Code_of_Thesis(2s_1sec_overlap)_10.ipynb

{
  "cells": [
    {
      "cell_type": "code",
      "source": [
        "from tensorflow.python.client import device_lib\n",
        "device_lib.list_local_devices()"
      ],
      "metadata": {
        "id": "tJiwtyiYOwr7",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "8c26eef3-9f71-46e7-9e65-e4809f8151af"
      },
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[name: \"/device:CPU:0\"\n",
              " device_type: \"CPU\"\n",
              " memory_limit: 268435456\n",
              " locality {\n",
              " }\n",
              " incarnation: 13722673948569957018\n",
              " xla_global_id: -1, name: \"/device:GPU:0\"\n",
              " device_type: \"GPU\"\n",
              " memory_limit: 14391508992\n",
              " locality {\n",
              "   bus_id: 1\n",
              "   links {\n",
              "   }\n",
              " }\n",
              " incarnation: 7517019587670175547\n",
              " physical_device_desc: \"device: 0, name: Tesla T4, pci bus id: 0000:00:04.0, compute capability: 7.5\"\n",
              " xla_global_id: 416903419]"
            ]
          },
          "metadata": {},
          "execution_count": 1
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "metadata": {
        "id": "PyL1kvEKPD7e"
      },
      "outputs": [],
      "source": [
        "import matplotlib.pyplot as plt\n",
        "from sklearn import metrics\n",
        "from keras import layers\n",
        "from keras import models\n",
        "import tensorflow as tf\n",
        "from scipy import io\n",
        "import numpy as np\n",
        "import keras\n",
        "import cv2"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "metadata": {
        "id": "qIekqYpW55dE"
      },
      "outputs": [],
      "source": [
        "trs = training_sample_size=12000     #\n",
        "tes = test_sample_size=1798          #\n",
        "\n",
        "dataset = io.loadmat('/content/drive/MyDrive/uni project/2_Class/2sec(1sec_overlap)/D_2c_shuffle_normal_2s(1sec overlap).mat')\n",
        "\n",
        "\n",
        "x_train_original = np.squeeze(dataset['data'][:trs])\n",
        "x_test_original  = np.squeeze(dataset['data'][trs:trs+tes])\n",
        "\n",
        "y_train_original = np.squeeze(dataset['label'][:trs])\n",
        "y_test_original  = np.squeeze(dataset['label'][trs:trs+tes])\n",
        "\n",
        "x_train_dsize = [cv2.resize(img, dsize=(4, 1024)) for img in x_train_original]\n",
        "x_test_dsize = [cv2.resize(img, dsize= (4, 1024)) for img in x_test_original]\n",
        "\n",
        "x_train_original_re = [x.reshape(1024*4) for x in x_train_dsize]\n",
        "x_test_original_re = [x.reshape(1024*4) for x in x_test_dsize]\n",
        "\n",
        "x_train = np.array(x_train_original_re)\n",
        "x_test = np.array(x_test_original_re)\n",
        "\n",
        "x_train = x_train.astype('float32')\n",
        "x_test  = x_test.astype('float32')\n",
        "\n",
        "x_train = x_train.reshape(-1,1024,4,1)\n",
        "x_test  = x_test.reshape(-1,1024,4,1)\n",
        "\n",
        "y_train = keras.utils.to_categorical(y_train_original, num_classes=2)\n",
        "y_test = keras.utils.to_categorical(y_test_original, num_classes=2)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model = models.Sequential()\n",
        "\n",
        "model.add(layers.Conv2D(8,  kernel_size=(4, 4),activation='relu', padding='same',input_shape=(1024,4,1)))\n",
        "model.add(layers.BatchNormalization())\n",
        "\n",
        "model.add(layers.Conv2D(16, kernel_size=(3, 3),activation='relu', padding='same'))\n",
        "model.add(layers.MaxPooling2D((2,2)))\n",
        "model.add(layers.BatchNormalization())\n",
        "\n",
        "model.add(layers.Conv2D(32, kernel_size=(3, 3),activation='relu', padding='same'))\n",
        "model.add(layers.MaxPooling2D((2,2)))\n",
        "model.add(layers.BatchNormalization())\n",
        "\n",
        "model.add(layers.Dropout(0.5))\n",
        "model.add(layers.Flatten())\n",
        "\n",
        "model.add(layers.Dense(128, activation='relu'))\n",
        "model.add(layers.Dropout(0.5))\n",
        "\n",
        "model.add(layers.ActivityRegularization(l1=0.003, l2=0.003))\n",
        "\n",
        "model.add(layers.Dense(2, activation='softmax'))"
      ],
      "metadata": {
        "id": "AlmIVvEKIT72"
      },
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": 8,
      "metadata": {
        "id": "2-cNZStZaFXO",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "991ecdf6-3ff6-46b7-d70a-4de77d1f9295"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " conv2d (Conv2D)             (None, 1024, 4, 8)        136       \n",
            "                                                                 \n",
            " batch_normalization (BatchN  (None, 1024, 4, 8)       32        \n",
            " ormalization)                                                   \n",
            "                                                                 \n",
            " conv2d_1 (Conv2D)           (None, 1024, 4, 16)       1168      \n",
            "                                                                 \n",
            " max_pooling2d (MaxPooling2D  (None, 512, 2, 16)       0         \n",
            " )                                                               \n",
            "                                                                 \n",
            " batch_normalization_1 (Batc  (None, 512, 2, 16)       64        \n",
            " hNormalization)                                                 \n",
            "                                                                 \n",
            " conv2d_2 (Conv2D)           (None, 512, 2, 32)        4640      \n",
            "                                                                 \n",
            " max_pooling2d_1 (MaxPooling  (None, 256, 1, 32)       0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " batch_normalization_2 (Batc  (None, 256, 1, 32)       128       \n",
            " hNormalization)                                                 \n",
            "                                                                 \n",
            " dropout (Dropout)           (None, 256, 1, 32)        0         \n",
            "                                                                 \n",
            " flatten (Flatten)           (None, 8192)              0         \n",
            "                                                                 \n",
            " dense (Dense)               (None, 128)               1048704   \n",
            "                                                                 \n",
            " dropout_1 (Dropout)         (None, 128)               0         \n",
            "                                                                 \n",
            " activity_regularization (Ac  (None, 128)              0         \n",
            " tivityRegularization)                                           \n",
            "                                                                 \n",
            " dense_1 (Dense)             (None, 2)                 258       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 1,055,130\n",
            "Trainable params: 1,055,018\n",
            "Non-trainable params: 112\n",
            "_________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "model.summary()"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "dot_img_file = './drive/MyDrive/uni project/2sec(1sec_overlap)/model_CNN.png'\n",
        "tf.keras.utils.plot_model(model, to_file=dot_img_file, show_shapes=True)"
      ],
      "metadata": {
        "id": "BaHJ71X94UXs",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "db174aba-ef0c-4a14-88f2-53155f0884c4"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {},
          "execution_count": 3
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "-Jxza8aX2boL",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d4fbc891-390e-482f-8ef0-073552fdb18f"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/400\n",
            "160/160 [==============================] - 11s 14ms/step - loss: 1.5027 - accuracy: 0.6439 - val_loss: 5.5271 - val_accuracy: 0.5044\n",
            "Epoch 2/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.6431 - accuracy: 0.7031 - val_loss: 8.6072 - val_accuracy: 0.5044\n",
            "Epoch 3/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.6082 - accuracy: 0.7278 - val_loss: 3.6001 - val_accuracy: 0.5050\n",
            "Epoch 4/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.5865 - accuracy: 0.7441 - val_loss: 0.5866 - val_accuracy: 0.7272\n",
            "Epoch 5/400\n",
            "160/160 [==============================] - 2s 11ms/step - loss: 0.5695 - accuracy: 0.7625 - val_loss: 0.4918 - val_accuracy: 0.8172\n",
            "Epoch 6/400\n",
            "160/160 [==============================] - 2s 11ms/step - loss: 0.5470 - accuracy: 0.7884 - val_loss: 0.4826 - val_accuracy: 0.8294\n",
            "Epoch 7/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.5375 - accuracy: 0.7934 - val_loss: 0.4752 - val_accuracy: 0.8311\n",
            "Epoch 8/400\n",
            "160/160 [==============================] - 2s 11ms/step - loss: 0.5177 - accuracy: 0.8107 - val_loss: 0.4651 - val_accuracy: 0.8428\n",
            "Epoch 9/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.5030 - accuracy: 0.8186 - val_loss: 0.4552 - val_accuracy: 0.8400\n",
            "Epoch 10/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.4883 - accuracy: 0.8295 - val_loss: 0.4435 - val_accuracy: 0.8311\n",
            "Epoch 11/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.4748 - accuracy: 0.8448 - val_loss: 0.4346 - val_accuracy: 0.8489\n",
            "Epoch 12/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.4593 - accuracy: 0.8499 - val_loss: 0.4245 - val_accuracy: 0.8378\n",
            "Epoch 13/400\n",
            "160/160 [==============================] - 2s 11ms/step - loss: 0.4509 - accuracy: 0.8554 - val_loss: 0.3881 - val_accuracy: 0.8694\n",
            "Epoch 14/400\n",
            "160/160 [==============================] - 2s 11ms/step - loss: 0.4301 - accuracy: 0.8646 - val_loss: 0.3902 - val_accuracy: 0.8728\n",
            "Epoch 15/400\n",
            "160/160 [==============================] - 2s 11ms/step - loss: 0.4189 - accuracy: 0.8724 - val_loss: 0.3854 - val_accuracy: 0.8756\n",
            "Epoch 16/400\n",
            "160/160 [==============================] - 2s 11ms/step - loss: 0.4080 - accuracy: 0.8761 - val_loss: 0.3592 - val_accuracy: 0.8889\n",
            "Epoch 17/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.3877 - accuracy: 0.8878 - val_loss: 0.3602 - val_accuracy: 0.8772\n",
            "Epoch 18/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.3756 - accuracy: 0.8956 - val_loss: 0.3530 - val_accuracy: 0.8794\n",
            "Epoch 19/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.3573 - accuracy: 0.9027 - val_loss: 0.3402 - val_accuracy: 0.8906\n",
            "Epoch 20/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.3479 - accuracy: 0.9042 - val_loss: 0.3199 - val_accuracy: 0.9050\n",
            "Epoch 21/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.3378 - accuracy: 0.9061 - val_loss: 0.3477 - val_accuracy: 0.8739\n",
            "Epoch 22/400\n",
            "160/160 [==============================] - 2s 11ms/step - loss: 0.3259 - accuracy: 0.9149 - val_loss: 0.2951 - val_accuracy: 0.9111\n",
            "Epoch 23/400\n",
            "160/160 [==============================] - 2s 11ms/step - loss: 0.3135 - accuracy: 0.9202 - val_loss: 0.3210 - val_accuracy: 0.8989\n",
            "Epoch 24/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.3015 - accuracy: 0.9231 - val_loss: 0.2782 - val_accuracy: 0.9183\n",
            "Epoch 25/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.2919 - accuracy: 0.9266 - val_loss: 0.2819 - val_accuracy: 0.9100\n",
            "Epoch 26/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.2787 - accuracy: 0.9295 - val_loss: 0.2582 - val_accuracy: 0.9317\n",
            "Epoch 27/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.2762 - accuracy: 0.9329 - val_loss: 0.2928 - val_accuracy: 0.9111\n",
            "Epoch 28/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.2689 - accuracy: 0.9352 - val_loss: 0.2535 - val_accuracy: 0.9278\n",
            "Epoch 29/400\n",
            "160/160 [==============================] - 2s 11ms/step - loss: 0.2580 - accuracy: 0.9390 - val_loss: 0.2801 - val_accuracy: 0.9089\n",
            "Epoch 30/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.2490 - accuracy: 0.9397 - val_loss: 0.2757 - val_accuracy: 0.9139\n",
            "Epoch 31/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.2411 - accuracy: 0.9438 - val_loss: 0.2796 - val_accuracy: 0.9094\n",
            "Epoch 32/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.2291 - accuracy: 0.9499 - val_loss: 0.2321 - val_accuracy: 0.9372\n",
            "Epoch 33/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.2228 - accuracy: 0.9531 - val_loss: 0.2473 - val_accuracy: 0.9283\n",
            "Epoch 34/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.2196 - accuracy: 0.9535 - val_loss: 0.2330 - val_accuracy: 0.9356\n",
            "Epoch 35/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.2119 - accuracy: 0.9546 - val_loss: 0.2441 - val_accuracy: 0.9333\n",
            "Epoch 36/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.2115 - accuracy: 0.9542 - val_loss: 0.2872 - val_accuracy: 0.9128\n",
            "Epoch 37/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.2012 - accuracy: 0.9560 - val_loss: 0.2234 - val_accuracy: 0.9372\n",
            "Epoch 38/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.2028 - accuracy: 0.9561 - val_loss: 0.2229 - val_accuracy: 0.9417\n",
            "Epoch 39/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.1922 - accuracy: 0.9598 - val_loss: 0.2434 - val_accuracy: 0.9278\n",
            "Epoch 40/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.1931 - accuracy: 0.9597 - val_loss: 0.2243 - val_accuracy: 0.9322\n",
            "Epoch 41/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.1849 - accuracy: 0.9622 - val_loss: 0.2193 - val_accuracy: 0.9344\n",
            "Epoch 42/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.1759 - accuracy: 0.9675 - val_loss: 0.2201 - val_accuracy: 0.9428\n",
            "Epoch 43/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.1725 - accuracy: 0.9674 - val_loss: 0.2096 - val_accuracy: 0.9367\n",
            "Epoch 44/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.1733 - accuracy: 0.9653 - val_loss: 0.2198 - val_accuracy: 0.9383\n",
            "Epoch 45/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.1674 - accuracy: 0.9668 - val_loss: 0.2029 - val_accuracy: 0.9400\n",
            "Epoch 46/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.1619 - accuracy: 0.9694 - val_loss: 0.2053 - val_accuracy: 0.9422\n",
            "Epoch 47/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.1605 - accuracy: 0.9684 - val_loss: 0.2217 - val_accuracy: 0.9361\n",
            "Epoch 48/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.1579 - accuracy: 0.9711 - val_loss: 0.2141 - val_accuracy: 0.9406\n",
            "Epoch 49/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.1525 - accuracy: 0.9727 - val_loss: 0.2204 - val_accuracy: 0.9383\n",
            "Epoch 50/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.1520 - accuracy: 0.9710 - val_loss: 0.2001 - val_accuracy: 0.9461\n",
            "Epoch 51/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.1497 - accuracy: 0.9727 - val_loss: 0.2034 - val_accuracy: 0.9450\n",
            "Epoch 52/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.1495 - accuracy: 0.9718 - val_loss: 0.2132 - val_accuracy: 0.9394\n",
            "Epoch 53/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.1440 - accuracy: 0.9740 - val_loss: 0.2006 - val_accuracy: 0.9428\n",
            "Epoch 54/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.1371 - accuracy: 0.9769 - val_loss: 0.1946 - val_accuracy: 0.9472\n",
            "Epoch 55/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.1383 - accuracy: 0.9737 - val_loss: 0.1915 - val_accuracy: 0.9428\n",
            "Epoch 56/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.1349 - accuracy: 0.9763 - val_loss: 0.2187 - val_accuracy: 0.9394\n",
            "Epoch 57/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.1328 - accuracy: 0.9776 - val_loss: 0.1939 - val_accuracy: 0.9428\n",
            "Epoch 58/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.1318 - accuracy: 0.9767 - val_loss: 0.1919 - val_accuracy: 0.9506\n",
            "Epoch 59/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.1251 - accuracy: 0.9790 - val_loss: 0.1862 - val_accuracy: 0.9483\n",
            "Epoch 60/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.1293 - accuracy: 0.9765 - val_loss: 0.1882 - val_accuracy: 0.9472\n",
            "Epoch 61/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.1202 - accuracy: 0.9812 - val_loss: 0.2413 - val_accuracy: 0.9172\n",
            "Epoch 62/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.1268 - accuracy: 0.9766 - val_loss: 0.2317 - val_accuracy: 0.9328\n",
            "Epoch 63/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.1243 - accuracy: 0.9777 - val_loss: 0.1939 - val_accuracy: 0.9439\n",
            "Epoch 64/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.1211 - accuracy: 0.9775 - val_loss: 0.2007 - val_accuracy: 0.9439\n",
            "Epoch 65/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.1133 - accuracy: 0.9796 - val_loss: 0.1879 - val_accuracy: 0.9483\n",
            "Epoch 68/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.1093 - accuracy: 0.9819 - val_loss: 0.1987 - val_accuracy: 0.9417\n",
            "Epoch 69/400\n",
            "160/160 [==============================] - 2s 11ms/step - loss: 0.1116 - accuracy: 0.9802 - val_loss: 0.2121 - val_accuracy: 0.9433\n",
            "Epoch 70/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.1055 - accuracy: 0.9830 - val_loss: 0.1876 - val_accuracy: 0.9467\n",
            "Epoch 71/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.1060 - accuracy: 0.9820 - val_loss: 0.1864 - val_accuracy: 0.9461\n",
            "Epoch 72/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.1024 - accuracy: 0.9844 - val_loss: 0.1858 - val_accuracy: 0.9389\n",
            "Epoch 73/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.1047 - accuracy: 0.9837 - val_loss: 0.1976 - val_accuracy: 0.9428\n",
            "Epoch 74/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.1016 - accuracy: 0.9847 - val_loss: 0.1981 - val_accuracy: 0.9456\n",
            "Epoch 75/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.1025 - accuracy: 0.9825 - val_loss: 0.2508 - val_accuracy: 0.9328\n",
            "Epoch 76/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.1111 - accuracy: 0.9808 - val_loss: 0.2080 - val_accuracy: 0.9444\n",
            "Epoch 77/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.1033 - accuracy: 0.9816 - val_loss: 0.1963 - val_accuracy: 0.9478\n",
            "Epoch 78/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.1047 - accuracy: 0.9816 - val_loss: 0.1777 - val_accuracy: 0.9522\n",
            "Epoch 79/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0984 - accuracy: 0.9846 - val_loss: 0.1983 - val_accuracy: 0.9472\n",
            "Epoch 80/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0974 - accuracy: 0.9836 - val_loss: 0.1898 - val_accuracy: 0.9472\n",
            "Epoch 81/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0970 - accuracy: 0.9834 - val_loss: 0.1962 - val_accuracy: 0.9478\n",
            "Epoch 82/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0931 - accuracy: 0.9858 - val_loss: 0.1853 - val_accuracy: 0.9456\n",
            "Epoch 83/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0908 - accuracy: 0.9872 - val_loss: 0.2058 - val_accuracy: 0.9456\n",
            "Epoch 84/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0923 - accuracy: 0.9862 - val_loss: 0.2002 - val_accuracy: 0.9500\n",
            "Epoch 85/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0934 - accuracy: 0.9857 - val_loss: 0.2082 - val_accuracy: 0.9406\n",
            "Epoch 86/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0911 - accuracy: 0.9857 - val_loss: 0.1793 - val_accuracy: 0.9500\n",
            "Epoch 87/400\n",
            "160/160 [==============================] - 2s 14ms/step - loss: 0.0946 - accuracy: 0.9843 - val_loss: 0.1713 - val_accuracy: 0.9483\n",
            "Epoch 88/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0935 - accuracy: 0.9821 - val_loss: 0.1957 - val_accuracy: 0.9472\n",
            "Epoch 89/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0898 - accuracy: 0.9854 - val_loss: 0.2620 - val_accuracy: 0.9339\n",
            "Epoch 90/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0885 - accuracy: 0.9865 - val_loss: 0.1831 - val_accuracy: 0.9450\n",
            "Epoch 91/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0909 - accuracy: 0.9852 - val_loss: 0.1650 - val_accuracy: 0.9528\n",
            "Epoch 92/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0892 - accuracy: 0.9848 - val_loss: 0.1830 - val_accuracy: 0.9483\n",
            "Epoch 93/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0828 - accuracy: 0.9873 - val_loss: 0.1988 - val_accuracy: 0.9478\n",
            "Epoch 94/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0854 - accuracy: 0.9854 - val_loss: 0.2032 - val_accuracy: 0.9483\n",
            "Epoch 95/400\n",
            "160/160 [==============================] - 2s 14ms/step - loss: 0.0855 - accuracy: 0.9863 - val_loss: 0.2374 - val_accuracy: 0.9267\n",
            "Epoch 96/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0827 - accuracy: 0.9875 - val_loss: 0.1769 - val_accuracy: 0.9422\n",
            "Epoch 97/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0820 - accuracy: 0.9875 - val_loss: 0.1654 - val_accuracy: 0.9533\n",
            "Epoch 98/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0818 - accuracy: 0.9875 - val_loss: 0.1761 - val_accuracy: 0.9489\n",
            "Epoch 99/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0782 - accuracy: 0.9884 - val_loss: 0.1842 - val_accuracy: 0.9478\n",
            "Epoch 100/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0794 - accuracy: 0.9879 - val_loss: 0.1804 - val_accuracy: 0.9433\n",
            "Epoch 101/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0823 - accuracy: 0.9858 - val_loss: 0.1736 - val_accuracy: 0.9506\n",
            "Epoch 102/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0827 - accuracy: 0.9860 - val_loss: 0.1682 - val_accuracy: 0.9439\n",
            "Epoch 103/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0790 - accuracy: 0.9877 - val_loss: 0.1611 - val_accuracy: 0.9522\n",
            "Epoch 104/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0792 - accuracy: 0.9870 - val_loss: 0.1890 - val_accuracy: 0.9483\n",
            "Epoch 105/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0750 - accuracy: 0.9886 - val_loss: 0.1792 - val_accuracy: 0.9467\n",
            "Epoch 106/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0767 - accuracy: 0.9897 - val_loss: 0.1676 - val_accuracy: 0.9461\n",
            "Epoch 107/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0810 - accuracy: 0.9860 - val_loss: 0.1881 - val_accuracy: 0.9489\n",
            "Epoch 108/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0727 - accuracy: 0.9886 - val_loss: 0.1850 - val_accuracy: 0.9517\n",
            "Epoch 109/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0702 - accuracy: 0.9899 - val_loss: 0.1821 - val_accuracy: 0.9450\n",
            "Epoch 110/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0748 - accuracy: 0.9882 - val_loss: 0.1887 - val_accuracy: 0.9533\n",
            "Epoch 111/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0738 - accuracy: 0.9894 - val_loss: 0.1745 - val_accuracy: 0.9528\n",
            "Epoch 112/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0724 - accuracy: 0.9881 - val_loss: 0.1805 - val_accuracy: 0.9350\n",
            "Epoch 113/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0707 - accuracy: 0.9896 - val_loss: 0.1752 - val_accuracy: 0.9522\n",
            "Epoch 114/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0833 - accuracy: 0.9844 - val_loss: 0.1760 - val_accuracy: 0.9483\n",
            "Epoch 115/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0764 - accuracy: 0.9865 - val_loss: 0.2303 - val_accuracy: 0.9467\n",
            "Epoch 116/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0731 - accuracy: 0.9879 - val_loss: 0.1840 - val_accuracy: 0.9500\n",
            "Epoch 117/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0710 - accuracy: 0.9896 - val_loss: 0.2068 - val_accuracy: 0.9472\n",
            "Epoch 118/400\n",
            "160/160 [==============================] - 2s 11ms/step - loss: 0.0696 - accuracy: 0.9894 - val_loss: 0.1732 - val_accuracy: 0.9489\n",
            "Epoch 119/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0685 - accuracy: 0.9901 - val_loss: 0.1721 - val_accuracy: 0.9583\n",
            "Epoch 120/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0743 - accuracy: 0.9874 - val_loss: 0.1786 - val_accuracy: 0.9489\n",
            "Epoch 121/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0669 - accuracy: 0.9904 - val_loss: 0.1649 - val_accuracy: 0.9506\n",
            "Epoch 122/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0723 - accuracy: 0.9874 - val_loss: 0.2058 - val_accuracy: 0.9489\n",
            "Epoch 123/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0694 - accuracy: 0.9892 - val_loss: 0.2344 - val_accuracy: 0.9406\n",
            "Epoch 124/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0709 - accuracy: 0.9889 - val_loss: 0.1784 - val_accuracy: 0.9522\n",
            "Epoch 125/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0638 - accuracy: 0.9914 - val_loss: 0.2135 - val_accuracy: 0.9417\n",
            "Epoch 126/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0666 - accuracy: 0.9898 - val_loss: 0.2009 - val_accuracy: 0.9506\n",
            "Epoch 127/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0688 - accuracy: 0.9888 - val_loss: 0.1635 - val_accuracy: 0.9578\n",
            "Epoch 128/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0697 - accuracy: 0.9890 - val_loss: 0.1896 - val_accuracy: 0.9483\n",
            "Epoch 129/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0614 - accuracy: 0.9920 - val_loss: 0.1957 - val_accuracy: 0.9500\n",
            "Epoch 130/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0647 - accuracy: 0.9896 - val_loss: 0.1551 - val_accuracy: 0.9522\n",
            "Epoch 131/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0665 - accuracy: 0.9898 - val_loss: 0.2599 - val_accuracy: 0.9394\n",
            "Epoch 132/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0621 - accuracy: 0.9919 - val_loss: 0.1685 - val_accuracy: 0.9483\n",
            "Epoch 133/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0680 - accuracy: 0.9887 - val_loss: 0.1866 - val_accuracy: 0.9522\n",
            "Epoch 134/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0613 - accuracy: 0.9903 - val_loss: 0.1945 - val_accuracy: 0.9511\n",
            "Epoch 135/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0699 - accuracy: 0.9884 - val_loss: 0.2006 - val_accuracy: 0.9500\n",
            "Epoch 136/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0642 - accuracy: 0.9890 - val_loss: 0.1883 - val_accuracy: 0.9467\n",
            "Epoch 137/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0654 - accuracy: 0.9904 - val_loss: 0.1972 - val_accuracy: 0.9483\n",
            "Epoch 138/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0611 - accuracy: 0.9915 - val_loss: 0.2075 - val_accuracy: 0.9472\n",
            "Epoch 139/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0596 - accuracy: 0.9921 - val_loss: 0.1760 - val_accuracy: 0.9528\n",
            "Epoch 140/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0647 - accuracy: 0.9895 - val_loss: 0.1758 - val_accuracy: 0.9594\n",
            "Epoch 141/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0607 - accuracy: 0.9915 - val_loss: 0.1728 - val_accuracy: 0.9572\n",
            "Epoch 142/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0604 - accuracy: 0.9918 - val_loss: 0.1789 - val_accuracy: 0.9533\n",
            "Epoch 143/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0620 - accuracy: 0.9913 - val_loss: 0.1979 - val_accuracy: 0.9544\n",
            "Epoch 144/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0564 - accuracy: 0.9931 - val_loss: 0.1717 - val_accuracy: 0.9544\n",
            "Epoch 145/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0587 - accuracy: 0.9916 - val_loss: 0.2465 - val_accuracy: 0.9444\n",
            "Epoch 146/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0618 - accuracy: 0.9906 - val_loss: 0.1978 - val_accuracy: 0.9506\n",
            "Epoch 147/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0568 - accuracy: 0.9930 - val_loss: 0.1567 - val_accuracy: 0.9528\n",
            "Epoch 148/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0608 - accuracy: 0.9899 - val_loss: 0.2023 - val_accuracy: 0.9478\n",
            "Epoch 149/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0644 - accuracy: 0.9877 - val_loss: 0.2035 - val_accuracy: 0.9489\n",
            "Epoch 150/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0615 - accuracy: 0.9899 - val_loss: 0.2461 - val_accuracy: 0.9483\n",
            "Epoch 151/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0590 - accuracy: 0.9914 - val_loss: 0.1809 - val_accuracy: 0.9567\n",
            "Epoch 152/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0609 - accuracy: 0.9913 - val_loss: 0.1680 - val_accuracy: 0.9444\n",
            "Epoch 153/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0590 - accuracy: 0.9910 - val_loss: 0.1632 - val_accuracy: 0.9589\n",
            "Epoch 154/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0572 - accuracy: 0.9911 - val_loss: 0.1817 - val_accuracy: 0.9556\n",
            "Epoch 155/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0579 - accuracy: 0.9911 - val_loss: 0.2024 - val_accuracy: 0.9489\n",
            "Epoch 156/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0537 - accuracy: 0.9931 - val_loss: 0.2144 - val_accuracy: 0.9467\n",
            "Epoch 157/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0558 - accuracy: 0.9929 - val_loss: 0.1961 - val_accuracy: 0.9539\n",
            "Epoch 158/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0599 - accuracy: 0.9885 - val_loss: 0.1752 - val_accuracy: 0.9539\n",
            "Epoch 159/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0584 - accuracy: 0.9914 - val_loss: 0.1536 - val_accuracy: 0.9522\n",
            "Epoch 160/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0619 - accuracy: 0.9904 - val_loss: 0.2558 - val_accuracy: 0.9444\n",
            "Epoch 161/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0582 - accuracy: 0.9910 - val_loss: 0.1601 - val_accuracy: 0.9500\n",
            "Epoch 162/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0595 - accuracy: 0.9896 - val_loss: 0.1902 - val_accuracy: 0.9533\n",
            "Epoch 163/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0537 - accuracy: 0.9935 - val_loss: 0.1616 - val_accuracy: 0.9522\n",
            "Epoch 164/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0529 - accuracy: 0.9933 - val_loss: 0.1930 - val_accuracy: 0.9550\n",
            "Epoch 165/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0519 - accuracy: 0.9928 - val_loss: 0.1732 - val_accuracy: 0.9517\n",
            "Epoch 166/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0550 - accuracy: 0.9913 - val_loss: 0.1841 - val_accuracy: 0.9528\n",
            "Epoch 168/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0589 - accuracy: 0.9912 - val_loss: 0.3372 - val_accuracy: 0.9344\n",
            "Epoch 169/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0591 - accuracy: 0.9900 - val_loss: 0.1653 - val_accuracy: 0.9556\n",
            "Epoch 170/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0525 - accuracy: 0.9921 - val_loss: 0.1984 - val_accuracy: 0.9472\n",
            "Epoch 171/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0529 - accuracy: 0.9927 - val_loss: 0.1890 - val_accuracy: 0.9428\n",
            "Epoch 172/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0554 - accuracy: 0.9905 - val_loss: 0.1664 - val_accuracy: 0.9539\n",
            "Epoch 173/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0523 - accuracy: 0.9922 - val_loss: 0.1672 - val_accuracy: 0.9544\n",
            "Epoch 174/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0549 - accuracy: 0.9922 - val_loss: 0.1955 - val_accuracy: 0.9517\n",
            "Epoch 175/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0533 - accuracy: 0.9916 - val_loss: 0.1852 - val_accuracy: 0.9606\n",
            "Epoch 176/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0517 - accuracy: 0.9914 - val_loss: 0.1594 - val_accuracy: 0.9522\n",
            "Epoch 177/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0536 - accuracy: 0.9918 - val_loss: 0.1827 - val_accuracy: 0.9494\n",
            "Epoch 178/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0575 - accuracy: 0.9907 - val_loss: 0.2305 - val_accuracy: 0.9494\n",
            "Epoch 179/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0549 - accuracy: 0.9914 - val_loss: 0.1977 - val_accuracy: 0.9500\n",
            "Epoch 180/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0537 - accuracy: 0.9927 - val_loss: 0.2693 - val_accuracy: 0.9422\n",
            "Epoch 181/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0504 - accuracy: 0.9939 - val_loss: 0.1884 - val_accuracy: 0.9561\n",
            "Epoch 182/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0517 - accuracy: 0.9915 - val_loss: 0.2037 - val_accuracy: 0.9389\n",
            "Epoch 183/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0542 - accuracy: 0.9919 - val_loss: 0.1689 - val_accuracy: 0.9533\n",
            "Epoch 184/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0502 - accuracy: 0.9932 - val_loss: 0.1643 - val_accuracy: 0.9561\n",
            "Epoch 185/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0510 - accuracy: 0.9933 - val_loss: 0.2092 - val_accuracy: 0.9494\n",
            "Epoch 186/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0512 - accuracy: 0.9929 - val_loss: 0.2194 - val_accuracy: 0.9500\n",
            "Epoch 187/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0506 - accuracy: 0.9928 - val_loss: 0.1911 - val_accuracy: 0.9539\n",
            "Epoch 188/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0489 - accuracy: 0.9930 - val_loss: 0.2182 - val_accuracy: 0.9522\n",
            "Epoch 189/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0505 - accuracy: 0.9935 - val_loss: 0.2348 - val_accuracy: 0.9517\n",
            "Epoch 190/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0471 - accuracy: 0.9949 - val_loss: 0.1623 - val_accuracy: 0.9606\n",
            "Epoch 191/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0502 - accuracy: 0.9927 - val_loss: 0.2634 - val_accuracy: 0.9478\n",
            "Epoch 192/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0511 - accuracy: 0.9925 - val_loss: 0.2088 - val_accuracy: 0.9517\n",
            "Epoch 193/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0520 - accuracy: 0.9926 - val_loss: 0.1680 - val_accuracy: 0.9506\n",
            "Epoch 194/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0491 - accuracy: 0.9928 - val_loss: 0.3530 - val_accuracy: 0.9278\n",
            "Epoch 195/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0488 - accuracy: 0.9934 - val_loss: 0.2033 - val_accuracy: 0.9539\n",
            "Epoch 196/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0494 - accuracy: 0.9927 - val_loss: 0.1597 - val_accuracy: 0.9589\n",
            "Epoch 197/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0483 - accuracy: 0.9929 - val_loss: 0.1718 - val_accuracy: 0.9556\n",
            "Epoch 198/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0481 - accuracy: 0.9941 - val_loss: 0.1952 - val_accuracy: 0.9528\n",
            "Epoch 199/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0524 - accuracy: 0.9909 - val_loss: 0.1589 - val_accuracy: 0.9550\n",
            "Epoch 200/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0496 - accuracy: 0.9925 - val_loss: 0.1622 - val_accuracy: 0.9550\n",
            "Epoch 201/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0479 - accuracy: 0.9933 - val_loss: 0.1751 - val_accuracy: 0.9528\n",
            "Epoch 202/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0494 - accuracy: 0.9919 - val_loss: 0.1940 - val_accuracy: 0.9561\n",
            "Epoch 203/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0480 - accuracy: 0.9925 - val_loss: 0.2509 - val_accuracy: 0.9467\n",
            "Epoch 204/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0478 - accuracy: 0.9926 - val_loss: 0.1762 - val_accuracy: 0.9533\n",
            "Epoch 205/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0488 - accuracy: 0.9923 - val_loss: 0.1665 - val_accuracy: 0.9506\n",
            "Epoch 206/400\n",
            "160/160 [==============================] - 2s 14ms/step - loss: 0.0462 - accuracy: 0.9931 - val_loss: 0.1823 - val_accuracy: 0.9533\n",
            "Epoch 207/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0476 - accuracy: 0.9937 - val_loss: 0.1811 - val_accuracy: 0.9528\n",
            "Epoch 208/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0510 - accuracy: 0.9921 - val_loss: 0.1808 - val_accuracy: 0.9400\n",
            "Epoch 209/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0443 - accuracy: 0.9949 - val_loss: 0.1718 - val_accuracy: 0.9572\n",
            "Epoch 210/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0472 - accuracy: 0.9922 - val_loss: 0.2168 - val_accuracy: 0.9533\n",
            "Epoch 211/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0450 - accuracy: 0.9932 - val_loss: 0.2240 - val_accuracy: 0.9500\n",
            "Epoch 212/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0438 - accuracy: 0.9944 - val_loss: 0.2451 - val_accuracy: 0.9472\n",
            "Epoch 213/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0445 - accuracy: 0.9942 - val_loss: 0.2147 - val_accuracy: 0.9572\n",
            "Epoch 214/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0467 - accuracy: 0.9941 - val_loss: 0.2014 - val_accuracy: 0.9544\n",
            "Epoch 215/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0475 - accuracy: 0.9928 - val_loss: 0.2269 - val_accuracy: 0.9483\n",
            "Epoch 216/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0464 - accuracy: 0.9926 - val_loss: 0.1829 - val_accuracy: 0.9550\n",
            "Epoch 217/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0478 - accuracy: 0.9930 - val_loss: 0.1756 - val_accuracy: 0.9544\n",
            "Epoch 218/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0425 - accuracy: 0.9944 - val_loss: 0.2165 - val_accuracy: 0.9511\n",
            "Epoch 219/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0432 - accuracy: 0.9944 - val_loss: 0.1752 - val_accuracy: 0.9550\n",
            "Epoch 220/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0423 - accuracy: 0.9960 - val_loss: 0.1872 - val_accuracy: 0.9561\n",
            "Epoch 221/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0404 - accuracy: 0.9950 - val_loss: 0.2376 - val_accuracy: 0.9483\n",
            "Epoch 222/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0455 - accuracy: 0.9937 - val_loss: 0.1939 - val_accuracy: 0.9517\n",
            "Epoch 223/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0433 - accuracy: 0.9951 - val_loss: 0.2218 - val_accuracy: 0.9528\n",
            "Epoch 224/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0439 - accuracy: 0.9938 - val_loss: 0.2151 - val_accuracy: 0.9544\n",
            "Epoch 225/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0395 - accuracy: 0.9958 - val_loss: 0.1736 - val_accuracy: 0.9567\n",
            "Epoch 226/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0450 - accuracy: 0.9937 - val_loss: 0.2148 - val_accuracy: 0.9494\n",
            "Epoch 227/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0468 - accuracy: 0.9929 - val_loss: 0.1860 - val_accuracy: 0.9517\n",
            "Epoch 228/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0448 - accuracy: 0.9932 - val_loss: 0.1772 - val_accuracy: 0.9578\n",
            "Epoch 229/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0458 - accuracy: 0.9933 - val_loss: 0.2820 - val_accuracy: 0.9400\n",
            "Epoch 230/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0463 - accuracy: 0.9932 - val_loss: 0.1703 - val_accuracy: 0.9506\n",
            "Epoch 231/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0419 - accuracy: 0.9953 - val_loss: 0.1846 - val_accuracy: 0.9450\n",
            "Epoch 232/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0435 - accuracy: 0.9946 - val_loss: 0.2242 - val_accuracy: 0.9467\n",
            "Epoch 233/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0436 - accuracy: 0.9944 - val_loss: 0.2760 - val_accuracy: 0.9483\n",
            "Epoch 234/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0391 - accuracy: 0.9954 - val_loss: 0.2037 - val_accuracy: 0.9506\n",
            "Epoch 235/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0416 - accuracy: 0.9945 - val_loss: 0.3658 - val_accuracy: 0.9250\n",
            "Epoch 236/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0423 - accuracy: 0.9943 - val_loss: 0.1694 - val_accuracy: 0.9544\n",
            "Epoch 237/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0455 - accuracy: 0.9927 - val_loss: 0.1862 - val_accuracy: 0.9583\n",
            "Epoch 238/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0413 - accuracy: 0.9946 - val_loss: 0.2273 - val_accuracy: 0.9539\n",
            "Epoch 239/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0414 - accuracy: 0.9948 - val_loss: 0.1800 - val_accuracy: 0.9539\n",
            "Epoch 240/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0410 - accuracy: 0.9947 - val_loss: 0.1906 - val_accuracy: 0.9561\n",
            "Epoch 241/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0396 - accuracy: 0.9950 - val_loss: 0.2125 - val_accuracy: 0.9511\n",
            "Epoch 242/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0403 - accuracy: 0.9945 - val_loss: 0.1788 - val_accuracy: 0.9572\n",
            "Epoch 243/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0427 - accuracy: 0.9948 - val_loss: 0.2043 - val_accuracy: 0.9494\n",
            "Epoch 244/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0427 - accuracy: 0.9938 - val_loss: 0.2522 - val_accuracy: 0.9494\n",
            "Epoch 245/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0398 - accuracy: 0.9954 - val_loss: 0.1828 - val_accuracy: 0.9522\n",
            "Epoch 246/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0421 - accuracy: 0.9924 - val_loss: 0.1777 - val_accuracy: 0.9411\n",
            "Epoch 247/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0406 - accuracy: 0.9938 - val_loss: 0.1689 - val_accuracy: 0.9511\n",
            "Epoch 248/400\n",
            "160/160 [==============================] - 2s 14ms/step - loss: 0.0431 - accuracy: 0.9943 - val_loss: 0.1972 - val_accuracy: 0.9389\n",
            "Epoch 249/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0429 - accuracy: 0.9945 - val_loss: 0.2029 - val_accuracy: 0.9544\n",
            "Epoch 250/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0413 - accuracy: 0.9938 - val_loss: 0.2574 - val_accuracy: 0.9456\n",
            "Epoch 251/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0427 - accuracy: 0.9940 - val_loss: 0.2944 - val_accuracy: 0.9344\n",
            "Epoch 252/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0382 - accuracy: 0.9945 - val_loss: 0.2655 - val_accuracy: 0.9483\n",
            "Epoch 253/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0419 - accuracy: 0.9943 - val_loss: 0.1662 - val_accuracy: 0.9517\n",
            "Epoch 254/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0389 - accuracy: 0.9946 - val_loss: 0.1688 - val_accuracy: 0.9550\n",
            "Epoch 255/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0422 - accuracy: 0.9937 - val_loss: 0.1972 - val_accuracy: 0.9539\n",
            "Epoch 256/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0411 - accuracy: 0.9932 - val_loss: 0.1937 - val_accuracy: 0.9539\n",
            "Epoch 257/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0412 - accuracy: 0.9935 - val_loss: 0.1866 - val_accuracy: 0.9517\n",
            "Epoch 258/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0393 - accuracy: 0.9952 - val_loss: 0.2409 - val_accuracy: 0.9489\n",
            "Epoch 259/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0432 - accuracy: 0.9936 - val_loss: 0.2427 - val_accuracy: 0.9456\n",
            "Epoch 260/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0391 - accuracy: 0.9950 - val_loss: 0.1930 - val_accuracy: 0.9572\n",
            "Epoch 261/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0385 - accuracy: 0.9957 - val_loss: 0.1946 - val_accuracy: 0.9489\n",
            "Epoch 262/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0387 - accuracy: 0.9945 - val_loss: 0.1950 - val_accuracy: 0.9567\n",
            "Epoch 263/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0388 - accuracy: 0.9949 - val_loss: 0.2149 - val_accuracy: 0.9500\n",
            "Epoch 264/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0406 - accuracy: 0.9932 - val_loss: 0.2052 - val_accuracy: 0.9561\n",
            "Epoch 265/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0379 - accuracy: 0.9959 - val_loss: 0.2253 - val_accuracy: 0.9528\n",
            "Epoch 266/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0384 - accuracy: 0.9951 - val_loss: 0.2106 - val_accuracy: 0.9506\n",
            "Epoch 267/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0352 - accuracy: 0.9959 - val_loss: 0.2259 - val_accuracy: 0.9539\n",
            "Epoch 268/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0415 - accuracy: 0.9939 - val_loss: 0.2626 - val_accuracy: 0.9467\n",
            "Epoch 269/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0390 - accuracy: 0.9948 - val_loss: 0.1798 - val_accuracy: 0.9528\n",
            "Epoch 270/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0383 - accuracy: 0.9940 - val_loss: 0.3508 - val_accuracy: 0.9311\n",
            "Epoch 271/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0386 - accuracy: 0.9945 - val_loss: 0.1997 - val_accuracy: 0.9567\n",
            "Epoch 272/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0424 - accuracy: 0.9929 - val_loss: 0.2316 - val_accuracy: 0.9467\n",
            "Epoch 273/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0387 - accuracy: 0.9951 - val_loss: 0.2004 - val_accuracy: 0.9544\n",
            "Epoch 274/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0382 - accuracy: 0.9954 - val_loss: 0.2436 - val_accuracy: 0.9506\n",
            "Epoch 275/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0390 - accuracy: 0.9942 - val_loss: 0.2294 - val_accuracy: 0.9494\n",
            "Epoch 276/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0356 - accuracy: 0.9959 - val_loss: 0.1723 - val_accuracy: 0.9594\n",
            "Epoch 277/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0337 - accuracy: 0.9968 - val_loss: 0.1779 - val_accuracy: 0.9500\n",
            "Epoch 278/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0390 - accuracy: 0.9945 - val_loss: 0.2888 - val_accuracy: 0.9422\n",
            "Epoch 279/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0393 - accuracy: 0.9945 - val_loss: 0.1727 - val_accuracy: 0.9561\n",
            "Epoch 280/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0379 - accuracy: 0.9951 - val_loss: 0.1945 - val_accuracy: 0.9578\n",
            "Epoch 281/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0365 - accuracy: 0.9957 - val_loss: 0.1739 - val_accuracy: 0.9583\n",
            "Epoch 282/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0364 - accuracy: 0.9949 - val_loss: 0.1933 - val_accuracy: 0.9544\n",
            "Epoch 283/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0379 - accuracy: 0.9956 - val_loss: 0.2236 - val_accuracy: 0.9517\n",
            "Epoch 284/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0358 - accuracy: 0.9955 - val_loss: 0.1895 - val_accuracy: 0.9544\n",
            "Epoch 285/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0365 - accuracy: 0.9947 - val_loss: 0.1833 - val_accuracy: 0.9556\n",
            "Epoch 286/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0380 - accuracy: 0.9948 - val_loss: 0.1933 - val_accuracy: 0.9578\n",
            "Epoch 287/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0354 - accuracy: 0.9958 - val_loss: 0.1974 - val_accuracy: 0.9544\n",
            "Epoch 288/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0340 - accuracy: 0.9952 - val_loss: 0.2264 - val_accuracy: 0.9478\n",
            "Epoch 289/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0379 - accuracy: 0.9951 - val_loss: 0.2121 - val_accuracy: 0.9533\n",
            "Epoch 290/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0384 - accuracy: 0.9945 - val_loss: 0.1645 - val_accuracy: 0.9544\n",
            "Epoch 291/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0349 - accuracy: 0.9953 - val_loss: 0.2211 - val_accuracy: 0.9461\n",
            "Epoch 292/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0363 - accuracy: 0.9956 - val_loss: 0.1759 - val_accuracy: 0.9528\n",
            "Epoch 293/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0400 - accuracy: 0.9940 - val_loss: 0.2234 - val_accuracy: 0.9556\n",
            "Epoch 294/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0357 - accuracy: 0.9959 - val_loss: 0.2299 - val_accuracy: 0.9544\n",
            "Epoch 295/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0341 - accuracy: 0.9957 - val_loss: 0.1877 - val_accuracy: 0.9594\n",
            "Epoch 296/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0328 - accuracy: 0.9966 - val_loss: 0.2325 - val_accuracy: 0.9506\n",
            "Epoch 297/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0347 - accuracy: 0.9952 - val_loss: 0.1812 - val_accuracy: 0.9544\n",
            "Epoch 298/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0351 - accuracy: 0.9959 - val_loss: 0.1633 - val_accuracy: 0.9567\n",
            "Epoch 299/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0352 - accuracy: 0.9955 - val_loss: 0.1770 - val_accuracy: 0.9550\n",
            "Epoch 300/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0351 - accuracy: 0.9952 - val_loss: 0.1855 - val_accuracy: 0.9533\n",
            "Epoch 301/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0342 - accuracy: 0.9961 - val_loss: 0.1734 - val_accuracy: 0.9567\n",
            "Epoch 302/400\n",
            "160/160 [==============================] - 2s 14ms/step - loss: 0.0337 - accuracy: 0.9961 - val_loss: 0.2248 - val_accuracy: 0.9533\n",
            "Epoch 303/400\n",
            "160/160 [==============================] - 2s 16ms/step - loss: 0.0348 - accuracy: 0.9955 - val_loss: 0.2102 - val_accuracy: 0.9494\n",
            "Epoch 304/400\n",
            "160/160 [==============================] - 2s 15ms/step - loss: 0.0367 - accuracy: 0.9941 - val_loss: 0.2116 - val_accuracy: 0.9528\n",
            "Epoch 305/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0362 - accuracy: 0.9942 - val_loss: 0.2380 - val_accuracy: 0.9506\n",
            "Epoch 306/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0325 - accuracy: 0.9962 - val_loss: 0.2184 - val_accuracy: 0.9539\n",
            "Epoch 307/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0317 - accuracy: 0.9968 - val_loss: 0.2267 - val_accuracy: 0.9561\n",
            "Epoch 308/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0328 - accuracy: 0.9964 - val_loss: 0.1859 - val_accuracy: 0.9544\n",
            "Epoch 309/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0360 - accuracy: 0.9948 - val_loss: 0.2640 - val_accuracy: 0.9489\n",
            "Epoch 310/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0332 - accuracy: 0.9957 - val_loss: 0.2783 - val_accuracy: 0.9489\n",
            "Epoch 311/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0341 - accuracy: 0.9949 - val_loss: 0.1769 - val_accuracy: 0.9606\n",
            "Epoch 312/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0359 - accuracy: 0.9945 - val_loss: 0.2146 - val_accuracy: 0.9522\n",
            "Epoch 313/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0364 - accuracy: 0.9953 - val_loss: 0.1641 - val_accuracy: 0.9578\n",
            "Epoch 314/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0329 - accuracy: 0.9971 - val_loss: 0.1931 - val_accuracy: 0.9578\n",
            "Epoch 315/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0357 - accuracy: 0.9945 - val_loss: 0.2134 - val_accuracy: 0.9583\n",
            "Epoch 316/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0329 - accuracy: 0.9969 - val_loss: 0.1832 - val_accuracy: 0.9589\n",
            "Epoch 317/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0348 - accuracy: 0.9954 - val_loss: 0.2108 - val_accuracy: 0.9550\n",
            "Epoch 318/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0357 - accuracy: 0.9948 - val_loss: 0.2024 - val_accuracy: 0.9572\n",
            "Epoch 319/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0330 - accuracy: 0.9953 - val_loss: 0.2035 - val_accuracy: 0.9539\n",
            "Epoch 320/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0340 - accuracy: 0.9952 - val_loss: 0.1520 - val_accuracy: 0.9611\n",
            "Epoch 321/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0341 - accuracy: 0.9954 - val_loss: 0.3098 - val_accuracy: 0.9467\n",
            "Epoch 322/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0351 - accuracy: 0.9946 - val_loss: 0.2190 - val_accuracy: 0.9494\n",
            "Epoch 323/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0308 - accuracy: 0.9965 - val_loss: 0.1925 - val_accuracy: 0.9522\n",
            "Epoch 324/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0336 - accuracy: 0.9965 - val_loss: 0.2335 - val_accuracy: 0.9522\n",
            "Epoch 325/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0313 - accuracy: 0.9962 - val_loss: 0.1906 - val_accuracy: 0.9594\n",
            "Epoch 326/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0315 - accuracy: 0.9970 - val_loss: 0.1678 - val_accuracy: 0.9606\n",
            "Epoch 327/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0320 - accuracy: 0.9960 - val_loss: 0.1884 - val_accuracy: 0.9617\n",
            "Epoch 328/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0309 - accuracy: 0.9971 - val_loss: 0.1947 - val_accuracy: 0.9500\n",
            "Epoch 329/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0318 - accuracy: 0.9965 - val_loss: 0.2141 - val_accuracy: 0.9550\n",
            "Epoch 330/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0320 - accuracy: 0.9959 - val_loss: 0.1717 - val_accuracy: 0.9533\n",
            "Epoch 331/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0323 - accuracy: 0.9954 - val_loss: 0.2017 - val_accuracy: 0.9528\n",
            "Epoch 332/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0307 - accuracy: 0.9962 - val_loss: 0.1848 - val_accuracy: 0.9578\n",
            "Epoch 333/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0333 - accuracy: 0.9954 - val_loss: 0.1851 - val_accuracy: 0.9439\n",
            "Epoch 334/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0322 - accuracy: 0.9962 - val_loss: 0.2235 - val_accuracy: 0.9539\n",
            "Epoch 335/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0347 - accuracy: 0.9953 - val_loss: 0.1863 - val_accuracy: 0.9544\n",
            "Epoch 336/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0302 - accuracy: 0.9970 - val_loss: 0.2005 - val_accuracy: 0.9517\n",
            "Epoch 337/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0325 - accuracy: 0.9957 - val_loss: 0.2297 - val_accuracy: 0.9578\n",
            "Epoch 338/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0330 - accuracy: 0.9947 - val_loss: 0.2157 - val_accuracy: 0.9556\n",
            "Epoch 339/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0320 - accuracy: 0.9962 - val_loss: 0.2914 - val_accuracy: 0.9428\n",
            "Epoch 340/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0334 - accuracy: 0.9942 - val_loss: 0.1954 - val_accuracy: 0.9556\n",
            "Epoch 341/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0303 - accuracy: 0.9968 - val_loss: 0.1609 - val_accuracy: 0.9589\n",
            "Epoch 342/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0298 - accuracy: 0.9974 - val_loss: 0.1870 - val_accuracy: 0.9550\n",
            "Epoch 343/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0305 - accuracy: 0.9966 - val_loss: 0.2769 - val_accuracy: 0.9450\n",
            "Epoch 344/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0333 - accuracy: 0.9952 - val_loss: 0.1965 - val_accuracy: 0.9567\n",
            "Epoch 345/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0310 - accuracy: 0.9964 - val_loss: 0.2426 - val_accuracy: 0.9556\n",
            "Epoch 346/400\n",
            "160/160 [==============================] - 2s 14ms/step - loss: 0.0331 - accuracy: 0.9950 - val_loss: 0.1876 - val_accuracy: 0.9600\n",
            "Epoch 347/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0311 - accuracy: 0.9960 - val_loss: 0.1665 - val_accuracy: 0.9572\n",
            "Epoch 348/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0291 - accuracy: 0.9966 - val_loss: 0.1783 - val_accuracy: 0.9533\n",
            "Epoch 349/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0320 - accuracy: 0.9965 - val_loss: 0.2393 - val_accuracy: 0.9528\n",
            "Epoch 350/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0300 - accuracy: 0.9967 - val_loss: 0.1680 - val_accuracy: 0.9539\n",
            "Epoch 351/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0320 - accuracy: 0.9959 - val_loss: 0.1819 - val_accuracy: 0.9567\n",
            "Epoch 352/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0314 - accuracy: 0.9968 - val_loss: 0.2627 - val_accuracy: 0.9517\n",
            "Epoch 353/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0309 - accuracy: 0.9966 - val_loss: 0.1623 - val_accuracy: 0.9572\n",
            "Epoch 354/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0312 - accuracy: 0.9955 - val_loss: 0.1793 - val_accuracy: 0.9578\n",
            "Epoch 355/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0319 - accuracy: 0.9958 - val_loss: 0.2005 - val_accuracy: 0.9594\n",
            "Epoch 356/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0298 - accuracy: 0.9972 - val_loss: 0.1673 - val_accuracy: 0.9611\n",
            "Epoch 357/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0282 - accuracy: 0.9972 - val_loss: 0.2463 - val_accuracy: 0.9506\n",
            "Epoch 358/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0286 - accuracy: 0.9969 - val_loss: 0.2307 - val_accuracy: 0.9600\n",
            "Epoch 359/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0360 - accuracy: 0.9943 - val_loss: 0.2064 - val_accuracy: 0.9544\n",
            "Epoch 360/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0330 - accuracy: 0.9958 - val_loss: 0.1972 - val_accuracy: 0.9494\n",
            "Epoch 361/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0269 - accuracy: 0.9975 - val_loss: 0.2430 - val_accuracy: 0.9539\n",
            "Epoch 362/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0271 - accuracy: 0.9970 - val_loss: 0.1854 - val_accuracy: 0.9589\n",
            "Epoch 363/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0291 - accuracy: 0.9974 - val_loss: 0.1939 - val_accuracy: 0.9583\n",
            "Epoch 364/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0292 - accuracy: 0.9969 - val_loss: 0.2149 - val_accuracy: 0.9533\n",
            "Epoch 365/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0322 - accuracy: 0.9953 - val_loss: 0.1999 - val_accuracy: 0.9483\n",
            "Epoch 366/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0324 - accuracy: 0.9951 - val_loss: 0.1947 - val_accuracy: 0.9544\n",
            "Epoch 367/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0275 - accuracy: 0.9970 - val_loss: 0.2488 - val_accuracy: 0.9511\n",
            "Epoch 368/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0288 - accuracy: 0.9965 - val_loss: 0.1636 - val_accuracy: 0.9583\n",
            "Epoch 369/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0307 - accuracy: 0.9964 - val_loss: 0.2300 - val_accuracy: 0.9528\n",
            "Epoch 370/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0291 - accuracy: 0.9968 - val_loss: 0.1713 - val_accuracy: 0.9550\n",
            "Epoch 371/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0315 - accuracy: 0.9943 - val_loss: 0.1810 - val_accuracy: 0.9544\n",
            "Epoch 372/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0287 - accuracy: 0.9963 - val_loss: 0.2211 - val_accuracy: 0.9550\n",
            "Epoch 373/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0321 - accuracy: 0.9946 - val_loss: 0.2034 - val_accuracy: 0.9522\n",
            "Epoch 374/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0332 - accuracy: 0.9951 - val_loss: 0.1732 - val_accuracy: 0.9533\n",
            "Epoch 375/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0288 - accuracy: 0.9964 - val_loss: 0.2736 - val_accuracy: 0.9522\n",
            "Epoch 376/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0300 - accuracy: 0.9961 - val_loss: 0.2262 - val_accuracy: 0.9439\n",
            "Epoch 377/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0280 - accuracy: 0.9974 - val_loss: 0.1761 - val_accuracy: 0.9544\n",
            "Epoch 378/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0298 - accuracy: 0.9958 - val_loss: 0.1722 - val_accuracy: 0.9539\n",
            "Epoch 379/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0297 - accuracy: 0.9960 - val_loss: 0.1657 - val_accuracy: 0.9578\n",
            "Epoch 380/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0279 - accuracy: 0.9966 - val_loss: 0.2530 - val_accuracy: 0.9544\n",
            "Epoch 381/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0306 - accuracy: 0.9964 - val_loss: 0.2149 - val_accuracy: 0.9550\n",
            "Epoch 382/400\n",
            "160/160 [==============================] - 2s 14ms/step - loss: 0.0282 - accuracy: 0.9966 - val_loss: 0.2445 - val_accuracy: 0.9533\n",
            "Epoch 383/400\n",
            "160/160 [==============================] - 2s 14ms/step - loss: 0.0256 - accuracy: 0.9979 - val_loss: 0.1880 - val_accuracy: 0.9550\n",
            "Epoch 384/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0258 - accuracy: 0.9975 - val_loss: 0.2341 - val_accuracy: 0.9539\n",
            "Epoch 385/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0251 - accuracy: 0.9975 - val_loss: 0.1897 - val_accuracy: 0.9528\n",
            "Epoch 386/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0302 - accuracy: 0.9965 - val_loss: 0.2826 - val_accuracy: 0.9444\n",
            "Epoch 387/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0303 - accuracy: 0.9959 - val_loss: 0.1836 - val_accuracy: 0.9428\n",
            "Epoch 388/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0268 - accuracy: 0.9969 - val_loss: 0.2162 - val_accuracy: 0.9561\n",
            "Epoch 389/400\n",
            "160/160 [==============================] - 2s 13ms/step - loss: 0.0305 - accuracy: 0.9955 - val_loss: 0.2124 - val_accuracy: 0.9556\n",
            "Epoch 390/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0297 - accuracy: 0.9950 - val_loss: 0.1951 - val_accuracy: 0.9567\n",
            "Epoch 391/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0288 - accuracy: 0.9959 - val_loss: 0.2099 - val_accuracy: 0.9561\n",
            "Epoch 392/400\n",
            "160/160 [==============================] - 2s 12ms/step - loss: 0.0276 - accuracy: 0.9956 - val_loss: 0.2185 - val_accuracy: 0.9511\n",
            "Epoch 393/400\n",
            "138/160 [========================>.....] - ETA: 0s - loss: 0.0244 - accuracy: 0.9981"
          ]
        }
      ],
      "source": [
        "model.compile(optimizer=keras.optimizers.Adam(learning_rate=0.0002),\n",
        "              loss=keras.losses.CategoricalCrossentropy(),\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "history = model.fit(x_train, y_train,epochs=400, batch_size=64, validation_split=0.15)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "metadata": {
        "id": "bu-sX3o2X0on",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 545
        },
        "outputId": "f3ae9363-6916-45b7-fee5-0b2cb44fc5f8"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "%matplotlib inline\n",
        "acc = history.history['accuracy']\n",
        "val_acc = history.history['val_accuracy']\n",
        "loss = history.history['loss']\n",
        "val_loss = history.history['val_loss']\n",
        "\n",
        "epochs = range(len(acc))\n",
        "\n",
        "plt.plot(epochs, acc, 'bo', label='Training acc')\n",
        "plt.plot(epochs, val_acc, 'b', label='Validation acc')\n",
        "plt.title('Training and validation accuracy')\n",
        "plt.legend()\n",
        "\n",
        "plt.figure()\n",
        "\n",
        "plt.plot(epochs, loss, 'bo', label='Training loss')\n",
        "plt.plot(epochs, val_loss, 'b', label='Validation loss')\n",
        "plt.title('Training and validation loss')\n",
        "plt.legend()\n",
        "\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 11,
      "metadata": {
        "id": "wCQer9ERbaEe",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "673f0a8e-2249-4c53-982e-6c721566b82b"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "57/57 [==============================] - 0s 4ms/step - loss: 0.3214 - accuracy: 0.9338\n",
            "57/57 [==============================] - 0s 2ms/step\n",
            "\n",
            "[0 0 1 ... 1 0 1]\n",
            "[0 0 1 ... 1 0 1]\n"
          ]
        }
      ],
      "source": [
        "test_loss, test_acc = model.evaluate(x_test, y_test)\n",
        "test_labels_p_org = model.predict(x_test)\n",
        "test_labels_p = np.argmax(test_labels_p_org, axis=1)\n",
        "\n",
        "print()\n",
        "print(test_labels_p)\n",
        "print(y_test_original)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn import metrics\n",
        "confusion_matrix = metrics.confusion_matrix(y_test_original, test_labels_p)\n",
        "cm_display = metrics.ConfusionMatrixDisplay(confusion_matrix = confusion_matrix, display_labels = [False, True])\n",
        "\n",
        "cm_display.plot()\n",
        "plt.show()"
      ],
      "metadata": {
        "id": "7bvnOhFwek9r",
        "outputId": "00de4902-ae80-465d-f8dc-91ad7a6c61cd",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        }
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAUsAAAEGCAYAAADscbcsAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAfWklEQVR4nO3deZgdVZ3/8fcnG1nJQhZDEkgUFJGRTGSVnSAS9BEcUUEU1Di48AOVcQFlBBlUcFSEn9sEcAiIyCJgcGExAVmEQAgBQlASQkjCloUEErJ3f+ePOk1umu57q+nbXX27P6/nqaerTp1bdW46+eacOqfOUURgZmbldSu6AGZmtcDB0swsBwdLM7McHCzNzHJwsDQzy6FH0QVoC0OHdI+xY3oWXQxrgace61t0EayF1rBqRUQMa8013n9Yv1j5cl2uvA8/tvG2iDiqNfdrjU4ZLMeO6cmDt40puhjWAu/fcXzRRbAW+mvc8Gxrr7Hy5ToevG2nXHm7j5w/tLX3a41OGSzNrDYEUE990cXIxcHSzAoTBJsjXzO8aA6WZlYo1yzNzCoIgroaeeXawdLMClWPg6WZWVkB1DlYmplV5pqlmVkFAWz2M0szs/KCcDPczKyigLraiJUOlmZWnOwNntrgYGlmBRJ1qOhC5OJgaWaFyTp4HCzNzMrKxlk6WJqZVVTvmqWZWXmuWZqZ5RCIuhpZ3cbB0swKVSvN8NoI6WbWKQViU3TPtVUi6auSnpA0V9I1knpLGidppqQFkq6V1Cvl3S4dL0jnx1a6voOlmRUmG5TeLddWjqRRwOnAXhGxB9AdOB64ELgoInYBVgGT00cmA6tS+kUpX1kOlmZWqLo0ML3SlkMPoI+kHkBf4AXgcOCGdH4qcGzaPyYdk85PlFT2Jg6WZlaYCFEX3XJtwFBJs0q2U7ZeJ54DfgQsJguSrwAPA6sjYkvKthQYlfZHAUvSZ7ek/DuUK6s7eMysUPX5hw6tiIi9mjohaTBZbXEcsBq4HqjqGuMOlmZWmKyDpyph6AjgmYhYDiDpRuAAYJCkHqn2OBp4LuV/DhgDLE3N9oHAynI3cDPczApTrQ4esub3fpL6pmePE4F5wJ3AcSnPycAf0v60dEw6PyOi/CzErlmaWaHqqjDOMiJmSroBmA1sAR4BpgB/An4n6fyUdnn6yOXAVZIWAC+T9ZyX5WBpZoWp5hs8EXEOcE6j5IXAPk3k3QB8tCXXd7A0s0LVR208DXSwNLPCZBNpOFiamZUViM05XmXsCBwszawwETQMOO/wHCzNrEBqyaD0QjlYmllhAtcszcxycQePmVkFgWpm8l8HSzMrTLYUbm2EodoopZl1Urnnqiycg6WZFSbwGzxmZrm4ZmlmVkGEXLM0M6sk6+Dx645mZhWoZgal10YpzaxTyjp4lGsrR9I7JM0p2V6V9BVJQyTdIWl++jk45ZekS9K64Y9JmlCprA6WZlaoOrrl2sqJiH9GxPiIGA+8B1gH3AScCUyPiF2B6ekYYBKwa9pOAX5ZqZwOlmZWmIY3eFpbs2xkIvB0RDzLtuuDN143/MrIPEC2sNnIchf1M0szK1SOxcgaDJU0q+R4SkRMaSLf8cA1aX9ERLyQ9l8ERqT919cNTxrWFH+BZjhYmllhImBzfe5g2ey64Q0k9QI+BJz1xntFSCq7gmM5DpZmVpisGV7Vp4GTgNkR8VI6fknSyIh4ITWzl6X0hnXDG5SuKd4kP7M0s0LVpffDK205ncDWJjhsuz5443XDT0q94vsBr5Q015vkmmUHc+OUYfzlt0OQYNxuG/iPixYzb1Y/Lj1vRzZvFru+ez1n/Hgx3XvAjBsHc93PhxMBffrVc9oFS3jbuzYU/RUMGLbjJr5+8WIGDdsCAX/+zQ7cfPmwoovV4TQMHaoGSf2A9wGfL0m+ALhO0mTgWeBjKf3PwNHAArKe889Uun6bBUtJdcDjJUnHRsSiZvKujYj+bVWWWrHihZ7cfPlQLr3rH2zXJzj/8ztz502DufJHb+HC655m9Ns2MvWHb+GO64Zw1CdeZsSYjfz37xcwYFAdD80YwMXfGMMlf5pf9NcwoG6LmHLejix4vC99+tXxs1ufYvbdA1g8v3fRRetgqtcMj4jXgB0apa0k6x1vnDeAU1ty/bZshq9vGPeUtkVteK9Oo26L2LihG3VbYOP6bmzXt56evYLRb9sIwIRD1nDvnwcB8K691zFgUB0Au01Yx4oXehZWbtvWy8t6suDxvgCsf607Sxb0ZujIzQWXqmOqT+vwVNqK1m7PLCX1lzRd0mxJj0s6pok8IyXdnUbgz5V0UEo/UtL96bPXS+qUtdChIzdz3BeX8am9d+eE8XvQb0Adh3xoNXVbxFOP9gHg3j8OYvnzbwyKt14zhL0PW9PeRbYcRozexNv2WM8/ZvctuigdTtYb3j3XVrS2DJZ9Sl49ugnYAHw4IiYAhwE/ltT4v4tPALelUfh7AnMkDQXOBo5In50FnNH4ZpJOkTRL0qzlK+va8Gu1nTWru3P/bQOZOnMev31kLhvWdWfGjYM565eL+NU5ozjt6F3p07+Obo1+a3Pu689t1+zA5G8/X0zBrVm9+9bxn5ct4lff2ZF1a4v/B9/RtNGg9DbRlh0861PQA0BST+D7kg4G6skGgI4gGyja4CHg1ynvzRExR9IhwO7AfSm29gLub3yzNDh1CsBee/Z+02OpivTIPf15y5hNDNohC/YHHL2aebP6MfEjq/jJzQsAePiuASxduN3rn1k4rzc//doYzv/NQrYfUpv/SXRW3XsE/3nZImbcOJj7/jKo6OJ0WB2hiZ1He/aGnwgMA94TEZslLQK2edodEXenYPoB4ApJPwFWAXdExAntWNZCDB+1mSdn92XDOrFdn2DOvQN4+7vXsXpFDwYN3cKmjeK6XwznhNOzIWTLlvbkvM+N4+uXPPv6M03rKIIzfryEJfN7c+MU94I3p5q94W2tPYPlQGBZCpSHATs3ziBpZ2BpRFwqaTtgAvA94OeSdomIBWl4wKiIeKody94udpuwjoM+8Aqnvv8ddO8R7LLHeiZ9ciVTLxzJzL9uT9TDB05eyfgD1wJw9UVvYc2q7vzsrGxsbfcewc9u7XR/LDXpXfu8xhEfXcXCeb35xR3/BOB/fzCSh2ZsX3DJOp5amfxXWQ96G1y40XCg9OzxFqA/2XPH/YBJEbGoIa+kk4GvA5uBtcBJEfGMpMOBC4GG9ufZETGtuXvvtWfvePC2Mc2dtg7o/TuOr5zJOpS/xg0PV3r9sJLBuw2Pw399XK68Nx7wy1bfrzXarGbZeNxkRKwA9i+XNyKmsnWGkNLzM4C926CYZlYwN8PNzCrwM0szs5wcLM3MKmgYZ1kLHCzNrFAeZ2lmVkEEbMk/+W+hHCzNrFBuhpuZVeBnlmZmOUWNBMvaeFhgZp1WteazlDRI0g2S/iHpSUn7Sxoi6Q5J89PPwSmvJF0iaYGkxyRNqHR9B0szK0wE1Zyi7WLg1ojYjWyKxyeBM4HpEbErMD0dQ7aw2a5pOwX4ZaWLO1iaWYFEXX23XFvZq0gDgYOBywEiYlNErAaOYesr1FOBY9P+McCVkXkAGJRWf2yWg6WZFSpCuTZgaMME32k7peQy44DlwP9KekTSZWmGshElqza+SDaHLmTz6S4p+fzSlNYsd/CYWWFa+G74ijKzDvUgm9LxtIiYKelitja5s3tFhKQ3Pc2aa5ZmVpzInlvm2SpYSjYX7sx0fANZ8HypoXmdfi5L558DSudxHJ3SmuVgaWaFqkZveES8CCyR9I6UNBGYB0wDTk5pJwN/SPvTgJNSr/h+wCslzfUmuRluZoWJ1MFTJacBV0vqBSwEPkNWIbxO0mTgWeBjKe+fgaOBBcC6lLcsB0szK1S1FmuIiDlAU880JzaRN4BTW3J9B0szK1StvMHjYGlmhck6bxwszcwq8kQaZmY5tNECs1XnYGlmhQlEvSf/NTOrrEYqlg6WZlYgd/CYmeVUI1VLB0szK1TN1ywl/X/KxPyIOL1NSmRmXUYA9fU1HiyBWe1WCjPrmgKo9ZplREwtPZbUNyLWtX2RzKwrqZVxlhUHOKVFf+YB/0jHe0r6RZuXzMy6hsi5FSzPaNCfAu8HVgJExKNka12YmbVSviUlOkInUK7e8IhYIm1T2Lq2KY6ZdTkdoNaYR56a5RJJ7wVCUk9JXyNbYtLMrHUCol65tkokLZL0uKQ5kmaltHZdN/wLZJNkjgKeB8bTwkkzzcyap5xbLodFxPiShc2qtm54xWZ4RKwATsxbUjOzFmnbZvgxwKFpfypwF/BNStYNBx6QNEjSyHLr8OTpDX+rpFskLZe0TNIfJL211V/BzAxa0htebt3whivdLunhknPtum74b4GfAx9Ox8cD1wD75vismVnzWjYovdy64QAHRsRzkoYDd0j6xza3aod1w/tGxFURsSVtvwF6v9kbmpmVqtK64UTEc+nnMuAmYB/aY93w1Is0BPiLpDMljZW0s6RvkC0jaWbWevXKt5UhqZ+kAQ37wJHAXNpp3fCHySrJDaX8fMm5AM4qW3ozsxzefMN4GyOAm9J48B7AbyPiVkkP0dbrhkfEuNaV3cysgiq9yhgRC4E9m0hfSXuuGy5pD2B3Sp5VRsSVLbmRmdkbqfZnHWog6RyycUq7k1VdJwH3Ag6WZtZ6neh1x+PIqrEvRsRnyKq6A9u0VGbWddTn3AqWpxm+PiLqJW2RtD1Z1/uYSh8yM6uoM0z+W2KWpEHApWQ95GuB+9u0VGbWZVSpN7zN5Xk3/Etp91eSbgW2j4jH2rZYZtZl1HqwLDdlkaQJETG7bYpkZtbxlKtZ/rjMuQAOr3JZqmb+/CFMOvL4oothLXDq/GlFF8Fa6K+7VOc6Nd8Mj4jD2rMgZtYFBRVfZewocg1KNzNrM7VeszQzaw813ww3M2sXNRIs88yULkmflPSddLyTpH3avmhm1iV0onXDfwHsD5yQjteQzZxuZtYqivxb0fI0w/eNiAmSHgGIiFWSerVxucysq6iR3vA8NcvNkrqTKsKShtEhXms3s86gmjVLSd0lPSLpj+l4nKSZaX3waxsqepK2S8cL0vmxla6dJ1heQraexXBJ3yObnu37+YpuZlZBdZ9Zfhl4suT4QuCiiNgFWAVMTumTgVUp/aKUr6yKwTIirga+AfwAeAE4NiKuz110M7PmVPGZpaTRwAeAy9KxyN40vCFlmQocm/aPScek8xNT/mblmfx3J7I1Km4pTYuIxZWLb2ZWQfU6b35KVrEbkI53AFZHxJZ0XLo2+OvrhkfEFkmvpPwrmrt4ng6eP7F14bLewDjgn8C7WvQ1zMyaoPw9IEMlzSo5nhIRUwAkfRBYFhEPSzq0uiXM5Jmi7V9Kj9NsRF9qJruZWVtZERF7NXPuAOBDko4mq9RtD1wMDJLUI9UuS9cGb1g3fKmkHmSrP6wsd/M8HTzbSFOz7dvSz5mZNakKHTwRcVZEjI6IscDxwIyIOBG4k2xpHHjjuuEN64kfl/KXvUueZ5ZnlBx2AyYAz1f6nJlZRW0/4PybwO8knQ88Alye0i8HrpK0AHiZLMCWleeZ5YCS/S1kzzB/36Limpk1p8rBMiLuAu5K+wuBN7yeHREbgI+25Lplg2UajD4gIr7WkouameXWAV5lzKPcshI9Upf6Ae1ZIDPrOkSLesMLVa5m+SDZ88k5kqYB1wOvNZyMiBvbuGxm1tl1kEky8sjzzLI3WZf64WwdbxmAg6WZtV4nCJbDU0/4XLYGyQY18vXMrMOrkWhSLlh2B/qzbZBsUCNfz8w6us7QDH8hIs5rt5KYWdfUCYJlbczIaWa1KzpHb/jEdiuFmXVdtV6zjIiX27MgZtY1dYZnlmZmbc/B0sysgg6yzG0eDpZmVhjhZriZWS4OlmZmeThYmpnlUCPBssXLSpiZVU2VlsKV1FvSg5IelfSEpO+m9HGSZkpaIOlaSb1S+nbpeEE6P7ZSUR0szaxYVViDB9gIHB4RewLjgaMk7QdcCFwUEbsAq4DJKf9kYFVKvyjlK8vB0swKpfp8WzmRWZsOe6YtyKaWvCGlTwWOTfvHpGPS+YmSyr7i7WBpZoVqQTN8qKRZJdsp21xH6i5pDrAMuAN4GlidlsEFWAqMSvujgCUA6fwrwA7lyukOHjMrTssGpZdbN5yIqAPGSxoE3ATs1urylXDN0syKVZ1nllsvF7GabL3w/YFBkhoqhaOB59L+c8AYyNYbAwaSrQjRLAdLMytMwxs8VegNH5ZqlEjqA7wPeJIsaB6Xsp0M/CHtT0vHpPMzIqLsXdwMN7NCqb4qAy1HAlPT8t3dgOsi4o+S5gG/k3Q+8Ahwecp/OXCVpAXAy8DxlW7gYGlmxanSRBoR8Rjwr02kLwT2aSJ9A/DRltzDwdLMCuV3w83M8nCwNDOrzDVLM7M8HCzNzCroJKs7mpm1Kc+UbmaWV/mx4B2Gg6WZFco1S2uxr57xIPvs9zyrV2/HF0+ZBMCBBy3hk5+ay5idXuUrp72P+fOHAHDY4Yv4yEf/+fpnx41bzWlfOpKFCwcXUvauatXCntz25ZGvH7+6pCf7fnklO+67nr99ZzhbNopuPeCQc5cxYs8NzL50ME9NGwBA1IlVT/fiszOfpvegGnlwV21e3XFbknYApqfDtwB1wPJ0vE9EbGqPcnR0d9wxlmnTduFr35j5etqziwbyX+cdwOlfnrVN3jtnjOXOGWMBGDt2Nd85914HygIMfutmjr9lMQD1dXDFgW9l3JFruevbI9j7tJXsfMg6Ft3Vj7//cCgfvnopE/59FRP+fRUAz0zvx6NXDO66gTJxB0+JiFhJNnsxks4F1kbEjxrOS+pRMudclzX38eEMH/HaNmlLlmxf8XOHHLaYv921U1sVy3Ja+ve+DNxpM9uP2gKCTWuzeWo2relGv+Fv/Os9/48D2PWDa9q7mB2Og2UFkq4ANpC9z3mfpFcpCaKS5gIfjIhFkj4JnA70AmYCX0pz1xlwyCGL+e65BxZdjC5v/p+2Br8Dv72cWz47ir9fMIwI8W/XLt4m7+b1YvE9/Tj4nGVFFLXjCGqmg6foKdpGA++NiDOayyDpncDHgQMiYjxZE/7EJvKd0jCD8qYtrzU+3Wm9Y7eVbNjYg2cXDSq6KF1a3SZYNKM/u0zKguXc3w7kwG8t5+R7nuGAby3jzm+N2Cb/ohn9GDlhfZdvgkN1pmhrD0UHy+tz1BAnAu8BHkpTxk8E3to4U0RMiYi9ImKvXj36tUFRO6ZDDl3M3+50E7xoz97dj2G7b6Dv0Oyv8z9v2p63vj9bEmaXSWt56dHe2+Sf/6ft3QRvUOXJf9tK0cGytAq4hW3L0/C3S8DUiBiftndExLntVcCOTAoOOniJn1d2AI2fP/YbvoXnH+wDwNL7+zBo7ObXz21c043nH+zDuCPWvuE6XU21Jv9tDx1p6NAi4IMAkiYA41L6dOAPki6KiGWShgADIuLZYorZdr551v28+93L2H7gRq66ehpXXbUHa9f04otfms3AgRv57vl3s/DpwZz9rUMA2ONflrNieR9efLF/wSXv2javE0vu68eh/7X1+eOh33uJe88fTn2d6N6rnkPPf+n1cwtv78+YA1+jZ98OEAGKFlGVyX8ljQGuBEaQ1UOnRMTFKV5cC4wlizEfi4hVaSXHi4GjgXXApyNidrl7dKRg+XvgJElPkHXiPAUQEfMknQ3cLqkbsBk4Feh0wfLCH+zfZPrf7xvdZPrjjw3nq19+X1sWyXLo2Tf43ENPb5O2414b+NjNi5vM/86PvMo7P/JqexStNlTn/4wtwH9ExGxJA4CHJd0BfBqYHhEXSDoTOBP4JjAJ2DVt+wK/TD+b1e7BsrkmdESsB45s5ty1ZP87mFknU40mdkS8ALyQ9tdIepJsudtjgENTtqnAXWTB8hjgyrTuzgOSBkkama7TpI5UszSzriaA/M3woZJK386YEhFTGmeSNJZsSOJMYERJAHyRrJkOJeuGJw1rijtYmlkHVaV1wwEk9Sd7pPeViHg1ezSZbhMR0puvxxbdG25mXVy1esMl9SQLlFdHxI0p+SVJI9P5kUBDL9zr64YnpWuKN8nB0swKpfrItZW9RlaFvBx4MiJ+UnKqdH3wxuuGn6TMfsAr5Z5XgpvhZlak6g04PwD4FPB4enkF4FvABcB1kiaTjaD5WDr3Z7JhQwvIhg59ptINHCzNrDDZoPTWR8uIuDddrikTm8gfZEMQc3OwNLNi1cjr8Q6WZlaoatQs24ODpZkVp4NMkpGHg6WZFag674a3BwdLMyuWm+FmZhWEl5UwM8vHNUszsxxqI1Y6WJpZsVRfG+1wB0szK07gQelmZpWI8KB0M7NcHCzNzHJwsDQzq8DPLM3M8nFvuJlZRVEzzXAvK2FmxQmyYJlnq0DSryUtkzS3JG2IpDskzU8/B6d0SbpE0gJJj0maUOn6DpZmVqz6nFtlVwBHNUo7E5geEbsC09MxwCRg17SdAvyy0sUdLM2sUIrItVUSEXcDLzdKPgaYmvanAseWpF8ZmQeAQQ2rQDbHzyzNrFj5n1kOlTSr5HhKREyp8JkRJas2vgiMSPujgCUl+ZamtGZXeHSwNLPiREBd7t7wFRGx15u/VYSUZwXyprkZbmbFqlIHTzNeamhep5/LUvpzwJiSfKNTWrMcLM2sWG0bLKcBJ6f9k4E/lKSflHrF9wNeKWmuN8nNcDMrTgBVWoNH0jXAoWTPNpcC5wAXANdJmgw8C3wsZf8zcDSwAFgHfKbS9R0szaxAAVGdN3gi4oRmTk1sIm8Ap7bk+g6WZlacoCUdPIVysDSzYtXI644OlmZWLAdLM7NKamciDQdLMytOAJ6izcwsB9cszcwqadHrjoVysDSz4gRElcZZtjUHSzMrVpXe4GlrDpZmViw/szQzqyDCveFmZrm4ZmlmVkkQdXVFFyIXB0szK04Vp2hraw6WZlYsDx0yMysvgHDN0sysgqje5L9tzcHSzApVKx08ihrptm8JScvJ1tvojIYCK4ouhLVIZ/2d7RwRw1pzAUm3kv355LEiIo5qzf1ao1MGy85M0qzWrJ1s7c+/s87BS+GameXgYGlmloODZe2ZUnQBrMX8O+sE/MzSzCwH1yzNzHJwsDQzy8GD0gsmqQ54vCTp2IhY1EzetRHRv10KZmVJ2gGYng7fAtQBy9PxPhGxqZCCWZvxM8uCtSQAOlh2TJLOBdZGxI9K0npExJbiSmXV5mZ4ByOpv6TpkmZLelzSMU3kGSnpbklzJM2VdFBKP1LS/emz10tyYG1Hkq6Q9CtJM4EfSjpX0tdKzs+VNDbtf1LSg+l3+D+SuhdUbMvJwbJ4fdI/mDmSbgI2AB+OiAnAYcCPJanRZz4B3BYR44E9gTmShgJnA0ekz84Czmi/r2HJaOC9EdHsn72kdwIfBw5Iv8M64MR2Kp+9SX5mWbz16R8MAJJ6At+XdDBQD4wCRgAvlnzmIeDXKe/NETFH0iHA7sB9Kbb2Au5vp+9gW10fEZVmhpgIvAd4KP2u+gDL2rpg1joOlh3PicAw4D0RsVnSIqB3aYaIuDsF0w8AV0j6CbAKuCMiTmjvAts2XivZ38K2rbeG36OAqRFxVruVylrNzfCOZyCwLAXKw4CdG2eQtDPwUkRcClwGTAAeAA6QtEvK00/S29ux3PZGi8h+N0iaAIxL6dOB4yQNT+eGpN+pdWCuWXY8VwO3SHqc7LnjP5rIcyjwdUmbgbXASRGxXNKngWskbZfynQ081fZFtmb8HjhJ0hPATNLvIiLmSTobuF1SN2AzcCqdd1rBTsFDh8zMcnAz3MwsBwdLM7McHCzNzHJwsDQzy8HB0swsBwfLLkpSXcm75ddL6tuKa10h6bi0f5mk3cvkPVTSe9/EPRalVzpzpTfKs7aF99rmnW4zcLDsytZHxPiI2APYBHyh9KSkNzUGNyI+FxHzymQ5FGhxsDQrmoOlAdwD7JJqffdImgbMk9Rd0n9LekjSY5I+D6DMzyT9U9JfgeENF5J0l6S90v5RaQakR9NMSmPJgvJXU632IEnDJP0+3eMhSQekz+4g6XZJT0i6jOwVwbIk3Szp4fSZUxqduyilT5c0LKW9TdKt6TP3SNqtGn+Y1jn5DZ4uLtUgJwG3pqQJwB4R8UwKOK9ExN7praD7JN0O/CvwDrKJO0YA84BfN7ruMOBS4OB0rSER8bKkX1Ey96Ok3wIXRcS9knYCbgPeCZwD3BsR50n6ADA5x9f5bLpHH7JJKn4fESuBfsCsiPiqpO+ka/8/soXEvhAR8yXtC/wCOPxN/DFaF+Bg2XX1kTQn7d8DXE7WPH4wIp5J6UcC7254Hkn23vquwMHANWl2neclzWji+vsBdzdcKyJebqYcRwC7l8xCt32ah/Ng4N/SZ/8kaVWO73S6pA+n/TGprCvJZm+6NqX/Brgx3eO9wPUl994Os2Y4WHZd20wNB5CCRumsOQJOi4jbGuU7uorl6AbsFxEbmihLbpIOJQu8+0fEOkl30Wi2phKR7ru68Z+BWXP8zNLKuQ34Ypo3E0lvl9QPuBv4eHqmOZJskuLGHgAOljQufXZISl8DDCjJdztwWsOBpIbgdTfZJMdImgQMrlDWgcCqFCh3I6vZNugGNNSOP0HWvH8VeEbSR9M9JGnPCvewLszB0sq5jOx55GxJc4H/IWuN3ATMT+eupIlJhiNiOXAKWZP3UbY2g28BPtzQwQOcDuyVOpDmsbVX/rtkwfYJsub44gplvRXoIelJ4AKyYN3gNWCf9B0OB85L6ScCk1P5ngDesISHWQPPOmRmloNrlmZmOThYmpnl4GBpZpaDg6WZWQ4OlmZmOThYmpnl4GBpZpbD/wHCkctkrUy86AAAAABJRU5ErkJggg==\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "_L5b6X7qc4Xa"
      },
      "outputs": [],
      "source": [
        "model.save_weights('/content/drive/MyDrive/uni project/weight')\n",
        "model.save('/content/drive/MyDrive/uni project/model.h5')"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#(True Positive + True Negative) / Total Predictions\n",
        "Accuracy = metrics.accuracy_score(y_test_original, test_labels_p)\n",
        "\n",
        "#True Positive / (True Positive + False Positive)\n",
        "Precision = metrics.precision_score(y_test_original, test_labels_p)\n",
        "\n",
        "#True Positive / (True Positive + False Negative)\n",
        "Sensitivity_recall = metrics.recall_score(y_test_original, test_labels_p)\n",
        "\n",
        "#True Negative / (True Negative + False Positive)\n",
        "Specificity = metrics.recall_score(y_test_original, test_labels_p, pos_label=0)\n",
        "\n",
        "# 2 * ((Precision * Sensitivity) / (Precision + Sensitivity))\n",
        "F1_score = metrics.f1_score(y_test_original, test_labels_p)\n",
        "\n",
        "\n",
        "print({\"Accuracy\":Accuracy})\n",
        "print({\"Precision\":Precision})\n",
        "print({\"Sensitivity_recall\":Sensitivity_recall})\n",
        "print({\"Specificity\":Specificity})\n",
        "print({\"F1_score\":F1_score})"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "kXe_VYMt8D7D",
        "outputId": "60dbaaa0-e006-462f-de70-15efa19c444d"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "{'Accuracy': 0.9521690767519466}\n",
            "{'Precision': 0.9733796296296297}\n",
            "{'Sensitivity_recall': 0.9303097345132744}\n",
            "{'Specificity': 0.9742729306487695}\n",
            "{'F1_score': 0.9513574660633485}\n"
          ]
        }
      ]
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "provenance": []
    },
    "gpuClass": "standard",
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}