ANN_crossvalidated.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Untitled14.ipynb",
      "version": "0.3.2",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "WkKEois8a02Y",
        "colab_type": "code",
        "outputId": "8240807a-5efe-47d8-bf5b-3e20445ecbfd",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "import pandas as pd\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dense\n",
        "import numpy as np\n",
        "from sklearn.preprocessing import StandardScaler\n",
        "from sklearn.preprocessing import OneHotEncoder\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.metrics import confusion_matrix\n",
        "from sklearn import metrics\n",
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "import os\n",
        "from sklearn.model_selection import KFold\n",
        "\n",
        "# fix random seed for reproducibility\n",
        "np.random.seed(7)\n",
        "\n",
        "# import dataset and transform it to a numpy array\n",
        "dataset = pd.read_pickle('fullDataFrame')\n",
        "dataset = dataset.replace(r'\\\\n',' ', regex=True)\n",
        "dataset['label'] = dataset['label'].astype(int)\n",
        "#print(dataset)\n",
        "dataset = dataset.values\n",
        "\n",
        "# split dataset in X (input) and Y (labels)\n",
        "X = dataset[:, 0:36]\n",
        "Y = dataset[:, 37]\n",
        "Y = Y.astype(int)\n",
        "\n",
        "\n",
        "# Feature Scaling (all values between 0.0 and 1.0)\n",
        "sc = StandardScaler(with_mean=True, with_std=True)\n",
        "X = sc.fit_transform(X)\n",
        "X = np.nan_to_num(X)\n",
        "\n",
        "# one hot encoding of labels\n",
        "onehot_encoder = OneHotEncoder(sparse=False)\n",
        "Y = Y.reshape(len(Y), 1)\n",
        "Y = onehot_encoder.fit_transform(Y)\n",
        "\n",
        "# split in train / test set\n",
        "X_train, X_test, y_train, y_test = train_test_split(X, Y, test_size=0.33, random_state=42, shuffle=False)\n",
        "\n",
        "\n",
        "\n",
        "# Cross-validate\n",
        "kf = KFold(5, shuffle=True, random_state=42)\n",
        "    \n",
        "oos_y = []\n",
        "oos_pred = []\n",
        "fold = 0\n",
        "\n",
        "for train, test in kf.split(X):\n",
        "    fold+=1\n",
        "    print(f\"Fold #{fold}\")\n",
        "        \n",
        "    x_train = X[train]\n",
        "    y_train = Y[train]\n",
        "    x_test = X[test]\n",
        "    y_test = Y[test]\n",
        "    \n",
        "    #kernel_regularizer=regularizers.l2(0.01),\n",
        "    \n",
        "    # create model\n",
        "    classes=3\n",
        "    model = Sequential()\n",
        "    model.add(Dense(36, input_dim=36, activation='relu'))\n",
        "    model.add(Dense(18, activation='relu'))\n",
        "    model.add(Dense(9, activation='relu'))\n",
        "    model.add(Dense(classes, activation='softmax'))  # we use Softmax, so the outputs represent accuracies that add up to 1\n",
        "\n",
        "\n",
        "    # Compile model\n",
        "    model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
        "\n",
        "    model.fit(x_train,y_train,validation_data=(x_test,y_test),verbose=1,epochs=100,batch_size=10,shuffle=True)\n",
        "    \n",
        "    pred = model.predict(x_test)\n",
        "    \n",
        "    oos_y.append(y_test)\n",
        "    pred = np.argmax(pred,axis=1) # raw probabilities to chosen class (highest probability)\n",
        "    oos_pred.append(pred)        \n",
        "\n",
        "    # Measure this fold's accuracy\n",
        "    y_compare = np.argmax(y_test,axis=1) # For accuracy calculation\n",
        "    score = metrics.accuracy_score(y_compare, pred)\n",
        "    print(f\"Fold score (accuracy): {score}\")\n",
        "\n",
        "\n",
        "# Build the oos prediction list and calculate the error.\n",
        "oos_y = np.concatenate(oos_y)\n",
        "oos_pred = np.concatenate(oos_pred)\n",
        "oos_y_compare = np.argmax(oos_y,axis=1) # For accuracy calculation\n",
        "\n",
        "score = metrics.accuracy_score(oos_y_compare, oos_pred)\n",
        "print(f\"Final score (accuracy): {score}\") \n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/sklearn/preprocessing/_encoders.py:415: FutureWarning: The handling of integer data will change in version 0.22. Currently, the categories are determined based on the range [0, max(values)], while in the future they will be determined based on the unique values.\n",
            "If you want the future behaviour and silence this warning, you can specify \"categories='auto'\".\n",
            "In case you used a LabelEncoder before this OneHotEncoder to convert the categories to integers, then you can now use the OneHotEncoder directly.\n",
            "  warnings.warn(msg, FutureWarning)\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "Fold #1\n",
            "Train on 5537 samples, validate on 1385 samples\n",
            "Epoch 1/100\n",
            "5537/5537 [==============================] - 2s 436us/step - loss: 0.8275 - acc: 0.6402 - val_loss: 0.7292 - val_acc: 0.7011\n",
            "Epoch 2/100\n",
            "5537/5537 [==============================] - 2s 343us/step - loss: 0.6493 - acc: 0.7379 - val_loss: 0.6206 - val_acc: 0.7502\n",
            "Epoch 3/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.5660 - acc: 0.7818 - val_loss: 0.5613 - val_acc: 0.7711\n",
            "Epoch 4/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.5227 - acc: 0.7992 - val_loss: 0.5553 - val_acc: 0.7588\n",
            "Epoch 5/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.4924 - acc: 0.8095 - val_loss: 0.5310 - val_acc: 0.7935\n",
            "Epoch 6/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.4701 - acc: 0.8163 - val_loss: 0.4823 - val_acc: 0.8043\n",
            "Epoch 7/100\n",
            "5537/5537 [==============================] - 2s 350us/step - loss: 0.4526 - acc: 0.8239 - val_loss: 0.4927 - val_acc: 0.7921\n",
            "Epoch 8/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.4429 - acc: 0.8324 - val_loss: 0.4778 - val_acc: 0.8195\n",
            "Epoch 9/100\n",
            "5537/5537 [==============================] - 2s 340us/step - loss: 0.4280 - acc: 0.8302 - val_loss: 0.4603 - val_acc: 0.8209\n",
            "Epoch 10/100\n",
            "5537/5537 [==============================] - 2s 343us/step - loss: 0.4172 - acc: 0.8378 - val_loss: 0.4523 - val_acc: 0.8173\n",
            "Epoch 11/100\n",
            "5537/5537 [==============================] - 2s 346us/step - loss: 0.4070 - acc: 0.8431 - val_loss: 0.4374 - val_acc: 0.8231\n",
            "Epoch 12/100\n",
            "5537/5537 [==============================] - 2s 349us/step - loss: 0.4001 - acc: 0.8494 - val_loss: 0.4538 - val_acc: 0.8390\n",
            "Epoch 13/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.3979 - acc: 0.8470 - val_loss: 0.4709 - val_acc: 0.8310\n",
            "Epoch 14/100\n",
            "5537/5537 [==============================] - 2s 346us/step - loss: 0.3869 - acc: 0.8490 - val_loss: 0.4304 - val_acc: 0.8375\n",
            "Epoch 15/100\n",
            "5537/5537 [==============================] - 2s 346us/step - loss: 0.3736 - acc: 0.8589 - val_loss: 0.4550 - val_acc: 0.8361\n",
            "Epoch 16/100\n",
            "5537/5537 [==============================] - 2s 347us/step - loss: 0.3735 - acc: 0.8543 - val_loss: 0.4182 - val_acc: 0.8484\n",
            "Epoch 17/100\n",
            "5537/5537 [==============================] - 2s 344us/step - loss: 0.3690 - acc: 0.8586 - val_loss: 0.4284 - val_acc: 0.8375\n",
            "Epoch 18/100\n",
            "5537/5537 [==============================] - 2s 347us/step - loss: 0.3697 - acc: 0.8553 - val_loss: 0.4193 - val_acc: 0.8520\n",
            "Epoch 19/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.3633 - acc: 0.8589 - val_loss: 0.4343 - val_acc: 0.8426\n",
            "Epoch 20/100\n",
            "5537/5537 [==============================] - 2s 347us/step - loss: 0.3543 - acc: 0.8617 - val_loss: 0.4450 - val_acc: 0.8433\n",
            "Epoch 21/100\n",
            "5537/5537 [==============================] - 2s 344us/step - loss: 0.3539 - acc: 0.8647 - val_loss: 0.4070 - val_acc: 0.8513\n",
            "Epoch 22/100\n",
            "5537/5537 [==============================] - 2s 344us/step - loss: 0.3452 - acc: 0.8629 - val_loss: 0.4110 - val_acc: 0.8469\n",
            "Epoch 23/100\n",
            "5537/5537 [==============================] - 2s 348us/step - loss: 0.3433 - acc: 0.8662 - val_loss: 0.4324 - val_acc: 0.8527\n",
            "Epoch 24/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.3398 - acc: 0.8673 - val_loss: 0.4232 - val_acc: 0.8383\n",
            "Epoch 25/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.3348 - acc: 0.8718 - val_loss: 0.4027 - val_acc: 0.8527\n",
            "Epoch 26/100\n",
            "5537/5537 [==============================] - 2s 350us/step - loss: 0.3319 - acc: 0.8718 - val_loss: 0.4048 - val_acc: 0.8563\n",
            "Epoch 27/100\n",
            "5537/5537 [==============================] - 2s 349us/step - loss: 0.3287 - acc: 0.8743 - val_loss: 0.4092 - val_acc: 0.8520\n",
            "Epoch 28/100\n",
            "5537/5537 [==============================] - 2s 359us/step - loss: 0.3225 - acc: 0.8700 - val_loss: 0.4069 - val_acc: 0.8426\n",
            "Epoch 29/100\n",
            "5537/5537 [==============================] - 2s 355us/step - loss: 0.3235 - acc: 0.8748 - val_loss: 0.3939 - val_acc: 0.8563\n",
            "Epoch 30/100\n",
            "5537/5537 [==============================] - 2s 358us/step - loss: 0.3203 - acc: 0.8741 - val_loss: 0.4120 - val_acc: 0.8440\n",
            "Epoch 31/100\n",
            "5537/5537 [==============================] - 2s 355us/step - loss: 0.3188 - acc: 0.8743 - val_loss: 0.4078 - val_acc: 0.8484\n",
            "Epoch 32/100\n",
            "5537/5537 [==============================] - 2s 356us/step - loss: 0.3187 - acc: 0.8757 - val_loss: 0.4106 - val_acc: 0.8556\n",
            "Epoch 33/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.3127 - acc: 0.8781 - val_loss: 0.3940 - val_acc: 0.8556\n",
            "Epoch 34/100\n",
            "5537/5537 [==============================] - 2s 347us/step - loss: 0.3133 - acc: 0.8790 - val_loss: 0.3961 - val_acc: 0.8484\n",
            "Epoch 35/100\n",
            "5537/5537 [==============================] - 2s 344us/step - loss: 0.3085 - acc: 0.8806 - val_loss: 0.4234 - val_acc: 0.8477\n",
            "Epoch 36/100\n",
            "5537/5537 [==============================] - 2s 344us/step - loss: 0.3090 - acc: 0.8790 - val_loss: 0.4287 - val_acc: 0.8347\n",
            "Epoch 37/100\n",
            "5537/5537 [==============================] - 2s 346us/step - loss: 0.3083 - acc: 0.8799 - val_loss: 0.3958 - val_acc: 0.8606\n",
            "Epoch 38/100\n",
            "5537/5537 [==============================] - 2s 350us/step - loss: 0.3055 - acc: 0.8830 - val_loss: 0.4106 - val_acc: 0.8534\n",
            "Epoch 39/100\n",
            "5537/5537 [==============================] - 2s 342us/step - loss: 0.3023 - acc: 0.8822 - val_loss: 0.4030 - val_acc: 0.8469\n",
            "Epoch 40/100\n",
            "5537/5537 [==============================] - 2s 339us/step - loss: 0.2991 - acc: 0.8810 - val_loss: 0.3967 - val_acc: 0.8643\n",
            "Epoch 41/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.2966 - acc: 0.8837 - val_loss: 0.4359 - val_acc: 0.8433\n",
            "Epoch 42/100\n",
            "5537/5537 [==============================] - 2s 343us/step - loss: 0.3001 - acc: 0.8810 - val_loss: 0.3923 - val_acc: 0.8628\n",
            "Epoch 43/100\n",
            "5537/5537 [==============================] - 2s 346us/step - loss: 0.3002 - acc: 0.8817 - val_loss: 0.3859 - val_acc: 0.8650\n",
            "Epoch 44/100\n",
            "5537/5537 [==============================] - 2s 346us/step - loss: 0.2971 - acc: 0.8826 - val_loss: 0.3855 - val_acc: 0.8650\n",
            "Epoch 45/100\n",
            "5537/5537 [==============================] - 2s 348us/step - loss: 0.2899 - acc: 0.8877 - val_loss: 0.4191 - val_acc: 0.8412\n",
            "Epoch 46/100\n",
            "5537/5537 [==============================] - 2s 342us/step - loss: 0.2904 - acc: 0.8844 - val_loss: 0.3869 - val_acc: 0.8462\n",
            "Epoch 47/100\n",
            "5537/5537 [==============================] - 2s 341us/step - loss: 0.2890 - acc: 0.8855 - val_loss: 0.3827 - val_acc: 0.8563\n",
            "Epoch 48/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.2887 - acc: 0.8875 - val_loss: 0.4053 - val_acc: 0.8448\n",
            "Epoch 49/100\n",
            "5537/5537 [==============================] - 2s 346us/step - loss: 0.2863 - acc: 0.8886 - val_loss: 0.4055 - val_acc: 0.8527\n",
            "Epoch 50/100\n",
            "5537/5537 [==============================] - 2s 341us/step - loss: 0.2876 - acc: 0.8893 - val_loss: 0.3918 - val_acc: 0.8614\n",
            "Epoch 51/100\n",
            "5537/5537 [==============================] - 2s 343us/step - loss: 0.2820 - acc: 0.8869 - val_loss: 0.3731 - val_acc: 0.8635\n",
            "Epoch 52/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.2887 - acc: 0.8887 - val_loss: 0.3869 - val_acc: 0.8585\n",
            "Epoch 53/100\n",
            "5537/5537 [==============================] - 2s 346us/step - loss: 0.2804 - acc: 0.8904 - val_loss: 0.4051 - val_acc: 0.8534\n",
            "Epoch 54/100\n",
            "5537/5537 [==============================] - 2s 341us/step - loss: 0.2786 - acc: 0.8922 - val_loss: 0.3869 - val_acc: 0.8513\n",
            "Epoch 55/100\n",
            "5537/5537 [==============================] - 2s 344us/step - loss: 0.2791 - acc: 0.8895 - val_loss: 0.4414 - val_acc: 0.8440\n",
            "Epoch 56/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.2802 - acc: 0.8902 - val_loss: 0.4059 - val_acc: 0.8527\n",
            "Epoch 57/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.2766 - acc: 0.8907 - val_loss: 0.3911 - val_acc: 0.8621\n",
            "Epoch 58/100\n",
            "5537/5537 [==============================] - 2s 347us/step - loss: 0.2759 - acc: 0.8918 - val_loss: 0.4227 - val_acc: 0.8477\n",
            "Epoch 59/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.2715 - acc: 0.8942 - val_loss: 0.4088 - val_acc: 0.8462\n",
            "Epoch 60/100\n",
            "5537/5537 [==============================] - 2s 344us/step - loss: 0.2751 - acc: 0.8911 - val_loss: 0.3903 - val_acc: 0.8599\n",
            "Epoch 61/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.2700 - acc: 0.8882 - val_loss: 0.3850 - val_acc: 0.8657\n",
            "Epoch 62/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.2673 - acc: 0.8958 - val_loss: 0.4174 - val_acc: 0.8491\n",
            "Epoch 63/100\n",
            "5537/5537 [==============================] - 2s 347us/step - loss: 0.2656 - acc: 0.8947 - val_loss: 0.3952 - val_acc: 0.8563\n",
            "Epoch 64/100\n",
            "5537/5537 [==============================] - 2s 343us/step - loss: 0.2698 - acc: 0.8940 - val_loss: 0.4025 - val_acc: 0.8599\n",
            "Epoch 65/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.2640 - acc: 0.8998 - val_loss: 0.4028 - val_acc: 0.8585\n",
            "Epoch 66/100\n",
            "5537/5537 [==============================] - 2s 342us/step - loss: 0.2671 - acc: 0.8920 - val_loss: 0.3969 - val_acc: 0.8549\n",
            "Epoch 67/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.2670 - acc: 0.8949 - val_loss: 0.4025 - val_acc: 0.8534\n",
            "Epoch 68/100\n",
            "5537/5537 [==============================] - 2s 342us/step - loss: 0.2672 - acc: 0.8922 - val_loss: 0.3981 - val_acc: 0.8578\n",
            "Epoch 69/100\n",
            "5537/5537 [==============================] - 2s 348us/step - loss: 0.2621 - acc: 0.8953 - val_loss: 0.3997 - val_acc: 0.8563\n",
            "Epoch 70/100\n",
            "5537/5537 [==============================] - 2s 348us/step - loss: 0.2654 - acc: 0.8963 - val_loss: 0.4055 - val_acc: 0.8563\n",
            "Epoch 71/100\n",
            "5537/5537 [==============================] - 2s 346us/step - loss: 0.2680 - acc: 0.8983 - val_loss: 0.3863 - val_acc: 0.8657\n",
            "Epoch 72/100\n",
            "5537/5537 [==============================] - 2s 346us/step - loss: 0.2619 - acc: 0.8987 - val_loss: 0.3812 - val_acc: 0.8664\n",
            "Epoch 73/100\n",
            "5537/5537 [==============================] - 2s 346us/step - loss: 0.2598 - acc: 0.8963 - val_loss: 0.4025 - val_acc: 0.8542\n",
            "Epoch 74/100\n",
            "5537/5537 [==============================] - 2s 344us/step - loss: 0.2601 - acc: 0.8972 - val_loss: 0.4286 - val_acc: 0.8505\n",
            "Epoch 75/100\n",
            "5537/5537 [==============================] - 2s 343us/step - loss: 0.2585 - acc: 0.8969 - val_loss: 0.3922 - val_acc: 0.8643\n",
            "Epoch 76/100\n",
            "5537/5537 [==============================] - 2s 342us/step - loss: 0.2570 - acc: 0.8996 - val_loss: 0.4204 - val_acc: 0.8477\n",
            "Epoch 77/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.2529 - acc: 0.8989 - val_loss: 0.4169 - val_acc: 0.8556\n",
            "Epoch 78/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.2535 - acc: 0.9041 - val_loss: 0.3945 - val_acc: 0.8635\n",
            "Epoch 79/100\n",
            "5537/5537 [==============================] - 2s 340us/step - loss: 0.2548 - acc: 0.9021 - val_loss: 0.3770 - val_acc: 0.8650\n",
            "Epoch 80/100\n",
            "5537/5537 [==============================] - 2s 343us/step - loss: 0.2527 - acc: 0.9012 - val_loss: 0.3912 - val_acc: 0.8570\n",
            "Epoch 81/100\n",
            "5537/5537 [==============================] - 2s 343us/step - loss: 0.2543 - acc: 0.9032 - val_loss: 0.3846 - val_acc: 0.8578\n",
            "Epoch 82/100\n",
            "5537/5537 [==============================] - 2s 340us/step - loss: 0.2498 - acc: 0.8996 - val_loss: 0.3988 - val_acc: 0.8635\n",
            "Epoch 83/100\n",
            "5537/5537 [==============================] - 2s 349us/step - loss: 0.2521 - acc: 0.8987 - val_loss: 0.3765 - val_acc: 0.8679\n",
            "Epoch 84/100\n",
            "5537/5537 [==============================] - 2s 342us/step - loss: 0.2558 - acc: 0.8980 - val_loss: 0.3807 - val_acc: 0.8621\n",
            "Epoch 85/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.2522 - acc: 0.8974 - val_loss: 0.3905 - val_acc: 0.8621\n",
            "Epoch 86/100\n",
            "5537/5537 [==============================] - 2s 340us/step - loss: 0.2475 - acc: 0.9023 - val_loss: 0.3992 - val_acc: 0.8635\n",
            "Epoch 87/100\n",
            "5537/5537 [==============================] - 2s 341us/step - loss: 0.2450 - acc: 0.9074 - val_loss: 0.4173 - val_acc: 0.8498\n",
            "Epoch 88/100\n",
            "5537/5537 [==============================] - 2s 337us/step - loss: 0.2439 - acc: 0.9057 - val_loss: 0.3946 - val_acc: 0.8578\n",
            "Epoch 89/100\n",
            "5537/5537 [==============================] - 2s 340us/step - loss: 0.2475 - acc: 0.9027 - val_loss: 0.3955 - val_acc: 0.8570\n",
            "Epoch 90/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.2475 - acc: 0.9028 - val_loss: 0.3844 - val_acc: 0.8635\n",
            "Epoch 91/100\n",
            "5537/5537 [==============================] - 2s 339us/step - loss: 0.2465 - acc: 0.9028 - val_loss: 0.3894 - val_acc: 0.8606\n",
            "Epoch 92/100\n",
            "5537/5537 [==============================] - 2s 337us/step - loss: 0.2423 - acc: 0.9023 - val_loss: 0.4300 - val_acc: 0.8404\n",
            "Epoch 93/100\n",
            "5537/5537 [==============================] - 2s 343us/step - loss: 0.2478 - acc: 0.8992 - val_loss: 0.3930 - val_acc: 0.8621\n",
            "Epoch 94/100\n",
            "5537/5537 [==============================] - 2s 339us/step - loss: 0.2392 - acc: 0.9061 - val_loss: 0.3905 - val_acc: 0.8671\n",
            "Epoch 95/100\n",
            "5537/5537 [==============================] - 2s 343us/step - loss: 0.2447 - acc: 0.9075 - val_loss: 0.3959 - val_acc: 0.8614\n",
            "Epoch 96/100\n",
            "5537/5537 [==============================] - 2s 344us/step - loss: 0.2384 - acc: 0.9059 - val_loss: 0.3829 - val_acc: 0.8657\n",
            "Epoch 97/100\n",
            "5537/5537 [==============================] - 2s 344us/step - loss: 0.2369 - acc: 0.9072 - val_loss: 0.4147 - val_acc: 0.8498\n",
            "Epoch 98/100\n",
            "5537/5537 [==============================] - 2s 343us/step - loss: 0.2449 - acc: 0.9016 - val_loss: 0.3850 - val_acc: 0.8715\n",
            "Epoch 99/100\n",
            "5537/5537 [==============================] - 2s 344us/step - loss: 0.2342 - acc: 0.9093 - val_loss: 0.4245 - val_acc: 0.8592\n",
            "Epoch 100/100\n",
            "5537/5537 [==============================] - 2s 347us/step - loss: 0.2404 - acc: 0.9063 - val_loss: 0.3820 - val_acc: 0.8671\n",
            "Fold score (accuracy): 0.8671480144404332\n",
            "Fold #2\n",
            "Train on 5537 samples, validate on 1385 samples\n",
            "Epoch 1/100\n",
            "5537/5537 [==============================] - 2s 404us/step - loss: 0.8389 - acc: 0.6218 - val_loss: 0.7297 - val_acc: 0.6939\n",
            "Epoch 2/100\n",
            "5537/5537 [==============================] - 2s 350us/step - loss: 0.6916 - acc: 0.7053 - val_loss: 0.6310 - val_acc: 0.7502\n",
            "Epoch 3/100\n",
            "5537/5537 [==============================] - 2s 347us/step - loss: 0.5879 - acc: 0.7674 - val_loss: 0.5406 - val_acc: 0.8007\n",
            "Epoch 4/100\n",
            "5537/5537 [==============================] - 2s 344us/step - loss: 0.5225 - acc: 0.7932 - val_loss: 0.5347 - val_acc: 0.7921\n",
            "Epoch 5/100\n",
            "5537/5537 [==============================] - 2s 342us/step - loss: 0.4887 - acc: 0.8053 - val_loss: 0.4806 - val_acc: 0.8245\n",
            "Epoch 6/100\n",
            "5537/5537 [==============================] - 2s 344us/step - loss: 0.4644 - acc: 0.8149 - val_loss: 0.4707 - val_acc: 0.8159\n",
            "Epoch 7/100\n",
            "5537/5537 [==============================] - 2s 346us/step - loss: 0.4434 - acc: 0.8223 - val_loss: 0.4704 - val_acc: 0.8238\n",
            "Epoch 8/100\n",
            "5537/5537 [==============================] - 2s 348us/step - loss: 0.4259 - acc: 0.8416 - val_loss: 0.4932 - val_acc: 0.7993\n",
            "Epoch 9/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.4174 - acc: 0.8346 - val_loss: 0.4511 - val_acc: 0.8368\n",
            "Epoch 10/100\n",
            "5537/5537 [==============================] - 2s 348us/step - loss: 0.4081 - acc: 0.8402 - val_loss: 0.4360 - val_acc: 0.8469\n",
            "Epoch 11/100\n",
            "5537/5537 [==============================] - 2s 344us/step - loss: 0.3954 - acc: 0.8456 - val_loss: 0.4380 - val_acc: 0.8390\n",
            "Epoch 12/100\n",
            "5537/5537 [==============================] - 2s 348us/step - loss: 0.3923 - acc: 0.8459 - val_loss: 0.4630 - val_acc: 0.8274\n",
            "Epoch 13/100\n",
            "5537/5537 [==============================] - 2s 346us/step - loss: 0.3827 - acc: 0.8517 - val_loss: 0.4194 - val_acc: 0.8534\n",
            "Epoch 14/100\n",
            "5537/5537 [==============================] - 2s 346us/step - loss: 0.3716 - acc: 0.8566 - val_loss: 0.4457 - val_acc: 0.8412\n",
            "Epoch 15/100\n",
            "5537/5537 [==============================] - 2s 348us/step - loss: 0.3717 - acc: 0.8541 - val_loss: 0.4229 - val_acc: 0.8491\n",
            "Epoch 16/100\n",
            "5537/5537 [==============================] - 2s 346us/step - loss: 0.3713 - acc: 0.8552 - val_loss: 0.4267 - val_acc: 0.8477\n",
            "Epoch 17/100\n",
            "5537/5537 [==============================] - 2s 346us/step - loss: 0.3612 - acc: 0.8613 - val_loss: 0.4443 - val_acc: 0.8354\n",
            "Epoch 18/100\n",
            "5537/5537 [==============================] - 2s 341us/step - loss: 0.3567 - acc: 0.8644 - val_loss: 0.4138 - val_acc: 0.8534\n",
            "Epoch 19/100\n",
            "5537/5537 [==============================] - 2s 346us/step - loss: 0.3573 - acc: 0.8613 - val_loss: 0.4370 - val_acc: 0.8455\n",
            "Epoch 20/100\n",
            "5537/5537 [==============================] - 2s 343us/step - loss: 0.3497 - acc: 0.8676 - val_loss: 0.4161 - val_acc: 0.8556\n",
            "Epoch 21/100\n",
            "5537/5537 [==============================] - 2s 347us/step - loss: 0.3525 - acc: 0.8644 - val_loss: 0.4017 - val_acc: 0.8563\n",
            "Epoch 22/100\n",
            "5537/5537 [==============================] - 2s 347us/step - loss: 0.3449 - acc: 0.8651 - val_loss: 0.3997 - val_acc: 0.8542\n",
            "Epoch 23/100\n",
            "5537/5537 [==============================] - 2s 342us/step - loss: 0.3477 - acc: 0.8656 - val_loss: 0.4125 - val_acc: 0.8513\n",
            "Epoch 24/100\n",
            "5537/5537 [==============================] - 2s 346us/step - loss: 0.3360 - acc: 0.8714 - val_loss: 0.3999 - val_acc: 0.8635\n",
            "Epoch 25/100\n",
            "5537/5537 [==============================] - 2s 341us/step - loss: 0.3360 - acc: 0.8703 - val_loss: 0.3909 - val_acc: 0.8549\n",
            "Epoch 26/100\n",
            "5537/5537 [==============================] - 2s 340us/step - loss: 0.3348 - acc: 0.8716 - val_loss: 0.4039 - val_acc: 0.8462\n",
            "Epoch 27/100\n",
            "5537/5537 [==============================] - 2s 340us/step - loss: 0.3349 - acc: 0.8687 - val_loss: 0.4125 - val_acc: 0.8534\n",
            "Epoch 28/100\n",
            "5537/5537 [==============================] - 2s 343us/step - loss: 0.3286 - acc: 0.8712 - val_loss: 0.4086 - val_acc: 0.8513\n",
            "Epoch 29/100\n",
            "5537/5537 [==============================] - 2s 340us/step - loss: 0.3247 - acc: 0.8729 - val_loss: 0.4012 - val_acc: 0.8585\n",
            "Epoch 30/100\n",
            "5537/5537 [==============================] - 2s 344us/step - loss: 0.3220 - acc: 0.8752 - val_loss: 0.4087 - val_acc: 0.8469\n",
            "Epoch 31/100\n",
            "5537/5537 [==============================] - 2s 343us/step - loss: 0.3230 - acc: 0.8732 - val_loss: 0.4069 - val_acc: 0.8513\n",
            "Epoch 32/100\n",
            "5537/5537 [==============================] - 2s 342us/step - loss: 0.3169 - acc: 0.8732 - val_loss: 0.3833 - val_acc: 0.8650\n",
            "Epoch 33/100\n",
            "5537/5537 [==============================] - 2s 343us/step - loss: 0.3162 - acc: 0.8736 - val_loss: 0.3974 - val_acc: 0.8679\n",
            "Epoch 34/100\n",
            "5537/5537 [==============================] - 2s 341us/step - loss: 0.3121 - acc: 0.8743 - val_loss: 0.4072 - val_acc: 0.8585\n",
            "Epoch 35/100\n",
            "5537/5537 [==============================] - 2s 343us/step - loss: 0.3093 - acc: 0.8817 - val_loss: 0.4091 - val_acc: 0.8592\n",
            "Epoch 36/100\n",
            "5537/5537 [==============================] - 2s 343us/step - loss: 0.3124 - acc: 0.8752 - val_loss: 0.4045 - val_acc: 0.8563\n",
            "Epoch 37/100\n",
            "5537/5537 [==============================] - 2s 344us/step - loss: 0.3049 - acc: 0.8808 - val_loss: 0.3978 - val_acc: 0.8570\n",
            "Epoch 38/100\n",
            "5537/5537 [==============================] - 2s 348us/step - loss: 0.3084 - acc: 0.8779 - val_loss: 0.3887 - val_acc: 0.8621\n",
            "Epoch 39/100\n",
            "5537/5537 [==============================] - 2s 342us/step - loss: 0.3055 - acc: 0.8788 - val_loss: 0.4102 - val_acc: 0.8462\n",
            "Epoch 40/100\n",
            "5537/5537 [==============================] - 2s 341us/step - loss: 0.2998 - acc: 0.8828 - val_loss: 0.4034 - val_acc: 0.8570\n",
            "Epoch 41/100\n",
            "5537/5537 [==============================] - 2s 349us/step - loss: 0.3123 - acc: 0.8765 - val_loss: 0.4060 - val_acc: 0.8592\n",
            "Epoch 42/100\n",
            "5537/5537 [==============================] - 2s 350us/step - loss: 0.2986 - acc: 0.8776 - val_loss: 0.4095 - val_acc: 0.8491\n",
            "Epoch 43/100\n",
            "5537/5537 [==============================] - 2s 355us/step - loss: 0.2973 - acc: 0.8795 - val_loss: 0.4023 - val_acc: 0.8549\n",
            "Epoch 44/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.2975 - acc: 0.8806 - val_loss: 0.3918 - val_acc: 0.8708\n",
            "Epoch 45/100\n",
            "5537/5537 [==============================] - 2s 348us/step - loss: 0.2958 - acc: 0.8828 - val_loss: 0.4276 - val_acc: 0.8455\n",
            "Epoch 46/100\n",
            "5537/5537 [==============================] - 2s 342us/step - loss: 0.2968 - acc: 0.8795 - val_loss: 0.4243 - val_acc: 0.8556\n",
            "Epoch 47/100\n",
            "5537/5537 [==============================] - 2s 347us/step - loss: 0.2974 - acc: 0.8747 - val_loss: 0.4051 - val_acc: 0.8614\n",
            "Epoch 48/100\n",
            "5537/5537 [==============================] - 2s 342us/step - loss: 0.2928 - acc: 0.8813 - val_loss: 0.4189 - val_acc: 0.8563\n",
            "Epoch 49/100\n",
            "5537/5537 [==============================] - 2s 347us/step - loss: 0.2890 - acc: 0.8831 - val_loss: 0.3930 - val_acc: 0.8650\n",
            "Epoch 50/100\n",
            "5537/5537 [==============================] - 2s 346us/step - loss: 0.2920 - acc: 0.8819 - val_loss: 0.3932 - val_acc: 0.8635\n",
            "Epoch 51/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.2838 - acc: 0.8850 - val_loss: 0.4297 - val_acc: 0.8404\n",
            "Epoch 52/100\n",
            "5537/5537 [==============================] - 2s 346us/step - loss: 0.2850 - acc: 0.8859 - val_loss: 0.4239 - val_acc: 0.8578\n",
            "Epoch 53/100\n",
            "5537/5537 [==============================] - 2s 351us/step - loss: 0.2825 - acc: 0.8880 - val_loss: 0.4270 - val_acc: 0.8570\n",
            "Epoch 54/100\n",
            "5537/5537 [==============================] - 2s 351us/step - loss: 0.2881 - acc: 0.8828 - val_loss: 0.3856 - val_acc: 0.8693\n",
            "Epoch 55/100\n",
            "5537/5537 [==============================] - 2s 351us/step - loss: 0.2817 - acc: 0.8824 - val_loss: 0.4103 - val_acc: 0.8527\n",
            "Epoch 56/100\n",
            "5537/5537 [==============================] - 2s 346us/step - loss: 0.2774 - acc: 0.8895 - val_loss: 0.4218 - val_acc: 0.8643\n",
            "Epoch 57/100\n",
            "5537/5537 [==============================] - 2s 347us/step - loss: 0.2833 - acc: 0.8844 - val_loss: 0.4137 - val_acc: 0.8650\n",
            "Epoch 58/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.2772 - acc: 0.8889 - val_loss: 0.3899 - val_acc: 0.8722\n",
            "Epoch 59/100\n",
            "5537/5537 [==============================] - 2s 347us/step - loss: 0.2790 - acc: 0.8871 - val_loss: 0.3974 - val_acc: 0.8621\n",
            "Epoch 60/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.2712 - acc: 0.8898 - val_loss: 0.4000 - val_acc: 0.8679\n",
            "Epoch 61/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.2779 - acc: 0.8897 - val_loss: 0.4031 - val_acc: 0.8592\n",
            "Epoch 62/100\n",
            "5537/5537 [==============================] - 2s 348us/step - loss: 0.2739 - acc: 0.8882 - val_loss: 0.4052 - val_acc: 0.8671\n",
            "Epoch 63/100\n",
            "5537/5537 [==============================] - 2s 346us/step - loss: 0.2759 - acc: 0.8853 - val_loss: 0.4036 - val_acc: 0.8621\n",
            "Epoch 64/100\n",
            "5537/5537 [==============================] - 2s 340us/step - loss: 0.2637 - acc: 0.8954 - val_loss: 0.3955 - val_acc: 0.8686\n",
            "Epoch 65/100\n",
            "5537/5537 [==============================] - 2s 346us/step - loss: 0.2678 - acc: 0.8922 - val_loss: 0.4149 - val_acc: 0.8527\n",
            "Epoch 66/100\n",
            "5537/5537 [==============================] - 2s 346us/step - loss: 0.2693 - acc: 0.8920 - val_loss: 0.4455 - val_acc: 0.8599\n",
            "Epoch 67/100\n",
            "5537/5537 [==============================] - 2s 340us/step - loss: 0.2706 - acc: 0.8951 - val_loss: 0.4078 - val_acc: 0.8643\n",
            "Epoch 68/100\n",
            "5537/5537 [==============================] - 2s 342us/step - loss: 0.2675 - acc: 0.8942 - val_loss: 0.4263 - val_acc: 0.8542\n",
            "Epoch 69/100\n",
            "5537/5537 [==============================] - 2s 343us/step - loss: 0.2629 - acc: 0.8924 - val_loss: 0.3896 - val_acc: 0.8686\n",
            "Epoch 70/100\n",
            "5537/5537 [==============================] - 2s 344us/step - loss: 0.2595 - acc: 0.8933 - val_loss: 0.4267 - val_acc: 0.8455\n",
            "Epoch 71/100\n",
            "5537/5537 [==============================] - 2s 346us/step - loss: 0.2655 - acc: 0.8934 - val_loss: 0.4047 - val_acc: 0.8614\n",
            "Epoch 72/100\n",
            "5537/5537 [==============================] - 2s 339us/step - loss: 0.2647 - acc: 0.8915 - val_loss: 0.3988 - val_acc: 0.8686\n",
            "Epoch 73/100\n",
            "5537/5537 [==============================] - 2s 346us/step - loss: 0.2614 - acc: 0.8953 - val_loss: 0.4030 - val_acc: 0.8686\n",
            "Epoch 74/100\n",
            "5537/5537 [==============================] - 2s 342us/step - loss: 0.2609 - acc: 0.8942 - val_loss: 0.4118 - val_acc: 0.8635\n",
            "Epoch 75/100\n",
            "5537/5537 [==============================] - 2s 337us/step - loss: 0.2569 - acc: 0.8980 - val_loss: 0.4475 - val_acc: 0.8440\n",
            "Epoch 76/100\n",
            "5537/5537 [==============================] - 2s 343us/step - loss: 0.2631 - acc: 0.8916 - val_loss: 0.4005 - val_acc: 0.8643\n",
            "Epoch 77/100\n",
            "5537/5537 [==============================] - 2s 340us/step - loss: 0.2573 - acc: 0.8940 - val_loss: 0.4064 - val_acc: 0.8643\n",
            "Epoch 78/100\n",
            "5537/5537 [==============================] - 2s 342us/step - loss: 0.2563 - acc: 0.8940 - val_loss: 0.3972 - val_acc: 0.8708\n",
            "Epoch 79/100\n",
            "5537/5537 [==============================] - 2s 341us/step - loss: 0.2615 - acc: 0.8936 - val_loss: 0.4038 - val_acc: 0.8628\n",
            "Epoch 80/100\n",
            "5537/5537 [==============================] - 2s 340us/step - loss: 0.2543 - acc: 0.8981 - val_loss: 0.4071 - val_acc: 0.8729\n",
            "Epoch 81/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.2494 - acc: 0.8971 - val_loss: 0.4087 - val_acc: 0.8736\n",
            "Epoch 82/100\n",
            "5537/5537 [==============================] - 2s 341us/step - loss: 0.2594 - acc: 0.8951 - val_loss: 0.4195 - val_acc: 0.8542\n",
            "Epoch 83/100\n",
            "5537/5537 [==============================] - 2s 339us/step - loss: 0.2558 - acc: 0.8967 - val_loss: 0.4299 - val_acc: 0.8592\n",
            "Epoch 84/100\n",
            "5537/5537 [==============================] - 2s 343us/step - loss: 0.2602 - acc: 0.8954 - val_loss: 0.3991 - val_acc: 0.8686\n",
            "Epoch 85/100\n",
            "5537/5537 [==============================] - 2s 342us/step - loss: 0.2555 - acc: 0.8972 - val_loss: 0.3973 - val_acc: 0.8736\n",
            "Epoch 86/100\n",
            "5537/5537 [==============================] - 2s 339us/step - loss: 0.2538 - acc: 0.8909 - val_loss: 0.4054 - val_acc: 0.8556\n",
            "Epoch 87/100\n",
            "5537/5537 [==============================] - 2s 345us/step - loss: 0.2547 - acc: 0.8924 - val_loss: 0.4068 - val_acc: 0.8744\n",
            "Epoch 88/100\n",
            "5537/5537 [==============================] - 2s 343us/step - loss: 0.2519 - acc: 0.8962 - val_loss: 0.4128 - val_acc: 0.8722\n",
            "Epoch 89/100\n",
            "5537/5537 [==============================] - 2s 339us/step - loss: 0.2513 - acc: 0.8985 - val_loss: 0.4061 - val_acc: 0.8744\n",
            "Epoch 90/100\n",
            "5537/5537 [==============================] - 2s 356us/step - loss: 0.2490 - acc: 0.8980 - val_loss: 0.4163 - val_acc: 0.8534\n",
            "Epoch 91/100\n",
            "5537/5537 [==============================] - 2s 350us/step - loss: 0.2464 - acc: 0.9030 - val_loss: 0.4205 - val_acc: 0.8440\n",
            "Epoch 92/100\n",
            "5537/5537 [==============================] - 2s 356us/step - loss: 0.2502 - acc: 0.8958 - val_loss: 0.4227 - val_acc: 0.8563\n",
            "Epoch 93/100\n",
            "5537/5537 [==============================] - 2s 354us/step - loss: 0.2525 - acc: 0.8965 - val_loss: 0.3993 - val_acc: 0.8715\n",
            "Epoch 94/100\n",
            "5537/5537 [==============================] - 2s 355us/step - loss: 0.2487 - acc: 0.8985 - val_loss: 0.4244 - val_acc: 0.8563\n",
            "Epoch 95/100\n",
            "5537/5537 [==============================] - 2s 352us/step - loss: 0.2496 - acc: 0.9003 - val_loss: 0.4226 - val_acc: 0.8614\n",
            "Epoch 96/100\n",
            "5537/5537 [==============================] - 2s 343us/step - loss: 0.2417 - acc: 0.9023 - val_loss: 0.4324 - val_acc: 0.8628\n",
            "Epoch 97/100\n",
            "5537/5537 [==============================] - 2s 343us/step - loss: 0.2471 - acc: 0.8998 - val_loss: 0.3828 - val_acc: 0.8715\n",
            "Epoch 98/100\n",
            "5537/5537 [==============================] - 2s 348us/step - loss: 0.2416 - acc: 0.9014 - val_loss: 0.4057 - val_acc: 0.8700\n",
            "Epoch 99/100\n",
            "5537/5537 [==============================] - 2s 344us/step - loss: 0.2403 - acc: 0.9046 - val_loss: 0.4095 - val_acc: 0.8751\n",
            "Epoch 100/100\n",
            "5537/5537 [==============================] - 2s 340us/step - loss: 0.2481 - acc: 0.8990 - val_loss: 0.4141 - val_acc: 0.8570\n",
            "Fold score (accuracy): 0.8570397111913357\n",
            "Fold #3\n",
            "Train on 5538 samples, validate on 1384 samples\n",
            "Epoch 1/100\n",
            "5538/5538 [==============================] - 2s 412us/step - loss: 0.8701 - acc: 0.5917 - val_loss: 0.7558 - val_acc: 0.6640\n",
            "Epoch 2/100\n",
            "5538/5538 [==============================] - 2s 346us/step - loss: 0.6907 - acc: 0.7091 - val_loss: 0.6414 - val_acc: 0.7478\n",
            "Epoch 3/100\n",
            "5538/5538 [==============================] - 2s 345us/step - loss: 0.5954 - acc: 0.7754 - val_loss: 0.5848 - val_acc: 0.7659\n",
            "Epoch 4/100\n",
            "5538/5538 [==============================] - 2s 342us/step - loss: 0.5442 - acc: 0.7905 - val_loss: 0.5505 - val_acc: 0.7840\n",
            "Epoch 5/100\n",
            "5538/5538 [==============================] - 2s 339us/step - loss: 0.5129 - acc: 0.8082 - val_loss: 0.5051 - val_acc: 0.8056\n",
            "Epoch 6/100\n",
            "5538/5538 [==============================] - 2s 343us/step - loss: 0.4925 - acc: 0.8053 - val_loss: 0.4787 - val_acc: 0.8143\n",
            "Epoch 7/100\n",
            "5538/5538 [==============================] - 2s 341us/step - loss: 0.4720 - acc: 0.8174 - val_loss: 0.4600 - val_acc: 0.8136\n",
            "Epoch 8/100\n",
            "5538/5538 [==============================] - 2s 349us/step - loss: 0.4548 - acc: 0.8230 - val_loss: 0.4709 - val_acc: 0.8071\n",
            "Epoch 9/100\n",
            "5538/5538 [==============================] - 2s 341us/step - loss: 0.4430 - acc: 0.8281 - val_loss: 0.4485 - val_acc: 0.8150\n",
            "Epoch 10/100\n",
            "5538/5538 [==============================] - 2s 340us/step - loss: 0.4323 - acc: 0.8304 - val_loss: 0.4524 - val_acc: 0.8259\n",
            "Epoch 11/100\n",
            "5538/5538 [==============================] - 2s 341us/step - loss: 0.4194 - acc: 0.8413 - val_loss: 0.4347 - val_acc: 0.8345\n",
            "Epoch 12/100\n",
            "5538/5538 [==============================] - 2s 339us/step - loss: 0.4118 - acc: 0.8407 - val_loss: 0.4386 - val_acc: 0.8244\n",
            "Epoch 13/100\n",
            "5538/5538 [==============================] - 2s 345us/step - loss: 0.4037 - acc: 0.8438 - val_loss: 0.4117 - val_acc: 0.8504\n",
            "Epoch 14/100\n",
            "5538/5538 [==============================] - 2s 348us/step - loss: 0.4008 - acc: 0.8443 - val_loss: 0.4168 - val_acc: 0.8324\n",
            "Epoch 15/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.3900 - acc: 0.8492 - val_loss: 0.4144 - val_acc: 0.8512\n",
            "Epoch 16/100\n",
            "5538/5538 [==============================] - 2s 341us/step - loss: 0.3864 - acc: 0.8530 - val_loss: 0.4106 - val_acc: 0.8504\n",
            "Epoch 17/100\n",
            "5538/5538 [==============================] - 2s 340us/step - loss: 0.3833 - acc: 0.8528 - val_loss: 0.4189 - val_acc: 0.8244\n",
            "Epoch 18/100\n",
            "5538/5538 [==============================] - 2s 346us/step - loss: 0.3787 - acc: 0.8550 - val_loss: 0.3971 - val_acc: 0.8497\n",
            "Epoch 19/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.3770 - acc: 0.8534 - val_loss: 0.3930 - val_acc: 0.8512\n",
            "Epoch 20/100\n",
            "5538/5538 [==============================] - 2s 339us/step - loss: 0.3646 - acc: 0.8617 - val_loss: 0.4221 - val_acc: 0.8316\n",
            "Epoch 21/100\n",
            "5538/5538 [==============================] - 2s 341us/step - loss: 0.3654 - acc: 0.8592 - val_loss: 0.3941 - val_acc: 0.8439\n",
            "Epoch 22/100\n",
            "5538/5538 [==============================] - 2s 339us/step - loss: 0.3595 - acc: 0.8584 - val_loss: 0.3949 - val_acc: 0.8410\n",
            "Epoch 23/100\n",
            "5538/5538 [==============================] - 2s 337us/step - loss: 0.3589 - acc: 0.8568 - val_loss: 0.4034 - val_acc: 0.8454\n",
            "Epoch 24/100\n",
            "5538/5538 [==============================] - 2s 339us/step - loss: 0.3491 - acc: 0.8671 - val_loss: 0.3880 - val_acc: 0.8468\n",
            "Epoch 25/100\n",
            "5538/5538 [==============================] - 2s 346us/step - loss: 0.3532 - acc: 0.8631 - val_loss: 0.3920 - val_acc: 0.8548\n",
            "Epoch 26/100\n",
            "5538/5538 [==============================] - 2s 343us/step - loss: 0.3518 - acc: 0.8660 - val_loss: 0.3984 - val_acc: 0.8382\n",
            "Epoch 27/100\n",
            "5538/5538 [==============================] - 2s 340us/step - loss: 0.3425 - acc: 0.8673 - val_loss: 0.3909 - val_acc: 0.8490\n",
            "Epoch 28/100\n",
            "5538/5538 [==============================] - 2s 343us/step - loss: 0.3421 - acc: 0.8657 - val_loss: 0.3976 - val_acc: 0.8447\n",
            "Epoch 29/100\n",
            "5538/5538 [==============================] - 2s 342us/step - loss: 0.3340 - acc: 0.8704 - val_loss: 0.3971 - val_acc: 0.8454\n",
            "Epoch 30/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.3389 - acc: 0.8685 - val_loss: 0.3825 - val_acc: 0.8389\n",
            "Epoch 31/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.3353 - acc: 0.8682 - val_loss: 0.3797 - val_acc: 0.8461\n",
            "Epoch 32/100\n",
            "5538/5538 [==============================] - 2s 340us/step - loss: 0.3343 - acc: 0.8731 - val_loss: 0.3881 - val_acc: 0.8504\n",
            "Epoch 33/100\n",
            "5538/5538 [==============================] - 2s 341us/step - loss: 0.3353 - acc: 0.8734 - val_loss: 0.3848 - val_acc: 0.8562\n",
            "Epoch 34/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.3297 - acc: 0.8729 - val_loss: 0.3836 - val_acc: 0.8540\n",
            "Epoch 35/100\n",
            "5538/5538 [==============================] - 2s 342us/step - loss: 0.3287 - acc: 0.8769 - val_loss: 0.3995 - val_acc: 0.8425\n",
            "Epoch 36/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.3291 - acc: 0.8704 - val_loss: 0.3712 - val_acc: 0.8526\n",
            "Epoch 37/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.3188 - acc: 0.8743 - val_loss: 0.4283 - val_acc: 0.8302\n",
            "Epoch 38/100\n",
            "5538/5538 [==============================] - 2s 343us/step - loss: 0.3191 - acc: 0.8750 - val_loss: 0.3825 - val_acc: 0.8418\n",
            "Epoch 39/100\n",
            "5538/5538 [==============================] - 2s 342us/step - loss: 0.3200 - acc: 0.8729 - val_loss: 0.3660 - val_acc: 0.8526\n",
            "Epoch 40/100\n",
            "5538/5538 [==============================] - 2s 341us/step - loss: 0.3148 - acc: 0.8763 - val_loss: 0.3636 - val_acc: 0.8584\n",
            "Epoch 41/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.3170 - acc: 0.8769 - val_loss: 0.3886 - val_acc: 0.8468\n",
            "Epoch 42/100\n",
            "5538/5538 [==============================] - 2s 341us/step - loss: 0.3158 - acc: 0.8727 - val_loss: 0.3600 - val_acc: 0.8598\n",
            "Epoch 43/100\n",
            "5538/5538 [==============================] - 2s 342us/step - loss: 0.3125 - acc: 0.8806 - val_loss: 0.3751 - val_acc: 0.8533\n",
            "Epoch 44/100\n",
            "5538/5538 [==============================] - 2s 343us/step - loss: 0.3119 - acc: 0.8749 - val_loss: 0.3820 - val_acc: 0.8584\n",
            "Epoch 45/100\n",
            "5538/5538 [==============================] - 2s 346us/step - loss: 0.3072 - acc: 0.8797 - val_loss: 0.3674 - val_acc: 0.8519\n",
            "Epoch 46/100\n",
            "5538/5538 [==============================] - 2s 343us/step - loss: 0.3102 - acc: 0.8783 - val_loss: 0.3688 - val_acc: 0.8555\n",
            "Epoch 47/100\n",
            "5538/5538 [==============================] - 2s 346us/step - loss: 0.3021 - acc: 0.8832 - val_loss: 0.3881 - val_acc: 0.8540\n",
            "Epoch 48/100\n",
            "5538/5538 [==============================] - 2s 342us/step - loss: 0.3101 - acc: 0.8801 - val_loss: 0.3728 - val_acc: 0.8548\n",
            "Epoch 49/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.3054 - acc: 0.8808 - val_loss: 0.3856 - val_acc: 0.8432\n",
            "Epoch 50/100\n",
            "5538/5538 [==============================] - 2s 343us/step - loss: 0.3026 - acc: 0.8853 - val_loss: 0.3708 - val_acc: 0.8504\n",
            "Epoch 51/100\n",
            "5538/5538 [==============================] - 2s 341us/step - loss: 0.3002 - acc: 0.8815 - val_loss: 0.3832 - val_acc: 0.8519\n",
            "Epoch 52/100\n",
            "5538/5538 [==============================] - 2s 342us/step - loss: 0.2973 - acc: 0.8864 - val_loss: 0.3672 - val_acc: 0.8591\n",
            "Epoch 53/100\n",
            "5538/5538 [==============================] - 2s 342us/step - loss: 0.2993 - acc: 0.8830 - val_loss: 0.3748 - val_acc: 0.8540\n",
            "Epoch 54/100\n",
            "5538/5538 [==============================] - 2s 341us/step - loss: 0.2938 - acc: 0.8843 - val_loss: 0.3606 - val_acc: 0.8663\n",
            "Epoch 55/100\n",
            "5538/5538 [==============================] - 2s 341us/step - loss: 0.2932 - acc: 0.8877 - val_loss: 0.3567 - val_acc: 0.8634\n",
            "Epoch 56/100\n",
            "5538/5538 [==============================] - 2s 338us/step - loss: 0.2966 - acc: 0.8837 - val_loss: 0.3617 - val_acc: 0.8555\n",
            "Epoch 57/100\n",
            "5538/5538 [==============================] - 2s 340us/step - loss: 0.2909 - acc: 0.8875 - val_loss: 0.3682 - val_acc: 0.8533\n",
            "Epoch 58/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.2892 - acc: 0.8832 - val_loss: 0.3670 - val_acc: 0.8627\n",
            "Epoch 59/100\n",
            "5538/5538 [==============================] - 2s 340us/step - loss: 0.2896 - acc: 0.8843 - val_loss: 0.3542 - val_acc: 0.8649\n",
            "Epoch 60/100\n",
            "5538/5538 [==============================] - 2s 341us/step - loss: 0.2922 - acc: 0.8808 - val_loss: 0.4113 - val_acc: 0.8367\n",
            "Epoch 61/100\n",
            "5538/5538 [==============================] - 2s 336us/step - loss: 0.2845 - acc: 0.8868 - val_loss: 0.3572 - val_acc: 0.8613\n",
            "Epoch 62/100\n",
            "5538/5538 [==============================] - 2s 338us/step - loss: 0.2833 - acc: 0.8891 - val_loss: 0.3545 - val_acc: 0.8649\n",
            "Epoch 63/100\n",
            "5538/5538 [==============================] - 2s 342us/step - loss: 0.2842 - acc: 0.8875 - val_loss: 0.3657 - val_acc: 0.8627\n",
            "Epoch 64/100\n",
            "5538/5538 [==============================] - 2s 340us/step - loss: 0.2786 - acc: 0.8893 - val_loss: 0.3817 - val_acc: 0.8569\n",
            "Epoch 65/100\n",
            "5538/5538 [==============================] - 2s 340us/step - loss: 0.2848 - acc: 0.8832 - val_loss: 0.3634 - val_acc: 0.8663\n",
            "Epoch 66/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.2825 - acc: 0.8884 - val_loss: 0.3591 - val_acc: 0.8634\n",
            "Epoch 67/100\n",
            "5538/5538 [==============================] - 2s 337us/step - loss: 0.2751 - acc: 0.8918 - val_loss: 0.3744 - val_acc: 0.8562\n",
            "Epoch 68/100\n",
            "5538/5538 [==============================] - 2s 341us/step - loss: 0.2803 - acc: 0.8879 - val_loss: 0.3755 - val_acc: 0.8685\n",
            "Epoch 69/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.2795 - acc: 0.8888 - val_loss: 0.3803 - val_acc: 0.8504\n",
            "Epoch 70/100\n",
            "5538/5538 [==============================] - 2s 340us/step - loss: 0.2766 - acc: 0.8875 - val_loss: 0.3553 - val_acc: 0.8504\n",
            "Epoch 71/100\n",
            "5538/5538 [==============================] - 2s 340us/step - loss: 0.2767 - acc: 0.8904 - val_loss: 0.3634 - val_acc: 0.8663\n",
            "Epoch 72/100\n",
            "5538/5538 [==============================] - 2s 335us/step - loss: 0.2685 - acc: 0.8924 - val_loss: 0.3785 - val_acc: 0.8678\n",
            "Epoch 73/100\n",
            "5538/5538 [==============================] - 2s 338us/step - loss: 0.2781 - acc: 0.8877 - val_loss: 0.3608 - val_acc: 0.8671\n",
            "Epoch 74/100\n",
            "5538/5538 [==============================] - 2s 338us/step - loss: 0.2708 - acc: 0.8915 - val_loss: 0.3692 - val_acc: 0.8649\n",
            "Epoch 75/100\n",
            "5538/5538 [==============================] - 2s 342us/step - loss: 0.2722 - acc: 0.8956 - val_loss: 0.3684 - val_acc: 0.8548\n",
            "Epoch 76/100\n",
            "5538/5538 [==============================] - 2s 338us/step - loss: 0.2776 - acc: 0.8873 - val_loss: 0.3940 - val_acc: 0.8519\n",
            "Epoch 77/100\n",
            "5538/5538 [==============================] - 2s 341us/step - loss: 0.2719 - acc: 0.8933 - val_loss: 0.3751 - val_acc: 0.8605\n",
            "Epoch 78/100\n",
            "5538/5538 [==============================] - 2s 337us/step - loss: 0.2737 - acc: 0.8936 - val_loss: 0.3590 - val_acc: 0.8671\n",
            "Epoch 79/100\n",
            "5538/5538 [==============================] - 2s 341us/step - loss: 0.2742 - acc: 0.8927 - val_loss: 0.3580 - val_acc: 0.8620\n",
            "Epoch 80/100\n",
            "5538/5538 [==============================] - 2s 345us/step - loss: 0.2672 - acc: 0.8944 - val_loss: 0.3711 - val_acc: 0.8671\n",
            "Epoch 81/100\n",
            "5538/5538 [==============================] - 2s 345us/step - loss: 0.2624 - acc: 0.8936 - val_loss: 0.3832 - val_acc: 0.8642\n",
            "Epoch 82/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.2676 - acc: 0.8960 - val_loss: 0.3562 - val_acc: 0.8627\n",
            "Epoch 83/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.2712 - acc: 0.8924 - val_loss: 0.3666 - val_acc: 0.8591\n",
            "Epoch 84/100\n",
            "5538/5538 [==============================] - 2s 341us/step - loss: 0.2664 - acc: 0.8915 - val_loss: 0.3494 - val_acc: 0.8707\n",
            "Epoch 85/100\n",
            "5538/5538 [==============================] - 2s 341us/step - loss: 0.2594 - acc: 0.8964 - val_loss: 0.3809 - val_acc: 0.8548\n",
            "Epoch 86/100\n",
            "5538/5538 [==============================] - 2s 337us/step - loss: 0.2625 - acc: 0.8951 - val_loss: 0.3556 - val_acc: 0.8692\n",
            "Epoch 87/100\n",
            "5538/5538 [==============================] - 2s 340us/step - loss: 0.2614 - acc: 0.8958 - val_loss: 0.3898 - val_acc: 0.8569\n",
            "Epoch 88/100\n",
            "5538/5538 [==============================] - 2s 339us/step - loss: 0.2674 - acc: 0.8938 - val_loss: 0.3852 - val_acc: 0.8642\n",
            "Epoch 89/100\n",
            "5538/5538 [==============================] - 2s 338us/step - loss: 0.2613 - acc: 0.8974 - val_loss: 0.3889 - val_acc: 0.8569\n",
            "Epoch 90/100\n",
            "5538/5538 [==============================] - 2s 339us/step - loss: 0.2594 - acc: 0.8980 - val_loss: 0.3527 - val_acc: 0.8743\n",
            "Epoch 91/100\n",
            "5538/5538 [==============================] - 2s 341us/step - loss: 0.2532 - acc: 0.8982 - val_loss: 0.3888 - val_acc: 0.8598\n",
            "Epoch 92/100\n",
            "5538/5538 [==============================] - 2s 339us/step - loss: 0.2625 - acc: 0.8949 - val_loss: 0.3682 - val_acc: 0.8663\n",
            "Epoch 93/100\n",
            "5538/5538 [==============================] - 2s 340us/step - loss: 0.2608 - acc: 0.8969 - val_loss: 0.3566 - val_acc: 0.8736\n",
            "Epoch 94/100\n",
            "5538/5538 [==============================] - 2s 345us/step - loss: 0.2564 - acc: 0.8953 - val_loss: 0.3595 - val_acc: 0.8736\n",
            "Epoch 95/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.2619 - acc: 0.8951 - val_loss: 0.3637 - val_acc: 0.8699\n",
            "Epoch 96/100\n",
            "5538/5538 [==============================] - 2s 338us/step - loss: 0.2553 - acc: 0.8982 - val_loss: 0.3602 - val_acc: 0.8707\n",
            "Epoch 97/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.2524 - acc: 0.9012 - val_loss: 0.3740 - val_acc: 0.8598\n",
            "Epoch 98/100\n",
            "5538/5538 [==============================] - 2s 340us/step - loss: 0.2584 - acc: 0.8951 - val_loss: 0.3819 - val_acc: 0.8490\n",
            "Epoch 99/100\n",
            "5538/5538 [==============================] - 2s 343us/step - loss: 0.2598 - acc: 0.8956 - val_loss: 0.3563 - val_acc: 0.8685\n",
            "Epoch 100/100\n",
            "5538/5538 [==============================] - 2s 341us/step - loss: 0.2563 - acc: 0.9003 - val_loss: 0.3538 - val_acc: 0.8685\n",
            "Fold score (accuracy): 0.8684971098265896\n",
            "Fold #4\n",
            "Train on 5538 samples, validate on 1384 samples\n",
            "Epoch 1/100\n",
            "5538/5538 [==============================] - 2s 431us/step - loss: 0.8258 - acc: 0.6331 - val_loss: 0.7067 - val_acc: 0.7327\n",
            "Epoch 2/100\n",
            "5538/5538 [==============================] - 2s 359us/step - loss: 0.6353 - acc: 0.7497 - val_loss: 0.6378 - val_acc: 0.7442\n",
            "Epoch 3/100\n",
            "5538/5538 [==============================] - 2s 349us/step - loss: 0.5572 - acc: 0.7777 - val_loss: 0.5578 - val_acc: 0.7919\n",
            "Epoch 4/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.5164 - acc: 0.7936 - val_loss: 0.5297 - val_acc: 0.8035\n",
            "Epoch 5/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.4943 - acc: 0.8113 - val_loss: 0.5169 - val_acc: 0.7991\n",
            "Epoch 6/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.4707 - acc: 0.8138 - val_loss: 0.5024 - val_acc: 0.8121\n",
            "Epoch 7/100\n",
            "5538/5538 [==============================] - 2s 352us/step - loss: 0.4529 - acc: 0.8259 - val_loss: 0.4916 - val_acc: 0.8078\n",
            "Epoch 8/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.4417 - acc: 0.8294 - val_loss: 0.4760 - val_acc: 0.8136\n",
            "Epoch 9/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.4279 - acc: 0.8312 - val_loss: 0.4811 - val_acc: 0.8114\n",
            "Epoch 10/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.4171 - acc: 0.8382 - val_loss: 0.4569 - val_acc: 0.8223\n",
            "Epoch 11/100\n",
            "5538/5538 [==============================] - 2s 342us/step - loss: 0.4064 - acc: 0.8404 - val_loss: 0.4559 - val_acc: 0.8150\n",
            "Epoch 12/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.4013 - acc: 0.8420 - val_loss: 0.4645 - val_acc: 0.8150\n",
            "Epoch 13/100\n",
            "5538/5538 [==============================] - 2s 348us/step - loss: 0.3952 - acc: 0.8508 - val_loss: 0.4414 - val_acc: 0.8259\n",
            "Epoch 14/100\n",
            "5538/5538 [==============================] - 2s 346us/step - loss: 0.3817 - acc: 0.8537 - val_loss: 0.4445 - val_acc: 0.8208\n",
            "Epoch 15/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.3759 - acc: 0.8534 - val_loss: 0.4865 - val_acc: 0.8136\n",
            "Epoch 16/100\n",
            "5538/5538 [==============================] - 2s 348us/step - loss: 0.3704 - acc: 0.8507 - val_loss: 0.4337 - val_acc: 0.8215\n",
            "Epoch 17/100\n",
            "5538/5538 [==============================] - 2s 346us/step - loss: 0.3603 - acc: 0.8638 - val_loss: 0.4208 - val_acc: 0.8324\n",
            "Epoch 18/100\n",
            "5538/5538 [==============================] - 2s 349us/step - loss: 0.3613 - acc: 0.8554 - val_loss: 0.4337 - val_acc: 0.8201\n",
            "Epoch 19/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.3539 - acc: 0.8624 - val_loss: 0.4193 - val_acc: 0.8345\n",
            "Epoch 20/100\n",
            "5538/5538 [==============================] - 2s 353us/step - loss: 0.3491 - acc: 0.8646 - val_loss: 0.4285 - val_acc: 0.8273\n",
            "Epoch 21/100\n",
            "5538/5538 [==============================] - 2s 350us/step - loss: 0.3448 - acc: 0.8704 - val_loss: 0.4380 - val_acc: 0.8338\n",
            "Epoch 22/100\n",
            "5538/5538 [==============================] - 2s 349us/step - loss: 0.3431 - acc: 0.8689 - val_loss: 0.4631 - val_acc: 0.8208\n",
            "Epoch 23/100\n",
            "5538/5538 [==============================] - 2s 354us/step - loss: 0.3391 - acc: 0.8704 - val_loss: 0.4251 - val_acc: 0.8439\n",
            "Epoch 24/100\n",
            "5538/5538 [==============================] - 2s 352us/step - loss: 0.3367 - acc: 0.8700 - val_loss: 0.4337 - val_acc: 0.8302\n",
            "Epoch 25/100\n",
            "5538/5538 [==============================] - 2s 355us/step - loss: 0.3339 - acc: 0.8716 - val_loss: 0.4142 - val_acc: 0.8302\n",
            "Epoch 26/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.3318 - acc: 0.8673 - val_loss: 0.4583 - val_acc: 0.8143\n",
            "Epoch 27/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.3244 - acc: 0.8720 - val_loss: 0.4258 - val_acc: 0.8374\n",
            "Epoch 28/100\n",
            "5538/5538 [==============================] - 2s 350us/step - loss: 0.3320 - acc: 0.8720 - val_loss: 0.4133 - val_acc: 0.8454\n",
            "Epoch 29/100\n",
            "5538/5538 [==============================] - 2s 346us/step - loss: 0.3247 - acc: 0.8752 - val_loss: 0.4064 - val_acc: 0.8439\n",
            "Epoch 30/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.3246 - acc: 0.8734 - val_loss: 0.4204 - val_acc: 0.8324\n",
            "Epoch 31/100\n",
            "5538/5538 [==============================] - 2s 346us/step - loss: 0.3146 - acc: 0.8792 - val_loss: 0.4286 - val_acc: 0.8295\n",
            "Epoch 32/100\n",
            "5538/5538 [==============================] - 2s 348us/step - loss: 0.3177 - acc: 0.8727 - val_loss: 0.4184 - val_acc: 0.8353\n",
            "Epoch 33/100\n",
            "5538/5538 [==============================] - 2s 349us/step - loss: 0.3113 - acc: 0.8812 - val_loss: 0.4355 - val_acc: 0.8302\n",
            "Epoch 34/100\n",
            "5538/5538 [==============================] - 2s 348us/step - loss: 0.3191 - acc: 0.8759 - val_loss: 0.4068 - val_acc: 0.8382\n",
            "Epoch 35/100\n",
            "5538/5538 [==============================] - 2s 345us/step - loss: 0.3126 - acc: 0.8787 - val_loss: 0.4087 - val_acc: 0.8418\n",
            "Epoch 36/100\n",
            "5538/5538 [==============================] - 2s 345us/step - loss: 0.3079 - acc: 0.8783 - val_loss: 0.4243 - val_acc: 0.8367\n",
            "Epoch 37/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.3030 - acc: 0.8846 - val_loss: 0.4277 - val_acc: 0.8316\n",
            "Epoch 38/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.3112 - acc: 0.8752 - val_loss: 0.4030 - val_acc: 0.8425\n",
            "Epoch 39/100\n",
            "5538/5538 [==============================] - 2s 343us/step - loss: 0.3037 - acc: 0.8834 - val_loss: 0.4107 - val_acc: 0.8418\n",
            "Epoch 40/100\n",
            "5538/5538 [==============================] - 2s 349us/step - loss: 0.3008 - acc: 0.8839 - val_loss: 0.4178 - val_acc: 0.8288\n",
            "Epoch 41/100\n",
            "5538/5538 [==============================] - 2s 346us/step - loss: 0.3006 - acc: 0.8835 - val_loss: 0.3954 - val_acc: 0.8425\n",
            "Epoch 42/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.3006 - acc: 0.8877 - val_loss: 0.4151 - val_acc: 0.8439\n",
            "Epoch 43/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.2957 - acc: 0.8850 - val_loss: 0.4336 - val_acc: 0.8425\n",
            "Epoch 44/100\n",
            "5538/5538 [==============================] - 2s 349us/step - loss: 0.2932 - acc: 0.8855 - val_loss: 0.4711 - val_acc: 0.8129\n",
            "Epoch 45/100\n",
            "5538/5538 [==============================] - 2s 350us/step - loss: 0.2935 - acc: 0.8882 - val_loss: 0.4056 - val_acc: 0.8367\n",
            "Epoch 46/100\n",
            "5538/5538 [==============================] - 2s 348us/step - loss: 0.2972 - acc: 0.8805 - val_loss: 0.4322 - val_acc: 0.8302\n",
            "Epoch 47/100\n",
            "5538/5538 [==============================] - 2s 346us/step - loss: 0.2905 - acc: 0.8882 - val_loss: 0.4030 - val_acc: 0.8367\n",
            "Epoch 48/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.2945 - acc: 0.8850 - val_loss: 0.4162 - val_acc: 0.8389\n",
            "Epoch 49/100\n",
            "5538/5538 [==============================] - 2s 348us/step - loss: 0.2832 - acc: 0.8888 - val_loss: 0.4117 - val_acc: 0.8425\n",
            "Epoch 50/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.2871 - acc: 0.8862 - val_loss: 0.4088 - val_acc: 0.8533\n",
            "Epoch 51/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.2848 - acc: 0.8852 - val_loss: 0.4262 - val_acc: 0.8309\n",
            "Epoch 52/100\n",
            "5538/5538 [==============================] - 2s 364us/step - loss: 0.2835 - acc: 0.8888 - val_loss: 0.4276 - val_acc: 0.8367\n",
            "Epoch 53/100\n",
            "5538/5538 [==============================] - 2s 355us/step - loss: 0.2813 - acc: 0.8909 - val_loss: 0.4229 - val_acc: 0.8447\n",
            "Epoch 54/100\n",
            "5538/5538 [==============================] - 2s 359us/step - loss: 0.2828 - acc: 0.8906 - val_loss: 0.4222 - val_acc: 0.8410\n",
            "Epoch 55/100\n",
            "5538/5538 [==============================] - 2s 357us/step - loss: 0.2809 - acc: 0.8899 - val_loss: 0.4014 - val_acc: 0.8468\n",
            "Epoch 56/100\n",
            "5538/5538 [==============================] - 2s 360us/step - loss: 0.2854 - acc: 0.8864 - val_loss: 0.4089 - val_acc: 0.8468\n",
            "Epoch 57/100\n",
            "5538/5538 [==============================] - 2s 355us/step - loss: 0.2827 - acc: 0.8908 - val_loss: 0.4139 - val_acc: 0.8410\n",
            "Epoch 58/100\n",
            "5538/5538 [==============================] - 2s 349us/step - loss: 0.2763 - acc: 0.8879 - val_loss: 0.4298 - val_acc: 0.8374\n",
            "Epoch 59/100\n",
            "5538/5538 [==============================] - 2s 352us/step - loss: 0.2813 - acc: 0.8920 - val_loss: 0.4259 - val_acc: 0.8360\n",
            "Epoch 60/100\n",
            "5538/5538 [==============================] - 2s 348us/step - loss: 0.2773 - acc: 0.8902 - val_loss: 0.3980 - val_acc: 0.8540\n",
            "Epoch 61/100\n",
            "5538/5538 [==============================] - 2s 354us/step - loss: 0.2806 - acc: 0.8929 - val_loss: 0.4005 - val_acc: 0.8475\n",
            "Epoch 62/100\n",
            "5538/5538 [==============================] - 2s 352us/step - loss: 0.2726 - acc: 0.8911 - val_loss: 0.4129 - val_acc: 0.8483\n",
            "Epoch 63/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.2759 - acc: 0.8908 - val_loss: 0.4068 - val_acc: 0.8447\n",
            "Epoch 64/100\n",
            "5538/5538 [==============================] - 2s 352us/step - loss: 0.2714 - acc: 0.8908 - val_loss: 0.4110 - val_acc: 0.8439\n",
            "Epoch 65/100\n",
            "5538/5538 [==============================] - 2s 349us/step - loss: 0.2703 - acc: 0.8924 - val_loss: 0.4076 - val_acc: 0.8468\n",
            "Epoch 66/100\n",
            "5538/5538 [==============================] - 2s 350us/step - loss: 0.2686 - acc: 0.8924 - val_loss: 0.3991 - val_acc: 0.8490\n",
            "Epoch 67/100\n",
            "5538/5538 [==============================] - 2s 349us/step - loss: 0.2703 - acc: 0.8931 - val_loss: 0.4109 - val_acc: 0.8454\n",
            "Epoch 68/100\n",
            "5538/5538 [==============================] - 2s 346us/step - loss: 0.2654 - acc: 0.8958 - val_loss: 0.4220 - val_acc: 0.8439\n",
            "Epoch 69/100\n",
            "5538/5538 [==============================] - 2s 348us/step - loss: 0.2751 - acc: 0.8931 - val_loss: 0.3967 - val_acc: 0.8540\n",
            "Epoch 70/100\n",
            "5538/5538 [==============================] - 2s 343us/step - loss: 0.2655 - acc: 0.8956 - val_loss: 0.3999 - val_acc: 0.8512\n",
            "Epoch 71/100\n",
            "5538/5538 [==============================] - 2s 342us/step - loss: 0.2680 - acc: 0.8953 - val_loss: 0.4334 - val_acc: 0.8403\n",
            "Epoch 72/100\n",
            "5538/5538 [==============================] - 2s 348us/step - loss: 0.2661 - acc: 0.8982 - val_loss: 0.4023 - val_acc: 0.8468\n",
            "Epoch 73/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.2638 - acc: 0.8918 - val_loss: 0.4135 - val_acc: 0.8483\n",
            "Epoch 74/100\n",
            "5538/5538 [==============================] - 2s 341us/step - loss: 0.2654 - acc: 0.8922 - val_loss: 0.4280 - val_acc: 0.8360\n",
            "Epoch 75/100\n",
            "5538/5538 [==============================] - 2s 345us/step - loss: 0.2633 - acc: 0.8951 - val_loss: 0.4397 - val_acc: 0.8309\n",
            "Epoch 76/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.2678 - acc: 0.8945 - val_loss: 0.4140 - val_acc: 0.8439\n",
            "Epoch 77/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.2609 - acc: 0.8956 - val_loss: 0.4030 - val_acc: 0.8461\n",
            "Epoch 78/100\n",
            "5538/5538 [==============================] - 2s 345us/step - loss: 0.2577 - acc: 0.8976 - val_loss: 0.4218 - val_acc: 0.8396\n",
            "Epoch 79/100\n",
            "5538/5538 [==============================] - 2s 340us/step - loss: 0.2615 - acc: 0.8954 - val_loss: 0.4212 - val_acc: 0.8439\n",
            "Epoch 80/100\n",
            "5538/5538 [==============================] - 2s 342us/step - loss: 0.2579 - acc: 0.8944 - val_loss: 0.4209 - val_acc: 0.8403\n",
            "Epoch 81/100\n",
            "5538/5538 [==============================] - 2s 345us/step - loss: 0.2584 - acc: 0.8992 - val_loss: 0.4279 - val_acc: 0.8432\n",
            "Epoch 82/100\n",
            "5538/5538 [==============================] - 2s 342us/step - loss: 0.2566 - acc: 0.8971 - val_loss: 0.4026 - val_acc: 0.8548\n",
            "Epoch 83/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.2593 - acc: 0.8947 - val_loss: 0.4332 - val_acc: 0.8360\n",
            "Epoch 84/100\n",
            "5538/5538 [==============================] - 2s 343us/step - loss: 0.2566 - acc: 0.8974 - val_loss: 0.4466 - val_acc: 0.8403\n",
            "Epoch 85/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.2545 - acc: 0.8985 - val_loss: 0.4170 - val_acc: 0.8512\n",
            "Epoch 86/100\n",
            "5538/5538 [==============================] - 2s 349us/step - loss: 0.2548 - acc: 0.8991 - val_loss: 0.4322 - val_acc: 0.8461\n",
            "Epoch 87/100\n",
            "5538/5538 [==============================] - 2s 345us/step - loss: 0.2533 - acc: 0.8996 - val_loss: 0.4318 - val_acc: 0.8454\n",
            "Epoch 88/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.2519 - acc: 0.8974 - val_loss: 0.4270 - val_acc: 0.8490\n",
            "Epoch 89/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.2536 - acc: 0.9009 - val_loss: 0.4147 - val_acc: 0.8555\n",
            "Epoch 90/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.2486 - acc: 0.9041 - val_loss: 0.4730 - val_acc: 0.8230\n",
            "Epoch 91/100\n",
            "5538/5538 [==============================] - 2s 348us/step - loss: 0.2500 - acc: 0.8996 - val_loss: 0.4454 - val_acc: 0.8316\n",
            "Epoch 92/100\n",
            "5538/5538 [==============================] - 2s 346us/step - loss: 0.2516 - acc: 0.9021 - val_loss: 0.4238 - val_acc: 0.8447\n",
            "Epoch 93/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.2474 - acc: 0.9010 - val_loss: 0.4160 - val_acc: 0.8526\n",
            "Epoch 94/100\n",
            "5538/5538 [==============================] - 2s 346us/step - loss: 0.2485 - acc: 0.9016 - val_loss: 0.4110 - val_acc: 0.8598\n",
            "Epoch 95/100\n",
            "5538/5538 [==============================] - 2s 345us/step - loss: 0.2490 - acc: 0.9025 - val_loss: 0.4209 - val_acc: 0.8475\n",
            "Epoch 96/100\n",
            "5538/5538 [==============================] - 2s 348us/step - loss: 0.2453 - acc: 0.9010 - val_loss: 0.4543 - val_acc: 0.8338\n",
            "Epoch 97/100\n",
            "5538/5538 [==============================] - 2s 345us/step - loss: 0.2446 - acc: 0.9010 - val_loss: 0.4318 - val_acc: 0.8410\n",
            "Epoch 98/100\n",
            "5538/5538 [==============================] - 2s 352us/step - loss: 0.2446 - acc: 0.9030 - val_loss: 0.4220 - val_acc: 0.8439\n",
            "Epoch 99/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.2422 - acc: 0.9020 - val_loss: 0.4259 - val_acc: 0.8468\n",
            "Epoch 100/100\n",
            "5538/5538 [==============================] - 2s 345us/step - loss: 0.2458 - acc: 0.8994 - val_loss: 0.4336 - val_acc: 0.8454\n",
            "Fold score (accuracy): 0.8453757225433526\n",
            "Fold #5\n",
            "Train on 5538 samples, validate on 1384 samples\n",
            "Epoch 1/100\n",
            "5538/5538 [==============================] - 2s 430us/step - loss: 0.8565 - acc: 0.6241 - val_loss: 0.6714 - val_acc: 0.7572\n",
            "Epoch 2/100\n",
            "5538/5538 [==============================] - 2s 348us/step - loss: 0.6347 - acc: 0.7647 - val_loss: 0.5816 - val_acc: 0.7832\n",
            "Epoch 3/100\n",
            "5538/5538 [==============================] - 2s 352us/step - loss: 0.5681 - acc: 0.7837 - val_loss: 0.5197 - val_acc: 0.8056\n",
            "Epoch 4/100\n",
            "5538/5538 [==============================] - 2s 350us/step - loss: 0.5261 - acc: 0.7974 - val_loss: 0.5228 - val_acc: 0.7934\n",
            "Epoch 5/100\n",
            "5538/5538 [==============================] - 2s 350us/step - loss: 0.4964 - acc: 0.8135 - val_loss: 0.4731 - val_acc: 0.8201\n",
            "Epoch 6/100\n",
            "5538/5538 [==============================] - 2s 349us/step - loss: 0.4809 - acc: 0.8196 - val_loss: 0.4896 - val_acc: 0.8143\n",
            "Epoch 7/100\n",
            "5538/5538 [==============================] - 2s 350us/step - loss: 0.4617 - acc: 0.8245 - val_loss: 0.4431 - val_acc: 0.8338\n",
            "Epoch 8/100\n",
            "5538/5538 [==============================] - 2s 349us/step - loss: 0.4502 - acc: 0.8292 - val_loss: 0.4519 - val_acc: 0.8129\n",
            "Epoch 9/100\n",
            "5538/5538 [==============================] - 2s 350us/step - loss: 0.4404 - acc: 0.8276 - val_loss: 0.4618 - val_acc: 0.8143\n",
            "Epoch 10/100\n",
            "5538/5538 [==============================] - 2s 350us/step - loss: 0.4294 - acc: 0.8288 - val_loss: 0.4263 - val_acc: 0.8353\n",
            "Epoch 11/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.4153 - acc: 0.8369 - val_loss: 0.4814 - val_acc: 0.8100\n",
            "Epoch 12/100\n",
            "5538/5538 [==============================] - 2s 346us/step - loss: 0.4111 - acc: 0.8391 - val_loss: 0.4125 - val_acc: 0.8490\n",
            "Epoch 13/100\n",
            "5538/5538 [==============================] - 2s 345us/step - loss: 0.4050 - acc: 0.8454 - val_loss: 0.4029 - val_acc: 0.8454\n",
            "Epoch 14/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.3973 - acc: 0.8438 - val_loss: 0.4422 - val_acc: 0.8324\n",
            "Epoch 15/100\n",
            "5538/5538 [==============================] - 2s 352us/step - loss: 0.3928 - acc: 0.8480 - val_loss: 0.4003 - val_acc: 0.8418\n",
            "Epoch 16/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.3801 - acc: 0.8525 - val_loss: 0.3895 - val_acc: 0.8555\n",
            "Epoch 17/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.3790 - acc: 0.8505 - val_loss: 0.3942 - val_acc: 0.8519\n",
            "Epoch 18/100\n",
            "5538/5538 [==============================] - 2s 346us/step - loss: 0.3740 - acc: 0.8550 - val_loss: 0.4080 - val_acc: 0.8490\n",
            "Epoch 19/100\n",
            "5538/5538 [==============================] - 2s 345us/step - loss: 0.3684 - acc: 0.8577 - val_loss: 0.3961 - val_acc: 0.8447\n",
            "Epoch 20/100\n",
            "5538/5538 [==============================] - 2s 346us/step - loss: 0.3670 - acc: 0.8548 - val_loss: 0.3881 - val_acc: 0.8382\n",
            "Epoch 21/100\n",
            "5538/5538 [==============================] - 2s 342us/step - loss: 0.3590 - acc: 0.8573 - val_loss: 0.3674 - val_acc: 0.8591\n",
            "Epoch 22/100\n",
            "5538/5538 [==============================] - 2s 345us/step - loss: 0.3582 - acc: 0.8552 - val_loss: 0.3626 - val_acc: 0.8663\n",
            "Epoch 23/100\n",
            "5538/5538 [==============================] - 2s 342us/step - loss: 0.3515 - acc: 0.8617 - val_loss: 0.3936 - val_acc: 0.8454\n",
            "Epoch 24/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.3526 - acc: 0.8597 - val_loss: 0.3642 - val_acc: 0.8584\n",
            "Epoch 25/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.3428 - acc: 0.8648 - val_loss: 0.3626 - val_acc: 0.8504\n",
            "Epoch 26/100\n",
            "5538/5538 [==============================] - 2s 349us/step - loss: 0.3415 - acc: 0.8619 - val_loss: 0.3817 - val_acc: 0.8569\n",
            "Epoch 27/100\n",
            "5538/5538 [==============================] - 2s 346us/step - loss: 0.3373 - acc: 0.8664 - val_loss: 0.3632 - val_acc: 0.8598\n",
            "Epoch 28/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.3415 - acc: 0.8635 - val_loss: 0.3500 - val_acc: 0.8620\n",
            "Epoch 29/100\n",
            "5538/5538 [==============================] - 2s 342us/step - loss: 0.3348 - acc: 0.8682 - val_loss: 0.3660 - val_acc: 0.8642\n",
            "Epoch 30/100\n",
            "5538/5538 [==============================] - 2s 342us/step - loss: 0.3301 - acc: 0.8680 - val_loss: 0.3626 - val_acc: 0.8642\n",
            "Epoch 31/100\n",
            "5538/5538 [==============================] - 2s 348us/step - loss: 0.3295 - acc: 0.8698 - val_loss: 0.3594 - val_acc: 0.8685\n",
            "Epoch 32/100\n",
            "5538/5538 [==============================] - 2s 349us/step - loss: 0.3288 - acc: 0.8653 - val_loss: 0.3654 - val_acc: 0.8634\n",
            "Epoch 33/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.3269 - acc: 0.8666 - val_loss: 0.3507 - val_acc: 0.8613\n",
            "Epoch 34/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.3240 - acc: 0.8713 - val_loss: 0.3605 - val_acc: 0.8736\n",
            "Epoch 35/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.3211 - acc: 0.8691 - val_loss: 0.3474 - val_acc: 0.8634\n",
            "Epoch 36/100\n",
            "5538/5538 [==============================] - 2s 341us/step - loss: 0.3224 - acc: 0.8754 - val_loss: 0.3370 - val_acc: 0.8786\n",
            "Epoch 37/100\n",
            "5538/5538 [==============================] - 2s 340us/step - loss: 0.3189 - acc: 0.8754 - val_loss: 0.3679 - val_acc: 0.8656\n",
            "Epoch 38/100\n",
            "5538/5538 [==============================] - 2s 348us/step - loss: 0.3171 - acc: 0.8745 - val_loss: 0.3692 - val_acc: 0.8678\n",
            "Epoch 39/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.3096 - acc: 0.8790 - val_loss: 0.3599 - val_acc: 0.8707\n",
            "Epoch 40/100\n",
            "5538/5538 [==============================] - 2s 346us/step - loss: 0.3073 - acc: 0.8806 - val_loss: 0.3415 - val_acc: 0.8736\n",
            "Epoch 41/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.3078 - acc: 0.8808 - val_loss: 0.3476 - val_acc: 0.8685\n",
            "Epoch 42/100\n",
            "5538/5538 [==============================] - 2s 345us/step - loss: 0.3051 - acc: 0.8788 - val_loss: 0.3500 - val_acc: 0.8678\n",
            "Epoch 43/100\n",
            "5538/5538 [==============================] - 2s 353us/step - loss: 0.3059 - acc: 0.8790 - val_loss: 0.3277 - val_acc: 0.8909\n",
            "Epoch 44/100\n",
            "5538/5538 [==============================] - 2s 341us/step - loss: 0.2983 - acc: 0.8844 - val_loss: 0.3545 - val_acc: 0.8743\n",
            "Epoch 45/100\n",
            "5538/5538 [==============================] - 2s 346us/step - loss: 0.3024 - acc: 0.8785 - val_loss: 0.3486 - val_acc: 0.8721\n",
            "Epoch 46/100\n",
            "5538/5538 [==============================] - 2s 348us/step - loss: 0.2988 - acc: 0.8805 - val_loss: 0.3836 - val_acc: 0.8591\n",
            "Epoch 47/100\n",
            "5538/5538 [==============================] - 2s 350us/step - loss: 0.2962 - acc: 0.8839 - val_loss: 0.3610 - val_acc: 0.8620\n",
            "Epoch 48/100\n",
            "5538/5538 [==============================] - 2s 346us/step - loss: 0.2954 - acc: 0.8799 - val_loss: 0.3468 - val_acc: 0.8779\n",
            "Epoch 49/100\n",
            "5538/5538 [==============================] - 2s 354us/step - loss: 0.2992 - acc: 0.8819 - val_loss: 0.3496 - val_acc: 0.8663\n",
            "Epoch 50/100\n",
            "5538/5538 [==============================] - 2s 349us/step - loss: 0.2990 - acc: 0.8797 - val_loss: 0.3393 - val_acc: 0.8757\n",
            "Epoch 51/100\n",
            "5538/5538 [==============================] - 2s 343us/step - loss: 0.2863 - acc: 0.8870 - val_loss: 0.3488 - val_acc: 0.8656\n",
            "Epoch 52/100\n",
            "5538/5538 [==============================] - 2s 345us/step - loss: 0.2883 - acc: 0.8857 - val_loss: 0.4419 - val_acc: 0.8367\n",
            "Epoch 53/100\n",
            "5538/5538 [==============================] - 2s 346us/step - loss: 0.2929 - acc: 0.8850 - val_loss: 0.3338 - val_acc: 0.8837\n",
            "Epoch 54/100\n",
            "5538/5538 [==============================] - 2s 346us/step - loss: 0.2846 - acc: 0.8873 - val_loss: 0.3617 - val_acc: 0.8627\n",
            "Epoch 55/100\n",
            "5538/5538 [==============================] - 2s 346us/step - loss: 0.2897 - acc: 0.8861 - val_loss: 0.3456 - val_acc: 0.8714\n",
            "Epoch 56/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.2838 - acc: 0.8873 - val_loss: 0.3499 - val_acc: 0.8649\n",
            "Epoch 57/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.2880 - acc: 0.8848 - val_loss: 0.3701 - val_acc: 0.8663\n",
            "Epoch 58/100\n",
            "5538/5538 [==============================] - 2s 349us/step - loss: 0.2798 - acc: 0.8897 - val_loss: 0.3587 - val_acc: 0.8743\n",
            "Epoch 59/100\n",
            "5538/5538 [==============================] - 2s 355us/step - loss: 0.2795 - acc: 0.8868 - val_loss: 0.3638 - val_acc: 0.8663\n",
            "Epoch 60/100\n",
            "5538/5538 [==============================] - 2s 358us/step - loss: 0.2752 - acc: 0.8918 - val_loss: 0.3584 - val_acc: 0.8692\n",
            "Epoch 61/100\n",
            "5538/5538 [==============================] - 2s 350us/step - loss: 0.2764 - acc: 0.8902 - val_loss: 0.3591 - val_acc: 0.8685\n",
            "Epoch 62/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.2793 - acc: 0.8871 - val_loss: 0.3624 - val_acc: 0.8663\n",
            "Epoch 63/100\n",
            "5538/5538 [==============================] - 2s 345us/step - loss: 0.2771 - acc: 0.8917 - val_loss: 0.4158 - val_acc: 0.8656\n",
            "Epoch 64/100\n",
            "5538/5538 [==============================] - 2s 348us/step - loss: 0.2779 - acc: 0.8879 - val_loss: 0.3772 - val_acc: 0.8605\n",
            "Epoch 65/100\n",
            "5538/5538 [==============================] - 2s 349us/step - loss: 0.2705 - acc: 0.8900 - val_loss: 0.3480 - val_acc: 0.8801\n",
            "Epoch 66/100\n",
            "5538/5538 [==============================] - 2s 346us/step - loss: 0.2765 - acc: 0.8886 - val_loss: 0.3441 - val_acc: 0.8728\n",
            "Epoch 67/100\n",
            "5538/5538 [==============================] - 2s 346us/step - loss: 0.2721 - acc: 0.8922 - val_loss: 0.3392 - val_acc: 0.8772\n",
            "Epoch 68/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.2673 - acc: 0.8958 - val_loss: 0.3813 - val_acc: 0.8649\n",
            "Epoch 69/100\n",
            "5538/5538 [==============================] - 2s 343us/step - loss: 0.2716 - acc: 0.8931 - val_loss: 0.3599 - val_acc: 0.8750\n",
            "Epoch 70/100\n",
            "5538/5538 [==============================] - 2s 351us/step - loss: 0.2673 - acc: 0.8942 - val_loss: 0.4415 - val_acc: 0.8504\n",
            "Epoch 71/100\n",
            "5538/5538 [==============================] - 2s 351us/step - loss: 0.2731 - acc: 0.8909 - val_loss: 0.3927 - val_acc: 0.8497\n",
            "Epoch 72/100\n",
            "5538/5538 [==============================] - 2s 345us/step - loss: 0.2692 - acc: 0.8902 - val_loss: 0.3781 - val_acc: 0.8685\n",
            "Epoch 73/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.2669 - acc: 0.8915 - val_loss: 0.3709 - val_acc: 0.8634\n",
            "Epoch 74/100\n",
            "5538/5538 [==============================] - 2s 345us/step - loss: 0.2673 - acc: 0.8938 - val_loss: 0.3749 - val_acc: 0.8692\n",
            "Epoch 75/100\n",
            "5538/5538 [==============================] - 2s 346us/step - loss: 0.2658 - acc: 0.8889 - val_loss: 0.3671 - val_acc: 0.8822\n",
            "Epoch 76/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.2614 - acc: 0.8922 - val_loss: 0.3504 - val_acc: 0.8764\n",
            "Epoch 77/100\n",
            "5538/5538 [==============================] - 2s 343us/step - loss: 0.2607 - acc: 0.8976 - val_loss: 0.3669 - val_acc: 0.8728\n",
            "Epoch 78/100\n",
            "5538/5538 [==============================] - 2s 346us/step - loss: 0.2646 - acc: 0.8940 - val_loss: 0.3763 - val_acc: 0.8598\n",
            "Epoch 79/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.2569 - acc: 0.8987 - val_loss: 0.3479 - val_acc: 0.8764\n",
            "Epoch 80/100\n",
            "5538/5538 [==============================] - 2s 340us/step - loss: 0.2596 - acc: 0.8962 - val_loss: 0.3626 - val_acc: 0.8685\n",
            "Epoch 81/100\n",
            "5538/5538 [==============================] - 2s 348us/step - loss: 0.2612 - acc: 0.8958 - val_loss: 0.3429 - val_acc: 0.8866\n",
            "Epoch 82/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.2561 - acc: 0.8976 - val_loss: 0.3571 - val_acc: 0.8685\n",
            "Epoch 83/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.2596 - acc: 0.8967 - val_loss: 0.3524 - val_acc: 0.8699\n",
            "Epoch 84/100\n",
            "5538/5538 [==============================] - 2s 348us/step - loss: 0.2584 - acc: 0.8929 - val_loss: 0.3799 - val_acc: 0.8577\n",
            "Epoch 85/100\n",
            "5538/5538 [==============================] - 2s 348us/step - loss: 0.2516 - acc: 0.8978 - val_loss: 0.4059 - val_acc: 0.8649\n",
            "Epoch 86/100\n",
            "5538/5538 [==============================] - 2s 345us/step - loss: 0.2583 - acc: 0.8960 - val_loss: 0.4211 - val_acc: 0.8410\n",
            "Epoch 87/100\n",
            "5538/5538 [==============================] - 2s 352us/step - loss: 0.2477 - acc: 0.9034 - val_loss: 0.3553 - val_acc: 0.8699\n",
            "Epoch 88/100\n",
            "5538/5538 [==============================] - 2s 353us/step - loss: 0.2581 - acc: 0.8982 - val_loss: 0.3478 - val_acc: 0.8743\n",
            "Epoch 89/100\n",
            "5538/5538 [==============================] - 2s 351us/step - loss: 0.2544 - acc: 0.8967 - val_loss: 0.3787 - val_acc: 0.8692\n",
            "Epoch 90/100\n",
            "5538/5538 [==============================] - 2s 349us/step - loss: 0.2537 - acc: 0.8954 - val_loss: 0.3693 - val_acc: 0.8649\n",
            "Epoch 91/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.2530 - acc: 0.8974 - val_loss: 0.3641 - val_acc: 0.8699\n",
            "Epoch 92/100\n",
            "5538/5538 [==============================] - 2s 354us/step - loss: 0.2486 - acc: 0.8994 - val_loss: 0.4622 - val_acc: 0.8353\n",
            "Epoch 93/100\n",
            "5538/5538 [==============================] - 2s 345us/step - loss: 0.2560 - acc: 0.8933 - val_loss: 0.3714 - val_acc: 0.8627\n",
            "Epoch 94/100\n",
            "5538/5538 [==============================] - 2s 347us/step - loss: 0.2528 - acc: 0.8964 - val_loss: 0.3625 - val_acc: 0.8699\n",
            "Epoch 95/100\n",
            "5538/5538 [==============================] - 2s 345us/step - loss: 0.2478 - acc: 0.9014 - val_loss: 0.3498 - val_acc: 0.8844\n",
            "Epoch 96/100\n",
            "5538/5538 [==============================] - 2s 342us/step - loss: 0.2494 - acc: 0.8985 - val_loss: 0.3739 - val_acc: 0.8736\n",
            "Epoch 97/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.2484 - acc: 0.9009 - val_loss: 0.3647 - val_acc: 0.8692\n",
            "Epoch 98/100\n",
            "5538/5538 [==============================] - 2s 350us/step - loss: 0.2485 - acc: 0.9003 - val_loss: 0.3545 - val_acc: 0.8728\n",
            "Epoch 99/100\n",
            "5538/5538 [==============================] - 2s 344us/step - loss: 0.2475 - acc: 0.8992 - val_loss: 0.3685 - val_acc: 0.8801\n",
            "Epoch 100/100\n",
            "5538/5538 [==============================] - 2s 341us/step - loss: 0.2449 - acc: 0.8996 - val_loss: 0.3736 - val_acc: 0.8757\n",
            "Fold score (accuracy): 0.8757225433526011\n",
            "Final score (accuracy): 0.8627564287778099\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JUxzWQ9FbhkP",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}