6750\n",
      "Epoch 4/50\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "24/24 [==============================] - 14s 564ms/step - loss: 0.7148 - acc: 0.6199 - val_loss: 0.5289 - val_acc: 0.7188\n",
      "Epoch 5/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.6676 - acc: 0.6552 - val_loss: 0.4939 - val_acc: 0.7562\n",
      "Epoch 6/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.6122 - acc: 0.6967 - val_loss: 0.4578 - val_acc: 0.7812\n",
      "Epoch 7/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.5450 - acc: 0.7362 - val_loss: 0.4515 - val_acc: 0.8125\n",
      "Epoch 8/50\n",
      "24/24 [==============================] - 14s 566ms/step - loss: 0.5156 - acc: 0.7531 - val_loss: 0.4113 - val_acc: 0.8125\n",
      "Epoch 9/50\n",
      "24/24 [==============================] - 12s 514ms/step - loss: 0.4664 - acc: 0.7747 - val_loss: 0.4148 - val_acc: 0.7875\n",
      "Epoch 10/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.4529 - acc: 0.7819 - val_loss: 0.4109 - val_acc: 0.8063\n",
      "Epoch 11/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.4525 - acc: 0.7902 - val_loss: 0.4071 - val_acc: 0.8250\n",
      "Epoch 12/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.4056 - acc: 0.8109 - val_loss: 0.4170 - val_acc: 0.8063\n",
      "Epoch 13/50\n",
      "24/24 [==============================] - 14s 566ms/step - loss: 0.3898 - acc: 0.8210 - val_loss: 0.3971 - val_acc: 0.8250\n",
      "Epoch 14/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.3693 - acc: 0.8348 - val_loss: 0.4065 - val_acc: 0.8250\n",
      "Epoch 15/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3713 - acc: 0.8378 - val_loss: 0.4008 - val_acc: 0.8125\n",
      "Epoch 16/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.3857 - acc: 0.8311 - val_loss: 0.4115 - val_acc: 0.8125\n",
      "Epoch 17/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.3450 - acc: 0.8513 - val_loss: 0.4083 - val_acc: 0.8313\n",
      "Epoch 18/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.3542 - acc: 0.8466 - val_loss: 0.3871 - val_acc: 0.8500\n",
      "Epoch 19/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.3264 - acc: 0.8547 - val_loss: 0.3809 - val_acc: 0.8500\n",
      "Epoch 20/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.3396 - acc: 0.8523 - val_loss: 0.3852 - val_acc: 0.8438\n",
      "Epoch 21/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3300 - acc: 0.8616 - val_loss: 0.3871 - val_acc: 0.8438\n",
      "Epoch 22/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.3143 - acc: 0.8617 - val_loss: 0.3837 - val_acc: 0.8562\n",
      "Epoch 23/50\n",
      "24/24 [==============================] - 12s 508ms/step - loss: 0.3326 - acc: 0.8580 - val_loss: 0.3833 - val_acc: 0.8563\n",
      "Epoch 24/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.3184 - acc: 0.8656 - val_loss: 0.3898 - val_acc: 0.8438\n",
      "Epoch 25/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.3061 - acc: 0.8715 - val_loss: 0.3736 - val_acc: 0.8562\n",
      "Epoch 26/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.3139 - acc: 0.8630 - val_loss: 0.3689 - val_acc: 0.8688\n",
      "Epoch 27/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.3165 - acc: 0.8616 - val_loss: 0.3720 - val_acc: 0.8625\n",
      "Epoch 28/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3072 - acc: 0.8630 - val_loss: 0.3961 - val_acc: 0.8562\n",
      "Epoch 29/50\n",
      "24/24 [==============================] - 12s 517ms/step - loss: 0.2983 - acc: 0.8796 - val_loss: 0.3716 - val_acc: 0.8688\n",
      "Epoch 30/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2986 - acc: 0.8743 - val_loss: 0.3933 - val_acc: 0.8688\n",
      "Epoch 31/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2894 - acc: 0.8682 - val_loss: 0.3802 - val_acc: 0.8562\n",
      "Epoch 32/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2859 - acc: 0.8736 - val_loss: 0.3806 - val_acc: 0.8625\n",
      "Epoch 33/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.3066 - acc: 0.8640 - val_loss: 0.3539 - val_acc: 0.8750\n",
      "Epoch 34/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2752 - acc: 0.8806 - val_loss: 0.3590 - val_acc: 0.8750\n",
      "Epoch 35/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.3078 - acc: 0.8803 - val_loss: 0.3651 - val_acc: 0.8625\n",
      "Epoch 36/50\n",
      "24/24 [==============================] - 14s 566ms/step - loss: 0.2931 - acc: 0.8736 - val_loss: 0.3496 - val_acc: 0.8562\n",
      "Epoch 37/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2988 - acc: 0.8653 - val_loss: 0.3523 - val_acc: 0.8688\n",
      "Epoch 38/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2837 - acc: 0.8752 - val_loss: 0.3627 - val_acc: 0.8812\n",
      "Epoch 39/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2987 - acc: 0.8671 - val_loss: 0.3738 - val_acc: 0.8688\n",
      "Epoch 40/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2730 - acc: 0.8863 - val_loss: 0.3630 - val_acc: 0.8688\n",
      "Epoch 41/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2647 - acc: 0.8832 - val_loss: 0.3604 - val_acc: 0.8750\n",
      "Epoch 42/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2775 - acc: 0.8816 - val_loss: 0.3553 - val_acc: 0.8812\n",
      "Epoch 43/50\n",
      "24/24 [==============================] - 12s 517ms/step - loss: 0.2854 - acc: 0.8796 - val_loss: 0.3634 - val_acc: 0.8688\n",
      "Epoch 44/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2606 - acc: 0.8798 - val_loss: 0.3597 - val_acc: 0.8688\n",
      "Epoch 45/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2569 - acc: 0.8889 - val_loss: 0.3728 - val_acc: 0.8688\n",
      "Epoch 46/50\n",
      "24/24 [==============================] - 12s 509ms/step - loss: 0.2771 - acc: 0.8709 - val_loss: 0.3773 - val_acc: 0.8688\n",
      "Train loss: 0.202126998176\n",
      "Train accuracy: 0.918282548476\n",
      "Test loss: 0.349606382847\n",
      "Test accuracy: 0.85625\n",
      "\n",
      "===================FOLD= 7\n",
      "Epoch 1/50\n",
      "24/24 [==============================] - 16s 674ms/step - loss: 0.9624 - acc: 0.4867 - val_loss: 0.6995 - val_acc: 0.5625\n",
      "Epoch 2/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.7984 - acc: 0.5404 - val_loss: 0.6262 - val_acc: 0.6312\n",
      "Epoch 3/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.7329 - acc: 0.6016 - val_loss: 0.5911 - val_acc: 0.6375\n",
      "Epoch 4/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.7124 - acc: 0.6042 - val_loss: 0.5359 - val_acc: 0.7063\n",
      "Epoch 5/50\n",
      "24/24 [==============================] - 14s 566ms/step - loss: 0.6456 - acc: 0.6445 - val_loss: 0.4997 - val_acc: 0.7188\n",
      "Epoch 6/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.5841 - acc: 0.6840 - val_loss: 0.4422 - val_acc: 0.7875\n",
      "Epoch 7/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.5493 - acc: 0.7054 - val_loss: 0.3896 - val_acc: 0.8187\n",
      "Epoch 8/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.5216 - acc: 0.7329 - val_loss: 0.3516 - val_acc: 0.8562\n",
      "Epoch 9/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.5051 - acc: 0.7443 - val_loss: 0.3183 - val_acc: 0.8750\n",
      "Epoch 10/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.4582 - acc: 0.7710 - val_loss: 0.3491 - val_acc: 0.8125\n",
      "Epoch 11/50\n",
      "24/24 [==============================] - 14s 566ms/step - loss: 0.4346 - acc: 0.7914 - val_loss: 0.2719 - val_acc: 0.8875\n",
      "Epoch 12/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.4461 - acc: 0.7884 - val_loss: 0.2651 - val_acc: 0.9062\n",
      "Epoch 13/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.3902 - acc: 0.8189 - val_loss: 0.2740 - val_acc: 0.8938\n",
      "Epoch 14/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.3718 - acc: 0.8384 - val_loss: 0.2423 - val_acc: 0.8938\n",
      "Epoch 15/50\n",
      "24/24 [==============================] - 14s 566ms/step - loss: 0.3715 - acc: 0.8283 - val_loss: 0.2382 - val_acc: 0.8625\n",
      "Epoch 16/50\n",
      "24/24 [==============================] - 12s 518ms/step - loss: 0.3746 - acc: 0.8293 - val_loss: 0.2547 - val_acc: 0.8875\n",
      "Epoch 17/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.3442 - acc: 0.8552 - val_loss: 0.2358 - val_acc: 0.8625\n",
      "Epoch 18/50\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "24/24 [==============================] - 14s 564ms/step - loss: 0.3583 - acc: 0.8259 - val_loss: 0.2293 - val_acc: 0.8938\n",
      "Epoch 19/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.3315 - acc: 0.8612 - val_loss: 0.2248 - val_acc: 0.8812\n",
      "Epoch 20/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.3395 - acc: 0.8471 - val_loss: 0.2234 - val_acc: 0.9125\n",
      "Epoch 21/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.3266 - acc: 0.8624 - val_loss: 0.2243 - val_acc: 0.8938\n",
      "Epoch 22/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.3228 - acc: 0.8546 - val_loss: 0.2111 - val_acc: 0.9000\n",
      "Epoch 23/50\n",
      "24/24 [==============================] - 12s 510ms/step - loss: 0.2948 - acc: 0.8761 - val_loss: 0.2224 - val_acc: 0.9125\n",
      "Epoch 24/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3231 - acc: 0.8562 - val_loss: 0.2139 - val_acc: 0.9000\n",
      "Epoch 25/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3235 - acc: 0.8563 - val_loss: 0.2198 - val_acc: 0.9062\n",
      "Epoch 26/50\n",
      "24/24 [==============================] - 12s 514ms/step - loss: 0.3201 - acc: 0.8660 - val_loss: 0.2118 - val_acc: 0.9062\n",
      "Epoch 27/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2807 - acc: 0.8759 - val_loss: 0.2143 - val_acc: 0.9125\n",
      "Epoch 28/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3203 - acc: 0.8625 - val_loss: 0.2212 - val_acc: 0.9000\n",
      "Epoch 29/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3034 - acc: 0.8663 - val_loss: 0.2238 - val_acc: 0.9187\n",
      "Epoch 30/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2987 - acc: 0.8712 - val_loss: 0.2261 - val_acc: 0.9000\n",
      "Epoch 31/50\n",
      "24/24 [==============================] - 12s 518ms/step - loss: 0.2782 - acc: 0.8818 - val_loss: 0.2233 - val_acc: 0.9125\n",
      "Epoch 32/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.2740 - acc: 0.8806 - val_loss: 0.2016 - val_acc: 0.9062\n",
      "Epoch 33/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2852 - acc: 0.8795 - val_loss: 0.2123 - val_acc: 0.9062\n",
      "Epoch 34/50\n",
      "24/24 [==============================] - 12s 518ms/step - loss: 0.2657 - acc: 0.8811 - val_loss: 0.2145 - val_acc: 0.9000\n",
      "Epoch 35/50\n",
      "24/24 [==============================] - 12s 518ms/step - loss: 0.3096 - acc: 0.8671 - val_loss: 0.2086 - val_acc: 0.9187\n",
      "Epoch 36/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2501 - acc: 0.8879 - val_loss: 0.2041 - val_acc: 0.9062\n",
      "Epoch 37/50\n",
      "24/24 [==============================] - 12s 518ms/step - loss: 0.2565 - acc: 0.8923 - val_loss: 0.2110 - val_acc: 0.9125\n",
      "Epoch 38/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.2700 - acc: 0.8831 - val_loss: 0.1994 - val_acc: 0.9187\n",
      "Epoch 39/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2660 - acc: 0.8876 - val_loss: 0.2243 - val_acc: 0.9062\n",
      "Epoch 40/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.2300 - acc: 0.8987 - val_loss: 0.1930 - val_acc: 0.9187\n",
      "Epoch 41/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2581 - acc: 0.8873 - val_loss: 0.2030 - val_acc: 0.9000\n",
      "Epoch 42/50\n",
      "24/24 [==============================] - 12s 513ms/step - loss: 0.2573 - acc: 0.8896 - val_loss: 0.2173 - val_acc: 0.9000\n",
      "Epoch 43/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2502 - acc: 0.8853 - val_loss: 0.2074 - val_acc: 0.9062\n",
      "Epoch 44/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2450 - acc: 0.8935 - val_loss: 0.1973 - val_acc: 0.9187\n",
      "Epoch 45/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2296 - acc: 0.8985 - val_loss: 0.1976 - val_acc: 0.9062\n",
      "Epoch 46/50\n",
      "24/24 [==============================] - 12s 509ms/step - loss: 0.2516 - acc: 0.8925 - val_loss: 0.2326 - val_acc: 0.8938\n",
      "Epoch 47/50\n",
      "24/24 [==============================] - 12s 514ms/step - loss: 0.2574 - acc: 0.8870 - val_loss: 0.2072 - val_acc: 0.9125\n",
      "Epoch 48/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2534 - acc: 0.8932 - val_loss: 0.2100 - val_acc: 0.9062\n",
      "Epoch 49/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2405 - acc: 0.8989 - val_loss: 0.1983 - val_acc: 0.9187\n",
      "Epoch 50/50\n",
      "24/24 [==============================] - 12s 517ms/step - loss: 0.2472 - acc: 0.8979 - val_loss: 0.1937 - val_acc: 0.9062\n",
      "Train loss: 0.209626888356\n",
      "Train accuracy: 0.916897506925\n",
      "Test loss: 0.192986506224\n",
      "Test accuracy: 0.91875\n",
      "\n",
      "===================FOLD= 8\n",
      "Epoch 1/50\n",
      "24/24 [==============================] - 16s 669ms/step - loss: 0.9162 - acc: 0.5244 - val_loss: 0.6506 - val_acc: 0.5625\n",
      "Epoch 2/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.8315 - acc: 0.5663 - val_loss: 0.5931 - val_acc: 0.6500\n",
      "Epoch 3/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.7501 - acc: 0.6065 - val_loss: 0.5359 - val_acc: 0.7000\n",
      "Epoch 4/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.6526 - acc: 0.6524 - val_loss: 0.4964 - val_acc: 0.7188\n",
      "Epoch 5/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.6771 - acc: 0.6704 - val_loss: 0.4623 - val_acc: 0.7438\n",
      "Epoch 6/50\n",
      "24/24 [==============================] - 14s 566ms/step - loss: 0.5824 - acc: 0.7191 - val_loss: 0.4340 - val_acc: 0.7688\n",
      "Epoch 7/50\n",
      "24/24 [==============================] - 14s 566ms/step - loss: 0.5090 - acc: 0.7533 - val_loss: 0.4049 - val_acc: 0.7937\n",
      "Epoch 8/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.4794 - acc: 0.7712 - val_loss: 0.3786 - val_acc: 0.7937\n",
      "Epoch 9/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.4876 - acc: 0.7788 - val_loss: 0.3775 - val_acc: 0.7937\n",
      "Epoch 10/50\n",
      "24/24 [==============================] - 14s 568ms/step - loss: 0.4722 - acc: 0.7901 - val_loss: 0.3657 - val_acc: 0.8000\n",
      "Epoch 11/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.4272 - acc: 0.8010 - val_loss: 0.3724 - val_acc: 0.8063\n",
      "Epoch 12/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.4341 - acc: 0.8127 - val_loss: 0.3769 - val_acc: 0.8125\n",
      "Epoch 13/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.4128 - acc: 0.8127 - val_loss: 0.3700 - val_acc: 0.8000\n",
      "Epoch 14/50\n",
      "24/24 [==============================] - 14s 567ms/step - loss: 0.4138 - acc: 0.7980 - val_loss: 0.3532 - val_acc: 0.8500\n",
      "Epoch 15/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3594 - acc: 0.8397 - val_loss: 0.3617 - val_acc: 0.8438\n",
      "Epoch 16/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3557 - acc: 0.8344 - val_loss: 0.3679 - val_acc: 0.8063\n",
      "Epoch 17/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3463 - acc: 0.8448 - val_loss: 0.3837 - val_acc: 0.8187\n",
      "Epoch 18/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3737 - acc: 0.8505 - val_loss: 0.3824 - val_acc: 0.8000\n",
      "Epoch 19/50\n",
      "24/24 [==============================] - 12s 517ms/step - loss: 0.3486 - acc: 0.8371 - val_loss: 0.3721 - val_acc: 0.8187\n",
      "Epoch 20/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.3540 - acc: 0.8477 - val_loss: 0.3493 - val_acc: 0.8438\n",
      "Epoch 21/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3491 - acc: 0.8510 - val_loss: 0.3589 - val_acc: 0.8250\n",
      "Epoch 22/50\n",
      "24/24 [==============================] - 12s 514ms/step - loss: 0.3153 - acc: 0.8660 - val_loss: 0.3680 - val_acc: 0.8125\n",
      "Epoch 23/50\n",
      "24/24 [==============================] - 13s 557ms/step - loss: 0.3297 - acc: 0.8515 - val_loss: 0.3491 - val_acc: 0.8375\n",
      "Epoch 24/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.3501 - acc: 0.8480 - val_loss: 0.3448 - val_acc: 0.8500\n",
      "Epoch 25/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3179 - acc: 0.8621 - val_loss: 0.3544 - val_acc: 0.8438\n",
      "Epoch 26/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3392 - acc: 0.8544 - val_loss: 0.3715 - val_acc: 0.8250\n",
      "Epoch 27/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2968 - acc: 0.8718 - val_loss: 0.3556 - val_acc: 0.8313\n",
      "Epoch 28/50\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "24/24 [==============================] - 12s 517ms/step - loss: 0.3089 - acc: 0.8612 - val_loss: 0.4135 - val_acc: 0.8125\n",
      "Epoch 29/50\n",
      "24/24 [==============================] - 12s 517ms/step - loss: 0.2863 - acc: 0.8723 - val_loss: 0.3640 - val_acc: 0.8187\n",
      "Epoch 30/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2874 - acc: 0.8723 - val_loss: 0.3834 - val_acc: 0.8187\n",
      "Epoch 31/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2916 - acc: 0.8772 - val_loss: 0.3646 - val_acc: 0.8250\n",
      "Epoch 32/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2780 - acc: 0.8803 - val_loss: 0.3769 - val_acc: 0.8375\n",
      "Epoch 33/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.3227 - acc: 0.8585 - val_loss: 0.3732 - val_acc: 0.8375\n",
      "Epoch 34/50\n",
      "24/24 [==============================] - 12s 517ms/step - loss: 0.2570 - acc: 0.8930 - val_loss: 0.4384 - val_acc: 0.8125\n",
      "Train loss: 0.22365373664\n",
      "Train accuracy: 0.91135734072\n",
      "Test loss: 0.344817435741\n",
      "Test accuracy: 0.85\n",
      "\n",
      "===================FOLD= 9\n",
      "Epoch 1/50\n",
      "24/24 [==============================] - 16s 683ms/step - loss: 0.8628 - acc: 0.5231 - val_loss: 0.6290 - val_acc: 0.6312\n",
      "Epoch 2/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.7105 - acc: 0.5709 - val_loss: 0.5621 - val_acc: 0.7750\n",
      "Epoch 3/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.6724 - acc: 0.6190 - val_loss: 0.4961 - val_acc: 0.8063\n",
      "Epoch 4/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.5818 - acc: 0.6850 - val_loss: 0.4557 - val_acc: 0.7875\n",
      "Epoch 5/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.5728 - acc: 0.6876 - val_loss: 0.4225 - val_acc: 0.8000\n",
      "Epoch 6/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.5008 - acc: 0.7397 - val_loss: 0.3917 - val_acc: 0.8250\n",
      "Epoch 7/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.4755 - acc: 0.7493 - val_loss: 0.3695 - val_acc: 0.8313\n",
      "Epoch 8/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.4291 - acc: 0.7837 - val_loss: 0.3501 - val_acc: 0.8500\n",
      "Epoch 9/50\n",
      "24/24 [==============================] - 13s 562ms/step - loss: 0.3964 - acc: 0.8075 - val_loss: 0.3267 - val_acc: 0.8625\n",
      "Epoch 10/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.3970 - acc: 0.8104 - val_loss: 0.3164 - val_acc: 0.8500\n",
      "Epoch 11/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.3550 - acc: 0.8358 - val_loss: 0.3149 - val_acc: 0.8438\n",
      "Epoch 12/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.3557 - acc: 0.8218 - val_loss: 0.3161 - val_acc: 0.8375\n",
      "Epoch 13/50\n",
      "24/24 [==============================] - 14s 566ms/step - loss: 0.3867 - acc: 0.8212 - val_loss: 0.3042 - val_acc: 0.8750\n",
      "Epoch 14/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.3414 - acc: 0.8366 - val_loss: 0.2976 - val_acc: 0.8562\n",
      "Epoch 15/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.3687 - acc: 0.8279 - val_loss: 0.2910 - val_acc: 0.8875\n",
      "Epoch 16/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.3327 - acc: 0.8419 - val_loss: 0.2870 - val_acc: 0.8625\n",
      "Epoch 17/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.3343 - acc: 0.8513 - val_loss: 0.2833 - val_acc: 0.8875\n",
      "Epoch 18/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.3198 - acc: 0.8519 - val_loss: 0.2908 - val_acc: 0.8625\n",
      "Epoch 19/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.3274 - acc: 0.8511 - val_loss: 0.2796 - val_acc: 0.8938\n",
      "Epoch 20/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.3274 - acc: 0.8542 - val_loss: 0.2745 - val_acc: 0.8625\n",
      "Epoch 21/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.3216 - acc: 0.8539 - val_loss: 0.2702 - val_acc: 0.9000\n",
      "Epoch 22/50\n",
      "24/24 [==============================] - 13s 562ms/step - loss: 0.2929 - acc: 0.8594 - val_loss: 0.2606 - val_acc: 0.9062\n",
      "Epoch 23/50\n",
      "24/24 [==============================] - 12s 508ms/step - loss: 0.2917 - acc: 0.8712 - val_loss: 0.2715 - val_acc: 0.8875\n",
      "Epoch 24/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2919 - acc: 0.8648 - val_loss: 0.2677 - val_acc: 0.9000\n",
      "Epoch 25/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2915 - acc: 0.8642 - val_loss: 0.2736 - val_acc: 0.8875\n",
      "Epoch 26/50\n",
      "24/24 [==============================] - 12s 517ms/step - loss: 0.2959 - acc: 0.8715 - val_loss: 0.2633 - val_acc: 0.8938\n",
      "Epoch 27/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.2789 - acc: 0.8746 - val_loss: 0.2510 - val_acc: 0.9125\n",
      "Epoch 28/50\n",
      "24/24 [==============================] - 12s 513ms/step - loss: 0.2937 - acc: 0.8664 - val_loss: 0.2525 - val_acc: 0.8875\n",
      "Epoch 29/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2696 - acc: 0.8772 - val_loss: 0.2569 - val_acc: 0.8938\n",
      "Epoch 30/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.2673 - acc: 0.8824 - val_loss: 0.2490 - val_acc: 0.8938\n",
      "Epoch 31/50\n",
      "24/24 [==============================] - 12s 514ms/step - loss: 0.2460 - acc: 0.8910 - val_loss: 0.2520 - val_acc: 0.9125\n",
      "Epoch 32/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2910 - acc: 0.8720 - val_loss: 0.2662 - val_acc: 0.8938\n",
      "Epoch 33/50\n",
      "24/24 [==============================] - 12s 517ms/step - loss: 0.2603 - acc: 0.8847 - val_loss: 0.2504 - val_acc: 0.8938\n",
      "Epoch 34/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2553 - acc: 0.8848 - val_loss: 0.2491 - val_acc: 0.9125\n",
      "Epoch 35/50\n",
      "24/24 [==============================] - 13s 562ms/step - loss: 0.2580 - acc: 0.8879 - val_loss: 0.2459 - val_acc: 0.9062\n",
      "Epoch 36/50\n",
      "24/24 [==============================] - 12s 514ms/step - loss: 0.2600 - acc: 0.8865 - val_loss: 0.2491 - val_acc: 0.9125\n",
      "Epoch 37/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2588 - acc: 0.8790 - val_loss: 0.2498 - val_acc: 0.8875\n",
      "Epoch 38/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2575 - acc: 0.8930 - val_loss: 0.2581 - val_acc: 0.8938\n",
      "Epoch 39/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2617 - acc: 0.8778 - val_loss: 0.2565 - val_acc: 0.9062\n",
      "Epoch 40/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2489 - acc: 0.8876 - val_loss: 0.2529 - val_acc: 0.8938\n",
      "Epoch 41/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.2588 - acc: 0.8832 - val_loss: 0.2404 - val_acc: 0.9187\n",
      "Epoch 42/50\n",
      "24/24 [==============================] - 12s 515ms/step - loss: 0.2592 - acc: 0.8857 - val_loss: 0.2517 - val_acc: 0.8875\n",
      "Epoch 43/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.2566 - acc: 0.8832 - val_loss: 0.2398 - val_acc: 0.9062\n",
      "Epoch 44/50\n",
      "24/24 [==============================] - 12s 514ms/step - loss: 0.2489 - acc: 0.8977 - val_loss: 0.2523 - val_acc: 0.8938\n",
      "Epoch 45/50\n",
      "24/24 [==============================] - 14s 564ms/step - loss: 0.2388 - acc: 0.9015 - val_loss: 0.2369 - val_acc: 0.8938\n",
      "Epoch 46/50\n",
      "24/24 [==============================] - 12s 508ms/step - loss: 0.2364 - acc: 0.8965 - val_loss: 0.2371 - val_acc: 0.8938\n",
      "Epoch 47/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2311 - acc: 0.9078 - val_loss: 0.2369 - val_acc: 0.8938\n",
      "Epoch 48/50\n",
      "24/24 [==============================] - 14s 565ms/step - loss: 0.2561 - acc: 0.8849 - val_loss: 0.2365 - val_acc: 0.9125\n",
      "Epoch 49/50\n",
      "24/24 [==============================] - 14s 563ms/step - loss: 0.2186 - acc: 0.9111 - val_loss: 0.2312 - val_acc: 0.9062\n",
      "Epoch 50/50\n",
      "24/24 [==============================] - 12s 516ms/step - loss: 0.2702 - acc: 0.8858 - val_loss: 0.2362 - val_acc: 0.8938\n",
      "Train loss: 0.19480495675\n",
      "Train accuracy: 0.915512465374\n",
      "Test loss: 0.231231492758\n",
      "Test accuracy: 0.90625\n",
      "\n",
      " Train Log Loss Validation=  0.193596459305\n",
      " Test Log Loss Validation=  0.236750899909\n"
     ]
    }
   ],
   "source": [
    "#def myAngleCV(X_train, X_angle, X_test):\n",
    "K=10\n",
    "folds = list(StratifiedKFold(n_splits=K, shuffle=True, random_state=16).split(X_train, target_train))\n",
    "y_test_pred_log = 0\n",
    "y_train_pred_log=0\n",
    "y_valid_pred_log = 0.0*target_train\n",
    "for j, (train_idx, test_idx) in enumerate(folds):\n",
    "    print('\\n===================FOLD=',j)\n",
    "    X_train_cv = X_train[train_idx]\n",
    "    y_train_cv = target_train[train_idx]\n",
    "    X_holdout = X_train[test_idx]\n",
    "    Y_holdout= target_train[test_idx]\n",
    "    \n",
    "    #Angle\n",
    "    X_angle_cv=X_angle[train_idx]\n",
    "    X_angle_hold=X_angle[test_idx]\n",
    "    #define file path and get callbacks\n",
    "    file_path = \"%s_aug_model_weights.hdf5\"%j\n",
    "    callbacks = get_callbacks(filepath=file_path, patience=5)\n",
    "    gen_flow = gen_flow_for_two_inputs(X_train_cv, X_angle_cv, y_train_cv)\n",
    "    galaxyModel= getVggAngleModel()\n",
    "    histr = galaxyModel.fit_generator(\n",
    "            gen_flow,\n",
    "            steps_per_epoch=24,\n",
    "            epochs=50,\n",
    "            shuffle=True,\n",
    "            verbose=1,\n",
    "            validation_data=([X_holdout,X_angle_hold], Y_holdout),\n",
    "            callbacks=callbacks)\n",
    "     #Getting the Best Model\n",
    "    galaxyModel.load_weights(filepath=file_path)\n",
    "    #Getting Training Score\n",
    "    score = galaxyModel.evaluate([X_train_cv,X_angle_cv], y_train_cv, verbose=0)\n",
    "    print('Train loss:', score[0])\n",
    "    print('Train accuracy:', score[1])\n",
    "    #Getting Test Score\n",
    "    score = galaxyModel.evaluate([X_holdout,X_angle_hold], Y_holdout, verbose=0)\n",
    "    print('Test loss:', score[0])\n",
    "    print('Test accuracy:', score[1])\n",
    "\n",
    "    #Getting validation Score.\n",
    "    pred_valid=galaxyModel.predict([X_holdout,X_angle_hold])\n",
    "    y_valid_pred_log[test_idx] = pred_valid.reshape(pred_valid.shape[0])\n",
    "\n",
    "    #Getting Test Scores\n",
    "    temp_test=galaxyModel.predict([X_test, X_test_angle])\n",
    "    y_test_pred_log+=temp_test.reshape(temp_test.shape[0])\n",
    "\n",
    "    #Getting Train Scores\n",
    "    temp_train=galaxyModel.predict([X_train, X_angle])\n",
    "    y_train_pred_log+=temp_train.reshape(temp_train.shape[0])\n",
    "\n",
    "y_test_pred_log=y_test_pred_log/K\n",
    "y_train_pred_log=y_train_pred_log/K\n",
    "\n",
    "print('\\n Train Log Loss Validation= ',log_loss(target_train, y_train_pred_log))\n",
    "print(' Test Log Loss Validation= ',log_loss(target_train, y_valid_pred_log))\n",
    "#return y_test_pred_log"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "#preds=myAngleCV(X_train, X_angle, X_test)\n",
    "preds = y_test_pred_log"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Submission to kaggle\n",
    "submission = pd.DataFrame()\n",
    "submission['id']=test['id']\n",
    "submission['is_iceberg']=preds\n",
    "submission.to_csv('sub_Adadelta_10folds.csv', index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "dict_keys(['acc', 'val_loss', 'loss', 'val_acc'])\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYsAAAEWCAYAAACXGLsWAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n/Rfb4rttyO/5IQAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7ff9650a9898>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEWCAYAAACJ0YulAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nxlT5/63uXc+ExmWPAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7ff9061af7b8>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# list all data in history\n",
    "print(histr.history.keys())\n",
    "# summarize history for accuracy\n",
    "plt.plot(histr.history['acc'])\n",
    "plt.plot(histr.history['val_acc'])\n",
    "plt.title('model accuracy')\n",
    "plt.ylabel('accuracy')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['train', 'test'], loc='upper left')\n",
    "plt.show()\n",
    "# summarize history for loss\n",
    "plt.plot(histr.history['loss'])\n",
    "plt.plot(histr.history['val_loss'])\n",
    "plt.title('model loss')\n",
    "plt.ylabel('loss')\n",
    "plt.xlabel('epoch')\n",
    "plt.legend(['train', 'test'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "---"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.image as mpimg\n",
    "\n",
    "min1 = np.min(X_holdout)\n",
    "max1 = np.max(X_holdout)\n",
    "diff1 = max1 - min1\n",
    "my_sum = lambda x: (x - min1)/diff1\n",
    "\n",
    "X_holdout2 = my_sum(X_holdout)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "13      1\n",
       "14      0\n",
       "38      1\n",
       "43      0\n",
       "44      0\n",
       "80      0\n",
       "104     1\n",
       "127     0\n",
       "137     0\n",
       "138     1\n",
       "141     1\n",
       "167     1\n",
       "195     1\n",
       "199     0\n",
       "207     1\n",
       "212     0\n",
       "213     1\n",
       "221     0\n",
       "240     0\n",
       "256     0\n",
       "267     1\n",
       "274     1\n",
       "276     1\n",
       "285     1\n",
       "290     0\n",
       "294     0\n",
       "300     1\n",
       "307     1\n",
       "311     1\n",
       "318     1\n",
       "       ..\n",
       "1363    1\n",
       "1364    0\n",
       "1373    1\n",
       "1380    1\n",
       "1394    0\n",
       "1411    0\n",
       "1435    0\n",
       "1440    1\n",
       "1445    1\n",
       "1447    1\n",
       "1460    1\n",
       "1470    0\n",
       "1488    0\n",
       "1495    0\n",
       "1504    1\n",
       "1516    0\n",
       "1525    0\n",
       "1533    0\n",
       "1537    0\n",
       "1539    0\n",
       "1541    0\n",
       "1542    0\n",
       "1549    0\n",
       "1552    0\n",
       "1559    0\n",
       "1566    0\n",
       "1567    0\n",
       "1577    0\n",
       "1590    0\n",
       "1602    0\n",
       "Name: is_iceberg, Length: 160, dtype: int64"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "Y_holdout"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x7ff9060904a8>"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAUUAAAEICAYAAADIsubvAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ngg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7ff9650be828>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(X_holdout2[0], cmap='gray', interpolation='nearest');\n",
    "tl = \"Actual label : \" +str(Y_holdout[13])+ \",\"+\" iceberg_probability : \"+str(pred_valid[0])\n",
    "plt.title(tl)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x7ff9060736a0>"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAUUAAAEICAYAAADIsubvAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nX+ohzmDcijZ5AAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7ff9060e6a90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(X_holdout2[2], cmap='gray', interpolation='nearest');\n",
    "tl = \"Actual label : \" +str(Y_holdout[38])+ \",\"+\" iceberg_probability : \"+str(pred_valid[2])\n",
    "plt.title(tl)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x7ff905fd0cc0>"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAUQAAAEICAYAAAAncI3RAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\njhPwCdFxHCfwH0QXzA0MixNoAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7ff906028400>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(X_holdout2[1], cmap='gray', interpolation='nearest');\n",
    "tl = \"Actual label : \" +str(Y_holdout[14])+ \",\"+\" iceberg_probability : \"+str(pred_valid[1])\n",
    "plt.title(tl)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x7ff905f11b70>"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAUQAAAEICAYAAAAncI3RAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\ngg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7ff905f60470>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(X_holdout2[3], cmap='gray', interpolation='nearest');\n",
    "tl = \"Actual label : \" +str(Y_holdout[43])+ \",\"+\" iceberg_probability : \"+str(pred_valid[3])\n",
    "plt.title(tl)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x7ff905e58e48>"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAUUAAAEICAYAAADIsubvAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nCAIQg2IQBAGIQTEIggD8D41Wm1fnBcIIAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7ff905eb62b0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(X_holdout2[6], cmap='gray', interpolation='nearest');\n",
    "tl = \"Actual label : \" +str(Y_holdout[104])+ \",\"+\" iceberg_probability : \"+str(pred_valid[6])\n",
    "plt.title(tl)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.text.Text at 0x7ff905e3e710>"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAUUAAAEICAYAAADIsubvAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n4JOi4ziO4JOi4ziO8F9oQDY5xu/q3AAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7ff905e13ef0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(X_holdout2[9], cmap='gray', interpolation='nearest');\n",
    "tl = \"Actual label : \" +str(Y_holdout[138])+ \",\"+\" iceberg_probability : \"+str(pred_valid[9])\n",
    "plt.title(tl)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.5.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}