Homework8_Dropout.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Homework8.Deep.2.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Tbg4G5wYlxEK"
      },
      "source": [
        "**This code is adding dropout layer**"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cv_30uZxqQey"
      },
      "source": [
        "## Loading the data and other imports"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "loClM54MqqWl"
      },
      "source": [
        "import pandas as pd\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "from sklearn.datasets import load_digits\n",
        "import keras\n",
        "from keras import layers\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dense, Dropout, Activation\n",
        "from sklearn.model_selection import RandomizedSearchCV \n",
        "from sklearn.model_selection import train_test_split\n",
        "import warnings \n",
        "warnings.filterwarnings(\"ignore\")\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "X3M4Hz87PLbB"
      },
      "source": [
        "digits = load_digits(return_X_y =True)\n",
        "X=digits[0]\n",
        "y=digits[1]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BuhOCCeeMvcx"
      },
      "source": [
        "## Train/Test Split"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qxpw7u6XwjtO"
      },
      "source": [
        "X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "H8qMvVBJ59JO"
      },
      "source": [
        "### Explore the data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7HXx5cknVQOl"
      },
      "source": [
        "X_train=X_train.reshape((X_train.shape[0],8,8,1))\n",
        "X_test=X_test.reshape((X_test.shape[0],8,8,1))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-zPzTPiv5AWK"
      },
      "source": [
        "X_test = np.array(X_test)\n",
        "y_test = np.array(y_test)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PgLqU-fN6sPS",
        "outputId": "e16e7dd6-ca43-4e6c-cae1-739b3d41797a",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "print(X_train.shape)\n",
        "print(y_train.shape)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(1437, 8, 8, 1)\n",
            "(1437,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5V22HClS5Aot",
        "outputId": "7415f4a7-0d22-4230-bec8-a63e19fbb23c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "print(X_test.shape)\n",
        "print(y_test.shape)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(360, 8, 8, 1)\n",
            "(360,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NczKVWqv_uYp"
      },
      "source": [
        "### Preprocess the data:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QdvdQTLT6iqq"
      },
      "source": [
        "X_train = X_train / 255.0\n",
        "X_test = X_test / 255.0"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Cnm-_UqV_2dj"
      },
      "source": [
        "### Creating the Model:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vbEkQsNkaFHs"
      },
      "source": [
        "def bl_model_drop(nf1,nf2,ks1,ks2,st1,st2,n_units,lrt,drop_check):\n",
        "  model=keras.models.Sequential()\n",
        "  model.add(keras.layers.Conv2D(filters=nf1, kernel_size=ks1, strides=st1, activation='elu', input_shape=(8,8,1), padding='same'))\n",
        "  model.add(keras.layers.Conv2D(filters=nf2, kernel_size=ks2, strides=st2, activation='relu'))\n",
        "  model.add(keras.layers.MaxPool2D(pool_size=(2,2), strides=2, padding='same' ))\n",
        "  model.add(keras.layers.Flatten())\n",
        "  \n",
        "  if drop_check== 1:\n",
        "    model.add(keras.layers.Dropout(0.5))\n",
        "    \n",
        "  model.add(keras.layers.Dense(n_units, activation='relu'))\n",
        "  model.add(keras.layers.Dense(10, activation='softmax'))\n",
        "  model.compile(optimizer=keras.optimizers.Adam(lr=lrt), loss='sparse_categorical_crossentropy', metrics=['accuracy'])\n",
        "  return(model)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "J7bholWoaPlb",
        "outputId": "2afb8667-c02b-4033-dbed-e89b5d124707",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 391
        }
      },
      "source": [
        "#Test the model with out randomized search\n",
        "#Model = bl_model_drop(nf1 = 50, nf2 = 150 ,ks1 = 3, ks2 = 3 , n_units = 100, lrt = 0.01,st1= 1, st2=1, drop_check=1)\n",
        "#Model.summary() # Get a summary of your model"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv2d (Conv2D)              (None, 8, 8, 50)          500       \n",
            "_________________________________________________________________\n",
            "conv2d_1 (Conv2D)            (None, 6, 6, 150)         67650     \n",
            "_________________________________________________________________\n",
            "max_pooling2d (MaxPooling2D) (None, 3, 3, 150)         0         \n",
            "_________________________________________________________________\n",
            "flatten (Flatten)            (None, 1350)              0         \n",
            "_________________________________________________________________\n",
            "dropout (Dropout)            (None, 1350)              0         \n",
            "_________________________________________________________________\n",
            "dense (Dense)                (None, 100)               135100    \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 10)                1010      \n",
            "=================================================================\n",
            "Total params: 204,260\n",
            "Trainable params: 204,260\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "91_rYB9LjnTB"
      },
      "source": [
        "#simple hyperparameters  to test\n",
        "#params={'nf1':[32],'nf2':[256],'ks1':[3],'ks2':[3],'n_units':[128],'lrt':[0.001],'st1':[1,2],\n",
        "  #  'st2':[1,2],'epochs':[5],'batch_size':[150],'drop_check':['True']}  #True/False"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zcvvCXa2abps"
      },
      "source": [
        "# Create hyperparameter space\n",
        "params={'nf1':[16, 32, 64, 128, 256, 400],\n",
        "        'nf2':[16, 32, 64, 128, 256, 400],\n",
        "        'ks1':[3,5],\n",
        "        'ks2':[3,5],\n",
        "        'n_units':[50,100,200],\n",
        "        'lrt':[0.001,0.05,0.01],\n",
        "        'st1':[1,2],\n",
        "        'st2':[1,2],\n",
        "        'epochs':[5,10,15,20],\n",
        "        'batch_size':[20, 50, 150],\n",
        "        'drop_check':[1]  # 1 means that we have dropout layer\n",
        "        }"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ozj1BV3daiWs"
      },
      "source": [
        "# Wrap Keras model so it can be used by scikit-learn\n",
        "from keras.wrappers.scikit_learn import KerasClassifier\n",
        "first_drop_model=KerasClassifier(bl_model_drop)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bF-JZIMlalVb",
        "outputId": "a9f1bc57-e811-420f-ebe6-7632fe5eacec",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "# Create grid search\n",
        "from sklearn.model_selection import RandomizedSearchCV\n",
        "cnn_drop_search=RandomizedSearchCV(first_drop_model,param_distributions=params, n_iter=100, scoring='accuracy', cv=5)\n",
        "# Fit grid search\n",
        "cnn_drop_search.fit(X_train,y_train,validation_split=0.1)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\u001b[1;30;43mStreaming output truncated to the last 5000 lines.\u001b[0m\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.3527 - accuracy: 0.5309 - val_loss: 0.9192 - val_accuracy: 0.8174\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.3357 - accuracy: 0.5155 - val_loss: 0.8509 - val_accuracy: 0.8174\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 0s 13ms/step - loss: 2.2987 - accuracy: 0.1053 - val_loss: 2.2867 - val_accuracy: 0.1043\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 2.2741 - accuracy: 0.1478 - val_loss: 2.2395 - val_accuracy: 0.1826\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 2.2181 - accuracy: 0.2068 - val_loss: 2.1357 - val_accuracy: 0.3043\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 2.0948 - accuracy: 0.2464 - val_loss: 1.9409 - val_accuracy: 0.3217\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.9262 - accuracy: 0.2754 - val_loss: 1.7084 - val_accuracy: 0.4783\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.7881 - accuracy: 0.3169 - val_loss: 1.5149 - val_accuracy: 0.6000\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.6634 - accuracy: 0.3710 - val_loss: 1.3985 - val_accuracy: 0.6522\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.5721 - accuracy: 0.4029 - val_loss: 1.2536 - val_accuracy: 0.7391\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.4826 - accuracy: 0.4193 - val_loss: 1.1338 - val_accuracy: 0.7478\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.5217 - accuracy: 0.3990 - val_loss: 1.1258 - val_accuracy: 0.7043\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.4439 - accuracy: 0.4531 - val_loss: 1.0560 - val_accuracy: 0.7565\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.4218 - accuracy: 0.4319 - val_loss: 1.0105 - val_accuracy: 0.7826\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 1.4207 - accuracy: 0.4396 - val_loss: 0.9881 - val_accuracy: 0.7826\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.3657 - accuracy: 0.4415 - val_loss: 0.9292 - val_accuracy: 0.7739\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.3505 - accuracy: 0.4628 - val_loss: 0.9238 - val_accuracy: 0.8000\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 0s 13ms/step - loss: 2.2995 - accuracy: 0.1092 - val_loss: 2.2918 - val_accuracy: 0.1391\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 2.2827 - accuracy: 0.1507 - val_loss: 2.2624 - val_accuracy: 0.2000\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 2.2395 - accuracy: 0.1942 - val_loss: 2.1822 - val_accuracy: 0.4435\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 2.1370 - accuracy: 0.2754 - val_loss: 2.0309 - val_accuracy: 0.4348\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.9904 - accuracy: 0.2928 - val_loss: 1.7941 - val_accuracy: 0.5391\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.8507 - accuracy: 0.2976 - val_loss: 1.5948 - val_accuracy: 0.6261\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 1.6870 - accuracy: 0.3643 - val_loss: 1.3823 - val_accuracy: 0.6174\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.5580 - accuracy: 0.3942 - val_loss: 1.2500 - val_accuracy: 0.6783\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 1.4895 - accuracy: 0.4357 - val_loss: 1.1534 - val_accuracy: 0.6957\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.4579 - accuracy: 0.4271 - val_loss: 1.0688 - val_accuracy: 0.7217\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.4119 - accuracy: 0.4676 - val_loss: 1.0184 - val_accuracy: 0.7739\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 0s 9ms/step - loss: 1.3166 - accuracy: 0.5063 - val_loss: 0.9568 - val_accuracy: 0.7478\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 1.3447 - accuracy: 0.4812 - val_loss: 0.8781 - val_accuracy: 0.7739\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.2710 - accuracy: 0.5324 - val_loss: 0.8367 - val_accuracy: 0.7826\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 1.2802 - accuracy: 0.5246 - val_loss: 0.7882 - val_accuracy: 0.7913\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 0s 14ms/step - loss: 2.2969 - accuracy: 0.1285 - val_loss: 2.2804 - val_accuracy: 0.2000\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 2.2714 - accuracy: 0.1816 - val_loss: 2.2166 - val_accuracy: 0.3043\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 2.2022 - accuracy: 0.1884 - val_loss: 2.0590 - val_accuracy: 0.3739\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 0s 9ms/step - loss: 2.0677 - accuracy: 0.2174 - val_loss: 1.8172 - val_accuracy: 0.4000\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 1.9048 - accuracy: 0.2773 - val_loss: 1.5916 - val_accuracy: 0.4435\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 1.7374 - accuracy: 0.3652 - val_loss: 1.4123 - val_accuracy: 0.5826\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.6131 - accuracy: 0.4406 - val_loss: 1.2663 - val_accuracy: 0.6783\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.5095 - accuracy: 0.4386 - val_loss: 1.1861 - val_accuracy: 0.7565\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 1.3767 - accuracy: 0.5111 - val_loss: 1.0769 - val_accuracy: 0.7217\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.3551 - accuracy: 0.5014 - val_loss: 0.9932 - val_accuracy: 0.8348\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 1.2729 - accuracy: 0.5420 - val_loss: 0.9521 - val_accuracy: 0.8261\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.2328 - accuracy: 0.5343 - val_loss: 0.8680 - val_accuracy: 0.8087\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.1583 - accuracy: 0.5952 - val_loss: 0.7674 - val_accuracy: 0.8696\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 1.1713 - accuracy: 0.5614 - val_loss: 0.7724 - val_accuracy: 0.8522\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.1373 - accuracy: 0.5758 - val_loss: 0.7133 - val_accuracy: 0.8696\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 2.1091 - accuracy: 0.2079 - val_loss: 1.2808 - val_accuracy: 0.6174\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 1s 24ms/step - loss: 0.9819 - accuracy: 0.6576 - val_loss: 0.5045 - val_accuracy: 0.8522\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 0s 23ms/step - loss: 0.6280 - accuracy: 0.8008 - val_loss: 0.3776 - val_accuracy: 0.9304\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 0s 23ms/step - loss: 0.5391 - accuracy: 0.8385 - val_loss: 0.3113 - val_accuracy: 0.9391\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 0s 23ms/step - loss: 0.4378 - accuracy: 0.8540 - val_loss: 0.2174 - val_accuracy: 0.9478\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 1s 24ms/step - loss: 0.3159 - accuracy: 0.8907 - val_loss: 0.1486 - val_accuracy: 0.9826\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 0.4023 - accuracy: 0.8859 - val_loss: 0.2846 - val_accuracy: 0.9478\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 0s 24ms/step - loss: 0.3424 - accuracy: 0.8917 - val_loss: 0.1644 - val_accuracy: 0.9565\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 1s 24ms/step - loss: 0.2722 - accuracy: 0.9033 - val_loss: 0.2609 - val_accuracy: 0.9913\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 0s 24ms/step - loss: 0.3234 - accuracy: 0.8965 - val_loss: 0.2800 - val_accuracy: 0.9478\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 0s 22ms/step - loss: 0.2908 - accuracy: 0.8926 - val_loss: 0.3762 - val_accuracy: 0.9478\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 0s 24ms/step - loss: 0.3041 - accuracy: 0.9033 - val_loss: 0.1855 - val_accuracy: 0.9478\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 1s 24ms/step - loss: 0.2227 - accuracy: 0.9197 - val_loss: 0.2556 - val_accuracy: 0.9652\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 0s 23ms/step - loss: 0.2420 - accuracy: 0.9168 - val_loss: 0.2797 - val_accuracy: 0.9478\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 1s 24ms/step - loss: 0.2926 - accuracy: 0.9091 - val_loss: 0.1095 - val_accuracy: 0.9826\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 2.2088 - accuracy: 0.1654 - val_loss: 1.3976 - val_accuracy: 0.5130\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 0s 23ms/step - loss: 1.0858 - accuracy: 0.6122 - val_loss: 0.5414 - val_accuracy: 0.8435\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 1s 24ms/step - loss: 0.5827 - accuracy: 0.7988 - val_loss: 0.3290 - val_accuracy: 0.9304\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 0s 24ms/step - loss: 0.4186 - accuracy: 0.8578 - val_loss: 0.2549 - val_accuracy: 0.9478\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 0.4646 - accuracy: 0.8588 - val_loss: 0.2843 - val_accuracy: 0.9217\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 1s 24ms/step - loss: 0.3144 - accuracy: 0.8965 - val_loss: 0.1670 - val_accuracy: 0.9478\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 0s 23ms/step - loss: 0.3208 - accuracy: 0.9081 - val_loss: 0.3324 - val_accuracy: 0.9478\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 0s 23ms/step - loss: 0.3438 - accuracy: 0.8956 - val_loss: 0.1716 - val_accuracy: 0.9652\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 0s 23ms/step - loss: 0.3330 - accuracy: 0.8975 - val_loss: 0.2043 - val_accuracy: 0.9565\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 1s 24ms/step - loss: 0.2826 - accuracy: 0.9139 - val_loss: 0.1963 - val_accuracy: 0.9652\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 1s 24ms/step - loss: 0.3330 - accuracy: 0.9072 - val_loss: 0.1358 - val_accuracy: 0.9391\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 0s 23ms/step - loss: 0.2486 - accuracy: 0.9217 - val_loss: 0.0813 - val_accuracy: 0.9826\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 0.2894 - accuracy: 0.9226 - val_loss: 0.1099 - val_accuracy: 0.9652\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 0s 23ms/step - loss: 0.2615 - accuracy: 0.9217 - val_loss: 0.1231 - val_accuracy: 0.9826\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 0.1936 - accuracy: 0.9400 - val_loss: 0.1451 - val_accuracy: 0.9739\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 1s 30ms/step - loss: 2.2665 - accuracy: 0.1372 - val_loss: 1.8531 - val_accuracy: 0.3739\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 1s 24ms/step - loss: 1.3794 - accuracy: 0.5237 - val_loss: 0.5927 - val_accuracy: 0.8174\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 0s 23ms/step - loss: 0.6669 - accuracy: 0.7855 - val_loss: 0.3636 - val_accuracy: 0.8609\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 0.5461 - accuracy: 0.8126 - val_loss: 0.3267 - val_accuracy: 0.9304\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 0s 24ms/step - loss: 0.3867 - accuracy: 0.8618 - val_loss: 0.3223 - val_accuracy: 0.9130\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 0.3839 - accuracy: 0.8725 - val_loss: 0.2625 - val_accuracy: 0.9391\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 0s 24ms/step - loss: 0.3872 - accuracy: 0.8792 - val_loss: 0.2247 - val_accuracy: 0.9391\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 0s 23ms/step - loss: 0.3286 - accuracy: 0.8899 - val_loss: 0.1924 - val_accuracy: 0.9391\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 0s 23ms/step - loss: 0.3034 - accuracy: 0.8966 - val_loss: 0.2368 - val_accuracy: 0.9217\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 0.3261 - accuracy: 0.9043 - val_loss: 0.2441 - val_accuracy: 0.9652\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 0s 23ms/step - loss: 0.2795 - accuracy: 0.9140 - val_loss: 0.2412 - val_accuracy: 0.9565\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 0s 24ms/step - loss: 0.2366 - accuracy: 0.9101 - val_loss: 0.2250 - val_accuracy: 0.9652\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 0s 23ms/step - loss: 0.2268 - accuracy: 0.9295 - val_loss: 0.2349 - val_accuracy: 0.9652\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 1s 24ms/step - loss: 0.2330 - accuracy: 0.9217 - val_loss: 0.1896 - val_accuracy: 0.9565\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 0s 24ms/step - loss: 0.2502 - accuracy: 0.9121 - val_loss: 0.2082 - val_accuracy: 0.9652\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 1s 30ms/step - loss: 2.2512 - accuracy: 0.1430 - val_loss: 1.7765 - val_accuracy: 0.3652\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 0s 23ms/step - loss: 1.1357 - accuracy: 0.6010 - val_loss: 0.5196 - val_accuracy: 0.8609\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 1s 24ms/step - loss: 0.6848 - accuracy: 0.7700 - val_loss: 0.2785 - val_accuracy: 0.9217\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 1s 24ms/step - loss: 0.5314 - accuracy: 0.8290 - val_loss: 0.2458 - val_accuracy: 0.9565\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 0s 24ms/step - loss: 0.3991 - accuracy: 0.8657 - val_loss: 0.1641 - val_accuracy: 0.9304\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 0s 24ms/step - loss: 0.4066 - accuracy: 0.8754 - val_loss: 0.2449 - val_accuracy: 0.9565\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 0.3679 - accuracy: 0.8744 - val_loss: 0.1905 - val_accuracy: 0.9478\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 0s 22ms/step - loss: 0.3464 - accuracy: 0.8918 - val_loss: 0.1832 - val_accuracy: 0.9565\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 0.2763 - accuracy: 0.9024 - val_loss: 0.2487 - val_accuracy: 0.9478\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 0.3440 - accuracy: 0.8860 - val_loss: 0.3165 - val_accuracy: 0.9565\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 0.3301 - accuracy: 0.8986 - val_loss: 0.1814 - val_accuracy: 0.9565\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 0s 22ms/step - loss: 0.3147 - accuracy: 0.9043 - val_loss: 0.2750 - val_accuracy: 0.9217\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 0s 23ms/step - loss: 0.3032 - accuracy: 0.9024 - val_loss: 0.2446 - val_accuracy: 0.9478\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 1s 24ms/step - loss: 0.3109 - accuracy: 0.9034 - val_loss: 0.2136 - val_accuracy: 0.9565\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 0.2974 - accuracy: 0.9092 - val_loss: 0.2975 - val_accuracy: 0.9565\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 1s 30ms/step - loss: 1.9048 - accuracy: 0.2879 - val_loss: 0.7833 - val_accuracy: 0.7391\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 0s 24ms/step - loss: 0.8924 - accuracy: 0.6850 - val_loss: 0.2690 - val_accuracy: 0.9391\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 0.5660 - accuracy: 0.8155 - val_loss: 0.1756 - val_accuracy: 0.9652\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 0s 22ms/step - loss: 0.4459 - accuracy: 0.8435 - val_loss: 0.1342 - val_accuracy: 0.9391\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 0s 24ms/step - loss: 0.3777 - accuracy: 0.8638 - val_loss: 0.1240 - val_accuracy: 0.9652\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 0s 23ms/step - loss: 0.3000 - accuracy: 0.8966 - val_loss: 0.0604 - val_accuracy: 0.9826\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 1s 24ms/step - loss: 0.3694 - accuracy: 0.8715 - val_loss: 0.0829 - val_accuracy: 0.9739\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 1s 24ms/step - loss: 0.3546 - accuracy: 0.8870 - val_loss: 0.0856 - val_accuracy: 0.9739\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 0s 23ms/step - loss: 0.3191 - accuracy: 0.8899 - val_loss: 0.0625 - val_accuracy: 0.9826\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 0s 24ms/step - loss: 0.3582 - accuracy: 0.8908 - val_loss: 0.0697 - val_accuracy: 0.9826\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 0.2857 - accuracy: 0.9024 - val_loss: 0.0332 - val_accuracy: 0.9913\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 0s 24ms/step - loss: 0.2911 - accuracy: 0.9053 - val_loss: 0.0795 - val_accuracy: 0.9739\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 0s 24ms/step - loss: 0.2299 - accuracy: 0.9198 - val_loss: 0.0480 - val_accuracy: 0.9913\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 0s 23ms/step - loss: 0.3528 - accuracy: 0.8783 - val_loss: 0.0618 - val_accuracy: 0.9652\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 0s 24ms/step - loss: 0.3640 - accuracy: 0.8995 - val_loss: 0.0600 - val_accuracy: 0.9913\n",
            "Epoch 1/5\n",
            "21/21 [==============================] - 0s 13ms/step - loss: 2.2967 - accuracy: 0.1286 - val_loss: 2.2844 - val_accuracy: 0.1391\n",
            "Epoch 2/5\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 2.2663 - accuracy: 0.2998 - val_loss: 2.2285 - val_accuracy: 0.4609\n",
            "Epoch 3/5\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 2.1641 - accuracy: 0.4149 - val_loss: 2.0475 - val_accuracy: 0.4783\n",
            "Epoch 4/5\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 1.8988 - accuracy: 0.4642 - val_loss: 1.6329 - val_accuracy: 0.6174\n",
            "Epoch 5/5\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 1.4755 - accuracy: 0.6267 - val_loss: 1.1752 - val_accuracy: 0.7391\n",
            "Epoch 1/5\n",
            "21/21 [==============================] - 0s 13ms/step - loss: 2.2952 - accuracy: 0.1605 - val_loss: 2.2798 - val_accuracy: 0.2000\n",
            "Epoch 2/5\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 2.2584 - accuracy: 0.2166 - val_loss: 2.2070 - val_accuracy: 0.3217\n",
            "Epoch 3/5\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 2.1363 - accuracy: 0.3946 - val_loss: 1.9829 - val_accuracy: 0.6087\n",
            "Epoch 4/5\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 1.8245 - accuracy: 0.5551 - val_loss: 1.5373 - val_accuracy: 0.7130\n",
            "Epoch 5/5\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 1.4235 - accuracy: 0.6219 - val_loss: 1.1257 - val_accuracy: 0.7565\n",
            "Epoch 1/5\n",
            "21/21 [==============================] - 0s 13ms/step - loss: 2.2978 - accuracy: 0.1188 - val_loss: 2.2855 - val_accuracy: 0.2000\n",
            "Epoch 2/5\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 2.2728 - accuracy: 0.1845 - val_loss: 2.2406 - val_accuracy: 0.3304\n",
            "Epoch 3/5\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 2.1979 - accuracy: 0.3787 - val_loss: 2.1066 - val_accuracy: 0.6348\n",
            "Epoch 4/5\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 1.9874 - accuracy: 0.5159 - val_loss: 1.7719 - val_accuracy: 0.6957\n",
            "Epoch 5/5\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 1.5841 - accuracy: 0.6386 - val_loss: 1.2744 - val_accuracy: 0.7565\n",
            "Epoch 1/5\n",
            "21/21 [==============================] - 0s 12ms/step - loss: 2.2972 - accuracy: 0.1556 - val_loss: 2.2853 - val_accuracy: 0.3217\n",
            "Epoch 2/5\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 2.2663 - accuracy: 0.3391 - val_loss: 2.2306 - val_accuracy: 0.4174\n",
            "Epoch 3/5\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 2.1638 - accuracy: 0.4522 - val_loss: 2.0490 - val_accuracy: 0.6348\n",
            "Epoch 4/5\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 1.8798 - accuracy: 0.5585 - val_loss: 1.6506 - val_accuracy: 0.6783\n",
            "Epoch 5/5\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 1.4615 - accuracy: 0.6126 - val_loss: 1.1793 - val_accuracy: 0.7652\n",
            "Epoch 1/5\n",
            "21/21 [==============================] - 0s 12ms/step - loss: 2.2957 - accuracy: 0.2145 - val_loss: 2.2858 - val_accuracy: 0.3478\n",
            "Epoch 2/5\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 2.2666 - accuracy: 0.2986 - val_loss: 2.2359 - val_accuracy: 0.3826\n",
            "Epoch 3/5\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 2.1738 - accuracy: 0.4444 - val_loss: 2.0916 - val_accuracy: 0.3826\n",
            "Epoch 4/5\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 1.9290 - accuracy: 0.4850 - val_loss: 1.7582 - val_accuracy: 0.5478\n",
            "Epoch 5/5\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 1.5518 - accuracy: 0.5961 - val_loss: 1.3705 - val_accuracy: 0.6783\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 5s 97ms/step - loss: 2.2310 - accuracy: 0.1528 - val_loss: 1.9830 - val_accuracy: 0.5130\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 3s 60ms/step - loss: 1.5324 - accuracy: 0.4874 - val_loss: 0.7876 - val_accuracy: 0.7739\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 3s 57ms/step - loss: 0.8024 - accuracy: 0.7524 - val_loss: 0.4552 - val_accuracy: 0.8783\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 3s 59ms/step - loss: 0.5490 - accuracy: 0.8269 - val_loss: 0.3354 - val_accuracy: 0.8957\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 3s 58ms/step - loss: 0.4222 - accuracy: 0.8665 - val_loss: 0.2375 - val_accuracy: 0.9391\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 3s 58ms/step - loss: 0.3237 - accuracy: 0.9101 - val_loss: 0.1895 - val_accuracy: 0.9565\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 3s 58ms/step - loss: 0.3394 - accuracy: 0.8917 - val_loss: 0.1993 - val_accuracy: 0.9478\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 3s 59ms/step - loss: 0.2792 - accuracy: 0.9120 - val_loss: 0.1654 - val_accuracy: 0.9478\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 3s 59ms/step - loss: 0.2353 - accuracy: 0.9265 - val_loss: 0.1194 - val_accuracy: 0.9652\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 3s 58ms/step - loss: 0.2346 - accuracy: 0.9236 - val_loss: 0.1284 - val_accuracy: 0.9565\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 3s 60ms/step - loss: 2.2200 - accuracy: 0.1750 - val_loss: 1.8524 - val_accuracy: 0.4261\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 3s 59ms/step - loss: 1.3160 - accuracy: 0.5725 - val_loss: 0.5909 - val_accuracy: 0.8609\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 3s 60ms/step - loss: 0.6679 - accuracy: 0.7959 - val_loss: 0.3804 - val_accuracy: 0.9304\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 3s 60ms/step - loss: 0.5015 - accuracy: 0.8366 - val_loss: 0.2851 - val_accuracy: 0.9304\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 3s 58ms/step - loss: 0.3936 - accuracy: 0.8830 - val_loss: 0.2341 - val_accuracy: 0.9478\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 3s 58ms/step - loss: 0.3309 - accuracy: 0.8888 - val_loss: 0.2679 - val_accuracy: 0.9391\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 3s 59ms/step - loss: 0.2636 - accuracy: 0.9120 - val_loss: 0.2027 - val_accuracy: 0.9478\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 3s 58ms/step - loss: 0.2310 - accuracy: 0.9188 - val_loss: 0.1730 - val_accuracy: 0.9652\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 3s 59ms/step - loss: 0.2275 - accuracy: 0.9275 - val_loss: 0.1709 - val_accuracy: 0.9652\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 3s 59ms/step - loss: 0.1905 - accuracy: 0.9400 - val_loss: 0.1731 - val_accuracy: 0.9478\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 3s 62ms/step - loss: 2.1855 - accuracy: 0.2570 - val_loss: 1.6586 - val_accuracy: 0.7217\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 3s 59ms/step - loss: 1.1360 - accuracy: 0.6560 - val_loss: 0.4975 - val_accuracy: 0.8870\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 3s 59ms/step - loss: 0.5911 - accuracy: 0.8000 - val_loss: 0.3747 - val_accuracy: 0.9043\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 3s 59ms/step - loss: 0.4696 - accuracy: 0.8541 - val_loss: 0.2995 - val_accuracy: 0.9304\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 3s 59ms/step - loss: 0.4053 - accuracy: 0.8763 - val_loss: 0.2168 - val_accuracy: 0.9391\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 3s 60ms/step - loss: 0.3247 - accuracy: 0.8966 - val_loss: 0.1784 - val_accuracy: 0.9565\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 3s 59ms/step - loss: 0.2641 - accuracy: 0.9179 - val_loss: 0.1474 - val_accuracy: 0.9739\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 3s 58ms/step - loss: 0.2265 - accuracy: 0.9304 - val_loss: 0.1267 - val_accuracy: 0.9913\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 3s 59ms/step - loss: 0.2247 - accuracy: 0.9295 - val_loss: 0.1308 - val_accuracy: 0.9652\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 3s 59ms/step - loss: 0.2029 - accuracy: 0.9382 - val_loss: 0.1232 - val_accuracy: 0.9652\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 3s 61ms/step - loss: 2.2188 - accuracy: 0.2019 - val_loss: 1.8907 - val_accuracy: 0.5652\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 3s 58ms/step - loss: 1.4239 - accuracy: 0.5304 - val_loss: 0.7376 - val_accuracy: 0.8261\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 3s 59ms/step - loss: 0.8257 - accuracy: 0.7440 - val_loss: 0.4389 - val_accuracy: 0.8783\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 3s 58ms/step - loss: 0.5595 - accuracy: 0.8290 - val_loss: 0.3171 - val_accuracy: 0.9304\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 3s 58ms/step - loss: 0.4484 - accuracy: 0.8599 - val_loss: 0.3228 - val_accuracy: 0.9043\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 3s 59ms/step - loss: 0.3518 - accuracy: 0.8860 - val_loss: 0.2266 - val_accuracy: 0.9478\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 3s 58ms/step - loss: 0.3123 - accuracy: 0.8995 - val_loss: 0.2552 - val_accuracy: 0.9391\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 3s 58ms/step - loss: 0.2965 - accuracy: 0.9101 - val_loss: 0.1557 - val_accuracy: 0.9652\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 3s 59ms/step - loss: 0.2418 - accuracy: 0.9208 - val_loss: 0.1529 - val_accuracy: 0.9652\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 3s 58ms/step - loss: 0.2253 - accuracy: 0.9285 - val_loss: 0.1497 - val_accuracy: 0.9739\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 3s 62ms/step - loss: 2.1091 - accuracy: 0.2831 - val_loss: 1.4942 - val_accuracy: 0.7652\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 3s 59ms/step - loss: 1.0852 - accuracy: 0.6599 - val_loss: 0.5454 - val_accuracy: 0.8870\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 3s 59ms/step - loss: 0.6267 - accuracy: 0.8126 - val_loss: 0.3048 - val_accuracy: 0.9130\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 3s 59ms/step - loss: 0.4541 - accuracy: 0.8560 - val_loss: 0.2616 - val_accuracy: 0.9217\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 3s 59ms/step - loss: 0.3886 - accuracy: 0.8802 - val_loss: 0.1323 - val_accuracy: 0.9826\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 3s 59ms/step - loss: 0.3220 - accuracy: 0.8918 - val_loss: 0.1023 - val_accuracy: 0.9652\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 3s 59ms/step - loss: 0.2754 - accuracy: 0.9101 - val_loss: 0.0819 - val_accuracy: 0.9826\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 3s 59ms/step - loss: 0.2518 - accuracy: 0.9188 - val_loss: 0.0717 - val_accuracy: 0.9739\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 3s 59ms/step - loss: 0.2092 - accuracy: 0.9304 - val_loss: 0.0364 - val_accuracy: 1.0000\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 3s 58ms/step - loss: 0.2014 - accuracy: 0.9353 - val_loss: 0.0569 - val_accuracy: 0.9913\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 1s 10ms/step - loss: 2.2761 - accuracy: 0.1596 - val_loss: 2.1914 - val_accuracy: 0.3217\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 2.0116 - accuracy: 0.2379 - val_loss: 1.6533 - val_accuracy: 0.4261\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 1.6660 - accuracy: 0.3665 - val_loss: 1.3071 - val_accuracy: 0.6783\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 0s 8ms/step - loss: 1.4230 - accuracy: 0.4768 - val_loss: 1.0480 - val_accuracy: 0.7217\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 1.2385 - accuracy: 0.5435 - val_loss: 0.8767 - val_accuracy: 0.8174\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 1.1348 - accuracy: 0.5986 - val_loss: 0.7549 - val_accuracy: 0.8348\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.9541 - accuracy: 0.6702 - val_loss: 0.6230 - val_accuracy: 0.8522\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.8346 - accuracy: 0.7060 - val_loss: 0.5198 - val_accuracy: 0.8783\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.7744 - accuracy: 0.7282 - val_loss: 0.4729 - val_accuracy: 0.8609\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.7021 - accuracy: 0.7515 - val_loss: 0.4260 - val_accuracy: 0.9304\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.6363 - accuracy: 0.7669 - val_loss: 0.4380 - val_accuracy: 0.9130\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.5818 - accuracy: 0.8056 - val_loss: 0.3713 - val_accuracy: 0.9391\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.5355 - accuracy: 0.8075 - val_loss: 0.3240 - val_accuracy: 0.9391\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.5536 - accuracy: 0.7988 - val_loss: 0.3179 - val_accuracy: 0.9217\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.4639 - accuracy: 0.8424 - val_loss: 0.2792 - val_accuracy: 0.9652\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 1s 10ms/step - loss: 2.2829 - accuracy: 0.1538 - val_loss: 2.1970 - val_accuracy: 0.2783\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 1.9364 - accuracy: 0.3327 - val_loss: 1.4082 - val_accuracy: 0.7130\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 1.4038 - accuracy: 0.4932 - val_loss: 0.8667 - val_accuracy: 0.8000\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 1.1091 - accuracy: 0.6132 - val_loss: 0.6592 - val_accuracy: 0.8261\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.9777 - accuracy: 0.6431 - val_loss: 0.5523 - val_accuracy: 0.8435\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.8376 - accuracy: 0.6886 - val_loss: 0.4382 - val_accuracy: 0.8783\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.7670 - accuracy: 0.7350 - val_loss: 0.3818 - val_accuracy: 0.9217\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 0s 8ms/step - loss: 0.6994 - accuracy: 0.7495 - val_loss: 0.3375 - val_accuracy: 0.9217\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.6389 - accuracy: 0.7785 - val_loss: 0.3091 - val_accuracy: 0.9217\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.6018 - accuracy: 0.7766 - val_loss: 0.2841 - val_accuracy: 0.9478\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.5256 - accuracy: 0.8221 - val_loss: 0.2461 - val_accuracy: 0.9478\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 0s 8ms/step - loss: 0.5144 - accuracy: 0.8279 - val_loss: 0.2456 - val_accuracy: 0.9478\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.5073 - accuracy: 0.8153 - val_loss: 0.2125 - val_accuracy: 0.9478\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.4789 - accuracy: 0.8259 - val_loss: 0.2242 - val_accuracy: 0.9565\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.4308 - accuracy: 0.8501 - val_loss: 0.2028 - val_accuracy: 0.9652\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 0s 9ms/step - loss: 2.2931 - accuracy: 0.1246 - val_loss: 2.2594 - val_accuracy: 0.2087\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 2.1386 - accuracy: 0.1787 - val_loss: 1.8289 - val_accuracy: 0.4522\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 1.8310 - accuracy: 0.2667 - val_loss: 1.4538 - val_accuracy: 0.6087\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 1.7066 - accuracy: 0.2928 - val_loss: 1.3124 - val_accuracy: 0.6174\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 1.5962 - accuracy: 0.3623 - val_loss: 1.1533 - val_accuracy: 0.7565\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 1.4585 - accuracy: 0.4019 - val_loss: 0.9744 - val_accuracy: 0.7217\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 1.1320 - accuracy: 0.5778 - val_loss: 0.6277 - val_accuracy: 0.8783\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.8552 - accuracy: 0.6696 - val_loss: 0.4210 - val_accuracy: 0.8957\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.7880 - accuracy: 0.7082 - val_loss: 0.3600 - val_accuracy: 0.9130\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 0s 8ms/step - loss: 0.6402 - accuracy: 0.7652 - val_loss: 0.2926 - val_accuracy: 0.9565\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 0s 8ms/step - loss: 0.6439 - accuracy: 0.7739 - val_loss: 0.2740 - val_accuracy: 0.9565\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.5748 - accuracy: 0.8048 - val_loss: 0.2266 - val_accuracy: 0.9478\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.5360 - accuracy: 0.8193 - val_loss: 0.2063 - val_accuracy: 0.9739\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.5314 - accuracy: 0.8029 - val_loss: 0.2404 - val_accuracy: 0.9652\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.5203 - accuracy: 0.8097 - val_loss: 0.2061 - val_accuracy: 0.9739\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 0s 10ms/step - loss: 2.2676 - accuracy: 0.1420 - val_loss: 2.1556 - val_accuracy: 0.2609\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 0s 8ms/step - loss: 1.9104 - accuracy: 0.2686 - val_loss: 1.5195 - val_accuracy: 0.4522\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 1.4001 - accuracy: 0.5053 - val_loss: 0.9231 - val_accuracy: 0.7913\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 1.0654 - accuracy: 0.6261 - val_loss: 0.6401 - val_accuracy: 0.8261\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.8669 - accuracy: 0.6966 - val_loss: 0.4833 - val_accuracy: 0.9130\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.7540 - accuracy: 0.7430 - val_loss: 0.4364 - val_accuracy: 0.9130\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.6949 - accuracy: 0.7488 - val_loss: 0.3645 - val_accuracy: 0.9391\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.5769 - accuracy: 0.7961 - val_loss: 0.3056 - val_accuracy: 0.9391\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.6064 - accuracy: 0.7952 - val_loss: 0.2800 - val_accuracy: 0.9565\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 0s 8ms/step - loss: 0.5357 - accuracy: 0.8077 - val_loss: 0.2749 - val_accuracy: 0.9478\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.5037 - accuracy: 0.8203 - val_loss: 0.2258 - val_accuracy: 0.9652\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.5042 - accuracy: 0.8251 - val_loss: 0.2342 - val_accuracy: 0.9565\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 0s 8ms/step - loss: 0.4401 - accuracy: 0.8435 - val_loss: 0.2002 - val_accuracy: 0.9739\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.4172 - accuracy: 0.8609 - val_loss: 0.2064 - val_accuracy: 0.9739\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.4231 - accuracy: 0.8531 - val_loss: 0.1993 - val_accuracy: 0.9652\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 1s 10ms/step - loss: 2.2751 - accuracy: 0.1362 - val_loss: 2.1741 - val_accuracy: 0.2783\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 1.9680 - accuracy: 0.2976 - val_loss: 1.5459 - val_accuracy: 0.5652\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 0s 8ms/step - loss: 1.3799 - accuracy: 0.4957 - val_loss: 0.9324 - val_accuracy: 0.7652\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 1.1305 - accuracy: 0.6077 - val_loss: 0.6663 - val_accuracy: 0.8957\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.9088 - accuracy: 0.6696 - val_loss: 0.4719 - val_accuracy: 0.8870\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.7994 - accuracy: 0.7217 - val_loss: 0.4208 - val_accuracy: 0.9391\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.7138 - accuracy: 0.7527 - val_loss: 0.3496 - val_accuracy: 0.9217\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.6322 - accuracy: 0.7826 - val_loss: 0.2864 - val_accuracy: 0.9478\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 0s 8ms/step - loss: 0.5758 - accuracy: 0.7923 - val_loss: 0.2347 - val_accuracy: 0.9391\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.5870 - accuracy: 0.7816 - val_loss: 0.2136 - val_accuracy: 0.9565\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 0s 8ms/step - loss: 0.5293 - accuracy: 0.8135 - val_loss: 0.2164 - val_accuracy: 0.9304\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 0s 8ms/step - loss: 0.4995 - accuracy: 0.8135 - val_loss: 0.1968 - val_accuracy: 0.9391\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.4853 - accuracy: 0.8319 - val_loss: 0.1463 - val_accuracy: 0.9739\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 0s 8ms/step - loss: 0.4603 - accuracy: 0.8348 - val_loss: 0.1608 - val_accuracy: 0.9565\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.4150 - accuracy: 0.8483 - val_loss: 0.1331 - val_accuracy: 0.9478\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 3s 130ms/step - loss: 2.2906 - accuracy: 0.1712 - val_loss: 1.8833 - val_accuracy: 0.3304\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 3s 125ms/step - loss: 1.2531 - accuracy: 0.5513 - val_loss: 0.7305 - val_accuracy: 0.7652\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 3s 128ms/step - loss: 0.5853 - accuracy: 0.8095 - val_loss: 0.2964 - val_accuracy: 0.9391\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 3s 127ms/step - loss: 0.4320 - accuracy: 0.8578 - val_loss: 0.2729 - val_accuracy: 0.9130\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 3s 128ms/step - loss: 0.3451 - accuracy: 0.8868 - val_loss: 0.3068 - val_accuracy: 0.9130\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 3s 128ms/step - loss: 0.3884 - accuracy: 0.8714 - val_loss: 0.2806 - val_accuracy: 0.9478\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 3s 126ms/step - loss: 0.3313 - accuracy: 0.9004 - val_loss: 0.3093 - val_accuracy: 0.9478\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 3s 125ms/step - loss: 0.3431 - accuracy: 0.8985 - val_loss: 0.3426 - val_accuracy: 0.9478\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 3s 125ms/step - loss: 0.2324 - accuracy: 0.9304 - val_loss: 0.2102 - val_accuracy: 0.9652\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 3s 126ms/step - loss: 0.1929 - accuracy: 0.9420 - val_loss: 0.1593 - val_accuracy: 0.9652\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 3s 126ms/step - loss: 0.2129 - accuracy: 0.9400 - val_loss: 0.2792 - val_accuracy: 0.9652\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 3s 126ms/step - loss: 0.1942 - accuracy: 0.9449 - val_loss: 0.1426 - val_accuracy: 0.9652\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 3s 125ms/step - loss: 0.2411 - accuracy: 0.9236 - val_loss: 0.2408 - val_accuracy: 0.9565\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 3s 125ms/step - loss: 0.2617 - accuracy: 0.9304 - val_loss: 0.1590 - val_accuracy: 0.9652\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 3s 125ms/step - loss: 0.3384 - accuracy: 0.9207 - val_loss: 0.2839 - val_accuracy: 0.9565\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 3s 129ms/step - loss: 2.4083 - accuracy: 0.1132 - val_loss: 2.2384 - val_accuracy: 0.4174\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 3s 123ms/step - loss: 2.0864 - accuracy: 0.2573 - val_loss: 1.4286 - val_accuracy: 0.5217\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 3s 125ms/step - loss: 1.0924 - accuracy: 0.6277 - val_loss: 0.5025 - val_accuracy: 0.8522\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 3s 124ms/step - loss: 0.6398 - accuracy: 0.7824 - val_loss: 0.4220 - val_accuracy: 0.9130\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 3s 124ms/step - loss: 0.4497 - accuracy: 0.8578 - val_loss: 0.3577 - val_accuracy: 0.9304\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 3s 124ms/step - loss: 0.3500 - accuracy: 0.8733 - val_loss: 0.2735 - val_accuracy: 0.9217\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 3s 124ms/step - loss: 0.2860 - accuracy: 0.9091 - val_loss: 0.3030 - val_accuracy: 0.9304\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 3s 123ms/step - loss: 0.2611 - accuracy: 0.9139 - val_loss: 0.2289 - val_accuracy: 0.9565\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 3s 124ms/step - loss: 0.2161 - accuracy: 0.9207 - val_loss: 0.2416 - val_accuracy: 0.9478\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 3s 124ms/step - loss: 0.3212 - accuracy: 0.9101 - val_loss: 0.2661 - val_accuracy: 0.9391\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 3s 123ms/step - loss: 0.2580 - accuracy: 0.9226 - val_loss: 0.2473 - val_accuracy: 0.9217\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 3s 124ms/step - loss: 0.2340 - accuracy: 0.9284 - val_loss: 0.2073 - val_accuracy: 0.9391\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 3s 123ms/step - loss: 0.2405 - accuracy: 0.9207 - val_loss: 0.1325 - val_accuracy: 0.9739\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 3s 124ms/step - loss: 0.2744 - accuracy: 0.9207 - val_loss: 0.2528 - val_accuracy: 0.9391\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 3s 124ms/step - loss: 0.2784 - accuracy: 0.9139 - val_loss: 0.1538 - val_accuracy: 0.9565\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 3s 131ms/step - loss: 2.3731 - accuracy: 0.0957 - val_loss: 2.2905 - val_accuracy: 0.1304\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 3s 121ms/step - loss: 2.0234 - accuracy: 0.2850 - val_loss: 1.1871 - val_accuracy: 0.6348\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 3s 123ms/step - loss: 0.9646 - accuracy: 0.6821 - val_loss: 0.4218 - val_accuracy: 0.8696\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 3s 122ms/step - loss: 0.4939 - accuracy: 0.8396 - val_loss: 0.3337 - val_accuracy: 0.9217\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 3s 121ms/step - loss: 0.4067 - accuracy: 0.8686 - val_loss: 0.2864 - val_accuracy: 0.9304\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 3s 122ms/step - loss: 0.2976 - accuracy: 0.8995 - val_loss: 0.1954 - val_accuracy: 0.9565\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 3s 122ms/step - loss: 0.2711 - accuracy: 0.8986 - val_loss: 0.3140 - val_accuracy: 0.9304\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 3s 121ms/step - loss: 0.3480 - accuracy: 0.8870 - val_loss: 0.2775 - val_accuracy: 0.9130\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 3s 124ms/step - loss: 0.2919 - accuracy: 0.9101 - val_loss: 0.2277 - val_accuracy: 0.9391\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 3s 123ms/step - loss: 0.2248 - accuracy: 0.9266 - val_loss: 0.1370 - val_accuracy: 0.9652\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 3s 122ms/step - loss: 0.1988 - accuracy: 0.9459 - val_loss: 0.2147 - val_accuracy: 0.9652\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 3s 122ms/step - loss: 0.1992 - accuracy: 0.9372 - val_loss: 0.1604 - val_accuracy: 0.9739\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 3s 122ms/step - loss: 0.2242 - accuracy: 0.9198 - val_loss: 0.2653 - val_accuracy: 0.9391\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 3s 122ms/step - loss: 0.2661 - accuracy: 0.9130 - val_loss: 0.2298 - val_accuracy: 0.9652\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 3s 122ms/step - loss: 0.2504 - accuracy: 0.9295 - val_loss: 0.2151 - val_accuracy: 0.9739\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 3s 129ms/step - loss: 2.3624 - accuracy: 0.1053 - val_loss: 2.3069 - val_accuracy: 0.1043\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 3s 123ms/step - loss: 2.2671 - accuracy: 0.1469 - val_loss: 2.0742 - val_accuracy: 0.3043\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 3s 123ms/step - loss: 1.7374 - accuracy: 0.3565 - val_loss: 1.0962 - val_accuracy: 0.6696\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 3s 124ms/step - loss: 1.2286 - accuracy: 0.5459 - val_loss: 0.8490 - val_accuracy: 0.7565\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 3s 122ms/step - loss: 0.8392 - accuracy: 0.7208 - val_loss: 0.4688 - val_accuracy: 0.8609\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 3s 123ms/step - loss: 0.5694 - accuracy: 0.7961 - val_loss: 0.2805 - val_accuracy: 0.9304\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 3s 124ms/step - loss: 0.4615 - accuracy: 0.8406 - val_loss: 0.2841 - val_accuracy: 0.9478\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 3s 122ms/step - loss: 0.4866 - accuracy: 0.8473 - val_loss: 0.2951 - val_accuracy: 0.9391\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 3s 124ms/step - loss: 0.4394 - accuracy: 0.8435 - val_loss: 0.2581 - val_accuracy: 0.9217\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 3s 126ms/step - loss: 0.4325 - accuracy: 0.8638 - val_loss: 0.2333 - val_accuracy: 0.9217\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 3s 127ms/step - loss: 0.3154 - accuracy: 0.8995 - val_loss: 0.1399 - val_accuracy: 0.9565\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 3s 126ms/step - loss: 0.3231 - accuracy: 0.8928 - val_loss: 0.1713 - val_accuracy: 0.9304\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 3s 123ms/step - loss: 0.3740 - accuracy: 0.8957 - val_loss: 0.3637 - val_accuracy: 0.8957\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 3s 124ms/step - loss: 0.4116 - accuracy: 0.8860 - val_loss: 0.2571 - val_accuracy: 0.9304\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 3s 124ms/step - loss: 0.4138 - accuracy: 0.8812 - val_loss: 0.1399 - val_accuracy: 0.9826\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 3s 130ms/step - loss: 2.3421 - accuracy: 0.1575 - val_loss: 2.0132 - val_accuracy: 0.2783\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 3s 124ms/step - loss: 1.3159 - accuracy: 0.5159 - val_loss: 0.7445 - val_accuracy: 0.7391\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 3s 124ms/step - loss: 0.6174 - accuracy: 0.7884 - val_loss: 0.3700 - val_accuracy: 0.8435\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 3s 123ms/step - loss: 0.4323 - accuracy: 0.8473 - val_loss: 0.2082 - val_accuracy: 0.9391\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 3s 122ms/step - loss: 0.3076 - accuracy: 0.8976 - val_loss: 0.2689 - val_accuracy: 0.9478\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 3s 124ms/step - loss: 0.2766 - accuracy: 0.9179 - val_loss: 0.1401 - val_accuracy: 0.9565\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 3s 122ms/step - loss: 0.2504 - accuracy: 0.9111 - val_loss: 0.0810 - val_accuracy: 0.9826\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 3s 123ms/step - loss: 0.2764 - accuracy: 0.9111 - val_loss: 0.0742 - val_accuracy: 0.9652\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 3s 124ms/step - loss: 0.2257 - accuracy: 0.9343 - val_loss: 0.0548 - val_accuracy: 0.9826\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 3s 125ms/step - loss: 0.2056 - accuracy: 0.9314 - val_loss: 0.0818 - val_accuracy: 0.9913\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 3s 123ms/step - loss: 0.2493 - accuracy: 0.9275 - val_loss: 0.0917 - val_accuracy: 0.9826\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 3s 124ms/step - loss: 0.2287 - accuracy: 0.9237 - val_loss: 0.0579 - val_accuracy: 0.9826\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 3s 125ms/step - loss: 0.1851 - accuracy: 0.9430 - val_loss: 0.0479 - val_accuracy: 0.9913\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 3s 124ms/step - loss: 0.2042 - accuracy: 0.9324 - val_loss: 0.0412 - val_accuracy: 0.9913\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 3s 125ms/step - loss: 0.2251 - accuracy: 0.9169 - val_loss: 0.0547 - val_accuracy: 0.9826\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 3s 63ms/step - loss: 2.1471 - accuracy: 0.2689 - val_loss: 1.6923 - val_accuracy: 0.5478\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 3s 61ms/step - loss: 1.2838 - accuracy: 0.6074 - val_loss: 0.7881 - val_accuracy: 0.7739\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 3s 61ms/step - loss: 0.7271 - accuracy: 0.7814 - val_loss: 0.5066 - val_accuracy: 0.8261\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 3s 61ms/step - loss: 0.5214 - accuracy: 0.8511 - val_loss: 0.3857 - val_accuracy: 0.9043\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 3s 62ms/step - loss: 0.3911 - accuracy: 0.8781 - val_loss: 0.2743 - val_accuracy: 0.9043\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 3s 62ms/step - loss: 0.3301 - accuracy: 0.8946 - val_loss: 0.2179 - val_accuracy: 0.9304\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 3s 62ms/step - loss: 0.2850 - accuracy: 0.9091 - val_loss: 0.1757 - val_accuracy: 0.9478\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 3s 62ms/step - loss: 0.2394 - accuracy: 0.9197 - val_loss: 0.1348 - val_accuracy: 0.9565\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 3s 61ms/step - loss: 0.1756 - accuracy: 0.9516 - val_loss: 0.1294 - val_accuracy: 0.9478\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 3s 61ms/step - loss: 0.1808 - accuracy: 0.9381 - val_loss: 0.1049 - val_accuracy: 0.9478\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 3s 64ms/step - loss: 2.1257 - accuracy: 0.3095 - val_loss: 1.6225 - val_accuracy: 0.5826\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 3s 61ms/step - loss: 1.2087 - accuracy: 0.6373 - val_loss: 0.7359 - val_accuracy: 0.8348\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 3s 62ms/step - loss: 0.7177 - accuracy: 0.7892 - val_loss: 0.5426 - val_accuracy: 0.8696\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 3s 61ms/step - loss: 0.4937 - accuracy: 0.8569 - val_loss: 0.2915 - val_accuracy: 0.9478\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 3s 61ms/step - loss: 0.3693 - accuracy: 0.8917 - val_loss: 0.2347 - val_accuracy: 0.9478\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 3s 61ms/step - loss: 0.2950 - accuracy: 0.9110 - val_loss: 0.1756 - val_accuracy: 0.9565\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 3s 61ms/step - loss: 0.2376 - accuracy: 0.9275 - val_loss: 0.1723 - val_accuracy: 0.9478\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 3s 61ms/step - loss: 0.2172 - accuracy: 0.9391 - val_loss: 0.1408 - val_accuracy: 0.9478\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 3s 60ms/step - loss: 0.1774 - accuracy: 0.9497 - val_loss: 0.1326 - val_accuracy: 0.9391\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 3s 62ms/step - loss: 0.1813 - accuracy: 0.9458 - val_loss: 0.1328 - val_accuracy: 0.9739\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 3s 63ms/step - loss: 2.1600 - accuracy: 0.2763 - val_loss: 1.6848 - val_accuracy: 0.6000\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 3s 61ms/step - loss: 1.2309 - accuracy: 0.6406 - val_loss: 0.7198 - val_accuracy: 0.7913\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 3s 62ms/step - loss: 0.6755 - accuracy: 0.8087 - val_loss: 0.4409 - val_accuracy: 0.8870\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 3s 62ms/step - loss: 0.4727 - accuracy: 0.8754 - val_loss: 0.3066 - val_accuracy: 0.9043\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 3s 62ms/step - loss: 0.3897 - accuracy: 0.8850 - val_loss: 0.2436 - val_accuracy: 0.9304\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 3s 61ms/step - loss: 0.3021 - accuracy: 0.9140 - val_loss: 0.1923 - val_accuracy: 0.9565\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 3s 62ms/step - loss: 0.2657 - accuracy: 0.9246 - val_loss: 0.1733 - val_accuracy: 0.9391\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 3s 62ms/step - loss: 0.2192 - accuracy: 0.9372 - val_loss: 0.1470 - val_accuracy: 0.9739\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 3s 62ms/step - loss: 0.1931 - accuracy: 0.9430 - val_loss: 0.1348 - val_accuracy: 0.9565\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 3s 60ms/step - loss: 0.1696 - accuracy: 0.9498 - val_loss: 0.1300 - val_accuracy: 0.9652\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 3s 63ms/step - loss: 2.1236 - accuracy: 0.3092 - val_loss: 1.6252 - val_accuracy: 0.6957\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 3s 61ms/step - loss: 1.2313 - accuracy: 0.6261 - val_loss: 0.7903 - val_accuracy: 0.7652\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 3s 62ms/step - loss: 0.7325 - accuracy: 0.7768 - val_loss: 0.6537 - val_accuracy: 0.8087\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 3s 63ms/step - loss: 0.5485 - accuracy: 0.8300 - val_loss: 0.4051 - val_accuracy: 0.8783\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 3s 63ms/step - loss: 0.3903 - accuracy: 0.8860 - val_loss: 0.2931 - val_accuracy: 0.9043\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 3s 63ms/step - loss: 0.3214 - accuracy: 0.9043 - val_loss: 0.2222 - val_accuracy: 0.9217\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 3s 62ms/step - loss: 0.2865 - accuracy: 0.9150 - val_loss: 0.1771 - val_accuracy: 0.9391\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 3s 62ms/step - loss: 0.2222 - accuracy: 0.9285 - val_loss: 0.1548 - val_accuracy: 0.9565\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 3s 61ms/step - loss: 0.1998 - accuracy: 0.9401 - val_loss: 0.1394 - val_accuracy: 0.9652\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 3s 62ms/step - loss: 0.1794 - accuracy: 0.9488 - val_loss: 0.1218 - val_accuracy: 0.9565\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 3s 64ms/step - loss: 2.1367 - accuracy: 0.3391 - val_loss: 1.6704 - val_accuracy: 0.5391\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 3s 61ms/step - loss: 1.1850 - accuracy: 0.6512 - val_loss: 0.6803 - val_accuracy: 0.8783\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 3s 61ms/step - loss: 0.6454 - accuracy: 0.7932 - val_loss: 0.3796 - val_accuracy: 0.9130\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 3s 61ms/step - loss: 0.4363 - accuracy: 0.8676 - val_loss: 0.2798 - val_accuracy: 0.9304\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 3s 60ms/step - loss: 0.3469 - accuracy: 0.8918 - val_loss: 0.2479 - val_accuracy: 0.9217\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 3s 61ms/step - loss: 0.2812 - accuracy: 0.9063 - val_loss: 0.1789 - val_accuracy: 0.9391\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 3s 62ms/step - loss: 0.2604 - accuracy: 0.9256 - val_loss: 0.1387 - val_accuracy: 0.9565\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 3s 62ms/step - loss: 0.2190 - accuracy: 0.9333 - val_loss: 0.1104 - val_accuracy: 0.9652\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 3s 62ms/step - loss: 0.1879 - accuracy: 0.9420 - val_loss: 0.0866 - val_accuracy: 0.9652\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 3s 60ms/step - loss: 0.1781 - accuracy: 0.9449 - val_loss: 0.0618 - val_accuracy: 0.9913\n",
            "Epoch 1/5\n",
            "7/7 [==============================] - 1s 203ms/step - loss: 2.2976 - accuracy: 0.1663 - val_loss: 2.2839 - val_accuracy: 0.2957\n",
            "Epoch 2/5\n",
            "7/7 [==============================] - 1s 191ms/step - loss: 2.2641 - accuracy: 0.2002 - val_loss: 2.2152 - val_accuracy: 0.3826\n",
            "Epoch 3/5\n",
            "7/7 [==============================] - 1s 187ms/step - loss: 2.1619 - accuracy: 0.3501 - val_loss: 2.0221 - val_accuracy: 0.4435\n",
            "Epoch 4/5\n",
            "7/7 [==============================] - 1s 192ms/step - loss: 1.9117 - accuracy: 0.5116 - val_loss: 1.6130 - val_accuracy: 0.7304\n",
            "Epoch 5/5\n",
            "7/7 [==============================] - 1s 189ms/step - loss: 1.4980 - accuracy: 0.6364 - val_loss: 1.1225 - val_accuracy: 0.8000\n",
            "Epoch 1/5\n",
            "7/7 [==============================] - 1s 207ms/step - loss: 2.2993 - accuracy: 0.1209 - val_loss: 2.2890 - val_accuracy: 0.1304\n",
            "Epoch 2/5\n",
            "7/7 [==============================] - 1s 191ms/step - loss: 2.2764 - accuracy: 0.1828 - val_loss: 2.2460 - val_accuracy: 0.2087\n",
            "Epoch 3/5\n",
            "7/7 [==============================] - 1s 188ms/step - loss: 2.2102 - accuracy: 0.2689 - val_loss: 2.1262 - val_accuracy: 0.3913\n",
            "Epoch 4/5\n",
            "7/7 [==============================] - 1s 192ms/step - loss: 2.0553 - accuracy: 0.3762 - val_loss: 1.8355 - val_accuracy: 0.6783\n",
            "Epoch 5/5\n",
            "7/7 [==============================] - 1s 186ms/step - loss: 1.7265 - accuracy: 0.5522 - val_loss: 1.3970 - val_accuracy: 0.7217\n",
            "Epoch 1/5\n",
            "7/7 [==============================] - 1s 207ms/step - loss: 2.2979 - accuracy: 0.1324 - val_loss: 2.2812 - val_accuracy: 0.2000\n",
            "Epoch 2/5\n",
            "7/7 [==============================] - 1s 193ms/step - loss: 2.2743 - accuracy: 0.1990 - val_loss: 2.2363 - val_accuracy: 0.3565\n",
            "Epoch 3/5\n",
            "7/7 [==============================] - 1s 188ms/step - loss: 2.2178 - accuracy: 0.2754 - val_loss: 2.1275 - val_accuracy: 0.4783\n",
            "Epoch 4/5\n",
            "7/7 [==============================] - 1s 193ms/step - loss: 2.0822 - accuracy: 0.3700 - val_loss: 1.8966 - val_accuracy: 0.6000\n",
            "Epoch 5/5\n",
            "7/7 [==============================] - 1s 188ms/step - loss: 1.8022 - accuracy: 0.5179 - val_loss: 1.4665 - val_accuracy: 0.7913\n",
            "Epoch 1/5\n",
            "7/7 [==============================] - 1s 211ms/step - loss: 2.2991 - accuracy: 0.1121 - val_loss: 2.2887 - val_accuracy: 0.1130\n",
            "Epoch 2/5\n",
            "7/7 [==============================] - 1s 188ms/step - loss: 2.2785 - accuracy: 0.1498 - val_loss: 2.2488 - val_accuracy: 0.1565\n",
            "Epoch 3/5\n",
            "7/7 [==============================] - 1s 191ms/step - loss: 2.2176 - accuracy: 0.2251 - val_loss: 2.1286 - val_accuracy: 0.3304\n",
            "Epoch 4/5\n",
            "7/7 [==============================] - 1s 190ms/step - loss: 2.0497 - accuracy: 0.3749 - val_loss: 1.8535 - val_accuracy: 0.7304\n",
            "Epoch 5/5\n",
            "7/7 [==============================] - 1s 186ms/step - loss: 1.7145 - accuracy: 0.5884 - val_loss: 1.3614 - val_accuracy: 0.7739\n",
            "WARNING:tensorflow:5 out of the last 22 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd09dfef28> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/5\n",
            "7/7 [==============================] - 1s 210ms/step - loss: 2.2994 - accuracy: 0.1159 - val_loss: 2.2895 - val_accuracy: 0.2522\n",
            "Epoch 2/5\n",
            "7/7 [==============================] - 1s 190ms/step - loss: 2.2801 - accuracy: 0.2087 - val_loss: 2.2572 - val_accuracy: 0.4435\n",
            "Epoch 3/5\n",
            "7/7 [==============================] - 1s 192ms/step - loss: 2.2285 - accuracy: 0.3053 - val_loss: 2.1654 - val_accuracy: 0.4435\n",
            "Epoch 4/5\n",
            "7/7 [==============================] - 1s 190ms/step - loss: 2.0979 - accuracy: 0.3903 - val_loss: 1.9399 - val_accuracy: 0.5478\n",
            "Epoch 5/5\n",
            "7/7 [==============================] - 1s 191ms/step - loss: 1.8304 - accuracy: 0.5024 - val_loss: 1.5232 - val_accuracy: 0.8348\n",
            "WARNING:tensorflow:6 out of the last 24 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd09dfe048> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 4s 79ms/step - loss: 2.3580 - accuracy: 0.0986 - val_loss: 2.3004 - val_accuracy: 0.1130\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 4s 76ms/step - loss: 2.3044 - accuracy: 0.0948 - val_loss: 2.3020 - val_accuracy: 0.1043\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 4s 78ms/step - loss: 2.3030 - accuracy: 0.1064 - val_loss: 2.3042 - val_accuracy: 0.1043\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 4s 77ms/step - loss: 2.3026 - accuracy: 0.1025 - val_loss: 2.3047 - val_accuracy: 0.1043\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 4s 78ms/step - loss: 2.3022 - accuracy: 0.1141 - val_loss: 2.3077 - val_accuracy: 0.1043\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 4s 76ms/step - loss: 2.3031 - accuracy: 0.1141 - val_loss: 2.3065 - val_accuracy: 0.1043\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 4s 76ms/step - loss: 2.3035 - accuracy: 0.1141 - val_loss: 2.3068 - val_accuracy: 0.1043\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 4s 76ms/step - loss: 2.3026 - accuracy: 0.1054 - val_loss: 2.3082 - val_accuracy: 0.1043\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 4s 77ms/step - loss: 2.3022 - accuracy: 0.1141 - val_loss: 2.3065 - val_accuracy: 0.1043\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 4s 78ms/step - loss: 2.3023 - accuracy: 0.1006 - val_loss: 2.3075 - val_accuracy: 0.1043\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 4s 80ms/step - loss: 2.3027 - accuracy: 0.1141 - val_loss: 2.3068 - val_accuracy: 0.1043\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 4s 77ms/step - loss: 2.3024 - accuracy: 0.1141 - val_loss: 2.3063 - val_accuracy: 0.1043\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 4s 77ms/step - loss: 2.3021 - accuracy: 0.1141 - val_loss: 2.3059 - val_accuracy: 0.1043\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 4s 77ms/step - loss: 2.3021 - accuracy: 0.1141 - val_loss: 2.3063 - val_accuracy: 0.1043\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 4s 78ms/step - loss: 2.3022 - accuracy: 0.1141 - val_loss: 2.3068 - val_accuracy: 0.1043\n",
            "WARNING:tensorflow:6 out of the last 11 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd0cb39ae8> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 4s 80ms/step - loss: 2.3445 - accuracy: 0.1054 - val_loss: 2.2023 - val_accuracy: 0.2087\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 4s 77ms/step - loss: 1.7302 - accuracy: 0.4023 - val_loss: 0.9948 - val_accuracy: 0.7130\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 4s 78ms/step - loss: 1.1433 - accuracy: 0.6170 - val_loss: 0.9591 - val_accuracy: 0.7565\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 4s 77ms/step - loss: 0.8798 - accuracy: 0.7215 - val_loss: 0.4938 - val_accuracy: 0.8609\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 4s 78ms/step - loss: 0.7522 - accuracy: 0.7785 - val_loss: 0.6433 - val_accuracy: 0.8261\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 4s 77ms/step - loss: 1.1452 - accuracy: 0.6828 - val_loss: 1.1720 - val_accuracy: 0.6348\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 4s 79ms/step - loss: 1.4561 - accuracy: 0.4884 - val_loss: 0.6491 - val_accuracy: 0.7913\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 4s 77ms/step - loss: 1.1702 - accuracy: 0.6431 - val_loss: 0.6924 - val_accuracy: 0.7739\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 4s 78ms/step - loss: 0.7234 - accuracy: 0.7737 - val_loss: 0.3943 - val_accuracy: 0.8783\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 4s 77ms/step - loss: 0.7561 - accuracy: 0.7959 - val_loss: 0.7220 - val_accuracy: 0.7739\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 4s 78ms/step - loss: 0.8344 - accuracy: 0.7669 - val_loss: 0.3469 - val_accuracy: 0.8957\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 4s 78ms/step - loss: 0.5892 - accuracy: 0.8375 - val_loss: 0.3773 - val_accuracy: 0.8957\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 4s 77ms/step - loss: 0.6503 - accuracy: 0.8162 - val_loss: 0.2816 - val_accuracy: 0.9217\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 4s 78ms/step - loss: 0.7947 - accuracy: 0.7988 - val_loss: 0.3453 - val_accuracy: 0.8957\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 4s 77ms/step - loss: 0.6878 - accuracy: 0.8317 - val_loss: 0.2901 - val_accuracy: 0.9130\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 4s 81ms/step - loss: 2.3806 - accuracy: 0.1063 - val_loss: 2.3032 - val_accuracy: 0.1043\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 4s 79ms/step - loss: 2.3039 - accuracy: 0.1111 - val_loss: 2.3010 - val_accuracy: 0.1043\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 4s 80ms/step - loss: 2.3036 - accuracy: 0.1111 - val_loss: 2.3002 - val_accuracy: 0.1043\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 4s 80ms/step - loss: 2.3035 - accuracy: 0.1111 - val_loss: 2.2996 - val_accuracy: 0.1043\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 4s 78ms/step - loss: 2.3035 - accuracy: 0.0957 - val_loss: 2.3001 - val_accuracy: 0.1043\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 4s 78ms/step - loss: 2.3031 - accuracy: 0.1005 - val_loss: 2.2993 - val_accuracy: 0.1043\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 4s 78ms/step - loss: 2.3032 - accuracy: 0.1111 - val_loss: 2.2999 - val_accuracy: 0.1043\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 4s 79ms/step - loss: 2.3034 - accuracy: 0.1014 - val_loss: 2.2995 - val_accuracy: 0.1043\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 4s 79ms/step - loss: 2.3030 - accuracy: 0.1111 - val_loss: 2.3000 - val_accuracy: 0.1043\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 4s 79ms/step - loss: 2.3027 - accuracy: 0.1111 - val_loss: 2.2996 - val_accuracy: 0.1043\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 4s 78ms/step - loss: 2.3036 - accuracy: 0.0986 - val_loss: 2.2992 - val_accuracy: 0.1043\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 4s 79ms/step - loss: 2.3032 - accuracy: 0.1111 - val_loss: 2.3013 - val_accuracy: 0.1043\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 4s 79ms/step - loss: 2.3029 - accuracy: 0.1111 - val_loss: 2.3008 - val_accuracy: 0.1043\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 4s 79ms/step - loss: 2.3025 - accuracy: 0.1111 - val_loss: 2.2994 - val_accuracy: 0.1043\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 4s 79ms/step - loss: 2.3038 - accuracy: 0.0947 - val_loss: 2.2991 - val_accuracy: 0.1043\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 4s 82ms/step - loss: 2.4622 - accuracy: 0.0899 - val_loss: 2.3035 - val_accuracy: 0.0957\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 4s 80ms/step - loss: 2.3065 - accuracy: 0.0966 - val_loss: 2.3025 - val_accuracy: 0.1304\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 4s 79ms/step - loss: 2.3049 - accuracy: 0.0821 - val_loss: 2.3049 - val_accuracy: 0.0957\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 4s 79ms/step - loss: 2.3050 - accuracy: 0.0937 - val_loss: 2.3031 - val_accuracy: 0.0957\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 4s 79ms/step - loss: 2.3047 - accuracy: 0.0889 - val_loss: 2.3039 - val_accuracy: 0.1130\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 4s 81ms/step - loss: 2.3044 - accuracy: 0.0908 - val_loss: 2.3046 - val_accuracy: 0.1130\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 4s 81ms/step - loss: 2.3045 - accuracy: 0.0899 - val_loss: 2.3041 - val_accuracy: 0.1130\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 4s 81ms/step - loss: 2.3041 - accuracy: 0.1014 - val_loss: 2.3049 - val_accuracy: 0.0957\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 4s 81ms/step - loss: 2.3043 - accuracy: 0.1053 - val_loss: 2.3066 - val_accuracy: 0.1130\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 4s 80ms/step - loss: 2.3044 - accuracy: 0.1005 - val_loss: 2.3056 - val_accuracy: 0.1130\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 4s 79ms/step - loss: 2.3047 - accuracy: 0.0947 - val_loss: 2.3044 - val_accuracy: 0.1130\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 4s 79ms/step - loss: 2.3039 - accuracy: 0.0995 - val_loss: 2.3050 - val_accuracy: 0.1130\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 4s 78ms/step - loss: 2.3058 - accuracy: 0.0841 - val_loss: 2.3035 - val_accuracy: 0.0957\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 4s 79ms/step - loss: 2.3038 - accuracy: 0.1063 - val_loss: 2.3046 - val_accuracy: 0.1130\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 4s 78ms/step - loss: 2.3047 - accuracy: 0.1063 - val_loss: 2.3045 - val_accuracy: 0.1130\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 4s 80ms/step - loss: 2.4589 - accuracy: 0.1179 - val_loss: 2.3153 - val_accuracy: 0.0870\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 4s 78ms/step - loss: 2.3064 - accuracy: 0.0928 - val_loss: 2.3045 - val_accuracy: 0.0609\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 4s 79ms/step - loss: 2.3038 - accuracy: 0.1072 - val_loss: 2.3057 - val_accuracy: 0.1043\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 4s 78ms/step - loss: 2.3033 - accuracy: 0.0976 - val_loss: 2.3087 - val_accuracy: 0.0609\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 4s 78ms/step - loss: 2.3034 - accuracy: 0.0870 - val_loss: 2.3086 - val_accuracy: 0.1043\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 4s 78ms/step - loss: 2.3040 - accuracy: 0.0995 - val_loss: 2.3112 - val_accuracy: 0.1043\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 4s 78ms/step - loss: 2.3038 - accuracy: 0.0899 - val_loss: 2.3075 - val_accuracy: 0.1043\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 4s 79ms/step - loss: 2.3037 - accuracy: 0.1053 - val_loss: 2.3139 - val_accuracy: 0.0870\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 4s 79ms/step - loss: 2.3039 - accuracy: 0.0841 - val_loss: 2.3094 - val_accuracy: 0.1043\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 4s 79ms/step - loss: 2.3037 - accuracy: 0.0976 - val_loss: 2.3106 - val_accuracy: 0.0870\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 4s 78ms/step - loss: 2.3035 - accuracy: 0.0928 - val_loss: 2.3096 - val_accuracy: 0.0609\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 4s 78ms/step - loss: 2.3034 - accuracy: 0.0986 - val_loss: 2.3107 - val_accuracy: 0.0870\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 4s 78ms/step - loss: 2.3035 - accuracy: 0.0986 - val_loss: 2.3064 - val_accuracy: 0.1043\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 4s 79ms/step - loss: 2.3039 - accuracy: 0.0986 - val_loss: 2.3063 - val_accuracy: 0.1043\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 4s 78ms/step - loss: 2.3039 - accuracy: 0.1053 - val_loss: 2.3105 - val_accuracy: 0.0870\n",
            "Epoch 1/10\n",
            "21/21 [==============================] - 16s 742ms/step - loss: 1.9675 - accuracy: 0.3395 - val_loss: 1.0945 - val_accuracy: 0.8174\n",
            "Epoch 2/10\n",
            "21/21 [==============================] - 15s 737ms/step - loss: 0.6617 - accuracy: 0.8269 - val_loss: 0.3182 - val_accuracy: 0.9304\n",
            "Epoch 3/10\n",
            "21/21 [==============================] - 16s 739ms/step - loss: 0.2829 - accuracy: 0.9168 - val_loss: 0.1797 - val_accuracy: 0.9478\n",
            "Epoch 4/10\n",
            "21/21 [==============================] - 15s 737ms/step - loss: 0.1708 - accuracy: 0.9555 - val_loss: 0.1579 - val_accuracy: 0.9478\n",
            "Epoch 5/10\n",
            "21/21 [==============================] - 15s 736ms/step - loss: 0.1268 - accuracy: 0.9623 - val_loss: 0.1148 - val_accuracy: 0.9652\n",
            "Epoch 6/10\n",
            "21/21 [==============================] - 15s 737ms/step - loss: 0.0925 - accuracy: 0.9768 - val_loss: 0.1200 - val_accuracy: 0.9652\n",
            "Epoch 7/10\n",
            "21/21 [==============================] - 15s 737ms/step - loss: 0.0710 - accuracy: 0.9749 - val_loss: 0.0736 - val_accuracy: 0.9739\n",
            "Epoch 8/10\n",
            "21/21 [==============================] - 15s 737ms/step - loss: 0.0646 - accuracy: 0.9778 - val_loss: 0.0493 - val_accuracy: 0.9913\n",
            "Epoch 9/10\n",
            "21/21 [==============================] - 15s 736ms/step - loss: 0.0401 - accuracy: 0.9913 - val_loss: 0.1047 - val_accuracy: 0.9739\n",
            "Epoch 10/10\n",
            "21/21 [==============================] - 16s 740ms/step - loss: 0.0377 - accuracy: 0.9855 - val_loss: 0.0844 - val_accuracy: 0.9913\n",
            "Epoch 1/10\n",
            "21/21 [==============================] - 16s 741ms/step - loss: 1.9283 - accuracy: 0.4178 - val_loss: 1.1497 - val_accuracy: 0.6870\n",
            "Epoch 2/10\n",
            "21/21 [==============================] - 16s 740ms/step - loss: 0.6262 - accuracy: 0.8133 - val_loss: 0.3429 - val_accuracy: 0.9043\n",
            "Epoch 3/10\n",
            "21/21 [==============================] - 15s 735ms/step - loss: 0.2710 - accuracy: 0.9110 - val_loss: 0.2024 - val_accuracy: 0.9652\n",
            "Epoch 4/10\n",
            "21/21 [==============================] - 16s 743ms/step - loss: 0.1699 - accuracy: 0.9478 - val_loss: 0.1714 - val_accuracy: 0.9739\n",
            "Epoch 5/10\n",
            "21/21 [==============================] - 16s 742ms/step - loss: 0.1155 - accuracy: 0.9662 - val_loss: 0.1453 - val_accuracy: 0.9739\n",
            "Epoch 6/10\n",
            "21/21 [==============================] - 15s 736ms/step - loss: 0.0914 - accuracy: 0.9671 - val_loss: 0.1159 - val_accuracy: 0.9739\n",
            "Epoch 7/10\n",
            "21/21 [==============================] - 16s 744ms/step - loss: 0.0726 - accuracy: 0.9778 - val_loss: 0.0923 - val_accuracy: 0.9739\n",
            "Epoch 8/10\n",
            "21/21 [==============================] - 15s 736ms/step - loss: 0.0574 - accuracy: 0.9816 - val_loss: 0.0650 - val_accuracy: 0.9826\n",
            "Epoch 9/10\n",
            "21/21 [==============================] - 16s 741ms/step - loss: 0.0503 - accuracy: 0.9874 - val_loss: 0.0730 - val_accuracy: 0.9826\n",
            "Epoch 10/10\n",
            "21/21 [==============================] - 15s 736ms/step - loss: 0.0390 - accuracy: 0.9913 - val_loss: 0.0909 - val_accuracy: 0.9826\n",
            "Epoch 1/10\n",
            "21/21 [==============================] - 16s 746ms/step - loss: 1.9523 - accuracy: 0.4280 - val_loss: 1.1014 - val_accuracy: 0.7652\n",
            "Epoch 2/10\n",
            "21/21 [==============================] - 16s 740ms/step - loss: 0.6952 - accuracy: 0.8000 - val_loss: 0.3347 - val_accuracy: 0.8957\n",
            "Epoch 3/10\n",
            "21/21 [==============================] - 16s 740ms/step - loss: 0.2598 - accuracy: 0.9188 - val_loss: 0.1792 - val_accuracy: 0.9652\n",
            "Epoch 4/10\n",
            "21/21 [==============================] - 15s 738ms/step - loss: 0.1656 - accuracy: 0.9546 - val_loss: 0.1537 - val_accuracy: 0.9739\n",
            "Epoch 5/10\n",
            "21/21 [==============================] - 16s 740ms/step - loss: 0.1213 - accuracy: 0.9749 - val_loss: 0.0921 - val_accuracy: 0.9826\n",
            "Epoch 6/10\n",
            "21/21 [==============================] - 16s 742ms/step - loss: 0.0732 - accuracy: 0.9797 - val_loss: 0.0858 - val_accuracy: 0.9913\n",
            "Epoch 7/10\n",
            "21/21 [==============================] - 16s 742ms/step - loss: 0.0671 - accuracy: 0.9807 - val_loss: 0.0583 - val_accuracy: 0.9913\n",
            "Epoch 8/10\n",
            "21/21 [==============================] - 16s 741ms/step - loss: 0.0553 - accuracy: 0.9826 - val_loss: 0.0584 - val_accuracy: 0.9913\n",
            "Epoch 9/10\n",
            "21/21 [==============================] - 16s 744ms/step - loss: 0.0480 - accuracy: 0.9836 - val_loss: 0.0458 - val_accuracy: 0.9826\n",
            "Epoch 10/10\n",
            "21/21 [==============================] - 16s 743ms/step - loss: 0.0375 - accuracy: 0.9894 - val_loss: 0.0637 - val_accuracy: 0.9826\n",
            "Epoch 1/10\n",
            "21/21 [==============================] - 16s 745ms/step - loss: 1.9131 - accuracy: 0.3942 - val_loss: 1.0182 - val_accuracy: 0.7304\n",
            "Epoch 2/10\n",
            "21/21 [==============================] - 16s 741ms/step - loss: 0.5875 - accuracy: 0.8300 - val_loss: 0.3871 - val_accuracy: 0.8957\n",
            "Epoch 3/10\n",
            "21/21 [==============================] - 16s 739ms/step - loss: 0.2635 - accuracy: 0.9246 - val_loss: 0.2461 - val_accuracy: 0.9391\n",
            "Epoch 4/10\n",
            "21/21 [==============================] - 16s 749ms/step - loss: 0.1624 - accuracy: 0.9546 - val_loss: 0.1938 - val_accuracy: 0.9478\n",
            "Epoch 5/10\n",
            "21/21 [==============================] - 16s 745ms/step - loss: 0.1089 - accuracy: 0.9691 - val_loss: 0.1369 - val_accuracy: 0.9652\n",
            "Epoch 6/10\n",
            "21/21 [==============================] - 16s 745ms/step - loss: 0.0833 - accuracy: 0.9749 - val_loss: 0.0985 - val_accuracy: 0.9826\n",
            "Epoch 7/10\n",
            "21/21 [==============================] - 16s 744ms/step - loss: 0.0540 - accuracy: 0.9855 - val_loss: 0.1045 - val_accuracy: 0.9826\n",
            "Epoch 8/10\n",
            "21/21 [==============================] - 16s 743ms/step - loss: 0.0516 - accuracy: 0.9845 - val_loss: 0.0740 - val_accuracy: 0.9739\n",
            "Epoch 9/10\n",
            "21/21 [==============================] - 16s 746ms/step - loss: 0.0484 - accuracy: 0.9826 - val_loss: 0.0825 - val_accuracy: 0.9826\n",
            "Epoch 10/10\n",
            "21/21 [==============================] - 16s 746ms/step - loss: 0.0470 - accuracy: 0.9816 - val_loss: 0.0909 - val_accuracy: 0.9826\n",
            "Epoch 1/10\n",
            "21/21 [==============================] - 16s 751ms/step - loss: 1.9468 - accuracy: 0.4077 - val_loss: 1.0214 - val_accuracy: 0.8174\n",
            "Epoch 2/10\n",
            "21/21 [==============================] - 16s 741ms/step - loss: 0.6071 - accuracy: 0.8338 - val_loss: 0.3204 - val_accuracy: 0.9304\n",
            "Epoch 3/10\n",
            "21/21 [==============================] - 16s 741ms/step - loss: 0.2811 - accuracy: 0.9130 - val_loss: 0.1262 - val_accuracy: 0.9652\n",
            "Epoch 4/10\n",
            "21/21 [==============================] - 16s 739ms/step - loss: 0.1789 - accuracy: 0.9459 - val_loss: 0.0569 - val_accuracy: 0.9913\n",
            "Epoch 5/10\n",
            "21/21 [==============================] - 16s 743ms/step - loss: 0.1131 - accuracy: 0.9623 - val_loss: 0.0408 - val_accuracy: 1.0000\n",
            "Epoch 6/10\n",
            "21/21 [==============================] - 16s 741ms/step - loss: 0.0933 - accuracy: 0.9729 - val_loss: 0.0370 - val_accuracy: 0.9913\n",
            "Epoch 7/10\n",
            "21/21 [==============================] - 16s 741ms/step - loss: 0.0667 - accuracy: 0.9797 - val_loss: 0.0162 - val_accuracy: 1.0000\n",
            "Epoch 8/10\n",
            "21/21 [==============================] - 16s 739ms/step - loss: 0.0508 - accuracy: 0.9855 - val_loss: 0.0147 - val_accuracy: 1.0000\n",
            "Epoch 9/10\n",
            "21/21 [==============================] - 16s 740ms/step - loss: 0.0381 - accuracy: 0.9913 - val_loss: 0.0179 - val_accuracy: 0.9913\n",
            "Epoch 10/10\n",
            "21/21 [==============================] - 15s 737ms/step - loss: 0.0348 - accuracy: 0.9894 - val_loss: 0.0175 - val_accuracy: 1.0000\n",
            "Epoch 1/5\n",
            "7/7 [==============================] - 1s 187ms/step - loss: 2.2895 - accuracy: 0.1983 - val_loss: 2.2582 - val_accuracy: 0.3391\n",
            "Epoch 2/5\n",
            "7/7 [==============================] - 1s 167ms/step - loss: 2.2127 - accuracy: 0.4130 - val_loss: 2.1151 - val_accuracy: 0.4696\n",
            "Epoch 3/5\n",
            "7/7 [==============================] - 1s 167ms/step - loss: 1.9915 - accuracy: 0.5919 - val_loss: 1.7547 - val_accuracy: 0.7304\n",
            "Epoch 4/5\n",
            "7/7 [==============================] - 1s 168ms/step - loss: 1.5271 - accuracy: 0.7331 - val_loss: 1.1874 - val_accuracy: 0.7913\n",
            "Epoch 5/5\n",
            "7/7 [==============================] - 1s 169ms/step - loss: 0.9724 - accuracy: 0.8104 - val_loss: 0.6400 - val_accuracy: 0.8522\n",
            "Epoch 1/5\n",
            "7/7 [==============================] - 1s 192ms/step - loss: 2.2922 - accuracy: 0.1209 - val_loss: 2.2632 - val_accuracy: 0.1913\n",
            "Epoch 2/5\n",
            "7/7 [==============================] - 1s 167ms/step - loss: 2.2218 - accuracy: 0.3559 - val_loss: 2.1282 - val_accuracy: 0.4957\n",
            "Epoch 3/5\n",
            "7/7 [==============================] - 1s 176ms/step - loss: 2.0121 - accuracy: 0.5532 - val_loss: 1.7775 - val_accuracy: 0.7130\n",
            "Epoch 4/5\n",
            "7/7 [==============================] - 1s 170ms/step - loss: 1.5471 - accuracy: 0.7060 - val_loss: 1.1599 - val_accuracy: 0.8261\n",
            "Epoch 5/5\n",
            "7/7 [==============================] - 1s 176ms/step - loss: 1.0018 - accuracy: 0.7863 - val_loss: 0.7161 - val_accuracy: 0.8000\n",
            "Epoch 1/5\n",
            "7/7 [==============================] - 1s 184ms/step - loss: 2.2902 - accuracy: 0.1469 - val_loss: 2.2507 - val_accuracy: 0.2348\n",
            "Epoch 2/5\n",
            "7/7 [==============================] - 1s 176ms/step - loss: 2.2138 - accuracy: 0.2425 - val_loss: 2.0981 - val_accuracy: 0.4870\n",
            "Epoch 3/5\n",
            "7/7 [==============================] - 1s 171ms/step - loss: 1.9848 - accuracy: 0.6386 - val_loss: 1.7106 - val_accuracy: 0.7478\n",
            "Epoch 4/5\n",
            "7/7 [==============================] - 1s 176ms/step - loss: 1.5220 - accuracy: 0.7063 - val_loss: 1.1520 - val_accuracy: 0.7739\n",
            "Epoch 5/5\n",
            "7/7 [==============================] - 1s 168ms/step - loss: 0.9882 - accuracy: 0.7758 - val_loss: 0.6620 - val_accuracy: 0.8348\n",
            "Epoch 1/5\n",
            "7/7 [==============================] - 1s 191ms/step - loss: 2.2926 - accuracy: 0.1816 - val_loss: 2.2654 - val_accuracy: 0.4261\n",
            "Epoch 2/5\n",
            "7/7 [==============================] - 1s 170ms/step - loss: 2.2222 - accuracy: 0.5159 - val_loss: 2.1352 - val_accuracy: 0.6087\n",
            "Epoch 3/5\n",
            "7/7 [==============================] - 1s 178ms/step - loss: 2.0020 - accuracy: 0.6135 - val_loss: 1.7714 - val_accuracy: 0.7565\n",
            "Epoch 4/5\n",
            "7/7 [==============================] - 1s 173ms/step - loss: 1.5426 - accuracy: 0.7198 - val_loss: 1.1797 - val_accuracy: 0.8174\n",
            "Epoch 5/5\n",
            "7/7 [==============================] - 1s 180ms/step - loss: 0.9853 - accuracy: 0.7787 - val_loss: 0.6653 - val_accuracy: 0.8261\n",
            "WARNING:tensorflow:5 out of the last 13 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd0e3c4400> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/5\n",
            "7/7 [==============================] - 1s 192ms/step - loss: 2.2900 - accuracy: 0.2000 - val_loss: 2.2546 - val_accuracy: 0.3043\n",
            "Epoch 2/5\n",
            "7/7 [==============================] - 1s 181ms/step - loss: 2.2085 - accuracy: 0.3923 - val_loss: 2.1038 - val_accuracy: 0.4696\n",
            "Epoch 3/5\n",
            "7/7 [==============================] - 1s 174ms/step - loss: 1.9698 - accuracy: 0.5275 - val_loss: 1.7347 - val_accuracy: 0.7130\n",
            "Epoch 4/5\n",
            "7/7 [==============================] - 1s 181ms/step - loss: 1.4965 - accuracy: 0.7266 - val_loss: 1.1797 - val_accuracy: 0.7739\n",
            "Epoch 5/5\n",
            "7/7 [==============================] - 1s 178ms/step - loss: 0.9701 - accuracy: 0.7585 - val_loss: 0.7528 - val_accuracy: 0.8435\n",
            "WARNING:tensorflow:6 out of the last 15 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd0e32fea0> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/5\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 2.2963 - accuracy: 0.1731 - val_loss: 2.2807 - val_accuracy: 0.3826\n",
            "Epoch 2/5\n",
            "21/21 [==============================] - 0s 20ms/step - loss: 2.2516 - accuracy: 0.3279 - val_loss: 2.1892 - val_accuracy: 0.4087\n",
            "Epoch 3/5\n",
            "21/21 [==============================] - 0s 21ms/step - loss: 2.0909 - accuracy: 0.4120 - val_loss: 1.8662 - val_accuracy: 0.6348\n",
            "Epoch 4/5\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 1.6310 - accuracy: 0.5619 - val_loss: 1.2413 - val_accuracy: 0.7043\n",
            "Epoch 5/5\n",
            "21/21 [==============================] - 0s 21ms/step - loss: 1.1226 - accuracy: 0.6779 - val_loss: 0.7892 - val_accuracy: 0.7826\n",
            "WARNING:tensorflow:6 out of the last 11 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd0e59ed08> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/5\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 2.2925 - accuracy: 0.1760 - val_loss: 2.2685 - val_accuracy: 0.4435\n",
            "Epoch 2/5\n",
            "21/21 [==============================] - 0s 21ms/step - loss: 2.2301 - accuracy: 0.3221 - val_loss: 2.1347 - val_accuracy: 0.4957\n",
            "Epoch 3/5\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 2.0020 - accuracy: 0.4652 - val_loss: 1.7230 - val_accuracy: 0.7043\n",
            "Epoch 4/5\n",
            "21/21 [==============================] - 0s 22ms/step - loss: 1.5076 - accuracy: 0.6093 - val_loss: 1.0920 - val_accuracy: 0.7826\n",
            "Epoch 5/5\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 1.0454 - accuracy: 0.7099 - val_loss: 0.7026 - val_accuracy: 0.8696\n",
            "Epoch 1/5\n",
            "21/21 [==============================] - 1s 26ms/step - loss: 2.2965 - accuracy: 0.1449 - val_loss: 2.2804 - val_accuracy: 0.1217\n",
            "Epoch 2/5\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 2.2652 - accuracy: 0.2599 - val_loss: 2.2156 - val_accuracy: 0.2870\n",
            "Epoch 3/5\n",
            "21/21 [==============================] - 0s 21ms/step - loss: 2.1454 - accuracy: 0.3855 - val_loss: 1.9763 - val_accuracy: 0.6435\n",
            "Epoch 4/5\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 1.7743 - accuracy: 0.5411 - val_loss: 1.4322 - val_accuracy: 0.6435\n",
            "Epoch 5/5\n",
            "21/21 [==============================] - 0s 21ms/step - loss: 1.2426 - accuracy: 0.6464 - val_loss: 0.8734 - val_accuracy: 0.7826\n",
            "Epoch 1/5\n",
            "21/21 [==============================] - 1s 24ms/step - loss: 2.2924 - accuracy: 0.1884 - val_loss: 2.2707 - val_accuracy: 0.2783\n",
            "Epoch 2/5\n",
            "21/21 [==============================] - 0s 21ms/step - loss: 2.2413 - accuracy: 0.3333 - val_loss: 2.1641 - val_accuracy: 0.6261\n",
            "Epoch 3/5\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 2.0363 - accuracy: 0.4705 - val_loss: 1.7940 - val_accuracy: 0.6261\n",
            "Epoch 4/5\n",
            "21/21 [==============================] - 0s 20ms/step - loss: 1.5792 - accuracy: 0.5932 - val_loss: 1.1410 - val_accuracy: 0.8261\n",
            "Epoch 5/5\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 1.0416 - accuracy: 0.7082 - val_loss: 0.6899 - val_accuracy: 0.8522\n",
            "Epoch 1/5\n",
            "21/21 [==============================] - 1s 26ms/step - loss: 2.2927 - accuracy: 0.1401 - val_loss: 2.2720 - val_accuracy: 0.1478\n",
            "Epoch 2/5\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 2.2266 - accuracy: 0.2657 - val_loss: 2.1551 - val_accuracy: 0.2696\n",
            "Epoch 3/5\n",
            "21/21 [==============================] - 0s 21ms/step - loss: 1.9680 - accuracy: 0.4435 - val_loss: 1.6949 - val_accuracy: 0.7043\n",
            "Epoch 4/5\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 1.4193 - accuracy: 0.6261 - val_loss: 1.0154 - val_accuracy: 0.8000\n",
            "Epoch 5/5\n",
            "21/21 [==============================] - 0s 22ms/step - loss: 0.9238 - accuracy: 0.7507 - val_loss: 0.6510 - val_accuracy: 0.9130\n",
            "Epoch 1/15\n",
            "7/7 [==============================] - 0s 25ms/step - loss: 2.3017 - accuracy: 0.1151 - val_loss: 2.2992 - val_accuracy: 0.1913\n",
            "Epoch 2/15\n",
            "7/7 [==============================] - 0s 8ms/step - loss: 2.2981 - accuracy: 0.1373 - val_loss: 2.2960 - val_accuracy: 0.1043\n",
            "Epoch 3/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 2.2935 - accuracy: 0.1460 - val_loss: 2.2900 - val_accuracy: 0.1130\n",
            "Epoch 4/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 2.2852 - accuracy: 0.1509 - val_loss: 2.2811 - val_accuracy: 0.1217\n",
            "Epoch 5/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 2.2754 - accuracy: 0.1847 - val_loss: 2.2670 - val_accuracy: 0.1913\n",
            "Epoch 6/15\n",
            "7/7 [==============================] - 0s 12ms/step - loss: 2.2573 - accuracy: 0.2099 - val_loss: 2.2444 - val_accuracy: 0.2174\n",
            "Epoch 7/15\n",
            "7/7 [==============================] - 0s 12ms/step - loss: 2.2360 - accuracy: 0.2195 - val_loss: 2.2120 - val_accuracy: 0.2783\n",
            "Epoch 8/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 2.1997 - accuracy: 0.2544 - val_loss: 2.1694 - val_accuracy: 0.2783\n",
            "Epoch 9/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 2.1598 - accuracy: 0.2128 - val_loss: 2.1120 - val_accuracy: 0.2957\n",
            "Epoch 10/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 2.1030 - accuracy: 0.2756 - val_loss: 2.0410 - val_accuracy: 0.3304\n",
            "Epoch 11/15\n",
            "7/7 [==============================] - 0s 8ms/step - loss: 2.0252 - accuracy: 0.2921 - val_loss: 1.9606 - val_accuracy: 0.3478\n",
            "Epoch 12/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 1.9600 - accuracy: 0.3104 - val_loss: 1.8712 - val_accuracy: 0.4000\n",
            "Epoch 13/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 1.8788 - accuracy: 0.3366 - val_loss: 1.7772 - val_accuracy: 0.4348\n",
            "Epoch 14/15\n",
            "7/7 [==============================] - 0s 11ms/step - loss: 1.8215 - accuracy: 0.3607 - val_loss: 1.6823 - val_accuracy: 0.4783\n",
            "Epoch 15/15\n",
            "7/7 [==============================] - 0s 13ms/step - loss: 1.7669 - accuracy: 0.3569 - val_loss: 1.6007 - val_accuracy: 0.5565\n",
            "Epoch 1/15\n",
            "7/7 [==============================] - 0s 26ms/step - loss: 2.3014 - accuracy: 0.1315 - val_loss: 2.2990 - val_accuracy: 0.1565\n",
            "Epoch 2/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 2.2972 - accuracy: 0.1644 - val_loss: 2.2946 - val_accuracy: 0.1304\n",
            "Epoch 3/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 2.2913 - accuracy: 0.1876 - val_loss: 2.2873 - val_accuracy: 0.2261\n",
            "Epoch 4/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 2.2821 - accuracy: 0.2263 - val_loss: 2.2755 - val_accuracy: 0.2696\n",
            "Epoch 5/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 2.2695 - accuracy: 0.2621 - val_loss: 2.2587 - val_accuracy: 0.2870\n",
            "Epoch 6/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 2.2503 - accuracy: 0.2747 - val_loss: 2.2323 - val_accuracy: 0.4000\n",
            "Epoch 7/15\n",
            "7/7 [==============================] - 0s 10ms/step - loss: 2.2189 - accuracy: 0.3075 - val_loss: 2.1925 - val_accuracy: 0.4696\n",
            "Epoch 8/15\n",
            "7/7 [==============================] - 0s 12ms/step - loss: 2.1777 - accuracy: 0.3191 - val_loss: 2.1364 - val_accuracy: 0.4870\n",
            "Epoch 9/15\n",
            "7/7 [==============================] - 0s 10ms/step - loss: 2.1139 - accuracy: 0.3636 - val_loss: 2.0587 - val_accuracy: 0.4870\n",
            "Epoch 10/15\n",
            "7/7 [==============================] - 0s 11ms/step - loss: 2.0480 - accuracy: 0.3530 - val_loss: 1.9600 - val_accuracy: 0.5304\n",
            "Epoch 11/15\n",
            "7/7 [==============================] - 0s 10ms/step - loss: 1.9445 - accuracy: 0.3723 - val_loss: 1.8483 - val_accuracy: 0.5913\n",
            "Epoch 12/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 1.8438 - accuracy: 0.4014 - val_loss: 1.7252 - val_accuracy: 0.6087\n",
            "Epoch 13/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 1.7701 - accuracy: 0.3839 - val_loss: 1.6138 - val_accuracy: 0.6261\n",
            "Epoch 14/15\n",
            "7/7 [==============================] - 0s 12ms/step - loss: 1.6714 - accuracy: 0.4342 - val_loss: 1.5064 - val_accuracy: 0.6522\n",
            "Epoch 15/15\n",
            "7/7 [==============================] - 0s 11ms/step - loss: 1.5875 - accuracy: 0.4333 - val_loss: 1.4193 - val_accuracy: 0.6696\n",
            "Epoch 1/15\n",
            "7/7 [==============================] - 0s 25ms/step - loss: 2.3013 - accuracy: 0.1546 - val_loss: 2.2980 - val_accuracy: 0.2261\n",
            "Epoch 2/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 2.2972 - accuracy: 0.2048 - val_loss: 2.2930 - val_accuracy: 0.2870\n",
            "Epoch 3/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 2.2920 - accuracy: 0.2415 - val_loss: 2.2860 - val_accuracy: 0.3652\n",
            "Epoch 4/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 2.2845 - accuracy: 0.2860 - val_loss: 2.2760 - val_accuracy: 0.3913\n",
            "Epoch 5/15\n",
            "7/7 [==============================] - 0s 14ms/step - loss: 2.2725 - accuracy: 0.3014 - val_loss: 2.2592 - val_accuracy: 0.4174\n",
            "Epoch 6/15\n",
            "7/7 [==============================] - 0s 10ms/step - loss: 2.2549 - accuracy: 0.3014 - val_loss: 2.2350 - val_accuracy: 0.4783\n",
            "Epoch 7/15\n",
            "7/7 [==============================] - 0s 10ms/step - loss: 2.2244 - accuracy: 0.3295 - val_loss: 2.1963 - val_accuracy: 0.4870\n",
            "Epoch 8/15\n",
            "7/7 [==============================] - 0s 12ms/step - loss: 2.1904 - accuracy: 0.3372 - val_loss: 2.1431 - val_accuracy: 0.4696\n",
            "Epoch 9/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 2.1372 - accuracy: 0.3546 - val_loss: 2.0702 - val_accuracy: 0.4957\n",
            "Epoch 10/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 2.0707 - accuracy: 0.3691 - val_loss: 1.9777 - val_accuracy: 0.5304\n",
            "Epoch 11/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 1.9805 - accuracy: 0.3681 - val_loss: 1.8651 - val_accuracy: 0.5217\n",
            "Epoch 12/15\n",
            "7/7 [==============================] - 0s 13ms/step - loss: 1.8989 - accuracy: 0.3845 - val_loss: 1.7422 - val_accuracy: 0.5217\n",
            "Epoch 13/15\n",
            "7/7 [==============================] - 0s 10ms/step - loss: 1.7972 - accuracy: 0.4010 - val_loss: 1.6178 - val_accuracy: 0.5391\n",
            "Epoch 14/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 1.7128 - accuracy: 0.4068 - val_loss: 1.5044 - val_accuracy: 0.6000\n",
            "Epoch 15/15\n",
            "7/7 [==============================] - 0s 11ms/step - loss: 1.6199 - accuracy: 0.4589 - val_loss: 1.3966 - val_accuracy: 0.6435\n",
            "Epoch 1/15\n",
            "7/7 [==============================] - 0s 26ms/step - loss: 2.3020 - accuracy: 0.1198 - val_loss: 2.2996 - val_accuracy: 0.2696\n",
            "Epoch 2/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 2.2987 - accuracy: 0.2135 - val_loss: 2.2956 - val_accuracy: 0.2522\n",
            "Epoch 3/15\n",
            "7/7 [==============================] - 0s 14ms/step - loss: 2.2935 - accuracy: 0.2338 - val_loss: 2.2891 - val_accuracy: 0.2609\n",
            "Epoch 4/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 2.2868 - accuracy: 0.2473 - val_loss: 2.2797 - val_accuracy: 0.3739\n",
            "Epoch 5/15\n",
            "7/7 [==============================] - 0s 10ms/step - loss: 2.2739 - accuracy: 0.2879 - val_loss: 2.2633 - val_accuracy: 0.4000\n",
            "Epoch 6/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 2.2559 - accuracy: 0.3121 - val_loss: 2.2371 - val_accuracy: 0.4435\n",
            "Epoch 7/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 2.2298 - accuracy: 0.3159 - val_loss: 2.1983 - val_accuracy: 0.5130\n",
            "Epoch 8/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 2.1922 - accuracy: 0.3150 - val_loss: 2.1422 - val_accuracy: 0.5217\n",
            "Epoch 9/15\n",
            "7/7 [==============================] - 0s 12ms/step - loss: 2.1321 - accuracy: 0.3382 - val_loss: 2.0651 - val_accuracy: 0.5130\n",
            "Epoch 10/15\n",
            "7/7 [==============================] - 0s 14ms/step - loss: 2.0708 - accuracy: 0.3246 - val_loss: 1.9687 - val_accuracy: 0.5304\n",
            "Epoch 11/15\n",
            "7/7 [==============================] - 0s 10ms/step - loss: 1.9858 - accuracy: 0.3720 - val_loss: 1.8581 - val_accuracy: 0.6174\n",
            "Epoch 12/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 1.8844 - accuracy: 0.4000 - val_loss: 1.7394 - val_accuracy: 0.6174\n",
            "Epoch 13/15\n",
            "7/7 [==============================] - 0s 11ms/step - loss: 1.7923 - accuracy: 0.3855 - val_loss: 1.6265 - val_accuracy: 0.6174\n",
            "Epoch 14/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 1.7330 - accuracy: 0.3894 - val_loss: 1.5353 - val_accuracy: 0.6174\n",
            "Epoch 15/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 1.6337 - accuracy: 0.4164 - val_loss: 1.4506 - val_accuracy: 0.6348\n",
            "WARNING:tensorflow:5 out of the last 13 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd19541b70> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/15\n",
            "7/7 [==============================] - 0s 33ms/step - loss: 2.3010 - accuracy: 0.1314 - val_loss: 2.2985 - val_accuracy: 0.2000\n",
            "Epoch 2/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 2.2962 - accuracy: 0.1739 - val_loss: 2.2968 - val_accuracy: 0.2000\n",
            "Epoch 3/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 2.2914 - accuracy: 0.1691 - val_loss: 2.2916 - val_accuracy: 0.2000\n",
            "Epoch 4/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 2.2836 - accuracy: 0.2280 - val_loss: 2.2839 - val_accuracy: 0.2348\n",
            "Epoch 5/15\n",
            "7/7 [==============================] - 0s 10ms/step - loss: 2.2736 - accuracy: 0.2203 - val_loss: 2.2708 - val_accuracy: 0.2261\n",
            "Epoch 6/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 2.2575 - accuracy: 0.2415 - val_loss: 2.2547 - val_accuracy: 0.2609\n",
            "Epoch 7/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 2.2323 - accuracy: 0.2628 - val_loss: 2.2304 - val_accuracy: 0.2435\n",
            "Epoch 8/15\n",
            "7/7 [==============================] - 0s 10ms/step - loss: 2.2020 - accuracy: 0.2493 - val_loss: 2.1933 - val_accuracy: 0.2957\n",
            "Epoch 9/15\n",
            "7/7 [==============================] - 0s 12ms/step - loss: 2.1593 - accuracy: 0.2783 - val_loss: 2.1446 - val_accuracy: 0.3130\n",
            "Epoch 10/15\n",
            "7/7 [==============================] - 0s 10ms/step - loss: 2.0995 - accuracy: 0.3034 - val_loss: 2.0866 - val_accuracy: 0.2957\n",
            "Epoch 11/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 2.0411 - accuracy: 0.3101 - val_loss: 2.0071 - val_accuracy: 0.3043\n",
            "Epoch 12/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 1.9610 - accuracy: 0.3024 - val_loss: 1.9169 - val_accuracy: 0.3391\n",
            "Epoch 13/15\n",
            "7/7 [==============================] - 0s 12ms/step - loss: 1.8860 - accuracy: 0.3430 - val_loss: 1.8283 - val_accuracy: 0.3826\n",
            "Epoch 14/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 1.7901 - accuracy: 0.3681 - val_loss: 1.7245 - val_accuracy: 0.4522\n",
            "Epoch 15/15\n",
            "7/7 [==============================] - 0s 9ms/step - loss: 1.7102 - accuracy: 0.3961 - val_loss: 1.6280 - val_accuracy: 0.5130\n",
            "WARNING:tensorflow:6 out of the last 15 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd0d28bbf8> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/10\n",
            "21/21 [==============================] - 0s 16ms/step - loss: 2.2968 - accuracy: 0.1238 - val_loss: 2.2873 - val_accuracy: 0.2609\n",
            "Epoch 2/10\n",
            "21/21 [==============================] - 0s 10ms/step - loss: 2.2670 - accuracy: 0.2505 - val_loss: 2.2353 - val_accuracy: 0.3478\n",
            "Epoch 3/10\n",
            "21/21 [==============================] - 0s 10ms/step - loss: 2.1745 - accuracy: 0.3153 - val_loss: 2.0557 - val_accuracy: 0.4609\n",
            "Epoch 4/10\n",
            "21/21 [==============================] - 0s 13ms/step - loss: 1.8916 - accuracy: 0.4787 - val_loss: 1.5787 - val_accuracy: 0.6609\n",
            "Epoch 5/10\n",
            "21/21 [==============================] - 0s 10ms/step - loss: 1.4505 - accuracy: 0.5793 - val_loss: 1.0526 - val_accuracy: 0.8261\n",
            "Epoch 6/10\n",
            "21/21 [==============================] - 0s 10ms/step - loss: 1.0486 - accuracy: 0.7060 - val_loss: 0.7579 - val_accuracy: 0.7826\n",
            "Epoch 7/10\n",
            "21/21 [==============================] - 0s 13ms/step - loss: 0.8403 - accuracy: 0.7447 - val_loss: 0.6046 - val_accuracy: 0.8174\n",
            "Epoch 8/10\n",
            "21/21 [==============================] - 0s 10ms/step - loss: 0.7204 - accuracy: 0.7834 - val_loss: 0.5252 - val_accuracy: 0.8609\n",
            "Epoch 9/10\n",
            "21/21 [==============================] - 0s 10ms/step - loss: 0.6573 - accuracy: 0.7959 - val_loss: 0.4480 - val_accuracy: 0.8783\n",
            "Epoch 10/10\n",
            "21/21 [==============================] - 0s 12ms/step - loss: 0.5641 - accuracy: 0.8221 - val_loss: 0.3951 - val_accuracy: 0.8957\n",
            "WARNING:tensorflow:6 out of the last 11 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd0a833598> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/10\n",
            "21/21 [==============================] - 0s 16ms/step - loss: 2.2972 - accuracy: 0.1238 - val_loss: 2.2862 - val_accuracy: 0.2000\n",
            "Epoch 2/10\n",
            "21/21 [==============================] - 0s 11ms/step - loss: 2.2675 - accuracy: 0.2070 - val_loss: 2.2235 - val_accuracy: 0.3217\n",
            "Epoch 3/10\n",
            "21/21 [==============================] - 0s 11ms/step - loss: 2.1517 - accuracy: 0.3250 - val_loss: 2.0020 - val_accuracy: 0.5739\n",
            "Epoch 4/10\n",
            "21/21 [==============================] - 0s 10ms/step - loss: 1.8301 - accuracy: 0.5348 - val_loss: 1.4746 - val_accuracy: 0.6696\n",
            "Epoch 5/10\n",
            "21/21 [==============================] - 0s 11ms/step - loss: 1.3199 - accuracy: 0.6567 - val_loss: 0.9597 - val_accuracy: 0.8087\n",
            "Epoch 6/10\n",
            "21/21 [==============================] - 0s 12ms/step - loss: 0.9559 - accuracy: 0.7398 - val_loss: 0.6648 - val_accuracy: 0.8261\n",
            "Epoch 7/10\n",
            "21/21 [==============================] - 0s 11ms/step - loss: 0.7749 - accuracy: 0.7515 - val_loss: 0.5589 - val_accuracy: 0.8783\n",
            "Epoch 8/10\n",
            "21/21 [==============================] - 0s 11ms/step - loss: 0.6522 - accuracy: 0.7863 - val_loss: 0.4736 - val_accuracy: 0.8870\n",
            "Epoch 9/10\n",
            "21/21 [==============================] - 0s 12ms/step - loss: 0.5867 - accuracy: 0.8250 - val_loss: 0.4116 - val_accuracy: 0.8957\n",
            "Epoch 10/10\n",
            "21/21 [==============================] - 0s 10ms/step - loss: 0.5162 - accuracy: 0.8404 - val_loss: 0.3491 - val_accuracy: 0.9043\n",
            "Epoch 1/10\n",
            "21/21 [==============================] - 0s 16ms/step - loss: 2.2948 - accuracy: 0.1420 - val_loss: 2.2749 - val_accuracy: 0.1217\n",
            "Epoch 2/10\n",
            "21/21 [==============================] - 0s 13ms/step - loss: 2.2538 - accuracy: 0.2309 - val_loss: 2.1901 - val_accuracy: 0.3565\n",
            "Epoch 3/10\n",
            "21/21 [==============================] - 0s 10ms/step - loss: 2.1132 - accuracy: 0.3671 - val_loss: 1.9147 - val_accuracy: 0.5652\n",
            "Epoch 4/10\n",
            "21/21 [==============================] - 0s 10ms/step - loss: 1.7445 - accuracy: 0.5391 - val_loss: 1.3843 - val_accuracy: 0.7043\n",
            "Epoch 5/10\n",
            "21/21 [==============================] - 0s 12ms/step - loss: 1.3229 - accuracy: 0.6068 - val_loss: 0.9893 - val_accuracy: 0.7739\n",
            "Epoch 6/10\n",
            "21/21 [==============================] - 0s 11ms/step - loss: 1.0296 - accuracy: 0.6966 - val_loss: 0.7437 - val_accuracy: 0.8174\n",
            "Epoch 7/10\n",
            "21/21 [==============================] - 0s 13ms/step - loss: 0.8333 - accuracy: 0.7623 - val_loss: 0.5881 - val_accuracy: 0.8696\n",
            "Epoch 8/10\n",
            "21/21 [==============================] - 0s 11ms/step - loss: 0.7224 - accuracy: 0.7884 - val_loss: 0.5040 - val_accuracy: 0.8609\n",
            "Epoch 9/10\n",
            "21/21 [==============================] - 0s 10ms/step - loss: 0.6521 - accuracy: 0.8106 - val_loss: 0.4588 - val_accuracy: 0.8696\n",
            "Epoch 10/10\n",
            "21/21 [==============================] - 0s 14ms/step - loss: 0.5814 - accuracy: 0.8184 - val_loss: 0.4315 - val_accuracy: 0.8696\n",
            "Epoch 1/10\n",
            "21/21 [==============================] - 0s 16ms/step - loss: 2.2949 - accuracy: 0.1990 - val_loss: 2.2797 - val_accuracy: 0.3043\n",
            "Epoch 2/10\n",
            "21/21 [==============================] - 0s 10ms/step - loss: 2.2541 - accuracy: 0.2976 - val_loss: 2.1901 - val_accuracy: 0.3652\n",
            "Epoch 3/10\n",
            "21/21 [==============================] - 0s 12ms/step - loss: 2.0884 - accuracy: 0.3710 - val_loss: 1.8817 - val_accuracy: 0.4435\n",
            "Epoch 4/10\n",
            "21/21 [==============================] - 0s 10ms/step - loss: 1.7017 - accuracy: 0.5188 - val_loss: 1.3845 - val_accuracy: 0.6609\n",
            "Epoch 5/10\n",
            "21/21 [==============================] - 0s 11ms/step - loss: 1.3504 - accuracy: 0.6077 - val_loss: 1.0358 - val_accuracy: 0.7913\n",
            "Epoch 6/10\n",
            "21/21 [==============================] - 0s 11ms/step - loss: 1.0351 - accuracy: 0.7072 - val_loss: 0.7414 - val_accuracy: 0.8261\n",
            "Epoch 7/10\n",
            "21/21 [==============================] - 0s 10ms/step - loss: 0.8502 - accuracy: 0.7324 - val_loss: 0.5990 - val_accuracy: 0.8435\n",
            "Epoch 8/10\n",
            "21/21 [==============================] - 0s 12ms/step - loss: 0.7535 - accuracy: 0.7643 - val_loss: 0.5007 - val_accuracy: 0.8870\n",
            "Epoch 9/10\n",
            "21/21 [==============================] - 0s 13ms/step - loss: 0.6413 - accuracy: 0.8058 - val_loss: 0.4637 - val_accuracy: 0.8783\n",
            "Epoch 10/10\n",
            "21/21 [==============================] - 0s 10ms/step - loss: 0.5875 - accuracy: 0.8164 - val_loss: 0.3966 - val_accuracy: 0.8870\n",
            "Epoch 1/10\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 2.2950 - accuracy: 0.1498 - val_loss: 2.2812 - val_accuracy: 0.3217\n",
            "Epoch 2/10\n",
            "21/21 [==============================] - 0s 12ms/step - loss: 2.2538 - accuracy: 0.2802 - val_loss: 2.1960 - val_accuracy: 0.3304\n",
            "Epoch 3/10\n",
            "21/21 [==============================] - 0s 10ms/step - loss: 2.1141 - accuracy: 0.3478 - val_loss: 1.9531 - val_accuracy: 0.4522\n",
            "Epoch 4/10\n",
            "21/21 [==============================] - 0s 11ms/step - loss: 1.7662 - accuracy: 0.5005 - val_loss: 1.4872 - val_accuracy: 0.7478\n",
            "Epoch 5/10\n",
            "21/21 [==============================] - 0s 13ms/step - loss: 1.2956 - accuracy: 0.6454 - val_loss: 0.9997 - val_accuracy: 0.8087\n",
            "Epoch 6/10\n",
            "21/21 [==============================] - 0s 11ms/step - loss: 0.9756 - accuracy: 0.7092 - val_loss: 0.7505 - val_accuracy: 0.8087\n",
            "Epoch 7/10\n",
            "21/21 [==============================] - 0s 10ms/step - loss: 0.7749 - accuracy: 0.7710 - val_loss: 0.5895 - val_accuracy: 0.8783\n",
            "Epoch 8/10\n",
            "21/21 [==============================] - 0s 12ms/step - loss: 0.6639 - accuracy: 0.7932 - val_loss: 0.5202 - val_accuracy: 0.9217\n",
            "Epoch 9/10\n",
            "21/21 [==============================] - 0s 11ms/step - loss: 0.6030 - accuracy: 0.8164 - val_loss: 0.4162 - val_accuracy: 0.9217\n",
            "Epoch 10/10\n",
            "21/21 [==============================] - 0s 10ms/step - loss: 0.5179 - accuracy: 0.8483 - val_loss: 0.3751 - val_accuracy: 0.9391\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 2s 43ms/step - loss: 2.2436 - accuracy: 0.1954 - val_loss: 1.9690 - val_accuracy: 0.6000\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 2s 41ms/step - loss: 1.4418 - accuracy: 0.5397 - val_loss: 0.6898 - val_accuracy: 0.8348\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 2s 41ms/step - loss: 0.7315 - accuracy: 0.7660 - val_loss: 0.3891 - val_accuracy: 0.9043\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 0.5083 - accuracy: 0.8443 - val_loss: 0.2491 - val_accuracy: 0.9304\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 2s 40ms/step - loss: 0.3792 - accuracy: 0.8820 - val_loss: 0.1868 - val_accuracy: 0.9391\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 2s 41ms/step - loss: 0.3125 - accuracy: 0.9004 - val_loss: 0.1752 - val_accuracy: 0.9565\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 0.2772 - accuracy: 0.9120 - val_loss: 0.1289 - val_accuracy: 0.9652\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 2s 41ms/step - loss: 0.2472 - accuracy: 0.9178 - val_loss: 0.1120 - val_accuracy: 0.9739\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 2s 41ms/step - loss: 0.2132 - accuracy: 0.9342 - val_loss: 0.1162 - val_accuracy: 0.9652\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 2s 41ms/step - loss: 0.2234 - accuracy: 0.9284 - val_loss: 0.0971 - val_accuracy: 0.9826\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 0.1561 - accuracy: 0.9526 - val_loss: 0.0963 - val_accuracy: 0.9652\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 2s 41ms/step - loss: 0.1684 - accuracy: 0.9497 - val_loss: 0.1009 - val_accuracy: 0.9652\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 2s 40ms/step - loss: 0.1724 - accuracy: 0.9468 - val_loss: 0.1023 - val_accuracy: 0.9652\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 2s 40ms/step - loss: 0.1517 - accuracy: 0.9516 - val_loss: 0.0921 - val_accuracy: 0.9739\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 2s 41ms/step - loss: 0.1340 - accuracy: 0.9565 - val_loss: 0.0864 - val_accuracy: 0.9652\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 2s 43ms/step - loss: 2.2603 - accuracy: 0.1818 - val_loss: 2.0737 - val_accuracy: 0.4348\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 2s 41ms/step - loss: 1.5225 - accuracy: 0.5300 - val_loss: 0.7517 - val_accuracy: 0.8348\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 2s 41ms/step - loss: 0.7551 - accuracy: 0.7553 - val_loss: 0.4007 - val_accuracy: 0.9130\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 2s 40ms/step - loss: 0.5060 - accuracy: 0.8424 - val_loss: 0.2714 - val_accuracy: 0.9391\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 2s 40ms/step - loss: 0.3952 - accuracy: 0.8849 - val_loss: 0.2407 - val_accuracy: 0.9478\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 2s 40ms/step - loss: 0.3094 - accuracy: 0.9043 - val_loss: 0.1785 - val_accuracy: 0.9565\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 2s 41ms/step - loss: 0.2780 - accuracy: 0.9168 - val_loss: 0.1633 - val_accuracy: 0.9652\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 2s 44ms/step - loss: 0.2201 - accuracy: 0.9371 - val_loss: 0.1522 - val_accuracy: 0.9739\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 2s 44ms/step - loss: 0.2269 - accuracy: 0.9265 - val_loss: 0.1454 - val_accuracy: 0.9739\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 2s 40ms/step - loss: 0.1980 - accuracy: 0.9313 - val_loss: 0.1528 - val_accuracy: 0.9652\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 2s 40ms/step - loss: 0.1849 - accuracy: 0.9429 - val_loss: 0.1262 - val_accuracy: 0.9652\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 0.1598 - accuracy: 0.9468 - val_loss: 0.1037 - val_accuracy: 0.9739\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 2s 41ms/step - loss: 0.1629 - accuracy: 0.9400 - val_loss: 0.1006 - val_accuracy: 0.9826\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 2s 40ms/step - loss: 0.1398 - accuracy: 0.9613 - val_loss: 0.1015 - val_accuracy: 0.9826\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 2s 41ms/step - loss: 0.1076 - accuracy: 0.9671 - val_loss: 0.0819 - val_accuracy: 0.9826\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 2s 45ms/step - loss: 2.2123 - accuracy: 0.2512 - val_loss: 1.8162 - val_accuracy: 0.5913\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 1.2405 - accuracy: 0.6193 - val_loss: 0.5774 - val_accuracy: 0.8174\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 2s 41ms/step - loss: 0.6712 - accuracy: 0.7855 - val_loss: 0.3598 - val_accuracy: 0.8870\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 0.4880 - accuracy: 0.8415 - val_loss: 0.2799 - val_accuracy: 0.9391\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 0.4266 - accuracy: 0.8609 - val_loss: 0.2285 - val_accuracy: 0.9565\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 0.2923 - accuracy: 0.9101 - val_loss: 0.1714 - val_accuracy: 0.9652\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 0.2961 - accuracy: 0.9121 - val_loss: 0.1791 - val_accuracy: 0.9391\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 0.2564 - accuracy: 0.9150 - val_loss: 0.1807 - val_accuracy: 0.9565\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 0.2113 - accuracy: 0.9343 - val_loss: 0.1434 - val_accuracy: 0.9739\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 2s 41ms/step - loss: 0.1976 - accuracy: 0.9372 - val_loss: 0.1088 - val_accuracy: 0.9739\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 2s 41ms/step - loss: 0.1880 - accuracy: 0.9449 - val_loss: 0.1038 - val_accuracy: 0.9652\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 2s 41ms/step - loss: 0.1706 - accuracy: 0.9469 - val_loss: 0.1139 - val_accuracy: 0.9739\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 0.1511 - accuracy: 0.9478 - val_loss: 0.0900 - val_accuracy: 0.9739\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 0.1199 - accuracy: 0.9720 - val_loss: 0.0805 - val_accuracy: 0.9826\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 2s 41ms/step - loss: 0.1411 - accuracy: 0.9556 - val_loss: 0.1115 - val_accuracy: 0.9739\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 2s 44ms/step - loss: 2.2138 - accuracy: 0.2406 - val_loss: 1.8899 - val_accuracy: 0.3826\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 2s 43ms/step - loss: 1.3605 - accuracy: 0.5691 - val_loss: 0.7347 - val_accuracy: 0.8000\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 0.7395 - accuracy: 0.7623 - val_loss: 0.4068 - val_accuracy: 0.8870\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 0.5071 - accuracy: 0.8464 - val_loss: 0.2925 - val_accuracy: 0.9130\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 0.4333 - accuracy: 0.8647 - val_loss: 0.2260 - val_accuracy: 0.9565\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 0.3590 - accuracy: 0.8812 - val_loss: 0.2081 - val_accuracy: 0.9565\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 0.2984 - accuracy: 0.8976 - val_loss: 0.1880 - val_accuracy: 0.9652\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 0.2705 - accuracy: 0.8995 - val_loss: 0.2059 - val_accuracy: 0.9565\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 2s 41ms/step - loss: 0.2452 - accuracy: 0.9217 - val_loss: 0.1493 - val_accuracy: 0.9652\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 2s 41ms/step - loss: 0.1963 - accuracy: 0.9382 - val_loss: 0.1543 - val_accuracy: 0.9739\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 2s 41ms/step - loss: 0.1922 - accuracy: 0.9391 - val_loss: 0.1295 - val_accuracy: 0.9652\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 0.1804 - accuracy: 0.9382 - val_loss: 0.1177 - val_accuracy: 0.9826\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 2s 41ms/step - loss: 0.1918 - accuracy: 0.9362 - val_loss: 0.1259 - val_accuracy: 0.9739\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 2s 41ms/step - loss: 0.1587 - accuracy: 0.9527 - val_loss: 0.1179 - val_accuracy: 0.9826\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 0.1638 - accuracy: 0.9459 - val_loss: 0.1013 - val_accuracy: 0.9739\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 2s 44ms/step - loss: 2.2347 - accuracy: 0.2097 - val_loss: 1.9654 - val_accuracy: 0.4609\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 1.3069 - accuracy: 0.5961 - val_loss: 0.5489 - val_accuracy: 0.9217\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 0.6729 - accuracy: 0.7778 - val_loss: 0.3817 - val_accuracy: 0.9130\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 2s 41ms/step - loss: 0.4991 - accuracy: 0.8483 - val_loss: 0.1879 - val_accuracy: 0.9652\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 2s 43ms/step - loss: 0.3783 - accuracy: 0.8783 - val_loss: 0.1691 - val_accuracy: 0.9478\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 2s 41ms/step - loss: 0.3327 - accuracy: 0.8908 - val_loss: 0.0914 - val_accuracy: 0.9826\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 0.2523 - accuracy: 0.9227 - val_loss: 0.0730 - val_accuracy: 0.9826\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 2s 41ms/step - loss: 0.2530 - accuracy: 0.9140 - val_loss: 0.0577 - val_accuracy: 0.9913\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 2s 41ms/step - loss: 0.2134 - accuracy: 0.9314 - val_loss: 0.0732 - val_accuracy: 0.9739\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 2s 43ms/step - loss: 0.2000 - accuracy: 0.9256 - val_loss: 0.0542 - val_accuracy: 0.9913\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 2s 41ms/step - loss: 0.1707 - accuracy: 0.9469 - val_loss: 0.0412 - val_accuracy: 0.9913\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 2s 41ms/step - loss: 0.1659 - accuracy: 0.9536 - val_loss: 0.0357 - val_accuracy: 1.0000\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 2s 41ms/step - loss: 0.1497 - accuracy: 0.9594 - val_loss: 0.0201 - val_accuracy: 1.0000\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 2s 43ms/step - loss: 0.1746 - accuracy: 0.9372 - val_loss: 0.0187 - val_accuracy: 1.0000\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 0.1546 - accuracy: 0.9478 - val_loss: 0.0146 - val_accuracy: 1.0000\n",
            "Epoch 1/10\n",
            "7/7 [==============================] - 1s 169ms/step - loss: 2.2574 - accuracy: 0.1344 - val_loss: 1.9289 - val_accuracy: 0.3130\n",
            "Epoch 2/10\n",
            "7/7 [==============================] - 1s 161ms/step - loss: 1.4153 - accuracy: 0.5416 - val_loss: 0.8623 - val_accuracy: 0.7391\n",
            "Epoch 3/10\n",
            "7/7 [==============================] - 1s 152ms/step - loss: 0.7890 - accuracy: 0.7718 - val_loss: 0.3910 - val_accuracy: 0.9217\n",
            "Epoch 4/10\n",
            "7/7 [==============================] - 1s 156ms/step - loss: 0.3928 - accuracy: 0.8733 - val_loss: 0.2812 - val_accuracy: 0.9130\n",
            "Epoch 5/10\n",
            "7/7 [==============================] - 1s 153ms/step - loss: 0.2959 - accuracy: 0.9072 - val_loss: 0.3332 - val_accuracy: 0.8957\n",
            "Epoch 6/10\n",
            "7/7 [==============================] - 1s 159ms/step - loss: 0.2231 - accuracy: 0.9246 - val_loss: 0.1403 - val_accuracy: 0.9391\n",
            "Epoch 7/10\n",
            "7/7 [==============================] - 1s 152ms/step - loss: 0.1294 - accuracy: 0.9565 - val_loss: 0.2254 - val_accuracy: 0.9478\n",
            "Epoch 8/10\n",
            "7/7 [==============================] - 1s 159ms/step - loss: 0.1218 - accuracy: 0.9526 - val_loss: 0.1180 - val_accuracy: 0.9652\n",
            "Epoch 9/10\n",
            "7/7 [==============================] - 1s 152ms/step - loss: 0.0849 - accuracy: 0.9662 - val_loss: 0.1758 - val_accuracy: 0.9565\n",
            "Epoch 10/10\n",
            "7/7 [==============================] - 1s 155ms/step - loss: 0.0851 - accuracy: 0.9642 - val_loss: 0.1268 - val_accuracy: 0.9565\n",
            "Epoch 1/10\n",
            "7/7 [==============================] - 1s 168ms/step - loss: 2.2816 - accuracy: 0.1528 - val_loss: 2.1597 - val_accuracy: 0.3043\n",
            "Epoch 2/10\n",
            "7/7 [==============================] - 1s 155ms/step - loss: 1.6249 - accuracy: 0.4584 - val_loss: 0.8603 - val_accuracy: 0.7217\n",
            "Epoch 3/10\n",
            "7/7 [==============================] - 1s 152ms/step - loss: 0.7112 - accuracy: 0.7795 - val_loss: 0.4370 - val_accuracy: 0.8609\n",
            "Epoch 4/10\n",
            "7/7 [==============================] - 1s 159ms/step - loss: 0.3994 - accuracy: 0.8694 - val_loss: 0.3021 - val_accuracy: 0.9043\n",
            "Epoch 5/10\n",
            "7/7 [==============================] - 1s 152ms/step - loss: 0.2732 - accuracy: 0.9159 - val_loss: 0.3138 - val_accuracy: 0.9391\n",
            "Epoch 6/10\n",
            "7/7 [==============================] - 1s 158ms/step - loss: 0.2205 - accuracy: 0.9197 - val_loss: 0.2135 - val_accuracy: 0.9565\n",
            "Epoch 7/10\n",
            "7/7 [==============================] - 1s 153ms/step - loss: 0.1745 - accuracy: 0.9391 - val_loss: 0.1942 - val_accuracy: 0.9565\n",
            "Epoch 8/10\n",
            "7/7 [==============================] - 1s 151ms/step - loss: 0.1515 - accuracy: 0.9487 - val_loss: 0.1686 - val_accuracy: 0.9565\n",
            "Epoch 9/10\n",
            "7/7 [==============================] - 1s 154ms/step - loss: 0.1052 - accuracy: 0.9584 - val_loss: 0.1826 - val_accuracy: 0.9652\n",
            "Epoch 10/10\n",
            "7/7 [==============================] - 1s 151ms/step - loss: 0.1130 - accuracy: 0.9642 - val_loss: 0.1345 - val_accuracy: 0.9652\n",
            "Epoch 1/10\n",
            "7/7 [==============================] - 1s 171ms/step - loss: 2.3168 - accuracy: 0.1295 - val_loss: 2.1789 - val_accuracy: 0.2000\n",
            "Epoch 2/10\n",
            "7/7 [==============================] - 1s 153ms/step - loss: 1.8851 - accuracy: 0.3768 - val_loss: 1.1985 - val_accuracy: 0.6609\n",
            "Epoch 3/10\n",
            "7/7 [==============================] - 1s 152ms/step - loss: 0.8583 - accuracy: 0.7295 - val_loss: 0.5006 - val_accuracy: 0.8435\n",
            "Epoch 4/10\n",
            "7/7 [==============================] - 1s 154ms/step - loss: 0.4516 - accuracy: 0.8522 - val_loss: 0.3512 - val_accuracy: 0.8783\n",
            "Epoch 5/10\n",
            "7/7 [==============================] - 1s 159ms/step - loss: 0.3128 - accuracy: 0.8908 - val_loss: 0.2053 - val_accuracy: 0.9043\n",
            "Epoch 6/10\n",
            "7/7 [==============================] - 1s 152ms/step - loss: 0.2270 - accuracy: 0.9295 - val_loss: 0.2179 - val_accuracy: 0.9478\n",
            "Epoch 7/10\n",
            "7/7 [==============================] - 1s 156ms/step - loss: 0.1842 - accuracy: 0.9391 - val_loss: 0.1398 - val_accuracy: 0.9652\n",
            "Epoch 8/10\n",
            "7/7 [==============================] - 1s 154ms/step - loss: 0.1427 - accuracy: 0.9498 - val_loss: 0.1288 - val_accuracy: 0.9478\n",
            "Epoch 9/10\n",
            "7/7 [==============================] - 1s 159ms/step - loss: 0.1186 - accuracy: 0.9565 - val_loss: 0.0627 - val_accuracy: 0.9826\n",
            "Epoch 10/10\n",
            "7/7 [==============================] - 1s 152ms/step - loss: 0.1177 - accuracy: 0.9527 - val_loss: 0.1464 - val_accuracy: 0.9652\n",
            "Epoch 1/10\n",
            "7/7 [==============================] - 1s 177ms/step - loss: 2.3244 - accuracy: 0.1314 - val_loss: 2.2382 - val_accuracy: 0.1391\n",
            "Epoch 2/10\n",
            "7/7 [==============================] - 1s 152ms/step - loss: 2.1386 - accuracy: 0.2271 - val_loss: 1.7202 - val_accuracy: 0.4174\n",
            "Epoch 3/10\n",
            "7/7 [==============================] - 1s 157ms/step - loss: 1.4605 - accuracy: 0.4986 - val_loss: 0.8921 - val_accuracy: 0.7652\n",
            "Epoch 4/10\n",
            "7/7 [==============================] - 1s 153ms/step - loss: 0.8828 - accuracy: 0.7024 - val_loss: 0.6370 - val_accuracy: 0.7826\n",
            "Epoch 5/10\n",
            "7/7 [==============================] - 1s 151ms/step - loss: 0.6239 - accuracy: 0.8077 - val_loss: 0.5451 - val_accuracy: 0.8435\n",
            "Epoch 6/10\n",
            "7/7 [==============================] - 1s 160ms/step - loss: 0.5381 - accuracy: 0.8290 - val_loss: 0.4097 - val_accuracy: 0.9130\n",
            "Epoch 7/10\n",
            "7/7 [==============================] - 1s 153ms/step - loss: 0.3786 - accuracy: 0.8821 - val_loss: 0.5057 - val_accuracy: 0.8522\n",
            "Epoch 8/10\n",
            "7/7 [==============================] - 1s 157ms/step - loss: 0.3508 - accuracy: 0.8908 - val_loss: 0.3416 - val_accuracy: 0.9217\n",
            "Epoch 9/10\n",
            "7/7 [==============================] - 1s 153ms/step - loss: 0.2558 - accuracy: 0.9159 - val_loss: 0.2194 - val_accuracy: 0.9478\n",
            "Epoch 10/10\n",
            "7/7 [==============================] - 1s 156ms/step - loss: 0.1876 - accuracy: 0.9401 - val_loss: 0.1849 - val_accuracy: 0.9739\n",
            "WARNING:tensorflow:5 out of the last 22 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd0df1d950> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/10\n",
            "7/7 [==============================] - 1s 171ms/step - loss: 2.2774 - accuracy: 0.1691 - val_loss: 1.8092 - val_accuracy: 0.3478\n",
            "Epoch 2/10\n",
            "7/7 [==============================] - 1s 159ms/step - loss: 1.5693 - accuracy: 0.4705 - val_loss: 0.8202 - val_accuracy: 0.8000\n",
            "Epoch 3/10\n",
            "7/7 [==============================] - 1s 153ms/step - loss: 0.6529 - accuracy: 0.8126 - val_loss: 0.4524 - val_accuracy: 0.8957\n",
            "Epoch 4/10\n",
            "7/7 [==============================] - 1s 158ms/step - loss: 0.3530 - accuracy: 0.8754 - val_loss: 0.1436 - val_accuracy: 0.9739\n",
            "Epoch 5/10\n",
            "7/7 [==============================] - 1s 152ms/step - loss: 0.2797 - accuracy: 0.8947 - val_loss: 0.1857 - val_accuracy: 0.9391\n",
            "Epoch 6/10\n",
            "7/7 [==============================] - 1s 158ms/step - loss: 0.1905 - accuracy: 0.9324 - val_loss: 0.1098 - val_accuracy: 0.9739\n",
            "Epoch 7/10\n",
            "7/7 [==============================] - 1s 153ms/step - loss: 0.1435 - accuracy: 0.9546 - val_loss: 0.0658 - val_accuracy: 0.9739\n",
            "Epoch 8/10\n",
            "7/7 [==============================] - 1s 158ms/step - loss: 0.1192 - accuracy: 0.9614 - val_loss: 0.0336 - val_accuracy: 0.9826\n",
            "Epoch 9/10\n",
            "7/7 [==============================] - 1s 153ms/step - loss: 0.0821 - accuracy: 0.9729 - val_loss: 0.0270 - val_accuracy: 1.0000\n",
            "Epoch 10/10\n",
            "7/7 [==============================] - 1s 160ms/step - loss: 0.0766 - accuracy: 0.9700 - val_loss: 0.0356 - val_accuracy: 0.9913\n",
            "WARNING:tensorflow:6 out of the last 24 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd0ae0bb70> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 2s 45ms/step - loss: 2.1138 - accuracy: 0.2166 - val_loss: 1.0620 - val_accuracy: 0.6696\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 2s 43ms/step - loss: 0.8867 - accuracy: 0.6847 - val_loss: 0.2845 - val_accuracy: 0.8870\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 2s 43ms/step - loss: 0.6378 - accuracy: 0.7921 - val_loss: 0.1545 - val_accuracy: 0.9652\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 2s 43ms/step - loss: 0.5107 - accuracy: 0.8578 - val_loss: 0.2675 - val_accuracy: 0.9565\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 2s 43ms/step - loss: 0.5378 - accuracy: 0.8356 - val_loss: 0.1499 - val_accuracy: 0.9391\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 2s 41ms/step - loss: 0.5110 - accuracy: 0.8511 - val_loss: 0.1995 - val_accuracy: 0.9304\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 0.4606 - accuracy: 0.8520 - val_loss: 0.1368 - val_accuracy: 0.9565\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 0.4600 - accuracy: 0.8636 - val_loss: 0.1507 - val_accuracy: 0.9652\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 0.3299 - accuracy: 0.9023 - val_loss: 0.1114 - val_accuracy: 0.9565\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 0.3067 - accuracy: 0.9062 - val_loss: 0.0511 - val_accuracy: 0.9826\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 2s 45ms/step - loss: 0.2713 - accuracy: 0.9091 - val_loss: 0.1962 - val_accuracy: 0.9652\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 2s 45ms/step - loss: 0.2894 - accuracy: 0.9052 - val_loss: 0.0754 - val_accuracy: 0.9652\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 2s 44ms/step - loss: 0.2991 - accuracy: 0.9062 - val_loss: 0.1263 - val_accuracy: 0.9478\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 2s 47ms/step - loss: 0.7426 - accuracy: 0.8085 - val_loss: 0.1207 - val_accuracy: 0.9565\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 0.5161 - accuracy: 0.8578 - val_loss: 0.1459 - val_accuracy: 0.9565\n",
            "WARNING:tensorflow:6 out of the last 11 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd064d1ea0> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 2s 45ms/step - loss: 2.1996 - accuracy: 0.1905 - val_loss: 1.3880 - val_accuracy: 0.5391\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 2s 44ms/step - loss: 1.2609 - accuracy: 0.5812 - val_loss: 0.7122 - val_accuracy: 0.7739\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 2s 46ms/step - loss: 0.9061 - accuracy: 0.6963 - val_loss: 0.5356 - val_accuracy: 0.8174\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 2s 47ms/step - loss: 0.7376 - accuracy: 0.7611 - val_loss: 0.3133 - val_accuracy: 0.8957\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 2s 44ms/step - loss: 0.5260 - accuracy: 0.8366 - val_loss: 0.1470 - val_accuracy: 0.9652\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 2s 44ms/step - loss: 0.4226 - accuracy: 0.8781 - val_loss: 0.2680 - val_accuracy: 0.8957\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 0.4412 - accuracy: 0.8723 - val_loss: 0.1227 - val_accuracy: 0.9652\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 0.3470 - accuracy: 0.8907 - val_loss: 0.0805 - val_accuracy: 0.9652\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 2s 44ms/step - loss: 0.4391 - accuracy: 0.8752 - val_loss: 0.1269 - val_accuracy: 0.9565\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 2s 43ms/step - loss: 0.3338 - accuracy: 0.9023 - val_loss: 0.0956 - val_accuracy: 0.9739\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 0.3241 - accuracy: 0.8965 - val_loss: 0.1568 - val_accuracy: 0.9478\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 2s 44ms/step - loss: 0.3241 - accuracy: 0.9062 - val_loss: 0.1578 - val_accuracy: 0.9826\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 2s 43ms/step - loss: 0.2995 - accuracy: 0.9004 - val_loss: 0.2150 - val_accuracy: 0.9478\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 2s 45ms/step - loss: 0.4208 - accuracy: 0.8810 - val_loss: 0.0567 - val_accuracy: 0.9739\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 3s 56ms/step - loss: 0.3400 - accuracy: 0.9033 - val_loss: 0.0993 - val_accuracy: 0.9652\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 3s 49ms/step - loss: 2.3208 - accuracy: 0.1150 - val_loss: 2.3001 - val_accuracy: 0.1043\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 3s 51ms/step - loss: 2.3049 - accuracy: 0.0957 - val_loss: 2.2994 - val_accuracy: 0.1043\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 3s 49ms/step - loss: 2.3039 - accuracy: 0.1034 - val_loss: 2.2980 - val_accuracy: 0.1043\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 3s 51ms/step - loss: 2.3032 - accuracy: 0.1111 - val_loss: 2.3004 - val_accuracy: 0.1043\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 2s 45ms/step - loss: 2.3030 - accuracy: 0.1005 - val_loss: 2.2987 - val_accuracy: 0.1043\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 2s 45ms/step - loss: 2.3030 - accuracy: 0.1111 - val_loss: 2.3011 - val_accuracy: 0.1043\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 2s 43ms/step - loss: 2.3041 - accuracy: 0.0918 - val_loss: 2.2983 - val_accuracy: 0.1043\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 2s 47ms/step - loss: 2.3036 - accuracy: 0.1111 - val_loss: 2.3016 - val_accuracy: 0.1043\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 2s 46ms/step - loss: 2.3029 - accuracy: 0.1111 - val_loss: 2.2983 - val_accuracy: 0.1043\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 2s 47ms/step - loss: 2.3030 - accuracy: 0.0947 - val_loss: 2.3010 - val_accuracy: 0.1043\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 2s 42ms/step - loss: 2.3030 - accuracy: 0.1024 - val_loss: 2.3006 - val_accuracy: 0.1043\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 2s 44ms/step - loss: 2.3039 - accuracy: 0.1111 - val_loss: 2.2990 - val_accuracy: 0.1043\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 2s 45ms/step - loss: 2.3033 - accuracy: 0.0937 - val_loss: 2.3008 - val_accuracy: 0.1043\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 2s 45ms/step - loss: 2.3028 - accuracy: 0.1111 - val_loss: 2.2994 - val_accuracy: 0.1043\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 2s 45ms/step - loss: 2.3041 - accuracy: 0.1111 - val_loss: 2.3005 - val_accuracy: 0.1043\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 2s 48ms/step - loss: 2.2150 - accuracy: 0.1565 - val_loss: 1.6262 - val_accuracy: 0.4870\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 2s 44ms/step - loss: 1.3018 - accuracy: 0.5556 - val_loss: 0.6773 - val_accuracy: 0.8087\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 2s 45ms/step - loss: 0.8121 - accuracy: 0.7372 - val_loss: 0.4220 - val_accuracy: 0.8783\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 2s 47ms/step - loss: 0.6925 - accuracy: 0.7797 - val_loss: 0.3126 - val_accuracy: 0.8957\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 3s 49ms/step - loss: 0.6269 - accuracy: 0.8048 - val_loss: 0.2276 - val_accuracy: 0.9478\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 3s 49ms/step - loss: 0.6114 - accuracy: 0.8077 - val_loss: 0.4876 - val_accuracy: 0.8348\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 2s 47ms/step - loss: 0.5415 - accuracy: 0.8367 - val_loss: 0.2540 - val_accuracy: 0.9304\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 2s 44ms/step - loss: 0.4907 - accuracy: 0.8483 - val_loss: 0.1095 - val_accuracy: 0.9565\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 2s 45ms/step - loss: 0.4583 - accuracy: 0.8502 - val_loss: 0.1633 - val_accuracy: 0.9391\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 2s 44ms/step - loss: 0.4411 - accuracy: 0.8560 - val_loss: 0.1112 - val_accuracy: 0.9739\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 2s 45ms/step - loss: 0.4839 - accuracy: 0.8580 - val_loss: 0.1170 - val_accuracy: 0.9826\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 2s 44ms/step - loss: 0.4072 - accuracy: 0.8628 - val_loss: 0.1090 - val_accuracy: 0.9913\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 2s 45ms/step - loss: 0.3984 - accuracy: 0.8647 - val_loss: 0.1148 - val_accuracy: 0.9652\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 2s 44ms/step - loss: 0.4635 - accuracy: 0.8589 - val_loss: 0.1732 - val_accuracy: 0.9130\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 2s 45ms/step - loss: 0.3658 - accuracy: 0.8676 - val_loss: 0.1595 - val_accuracy: 0.9478\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 2s 45ms/step - loss: 2.3647 - accuracy: 0.1198 - val_loss: 2.3064 - val_accuracy: 0.1043\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 2s 44ms/step - loss: 2.3054 - accuracy: 0.0879 - val_loss: 2.3058 - val_accuracy: 0.0870\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 2s 44ms/step - loss: 2.3036 - accuracy: 0.0879 - val_loss: 2.3081 - val_accuracy: 0.0870\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 2s 46ms/step - loss: 2.3036 - accuracy: 0.0821 - val_loss: 2.3139 - val_accuracy: 0.0870\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 2s 44ms/step - loss: 2.3043 - accuracy: 0.1063 - val_loss: 2.3099 - val_accuracy: 0.0870\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 2s 43ms/step - loss: 2.3039 - accuracy: 0.1063 - val_loss: 2.3083 - val_accuracy: 0.1043\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 2s 43ms/step - loss: 2.3032 - accuracy: 0.0918 - val_loss: 2.3082 - val_accuracy: 0.0870\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 2s 47ms/step - loss: 2.3035 - accuracy: 0.0937 - val_loss: 2.3084 - val_accuracy: 0.0870\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 2s 43ms/step - loss: 2.3034 - accuracy: 0.0899 - val_loss: 2.3069 - val_accuracy: 0.0870\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 2s 43ms/step - loss: 2.3036 - accuracy: 0.1053 - val_loss: 2.3107 - val_accuracy: 0.1043\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 2s 44ms/step - loss: 2.3035 - accuracy: 0.1014 - val_loss: 2.3111 - val_accuracy: 0.0870\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 2s 43ms/step - loss: 2.3031 - accuracy: 0.1024 - val_loss: 2.3072 - val_accuracy: 0.1043\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 2s 45ms/step - loss: 2.3029 - accuracy: 0.0870 - val_loss: 2.3050 - val_accuracy: 0.1043\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 2s 45ms/step - loss: 2.3041 - accuracy: 0.1072 - val_loss: 2.3079 - val_accuracy: 0.1043\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 2s 44ms/step - loss: 2.3032 - accuracy: 0.1014 - val_loss: 2.3091 - val_accuracy: 0.1043\n",
            "Epoch 1/5\n",
            "7/7 [==============================] - 1s 135ms/step - loss: 2.2998 - accuracy: 0.1083 - val_loss: 2.2946 - val_accuracy: 0.1217\n",
            "Epoch 2/5\n",
            "7/7 [==============================] - 1s 121ms/step - loss: 2.2875 - accuracy: 0.1209 - val_loss: 2.2771 - val_accuracy: 0.1043\n",
            "Epoch 3/5\n",
            "7/7 [==============================] - 1s 116ms/step - loss: 2.2607 - accuracy: 0.1296 - val_loss: 2.2347 - val_accuracy: 0.1043\n",
            "Epoch 4/5\n",
            "7/7 [==============================] - 1s 121ms/step - loss: 2.2046 - accuracy: 0.1673 - val_loss: 2.1446 - val_accuracy: 0.2000\n",
            "Epoch 5/5\n",
            "7/7 [==============================] - 1s 119ms/step - loss: 2.1011 - accuracy: 0.2669 - val_loss: 1.9803 - val_accuracy: 0.3565\n",
            "Epoch 1/5\n",
            "7/7 [==============================] - 1s 138ms/step - loss: 2.2993 - accuracy: 0.1354 - val_loss: 2.2942 - val_accuracy: 0.2957\n",
            "Epoch 2/5\n",
            "7/7 [==============================] - 1s 116ms/step - loss: 2.2882 - accuracy: 0.1983 - val_loss: 2.2758 - val_accuracy: 0.2522\n",
            "Epoch 3/5\n",
            "7/7 [==============================] - 1s 120ms/step - loss: 2.2607 - accuracy: 0.2669 - val_loss: 2.2319 - val_accuracy: 0.4522\n",
            "Epoch 4/5\n",
            "7/7 [==============================] - 1s 117ms/step - loss: 2.2072 - accuracy: 0.3133 - val_loss: 2.1433 - val_accuracy: 0.4870\n",
            "Epoch 5/5\n",
            "7/7 [==============================] - 1s 120ms/step - loss: 2.1017 - accuracy: 0.3868 - val_loss: 1.9765 - val_accuracy: 0.6174\n",
            "Epoch 1/5\n",
            "7/7 [==============================] - 1s 135ms/step - loss: 2.2999 - accuracy: 0.1179 - val_loss: 2.2901 - val_accuracy: 0.1826\n",
            "Epoch 2/5\n",
            "7/7 [==============================] - 1s 122ms/step - loss: 2.2872 - accuracy: 0.1913 - val_loss: 2.2669 - val_accuracy: 0.1826\n",
            "Epoch 3/5\n",
            "7/7 [==============================] - 1s 116ms/step - loss: 2.2593 - accuracy: 0.2135 - val_loss: 2.2171 - val_accuracy: 0.2609\n",
            "Epoch 4/5\n",
            "7/7 [==============================] - 1s 119ms/step - loss: 2.1940 - accuracy: 0.3014 - val_loss: 2.1158 - val_accuracy: 0.4087\n",
            "Epoch 5/5\n",
            "7/7 [==============================] - 1s 116ms/step - loss: 2.0751 - accuracy: 0.3691 - val_loss: 1.9214 - val_accuracy: 0.4957\n",
            "Epoch 1/5\n",
            "7/7 [==============================] - 1s 136ms/step - loss: 2.2996 - accuracy: 0.1266 - val_loss: 2.2941 - val_accuracy: 0.3304\n",
            "Epoch 2/5\n",
            "7/7 [==============================] - 1s 116ms/step - loss: 2.2847 - accuracy: 0.2010 - val_loss: 2.2714 - val_accuracy: 0.1652\n",
            "Epoch 3/5\n",
            "7/7 [==============================] - 1s 120ms/step - loss: 2.2569 - accuracy: 0.2531 - val_loss: 2.2275 - val_accuracy: 0.2522\n",
            "Epoch 4/5\n",
            "7/7 [==============================] - 1s 118ms/step - loss: 2.1921 - accuracy: 0.3420 - val_loss: 2.1282 - val_accuracy: 0.3739\n",
            "Epoch 5/5\n",
            "7/7 [==============================] - 1s 119ms/step - loss: 2.0708 - accuracy: 0.4077 - val_loss: 1.9557 - val_accuracy: 0.5043\n",
            "WARNING:tensorflow:5 out of the last 22 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd0cb93488> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/5\n",
            "7/7 [==============================] - 1s 133ms/step - loss: 2.2992 - accuracy: 0.1353 - val_loss: 2.2886 - val_accuracy: 0.1739\n",
            "Epoch 2/5\n",
            "7/7 [==============================] - 1s 122ms/step - loss: 2.2823 - accuracy: 0.1739 - val_loss: 2.2637 - val_accuracy: 0.1652\n",
            "Epoch 3/5\n",
            "7/7 [==============================] - 1s 116ms/step - loss: 2.2464 - accuracy: 0.2068 - val_loss: 2.1996 - val_accuracy: 0.4696\n",
            "Epoch 4/5\n",
            "7/7 [==============================] - 1s 122ms/step - loss: 2.1660 - accuracy: 0.3217 - val_loss: 2.0671 - val_accuracy: 0.4870\n",
            "Epoch 5/5\n",
            "7/7 [==============================] - 1s 115ms/step - loss: 2.0117 - accuracy: 0.3816 - val_loss: 1.8484 - val_accuracy: 0.5391\n",
            "WARNING:tensorflow:6 out of the last 24 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd0e184268> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 1s 35ms/step - loss: 2.2722 - accuracy: 0.1480 - val_loss: 2.1902 - val_accuracy: 0.3391\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 1s 27ms/step - loss: 1.9820 - accuracy: 0.4410 - val_loss: 1.5986 - val_accuracy: 0.6087\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 1s 28ms/step - loss: 1.1688 - accuracy: 0.7534 - val_loss: 0.7125 - val_accuracy: 0.8609\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 1s 30ms/step - loss: 0.6338 - accuracy: 0.8346 - val_loss: 0.4514 - val_accuracy: 0.8696\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 0.4581 - accuracy: 0.8762 - val_loss: 0.3463 - val_accuracy: 0.8957\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 1s 28ms/step - loss: 0.3690 - accuracy: 0.8965 - val_loss: 0.2934 - val_accuracy: 0.9130\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 1s 28ms/step - loss: 0.2799 - accuracy: 0.9265 - val_loss: 0.2642 - val_accuracy: 0.9391\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 1s 30ms/step - loss: 0.2518 - accuracy: 0.9275 - val_loss: 0.2161 - val_accuracy: 0.9478\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 1s 27ms/step - loss: 0.2172 - accuracy: 0.9449 - val_loss: 0.1851 - val_accuracy: 0.9478\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 1s 30ms/step - loss: 0.1950 - accuracy: 0.9449 - val_loss: 0.1612 - val_accuracy: 0.9652\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 1s 28ms/step - loss: 0.1676 - accuracy: 0.9565 - val_loss: 0.1395 - val_accuracy: 0.9826\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 1s 28ms/step - loss: 0.1672 - accuracy: 0.9497 - val_loss: 0.1533 - val_accuracy: 0.9652\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 1s 28ms/step - loss: 0.1303 - accuracy: 0.9613 - val_loss: 0.1136 - val_accuracy: 0.9739\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 1s 30ms/step - loss: 0.1256 - accuracy: 0.9652 - val_loss: 0.1045 - val_accuracy: 0.9826\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 1s 28ms/step - loss: 0.1017 - accuracy: 0.9749 - val_loss: 0.0980 - val_accuracy: 0.9826\n",
            "WARNING:tensorflow:6 out of the last 11 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd0acb5620> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 1s 33ms/step - loss: 2.2773 - accuracy: 0.2321 - val_loss: 2.2137 - val_accuracy: 0.5043\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 1s 30ms/step - loss: 2.0634 - accuracy: 0.4807 - val_loss: 1.7670 - val_accuracy: 0.7478\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 1.3975 - accuracy: 0.7302 - val_loss: 0.9751 - val_accuracy: 0.7652\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 1s 28ms/step - loss: 0.7812 - accuracy: 0.8008 - val_loss: 0.5265 - val_accuracy: 0.8783\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 1s 28ms/step - loss: 0.5035 - accuracy: 0.8491 - val_loss: 0.3743 - val_accuracy: 0.9130\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 1s 28ms/step - loss: 0.4163 - accuracy: 0.8781 - val_loss: 0.2897 - val_accuracy: 0.9565\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 1s 28ms/step - loss: 0.3284 - accuracy: 0.8946 - val_loss: 0.2936 - val_accuracy: 0.9391\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 0.2696 - accuracy: 0.9207 - val_loss: 0.2052 - val_accuracy: 0.9565\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 0.2295 - accuracy: 0.9352 - val_loss: 0.2113 - val_accuracy: 0.9565\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 1s 27ms/step - loss: 0.2031 - accuracy: 0.9468 - val_loss: 0.1610 - val_accuracy: 0.9652\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 0.1741 - accuracy: 0.9497 - val_loss: 0.1507 - val_accuracy: 0.9652\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 0.1752 - accuracy: 0.9458 - val_loss: 0.1346 - val_accuracy: 0.9739\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 1s 27ms/step - loss: 0.1438 - accuracy: 0.9603 - val_loss: 0.1273 - val_accuracy: 0.9826\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 1s 27ms/step - loss: 0.1371 - accuracy: 0.9574 - val_loss: 0.1126 - val_accuracy: 0.9826\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 0.1186 - accuracy: 0.9652 - val_loss: 0.0957 - val_accuracy: 0.9913\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 1s 34ms/step - loss: 2.2795 - accuracy: 0.1768 - val_loss: 2.2196 - val_accuracy: 0.3304\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 1s 27ms/step - loss: 2.0676 - accuracy: 0.5101 - val_loss: 1.7400 - val_accuracy: 0.7565\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 1.3556 - accuracy: 0.7507 - val_loss: 0.8386 - val_accuracy: 0.8609\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 0.6979 - accuracy: 0.8232 - val_loss: 0.5222 - val_accuracy: 0.8522\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 1s 27ms/step - loss: 0.4696 - accuracy: 0.8599 - val_loss: 0.3543 - val_accuracy: 0.9043\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 1s 30ms/step - loss: 0.3574 - accuracy: 0.9034 - val_loss: 0.2981 - val_accuracy: 0.9217\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 0.2969 - accuracy: 0.9130 - val_loss: 0.2462 - val_accuracy: 0.9478\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 1s 27ms/step - loss: 0.2595 - accuracy: 0.9179 - val_loss: 0.2220 - val_accuracy: 0.9391\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 0.2170 - accuracy: 0.9391 - val_loss: 0.1841 - val_accuracy: 0.9565\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 0.1952 - accuracy: 0.9430 - val_loss: 0.1676 - val_accuracy: 0.9391\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 1s 27ms/step - loss: 0.1831 - accuracy: 0.9498 - val_loss: 0.1385 - val_accuracy: 0.9739\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 0.1438 - accuracy: 0.9575 - val_loss: 0.1318 - val_accuracy: 0.9739\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 0.1460 - accuracy: 0.9527 - val_loss: 0.1482 - val_accuracy: 0.9304\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 1s 26ms/step - loss: 0.1413 - accuracy: 0.9652 - val_loss: 0.1084 - val_accuracy: 0.9739\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 1s 30ms/step - loss: 0.1223 - accuracy: 0.9652 - val_loss: 0.1068 - val_accuracy: 0.9913\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 1s 33ms/step - loss: 2.2789 - accuracy: 0.2367 - val_loss: 2.2164 - val_accuracy: 0.4870\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 1s 28ms/step - loss: 2.0540 - accuracy: 0.5739 - val_loss: 1.7341 - val_accuracy: 0.8696\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 1s 28ms/step - loss: 1.3172 - accuracy: 0.7594 - val_loss: 0.8511 - val_accuracy: 0.8609\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 0.6792 - accuracy: 0.8290 - val_loss: 0.4770 - val_accuracy: 0.8696\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 0.4614 - accuracy: 0.8657 - val_loss: 0.3700 - val_accuracy: 0.8870\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 1s 27ms/step - loss: 0.3638 - accuracy: 0.8957 - val_loss: 0.2691 - val_accuracy: 0.9565\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 1s 30ms/step - loss: 0.2804 - accuracy: 0.9159 - val_loss: 0.2193 - val_accuracy: 0.9565\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 1s 28ms/step - loss: 0.2511 - accuracy: 0.9256 - val_loss: 0.1837 - val_accuracy: 0.9739\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 1s 27ms/step - loss: 0.2093 - accuracy: 0.9324 - val_loss: 0.1802 - val_accuracy: 0.9478\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 0.1706 - accuracy: 0.9507 - val_loss: 0.1506 - val_accuracy: 0.9739\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 1s 31ms/step - loss: 0.1646 - accuracy: 0.9488 - val_loss: 0.1432 - val_accuracy: 0.9652\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 1s 28ms/step - loss: 0.1507 - accuracy: 0.9488 - val_loss: 0.1247 - val_accuracy: 0.9652\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 0.1377 - accuracy: 0.9604 - val_loss: 0.0985 - val_accuracy: 0.9739\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 0.1242 - accuracy: 0.9565 - val_loss: 0.1039 - val_accuracy: 0.9739\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 1s 27ms/step - loss: 0.1239 - accuracy: 0.9652 - val_loss: 0.1004 - val_accuracy: 0.9826\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 1s 33ms/step - loss: 2.2636 - accuracy: 0.3295 - val_loss: 2.1809 - val_accuracy: 0.4174\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 1s 30ms/step - loss: 1.9427 - accuracy: 0.5246 - val_loss: 1.5705 - val_accuracy: 0.7043\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 1.1510 - accuracy: 0.7768 - val_loss: 0.7633 - val_accuracy: 0.8783\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 1s 27ms/step - loss: 0.6276 - accuracy: 0.8300 - val_loss: 0.4538 - val_accuracy: 0.9217\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 0.4417 - accuracy: 0.8763 - val_loss: 0.3266 - val_accuracy: 0.9217\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 0.3436 - accuracy: 0.9014 - val_loss: 0.2576 - val_accuracy: 0.9478\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 0.2846 - accuracy: 0.9179 - val_loss: 0.2178 - val_accuracy: 0.9565\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 1s 30ms/step - loss: 0.2582 - accuracy: 0.9285 - val_loss: 0.1470 - val_accuracy: 0.9565\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 0.2335 - accuracy: 0.9285 - val_loss: 0.1695 - val_accuracy: 0.9217\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 1s 28ms/step - loss: 0.2095 - accuracy: 0.9430 - val_loss: 0.1365 - val_accuracy: 0.9478\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 0.1738 - accuracy: 0.9536 - val_loss: 0.0854 - val_accuracy: 0.9826\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 0.1383 - accuracy: 0.9604 - val_loss: 0.0853 - val_accuracy: 0.9739\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 0.1361 - accuracy: 0.9614 - val_loss: 0.0714 - val_accuracy: 0.9826\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 1s 28ms/step - loss: 0.1121 - accuracy: 0.9681 - val_loss: 0.0505 - val_accuracy: 0.9826\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 0.1273 - accuracy: 0.9614 - val_loss: 0.0421 - val_accuracy: 0.9913\n",
            "Epoch 1/10\n",
            "21/21 [==============================] - 1s 43ms/step - loss: 2.2472 - accuracy: 0.1760 - val_loss: 1.7288 - val_accuracy: 0.4522\n",
            "Epoch 2/10\n",
            "21/21 [==============================] - 1s 36ms/step - loss: 1.0973 - accuracy: 0.6431 - val_loss: 0.5466 - val_accuracy: 0.8087\n",
            "Epoch 3/10\n",
            "21/21 [==============================] - 1s 34ms/step - loss: 0.5255 - accuracy: 0.8259 - val_loss: 0.2445 - val_accuracy: 0.9565\n",
            "Epoch 4/10\n",
            "21/21 [==============================] - 1s 36ms/step - loss: 0.3403 - accuracy: 0.8975 - val_loss: 0.2386 - val_accuracy: 0.8957\n",
            "Epoch 5/10\n",
            "21/21 [==============================] - 1s 38ms/step - loss: 0.3189 - accuracy: 0.8975 - val_loss: 0.1450 - val_accuracy: 0.9739\n",
            "Epoch 6/10\n",
            "21/21 [==============================] - 1s 36ms/step - loss: 0.2063 - accuracy: 0.9265 - val_loss: 0.1765 - val_accuracy: 0.9217\n",
            "Epoch 7/10\n",
            "21/21 [==============================] - 1s 35ms/step - loss: 0.2453 - accuracy: 0.9130 - val_loss: 0.1599 - val_accuracy: 0.9565\n",
            "Epoch 8/10\n",
            "21/21 [==============================] - 1s 37ms/step - loss: 0.2083 - accuracy: 0.9284 - val_loss: 0.1356 - val_accuracy: 0.9478\n",
            "Epoch 9/10\n",
            "21/21 [==============================] - 1s 37ms/step - loss: 0.1757 - accuracy: 0.9400 - val_loss: 0.1565 - val_accuracy: 0.9652\n",
            "Epoch 10/10\n",
            "21/21 [==============================] - 1s 36ms/step - loss: 0.2192 - accuracy: 0.9391 - val_loss: 0.2299 - val_accuracy: 0.9565\n",
            "Epoch 1/10\n",
            "21/21 [==============================] - 1s 41ms/step - loss: 2.2755 - accuracy: 0.1383 - val_loss: 1.9750 - val_accuracy: 0.5565\n",
            "Epoch 2/10\n",
            "21/21 [==============================] - 1s 37ms/step - loss: 1.2610 - accuracy: 0.5996 - val_loss: 0.6650 - val_accuracy: 0.7739\n",
            "Epoch 3/10\n",
            "21/21 [==============================] - 1s 36ms/step - loss: 0.5462 - accuracy: 0.8162 - val_loss: 0.3526 - val_accuracy: 0.9043\n",
            "Epoch 4/10\n",
            "21/21 [==============================] - 1s 37ms/step - loss: 0.3666 - accuracy: 0.8733 - val_loss: 0.1840 - val_accuracy: 0.9217\n",
            "Epoch 5/10\n",
            "21/21 [==============================] - 1s 35ms/step - loss: 0.2844 - accuracy: 0.9110 - val_loss: 0.1782 - val_accuracy: 0.9565\n",
            "Epoch 6/10\n",
            "21/21 [==============================] - 1s 34ms/step - loss: 0.2559 - accuracy: 0.9226 - val_loss: 0.2137 - val_accuracy: 0.9130\n",
            "Epoch 7/10\n",
            "21/21 [==============================] - 1s 37ms/step - loss: 0.2821 - accuracy: 0.9072 - val_loss: 0.2173 - val_accuracy: 0.9478\n",
            "Epoch 8/10\n",
            "21/21 [==============================] - 1s 37ms/step - loss: 0.2661 - accuracy: 0.9062 - val_loss: 0.1714 - val_accuracy: 0.9565\n",
            "Epoch 9/10\n",
            "21/21 [==============================] - 1s 35ms/step - loss: 0.1776 - accuracy: 0.9420 - val_loss: 0.0918 - val_accuracy: 0.9739\n",
            "Epoch 10/10\n",
            "21/21 [==============================] - 1s 35ms/step - loss: 0.2191 - accuracy: 0.9275 - val_loss: 0.1422 - val_accuracy: 0.9652\n",
            "Epoch 1/10\n",
            "21/21 [==============================] - 1s 45ms/step - loss: 2.2090 - accuracy: 0.1652 - val_loss: 1.5867 - val_accuracy: 0.5217\n",
            "Epoch 2/10\n",
            "21/21 [==============================] - 1s 37ms/step - loss: 1.0724 - accuracy: 0.6502 - val_loss: 0.4166 - val_accuracy: 0.8609\n",
            "Epoch 3/10\n",
            "21/21 [==============================] - 1s 36ms/step - loss: 0.5264 - accuracy: 0.8309 - val_loss: 0.2862 - val_accuracy: 0.9304\n",
            "Epoch 4/10\n",
            "21/21 [==============================] - 1s 34ms/step - loss: 0.3603 - accuracy: 0.8860 - val_loss: 0.1994 - val_accuracy: 0.9391\n",
            "Epoch 5/10\n",
            "21/21 [==============================] - 1s 37ms/step - loss: 0.3082 - accuracy: 0.9072 - val_loss: 0.2218 - val_accuracy: 0.9304\n",
            "Epoch 6/10\n",
            "21/21 [==============================] - 1s 36ms/step - loss: 0.2547 - accuracy: 0.9111 - val_loss: 0.1892 - val_accuracy: 0.9565\n",
            "Epoch 7/10\n",
            "21/21 [==============================] - 1s 38ms/step - loss: 0.2412 - accuracy: 0.9063 - val_loss: 0.2098 - val_accuracy: 0.9565\n",
            "Epoch 8/10\n",
            "21/21 [==============================] - 1s 35ms/step - loss: 0.2258 - accuracy: 0.9285 - val_loss: 0.1343 - val_accuracy: 0.9739\n",
            "Epoch 9/10\n",
            "21/21 [==============================] - 1s 36ms/step - loss: 0.2125 - accuracy: 0.9324 - val_loss: 0.1722 - val_accuracy: 0.9565\n",
            "Epoch 10/10\n",
            "21/21 [==============================] - 1s 38ms/step - loss: 0.1904 - accuracy: 0.9285 - val_loss: 0.1935 - val_accuracy: 0.9652\n",
            "Epoch 1/10\n",
            "21/21 [==============================] - 1s 44ms/step - loss: 2.2268 - accuracy: 0.1913 - val_loss: 1.6585 - val_accuracy: 0.5130\n",
            "Epoch 2/10\n",
            "21/21 [==============================] - 1s 37ms/step - loss: 1.0479 - accuracy: 0.6464 - val_loss: 0.5315 - val_accuracy: 0.8261\n",
            "Epoch 3/10\n",
            "21/21 [==============================] - 1s 34ms/step - loss: 0.5672 - accuracy: 0.8068 - val_loss: 0.3156 - val_accuracy: 0.9130\n",
            "Epoch 4/10\n",
            "21/21 [==============================] - 1s 36ms/step - loss: 0.3346 - accuracy: 0.8792 - val_loss: 0.3139 - val_accuracy: 0.9304\n",
            "Epoch 5/10\n",
            "21/21 [==============================] - 1s 37ms/step - loss: 0.3228 - accuracy: 0.8976 - val_loss: 0.2735 - val_accuracy: 0.9304\n",
            "Epoch 6/10\n",
            "21/21 [==============================] - 1s 36ms/step - loss: 0.2344 - accuracy: 0.9256 - val_loss: 0.3332 - val_accuracy: 0.9304\n",
            "Epoch 7/10\n",
            "21/21 [==============================] - 1s 33ms/step - loss: 0.2245 - accuracy: 0.9275 - val_loss: 0.2706 - val_accuracy: 0.9565\n",
            "Epoch 8/10\n",
            "21/21 [==============================] - 1s 38ms/step - loss: 0.1654 - accuracy: 0.9459 - val_loss: 0.2401 - val_accuracy: 0.9565\n",
            "Epoch 9/10\n",
            "21/21 [==============================] - 1s 37ms/step - loss: 0.1996 - accuracy: 0.9343 - val_loss: 0.1296 - val_accuracy: 0.9826\n",
            "Epoch 10/10\n",
            "21/21 [==============================] - 1s 37ms/step - loss: 0.1831 - accuracy: 0.9411 - val_loss: 0.2360 - val_accuracy: 0.9304\n",
            "Epoch 1/10\n",
            "21/21 [==============================] - 1s 41ms/step - loss: 2.3131 - accuracy: 0.1246 - val_loss: 2.2926 - val_accuracy: 0.2174\n",
            "Epoch 2/10\n",
            "21/21 [==============================] - 1s 34ms/step - loss: 1.9095 - accuracy: 0.3304 - val_loss: 1.1370 - val_accuracy: 0.7043\n",
            "Epoch 3/10\n",
            "21/21 [==============================] - 1s 38ms/step - loss: 0.8973 - accuracy: 0.7121 - val_loss: 0.4229 - val_accuracy: 0.8957\n",
            "Epoch 4/10\n",
            "21/21 [==============================] - 1s 37ms/step - loss: 0.4458 - accuracy: 0.8560 - val_loss: 0.2004 - val_accuracy: 0.9391\n",
            "Epoch 5/10\n",
            "21/21 [==============================] - 1s 36ms/step - loss: 0.3392 - accuracy: 0.8908 - val_loss: 0.1741 - val_accuracy: 0.9565\n",
            "Epoch 6/10\n",
            "21/21 [==============================] - 1s 34ms/step - loss: 0.2869 - accuracy: 0.9159 - val_loss: 0.0699 - val_accuracy: 0.9739\n",
            "Epoch 7/10\n",
            "21/21 [==============================] - 1s 37ms/step - loss: 0.2630 - accuracy: 0.9111 - val_loss: 0.0895 - val_accuracy: 0.9739\n",
            "Epoch 8/10\n",
            "21/21 [==============================] - 1s 37ms/step - loss: 0.2633 - accuracy: 0.9295 - val_loss: 0.1781 - val_accuracy: 0.9652\n",
            "Epoch 9/10\n",
            "21/21 [==============================] - 1s 37ms/step - loss: 0.2365 - accuracy: 0.9227 - val_loss: 0.0561 - val_accuracy: 0.9913\n",
            "Epoch 10/10\n",
            "21/21 [==============================] - 1s 34ms/step - loss: 0.2690 - accuracy: 0.9188 - val_loss: 0.0542 - val_accuracy: 1.0000\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 5s 103ms/step - loss: 1.9878 - accuracy: 0.3791 - val_loss: 1.0247 - val_accuracy: 0.6609\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 5s 102ms/step - loss: 0.5532 - accuracy: 0.8317 - val_loss: 0.2731 - val_accuracy: 0.9391\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 5s 101ms/step - loss: 0.2601 - accuracy: 0.9178 - val_loss: 0.1638 - val_accuracy: 0.9652\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 5s 103ms/step - loss: 0.1620 - accuracy: 0.9516 - val_loss: 0.1114 - val_accuracy: 0.9739\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 5s 101ms/step - loss: 0.1190 - accuracy: 0.9623 - val_loss: 0.1495 - val_accuracy: 0.9478\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 5s 101ms/step - loss: 0.1027 - accuracy: 0.9710 - val_loss: 0.1013 - val_accuracy: 0.9826\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 5s 101ms/step - loss: 0.0828 - accuracy: 0.9691 - val_loss: 0.0578 - val_accuracy: 0.9739\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 5s 104ms/step - loss: 0.0578 - accuracy: 0.9778 - val_loss: 0.0809 - val_accuracy: 0.9739\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 5s 103ms/step - loss: 0.0467 - accuracy: 0.9787 - val_loss: 0.0597 - val_accuracy: 0.9739\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 5s 103ms/step - loss: 0.0643 - accuracy: 0.9749 - val_loss: 0.0553 - val_accuracy: 0.9826\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 5s 103ms/step - loss: 0.0312 - accuracy: 0.9932 - val_loss: 0.0590 - val_accuracy: 0.9739\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 5s 103ms/step - loss: 0.0322 - accuracy: 0.9903 - val_loss: 0.0634 - val_accuracy: 0.9826\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 5s 101ms/step - loss: 0.0178 - accuracy: 0.9971 - val_loss: 0.0335 - val_accuracy: 0.9913\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 5s 102ms/step - loss: 0.0172 - accuracy: 0.9932 - val_loss: 0.0636 - val_accuracy: 0.9739\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 5s 101ms/step - loss: 0.0439 - accuracy: 0.9816 - val_loss: 0.0533 - val_accuracy: 0.9739\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 5s 103ms/step - loss: 1.8790 - accuracy: 0.4265 - val_loss: 0.7897 - val_accuracy: 0.7478\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 5s 100ms/step - loss: 0.4839 - accuracy: 0.8482 - val_loss: 0.3015 - val_accuracy: 0.9130\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 5s 101ms/step - loss: 0.2240 - accuracy: 0.9294 - val_loss: 0.1777 - val_accuracy: 0.9652\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 5s 101ms/step - loss: 0.1605 - accuracy: 0.9507 - val_loss: 0.1035 - val_accuracy: 0.9826\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 5s 101ms/step - loss: 0.1096 - accuracy: 0.9739 - val_loss: 0.1045 - val_accuracy: 0.9826\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 5s 102ms/step - loss: 0.0786 - accuracy: 0.9778 - val_loss: 0.0846 - val_accuracy: 0.9739\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 5s 101ms/step - loss: 0.0679 - accuracy: 0.9797 - val_loss: 0.0848 - val_accuracy: 0.9826\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 5s 104ms/step - loss: 0.0493 - accuracy: 0.9874 - val_loss: 0.0848 - val_accuracy: 0.9739\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 5s 101ms/step - loss: 0.0601 - accuracy: 0.9816 - val_loss: 0.0424 - val_accuracy: 0.9826\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 5s 100ms/step - loss: 0.0367 - accuracy: 0.9894 - val_loss: 0.0527 - val_accuracy: 0.9739\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 5s 100ms/step - loss: 0.0445 - accuracy: 0.9884 - val_loss: 0.0618 - val_accuracy: 0.9826\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 5s 100ms/step - loss: 0.0367 - accuracy: 0.9865 - val_loss: 0.0570 - val_accuracy: 0.9739\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 5s 102ms/step - loss: 0.0277 - accuracy: 0.9923 - val_loss: 0.0372 - val_accuracy: 0.9913\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 5s 101ms/step - loss: 0.0346 - accuracy: 0.9874 - val_loss: 0.0798 - val_accuracy: 0.9652\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 5s 101ms/step - loss: 0.0206 - accuracy: 0.9942 - val_loss: 0.0313 - val_accuracy: 0.9913\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 5s 103ms/step - loss: 1.9338 - accuracy: 0.3729 - val_loss: 0.8904 - val_accuracy: 0.7826\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 5s 101ms/step - loss: 0.5523 - accuracy: 0.8377 - val_loss: 0.2519 - val_accuracy: 0.9391\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 5s 101ms/step - loss: 0.2338 - accuracy: 0.9343 - val_loss: 0.2096 - val_accuracy: 0.9478\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 5s 101ms/step - loss: 0.1703 - accuracy: 0.9478 - val_loss: 0.1150 - val_accuracy: 0.9739\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 5s 101ms/step - loss: 0.1305 - accuracy: 0.9643 - val_loss: 0.0862 - val_accuracy: 0.9739\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 5s 101ms/step - loss: 0.1012 - accuracy: 0.9633 - val_loss: 0.1231 - val_accuracy: 0.9739\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 5s 101ms/step - loss: 0.1036 - accuracy: 0.9623 - val_loss: 0.1126 - val_accuracy: 0.9739\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 5s 101ms/step - loss: 0.0728 - accuracy: 0.9807 - val_loss: 0.1069 - val_accuracy: 0.9478\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 5s 101ms/step - loss: 0.0429 - accuracy: 0.9874 - val_loss: 0.0594 - val_accuracy: 0.9826\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 5s 101ms/step - loss: 0.0303 - accuracy: 0.9932 - val_loss: 0.0436 - val_accuracy: 0.9913\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 5s 101ms/step - loss: 0.0403 - accuracy: 0.9845 - val_loss: 0.0461 - val_accuracy: 0.9826\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 5s 101ms/step - loss: 0.0313 - accuracy: 0.9874 - val_loss: 0.0468 - val_accuracy: 0.9826\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 5s 102ms/step - loss: 0.0406 - accuracy: 0.9865 - val_loss: 0.0703 - val_accuracy: 0.9826\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 5s 102ms/step - loss: 0.0265 - accuracy: 0.9913 - val_loss: 0.0263 - val_accuracy: 1.0000\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 5s 101ms/step - loss: 0.0236 - accuracy: 0.9932 - val_loss: 0.0208 - val_accuracy: 0.9913\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 5s 104ms/step - loss: 1.8118 - accuracy: 0.4560 - val_loss: 0.7511 - val_accuracy: 0.7565\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 5s 102ms/step - loss: 0.4617 - accuracy: 0.8667 - val_loss: 0.2973 - val_accuracy: 0.9304\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 5s 102ms/step - loss: 0.2572 - accuracy: 0.9227 - val_loss: 0.1681 - val_accuracy: 0.9565\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 5s 101ms/step - loss: 0.1718 - accuracy: 0.9517 - val_loss: 0.1430 - val_accuracy: 0.9565\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 5s 102ms/step - loss: 0.1391 - accuracy: 0.9594 - val_loss: 0.1071 - val_accuracy: 0.9652\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 5s 102ms/step - loss: 0.1117 - accuracy: 0.9594 - val_loss: 0.1218 - val_accuracy: 0.9652\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 5s 102ms/step - loss: 0.0879 - accuracy: 0.9720 - val_loss: 0.0934 - val_accuracy: 0.9739\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 5s 102ms/step - loss: 0.0750 - accuracy: 0.9749 - val_loss: 0.0592 - val_accuracy: 0.9913\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 5s 103ms/step - loss: 0.0635 - accuracy: 0.9797 - val_loss: 0.0527 - val_accuracy: 0.9826\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 5s 102ms/step - loss: 0.0497 - accuracy: 0.9874 - val_loss: 0.0353 - val_accuracy: 0.9913\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 5s 101ms/step - loss: 0.0435 - accuracy: 0.9874 - val_loss: 0.0788 - val_accuracy: 0.9739\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 5s 102ms/step - loss: 0.0304 - accuracy: 0.9894 - val_loss: 0.0401 - val_accuracy: 0.9826\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 5s 101ms/step - loss: 0.0436 - accuracy: 0.9855 - val_loss: 0.0232 - val_accuracy: 0.9913\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 5s 104ms/step - loss: 0.0376 - accuracy: 0.9874 - val_loss: 0.0524 - val_accuracy: 0.9739\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 5s 102ms/step - loss: 0.0262 - accuracy: 0.9923 - val_loss: 0.0344 - val_accuracy: 0.9826\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 5s 104ms/step - loss: 1.9213 - accuracy: 0.3913 - val_loss: 0.8922 - val_accuracy: 0.7217\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 5s 102ms/step - loss: 0.5787 - accuracy: 0.8222 - val_loss: 0.3379 - val_accuracy: 0.9130\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 5s 102ms/step - loss: 0.2554 - accuracy: 0.9275 - val_loss: 0.0836 - val_accuracy: 0.9739\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 5s 102ms/step - loss: 0.1806 - accuracy: 0.9449 - val_loss: 0.0506 - val_accuracy: 1.0000\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 5s 102ms/step - loss: 0.1352 - accuracy: 0.9623 - val_loss: 0.0388 - val_accuracy: 1.0000\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 5s 102ms/step - loss: 0.0953 - accuracy: 0.9681 - val_loss: 0.0355 - val_accuracy: 0.9913\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 5s 103ms/step - loss: 0.0720 - accuracy: 0.9749 - val_loss: 0.0218 - val_accuracy: 1.0000\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 5s 104ms/step - loss: 0.0758 - accuracy: 0.9749 - val_loss: 0.0215 - val_accuracy: 1.0000\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 5s 103ms/step - loss: 0.0580 - accuracy: 0.9807 - val_loss: 0.0085 - val_accuracy: 1.0000\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 5s 104ms/step - loss: 0.0532 - accuracy: 0.9826 - val_loss: 0.0149 - val_accuracy: 1.0000\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 5s 103ms/step - loss: 0.0501 - accuracy: 0.9874 - val_loss: 0.0132 - val_accuracy: 1.0000\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 5s 102ms/step - loss: 0.0314 - accuracy: 0.9923 - val_loss: 0.0035 - val_accuracy: 1.0000\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 5s 102ms/step - loss: 0.0195 - accuracy: 0.9942 - val_loss: 0.0228 - val_accuracy: 0.9913\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 5s 102ms/step - loss: 0.0365 - accuracy: 0.9865 - val_loss: 0.0073 - val_accuracy: 1.0000\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 5s 102ms/step - loss: 0.0297 - accuracy: 0.9903 - val_loss: 0.0066 - val_accuracy: 1.0000\n",
            "Epoch 1/15\n",
            "7/7 [==============================] - 1s 77ms/step - loss: 2.2854 - accuracy: 0.1402 - val_loss: 2.1227 - val_accuracy: 0.2870\n",
            "Epoch 2/15\n",
            "7/7 [==============================] - 0s 54ms/step - loss: 1.7440 - accuracy: 0.4130 - val_loss: 0.9556 - val_accuracy: 0.7304\n",
            "Epoch 3/15\n",
            "7/7 [==============================] - 0s 58ms/step - loss: 0.9081 - accuracy: 0.7079 - val_loss: 0.6124 - val_accuracy: 0.8174\n",
            "Epoch 4/15\n",
            "7/7 [==============================] - 0s 55ms/step - loss: 0.5823 - accuracy: 0.8056 - val_loss: 0.4297 - val_accuracy: 0.8609\n",
            "Epoch 5/15\n",
            "7/7 [==============================] - 0s 60ms/step - loss: 0.4657 - accuracy: 0.8453 - val_loss: 0.2970 - val_accuracy: 0.9478\n",
            "Epoch 6/15\n",
            "7/7 [==============================] - 0s 55ms/step - loss: 0.3422 - accuracy: 0.8917 - val_loss: 0.2629 - val_accuracy: 0.9391\n",
            "Epoch 7/15\n",
            "7/7 [==============================] - 0s 55ms/step - loss: 0.2708 - accuracy: 0.9101 - val_loss: 0.2260 - val_accuracy: 0.9304\n",
            "Epoch 8/15\n",
            "7/7 [==============================] - 0s 57ms/step - loss: 0.2041 - accuracy: 0.9226 - val_loss: 0.2086 - val_accuracy: 0.9304\n",
            "Epoch 9/15\n",
            "7/7 [==============================] - 0s 53ms/step - loss: 0.1692 - accuracy: 0.9487 - val_loss: 0.1896 - val_accuracy: 0.9565\n",
            "Epoch 10/15\n",
            "7/7 [==============================] - 0s 59ms/step - loss: 0.1545 - accuracy: 0.9565 - val_loss: 0.1501 - val_accuracy: 0.9565\n",
            "Epoch 11/15\n",
            "7/7 [==============================] - 0s 53ms/step - loss: 0.1444 - accuracy: 0.9536 - val_loss: 0.1428 - val_accuracy: 0.9565\n",
            "Epoch 12/15\n",
            "7/7 [==============================] - 0s 59ms/step - loss: 0.1191 - accuracy: 0.9623 - val_loss: 0.1233 - val_accuracy: 0.9826\n",
            "Epoch 13/15\n",
            "7/7 [==============================] - 0s 55ms/step - loss: 0.1081 - accuracy: 0.9662 - val_loss: 0.1559 - val_accuracy: 0.9652\n",
            "Epoch 14/15\n",
            "7/7 [==============================] - 0s 61ms/step - loss: 0.0994 - accuracy: 0.9652 - val_loss: 0.1902 - val_accuracy: 0.9478\n",
            "Epoch 15/15\n",
            "7/7 [==============================] - 0s 53ms/step - loss: 0.0958 - accuracy: 0.9603 - val_loss: 0.1211 - val_accuracy: 0.9652\n",
            "Epoch 1/15\n",
            "7/7 [==============================] - 1s 148ms/step - loss: 2.2917 - accuracy: 0.1093 - val_loss: 2.1820 - val_accuracy: 0.2870\n",
            "Epoch 2/15\n",
            "7/7 [==============================] - 0s 53ms/step - loss: 1.7725 - accuracy: 0.4139 - val_loss: 1.0174 - val_accuracy: 0.7565\n",
            "Epoch 3/15\n",
            "7/7 [==============================] - 0s 60ms/step - loss: 0.9633 - accuracy: 0.7041 - val_loss: 0.5382 - val_accuracy: 0.8435\n",
            "Epoch 4/15\n",
            "7/7 [==============================] - 0s 54ms/step - loss: 0.5884 - accuracy: 0.8230 - val_loss: 0.3321 - val_accuracy: 0.9130\n",
            "Epoch 5/15\n",
            "7/7 [==============================] - 0s 61ms/step - loss: 0.3842 - accuracy: 0.8733 - val_loss: 0.2600 - val_accuracy: 0.9391\n",
            "Epoch 6/15\n",
            "7/7 [==============================] - 0s 53ms/step - loss: 0.3197 - accuracy: 0.8965 - val_loss: 0.2372 - val_accuracy: 0.9478\n",
            "Epoch 7/15\n",
            "7/7 [==============================] - 0s 62ms/step - loss: 0.2609 - accuracy: 0.9197 - val_loss: 0.1867 - val_accuracy: 0.9652\n",
            "Epoch 8/15\n",
            "7/7 [==============================] - 0s 54ms/step - loss: 0.2071 - accuracy: 0.9217 - val_loss: 0.1858 - val_accuracy: 0.9565\n",
            "Epoch 9/15\n",
            "7/7 [==============================] - 0s 49ms/step - loss: 0.1880 - accuracy: 0.9391 - val_loss: 0.2339 - val_accuracy: 0.9652\n",
            "Epoch 10/15\n",
            "7/7 [==============================] - 0s 62ms/step - loss: 0.1852 - accuracy: 0.9362 - val_loss: 0.1382 - val_accuracy: 0.9565\n",
            "Epoch 11/15\n",
            "7/7 [==============================] - 0s 54ms/step - loss: 0.1563 - accuracy: 0.9487 - val_loss: 0.1317 - val_accuracy: 0.9826\n",
            "Epoch 12/15\n",
            "7/7 [==============================] - 0s 61ms/step - loss: 0.1485 - accuracy: 0.9497 - val_loss: 0.1108 - val_accuracy: 0.9739\n",
            "Epoch 13/15\n",
            "7/7 [==============================] - 0s 53ms/step - loss: 0.1446 - accuracy: 0.9584 - val_loss: 0.0989 - val_accuracy: 0.9739\n",
            "Epoch 14/15\n",
            "7/7 [==============================] - 0s 58ms/step - loss: 0.1333 - accuracy: 0.9555 - val_loss: 0.1341 - val_accuracy: 0.9652\n",
            "Epoch 15/15\n",
            "7/7 [==============================] - 0s 54ms/step - loss: 0.1459 - accuracy: 0.9507 - val_loss: 0.1045 - val_accuracy: 0.9652\n",
            "Epoch 1/15\n",
            "7/7 [==============================] - 1s 75ms/step - loss: 2.2799 - accuracy: 0.1517 - val_loss: 2.1174 - val_accuracy: 0.3217\n",
            "Epoch 2/15\n",
            "7/7 [==============================] - 0s 56ms/step - loss: 1.7277 - accuracy: 0.4386 - val_loss: 0.8724 - val_accuracy: 0.8000\n",
            "Epoch 3/15\n",
            "7/7 [==============================] - 0s 59ms/step - loss: 0.8803 - accuracy: 0.7111 - val_loss: 0.5342 - val_accuracy: 0.8348\n",
            "Epoch 4/15\n",
            "7/7 [==============================] - 0s 56ms/step - loss: 0.5377 - accuracy: 0.8222 - val_loss: 0.5584 - val_accuracy: 0.8261\n",
            "Epoch 5/15\n",
            "7/7 [==============================] - 0s 57ms/step - loss: 0.4361 - accuracy: 0.8502 - val_loss: 0.3389 - val_accuracy: 0.9130\n",
            "Epoch 6/15\n",
            "7/7 [==============================] - 0s 56ms/step - loss: 0.3393 - accuracy: 0.8986 - val_loss: 0.2513 - val_accuracy: 0.9391\n",
            "Epoch 7/15\n",
            "7/7 [==============================] - 0s 60ms/step - loss: 0.2489 - accuracy: 0.9130 - val_loss: 0.1839 - val_accuracy: 0.9565\n",
            "Epoch 8/15\n",
            "7/7 [==============================] - 0s 56ms/step - loss: 0.2588 - accuracy: 0.9082 - val_loss: 0.1366 - val_accuracy: 0.9652\n",
            "Epoch 9/15\n",
            "7/7 [==============================] - 0s 56ms/step - loss: 0.1889 - accuracy: 0.9391 - val_loss: 0.1100 - val_accuracy: 0.9739\n",
            "Epoch 10/15\n",
            "7/7 [==============================] - 0s 61ms/step - loss: 0.1881 - accuracy: 0.9382 - val_loss: 0.0958 - val_accuracy: 0.9826\n",
            "Epoch 11/15\n",
            "7/7 [==============================] - 0s 56ms/step - loss: 0.1562 - accuracy: 0.9449 - val_loss: 0.1137 - val_accuracy: 0.9739\n",
            "Epoch 12/15\n",
            "7/7 [==============================] - 0s 59ms/step - loss: 0.1408 - accuracy: 0.9498 - val_loss: 0.0954 - val_accuracy: 0.9739\n",
            "Epoch 13/15\n",
            "7/7 [==============================] - 0s 56ms/step - loss: 0.1408 - accuracy: 0.9469 - val_loss: 0.0801 - val_accuracy: 0.9739\n",
            "Epoch 14/15\n",
            "7/7 [==============================] - 0s 62ms/step - loss: 0.1105 - accuracy: 0.9585 - val_loss: 0.0789 - val_accuracy: 0.9826\n",
            "Epoch 15/15\n",
            "7/7 [==============================] - 0s 56ms/step - loss: 0.1284 - accuracy: 0.9585 - val_loss: 0.0823 - val_accuracy: 0.9739\n",
            "Epoch 1/15\n",
            "7/7 [==============================] - 1s 81ms/step - loss: 2.2719 - accuracy: 0.1517 - val_loss: 2.0963 - val_accuracy: 0.3130\n",
            "Epoch 2/15\n",
            "7/7 [==============================] - 0s 51ms/step - loss: 1.5987 - accuracy: 0.4812 - val_loss: 1.0427 - val_accuracy: 0.6783\n",
            "Epoch 3/15\n",
            "7/7 [==============================] - 0s 62ms/step - loss: 0.8311 - accuracy: 0.7382 - val_loss: 0.6012 - val_accuracy: 0.8087\n",
            "Epoch 4/15\n",
            "7/7 [==============================] - 0s 51ms/step - loss: 0.5322 - accuracy: 0.8357 - val_loss: 0.3567 - val_accuracy: 0.9043\n",
            "Epoch 5/15\n",
            "7/7 [==============================] - 0s 56ms/step - loss: 0.3918 - accuracy: 0.8783 - val_loss: 0.2945 - val_accuracy: 0.9043\n",
            "Epoch 6/15\n",
            "7/7 [==============================] - 0s 58ms/step - loss: 0.2689 - accuracy: 0.9159 - val_loss: 0.2582 - val_accuracy: 0.9478\n",
            "Epoch 7/15\n",
            "7/7 [==============================] - 0s 55ms/step - loss: 0.2318 - accuracy: 0.9159 - val_loss: 0.2028 - val_accuracy: 0.9478\n",
            "Epoch 8/15\n",
            "7/7 [==============================] - 0s 61ms/step - loss: 0.2075 - accuracy: 0.9324 - val_loss: 0.1785 - val_accuracy: 0.9739\n",
            "Epoch 9/15\n",
            "7/7 [==============================] - 0s 55ms/step - loss: 0.1572 - accuracy: 0.9536 - val_loss: 0.1306 - val_accuracy: 0.9826\n",
            "Epoch 10/15\n",
            "7/7 [==============================] - 0s 61ms/step - loss: 0.1797 - accuracy: 0.9420 - val_loss: 0.1446 - val_accuracy: 0.9739\n",
            "Epoch 11/15\n",
            "7/7 [==============================] - 0s 55ms/step - loss: 0.1771 - accuracy: 0.9411 - val_loss: 0.1465 - val_accuracy: 0.9652\n",
            "Epoch 12/15\n",
            "7/7 [==============================] - 0s 61ms/step - loss: 0.1371 - accuracy: 0.9536 - val_loss: 0.1400 - val_accuracy: 0.9913\n",
            "Epoch 13/15\n",
            "7/7 [==============================] - 0s 55ms/step - loss: 0.1500 - accuracy: 0.9440 - val_loss: 0.1324 - val_accuracy: 0.9739\n",
            "Epoch 14/15\n",
            "7/7 [==============================] - 0s 60ms/step - loss: 0.1201 - accuracy: 0.9585 - val_loss: 0.1439 - val_accuracy: 0.9739\n",
            "Epoch 15/15\n",
            "7/7 [==============================] - 0s 54ms/step - loss: 0.1133 - accuracy: 0.9585 - val_loss: 0.1474 - val_accuracy: 0.9826\n",
            "WARNING:tensorflow:5 out of the last 22 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd0845d730> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/15\n",
            "7/7 [==============================] - 1s 80ms/step - loss: 2.2619 - accuracy: 0.1585 - val_loss: 1.9759 - val_accuracy: 0.3826\n",
            "Epoch 2/15\n",
            "7/7 [==============================] - 0s 54ms/step - loss: 1.5161 - accuracy: 0.5285 - val_loss: 0.9878 - val_accuracy: 0.7826\n",
            "Epoch 3/15\n",
            "7/7 [==============================] - 0s 59ms/step - loss: 0.7016 - accuracy: 0.7836 - val_loss: 0.3738 - val_accuracy: 0.8957\n",
            "Epoch 4/15\n",
            "7/7 [==============================] - 0s 57ms/step - loss: 0.4862 - accuracy: 0.8406 - val_loss: 0.2105 - val_accuracy: 0.9391\n",
            "Epoch 5/15\n",
            "7/7 [==============================] - 0s 60ms/step - loss: 0.3602 - accuracy: 0.8773 - val_loss: 0.1249 - val_accuracy: 0.9565\n",
            "Epoch 6/15\n",
            "7/7 [==============================] - 0s 57ms/step - loss: 0.2206 - accuracy: 0.9314 - val_loss: 0.0539 - val_accuracy: 0.9739\n",
            "Epoch 7/15\n",
            "7/7 [==============================] - 0s 60ms/step - loss: 0.2458 - accuracy: 0.9150 - val_loss: 0.0968 - val_accuracy: 0.9565\n",
            "Epoch 8/15\n",
            "7/7 [==============================] - 0s 57ms/step - loss: 0.1848 - accuracy: 0.9314 - val_loss: 0.0381 - val_accuracy: 0.9913\n",
            "Epoch 9/15\n",
            "7/7 [==============================] - 0s 54ms/step - loss: 0.1828 - accuracy: 0.9343 - val_loss: 0.0451 - val_accuracy: 0.9913\n",
            "Epoch 10/15\n",
            "7/7 [==============================] - 0s 58ms/step - loss: 0.1727 - accuracy: 0.9527 - val_loss: 0.0459 - val_accuracy: 0.9739\n",
            "Epoch 11/15\n",
            "7/7 [==============================] - 0s 56ms/step - loss: 0.1340 - accuracy: 0.9488 - val_loss: 0.0345 - val_accuracy: 0.9913\n",
            "Epoch 12/15\n",
            "7/7 [==============================] - 0s 62ms/step - loss: 0.1152 - accuracy: 0.9614 - val_loss: 0.0229 - val_accuracy: 0.9913\n",
            "Epoch 13/15\n",
            "7/7 [==============================] - 0s 53ms/step - loss: 0.1209 - accuracy: 0.9604 - val_loss: 0.0340 - val_accuracy: 0.9913\n",
            "Epoch 14/15\n",
            "7/7 [==============================] - 0s 59ms/step - loss: 0.1008 - accuracy: 0.9662 - val_loss: 0.0112 - val_accuracy: 1.0000\n",
            "Epoch 15/15\n",
            "7/7 [==============================] - 0s 54ms/step - loss: 0.0805 - accuracy: 0.9729 - val_loss: 0.0120 - val_accuracy: 1.0000\n",
            "WARNING:tensorflow:6 out of the last 24 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd0a8e1488> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/5\n",
            "21/21 [==============================] - 1s 61ms/step - loss: 2.1003 - accuracy: 0.2389 - val_loss: 1.2924 - val_accuracy: 0.5130\n",
            "Epoch 2/5\n",
            "21/21 [==============================] - 1s 49ms/step - loss: 0.7405 - accuracy: 0.7863 - val_loss: 0.2313 - val_accuracy: 0.9043\n",
            "Epoch 3/5\n",
            "21/21 [==============================] - 1s 52ms/step - loss: 0.3483 - accuracy: 0.8888 - val_loss: 0.2096 - val_accuracy: 0.9304\n",
            "Epoch 4/5\n",
            "21/21 [==============================] - 1s 50ms/step - loss: 0.2260 - accuracy: 0.9313 - val_loss: 0.1361 - val_accuracy: 0.9565\n",
            "Epoch 5/5\n",
            "21/21 [==============================] - 1s 54ms/step - loss: 0.2180 - accuracy: 0.9236 - val_loss: 0.1578 - val_accuracy: 0.9565\n",
            "WARNING:tensorflow:6 out of the last 11 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd0a992378> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/5\n",
            "21/21 [==============================] - 1s 57ms/step - loss: 2.0218 - accuracy: 0.2795 - val_loss: 0.9055 - val_accuracy: 0.6783\n",
            "Epoch 2/5\n",
            "21/21 [==============================] - 1s 52ms/step - loss: 0.6430 - accuracy: 0.7911 - val_loss: 0.3002 - val_accuracy: 0.9217\n",
            "Epoch 3/5\n",
            "21/21 [==============================] - 1s 54ms/step - loss: 0.3652 - accuracy: 0.8888 - val_loss: 0.1578 - val_accuracy: 0.9652\n",
            "Epoch 4/5\n",
            "21/21 [==============================] - 1s 52ms/step - loss: 0.2229 - accuracy: 0.9246 - val_loss: 0.1446 - val_accuracy: 0.9565\n",
            "Epoch 5/5\n",
            "21/21 [==============================] - 1s 49ms/step - loss: 0.2396 - accuracy: 0.9246 - val_loss: 0.1742 - val_accuracy: 0.9478\n",
            "Epoch 1/5\n",
            "21/21 [==============================] - 1s 59ms/step - loss: 2.0860 - accuracy: 0.2348 - val_loss: 1.0292 - val_accuracy: 0.6696\n",
            "Epoch 2/5\n",
            "21/21 [==============================] - 1s 49ms/step - loss: 0.7657 - accuracy: 0.7401 - val_loss: 0.2997 - val_accuracy: 0.9130\n",
            "Epoch 3/5\n",
            "21/21 [==============================] - 1s 50ms/step - loss: 0.3484 - accuracy: 0.8802 - val_loss: 0.2409 - val_accuracy: 0.9304\n",
            "Epoch 4/5\n",
            "21/21 [==============================] - 1s 51ms/step - loss: 0.2211 - accuracy: 0.9227 - val_loss: 0.1414 - val_accuracy: 0.9652\n",
            "Epoch 5/5\n",
            "21/21 [==============================] - 1s 51ms/step - loss: 0.1936 - accuracy: 0.9343 - val_loss: 0.1493 - val_accuracy: 0.9652\n",
            "Epoch 1/5\n",
            "21/21 [==============================] - 1s 59ms/step - loss: 2.2757 - accuracy: 0.1459 - val_loss: 2.0221 - val_accuracy: 0.4435\n",
            "Epoch 2/5\n",
            "21/21 [==============================] - 1s 52ms/step - loss: 1.3689 - accuracy: 0.5227 - val_loss: 0.4841 - val_accuracy: 0.8435\n",
            "Epoch 3/5\n",
            "21/21 [==============================] - 1s 54ms/step - loss: 0.6099 - accuracy: 0.7913 - val_loss: 0.3477 - val_accuracy: 0.8696\n",
            "Epoch 4/5\n",
            "21/21 [==============================] - 1s 51ms/step - loss: 0.3773 - accuracy: 0.8676 - val_loss: 0.3534 - val_accuracy: 0.8957\n",
            "Epoch 5/5\n",
            "21/21 [==============================] - 1s 51ms/step - loss: 0.3394 - accuracy: 0.8899 - val_loss: 0.1417 - val_accuracy: 0.9478\n",
            "Epoch 1/5\n",
            "21/21 [==============================] - 1s 60ms/step - loss: 2.3369 - accuracy: 0.1034 - val_loss: 2.3323 - val_accuracy: 0.1304\n",
            "Epoch 2/5\n",
            "21/21 [==============================] - 1s 53ms/step - loss: 2.2003 - accuracy: 0.2000 - val_loss: 1.8831 - val_accuracy: 0.5565\n",
            "Epoch 3/5\n",
            "21/21 [==============================] - 1s 49ms/step - loss: 1.5733 - accuracy: 0.4812 - val_loss: 1.0046 - val_accuracy: 0.7391\n",
            "Epoch 4/5\n",
            "21/21 [==============================] - 1s 54ms/step - loss: 1.1220 - accuracy: 0.6271 - val_loss: 0.6780 - val_accuracy: 0.8696\n",
            "Epoch 5/5\n",
            "21/21 [==============================] - 1s 53ms/step - loss: 0.9185 - accuracy: 0.6850 - val_loss: 0.3526 - val_accuracy: 0.9043\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 1s 16ms/step - loss: 3.0470 - accuracy: 0.1006 - val_loss: 2.3047 - val_accuracy: 0.1130\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3099 - accuracy: 0.0841 - val_loss: 2.3071 - val_accuracy: 0.1304\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 1s 14ms/step - loss: 2.3077 - accuracy: 0.1074 - val_loss: 2.3049 - val_accuracy: 0.1043\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 1s 14ms/step - loss: 2.3106 - accuracy: 0.0996 - val_loss: 2.3029 - val_accuracy: 0.1304\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3094 - accuracy: 0.0967 - val_loss: 2.3080 - val_accuracy: 0.0696\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 1s 14ms/step - loss: 2.3131 - accuracy: 0.0938 - val_loss: 2.3071 - val_accuracy: 0.0696\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3082 - accuracy: 0.1122 - val_loss: 2.2976 - val_accuracy: 0.1043\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3156 - accuracy: 0.0841 - val_loss: 2.2937 - val_accuracy: 0.1043\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 1s 14ms/step - loss: 2.3110 - accuracy: 0.1074 - val_loss: 2.3146 - val_accuracy: 0.1043\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 1s 14ms/step - loss: 2.3106 - accuracy: 0.1074 - val_loss: 2.3060 - val_accuracy: 0.1043\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3111 - accuracy: 0.1044 - val_loss: 2.3178 - val_accuracy: 0.1043\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 1s 14ms/step - loss: 2.3116 - accuracy: 0.1006 - val_loss: 2.3045 - val_accuracy: 0.0696\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 1s 14ms/step - loss: 2.3106 - accuracy: 0.0977 - val_loss: 2.3053 - val_accuracy: 0.1043\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3131 - accuracy: 0.0870 - val_loss: 2.3041 - val_accuracy: 0.1043\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 1s 14ms/step - loss: 2.3097 - accuracy: 0.0996 - val_loss: 2.3083 - val_accuracy: 0.0696\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 1s 16ms/step - loss: 2.7481 - accuracy: 0.0909 - val_loss: 2.3112 - val_accuracy: 0.1043\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3137 - accuracy: 0.0793 - val_loss: 2.3086 - val_accuracy: 0.1043\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3157 - accuracy: 0.1015 - val_loss: 2.3120 - val_accuracy: 0.0696\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3104 - accuracy: 0.0899 - val_loss: 2.3024 - val_accuracy: 0.1043\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3093 - accuracy: 0.0822 - val_loss: 2.3086 - val_accuracy: 0.0696\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 1s 14ms/step - loss: 2.3106 - accuracy: 0.0938 - val_loss: 2.3032 - val_accuracy: 0.0696\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3131 - accuracy: 0.0909 - val_loss: 2.3147 - val_accuracy: 0.1043\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 1s 14ms/step - loss: 2.3107 - accuracy: 0.0909 - val_loss: 2.3089 - val_accuracy: 0.0696\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3106 - accuracy: 0.0899 - val_loss: 2.3099 - val_accuracy: 0.0957\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3122 - accuracy: 0.0890 - val_loss: 2.3203 - val_accuracy: 0.1130\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3124 - accuracy: 0.0948 - val_loss: 2.3148 - val_accuracy: 0.1043\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3126 - accuracy: 0.0977 - val_loss: 2.3135 - val_accuracy: 0.1043\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 1s 14ms/step - loss: 2.3142 - accuracy: 0.1044 - val_loss: 2.3002 - val_accuracy: 0.1043\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3132 - accuracy: 0.1025 - val_loss: 2.3122 - val_accuracy: 0.0696\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 1s 14ms/step - loss: 2.3104 - accuracy: 0.0909 - val_loss: 2.3143 - val_accuracy: 0.1043\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 1s 17ms/step - loss: 3.2505 - accuracy: 0.0986 - val_loss: 2.3038 - val_accuracy: 0.1304\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3094 - accuracy: 0.0966 - val_loss: 2.2975 - val_accuracy: 0.1304\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3074 - accuracy: 0.1053 - val_loss: 2.3027 - val_accuracy: 0.1043\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 1s 14ms/step - loss: 2.3096 - accuracy: 0.1014 - val_loss: 2.3040 - val_accuracy: 0.1304\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3100 - accuracy: 0.1024 - val_loss: 2.3063 - val_accuracy: 0.0696\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 1s 14ms/step - loss: 2.3103 - accuracy: 0.1024 - val_loss: 2.2986 - val_accuracy: 0.1130\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3109 - accuracy: 0.0986 - val_loss: 2.3064 - val_accuracy: 0.1130\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3101 - accuracy: 0.1005 - val_loss: 2.2969 - val_accuracy: 0.1043\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 1s 14ms/step - loss: 2.3111 - accuracy: 0.1034 - val_loss: 2.3132 - val_accuracy: 0.1130\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 1s 14ms/step - loss: 2.3138 - accuracy: 0.0850 - val_loss: 2.3027 - val_accuracy: 0.0696\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3123 - accuracy: 0.1063 - val_loss: 2.3026 - val_accuracy: 0.1043\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3104 - accuracy: 0.0928 - val_loss: 2.3009 - val_accuracy: 0.1043\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3095 - accuracy: 0.1072 - val_loss: 2.2973 - val_accuracy: 0.1304\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3097 - accuracy: 0.0908 - val_loss: 2.3028 - val_accuracy: 0.0957\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3106 - accuracy: 0.0986 - val_loss: 2.3043 - val_accuracy: 0.0696\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 1s 18ms/step - loss: 2.9317 - accuracy: 0.0860 - val_loss: 2.3080 - val_accuracy: 0.1043\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 1s 14ms/step - loss: 2.3095 - accuracy: 0.0821 - val_loss: 2.2998 - val_accuracy: 0.0957\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3095 - accuracy: 0.0860 - val_loss: 2.3088 - val_accuracy: 0.1304\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3118 - accuracy: 0.0850 - val_loss: 2.3019 - val_accuracy: 0.1043\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3137 - accuracy: 0.0947 - val_loss: 2.3273 - val_accuracy: 0.0957\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3139 - accuracy: 0.0928 - val_loss: 2.3005 - val_accuracy: 0.1130\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3115 - accuracy: 0.0957 - val_loss: 2.3063 - val_accuracy: 0.0957\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 1s 14ms/step - loss: 2.3113 - accuracy: 0.0976 - val_loss: 2.3164 - val_accuracy: 0.0696\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3158 - accuracy: 0.0860 - val_loss: 2.3223 - val_accuracy: 0.0957\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3125 - accuracy: 0.0850 - val_loss: 2.3090 - val_accuracy: 0.1130\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3078 - accuracy: 0.0889 - val_loss: 2.3011 - val_accuracy: 0.0957\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3158 - accuracy: 0.0860 - val_loss: 2.3057 - val_accuracy: 0.0696\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3129 - accuracy: 0.0966 - val_loss: 2.3016 - val_accuracy: 0.0957\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 1s 14ms/step - loss: 2.3110 - accuracy: 0.0860 - val_loss: 2.3156 - val_accuracy: 0.0696\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 1s 14ms/step - loss: 2.3106 - accuracy: 0.0841 - val_loss: 2.3023 - val_accuracy: 0.0696\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 1s 17ms/step - loss: 3.4401 - accuracy: 0.1072 - val_loss: 2.2994 - val_accuracy: 0.0870\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3094 - accuracy: 0.0879 - val_loss: 2.3183 - val_accuracy: 0.0870\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3106 - accuracy: 0.0957 - val_loss: 2.3136 - val_accuracy: 0.0609\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3082 - accuracy: 0.0831 - val_loss: 2.3073 - val_accuracy: 0.1043\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 1s 14ms/step - loss: 2.3084 - accuracy: 0.1024 - val_loss: 2.3120 - val_accuracy: 0.0870\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3089 - accuracy: 0.0947 - val_loss: 2.3119 - val_accuracy: 0.1913\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 1s 14ms/step - loss: 2.3096 - accuracy: 0.0957 - val_loss: 2.3174 - val_accuracy: 0.0783\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3109 - accuracy: 0.0937 - val_loss: 2.3204 - val_accuracy: 0.0870\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3109 - accuracy: 0.0947 - val_loss: 2.2943 - val_accuracy: 0.1043\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3121 - accuracy: 0.0918 - val_loss: 2.3147 - val_accuracy: 0.1304\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3100 - accuracy: 0.1005 - val_loss: 2.3085 - val_accuracy: 0.0957\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 1s 14ms/step - loss: 2.3112 - accuracy: 0.0744 - val_loss: 2.3271 - val_accuracy: 0.0609\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3125 - accuracy: 0.0860 - val_loss: 2.3122 - val_accuracy: 0.0609\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3126 - accuracy: 0.0947 - val_loss: 2.3123 - val_accuracy: 0.1043\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3080 - accuracy: 0.0995 - val_loss: 2.2943 - val_accuracy: 0.1043\n",
            "Epoch 1/10\n",
            "7/7 [==============================] - 1s 198ms/step - loss: 2.2784 - accuracy: 0.2485 - val_loss: 2.2318 - val_accuracy: 0.3391\n",
            "Epoch 2/10\n",
            "7/7 [==============================] - 1s 196ms/step - loss: 2.1538 - accuracy: 0.4294 - val_loss: 2.0492 - val_accuracy: 0.4522\n",
            "Epoch 3/10\n",
            "7/7 [==============================] - 1s 185ms/step - loss: 1.8870 - accuracy: 0.5648 - val_loss: 1.6897 - val_accuracy: 0.6609\n",
            "Epoch 4/10\n",
            "7/7 [==============================] - 1s 190ms/step - loss: 1.4979 - accuracy: 0.6799 - val_loss: 1.2492 - val_accuracy: 0.7565\n",
            "Epoch 5/10\n",
            "7/7 [==============================] - 1s 185ms/step - loss: 1.0962 - accuracy: 0.7718 - val_loss: 0.8994 - val_accuracy: 0.7565\n",
            "Epoch 6/10\n",
            "7/7 [==============================] - 1s 191ms/step - loss: 0.8130 - accuracy: 0.8191 - val_loss: 0.6867 - val_accuracy: 0.8174\n",
            "Epoch 7/10\n",
            "7/7 [==============================] - 1s 184ms/step - loss: 0.6672 - accuracy: 0.8114 - val_loss: 0.5207 - val_accuracy: 0.8348\n",
            "Epoch 8/10\n",
            "7/7 [==============================] - 1s 192ms/step - loss: 0.5050 - accuracy: 0.8598 - val_loss: 0.3870 - val_accuracy: 0.9043\n",
            "Epoch 9/10\n",
            "7/7 [==============================] - 1s 183ms/step - loss: 0.4018 - accuracy: 0.8956 - val_loss: 0.3353 - val_accuracy: 0.9130\n",
            "Epoch 10/10\n",
            "7/7 [==============================] - 1s 192ms/step - loss: 0.3446 - accuracy: 0.8975 - val_loss: 0.2726 - val_accuracy: 0.9304\n",
            "Epoch 1/10\n",
            "7/7 [==============================] - 1s 201ms/step - loss: 2.2806 - accuracy: 0.3598 - val_loss: 2.2325 - val_accuracy: 0.4870\n",
            "Epoch 2/10\n",
            "7/7 [==============================] - 1s 192ms/step - loss: 2.1639 - accuracy: 0.5880 - val_loss: 2.0480 - val_accuracy: 0.6522\n",
            "Epoch 3/10\n",
            "7/7 [==============================] - 1s 186ms/step - loss: 1.9120 - accuracy: 0.6112 - val_loss: 1.7000 - val_accuracy: 0.7478\n",
            "Epoch 4/10\n",
            "7/7 [==============================] - 1s 190ms/step - loss: 1.5016 - accuracy: 0.7485 - val_loss: 1.2318 - val_accuracy: 0.8696\n",
            "Epoch 5/10\n",
            "7/7 [==============================] - 1s 184ms/step - loss: 1.0686 - accuracy: 0.8250 - val_loss: 0.8093 - val_accuracy: 0.8435\n",
            "Epoch 6/10\n",
            "7/7 [==============================] - 1s 184ms/step - loss: 0.7430 - accuracy: 0.8482 - val_loss: 0.5945 - val_accuracy: 0.8783\n",
            "Epoch 7/10\n",
            "7/7 [==============================] - 1s 184ms/step - loss: 0.5566 - accuracy: 0.8540 - val_loss: 0.4438 - val_accuracy: 0.8783\n",
            "Epoch 8/10\n",
            "7/7 [==============================] - 1s 184ms/step - loss: 0.4328 - accuracy: 0.8820 - val_loss: 0.3444 - val_accuracy: 0.9304\n",
            "Epoch 9/10\n",
            "7/7 [==============================] - 1s 193ms/step - loss: 0.3475 - accuracy: 0.8994 - val_loss: 0.2972 - val_accuracy: 0.9304\n",
            "Epoch 10/10\n",
            "7/7 [==============================] - 1s 186ms/step - loss: 0.3030 - accuracy: 0.9120 - val_loss: 0.2588 - val_accuracy: 0.9391\n",
            "Epoch 1/10\n",
            "7/7 [==============================] - 1s 210ms/step - loss: 2.2853 - accuracy: 0.2464 - val_loss: 2.2380 - val_accuracy: 0.6783\n",
            "Epoch 2/10\n",
            "7/7 [==============================] - 1s 181ms/step - loss: 2.1781 - accuracy: 0.6271 - val_loss: 2.0560 - val_accuracy: 0.6435\n",
            "Epoch 3/10\n",
            "7/7 [==============================] - 1s 192ms/step - loss: 1.9210 - accuracy: 0.6386 - val_loss: 1.6899 - val_accuracy: 0.7217\n",
            "Epoch 4/10\n",
            "7/7 [==============================] - 1s 185ms/step - loss: 1.5145 - accuracy: 0.7275 - val_loss: 1.2581 - val_accuracy: 0.8000\n",
            "Epoch 5/10\n",
            "7/7 [==============================] - 1s 195ms/step - loss: 1.1169 - accuracy: 0.7739 - val_loss: 0.8515 - val_accuracy: 0.8522\n",
            "Epoch 6/10\n",
            "7/7 [==============================] - 1s 186ms/step - loss: 0.7761 - accuracy: 0.8541 - val_loss: 0.6089 - val_accuracy: 0.8609\n",
            "Epoch 7/10\n",
            "7/7 [==============================] - 1s 187ms/step - loss: 0.5982 - accuracy: 0.8444 - val_loss: 0.5054 - val_accuracy: 0.8783\n",
            "Epoch 8/10\n",
            "7/7 [==============================] - 1s 183ms/step - loss: 0.4485 - accuracy: 0.8957 - val_loss: 0.3775 - val_accuracy: 0.9217\n",
            "Epoch 9/10\n",
            "7/7 [==============================] - 1s 189ms/step - loss: 0.3892 - accuracy: 0.9005 - val_loss: 0.3211 - val_accuracy: 0.8957\n",
            "Epoch 10/10\n",
            "7/7 [==============================] - 1s 185ms/step - loss: 0.3316 - accuracy: 0.9043 - val_loss: 0.2829 - val_accuracy: 0.9478\n",
            "Epoch 1/10\n",
            "7/7 [==============================] - 1s 200ms/step - loss: 2.2789 - accuracy: 0.3092 - val_loss: 2.2268 - val_accuracy: 0.5217\n",
            "Epoch 2/10\n",
            "7/7 [==============================] - 1s 183ms/step - loss: 2.1597 - accuracy: 0.5971 - val_loss: 2.0367 - val_accuracy: 0.6087\n",
            "Epoch 3/10\n",
            "7/7 [==============================] - 1s 184ms/step - loss: 1.8916 - accuracy: 0.6329 - val_loss: 1.6737 - val_accuracy: 0.7043\n",
            "Epoch 4/10\n",
            "7/7 [==============================] - 1s 185ms/step - loss: 1.4962 - accuracy: 0.6628 - val_loss: 1.2070 - val_accuracy: 0.7652\n",
            "Epoch 5/10\n",
            "7/7 [==============================] - 1s 182ms/step - loss: 1.0632 - accuracy: 0.8106 - val_loss: 0.8340 - val_accuracy: 0.8261\n",
            "Epoch 6/10\n",
            "7/7 [==============================] - 1s 185ms/step - loss: 0.7596 - accuracy: 0.8203 - val_loss: 0.5949 - val_accuracy: 0.8609\n",
            "Epoch 7/10\n",
            "7/7 [==============================] - 1s 186ms/step - loss: 0.5923 - accuracy: 0.8589 - val_loss: 0.4902 - val_accuracy: 0.8783\n",
            "Epoch 8/10\n",
            "7/7 [==============================] - 1s 185ms/step - loss: 0.4658 - accuracy: 0.8734 - val_loss: 0.3993 - val_accuracy: 0.9043\n",
            "Epoch 9/10\n",
            "7/7 [==============================] - 1s 184ms/step - loss: 0.3937 - accuracy: 0.8870 - val_loss: 0.3547 - val_accuracy: 0.9043\n",
            "Epoch 10/10\n",
            "7/7 [==============================] - 1s 186ms/step - loss: 0.3351 - accuracy: 0.9101 - val_loss: 0.3268 - val_accuracy: 0.9304\n",
            "WARNING:tensorflow:5 out of the last 22 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd0cb39bf8> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/10\n",
            "7/7 [==============================] - 1s 200ms/step - loss: 2.2829 - accuracy: 0.3063 - val_loss: 2.2432 - val_accuracy: 0.4609\n",
            "Epoch 2/10\n",
            "7/7 [==============================] - 1s 189ms/step - loss: 2.1740 - accuracy: 0.4957 - val_loss: 2.0964 - val_accuracy: 0.4348\n",
            "Epoch 3/10\n",
            "7/7 [==============================] - 1s 187ms/step - loss: 1.9265 - accuracy: 0.5749 - val_loss: 1.7734 - val_accuracy: 0.6696\n",
            "Epoch 4/10\n",
            "7/7 [==============================] - 1s 184ms/step - loss: 1.5284 - accuracy: 0.6754 - val_loss: 1.3305 - val_accuracy: 0.7652\n",
            "Epoch 5/10\n",
            "7/7 [==============================] - 1s 193ms/step - loss: 1.1164 - accuracy: 0.7633 - val_loss: 0.9685 - val_accuracy: 0.8609\n",
            "Epoch 6/10\n",
            "7/7 [==============================] - 1s 186ms/step - loss: 0.8130 - accuracy: 0.8097 - val_loss: 0.7664 - val_accuracy: 0.8522\n",
            "Epoch 7/10\n",
            "7/7 [==============================] - 1s 194ms/step - loss: 0.6206 - accuracy: 0.8280 - val_loss: 0.5371 - val_accuracy: 0.8957\n",
            "Epoch 8/10\n",
            "7/7 [==============================] - 1s 186ms/step - loss: 0.4930 - accuracy: 0.8744 - val_loss: 0.4149 - val_accuracy: 0.9130\n",
            "Epoch 9/10\n",
            "7/7 [==============================] - 1s 194ms/step - loss: 0.4153 - accuracy: 0.8783 - val_loss: 0.3795 - val_accuracy: 0.9304\n",
            "Epoch 10/10\n",
            "7/7 [==============================] - 1s 185ms/step - loss: 0.3598 - accuracy: 0.8928 - val_loss: 0.2719 - val_accuracy: 0.9478\n",
            "WARNING:tensorflow:6 out of the last 24 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd0ae568c8> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 2s 47ms/step - loss: 8.3584 - accuracy: 0.1015 - val_loss: 2.3130 - val_accuracy: 0.1043\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 2s 46ms/step - loss: 2.3382 - accuracy: 0.0938 - val_loss: 2.3051 - val_accuracy: 0.1043\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 2s 46ms/step - loss: 2.3080 - accuracy: 0.1141 - val_loss: 2.3007 - val_accuracy: 0.1043\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 2s 47ms/step - loss: 2.3069 - accuracy: 0.0986 - val_loss: 2.3128 - val_accuracy: 0.1043\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 3s 48ms/step - loss: 2.3080 - accuracy: 0.0996 - val_loss: 2.3044 - val_accuracy: 0.1043\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 2s 46ms/step - loss: 2.3070 - accuracy: 0.1141 - val_loss: 2.3086 - val_accuracy: 0.1043\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 2s 46ms/step - loss: 2.3080 - accuracy: 0.0919 - val_loss: 2.3109 - val_accuracy: 0.1043\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 2s 47ms/step - loss: 2.3102 - accuracy: 0.0996 - val_loss: 2.3093 - val_accuracy: 0.1043\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 2s 45ms/step - loss: 2.3096 - accuracy: 0.0899 - val_loss: 2.3193 - val_accuracy: 0.1043\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 3s 48ms/step - loss: 2.3086 - accuracy: 0.1006 - val_loss: 2.3064 - val_accuracy: 0.1304\n",
            "WARNING:tensorflow:6 out of the last 11 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd064d1378> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 3s 48ms/step - loss: 4.9397 - accuracy: 0.1074 - val_loss: 2.3050 - val_accuracy: 0.0696\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 2s 46ms/step - loss: 2.3115 - accuracy: 0.0841 - val_loss: 2.3063 - val_accuracy: 0.0696\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 2s 46ms/step - loss: 2.3113 - accuracy: 0.0967 - val_loss: 2.3027 - val_accuracy: 0.1304\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 2s 46ms/step - loss: 2.3120 - accuracy: 0.0909 - val_loss: 2.3086 - val_accuracy: 0.1043\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 2s 46ms/step - loss: 2.3130 - accuracy: 0.0803 - val_loss: 2.3108 - val_accuracy: 0.1304\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 2s 45ms/step - loss: 2.3095 - accuracy: 0.1103 - val_loss: 2.3091 - val_accuracy: 0.1043\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 2s 46ms/step - loss: 2.3110 - accuracy: 0.0928 - val_loss: 2.3095 - val_accuracy: 0.1043\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 2s 46ms/step - loss: 2.3101 - accuracy: 0.1015 - val_loss: 2.3074 - val_accuracy: 0.1043\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 2s 46ms/step - loss: 2.3092 - accuracy: 0.0977 - val_loss: 2.3033 - val_accuracy: 0.1304\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 2s 46ms/step - loss: 2.3096 - accuracy: 0.0938 - val_loss: 2.3109 - val_accuracy: 0.1043\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 2s 47ms/step - loss: 7.3745 - accuracy: 0.1005 - val_loss: 2.3050 - val_accuracy: 0.1043\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 2s 46ms/step - loss: 2.3073 - accuracy: 0.0899 - val_loss: 2.2961 - val_accuracy: 0.1304\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 2s 46ms/step - loss: 2.3100 - accuracy: 0.0947 - val_loss: 2.2996 - val_accuracy: 0.1304\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 3s 48ms/step - loss: 2.3078 - accuracy: 0.0918 - val_loss: 2.2986 - val_accuracy: 0.1043\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 2s 48ms/step - loss: 2.3100 - accuracy: 0.0937 - val_loss: 2.3004 - val_accuracy: 0.1130\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 3s 48ms/step - loss: 2.3123 - accuracy: 0.0986 - val_loss: 2.3033 - val_accuracy: 0.1130\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 2s 47ms/step - loss: 2.3103 - accuracy: 0.0879 - val_loss: 2.2964 - val_accuracy: 0.1304\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 3s 50ms/step - loss: 2.3099 - accuracy: 0.0889 - val_loss: 2.2949 - val_accuracy: 0.1304\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 2s 47ms/step - loss: 2.3102 - accuracy: 0.1111 - val_loss: 2.3054 - val_accuracy: 0.1043\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 2s 46ms/step - loss: 2.3119 - accuracy: 0.0908 - val_loss: 2.3009 - val_accuracy: 0.1043\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 2s 48ms/step - loss: 4.8829 - accuracy: 0.0986 - val_loss: 2.3154 - val_accuracy: 0.0696\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 2s 46ms/step - loss: 2.3100 - accuracy: 0.0899 - val_loss: 2.3061 - val_accuracy: 0.0696\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 2s 44ms/step - loss: 2.3092 - accuracy: 0.0966 - val_loss: 2.3122 - val_accuracy: 0.0696\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 2s 44ms/step - loss: 2.3091 - accuracy: 0.0792 - val_loss: 2.3079 - val_accuracy: 0.0696\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 2s 46ms/step - loss: 2.3133 - accuracy: 0.0850 - val_loss: 2.3049 - val_accuracy: 0.0696\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 2s 47ms/step - loss: 2.3119 - accuracy: 0.0879 - val_loss: 2.3021 - val_accuracy: 0.1304\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 2s 47ms/step - loss: 2.3132 - accuracy: 0.0947 - val_loss: 2.3059 - val_accuracy: 0.1043\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 2s 46ms/step - loss: 2.3109 - accuracy: 0.0850 - val_loss: 2.3093 - val_accuracy: 0.0696\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 2s 45ms/step - loss: 2.3117 - accuracy: 0.0928 - val_loss: 2.3085 - val_accuracy: 0.0696\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 2s 45ms/step - loss: 2.3135 - accuracy: 0.0966 - val_loss: 2.3066 - val_accuracy: 0.1304\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 3s 49ms/step - loss: 4.7020 - accuracy: 0.1063 - val_loss: 2.3119 - val_accuracy: 0.1043\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 2s 45ms/step - loss: 2.6464 - accuracy: 0.0908 - val_loss: 2.3094 - val_accuracy: 0.0609\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 2s 47ms/step - loss: 2.5083 - accuracy: 0.0918 - val_loss: 2.3123 - val_accuracy: 0.0348\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 2s 47ms/step - loss: 2.5049 - accuracy: 0.0957 - val_loss: 2.3121 - val_accuracy: 0.0957\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 3s 50ms/step - loss: 2.3357 - accuracy: 0.0860 - val_loss: 2.3019 - val_accuracy: 0.0870\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 3s 51ms/step - loss: 2.3094 - accuracy: 0.0918 - val_loss: 2.2938 - val_accuracy: 0.1913\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 2s 46ms/step - loss: 2.3100 - accuracy: 0.0937 - val_loss: 2.3120 - val_accuracy: 0.1043\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 2s 45ms/step - loss: 2.3108 - accuracy: 0.1024 - val_loss: 2.3059 - val_accuracy: 0.1043\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 2s 46ms/step - loss: 2.3148 - accuracy: 0.1005 - val_loss: 2.3129 - val_accuracy: 0.0870\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 2s 48ms/step - loss: 2.3120 - accuracy: 0.0947 - val_loss: 2.3214 - val_accuracy: 0.1043\n",
            "Epoch 1/5\n",
            "21/21 [==============================] - 3s 135ms/step - loss: 2.2525 - accuracy: 0.2273 - val_loss: 2.1170 - val_accuracy: 0.3304\n",
            "Epoch 2/5\n",
            "21/21 [==============================] - 3s 133ms/step - loss: 1.8452 - accuracy: 0.3965 - val_loss: 1.3647 - val_accuracy: 0.7130\n",
            "Epoch 3/5\n",
            "21/21 [==============================] - 3s 128ms/step - loss: 1.1404 - accuracy: 0.6838 - val_loss: 0.7312 - val_accuracy: 0.8435\n",
            "Epoch 4/5\n",
            "21/21 [==============================] - 3s 127ms/step - loss: 0.6905 - accuracy: 0.8095 - val_loss: 0.4766 - val_accuracy: 0.8783\n",
            "Epoch 5/5\n",
            "21/21 [==============================] - 3s 129ms/step - loss: 0.5076 - accuracy: 0.8549 - val_loss: 0.3673 - val_accuracy: 0.9043\n",
            "Epoch 1/5\n",
            "21/21 [==============================] - 3s 132ms/step - loss: 2.2534 - accuracy: 0.2650 - val_loss: 2.1302 - val_accuracy: 0.4174\n",
            "Epoch 2/5\n",
            "21/21 [==============================] - 3s 129ms/step - loss: 1.8140 - accuracy: 0.4836 - val_loss: 1.2769 - val_accuracy: 0.7826\n",
            "Epoch 3/5\n",
            "21/21 [==============================] - 3s 128ms/step - loss: 1.0433 - accuracy: 0.7331 - val_loss: 0.6604 - val_accuracy: 0.8261\n",
            "Epoch 4/5\n",
            "21/21 [==============================] - 3s 131ms/step - loss: 0.6601 - accuracy: 0.7921 - val_loss: 0.4590 - val_accuracy: 0.8696\n",
            "Epoch 5/5\n",
            "21/21 [==============================] - 3s 127ms/step - loss: 0.4953 - accuracy: 0.8482 - val_loss: 0.3678 - val_accuracy: 0.8870\n",
            "Epoch 1/5\n",
            "21/21 [==============================] - 3s 132ms/step - loss: 2.2461 - accuracy: 0.2754 - val_loss: 2.0722 - val_accuracy: 0.6000\n",
            "Epoch 2/5\n",
            "21/21 [==============================] - 3s 127ms/step - loss: 1.7505 - accuracy: 0.5826 - val_loss: 1.2389 - val_accuracy: 0.7478\n",
            "Epoch 3/5\n",
            "21/21 [==============================] - 3s 128ms/step - loss: 0.9839 - accuracy: 0.7217 - val_loss: 0.6377 - val_accuracy: 0.8174\n",
            "Epoch 4/5\n",
            "21/21 [==============================] - 3s 126ms/step - loss: 0.6519 - accuracy: 0.8010 - val_loss: 0.4747 - val_accuracy: 0.8522\n",
            "Epoch 5/5\n",
            "21/21 [==============================] - 3s 129ms/step - loss: 0.4891 - accuracy: 0.8580 - val_loss: 0.3607 - val_accuracy: 0.8783\n",
            "Epoch 1/5\n",
            "21/21 [==============================] - 3s 135ms/step - loss: 2.2514 - accuracy: 0.2077 - val_loss: 2.1277 - val_accuracy: 0.3130\n",
            "Epoch 2/5\n",
            "21/21 [==============================] - 3s 127ms/step - loss: 1.8094 - accuracy: 0.4966 - val_loss: 1.3350 - val_accuracy: 0.8000\n",
            "Epoch 3/5\n",
            "21/21 [==============================] - 3s 127ms/step - loss: 1.1217 - accuracy: 0.6879 - val_loss: 0.7744 - val_accuracy: 0.8261\n",
            "Epoch 4/5\n",
            "21/21 [==============================] - 3s 128ms/step - loss: 0.7522 - accuracy: 0.7787 - val_loss: 0.5423 - val_accuracy: 0.8435\n",
            "Epoch 5/5\n",
            "21/21 [==============================] - 3s 131ms/step - loss: 0.5311 - accuracy: 0.8425 - val_loss: 0.4496 - val_accuracy: 0.8609\n",
            "Epoch 1/5\n",
            "21/21 [==============================] - 3s 137ms/step - loss: 2.2507 - accuracy: 0.2715 - val_loss: 2.1015 - val_accuracy: 0.5304\n",
            "Epoch 2/5\n",
            "21/21 [==============================] - 3s 131ms/step - loss: 1.7825 - accuracy: 0.4986 - val_loss: 1.3040 - val_accuracy: 0.7739\n",
            "Epoch 3/5\n",
            "21/21 [==============================] - 3s 131ms/step - loss: 1.0870 - accuracy: 0.6734 - val_loss: 0.7980 - val_accuracy: 0.8435\n",
            "Epoch 4/5\n",
            "21/21 [==============================] - 3s 129ms/step - loss: 0.7111 - accuracy: 0.7913 - val_loss: 0.4300 - val_accuracy: 0.9043\n",
            "Epoch 5/5\n",
            "21/21 [==============================] - 3s 129ms/step - loss: 0.4949 - accuracy: 0.8541 - val_loss: 0.3079 - val_accuracy: 0.9391\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 1s 16ms/step - loss: 2.4562 - accuracy: 0.0841 - val_loss: 2.3019 - val_accuracy: 0.1043\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 1s 13ms/step - loss: 2.3081 - accuracy: 0.1006 - val_loss: 2.3077 - val_accuracy: 0.0696\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 1s 12ms/step - loss: 2.3084 - accuracy: 0.1083 - val_loss: 2.3096 - val_accuracy: 0.1043\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 1s 14ms/step - loss: 2.3107 - accuracy: 0.0948 - val_loss: 2.3112 - val_accuracy: 0.1043\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 1s 13ms/step - loss: 2.3097 - accuracy: 0.0919 - val_loss: 2.3046 - val_accuracy: 0.1043\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 1s 13ms/step - loss: 2.3097 - accuracy: 0.1122 - val_loss: 2.3138 - val_accuracy: 0.1043\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 1s 12ms/step - loss: 2.3106 - accuracy: 0.0832 - val_loss: 2.3173 - val_accuracy: 0.0696\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 1s 13ms/step - loss: 2.3115 - accuracy: 0.1025 - val_loss: 2.3166 - val_accuracy: 0.1043\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 1s 13ms/step - loss: 2.3106 - accuracy: 0.1074 - val_loss: 2.3090 - val_accuracy: 0.1043\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 1s 12ms/step - loss: 2.3117 - accuracy: 0.1035 - val_loss: 2.3128 - val_accuracy: 0.1043\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 1s 16ms/step - loss: 2.8549 - accuracy: 0.0919 - val_loss: 2.3059 - val_accuracy: 0.0783\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 1s 13ms/step - loss: 2.3108 - accuracy: 0.1025 - val_loss: 2.3182 - val_accuracy: 0.0696\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 1s 13ms/step - loss: 2.3064 - accuracy: 0.0890 - val_loss: 2.3089 - val_accuracy: 0.1217\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 1s 13ms/step - loss: 2.3137 - accuracy: 0.1025 - val_loss: 2.3062 - val_accuracy: 0.1043\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 1s 13ms/step - loss: 2.3116 - accuracy: 0.0928 - val_loss: 2.3092 - val_accuracy: 0.1304\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 1s 13ms/step - loss: 2.3128 - accuracy: 0.0899 - val_loss: 2.3061 - val_accuracy: 0.0957\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 1s 13ms/step - loss: 2.3119 - accuracy: 0.1015 - val_loss: 2.2995 - val_accuracy: 0.1304\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 1s 13ms/step - loss: 2.3126 - accuracy: 0.0861 - val_loss: 2.3081 - val_accuracy: 0.1130\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 1s 12ms/step - loss: 2.3118 - accuracy: 0.0996 - val_loss: 2.3043 - val_accuracy: 0.1217\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 1s 12ms/step - loss: 2.3147 - accuracy: 0.0890 - val_loss: 2.3051 - val_accuracy: 0.1043\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 1s 14ms/step - loss: 2.6149 - accuracy: 0.1024 - val_loss: 2.3087 - val_accuracy: 0.0696\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 1s 13ms/step - loss: 2.3084 - accuracy: 0.1014 - val_loss: 2.2962 - val_accuracy: 0.1043\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 1s 13ms/step - loss: 2.3138 - accuracy: 0.1005 - val_loss: 2.3033 - val_accuracy: 0.0696\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 1s 13ms/step - loss: 2.3099 - accuracy: 0.0976 - val_loss: 2.3022 - val_accuracy: 0.1304\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 1s 13ms/step - loss: 2.3106 - accuracy: 0.1005 - val_loss: 2.2953 - val_accuracy: 0.1043\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 1s 13ms/step - loss: 2.3117 - accuracy: 0.0957 - val_loss: 2.3068 - val_accuracy: 0.0696\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 1s 13ms/step - loss: 2.3104 - accuracy: 0.0957 - val_loss: 2.3098 - val_accuracy: 0.0957\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 1s 14ms/step - loss: 2.3096 - accuracy: 0.1034 - val_loss: 2.2996 - val_accuracy: 0.1130\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 1s 12ms/step - loss: 2.3112 - accuracy: 0.1043 - val_loss: 2.2979 - val_accuracy: 0.1043\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 1s 13ms/step - loss: 2.3094 - accuracy: 0.1024 - val_loss: 2.3062 - val_accuracy: 0.1043\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.3930 - accuracy: 0.0966 - val_loss: 2.2986 - val_accuracy: 0.1217\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 1s 12ms/step - loss: 2.3124 - accuracy: 0.0918 - val_loss: 2.3102 - val_accuracy: 0.0957\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 1s 13ms/step - loss: 2.3099 - accuracy: 0.0899 - val_loss: 2.3068 - val_accuracy: 0.0696\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 1s 13ms/step - loss: 2.3127 - accuracy: 0.0870 - val_loss: 2.3081 - val_accuracy: 0.0696\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 1s 12ms/step - loss: 2.3121 - accuracy: 0.0792 - val_loss: 2.3084 - val_accuracy: 0.0696\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 1s 13ms/step - loss: 2.3128 - accuracy: 0.0889 - val_loss: 2.3164 - val_accuracy: 0.1304\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 1s 13ms/step - loss: 2.3135 - accuracy: 0.0850 - val_loss: 2.3028 - val_accuracy: 0.1130\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 1s 12ms/step - loss: 2.3122 - accuracy: 0.0889 - val_loss: 2.3183 - val_accuracy: 0.0696\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 1s 14ms/step - loss: 2.3131 - accuracy: 0.1005 - val_loss: 2.2980 - val_accuracy: 0.0957\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 1s 13ms/step - loss: 2.3143 - accuracy: 0.0908 - val_loss: 2.3065 - val_accuracy: 0.1043\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 1s 15ms/step - loss: 2.7118 - accuracy: 0.1024 - val_loss: 2.3081 - val_accuracy: 0.1043\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 1s 12ms/step - loss: 2.3119 - accuracy: 0.0947 - val_loss: 2.2907 - val_accuracy: 0.1913\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 1s 12ms/step - loss: 2.3146 - accuracy: 0.0986 - val_loss: 2.3038 - val_accuracy: 0.1913\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 1s 13ms/step - loss: 2.3115 - accuracy: 0.0899 - val_loss: 2.3165 - val_accuracy: 0.0609\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 1s 13ms/step - loss: 2.3086 - accuracy: 0.0879 - val_loss: 2.3238 - val_accuracy: 0.0348\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 1s 12ms/step - loss: 2.3093 - accuracy: 0.0995 - val_loss: 2.3040 - val_accuracy: 0.1304\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 1s 13ms/step - loss: 2.3118 - accuracy: 0.0908 - val_loss: 2.2980 - val_accuracy: 0.0870\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 1s 13ms/step - loss: 2.3101 - accuracy: 0.0860 - val_loss: 2.3116 - val_accuracy: 0.0783\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 1s 12ms/step - loss: 2.3120 - accuracy: 0.0918 - val_loss: 2.2983 - val_accuracy: 0.1043\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 1s 14ms/step - loss: 2.3107 - accuracy: 0.1024 - val_loss: 2.3154 - val_accuracy: 0.0870\n",
            "Epoch 1/10\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 2.1515 - accuracy: 0.2437 - val_loss: 1.4039 - val_accuracy: 0.6783\n",
            "Epoch 2/10\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.8864 - accuracy: 0.7089 - val_loss: 0.3205 - val_accuracy: 0.8957\n",
            "Epoch 3/10\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.4773 - accuracy: 0.8501 - val_loss: 0.2379 - val_accuracy: 0.9391\n",
            "Epoch 4/10\n",
            "21/21 [==============================] - 0s 17ms/step - loss: 0.3690 - accuracy: 0.8791 - val_loss: 0.1376 - val_accuracy: 0.9565\n",
            "Epoch 5/10\n",
            "21/21 [==============================] - 0s 14ms/step - loss: 0.3386 - accuracy: 0.8868 - val_loss: 0.1904 - val_accuracy: 0.9304\n",
            "Epoch 6/10\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.2905 - accuracy: 0.9052 - val_loss: 0.2375 - val_accuracy: 0.9217\n",
            "Epoch 7/10\n",
            "21/21 [==============================] - 0s 14ms/step - loss: 0.2284 - accuracy: 0.9313 - val_loss: 0.1480 - val_accuracy: 0.9565\n",
            "Epoch 8/10\n",
            "21/21 [==============================] - 0s 14ms/step - loss: 0.1825 - accuracy: 0.9400 - val_loss: 0.2579 - val_accuracy: 0.9130\n",
            "Epoch 9/10\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.1738 - accuracy: 0.9429 - val_loss: 0.1779 - val_accuracy: 0.9652\n",
            "Epoch 10/10\n",
            "21/21 [==============================] - 0s 14ms/step - loss: 0.1828 - accuracy: 0.9333 - val_loss: 0.0954 - val_accuracy: 0.9652\n",
            "Epoch 1/10\n",
            "21/21 [==============================] - 0s 20ms/step - loss: 2.1612 - accuracy: 0.2263 - val_loss: 1.4274 - val_accuracy: 0.7304\n",
            "Epoch 2/10\n",
            "21/21 [==============================] - 0s 17ms/step - loss: 1.0117 - accuracy: 0.6538 - val_loss: 0.4056 - val_accuracy: 0.9130\n",
            "Epoch 3/10\n",
            "21/21 [==============================] - 0s 14ms/step - loss: 0.5524 - accuracy: 0.8298 - val_loss: 0.2109 - val_accuracy: 0.9565\n",
            "Epoch 4/10\n",
            "21/21 [==============================] - 0s 14ms/step - loss: 0.3441 - accuracy: 0.8936 - val_loss: 0.1967 - val_accuracy: 0.9478\n",
            "Epoch 5/10\n",
            "21/21 [==============================] - 0s 13ms/step - loss: 0.3143 - accuracy: 0.8975 - val_loss: 0.1216 - val_accuracy: 0.9739\n",
            "Epoch 6/10\n",
            "21/21 [==============================] - 0s 14ms/step - loss: 0.2483 - accuracy: 0.9178 - val_loss: 0.1092 - val_accuracy: 0.9739\n",
            "Epoch 7/10\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.2386 - accuracy: 0.9265 - val_loss: 0.1184 - val_accuracy: 0.9826\n",
            "Epoch 8/10\n",
            "21/21 [==============================] - 0s 16ms/step - loss: 0.2135 - accuracy: 0.9333 - val_loss: 0.1712 - val_accuracy: 0.9565\n",
            "Epoch 9/10\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.2136 - accuracy: 0.9391 - val_loss: 0.1044 - val_accuracy: 0.9739\n",
            "Epoch 10/10\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.1468 - accuracy: 0.9478 - val_loss: 0.1012 - val_accuracy: 0.9826\n",
            "Epoch 1/10\n",
            "21/21 [==============================] - 0s 18ms/step - loss: 2.2228 - accuracy: 0.1797 - val_loss: 1.7027 - val_accuracy: 0.5652\n",
            "Epoch 2/10\n",
            "21/21 [==============================] - 0s 14ms/step - loss: 1.2762 - accuracy: 0.5614 - val_loss: 0.5197 - val_accuracy: 0.8261\n",
            "Epoch 3/10\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.6252 - accuracy: 0.8135 - val_loss: 0.3534 - val_accuracy: 0.8870\n",
            "Epoch 4/10\n",
            "21/21 [==============================] - 0s 14ms/step - loss: 0.4391 - accuracy: 0.8522 - val_loss: 0.2215 - val_accuracy: 0.9391\n",
            "Epoch 5/10\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.3599 - accuracy: 0.8850 - val_loss: 0.1750 - val_accuracy: 0.9565\n",
            "Epoch 6/10\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.2696 - accuracy: 0.9053 - val_loss: 0.1560 - val_accuracy: 0.9565\n",
            "Epoch 7/10\n",
            "21/21 [==============================] - 0s 14ms/step - loss: 0.2746 - accuracy: 0.9198 - val_loss: 0.1259 - val_accuracy: 0.9739\n",
            "Epoch 8/10\n",
            "21/21 [==============================] - 0s 16ms/step - loss: 0.2285 - accuracy: 0.9208 - val_loss: 0.1089 - val_accuracy: 0.9652\n",
            "Epoch 9/10\n",
            "21/21 [==============================] - 0s 14ms/step - loss: 0.2021 - accuracy: 0.9430 - val_loss: 0.0915 - val_accuracy: 0.9739\n",
            "Epoch 10/10\n",
            "21/21 [==============================] - 0s 14ms/step - loss: 0.1816 - accuracy: 0.9362 - val_loss: 0.0624 - val_accuracy: 0.9826\n",
            "Epoch 1/10\n",
            "21/21 [==============================] - 0s 18ms/step - loss: 2.1867 - accuracy: 0.2174 - val_loss: 1.5879 - val_accuracy: 0.5652\n",
            "Epoch 2/10\n",
            "21/21 [==============================] - 0s 14ms/step - loss: 1.0425 - accuracy: 0.6396 - val_loss: 0.6006 - val_accuracy: 0.8435\n",
            "Epoch 3/10\n",
            "21/21 [==============================] - 0s 16ms/step - loss: 0.6000 - accuracy: 0.8048 - val_loss: 0.4306 - val_accuracy: 0.8783\n",
            "Epoch 4/10\n",
            "21/21 [==============================] - 0s 16ms/step - loss: 0.4596 - accuracy: 0.8425 - val_loss: 0.2731 - val_accuracy: 0.9130\n",
            "Epoch 5/10\n",
            "21/21 [==============================] - 0s 14ms/step - loss: 0.3455 - accuracy: 0.8870 - val_loss: 0.2020 - val_accuracy: 0.9217\n",
            "Epoch 6/10\n",
            "21/21 [==============================] - 0s 14ms/step - loss: 0.2713 - accuracy: 0.9130 - val_loss: 0.1607 - val_accuracy: 0.9652\n",
            "Epoch 7/10\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.2201 - accuracy: 0.9275 - val_loss: 0.1401 - val_accuracy: 0.9652\n",
            "Epoch 8/10\n",
            "21/21 [==============================] - 0s 16ms/step - loss: 0.1968 - accuracy: 0.9333 - val_loss: 0.1739 - val_accuracy: 0.9478\n",
            "Epoch 9/10\n",
            "21/21 [==============================] - 0s 13ms/step - loss: 0.2037 - accuracy: 0.9353 - val_loss: 0.1362 - val_accuracy: 0.9565\n",
            "Epoch 10/10\n",
            "21/21 [==============================] - 0s 16ms/step - loss: 0.2179 - accuracy: 0.9295 - val_loss: 0.1441 - val_accuracy: 0.9739\n",
            "Epoch 1/10\n",
            "21/21 [==============================] - 1s 44ms/step - loss: 2.1443 - accuracy: 0.2348 - val_loss: 1.3069 - val_accuracy: 0.7304\n",
            "Epoch 2/10\n",
            "21/21 [==============================] - 0s 13ms/step - loss: 0.8872 - accuracy: 0.7140 - val_loss: 0.3682 - val_accuracy: 0.8957\n",
            "Epoch 3/10\n",
            "21/21 [==============================] - 0s 16ms/step - loss: 0.5453 - accuracy: 0.8106 - val_loss: 0.3038 - val_accuracy: 0.9304\n",
            "Epoch 4/10\n",
            "21/21 [==============================] - 0s 14ms/step - loss: 0.4731 - accuracy: 0.8570 - val_loss: 0.1586 - val_accuracy: 0.9565\n",
            "Epoch 5/10\n",
            "21/21 [==============================] - 0s 16ms/step - loss: 0.2984 - accuracy: 0.9005 - val_loss: 0.1340 - val_accuracy: 0.9739\n",
            "Epoch 6/10\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.2700 - accuracy: 0.9179 - val_loss: 0.0728 - val_accuracy: 0.9913\n",
            "Epoch 7/10\n",
            "21/21 [==============================] - 0s 13ms/step - loss: 0.3066 - accuracy: 0.9014 - val_loss: 0.1560 - val_accuracy: 0.9478\n",
            "Epoch 8/10\n",
            "21/21 [==============================] - 0s 14ms/step - loss: 0.2317 - accuracy: 0.9198 - val_loss: 0.0447 - val_accuracy: 1.0000\n",
            "Epoch 9/10\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 0.2206 - accuracy: 0.9208 - val_loss: 0.0577 - val_accuracy: 0.9913\n",
            "Epoch 10/10\n",
            "21/21 [==============================] - 0s 13ms/step - loss: 0.1725 - accuracy: 0.9411 - val_loss: 0.0391 - val_accuracy: 1.0000\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 1s 10ms/step - loss: 2.2852 - accuracy: 0.1634 - val_loss: 2.2354 - val_accuracy: 0.3478\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 0s 6ms/step - loss: 2.0580 - accuracy: 0.3317 - val_loss: 1.7184 - val_accuracy: 0.4783\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 0s 8ms/step - loss: 1.5746 - accuracy: 0.4536 - val_loss: 1.1770 - val_accuracy: 0.6870\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 0s 6ms/step - loss: 1.2347 - accuracy: 0.5803 - val_loss: 0.9248 - val_accuracy: 0.7217\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 0s 6ms/step - loss: 1.0381 - accuracy: 0.6547 - val_loss: 0.7678 - val_accuracy: 0.7826\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 0s 6ms/step - loss: 0.9334 - accuracy: 0.6963 - val_loss: 0.6516 - val_accuracy: 0.8261\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.8348 - accuracy: 0.7292 - val_loss: 0.5723 - val_accuracy: 0.8174\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 0s 8ms/step - loss: 0.6942 - accuracy: 0.7892 - val_loss: 0.4734 - val_accuracy: 0.8783\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 0s 6ms/step - loss: 0.6478 - accuracy: 0.7950 - val_loss: 0.4002 - val_accuracy: 0.8957\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.5553 - accuracy: 0.8211 - val_loss: 0.3277 - val_accuracy: 0.9304\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 0s 9ms/step - loss: 2.2723 - accuracy: 0.1886 - val_loss: 2.1879 - val_accuracy: 0.4783\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 0s 6ms/step - loss: 1.9560 - accuracy: 0.3897 - val_loss: 1.5294 - val_accuracy: 0.5739\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 1.4237 - accuracy: 0.5358 - val_loss: 1.0433 - val_accuracy: 0.7739\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 0s 6ms/step - loss: 1.1201 - accuracy: 0.6296 - val_loss: 0.8153 - val_accuracy: 0.7913\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 0s 6ms/step - loss: 0.9852 - accuracy: 0.6886 - val_loss: 0.6740 - val_accuracy: 0.8348\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 0s 8ms/step - loss: 0.8618 - accuracy: 0.7244 - val_loss: 0.5500 - val_accuracy: 0.8435\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.7643 - accuracy: 0.7631 - val_loss: 0.4929 - val_accuracy: 0.8522\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.6944 - accuracy: 0.7911 - val_loss: 0.3976 - val_accuracy: 0.8870\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.5901 - accuracy: 0.8153 - val_loss: 0.3673 - val_accuracy: 0.8783\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 0s 6ms/step - loss: 0.5855 - accuracy: 0.8066 - val_loss: 0.3095 - val_accuracy: 0.9130\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 0s 9ms/step - loss: 2.2836 - accuracy: 0.1633 - val_loss: 2.2320 - val_accuracy: 0.2609\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 2.0863 - accuracy: 0.3082 - val_loss: 1.7739 - val_accuracy: 0.4348\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 1.5834 - accuracy: 0.4589 - val_loss: 1.1873 - val_accuracy: 0.6870\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 1.1693 - accuracy: 0.6193 - val_loss: 0.8534 - val_accuracy: 0.7913\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.9645 - accuracy: 0.6870 - val_loss: 0.6696 - val_accuracy: 0.8522\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 0s 6ms/step - loss: 0.8648 - accuracy: 0.7324 - val_loss: 0.5079 - val_accuracy: 0.8957\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 0s 6ms/step - loss: 0.7592 - accuracy: 0.7691 - val_loss: 0.4190 - val_accuracy: 0.9391\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.6626 - accuracy: 0.7971 - val_loss: 0.3370 - val_accuracy: 0.9304\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 0s 6ms/step - loss: 0.5755 - accuracy: 0.8367 - val_loss: 0.3041 - val_accuracy: 0.9391\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 0s 6ms/step - loss: 0.5523 - accuracy: 0.8309 - val_loss: 0.2578 - val_accuracy: 0.9478\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 0s 9ms/step - loss: 2.2753 - accuracy: 0.1855 - val_loss: 2.1913 - val_accuracy: 0.3913\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 0s 6ms/step - loss: 2.0162 - accuracy: 0.3507 - val_loss: 1.6023 - val_accuracy: 0.6696\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 0s 6ms/step - loss: 1.4881 - accuracy: 0.4957 - val_loss: 1.1027 - val_accuracy: 0.7391\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 1.1514 - accuracy: 0.6145 - val_loss: 0.8032 - val_accuracy: 0.7739\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.9581 - accuracy: 0.6976 - val_loss: 0.6590 - val_accuracy: 0.8348\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 0s 8ms/step - loss: 0.8411 - accuracy: 0.7150 - val_loss: 0.5375 - val_accuracy: 0.8609\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.7265 - accuracy: 0.7488 - val_loss: 0.4625 - val_accuracy: 0.8696\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.6826 - accuracy: 0.7739 - val_loss: 0.3959 - val_accuracy: 0.9130\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 0s 8ms/step - loss: 0.5765 - accuracy: 0.8135 - val_loss: 0.3700 - val_accuracy: 0.8870\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.5221 - accuracy: 0.8184 - val_loss: 0.2989 - val_accuracy: 0.9304\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 1s 10ms/step - loss: 2.2803 - accuracy: 0.1565 - val_loss: 2.2065 - val_accuracy: 0.4609\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 0s 6ms/step - loss: 2.0501 - accuracy: 0.3681 - val_loss: 1.6782 - val_accuracy: 0.4783\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 1.5208 - accuracy: 0.4850 - val_loss: 1.1579 - val_accuracy: 0.7043\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 0s 6ms/step - loss: 1.2282 - accuracy: 0.5671 - val_loss: 0.8985 - val_accuracy: 0.8522\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 0s 6ms/step - loss: 1.0477 - accuracy: 0.6242 - val_loss: 0.7636 - val_accuracy: 0.8174\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.9540 - accuracy: 0.6860 - val_loss: 0.6311 - val_accuracy: 0.8783\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.8454 - accuracy: 0.7227 - val_loss: 0.5258 - val_accuracy: 0.8783\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.7523 - accuracy: 0.7633 - val_loss: 0.4517 - val_accuracy: 0.9043\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.6788 - accuracy: 0.7768 - val_loss: 0.3769 - val_accuracy: 0.9130\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 0s 7ms/step - loss: 0.5987 - accuracy: 0.8077 - val_loss: 0.3243 - val_accuracy: 0.9217\n",
            "Epoch 1/15\n",
            "7/7 [==============================] - 0s 39ms/step - loss: 2.8138 - accuracy: 0.0870 - val_loss: 2.3032 - val_accuracy: 0.1217\n",
            "Epoch 2/15\n",
            "7/7 [==============================] - 0s 22ms/step - loss: 2.5729 - accuracy: 0.0996 - val_loss: 2.3024 - val_accuracy: 0.1043\n",
            "Epoch 3/15\n",
            "7/7 [==============================] - 0s 23ms/step - loss: 2.7368 - accuracy: 0.1083 - val_loss: 2.3129 - val_accuracy: 0.0696\n",
            "Epoch 4/15\n",
            "7/7 [==============================] - 0s 26ms/step - loss: 2.3061 - accuracy: 0.0880 - val_loss: 2.3179 - val_accuracy: 0.1043\n",
            "Epoch 5/15\n",
            "7/7 [==============================] - 0s 24ms/step - loss: 2.3045 - accuracy: 0.0928 - val_loss: 2.3151 - val_accuracy: 0.1043\n",
            "Epoch 6/15\n",
            "7/7 [==============================] - 0s 23ms/step - loss: 2.3029 - accuracy: 0.0977 - val_loss: 2.3106 - val_accuracy: 0.1043\n",
            "Epoch 7/15\n",
            "7/7 [==============================] - 0s 24ms/step - loss: 2.3018 - accuracy: 0.1093 - val_loss: 2.3089 - val_accuracy: 0.1043\n",
            "Epoch 8/15\n",
            "7/7 [==============================] - 0s 23ms/step - loss: 2.3013 - accuracy: 0.1141 - val_loss: 2.3091 - val_accuracy: 0.1043\n",
            "Epoch 9/15\n",
            "7/7 [==============================] - 0s 29ms/step - loss: 2.3013 - accuracy: 0.1141 - val_loss: 2.3091 - val_accuracy: 0.1043\n",
            "Epoch 10/15\n",
            "7/7 [==============================] - 0s 23ms/step - loss: 2.3013 - accuracy: 0.1141 - val_loss: 2.3060 - val_accuracy: 0.1043\n",
            "Epoch 11/15\n",
            "7/7 [==============================] - 0s 24ms/step - loss: 2.3010 - accuracy: 0.1141 - val_loss: 2.3049 - val_accuracy: 0.1043\n",
            "Epoch 12/15\n",
            "7/7 [==============================] - 0s 23ms/step - loss: 2.3009 - accuracy: 0.1141 - val_loss: 2.3070 - val_accuracy: 0.1043\n",
            "Epoch 13/15\n",
            "7/7 [==============================] - 0s 26ms/step - loss: 2.3010 - accuracy: 0.1141 - val_loss: 2.3087 - val_accuracy: 0.1043\n",
            "Epoch 14/15\n",
            "7/7 [==============================] - 0s 24ms/step - loss: 2.3015 - accuracy: 0.1141 - val_loss: 2.3056 - val_accuracy: 0.1043\n",
            "Epoch 15/15\n",
            "7/7 [==============================] - 0s 24ms/step - loss: 2.3014 - accuracy: 0.1141 - val_loss: 2.3057 - val_accuracy: 0.1043\n",
            "Epoch 1/15\n",
            "7/7 [==============================] - 0s 49ms/step - loss: 2.7179 - accuracy: 0.1103 - val_loss: 2.3082 - val_accuracy: 0.0957\n",
            "Epoch 2/15\n",
            "7/7 [==============================] - 0s 22ms/step - loss: 2.3080 - accuracy: 0.0928 - val_loss: 2.3081 - val_accuracy: 0.0696\n",
            "Epoch 3/15\n",
            "7/7 [==============================] - 0s 23ms/step - loss: 2.3039 - accuracy: 0.1035 - val_loss: 2.3054 - val_accuracy: 0.1304\n",
            "Epoch 4/15\n",
            "7/7 [==============================] - 0s 21ms/step - loss: 2.3029 - accuracy: 0.1044 - val_loss: 2.3037 - val_accuracy: 0.1304\n",
            "Epoch 5/15\n",
            "7/7 [==============================] - 0s 22ms/step - loss: 2.3029 - accuracy: 0.0996 - val_loss: 2.3034 - val_accuracy: 0.0696\n",
            "Epoch 6/15\n",
            "7/7 [==============================] - 0s 24ms/step - loss: 3.7429 - accuracy: 0.0909 - val_loss: 2.3062 - val_accuracy: 0.0696\n",
            "Epoch 7/15\n",
            "7/7 [==============================] - 0s 23ms/step - loss: 2.3328 - accuracy: 0.0890 - val_loss: 2.3066 - val_accuracy: 0.1043\n",
            "Epoch 8/15\n",
            "7/7 [==============================] - 0s 24ms/step - loss: 2.3056 - accuracy: 0.1074 - val_loss: 2.3083 - val_accuracy: 0.1043\n",
            "Epoch 9/15\n",
            "7/7 [==============================] - 0s 24ms/step - loss: 2.3050 - accuracy: 0.1074 - val_loss: 2.3070 - val_accuracy: 0.1043\n",
            "Epoch 10/15\n",
            "7/7 [==============================] - 0s 28ms/step - loss: 2.3035 - accuracy: 0.1006 - val_loss: 2.3065 - val_accuracy: 0.0696\n",
            "Epoch 11/15\n",
            "7/7 [==============================] - 0s 23ms/step - loss: 2.3027 - accuracy: 0.1064 - val_loss: 2.3067 - val_accuracy: 0.0696\n",
            "Epoch 12/15\n",
            "7/7 [==============================] - 0s 24ms/step - loss: 2.3029 - accuracy: 0.1064 - val_loss: 2.3078 - val_accuracy: 0.1043\n",
            "Epoch 13/15\n",
            "7/7 [==============================] - 0s 25ms/step - loss: 2.3027 - accuracy: 0.1074 - val_loss: 2.3030 - val_accuracy: 0.1043\n",
            "Epoch 14/15\n",
            "7/7 [==============================] - 0s 24ms/step - loss: 2.3031 - accuracy: 0.0996 - val_loss: 2.3032 - val_accuracy: 0.1043\n",
            "Epoch 15/15\n",
            "7/7 [==============================] - 0s 24ms/step - loss: 2.3034 - accuracy: 0.1074 - val_loss: 2.3051 - val_accuracy: 0.1043\n",
            "Epoch 1/15\n",
            "7/7 [==============================] - 0s 39ms/step - loss: 3.7483 - accuracy: 0.1034 - val_loss: 2.3112 - val_accuracy: 0.1304\n",
            "Epoch 2/15\n",
            "7/7 [==============================] - 0s 21ms/step - loss: 2.3253 - accuracy: 0.1034 - val_loss: 2.3093 - val_accuracy: 0.1043\n",
            "Epoch 3/15\n",
            "7/7 [==============================] - 0s 23ms/step - loss: 2.3043 - accuracy: 0.1111 - val_loss: 2.3045 - val_accuracy: 0.1043\n",
            "Epoch 4/15\n",
            "7/7 [==============================] - 0s 22ms/step - loss: 2.3032 - accuracy: 0.1111 - val_loss: 2.3020 - val_accuracy: 0.1043\n",
            "Epoch 5/15\n",
            "7/7 [==============================] - 0s 24ms/step - loss: 2.3021 - accuracy: 0.1111 - val_loss: 2.2991 - val_accuracy: 0.1043\n",
            "Epoch 6/15\n",
            "7/7 [==============================] - 0s 28ms/step - loss: 2.3024 - accuracy: 0.1111 - val_loss: 2.2997 - val_accuracy: 0.1043\n",
            "Epoch 7/15\n",
            "7/7 [==============================] - 0s 24ms/step - loss: 2.3028 - accuracy: 0.1111 - val_loss: 2.2989 - val_accuracy: 0.1043\n",
            "Epoch 8/15\n",
            "7/7 [==============================] - 0s 24ms/step - loss: 2.3016 - accuracy: 0.1111 - val_loss: 2.2995 - val_accuracy: 0.1043\n",
            "Epoch 9/15\n",
            "7/7 [==============================] - 0s 28ms/step - loss: 2.3027 - accuracy: 0.1111 - val_loss: 2.3022 - val_accuracy: 0.1043\n",
            "Epoch 10/15\n",
            "7/7 [==============================] - 0s 28ms/step - loss: 2.3021 - accuracy: 0.1111 - val_loss: 2.3012 - val_accuracy: 0.1043\n",
            "Epoch 11/15\n",
            "7/7 [==============================] - 0s 24ms/step - loss: 2.3019 - accuracy: 0.1111 - val_loss: 2.3005 - val_accuracy: 0.1043\n",
            "Epoch 12/15\n",
            "7/7 [==============================] - 0s 25ms/step - loss: 2.3023 - accuracy: 0.1111 - val_loss: 2.2987 - val_accuracy: 0.1043\n",
            "Epoch 13/15\n",
            "7/7 [==============================] - 0s 24ms/step - loss: 2.3029 - accuracy: 0.0966 - val_loss: 2.3001 - val_accuracy: 0.1043\n",
            "Epoch 14/15\n",
            "7/7 [==============================] - 0s 22ms/step - loss: 2.3023 - accuracy: 0.1111 - val_loss: 2.3020 - val_accuracy: 0.1043\n",
            "Epoch 15/15\n",
            "7/7 [==============================] - 0s 23ms/step - loss: 2.3018 - accuracy: 0.1111 - val_loss: 2.3008 - val_accuracy: 0.1043\n",
            "Epoch 1/15\n",
            "7/7 [==============================] - 0s 41ms/step - loss: 2.7929 - accuracy: 0.0995 - val_loss: 2.3118 - val_accuracy: 0.0783\n",
            "Epoch 2/15\n",
            "7/7 [==============================] - 0s 26ms/step - loss: 2.3678 - accuracy: 0.0918 - val_loss: 2.3011 - val_accuracy: 0.1304\n",
            "Epoch 3/15\n",
            "7/7 [==============================] - 0s 23ms/step - loss: 2.3044 - accuracy: 0.1005 - val_loss: 2.2993 - val_accuracy: 0.1130\n",
            "Epoch 4/15\n",
            "7/7 [==============================] - 0s 23ms/step - loss: 2.3041 - accuracy: 0.1063 - val_loss: 2.3018 - val_accuracy: 0.1130\n",
            "Epoch 5/15\n",
            "7/7 [==============================] - 0s 24ms/step - loss: 2.3036 - accuracy: 0.1063 - val_loss: 2.3097 - val_accuracy: 0.1130\n",
            "Epoch 6/15\n",
            "7/7 [==============================] - 0s 27ms/step - loss: 2.3038 - accuracy: 0.1005 - val_loss: 2.3080 - val_accuracy: 0.0696\n",
            "Epoch 7/15\n",
            "7/7 [==============================] - 0s 21ms/step - loss: 2.3037 - accuracy: 0.0986 - val_loss: 2.3061 - val_accuracy: 0.0957\n",
            "Epoch 8/15\n",
            "7/7 [==============================] - 0s 25ms/step - loss: 2.3028 - accuracy: 0.1043 - val_loss: 2.3042 - val_accuracy: 0.1130\n",
            "Epoch 9/15\n",
            "7/7 [==============================] - 0s 23ms/step - loss: 2.3039 - accuracy: 0.1063 - val_loss: 2.3059 - val_accuracy: 0.1130\n",
            "Epoch 10/15\n",
            "7/7 [==============================] - 0s 24ms/step - loss: 2.3036 - accuracy: 0.1063 - val_loss: 2.3058 - val_accuracy: 0.1130\n",
            "Epoch 11/15\n",
            "7/7 [==============================] - 0s 26ms/step - loss: 2.3039 - accuracy: 0.0976 - val_loss: 2.3056 - val_accuracy: 0.1130\n",
            "Epoch 12/15\n",
            "7/7 [==============================] - 0s 24ms/step - loss: 2.3031 - accuracy: 0.1063 - val_loss: 2.3042 - val_accuracy: 0.1130\n",
            "Epoch 13/15\n",
            "7/7 [==============================] - 0s 24ms/step - loss: 2.3035 - accuracy: 0.0995 - val_loss: 2.3065 - val_accuracy: 0.0957\n",
            "Epoch 14/15\n",
            "7/7 [==============================] - 0s 25ms/step - loss: 2.3036 - accuracy: 0.1043 - val_loss: 2.3048 - val_accuracy: 0.0957\n",
            "Epoch 15/15\n",
            "7/7 [==============================] - 0s 24ms/step - loss: 2.3039 - accuracy: 0.0957 - val_loss: 2.3065 - val_accuracy: 0.1130\n",
            "WARNING:tensorflow:5 out of the last 22 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd19e65268> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/15\n",
            "7/7 [==============================] - 0s 40ms/step - loss: 2.5759 - accuracy: 0.1014 - val_loss: 2.2679 - val_accuracy: 0.1913\n",
            "Epoch 2/15\n",
            "7/7 [==============================] - 0s 28ms/step - loss: 2.3190 - accuracy: 0.0986 - val_loss: 2.2940 - val_accuracy: 0.1043\n",
            "Epoch 3/15\n",
            "7/7 [==============================] - 0s 21ms/step - loss: 2.3050 - accuracy: 0.1014 - val_loss: 2.3156 - val_accuracy: 0.0609\n",
            "Epoch 4/15\n",
            "7/7 [==============================] - 0s 22ms/step - loss: 2.3038 - accuracy: 0.1053 - val_loss: 2.3063 - val_accuracy: 0.1043\n",
            "Epoch 5/15\n",
            "7/7 [==============================] - 0s 24ms/step - loss: 2.3044 - accuracy: 0.1072 - val_loss: 2.3029 - val_accuracy: 0.1043\n",
            "Epoch 6/15\n",
            "7/7 [==============================] - 0s 23ms/step - loss: 2.3029 - accuracy: 0.1072 - val_loss: 2.3086 - val_accuracy: 0.0348\n",
            "Epoch 7/15\n",
            "7/7 [==============================] - 0s 26ms/step - loss: 2.3025 - accuracy: 0.1005 - val_loss: 2.3109 - val_accuracy: 0.0870\n",
            "Epoch 8/15\n",
            "7/7 [==============================] - 0s 23ms/step - loss: 2.3025 - accuracy: 0.1063 - val_loss: 2.3073 - val_accuracy: 0.0870\n",
            "Epoch 9/15\n",
            "7/7 [==============================] - 0s 23ms/step - loss: 2.3030 - accuracy: 0.0889 - val_loss: 2.3040 - val_accuracy: 0.1043\n",
            "Epoch 10/15\n",
            "7/7 [==============================] - 0s 24ms/step - loss: 2.3032 - accuracy: 0.0899 - val_loss: 2.3133 - val_accuracy: 0.0609\n",
            "Epoch 11/15\n",
            "7/7 [==============================] - 0s 28ms/step - loss: 2.3028 - accuracy: 0.1053 - val_loss: 2.3110 - val_accuracy: 0.1043\n",
            "Epoch 12/15\n",
            "7/7 [==============================] - 0s 21ms/step - loss: 2.3030 - accuracy: 0.1072 - val_loss: 2.3059 - val_accuracy: 0.1043\n",
            "Epoch 13/15\n",
            "7/7 [==============================] - 0s 21ms/step - loss: 2.3032 - accuracy: 0.1014 - val_loss: 2.3045 - val_accuracy: 0.0870\n",
            "Epoch 14/15\n",
            "7/7 [==============================] - 0s 23ms/step - loss: 2.3020 - accuracy: 0.1063 - val_loss: 2.3062 - val_accuracy: 0.0870\n",
            "Epoch 15/15\n",
            "7/7 [==============================] - 0s 29ms/step - loss: 2.3020 - accuracy: 0.0976 - val_loss: 2.3107 - val_accuracy: 0.1043\n",
            "WARNING:tensorflow:6 out of the last 24 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd075d07b8> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/5\n",
            "52/52 [==============================] - 7s 133ms/step - loss: 8.5003 - accuracy: 0.0928 - val_loss: 2.3086 - val_accuracy: 0.0696\n",
            "Epoch 2/5\n",
            "52/52 [==============================] - 7s 132ms/step - loss: 2.3113 - accuracy: 0.0986 - val_loss: 2.3135 - val_accuracy: 0.0696\n",
            "Epoch 3/5\n",
            "52/52 [==============================] - 7s 129ms/step - loss: 2.3085 - accuracy: 0.0832 - val_loss: 2.3067 - val_accuracy: 0.1043\n",
            "Epoch 4/5\n",
            "52/52 [==============================] - 7s 131ms/step - loss: 2.3101 - accuracy: 0.0967 - val_loss: 2.3128 - val_accuracy: 0.1130\n",
            "Epoch 5/5\n",
            "52/52 [==============================] - 7s 132ms/step - loss: 2.3106 - accuracy: 0.0851 - val_loss: 2.3040 - val_accuracy: 0.1130\n",
            "WARNING:tensorflow:6 out of the last 11 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd0dd07a60> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/5\n",
            "52/52 [==============================] - 7s 132ms/step - loss: 16.9816 - accuracy: 0.0890 - val_loss: 2.3088 - val_accuracy: 0.0696\n",
            "Epoch 2/5\n",
            "52/52 [==============================] - 7s 130ms/step - loss: 2.3089 - accuracy: 0.0861 - val_loss: 2.3059 - val_accuracy: 0.1043\n",
            "Epoch 3/5\n",
            "52/52 [==============================] - 7s 130ms/step - loss: 2.3125 - accuracy: 0.0861 - val_loss: 2.3054 - val_accuracy: 0.1043\n",
            "Epoch 4/5\n",
            "52/52 [==============================] - 7s 130ms/step - loss: 2.3107 - accuracy: 0.0957 - val_loss: 2.3117 - val_accuracy: 0.1304\n",
            "Epoch 5/5\n",
            "52/52 [==============================] - 7s 130ms/step - loss: 2.3088 - accuracy: 0.0967 - val_loss: 2.3087 - val_accuracy: 0.0696\n",
            "Epoch 1/5\n",
            "52/52 [==============================] - 7s 134ms/step - loss: 8.5128 - accuracy: 0.0976 - val_loss: 2.3101 - val_accuracy: 0.0696\n",
            "Epoch 2/5\n",
            "52/52 [==============================] - 7s 132ms/step - loss: 2.3093 - accuracy: 0.0908 - val_loss: 2.2975 - val_accuracy: 0.1130\n",
            "Epoch 3/5\n",
            "52/52 [==============================] - 7s 132ms/step - loss: 2.3115 - accuracy: 0.0986 - val_loss: 2.2964 - val_accuracy: 0.1304\n",
            "Epoch 4/5\n",
            "52/52 [==============================] - 7s 131ms/step - loss: 2.3102 - accuracy: 0.0899 - val_loss: 2.3049 - val_accuracy: 0.1043\n",
            "Epoch 5/5\n",
            "52/52 [==============================] - 7s 130ms/step - loss: 2.3136 - accuracy: 0.0802 - val_loss: 2.2933 - val_accuracy: 0.1043\n",
            "Epoch 1/5\n",
            "52/52 [==============================] - 7s 133ms/step - loss: 12.0988 - accuracy: 0.1169 - val_loss: 2.3037 - val_accuracy: 0.1130\n",
            "Epoch 2/5\n",
            "52/52 [==============================] - 7s 133ms/step - loss: 2.3121 - accuracy: 0.0937 - val_loss: 2.3044 - val_accuracy: 0.0957\n",
            "Epoch 3/5\n",
            "52/52 [==============================] - 7s 131ms/step - loss: 2.3122 - accuracy: 0.0812 - val_loss: 2.3155 - val_accuracy: 0.0783\n",
            "Epoch 4/5\n",
            "52/52 [==============================] - 7s 131ms/step - loss: 2.3113 - accuracy: 0.0957 - val_loss: 2.3015 - val_accuracy: 0.1304\n",
            "Epoch 5/5\n",
            "52/52 [==============================] - 7s 131ms/step - loss: 2.3119 - accuracy: 0.0841 - val_loss: 2.3034 - val_accuracy: 0.1304\n",
            "Epoch 1/5\n",
            "52/52 [==============================] - 7s 132ms/step - loss: 7.9698 - accuracy: 0.0986 - val_loss: 2.3529 - val_accuracy: 0.0348\n",
            "Epoch 2/5\n",
            "52/52 [==============================] - 7s 130ms/step - loss: 2.3093 - accuracy: 0.0918 - val_loss: 2.2982 - val_accuracy: 0.0348\n",
            "Epoch 3/5\n",
            "52/52 [==============================] - 7s 130ms/step - loss: 2.3094 - accuracy: 0.1014 - val_loss: 2.3153 - val_accuracy: 0.1043\n",
            "Epoch 4/5\n",
            "52/52 [==============================] - 7s 129ms/step - loss: 2.3108 - accuracy: 0.0995 - val_loss: 2.3107 - val_accuracy: 0.0870\n",
            "Epoch 5/5\n",
            "52/52 [==============================] - 7s 131ms/step - loss: 2.3116 - accuracy: 0.0966 - val_loss: 2.3048 - val_accuracy: 0.0609\n",
            "Epoch 1/10\n",
            "21/21 [==============================] - 1s 70ms/step - loss: 3.9736 - accuracy: 0.0957 - val_loss: 2.2994 - val_accuracy: 0.1304\n",
            "Epoch 2/10\n",
            "21/21 [==============================] - 1s 65ms/step - loss: 2.3076 - accuracy: 0.0977 - val_loss: 2.3120 - val_accuracy: 0.1043\n",
            "Epoch 3/10\n",
            "21/21 [==============================] - 1s 66ms/step - loss: 2.3034 - accuracy: 0.1141 - val_loss: 2.3050 - val_accuracy: 0.1043\n",
            "Epoch 4/10\n",
            "21/21 [==============================] - 1s 67ms/step - loss: 2.3055 - accuracy: 0.0977 - val_loss: 2.3111 - val_accuracy: 0.1043\n",
            "Epoch 5/10\n",
            "21/21 [==============================] - 1s 67ms/step - loss: 2.3047 - accuracy: 0.1141 - val_loss: 2.3021 - val_accuracy: 0.1043\n",
            "Epoch 6/10\n",
            "21/21 [==============================] - 1s 66ms/step - loss: 2.3045 - accuracy: 0.1025 - val_loss: 2.3109 - val_accuracy: 0.0696\n",
            "Epoch 7/10\n",
            "21/21 [==============================] - 1s 68ms/step - loss: 2.3053 - accuracy: 0.1122 - val_loss: 2.3081 - val_accuracy: 0.1043\n",
            "Epoch 8/10\n",
            "21/21 [==============================] - 1s 65ms/step - loss: 2.3041 - accuracy: 0.1074 - val_loss: 2.3077 - val_accuracy: 0.0696\n",
            "Epoch 9/10\n",
            "21/21 [==============================] - 1s 67ms/step - loss: 2.3041 - accuracy: 0.0899 - val_loss: 2.3040 - val_accuracy: 0.1130\n",
            "Epoch 10/10\n",
            "21/21 [==============================] - 1s 66ms/step - loss: 2.3060 - accuracy: 0.1161 - val_loss: 2.3072 - val_accuracy: 0.1043\n",
            "Epoch 1/10\n",
            "21/21 [==============================] - 1s 71ms/step - loss: 4.3246 - accuracy: 0.0986 - val_loss: 2.3127 - val_accuracy: 0.0957\n",
            "Epoch 2/10\n",
            "21/21 [==============================] - 1s 65ms/step - loss: 2.3143 - accuracy: 0.0986 - val_loss: 2.3050 - val_accuracy: 0.1043\n",
            "Epoch 3/10\n",
            "21/21 [==============================] - 1s 65ms/step - loss: 2.3092 - accuracy: 0.0996 - val_loss: 2.3047 - val_accuracy: 0.1304\n",
            "Epoch 4/10\n",
            "21/21 [==============================] - 1s 66ms/step - loss: 2.3095 - accuracy: 0.0986 - val_loss: 2.3112 - val_accuracy: 0.1043\n",
            "Epoch 5/10\n",
            "21/21 [==============================] - 1s 64ms/step - loss: 2.3114 - accuracy: 0.1035 - val_loss: 2.3066 - val_accuracy: 0.0696\n",
            "Epoch 6/10\n",
            "21/21 [==============================] - 1s 65ms/step - loss: 2.3104 - accuracy: 0.0919 - val_loss: 2.3051 - val_accuracy: 0.0696\n",
            "Epoch 7/10\n",
            "21/21 [==============================] - 1s 67ms/step - loss: 2.3072 - accuracy: 0.0977 - val_loss: 2.3085 - val_accuracy: 0.1043\n",
            "Epoch 8/10\n",
            "21/21 [==============================] - 1s 69ms/step - loss: 2.3058 - accuracy: 0.0986 - val_loss: 2.3047 - val_accuracy: 0.0696\n",
            "Epoch 9/10\n",
            "21/21 [==============================] - 1s 66ms/step - loss: 2.3066 - accuracy: 0.0977 - val_loss: 2.3039 - val_accuracy: 0.1043\n",
            "Epoch 10/10\n",
            "21/21 [==============================] - 1s 65ms/step - loss: 2.3049 - accuracy: 0.1074 - val_loss: 2.3067 - val_accuracy: 0.1043\n",
            "Epoch 1/10\n",
            "21/21 [==============================] - 2s 72ms/step - loss: 4.7597 - accuracy: 0.0908 - val_loss: 2.2976 - val_accuracy: 0.1130\n",
            "Epoch 2/10\n",
            "21/21 [==============================] - 1s 68ms/step - loss: 2.3055 - accuracy: 0.1063 - val_loss: 2.3010 - val_accuracy: 0.1043\n",
            "Epoch 3/10\n",
            "21/21 [==============================] - 1s 68ms/step - loss: 2.3041 - accuracy: 0.1111 - val_loss: 2.3027 - val_accuracy: 0.1043\n",
            "Epoch 4/10\n",
            "21/21 [==============================] - 1s 67ms/step - loss: 2.3049 - accuracy: 0.0986 - val_loss: 2.2973 - val_accuracy: 0.1304\n",
            "Epoch 5/10\n",
            "21/21 [==============================] - 1s 68ms/step - loss: 2.3041 - accuracy: 0.1063 - val_loss: 2.3040 - val_accuracy: 0.1043\n",
            "Epoch 6/10\n",
            "21/21 [==============================] - 1s 66ms/step - loss: 2.3044 - accuracy: 0.1111 - val_loss: 2.2995 - val_accuracy: 0.1043\n",
            "Epoch 7/10\n",
            "21/21 [==============================] - 1s 66ms/step - loss: 2.3035 - accuracy: 0.1072 - val_loss: 2.3005 - val_accuracy: 0.1304\n",
            "Epoch 8/10\n",
            "21/21 [==============================] - 1s 66ms/step - loss: 2.3050 - accuracy: 0.1072 - val_loss: 2.3020 - val_accuracy: 0.1043\n",
            "Epoch 9/10\n",
            "21/21 [==============================] - 1s 70ms/step - loss: 2.3051 - accuracy: 0.0937 - val_loss: 2.3028 - val_accuracy: 0.1043\n",
            "Epoch 10/10\n",
            "21/21 [==============================] - 1s 66ms/step - loss: 2.3046 - accuracy: 0.1014 - val_loss: 2.2978 - val_accuracy: 0.1043\n",
            "Epoch 1/10\n",
            "21/21 [==============================] - 2s 73ms/step - loss: 3.2103 - accuracy: 0.0850 - val_loss: 2.3070 - val_accuracy: 0.1130\n",
            "Epoch 2/10\n",
            "21/21 [==============================] - 1s 67ms/step - loss: 2.3089 - accuracy: 0.1024 - val_loss: 2.3081 - val_accuracy: 0.1130\n",
            "Epoch 3/10\n",
            "21/21 [==============================] - 1s 64ms/step - loss: 2.3058 - accuracy: 0.1053 - val_loss: 2.3078 - val_accuracy: 0.0696\n",
            "Epoch 4/10\n",
            "21/21 [==============================] - 1s 67ms/step - loss: 2.3062 - accuracy: 0.1043 - val_loss: 2.3034 - val_accuracy: 0.0957\n",
            "Epoch 5/10\n",
            "21/21 [==============================] - 1s 65ms/step - loss: 2.3059 - accuracy: 0.0812 - val_loss: 2.3048 - val_accuracy: 0.1130\n",
            "Epoch 6/10\n",
            "21/21 [==============================] - 1s 66ms/step - loss: 2.3064 - accuracy: 0.1063 - val_loss: 2.3029 - val_accuracy: 0.1130\n",
            "Epoch 7/10\n",
            "21/21 [==============================] - 1s 66ms/step - loss: 2.3063 - accuracy: 0.1014 - val_loss: 2.3041 - val_accuracy: 0.1130\n",
            "Epoch 8/10\n",
            "21/21 [==============================] - 1s 65ms/step - loss: 2.3059 - accuracy: 0.0870 - val_loss: 2.3081 - val_accuracy: 0.1130\n",
            "Epoch 9/10\n",
            "21/21 [==============================] - 1s 65ms/step - loss: 2.3063 - accuracy: 0.0947 - val_loss: 2.3041 - val_accuracy: 0.1130\n",
            "Epoch 10/10\n",
            "21/21 [==============================] - 1s 67ms/step - loss: 2.3078 - accuracy: 0.0947 - val_loss: 2.3018 - val_accuracy: 0.0957\n",
            "Epoch 1/10\n",
            "21/21 [==============================] - 2s 73ms/step - loss: 4.0006 - accuracy: 0.1005 - val_loss: 2.3201 - val_accuracy: 0.0609\n",
            "Epoch 2/10\n",
            "21/21 [==============================] - 1s 66ms/step - loss: 2.3078 - accuracy: 0.1043 - val_loss: 2.3080 - val_accuracy: 0.0609\n",
            "Epoch 3/10\n",
            "21/21 [==============================] - 1s 67ms/step - loss: 2.3049 - accuracy: 0.0850 - val_loss: 2.3117 - val_accuracy: 0.0870\n",
            "Epoch 4/10\n",
            "21/21 [==============================] - 1s 67ms/step - loss: 2.3041 - accuracy: 0.1063 - val_loss: 2.3094 - val_accuracy: 0.1043\n",
            "Epoch 5/10\n",
            "21/21 [==============================] - 1s 66ms/step - loss: 2.3059 - accuracy: 0.1072 - val_loss: 2.3011 - val_accuracy: 0.1043\n",
            "Epoch 6/10\n",
            "21/21 [==============================] - 1s 69ms/step - loss: 2.3051 - accuracy: 0.1043 - val_loss: 2.3169 - val_accuracy: 0.0348\n",
            "Epoch 7/10\n",
            "21/21 [==============================] - 1s 67ms/step - loss: 2.3052 - accuracy: 0.0937 - val_loss: 2.3072 - val_accuracy: 0.0870\n",
            "Epoch 8/10\n",
            "21/21 [==============================] - 1s 67ms/step - loss: 2.3065 - accuracy: 0.1053 - val_loss: 2.3157 - val_accuracy: 0.1043\n",
            "Epoch 9/10\n",
            "21/21 [==============================] - 1s 66ms/step - loss: 2.3056 - accuracy: 0.1005 - val_loss: 2.3086 - val_accuracy: 0.1043\n",
            "Epoch 10/10\n",
            "21/21 [==============================] - 1s 69ms/step - loss: 2.3063 - accuracy: 0.0870 - val_loss: 2.3114 - val_accuracy: 0.1043\n",
            "Epoch 1/5\n",
            "7/7 [==============================] - 0s 40ms/step - loss: 2.3058 - accuracy: 0.1093 - val_loss: 2.2919 - val_accuracy: 0.1304\n",
            "Epoch 2/5\n",
            "7/7 [==============================] - 0s 24ms/step - loss: 2.2440 - accuracy: 0.1489 - val_loss: 2.1417 - val_accuracy: 0.2174\n",
            "Epoch 3/5\n",
            "7/7 [==============================] - 0s 24ms/step - loss: 2.0250 - accuracy: 0.2689 - val_loss: 1.6587 - val_accuracy: 0.4783\n",
            "Epoch 4/5\n",
            "7/7 [==============================] - 0s 25ms/step - loss: 1.6811 - accuracy: 0.4004 - val_loss: 1.1632 - val_accuracy: 0.7478\n",
            "Epoch 5/5\n",
            "7/7 [==============================] - 0s 24ms/step - loss: 1.3686 - accuracy: 0.5280 - val_loss: 0.8162 - val_accuracy: 0.8087\n",
            "Epoch 1/5\n",
            "7/7 [==============================] - 0s 41ms/step - loss: 2.3074 - accuracy: 0.1141 - val_loss: 2.2707 - val_accuracy: 0.2348\n",
            "Epoch 2/5\n",
            "7/7 [==============================] - 0s 21ms/step - loss: 2.2270 - accuracy: 0.1741 - val_loss: 2.0675 - val_accuracy: 0.2957\n",
            "Epoch 3/5\n",
            "7/7 [==============================] - 0s 24ms/step - loss: 1.9671 - accuracy: 0.2940 - val_loss: 1.5487 - val_accuracy: 0.5565\n",
            "Epoch 4/5\n",
            "7/7 [==============================] - 0s 23ms/step - loss: 1.5234 - accuracy: 0.4671 - val_loss: 1.0198 - val_accuracy: 0.6696\n",
            "Epoch 5/5\n",
            "7/7 [==============================] - 0s 23ms/step - loss: 1.2711 - accuracy: 0.5474 - val_loss: 0.8289 - val_accuracy: 0.7304\n",
            "Epoch 1/5\n",
            "7/7 [==============================] - 0s 40ms/step - loss: 2.3030 - accuracy: 0.1121 - val_loss: 2.2981 - val_accuracy: 0.1043\n",
            "Epoch 2/5\n",
            "7/7 [==============================] - 0s 22ms/step - loss: 2.2715 - accuracy: 0.1295 - val_loss: 2.2196 - val_accuracy: 0.1652\n",
            "Epoch 3/5\n",
            "7/7 [==============================] - 0s 23ms/step - loss: 2.1281 - accuracy: 0.2213 - val_loss: 1.8806 - val_accuracy: 0.6609\n",
            "Epoch 4/5\n",
            "7/7 [==============================] - 0s 21ms/step - loss: 1.7501 - accuracy: 0.3903 - val_loss: 1.2851 - val_accuracy: 0.6435\n",
            "Epoch 5/5\n",
            "7/7 [==============================] - 0s 24ms/step - loss: 1.4274 - accuracy: 0.4763 - val_loss: 0.9607 - val_accuracy: 0.7826\n",
            "Epoch 1/5\n",
            "7/7 [==============================] - 0s 51ms/step - loss: 2.3079 - accuracy: 0.1005 - val_loss: 2.2974 - val_accuracy: 0.1130\n",
            "Epoch 2/5\n",
            "7/7 [==============================] - 0s 25ms/step - loss: 2.2833 - accuracy: 0.1546 - val_loss: 2.2284 - val_accuracy: 0.2522\n",
            "Epoch 3/5\n",
            "7/7 [==============================] - 0s 23ms/step - loss: 2.0807 - accuracy: 0.2580 - val_loss: 1.7234 - val_accuracy: 0.5130\n",
            "Epoch 4/5\n",
            "7/7 [==============================] - 0s 26ms/step - loss: 1.7152 - accuracy: 0.3623 - val_loss: 1.2659 - val_accuracy: 0.7043\n",
            "Epoch 5/5\n",
            "7/7 [==============================] - 0s 24ms/step - loss: 1.4363 - accuracy: 0.4686 - val_loss: 0.9795 - val_accuracy: 0.7739\n",
            "WARNING:tensorflow:5 out of the last 13 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd0cb93e18> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/5\n",
            "7/7 [==============================] - 0s 40ms/step - loss: 2.3021 - accuracy: 0.1005 - val_loss: 2.2948 - val_accuracy: 0.1391\n",
            "Epoch 2/5\n",
            "7/7 [==============================] - 0s 21ms/step - loss: 2.2438 - accuracy: 0.1652 - val_loss: 2.1385 - val_accuracy: 0.3130\n",
            "Epoch 3/5\n",
            "7/7 [==============================] - 0s 22ms/step - loss: 1.9506 - accuracy: 0.2879 - val_loss: 1.5682 - val_accuracy: 0.6000\n",
            "Epoch 4/5\n",
            "7/7 [==============================] - 0s 29ms/step - loss: 1.5467 - accuracy: 0.4348 - val_loss: 1.0655 - val_accuracy: 0.7130\n",
            "Epoch 5/5\n",
            "7/7 [==============================] - 0s 25ms/step - loss: 1.2536 - accuracy: 0.5391 - val_loss: 0.8454 - val_accuracy: 0.7478\n",
            "WARNING:tensorflow:6 out of the last 15 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd066a60d0> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 3s 54ms/step - loss: 2.1664 - accuracy: 0.2611 - val_loss: 1.5542 - val_accuracy: 0.5739\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 3s 55ms/step - loss: 0.9094 - accuracy: 0.7485 - val_loss: 0.4192 - val_accuracy: 0.8957\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 3s 54ms/step - loss: 0.4213 - accuracy: 0.8723 - val_loss: 0.2737 - val_accuracy: 0.9130\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 3s 54ms/step - loss: 0.3051 - accuracy: 0.9052 - val_loss: 0.1770 - val_accuracy: 0.9565\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 3s 51ms/step - loss: 0.2109 - accuracy: 0.9265 - val_loss: 0.1620 - val_accuracy: 0.9478\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 3s 56ms/step - loss: 0.1661 - accuracy: 0.9449 - val_loss: 0.1110 - val_accuracy: 0.9652\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 3s 57ms/step - loss: 0.1332 - accuracy: 0.9642 - val_loss: 0.1213 - val_accuracy: 0.9652\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 3s 57ms/step - loss: 0.1178 - accuracy: 0.9574 - val_loss: 0.1162 - val_accuracy: 0.9652\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 3s 57ms/step - loss: 0.1151 - accuracy: 0.9603 - val_loss: 0.0773 - val_accuracy: 0.9826\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 3s 56ms/step - loss: 0.1029 - accuracy: 0.9603 - val_loss: 0.0621 - val_accuracy: 0.9739\n",
            "WARNING:tensorflow:6 out of the last 11 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd07124730> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 3s 59ms/step - loss: 2.1261 - accuracy: 0.3588 - val_loss: 1.3938 - val_accuracy: 0.6609\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 3s 56ms/step - loss: 0.8140 - accuracy: 0.7563 - val_loss: 0.4019 - val_accuracy: 0.8870\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 3s 56ms/step - loss: 0.3944 - accuracy: 0.8762 - val_loss: 0.2559 - val_accuracy: 0.9043\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 3s 56ms/step - loss: 0.2888 - accuracy: 0.9072 - val_loss: 0.2806 - val_accuracy: 0.9391\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 3s 54ms/step - loss: 0.2304 - accuracy: 0.9255 - val_loss: 0.2521 - val_accuracy: 0.9304\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 3s 52ms/step - loss: 0.1861 - accuracy: 0.9410 - val_loss: 0.2543 - val_accuracy: 0.9130\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 3s 52ms/step - loss: 0.1534 - accuracy: 0.9555 - val_loss: 0.1649 - val_accuracy: 0.9565\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 3s 50ms/step - loss: 0.1285 - accuracy: 0.9574 - val_loss: 0.1524 - val_accuracy: 0.9304\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 3s 54ms/step - loss: 0.1297 - accuracy: 0.9526 - val_loss: 0.0980 - val_accuracy: 0.9826\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 3s 53ms/step - loss: 0.1115 - accuracy: 0.9594 - val_loss: 0.1449 - val_accuracy: 0.9565\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 3s 67ms/step - loss: 2.1541 - accuracy: 0.3159 - val_loss: 1.4491 - val_accuracy: 0.8087\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 3s 53ms/step - loss: 0.8208 - accuracy: 0.7797 - val_loss: 0.4245 - val_accuracy: 0.8783\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 3s 50ms/step - loss: 0.4048 - accuracy: 0.8754 - val_loss: 0.3395 - val_accuracy: 0.9043\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 3s 50ms/step - loss: 0.2617 - accuracy: 0.9198 - val_loss: 0.3129 - val_accuracy: 0.9130\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 3s 53ms/step - loss: 0.2425 - accuracy: 0.9208 - val_loss: 0.1339 - val_accuracy: 0.9739\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 3s 51ms/step - loss: 0.1466 - accuracy: 0.9507 - val_loss: 0.1352 - val_accuracy: 0.9739\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 3s 49ms/step - loss: 0.1274 - accuracy: 0.9623 - val_loss: 0.0966 - val_accuracy: 0.9826\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 3s 54ms/step - loss: 0.1051 - accuracy: 0.9633 - val_loss: 0.0789 - val_accuracy: 0.9913\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 2s 46ms/step - loss: 0.1186 - accuracy: 0.9623 - val_loss: 0.1362 - val_accuracy: 0.9739\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 3s 51ms/step - loss: 0.0938 - accuracy: 0.9662 - val_loss: 0.0719 - val_accuracy: 0.9913\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 3s 50ms/step - loss: 2.1276 - accuracy: 0.2889 - val_loss: 1.3887 - val_accuracy: 0.7478\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 3s 56ms/step - loss: 0.8889 - accuracy: 0.7391 - val_loss: 0.4549 - val_accuracy: 0.8609\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 3s 55ms/step - loss: 0.4256 - accuracy: 0.8802 - val_loss: 0.3769 - val_accuracy: 0.8609\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 3s 54ms/step - loss: 0.3094 - accuracy: 0.8957 - val_loss: 0.2566 - val_accuracy: 0.9130\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 3s 56ms/step - loss: 0.2141 - accuracy: 0.9314 - val_loss: 0.1947 - val_accuracy: 0.9565\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 3s 56ms/step - loss: 0.1897 - accuracy: 0.9353 - val_loss: 0.1651 - val_accuracy: 0.9478\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 3s 56ms/step - loss: 0.1542 - accuracy: 0.9565 - val_loss: 0.1253 - val_accuracy: 0.9565\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 3s 53ms/step - loss: 0.1240 - accuracy: 0.9652 - val_loss: 0.1280 - val_accuracy: 0.9478\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 3s 56ms/step - loss: 0.1003 - accuracy: 0.9681 - val_loss: 0.1378 - val_accuracy: 0.9478\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 3s 53ms/step - loss: 0.1080 - accuracy: 0.9662 - val_loss: 0.1498 - val_accuracy: 0.9652\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 3s 55ms/step - loss: 2.1980 - accuracy: 0.2338 - val_loss: 1.7585 - val_accuracy: 0.5565\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 3s 51ms/step - loss: 0.9885 - accuracy: 0.7082 - val_loss: 0.4708 - val_accuracy: 0.8870\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 3s 51ms/step - loss: 0.4261 - accuracy: 0.8705 - val_loss: 0.2818 - val_accuracy: 0.9304\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 3s 52ms/step - loss: 0.3210 - accuracy: 0.8995 - val_loss: 0.1733 - val_accuracy: 0.9652\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 3s 58ms/step - loss: 0.2414 - accuracy: 0.9179 - val_loss: 0.1003 - val_accuracy: 0.9739\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 3s 56ms/step - loss: 0.2009 - accuracy: 0.9353 - val_loss: 0.0834 - val_accuracy: 0.9826\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 3s 58ms/step - loss: 0.1532 - accuracy: 0.9507 - val_loss: 0.0613 - val_accuracy: 0.9913\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 3s 54ms/step - loss: 0.1097 - accuracy: 0.9681 - val_loss: 0.0375 - val_accuracy: 1.0000\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 3s 58ms/step - loss: 0.1133 - accuracy: 0.9652 - val_loss: 0.0275 - val_accuracy: 1.0000\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 3s 54ms/step - loss: 0.0953 - accuracy: 0.9671 - val_loss: 0.0211 - val_accuracy: 1.0000\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 1s 27ms/step - loss: 2.9943 - accuracy: 0.0928 - val_loss: 2.3045 - val_accuracy: 0.1043\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3130 - accuracy: 0.0977 - val_loss: 2.3119 - val_accuracy: 0.1043\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 1s 23ms/step - loss: 2.3085 - accuracy: 0.1103 - val_loss: 2.3063 - val_accuracy: 0.1043\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 1s 23ms/step - loss: 2.3098 - accuracy: 0.1006 - val_loss: 2.3081 - val_accuracy: 0.0696\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 1s 23ms/step - loss: 2.3083 - accuracy: 0.1035 - val_loss: 2.3117 - val_accuracy: 0.1043\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 1s 23ms/step - loss: 2.3086 - accuracy: 0.1151 - val_loss: 2.3141 - val_accuracy: 0.1043\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3085 - accuracy: 0.1006 - val_loss: 2.3110 - val_accuracy: 0.1043\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 1s 25ms/step - loss: 2.3113 - accuracy: 0.1025 - val_loss: 2.3110 - val_accuracy: 0.1043\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 1s 25ms/step - loss: 2.3083 - accuracy: 0.0938 - val_loss: 2.3108 - val_accuracy: 0.0696\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 1s 25ms/step - loss: 2.3096 - accuracy: 0.0909 - val_loss: 2.3085 - val_accuracy: 0.1130\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 1s 26ms/step - loss: 2.3119 - accuracy: 0.1074 - val_loss: 2.3148 - val_accuracy: 0.1043\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3091 - accuracy: 0.0812 - val_loss: 2.2950 - val_accuracy: 0.1043\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 1s 26ms/step - loss: 2.3100 - accuracy: 0.1054 - val_loss: 2.3060 - val_accuracy: 0.1043\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 1s 26ms/step - loss: 2.3097 - accuracy: 0.0948 - val_loss: 2.2993 - val_accuracy: 0.1043\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 1s 23ms/step - loss: 2.3088 - accuracy: 0.1054 - val_loss: 2.3080 - val_accuracy: 0.1043\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 1s 25ms/step - loss: 2.7852 - accuracy: 0.0919 - val_loss: 2.3182 - val_accuracy: 0.0696\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3108 - accuracy: 0.0928 - val_loss: 2.3162 - val_accuracy: 0.0696\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3124 - accuracy: 0.1054 - val_loss: 2.3090 - val_accuracy: 0.1043\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3148 - accuracy: 0.0841 - val_loss: 2.3012 - val_accuracy: 0.1043\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3109 - accuracy: 0.0919 - val_loss: 2.3075 - val_accuracy: 0.1043\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 1s 23ms/step - loss: 2.3101 - accuracy: 0.1093 - val_loss: 2.3087 - val_accuracy: 0.1304\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3121 - accuracy: 0.0851 - val_loss: 2.3071 - val_accuracy: 0.0696\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 1s 23ms/step - loss: 2.3127 - accuracy: 0.1025 - val_loss: 2.3027 - val_accuracy: 0.0696\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 1s 23ms/step - loss: 2.3102 - accuracy: 0.0948 - val_loss: 2.3020 - val_accuracy: 0.1043\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3108 - accuracy: 0.0996 - val_loss: 2.3060 - val_accuracy: 0.0783\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3166 - accuracy: 0.0832 - val_loss: 2.3082 - val_accuracy: 0.1043\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3126 - accuracy: 0.0774 - val_loss: 2.3102 - val_accuracy: 0.1043\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3124 - accuracy: 0.1006 - val_loss: 2.3016 - val_accuracy: 0.1304\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 1s 23ms/step - loss: 2.3126 - accuracy: 0.0909 - val_loss: 2.2974 - val_accuracy: 0.1304\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 1s 22ms/step - loss: 2.3100 - accuracy: 0.1035 - val_loss: 2.3047 - val_accuracy: 0.1304\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 1s 25ms/step - loss: 3.1248 - accuracy: 0.0995 - val_loss: 2.3012 - val_accuracy: 0.1130\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 1s 22ms/step - loss: 2.3092 - accuracy: 0.0918 - val_loss: 2.2979 - val_accuracy: 0.1304\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3111 - accuracy: 0.0986 - val_loss: 2.2985 - val_accuracy: 0.1304\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 1s 22ms/step - loss: 2.3097 - accuracy: 0.1092 - val_loss: 2.3081 - val_accuracy: 0.1043\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 1s 23ms/step - loss: 2.3112 - accuracy: 0.0995 - val_loss: 2.3056 - val_accuracy: 0.1130\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3097 - accuracy: 0.1101 - val_loss: 2.3023 - val_accuracy: 0.1043\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3071 - accuracy: 0.0918 - val_loss: 2.2986 - val_accuracy: 0.1304\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3081 - accuracy: 0.1014 - val_loss: 2.3090 - val_accuracy: 0.1043\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3102 - accuracy: 0.0995 - val_loss: 2.2980 - val_accuracy: 0.1304\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 1s 23ms/step - loss: 2.3132 - accuracy: 0.0841 - val_loss: 2.2962 - val_accuracy: 0.1130\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3110 - accuracy: 0.1024 - val_loss: 2.3061 - val_accuracy: 0.1043\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3130 - accuracy: 0.0860 - val_loss: 2.3095 - val_accuracy: 0.1130\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3137 - accuracy: 0.0986 - val_loss: 2.2969 - val_accuracy: 0.1304\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 1s 25ms/step - loss: 2.3081 - accuracy: 0.1082 - val_loss: 2.3090 - val_accuracy: 0.1043\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3109 - accuracy: 0.0947 - val_loss: 2.3126 - val_accuracy: 0.1043\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 1s 25ms/step - loss: 3.0862 - accuracy: 0.0966 - val_loss: 2.3097 - val_accuracy: 0.0957\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 1s 23ms/step - loss: 2.3133 - accuracy: 0.0947 - val_loss: 2.3102 - val_accuracy: 0.0957\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 1s 23ms/step - loss: 2.3093 - accuracy: 0.0763 - val_loss: 2.3010 - val_accuracy: 0.1304\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 1s 22ms/step - loss: 2.3152 - accuracy: 0.0754 - val_loss: 2.2988 - val_accuracy: 0.1043\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3112 - accuracy: 0.0899 - val_loss: 2.3084 - val_accuracy: 0.0957\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 1s 23ms/step - loss: 2.3109 - accuracy: 0.0928 - val_loss: 2.2989 - val_accuracy: 0.1043\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 1s 22ms/step - loss: 2.3115 - accuracy: 0.0937 - val_loss: 2.3077 - val_accuracy: 0.0783\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 1s 25ms/step - loss: 2.3099 - accuracy: 0.0879 - val_loss: 2.3087 - val_accuracy: 0.0783\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 1s 23ms/step - loss: 2.3096 - accuracy: 0.0995 - val_loss: 2.3120 - val_accuracy: 0.0696\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3102 - accuracy: 0.0773 - val_loss: 2.3140 - val_accuracy: 0.0696\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 1s 23ms/step - loss: 2.3099 - accuracy: 0.1024 - val_loss: 2.3044 - val_accuracy: 0.0957\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3123 - accuracy: 0.0899 - val_loss: 2.3129 - val_accuracy: 0.0957\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3116 - accuracy: 0.0812 - val_loss: 2.3126 - val_accuracy: 0.1130\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 1s 22ms/step - loss: 2.3090 - accuracy: 0.0918 - val_loss: 2.3011 - val_accuracy: 0.1043\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3134 - accuracy: 0.0918 - val_loss: 2.3151 - val_accuracy: 0.1130\n",
            "Epoch 1/15\n",
            "52/52 [==============================] - 1s 26ms/step - loss: 2.8760 - accuracy: 0.0879 - val_loss: 2.3152 - val_accuracy: 0.0870\n",
            "Epoch 2/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3139 - accuracy: 0.1101 - val_loss: 2.3141 - val_accuracy: 0.1043\n",
            "Epoch 3/15\n",
            "52/52 [==============================] - 1s 25ms/step - loss: 2.3100 - accuracy: 0.0928 - val_loss: 2.3116 - val_accuracy: 0.0609\n",
            "Epoch 4/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3101 - accuracy: 0.0947 - val_loss: 2.3033 - val_accuracy: 0.0348\n",
            "Epoch 5/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3082 - accuracy: 0.0947 - val_loss: 2.3293 - val_accuracy: 0.0609\n",
            "Epoch 6/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3141 - accuracy: 0.0812 - val_loss: 2.3190 - val_accuracy: 0.0348\n",
            "Epoch 7/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3095 - accuracy: 0.0986 - val_loss: 2.2902 - val_accuracy: 0.1043\n",
            "Epoch 8/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3118 - accuracy: 0.0928 - val_loss: 2.3149 - val_accuracy: 0.0870\n",
            "Epoch 9/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3088 - accuracy: 0.0986 - val_loss: 2.2919 - val_accuracy: 0.1043\n",
            "Epoch 10/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3106 - accuracy: 0.0860 - val_loss: 2.3055 - val_accuracy: 0.0870\n",
            "Epoch 11/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3080 - accuracy: 0.0986 - val_loss: 2.3022 - val_accuracy: 0.1043\n",
            "Epoch 12/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3128 - accuracy: 0.0860 - val_loss: 2.3149 - val_accuracy: 0.1043\n",
            "Epoch 13/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3078 - accuracy: 0.1092 - val_loss: 2.3164 - val_accuracy: 0.0870\n",
            "Epoch 14/15\n",
            "52/52 [==============================] - 1s 23ms/step - loss: 2.3106 - accuracy: 0.0966 - val_loss: 2.3126 - val_accuracy: 0.0870\n",
            "Epoch 15/15\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 2.3124 - accuracy: 0.0976 - val_loss: 2.3385 - val_accuracy: 0.0870\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 0s 17ms/step - loss: 2.3020 - accuracy: 0.0851 - val_loss: 2.2830 - val_accuracy: 0.1565\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 2.1730 - accuracy: 0.1789 - val_loss: 1.8430 - val_accuracy: 0.2522\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.7335 - accuracy: 0.3598 - val_loss: 1.2133 - val_accuracy: 0.5913\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.3633 - accuracy: 0.5155 - val_loss: 0.8675 - val_accuracy: 0.7217\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.2002 - accuracy: 0.5493 - val_loss: 0.7580 - val_accuracy: 0.8174\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 0s 9ms/step - loss: 1.0703 - accuracy: 0.6238 - val_loss: 0.5906 - val_accuracy: 0.8261\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 0s 9ms/step - loss: 1.0494 - accuracy: 0.6364 - val_loss: 0.6930 - val_accuracy: 0.8522\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 0.9974 - accuracy: 0.6518 - val_loss: 0.5184 - val_accuracy: 0.8870\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 0.9675 - accuracy: 0.6518 - val_loss: 0.5006 - val_accuracy: 0.8957\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 0.9297 - accuracy: 0.6770 - val_loss: 0.4856 - val_accuracy: 0.8783\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 0.8908 - accuracy: 0.6828 - val_loss: 0.5418 - val_accuracy: 0.8696\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 0.8633 - accuracy: 0.6876 - val_loss: 0.3966 - val_accuracy: 0.9217\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 0s 10ms/step - loss: 0.8730 - accuracy: 0.6886 - val_loss: 0.4518 - val_accuracy: 0.8957\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 0.8343 - accuracy: 0.7021 - val_loss: 0.4210 - val_accuracy: 0.9130\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 0.8279 - accuracy: 0.7157 - val_loss: 0.4341 - val_accuracy: 0.8957\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 0s 14ms/step - loss: 2.2999 - accuracy: 0.1170 - val_loss: 2.2414 - val_accuracy: 0.3130\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 2.0757 - accuracy: 0.2157 - val_loss: 1.6658 - val_accuracy: 0.5130\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.5840 - accuracy: 0.4207 - val_loss: 1.1372 - val_accuracy: 0.6609\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 0s 11ms/step - loss: 1.3714 - accuracy: 0.5242 - val_loss: 0.8824 - val_accuracy: 0.7217\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.2154 - accuracy: 0.5513 - val_loss: 0.6399 - val_accuracy: 0.8348\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 0s 9ms/step - loss: 1.1310 - accuracy: 0.5822 - val_loss: 0.6135 - val_accuracy: 0.8348\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.0523 - accuracy: 0.6151 - val_loss: 0.6176 - val_accuracy: 0.8609\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 0.9565 - accuracy: 0.6567 - val_loss: 0.5756 - val_accuracy: 0.8609\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 0.9466 - accuracy: 0.6683 - val_loss: 0.5307 - val_accuracy: 0.8870\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 0.9209 - accuracy: 0.6596 - val_loss: 0.4348 - val_accuracy: 0.9130\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 0.9037 - accuracy: 0.6934 - val_loss: 0.4217 - val_accuracy: 0.9043\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 0.8329 - accuracy: 0.7079 - val_loss: 0.3930 - val_accuracy: 0.9565\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 0.8897 - accuracy: 0.6944 - val_loss: 0.4059 - val_accuracy: 0.9130\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 0s 11ms/step - loss: 0.7704 - accuracy: 0.7263 - val_loss: 0.3894 - val_accuracy: 0.9217\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 0.8246 - accuracy: 0.7041 - val_loss: 0.4552 - val_accuracy: 0.9043\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 0s 13ms/step - loss: 2.3080 - accuracy: 0.0995 - val_loss: 2.3007 - val_accuracy: 0.1043\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 2.3085 - accuracy: 0.1266 - val_loss: 2.2940 - val_accuracy: 0.1130\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 2.2761 - accuracy: 0.1382 - val_loss: 2.1714 - val_accuracy: 0.3130\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 2.0681 - accuracy: 0.1894 - val_loss: 1.7432 - val_accuracy: 0.3913\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 0s 9ms/step - loss: 1.8829 - accuracy: 0.2512 - val_loss: 1.5816 - val_accuracy: 0.4696\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.6316 - accuracy: 0.3981 - val_loss: 1.1896 - val_accuracy: 0.7043\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 1.5530 - accuracy: 0.3971 - val_loss: 1.0402 - val_accuracy: 0.7043\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.4283 - accuracy: 0.4541 - val_loss: 1.0072 - val_accuracy: 0.7565\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.3908 - accuracy: 0.4638 - val_loss: 0.9989 - val_accuracy: 0.7391\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 0s 9ms/step - loss: 1.3151 - accuracy: 0.5121 - val_loss: 0.7763 - val_accuracy: 0.7913\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 0s 10ms/step - loss: 1.3530 - accuracy: 0.4870 - val_loss: 0.8008 - val_accuracy: 0.8348\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.2643 - accuracy: 0.5237 - val_loss: 0.9194 - val_accuracy: 0.8174\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.2013 - accuracy: 0.5546 - val_loss: 0.6965 - val_accuracy: 0.8609\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.1982 - accuracy: 0.5585 - val_loss: 0.7643 - val_accuracy: 0.8609\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.2379 - accuracy: 0.5401 - val_loss: 0.8661 - val_accuracy: 0.7565\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 0s 14ms/step - loss: 2.3080 - accuracy: 0.1072 - val_loss: 2.2876 - val_accuracy: 0.1130\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 0s 11ms/step - loss: 2.2127 - accuracy: 0.1700 - val_loss: 1.9326 - val_accuracy: 0.2435\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.7938 - accuracy: 0.3188 - val_loss: 1.2411 - val_accuracy: 0.6957\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.3796 - accuracy: 0.4986 - val_loss: 0.9258 - val_accuracy: 0.7043\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.2013 - accuracy: 0.5700 - val_loss: 0.8030 - val_accuracy: 0.8000\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.0969 - accuracy: 0.5913 - val_loss: 0.6348 - val_accuracy: 0.8174\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 0s 9ms/step - loss: 0.9516 - accuracy: 0.6570 - val_loss: 0.5073 - val_accuracy: 0.8696\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 0s 9ms/step - loss: 0.8655 - accuracy: 0.6831 - val_loss: 0.4614 - val_accuracy: 0.9130\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 0.9242 - accuracy: 0.6589 - val_loss: 0.4834 - val_accuracy: 0.8696\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 0.8500 - accuracy: 0.6937 - val_loss: 0.4443 - val_accuracy: 0.9304\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 0.8167 - accuracy: 0.7179 - val_loss: 0.4487 - val_accuracy: 0.8783\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 0s 9ms/step - loss: 0.8486 - accuracy: 0.6889 - val_loss: 0.4447 - val_accuracy: 0.8957\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 0.7733 - accuracy: 0.7227 - val_loss: 0.4238 - val_accuracy: 0.9043\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 0.7657 - accuracy: 0.7130 - val_loss: 0.3873 - val_accuracy: 0.8870\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 0.7767 - accuracy: 0.7208 - val_loss: 0.4238 - val_accuracy: 0.8783\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 0s 15ms/step - loss: 2.3034 - accuracy: 0.1024 - val_loss: 2.2640 - val_accuracy: 0.2087\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 2.1120 - accuracy: 0.2377 - val_loss: 1.7420 - val_accuracy: 0.4609\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 0s 9ms/step - loss: 1.5153 - accuracy: 0.4609 - val_loss: 0.9679 - val_accuracy: 0.7565\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 0s 9ms/step - loss: 1.1733 - accuracy: 0.5913 - val_loss: 0.7744 - val_accuracy: 0.7652\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 1.0338 - accuracy: 0.6290 - val_loss: 0.5753 - val_accuracy: 0.8696\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 0.9706 - accuracy: 0.6531 - val_loss: 0.6715 - val_accuracy: 0.8174\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 0s 9ms/step - loss: 0.8971 - accuracy: 0.6831 - val_loss: 0.4276 - val_accuracy: 0.8957\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 0s 9ms/step - loss: 0.7986 - accuracy: 0.7169 - val_loss: 0.3456 - val_accuracy: 0.9304\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 0s 10ms/step - loss: 0.8452 - accuracy: 0.7101 - val_loss: 0.4092 - val_accuracy: 0.9043\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 0.7864 - accuracy: 0.7208 - val_loss: 0.4183 - val_accuracy: 0.8609\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 0s 7ms/step - loss: 0.7628 - accuracy: 0.7198 - val_loss: 0.4963 - val_accuracy: 0.8870\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 0.7712 - accuracy: 0.7159 - val_loss: 0.4006 - val_accuracy: 0.9043\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 0s 8ms/step - loss: 0.7589 - accuracy: 0.7440 - val_loss: 0.3744 - val_accuracy: 0.8783\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 0s 9ms/step - loss: 0.7465 - accuracy: 0.7517 - val_loss: 0.2974 - val_accuracy: 0.9130\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 0s 9ms/step - loss: 0.7121 - accuracy: 0.7594 - val_loss: 0.4621 - val_accuracy: 0.8696\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 2s 79ms/step - loss: 2.2971 - accuracy: 0.1228 - val_loss: 2.2814 - val_accuracy: 0.1913\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 1s 71ms/step - loss: 2.2502 - accuracy: 0.2331 - val_loss: 2.1751 - val_accuracy: 0.3565\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 2s 73ms/step - loss: 2.0992 - accuracy: 0.2727 - val_loss: 1.8721 - val_accuracy: 0.5826\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 2s 74ms/step - loss: 1.8048 - accuracy: 0.3859 - val_loss: 1.4131 - val_accuracy: 0.6348\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 1s 70ms/step - loss: 1.4557 - accuracy: 0.5097 - val_loss: 1.0297 - val_accuracy: 0.7478\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 1s 71ms/step - loss: 1.1856 - accuracy: 0.6161 - val_loss: 0.7558 - val_accuracy: 0.8174\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 2s 72ms/step - loss: 0.9418 - accuracy: 0.6857 - val_loss: 0.5945 - val_accuracy: 0.8783\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 2s 73ms/step - loss: 0.8858 - accuracy: 0.7186 - val_loss: 0.4782 - val_accuracy: 0.8870\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 2s 72ms/step - loss: 0.7622 - accuracy: 0.7553 - val_loss: 0.4082 - val_accuracy: 0.9130\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 2s 72ms/step - loss: 0.7083 - accuracy: 0.7679 - val_loss: 0.3525 - val_accuracy: 0.9304\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 2s 73ms/step - loss: 0.6743 - accuracy: 0.7650 - val_loss: 0.3508 - val_accuracy: 0.9043\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 2s 73ms/step - loss: 0.6170 - accuracy: 0.7969 - val_loss: 0.3030 - val_accuracy: 0.9391\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 1s 70ms/step - loss: 0.5970 - accuracy: 0.8085 - val_loss: 0.2768 - val_accuracy: 0.9478\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 2s 74ms/step - loss: 0.5673 - accuracy: 0.8124 - val_loss: 0.2905 - val_accuracy: 0.9391\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 1s 71ms/step - loss: 0.5319 - accuracy: 0.8104 - val_loss: 0.2499 - val_accuracy: 0.9391\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 2s 78ms/step - loss: 2.2952 - accuracy: 0.1412 - val_loss: 2.2730 - val_accuracy: 0.2000\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 2s 73ms/step - loss: 2.2378 - accuracy: 0.2466 - val_loss: 2.1404 - val_accuracy: 0.4087\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 2s 75ms/step - loss: 1.9725 - accuracy: 0.3859 - val_loss: 1.6306 - val_accuracy: 0.7391\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 2s 73ms/step - loss: 1.4986 - accuracy: 0.5135 - val_loss: 1.0242 - val_accuracy: 0.7913\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 2s 73ms/step - loss: 1.1525 - accuracy: 0.6354 - val_loss: 0.7288 - val_accuracy: 0.8348\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 2s 75ms/step - loss: 0.9508 - accuracy: 0.6760 - val_loss: 0.5912 - val_accuracy: 0.8348\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 2s 74ms/step - loss: 0.8516 - accuracy: 0.7070 - val_loss: 0.5033 - val_accuracy: 0.8783\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 2s 73ms/step - loss: 0.7283 - accuracy: 0.7524 - val_loss: 0.4204 - val_accuracy: 0.9043\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 2s 72ms/step - loss: 0.6853 - accuracy: 0.7689 - val_loss: 0.4005 - val_accuracy: 0.8783\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 2s 74ms/step - loss: 0.6292 - accuracy: 0.8066 - val_loss: 0.3396 - val_accuracy: 0.9304\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 2s 75ms/step - loss: 0.5934 - accuracy: 0.8037 - val_loss: 0.3212 - val_accuracy: 0.9304\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 2s 73ms/step - loss: 0.5512 - accuracy: 0.8162 - val_loss: 0.2978 - val_accuracy: 0.9478\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 2s 75ms/step - loss: 0.5356 - accuracy: 0.8259 - val_loss: 0.2762 - val_accuracy: 0.9391\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 2s 73ms/step - loss: 0.4730 - accuracy: 0.8356 - val_loss: 0.2563 - val_accuracy: 0.9391\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 2s 72ms/step - loss: 0.4587 - accuracy: 0.8501 - val_loss: 0.2355 - val_accuracy: 0.9478\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 2s 78ms/step - loss: 2.2963 - accuracy: 0.1391 - val_loss: 2.2776 - val_accuracy: 0.2174\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 2s 74ms/step - loss: 2.2391 - accuracy: 0.2444 - val_loss: 2.1405 - val_accuracy: 0.3478\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 2s 74ms/step - loss: 1.9751 - accuracy: 0.3816 - val_loss: 1.6407 - val_accuracy: 0.6174\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 1s 71ms/step - loss: 1.4861 - accuracy: 0.5208 - val_loss: 1.1168 - val_accuracy: 0.7478\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 2s 72ms/step - loss: 1.1877 - accuracy: 0.6048 - val_loss: 0.8068 - val_accuracy: 0.8087\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 2s 74ms/step - loss: 0.9681 - accuracy: 0.6783 - val_loss: 0.6124 - val_accuracy: 0.8696\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 2s 73ms/step - loss: 0.8392 - accuracy: 0.7285 - val_loss: 0.4812 - val_accuracy: 0.8609\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 2s 74ms/step - loss: 0.7670 - accuracy: 0.7391 - val_loss: 0.3996 - val_accuracy: 0.9043\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 2s 74ms/step - loss: 0.6424 - accuracy: 0.7836 - val_loss: 0.3354 - val_accuracy: 0.9217\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 2s 74ms/step - loss: 0.5944 - accuracy: 0.8126 - val_loss: 0.3174 - val_accuracy: 0.9304\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 2s 72ms/step - loss: 0.5546 - accuracy: 0.8174 - val_loss: 0.2880 - val_accuracy: 0.9217\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 2s 75ms/step - loss: 0.5079 - accuracy: 0.8386 - val_loss: 0.2548 - val_accuracy: 0.9130\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 2s 74ms/step - loss: 0.5341 - accuracy: 0.8126 - val_loss: 0.2340 - val_accuracy: 0.9304\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 2s 72ms/step - loss: 0.4413 - accuracy: 0.8473 - val_loss: 0.2372 - val_accuracy: 0.9565\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 1s 71ms/step - loss: 0.4807 - accuracy: 0.8444 - val_loss: 0.2075 - val_accuracy: 0.9565\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 2s 81ms/step - loss: 2.2913 - accuracy: 0.1362 - val_loss: 2.2640 - val_accuracy: 0.1130\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 2s 73ms/step - loss: 2.2201 - accuracy: 0.2860 - val_loss: 2.1026 - val_accuracy: 0.5043\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 2s 72ms/step - loss: 1.9764 - accuracy: 0.4232 - val_loss: 1.6331 - val_accuracy: 0.7391\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 2s 73ms/step - loss: 1.5365 - accuracy: 0.5208 - val_loss: 1.1051 - val_accuracy: 0.7043\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 2s 72ms/step - loss: 1.1854 - accuracy: 0.5981 - val_loss: 0.7993 - val_accuracy: 0.8000\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 2s 73ms/step - loss: 0.9992 - accuracy: 0.6792 - val_loss: 0.6502 - val_accuracy: 0.8348\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 2s 72ms/step - loss: 0.8345 - accuracy: 0.7353 - val_loss: 0.5040 - val_accuracy: 0.8522\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 2s 74ms/step - loss: 0.7462 - accuracy: 0.7459 - val_loss: 0.4463 - val_accuracy: 0.9043\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 2s 73ms/step - loss: 0.6494 - accuracy: 0.7662 - val_loss: 0.4150 - val_accuracy: 0.9043\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 1s 71ms/step - loss: 0.6309 - accuracy: 0.8135 - val_loss: 0.3167 - val_accuracy: 0.9478\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 2s 73ms/step - loss: 0.5450 - accuracy: 0.8213 - val_loss: 0.2897 - val_accuracy: 0.9565\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 2s 73ms/step - loss: 0.5461 - accuracy: 0.8164 - val_loss: 0.2589 - val_accuracy: 0.9652\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 1s 71ms/step - loss: 0.4892 - accuracy: 0.8464 - val_loss: 0.2477 - val_accuracy: 0.9652\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 2s 73ms/step - loss: 0.4649 - accuracy: 0.8618 - val_loss: 0.2343 - val_accuracy: 0.9565\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 2s 74ms/step - loss: 0.4687 - accuracy: 0.8599 - val_loss: 0.2267 - val_accuracy: 0.9478\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 2s 78ms/step - loss: 2.2953 - accuracy: 0.1507 - val_loss: 2.2895 - val_accuracy: 0.0783\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 2s 72ms/step - loss: 2.2415 - accuracy: 0.2116 - val_loss: 2.1855 - val_accuracy: 0.1739\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 1s 71ms/step - loss: 2.0003 - accuracy: 0.3700 - val_loss: 1.7919 - val_accuracy: 0.5217\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 2s 74ms/step - loss: 1.6006 - accuracy: 0.4792 - val_loss: 1.2181 - val_accuracy: 0.7478\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 1s 70ms/step - loss: 1.2307 - accuracy: 0.5845 - val_loss: 0.8512 - val_accuracy: 0.8348\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 2s 73ms/step - loss: 1.0254 - accuracy: 0.6609 - val_loss: 0.6697 - val_accuracy: 0.8957\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 2s 74ms/step - loss: 0.8738 - accuracy: 0.6976 - val_loss: 0.5113 - val_accuracy: 0.9304\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 2s 72ms/step - loss: 0.7999 - accuracy: 0.7343 - val_loss: 0.4452 - val_accuracy: 0.8957\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 2s 73ms/step - loss: 0.7348 - accuracy: 0.7440 - val_loss: 0.4088 - val_accuracy: 0.9391\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 2s 72ms/step - loss: 0.7470 - accuracy: 0.7372 - val_loss: 0.3258 - val_accuracy: 0.9391\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 2s 74ms/step - loss: 0.6617 - accuracy: 0.7691 - val_loss: 0.3079 - val_accuracy: 0.9565\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 2s 73ms/step - loss: 0.6456 - accuracy: 0.7855 - val_loss: 0.2814 - val_accuracy: 0.9565\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 2s 73ms/step - loss: 0.5691 - accuracy: 0.8029 - val_loss: 0.2262 - val_accuracy: 0.9565\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 2s 72ms/step - loss: 0.5700 - accuracy: 0.8164 - val_loss: 0.2754 - val_accuracy: 0.9565\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 2s 75ms/step - loss: 0.5092 - accuracy: 0.8309 - val_loss: 0.2230 - val_accuracy: 0.9304\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 2s 97ms/step - loss: 5.5936 - accuracy: 0.1103 - val_loss: 2.3187 - val_accuracy: 0.1043\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 2s 92ms/step - loss: 2.3075 - accuracy: 0.1064 - val_loss: 2.3011 - val_accuracy: 0.1043\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 2s 92ms/step - loss: 2.3024 - accuracy: 0.1141 - val_loss: 2.3091 - val_accuracy: 0.1043\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 2s 93ms/step - loss: 2.3026 - accuracy: 0.1141 - val_loss: 2.3058 - val_accuracy: 0.1043\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 2s 93ms/step - loss: 2.3039 - accuracy: 0.0967 - val_loss: 2.3094 - val_accuracy: 0.1043\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 2s 92ms/step - loss: 2.3042 - accuracy: 0.1141 - val_loss: 2.3086 - val_accuracy: 0.1043\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 2s 92ms/step - loss: 2.3044 - accuracy: 0.0986 - val_loss: 2.3077 - val_accuracy: 0.1043\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 2s 96ms/step - loss: 2.3051 - accuracy: 0.0919 - val_loss: 2.3064 - val_accuracy: 0.1043\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 2s 93ms/step - loss: 2.3070 - accuracy: 0.0977 - val_loss: 2.3086 - val_accuracy: 0.1043\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 2s 93ms/step - loss: 2.3031 - accuracy: 0.1141 - val_loss: 2.3098 - val_accuracy: 0.1043\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 2s 94ms/step - loss: 2.3028 - accuracy: 0.1141 - val_loss: 2.3048 - val_accuracy: 0.1043\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 2s 93ms/step - loss: 2.3038 - accuracy: 0.1006 - val_loss: 2.3034 - val_accuracy: 0.1043\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 2s 92ms/step - loss: 2.3050 - accuracy: 0.1064 - val_loss: 2.3118 - val_accuracy: 0.0696\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 2s 96ms/step - loss: 2.3045 - accuracy: 0.1161 - val_loss: 2.3042 - val_accuracy: 0.1043\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 2s 93ms/step - loss: 2.3032 - accuracy: 0.1141 - val_loss: 2.3081 - val_accuracy: 0.1043\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 2s 98ms/step - loss: 9.3687 - accuracy: 0.0851 - val_loss: 2.3117 - val_accuracy: 0.1130\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 2s 93ms/step - loss: 2.3074 - accuracy: 0.1006 - val_loss: 2.3071 - val_accuracy: 0.0696\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 2s 93ms/step - loss: 2.3063 - accuracy: 0.1006 - val_loss: 2.3064 - val_accuracy: 0.1304\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 2s 94ms/step - loss: 2.3055 - accuracy: 0.0899 - val_loss: 2.3039 - val_accuracy: 0.1043\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 2s 93ms/step - loss: 2.3046 - accuracy: 0.1074 - val_loss: 2.3082 - val_accuracy: 0.0696\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 2s 97ms/step - loss: 2.3053 - accuracy: 0.1083 - val_loss: 2.3062 - val_accuracy: 0.1043\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 2s 93ms/step - loss: 2.3058 - accuracy: 0.0977 - val_loss: 2.3084 - val_accuracy: 0.0696\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 2s 92ms/step - loss: 2.3064 - accuracy: 0.0919 - val_loss: 2.3058 - val_accuracy: 0.1043\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 2s 94ms/step - loss: 2.3064 - accuracy: 0.0967 - val_loss: 2.3068 - val_accuracy: 0.0696\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 2s 94ms/step - loss: 2.3065 - accuracy: 0.0890 - val_loss: 2.3078 - val_accuracy: 0.1043\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 2s 93ms/step - loss: 2.3050 - accuracy: 0.1074 - val_loss: 2.3065 - val_accuracy: 0.0696\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 2s 93ms/step - loss: 2.3078 - accuracy: 0.0909 - val_loss: 2.3089 - val_accuracy: 0.0696\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 2s 96ms/step - loss: 2.3046 - accuracy: 0.0948 - val_loss: 2.3017 - val_accuracy: 0.1304\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 2s 92ms/step - loss: 2.3071 - accuracy: 0.0967 - val_loss: 2.3072 - val_accuracy: 0.1043\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 2s 94ms/step - loss: 2.3076 - accuracy: 0.0977 - val_loss: 2.3076 - val_accuracy: 0.1304\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 2s 105ms/step - loss: 6.9885 - accuracy: 0.0899 - val_loss: 2.3035 - val_accuracy: 0.1043\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 2s 95ms/step - loss: 2.3096 - accuracy: 0.0899 - val_loss: 2.2976 - val_accuracy: 0.1304\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 2s 94ms/step - loss: 2.3043 - accuracy: 0.1043 - val_loss: 2.3012 - val_accuracy: 0.1043\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 2s 95ms/step - loss: 2.3052 - accuracy: 0.1111 - val_loss: 2.2992 - val_accuracy: 0.1043\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 2s 95ms/step - loss: 2.3046 - accuracy: 0.0879 - val_loss: 2.3021 - val_accuracy: 0.1043\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 2s 95ms/step - loss: 2.3070 - accuracy: 0.1111 - val_loss: 2.3044 - val_accuracy: 0.1043\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 2s 98ms/step - loss: 2.3054 - accuracy: 0.0957 - val_loss: 2.2960 - val_accuracy: 0.1043\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 2s 106ms/step - loss: 2.3042 - accuracy: 0.1111 - val_loss: 2.3031 - val_accuracy: 0.1043\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 2s 102ms/step - loss: 2.3070 - accuracy: 0.0850 - val_loss: 2.2992 - val_accuracy: 0.1043\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 2s 99ms/step - loss: 2.3041 - accuracy: 0.1072 - val_loss: 2.2964 - val_accuracy: 0.1304\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 2s 96ms/step - loss: 2.3047 - accuracy: 0.0966 - val_loss: 2.3011 - val_accuracy: 0.1043\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 2s 99ms/step - loss: 2.3035 - accuracy: 0.1082 - val_loss: 2.3011 - val_accuracy: 0.1304\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 2s 97ms/step - loss: 2.3062 - accuracy: 0.0947 - val_loss: 2.3006 - val_accuracy: 0.1043\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 2s 94ms/step - loss: 2.3055 - accuracy: 0.0986 - val_loss: 2.3014 - val_accuracy: 0.1043\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 2s 96ms/step - loss: 2.3054 - accuracy: 0.0986 - val_loss: 2.2994 - val_accuracy: 0.1043\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 2s 98ms/step - loss: 7.5452 - accuracy: 0.0947 - val_loss: 2.3063 - val_accuracy: 0.0783\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 2s 93ms/step - loss: 2.3096 - accuracy: 0.0792 - val_loss: 2.3097 - val_accuracy: 0.0696\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 2s 97ms/step - loss: 2.3085 - accuracy: 0.0937 - val_loss: 2.3088 - val_accuracy: 0.0783\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 2s 93ms/step - loss: 2.3062 - accuracy: 0.0976 - val_loss: 2.3017 - val_accuracy: 0.1130\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 2s 93ms/step - loss: 2.3067 - accuracy: 0.1063 - val_loss: 2.3073 - val_accuracy: 0.0783\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 2s 93ms/step - loss: 2.3059 - accuracy: 0.0899 - val_loss: 2.3068 - val_accuracy: 0.0696\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 2s 93ms/step - loss: 2.3057 - accuracy: 0.0937 - val_loss: 2.3060 - val_accuracy: 0.1130\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 2s 94ms/step - loss: 2.3065 - accuracy: 0.0986 - val_loss: 2.3057 - val_accuracy: 0.0957\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 2s 93ms/step - loss: 2.3060 - accuracy: 0.0957 - val_loss: 2.3035 - val_accuracy: 0.1130\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 2s 93ms/step - loss: 2.3056 - accuracy: 0.0763 - val_loss: 2.3085 - val_accuracy: 0.0957\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 2s 95ms/step - loss: 2.3070 - accuracy: 0.0957 - val_loss: 2.3030 - val_accuracy: 0.1130\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 2s 94ms/step - loss: 2.3059 - accuracy: 0.0995 - val_loss: 2.3046 - val_accuracy: 0.0957\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 2s 94ms/step - loss: 2.3057 - accuracy: 0.0995 - val_loss: 2.3102 - val_accuracy: 0.0696\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 2s 97ms/step - loss: 2.3062 - accuracy: 0.0966 - val_loss: 2.3050 - val_accuracy: 0.1130\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 2s 94ms/step - loss: 2.3084 - accuracy: 0.0879 - val_loss: 2.3080 - val_accuracy: 0.0957\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 2s 100ms/step - loss: 5.7463 - accuracy: 0.0734 - val_loss: 19.9582 - val_accuracy: 0.1304\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 2s 94ms/step - loss: 3.1936 - accuracy: 0.0957 - val_loss: 2.3088 - val_accuracy: 0.0870\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 2s 97ms/step - loss: 2.3070 - accuracy: 0.0957 - val_loss: 2.3026 - val_accuracy: 0.0609\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 2s 95ms/step - loss: 2.3056 - accuracy: 0.0928 - val_loss: 2.3102 - val_accuracy: 0.1043\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 2s 94ms/step - loss: 2.3047 - accuracy: 0.1034 - val_loss: 2.3150 - val_accuracy: 0.0348\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 2s 94ms/step - loss: 2.3055 - accuracy: 0.0928 - val_loss: 2.3030 - val_accuracy: 0.1913\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 2s 95ms/step - loss: 2.3039 - accuracy: 0.1092 - val_loss: 2.3108 - val_accuracy: 0.1043\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 2s 93ms/step - loss: 2.3063 - accuracy: 0.0966 - val_loss: 2.3093 - val_accuracy: 0.0870\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 2s 94ms/step - loss: 2.3041 - accuracy: 0.1024 - val_loss: 2.3068 - val_accuracy: 0.1043\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 2s 93ms/step - loss: 2.3035 - accuracy: 0.1024 - val_loss: 2.3104 - val_accuracy: 0.0348\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 2s 95ms/step - loss: 2.3061 - accuracy: 0.1063 - val_loss: 2.3042 - val_accuracy: 0.1043\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 2s 93ms/step - loss: 2.3044 - accuracy: 0.0899 - val_loss: 2.3132 - val_accuracy: 0.0870\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 2s 95ms/step - loss: 2.3055 - accuracy: 0.0947 - val_loss: 2.3107 - val_accuracy: 0.0609\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 2s 93ms/step - loss: 2.3050 - accuracy: 0.1053 - val_loss: 2.3121 - val_accuracy: 0.1043\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 2s 94ms/step - loss: 2.3044 - accuracy: 0.1072 - val_loss: 2.3086 - val_accuracy: 0.1043\n",
            "Epoch 1/5\n",
            "52/52 [==============================] - 1s 26ms/step - loss: 2.2693 - accuracy: 0.1673 - val_loss: 2.1834 - val_accuracy: 0.3739\n",
            "Epoch 2/5\n",
            "52/52 [==============================] - 1s 23ms/step - loss: 1.8364 - accuracy: 0.4855 - val_loss: 1.3214 - val_accuracy: 0.6261\n",
            "Epoch 3/5\n",
            "52/52 [==============================] - 1s 23ms/step - loss: 0.9688 - accuracy: 0.7369 - val_loss: 0.6122 - val_accuracy: 0.8783\n",
            "Epoch 4/5\n",
            "52/52 [==============================] - 1s 23ms/step - loss: 0.5874 - accuracy: 0.8182 - val_loss: 0.3898 - val_accuracy: 0.9130\n",
            "Epoch 5/5\n",
            "52/52 [==============================] - 1s 23ms/step - loss: 0.4149 - accuracy: 0.8801 - val_loss: 0.2969 - val_accuracy: 0.9130\n",
            "Epoch 1/5\n",
            "52/52 [==============================] - 2s 29ms/step - loss: 2.2720 - accuracy: 0.2708 - val_loss: 2.1790 - val_accuracy: 0.4696\n",
            "Epoch 2/5\n",
            "52/52 [==============================] - 1s 27ms/step - loss: 1.7683 - accuracy: 0.6044 - val_loss: 1.1160 - val_accuracy: 0.7739\n",
            "Epoch 3/5\n",
            "52/52 [==============================] - 1s 26ms/step - loss: 0.8411 - accuracy: 0.7592 - val_loss: 0.5090 - val_accuracy: 0.8522\n",
            "Epoch 4/5\n",
            "52/52 [==============================] - 1s 27ms/step - loss: 0.4926 - accuracy: 0.8540 - val_loss: 0.3572 - val_accuracy: 0.9217\n",
            "Epoch 5/5\n",
            "52/52 [==============================] - 1s 26ms/step - loss: 0.3597 - accuracy: 0.8762 - val_loss: 0.2612 - val_accuracy: 0.9652\n",
            "Epoch 1/5\n",
            "52/52 [==============================] - 2s 29ms/step - loss: 2.2718 - accuracy: 0.2425 - val_loss: 2.1747 - val_accuracy: 0.4522\n",
            "Epoch 2/5\n",
            "52/52 [==============================] - 1s 27ms/step - loss: 1.7963 - accuracy: 0.5140 - val_loss: 1.1998 - val_accuracy: 0.7652\n",
            "Epoch 3/5\n",
            "52/52 [==============================] - 1s 27ms/step - loss: 0.8781 - accuracy: 0.7604 - val_loss: 0.5873 - val_accuracy: 0.8522\n",
            "Epoch 4/5\n",
            "52/52 [==============================] - 1s 26ms/step - loss: 0.5416 - accuracy: 0.8329 - val_loss: 0.3521 - val_accuracy: 0.9478\n",
            "Epoch 5/5\n",
            "52/52 [==============================] - 1s 26ms/step - loss: 0.3632 - accuracy: 0.9005 - val_loss: 0.2477 - val_accuracy: 0.9478\n",
            "Epoch 1/5\n",
            "52/52 [==============================] - 2s 29ms/step - loss: 2.2552 - accuracy: 0.2357 - val_loss: 2.1134 - val_accuracy: 0.4174\n",
            "Epoch 2/5\n",
            "52/52 [==============================] - 1s 24ms/step - loss: 1.6082 - accuracy: 0.5778 - val_loss: 0.9800 - val_accuracy: 0.7739\n",
            "Epoch 3/5\n",
            "52/52 [==============================] - 1s 27ms/step - loss: 0.7139 - accuracy: 0.7932 - val_loss: 0.4731 - val_accuracy: 0.8957\n",
            "Epoch 4/5\n",
            "52/52 [==============================] - 1s 26ms/step - loss: 0.4676 - accuracy: 0.8667 - val_loss: 0.3319 - val_accuracy: 0.9652\n",
            "Epoch 5/5\n",
            "52/52 [==============================] - 1s 25ms/step - loss: 0.3443 - accuracy: 0.9063 - val_loss: 0.2781 - val_accuracy: 0.9391\n",
            "Epoch 1/5\n",
            "52/52 [==============================] - 1s 28ms/step - loss: 2.2732 - accuracy: 0.1440 - val_loss: 2.1615 - val_accuracy: 0.4609\n",
            "Epoch 2/5\n",
            "52/52 [==============================] - 1s 25ms/step - loss: 1.8019 - accuracy: 0.5440 - val_loss: 1.1856 - val_accuracy: 0.7043\n",
            "Epoch 3/5\n",
            "52/52 [==============================] - 1s 25ms/step - loss: 0.8397 - accuracy: 0.7758 - val_loss: 0.6170 - val_accuracy: 0.8609\n",
            "Epoch 4/5\n",
            "52/52 [==============================] - 1s 26ms/step - loss: 0.5422 - accuracy: 0.8444 - val_loss: 0.4256 - val_accuracy: 0.9043\n",
            "Epoch 5/5\n",
            "52/52 [==============================] - 1s 25ms/step - loss: 0.3910 - accuracy: 0.8812 - val_loss: 0.2957 - val_accuracy: 0.9304\n",
            "Epoch 1/15\n",
            "7/7 [==============================] - 1s 117ms/step - loss: 6.7088 - accuracy: 0.1170 - val_loss: 2.9987 - val_accuracy: 0.0957\n",
            "Epoch 2/15\n",
            "7/7 [==============================] - 1s 104ms/step - loss: 2.7845 - accuracy: 0.0996 - val_loss: 2.2972 - val_accuracy: 0.1043\n",
            "Epoch 3/15\n",
            "7/7 [==============================] - 1s 94ms/step - loss: 2.3043 - accuracy: 0.1064 - val_loss: 2.3022 - val_accuracy: 0.1043\n",
            "Epoch 4/15\n",
            "7/7 [==============================] - 1s 96ms/step - loss: 2.3041 - accuracy: 0.1064 - val_loss: 2.3066 - val_accuracy: 0.1043\n",
            "Epoch 5/15\n",
            "7/7 [==============================] - 1s 99ms/step - loss: 2.3025 - accuracy: 0.1103 - val_loss: 2.3094 - val_accuracy: 0.1043\n",
            "Epoch 6/15\n",
            "7/7 [==============================] - 1s 95ms/step - loss: 2.3013 - accuracy: 0.1141 - val_loss: 2.3101 - val_accuracy: 0.1043\n",
            "Epoch 7/15\n",
            "7/7 [==============================] - 1s 101ms/step - loss: 2.3013 - accuracy: 0.1141 - val_loss: 2.3082 - val_accuracy: 0.1043\n",
            "Epoch 8/15\n",
            "7/7 [==============================] - 1s 95ms/step - loss: 2.3014 - accuracy: 0.1141 - val_loss: 2.3093 - val_accuracy: 0.1043\n",
            "Epoch 9/15\n",
            "7/7 [==============================] - 1s 103ms/step - loss: 2.3015 - accuracy: 0.1141 - val_loss: 2.3085 - val_accuracy: 0.1043\n",
            "Epoch 10/15\n",
            "7/7 [==============================] - 1s 96ms/step - loss: 2.3008 - accuracy: 0.1141 - val_loss: 2.3070 - val_accuracy: 0.1043\n",
            "Epoch 11/15\n",
            "7/7 [==============================] - 1s 105ms/step - loss: 2.3008 - accuracy: 0.1141 - val_loss: 2.3059 - val_accuracy: 0.1043\n",
            "Epoch 12/15\n",
            "7/7 [==============================] - 1s 94ms/step - loss: 2.3010 - accuracy: 0.1141 - val_loss: 2.3057 - val_accuracy: 0.1043\n",
            "Epoch 13/15\n",
            "7/7 [==============================] - 1s 99ms/step - loss: 2.3007 - accuracy: 0.1141 - val_loss: 2.3062 - val_accuracy: 0.1043\n",
            "Epoch 14/15\n",
            "7/7 [==============================] - 1s 93ms/step - loss: 2.3007 - accuracy: 0.1141 - val_loss: 2.3072 - val_accuracy: 0.1043\n",
            "Epoch 15/15\n",
            "7/7 [==============================] - 1s 103ms/step - loss: 2.3013 - accuracy: 0.1141 - val_loss: 2.3066 - val_accuracy: 0.1043\n",
            "Epoch 1/15\n",
            "7/7 [==============================] - 1s 114ms/step - loss: 4.6942 - accuracy: 0.1170 - val_loss: 2.4576 - val_accuracy: 0.0696\n",
            "Epoch 2/15\n",
            "7/7 [==============================] - 1s 101ms/step - loss: 2.4914 - accuracy: 0.0957 - val_loss: 2.3014 - val_accuracy: 0.1130\n",
            "Epoch 3/15\n",
            "7/7 [==============================] - 1s 93ms/step - loss: 2.3038 - accuracy: 0.1006 - val_loss: 2.3024 - val_accuracy: 0.0783\n",
            "Epoch 4/15\n",
            "7/7 [==============================] - 1s 94ms/step - loss: 2.3033 - accuracy: 0.1044 - val_loss: 2.3051 - val_accuracy: 0.0696\n",
            "Epoch 5/15\n",
            "7/7 [==============================] - 1s 100ms/step - loss: 2.3028 - accuracy: 0.1064 - val_loss: 2.3061 - val_accuracy: 0.0696\n",
            "Epoch 6/15\n",
            "7/7 [==============================] - 1s 93ms/step - loss: 2.3024 - accuracy: 0.0899 - val_loss: 2.3056 - val_accuracy: 0.1043\n",
            "Epoch 7/15\n",
            "7/7 [==============================] - 1s 104ms/step - loss: 2.3023 - accuracy: 0.0938 - val_loss: 2.3058 - val_accuracy: 0.1043\n",
            "Epoch 8/15\n",
            "7/7 [==============================] - 1s 94ms/step - loss: 2.3024 - accuracy: 0.1074 - val_loss: 2.3055 - val_accuracy: 0.1043\n",
            "Epoch 9/15\n",
            "7/7 [==============================] - 1s 102ms/step - loss: 2.3021 - accuracy: 0.1074 - val_loss: 2.3061 - val_accuracy: 0.1043\n",
            "Epoch 10/15\n",
            "7/7 [==============================] - 1s 94ms/step - loss: 2.3026 - accuracy: 0.1074 - val_loss: 2.3066 - val_accuracy: 0.1043\n",
            "Epoch 11/15\n",
            "7/7 [==============================] - 1s 101ms/step - loss: 2.3028 - accuracy: 0.0948 - val_loss: 2.3080 - val_accuracy: 0.0696\n",
            "Epoch 12/15\n",
            "7/7 [==============================] - 1s 91ms/step - loss: 2.3021 - accuracy: 0.1054 - val_loss: 2.3061 - val_accuracy: 0.1043\n",
            "Epoch 13/15\n",
            "7/7 [==============================] - 1s 99ms/step - loss: 2.3025 - accuracy: 0.1074 - val_loss: 2.3049 - val_accuracy: 0.1043\n",
            "Epoch 14/15\n",
            "7/7 [==============================] - 1s 93ms/step - loss: 2.3025 - accuracy: 0.1074 - val_loss: 2.3029 - val_accuracy: 0.1043\n",
            "Epoch 15/15\n",
            "7/7 [==============================] - 1s 100ms/step - loss: 2.3023 - accuracy: 0.1074 - val_loss: 2.3045 - val_accuracy: 0.1043\n",
            "Epoch 1/15\n",
            "7/7 [==============================] - 1s 109ms/step - loss: 3.3500 - accuracy: 0.1072 - val_loss: 2.8138 - val_accuracy: 0.0435\n",
            "Epoch 2/15\n",
            "7/7 [==============================] - 1s 99ms/step - loss: 2.4150 - accuracy: 0.0889 - val_loss: 2.3019 - val_accuracy: 0.1130\n",
            "Epoch 3/15\n",
            "7/7 [==============================] - 1s 97ms/step - loss: 2.3033 - accuracy: 0.0976 - val_loss: 2.2968 - val_accuracy: 0.0957\n",
            "Epoch 4/15\n",
            "7/7 [==============================] - 1s 93ms/step - loss: 2.3027 - accuracy: 0.1005 - val_loss: 2.2979 - val_accuracy: 0.1043\n",
            "Epoch 5/15\n",
            "7/7 [==============================] - 1s 99ms/step - loss: 2.3020 - accuracy: 0.1111 - val_loss: 2.3000 - val_accuracy: 0.1043\n",
            "Epoch 6/15\n",
            "7/7 [==============================] - 1s 93ms/step - loss: 2.3017 - accuracy: 0.1111 - val_loss: 2.3001 - val_accuracy: 0.1043\n",
            "Epoch 7/15\n",
            "7/7 [==============================] - 1s 104ms/step - loss: 2.3015 - accuracy: 0.1111 - val_loss: 2.2996 - val_accuracy: 0.1043\n",
            "Epoch 8/15\n",
            "7/7 [==============================] - 1s 92ms/step - loss: 2.3016 - accuracy: 0.1111 - val_loss: 2.2993 - val_accuracy: 0.1043\n",
            "Epoch 9/15\n",
            "7/7 [==============================] - 1s 103ms/step - loss: 2.3015 - accuracy: 0.1111 - val_loss: 2.2999 - val_accuracy: 0.1043\n",
            "Epoch 10/15\n",
            "7/7 [==============================] - 1s 95ms/step - loss: 2.3017 - accuracy: 0.1111 - val_loss: 2.3000 - val_accuracy: 0.1043\n",
            "Epoch 11/15\n",
            "7/7 [==============================] - 1s 99ms/step - loss: 2.3013 - accuracy: 0.1111 - val_loss: 2.3002 - val_accuracy: 0.1043\n",
            "Epoch 12/15\n",
            "7/7 [==============================] - 1s 93ms/step - loss: 2.3029 - accuracy: 0.1111 - val_loss: 2.3010 - val_accuracy: 0.1043\n",
            "Epoch 13/15\n",
            "7/7 [==============================] - 1s 103ms/step - loss: 2.3016 - accuracy: 0.1111 - val_loss: 2.3009 - val_accuracy: 0.1043\n",
            "Epoch 14/15\n",
            "7/7 [==============================] - 1s 95ms/step - loss: 2.3020 - accuracy: 0.1111 - val_loss: 2.3014 - val_accuracy: 0.1043\n",
            "Epoch 15/15\n",
            "7/7 [==============================] - 1s 104ms/step - loss: 2.3018 - accuracy: 0.1111 - val_loss: 2.2993 - val_accuracy: 0.1043\n",
            "Epoch 1/15\n",
            "7/7 [==============================] - 1s 111ms/step - loss: 3.5151 - accuracy: 0.1111 - val_loss: 2.3476 - val_accuracy: 0.0783\n",
            "Epoch 2/15\n",
            "7/7 [==============================] - 1s 103ms/step - loss: 2.3239 - accuracy: 0.0937 - val_loss: 2.3063 - val_accuracy: 0.1043\n",
            "Epoch 3/15\n",
            "7/7 [==============================] - 1s 95ms/step - loss: 2.3077 - accuracy: 0.1014 - val_loss: 2.3012 - val_accuracy: 0.1043\n",
            "Epoch 4/15\n",
            "7/7 [==============================] - 1s 104ms/step - loss: 2.3048 - accuracy: 0.1005 - val_loss: 2.3017 - val_accuracy: 0.0957\n",
            "Epoch 5/15\n",
            "7/7 [==============================] - 1s 92ms/step - loss: 2.3033 - accuracy: 0.1043 - val_loss: 2.3032 - val_accuracy: 0.1130\n",
            "Epoch 6/15\n",
            "7/7 [==============================] - 1s 96ms/step - loss: 2.3036 - accuracy: 0.1063 - val_loss: 2.3043 - val_accuracy: 0.1130\n",
            "Epoch 7/15\n",
            "7/7 [==============================] - 1s 97ms/step - loss: 2.3029 - accuracy: 0.1063 - val_loss: 2.3047 - val_accuracy: 0.1130\n",
            "Epoch 8/15\n",
            "7/7 [==============================] - 1s 95ms/step - loss: 2.3029 - accuracy: 0.1063 - val_loss: 2.3061 - val_accuracy: 0.1130\n",
            "Epoch 9/15\n",
            "7/7 [==============================] - 1s 103ms/step - loss: 2.3028 - accuracy: 0.1005 - val_loss: 2.3064 - val_accuracy: 0.0957\n",
            "Epoch 10/15\n",
            "7/7 [==============================] - 1s 94ms/step - loss: 2.3026 - accuracy: 0.1014 - val_loss: 2.3060 - val_accuracy: 0.1130\n",
            "Epoch 11/15\n",
            "7/7 [==============================] - 1s 104ms/step - loss: 2.3033 - accuracy: 0.1063 - val_loss: 2.3038 - val_accuracy: 0.1130\n",
            "Epoch 12/15\n",
            "7/7 [==============================] - 1s 94ms/step - loss: 2.3032 - accuracy: 0.1063 - val_loss: 2.3057 - val_accuracy: 0.1130\n",
            "Epoch 13/15\n",
            "7/7 [==============================] - 1s 105ms/step - loss: 2.3028 - accuracy: 0.1063 - val_loss: 2.3040 - val_accuracy: 0.1130\n",
            "Epoch 14/15\n",
            "7/7 [==============================] - 1s 93ms/step - loss: 2.3034 - accuracy: 0.1063 - val_loss: 2.3034 - val_accuracy: 0.1130\n",
            "Epoch 15/15\n",
            "7/7 [==============================] - 1s 104ms/step - loss: 2.3025 - accuracy: 0.1063 - val_loss: 2.3057 - val_accuracy: 0.1130\n",
            "WARNING:tensorflow:5 out of the last 22 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd04ac6048> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/15\n",
            "7/7 [==============================] - 1s 113ms/step - loss: 5.0286 - accuracy: 0.1121 - val_loss: 2.3013 - val_accuracy: 0.0609\n",
            "Epoch 2/15\n",
            "7/7 [==============================] - 1s 101ms/step - loss: 2.3094 - accuracy: 0.0966 - val_loss: 2.3161 - val_accuracy: 0.1043\n",
            "Epoch 3/15\n",
            "7/7 [==============================] - 1s 91ms/step - loss: 2.3096 - accuracy: 0.0986 - val_loss: 2.3216 - val_accuracy: 0.0609\n",
            "Epoch 4/15\n",
            "7/7 [==============================] - 1s 98ms/step - loss: 2.3064 - accuracy: 0.1043 - val_loss: 2.3178 - val_accuracy: 0.0609\n",
            "Epoch 5/15\n",
            "7/7 [==============================] - 1s 96ms/step - loss: 2.3038 - accuracy: 0.1063 - val_loss: 2.3154 - val_accuracy: 0.1043\n",
            "Epoch 6/15\n",
            "7/7 [==============================] - 1s 94ms/step - loss: 2.3026 - accuracy: 0.1072 - val_loss: 2.3137 - val_accuracy: 0.1043\n",
            "Epoch 7/15\n",
            "7/7 [==============================] - 1s 103ms/step - loss: 2.3020 - accuracy: 0.1101 - val_loss: 2.3127 - val_accuracy: 0.0870\n",
            "Epoch 8/15\n",
            "7/7 [==============================] - 1s 93ms/step - loss: 2.3024 - accuracy: 0.0986 - val_loss: 2.3111 - val_accuracy: 0.0870\n",
            "Epoch 9/15\n",
            "7/7 [==============================] - 1s 102ms/step - loss: 2.3021 - accuracy: 0.1063 - val_loss: 2.3097 - val_accuracy: 0.0870\n",
            "Epoch 10/15\n",
            "7/7 [==============================] - 1s 93ms/step - loss: 2.3021 - accuracy: 0.1063 - val_loss: 2.3105 - val_accuracy: 0.0870\n",
            "Epoch 11/15\n",
            "7/7 [==============================] - 1s 109ms/step - loss: 2.3021 - accuracy: 0.1005 - val_loss: 2.3093 - val_accuracy: 0.1043\n",
            "Epoch 12/15\n",
            "7/7 [==============================] - 1s 92ms/step - loss: 2.3016 - accuracy: 0.1072 - val_loss: 2.3072 - val_accuracy: 0.1043\n",
            "Epoch 13/15\n",
            "7/7 [==============================] - 1s 104ms/step - loss: 2.3020 - accuracy: 0.0995 - val_loss: 2.3078 - val_accuracy: 0.1043\n",
            "Epoch 14/15\n",
            "7/7 [==============================] - 1s 93ms/step - loss: 2.3019 - accuracy: 0.1072 - val_loss: 2.3078 - val_accuracy: 0.1043\n",
            "Epoch 15/15\n",
            "7/7 [==============================] - 1s 105ms/step - loss: 2.3022 - accuracy: 0.1072 - val_loss: 2.3069 - val_accuracy: 0.1043\n",
            "WARNING:tensorflow:6 out of the last 24 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd03646ae8> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/5\n",
            "21/21 [==============================] - 0s 23ms/step - loss: 2.2878 - accuracy: 0.1180 - val_loss: 2.0939 - val_accuracy: 0.2957\n",
            "Epoch 2/5\n",
            "21/21 [==============================] - 0s 16ms/step - loss: 1.6284 - accuracy: 0.4294 - val_loss: 1.0518 - val_accuracy: 0.5913\n",
            "Epoch 3/5\n",
            "21/21 [==============================] - 0s 18ms/step - loss: 1.1081 - accuracy: 0.5957 - val_loss: 0.6543 - val_accuracy: 0.7652\n",
            "Epoch 4/5\n",
            "21/21 [==============================] - 0s 18ms/step - loss: 0.9489 - accuracy: 0.6663 - val_loss: 0.4481 - val_accuracy: 0.8783\n",
            "Epoch 5/5\n",
            "21/21 [==============================] - 0s 18ms/step - loss: 0.8654 - accuracy: 0.7041 - val_loss: 0.4300 - val_accuracy: 0.8609\n",
            "WARNING:tensorflow:6 out of the last 11 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd081d2730> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/5\n",
            "21/21 [==============================] - 0s 23ms/step - loss: 2.3165 - accuracy: 0.0880 - val_loss: 2.3048 - val_accuracy: 0.0696\n",
            "Epoch 2/5\n",
            "21/21 [==============================] - 0s 20ms/step - loss: 2.3039 - accuracy: 0.0996 - val_loss: 2.3053 - val_accuracy: 0.1043\n",
            "Epoch 3/5\n",
            "21/21 [==============================] - 0s 18ms/step - loss: 2.3028 - accuracy: 0.1074 - val_loss: 2.3057 - val_accuracy: 0.1043\n",
            "Epoch 4/5\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 2.3029 - accuracy: 0.0928 - val_loss: 2.3048 - val_accuracy: 0.0696\n",
            "Epoch 5/5\n",
            "21/21 [==============================] - 0s 19ms/step - loss: 2.3029 - accuracy: 0.1044 - val_loss: 2.3055 - val_accuracy: 0.1043\n",
            "Epoch 1/5\n",
            "21/21 [==============================] - 1s 28ms/step - loss: 2.3196 - accuracy: 0.1014 - val_loss: 2.3127 - val_accuracy: 0.0957\n",
            "Epoch 2/5\n",
            "21/21 [==============================] - 0s 17ms/step - loss: 2.3109 - accuracy: 0.1227 - val_loss: 2.2546 - val_accuracy: 0.1565\n",
            "Epoch 3/5\n",
            "21/21 [==============================] - 0s 17ms/step - loss: 2.1272 - accuracy: 0.1884 - val_loss: 1.8828 - val_accuracy: 0.3913\n",
            "Epoch 4/5\n",
            "21/21 [==============================] - 0s 18ms/step - loss: 2.0522 - accuracy: 0.2164 - val_loss: 1.8883 - val_accuracy: 0.2522\n",
            "Epoch 5/5\n",
            "21/21 [==============================] - 0s 17ms/step - loss: 1.9301 - accuracy: 0.2522 - val_loss: 1.6438 - val_accuracy: 0.4435\n",
            "Epoch 1/5\n",
            "21/21 [==============================] - 1s 24ms/step - loss: 2.2945 - accuracy: 0.1208 - val_loss: 2.1681 - val_accuracy: 0.2522\n",
            "Epoch 2/5\n",
            "21/21 [==============================] - 0s 17ms/step - loss: 1.8156 - accuracy: 0.3246 - val_loss: 1.1493 - val_accuracy: 0.7304\n",
            "Epoch 3/5\n",
            "21/21 [==============================] - 0s 18ms/step - loss: 1.2799 - accuracy: 0.5430 - val_loss: 0.8238 - val_accuracy: 0.7739\n",
            "Epoch 4/5\n",
            "21/21 [==============================] - 0s 21ms/step - loss: 0.9883 - accuracy: 0.6531 - val_loss: 0.5750 - val_accuracy: 0.8435\n",
            "Epoch 5/5\n",
            "21/21 [==============================] - 0s 18ms/step - loss: 0.9555 - accuracy: 0.6580 - val_loss: 0.5544 - val_accuracy: 0.8261\n",
            "Epoch 1/5\n",
            "21/21 [==============================] - 1s 27ms/step - loss: 2.3166 - accuracy: 0.1159 - val_loss: 2.3140 - val_accuracy: 0.0957\n",
            "Epoch 2/5\n",
            "21/21 [==============================] - 0s 18ms/step - loss: 2.3049 - accuracy: 0.1188 - val_loss: 2.2603 - val_accuracy: 0.0957\n",
            "Epoch 3/5\n",
            "21/21 [==============================] - 0s 20ms/step - loss: 2.2210 - accuracy: 0.1430 - val_loss: 2.1623 - val_accuracy: 0.1478\n",
            "Epoch 4/5\n",
            "21/21 [==============================] - 0s 17ms/step - loss: 2.1819 - accuracy: 0.1729 - val_loss: 2.0750 - val_accuracy: 0.3043\n",
            "Epoch 5/5\n",
            "21/21 [==============================] - 0s 18ms/step - loss: 2.0153 - accuracy: 0.2386 - val_loss: 1.7788 - val_accuracy: 0.5739\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 8s 158ms/step - loss: 2.2351 - accuracy: 0.2166 - val_loss: 1.9794 - val_accuracy: 0.4348\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 8s 157ms/step - loss: 1.2420 - accuracy: 0.6625 - val_loss: 0.5660 - val_accuracy: 0.8435\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 8s 155ms/step - loss: 0.4750 - accuracy: 0.8578 - val_loss: 0.3478 - val_accuracy: 0.8957\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 8s 154ms/step - loss: 0.2932 - accuracy: 0.9023 - val_loss: 0.1898 - val_accuracy: 0.9478\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 8s 154ms/step - loss: 0.1963 - accuracy: 0.9458 - val_loss: 0.2303 - val_accuracy: 0.9478\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 8s 153ms/step - loss: 0.1590 - accuracy: 0.9507 - val_loss: 0.1987 - val_accuracy: 0.9391\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 8s 154ms/step - loss: 0.1352 - accuracy: 0.9565 - val_loss: 0.1271 - val_accuracy: 0.9565\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 8s 154ms/step - loss: 0.0978 - accuracy: 0.9671 - val_loss: 0.1283 - val_accuracy: 0.9565\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 8s 157ms/step - loss: 0.0849 - accuracy: 0.9739 - val_loss: 0.1233 - val_accuracy: 0.9478\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 8s 155ms/step - loss: 0.0728 - accuracy: 0.9768 - val_loss: 0.1324 - val_accuracy: 0.9739\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 8s 158ms/step - loss: 2.2135 - accuracy: 0.2640 - val_loss: 1.8394 - val_accuracy: 0.6174\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 8s 155ms/step - loss: 1.0711 - accuracy: 0.7389 - val_loss: 0.5092 - val_accuracy: 0.8522\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 8s 154ms/step - loss: 0.4146 - accuracy: 0.8791 - val_loss: 0.2847 - val_accuracy: 0.9478\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 9s 170ms/step - loss: 0.2896 - accuracy: 0.9062 - val_loss: 0.2165 - val_accuracy: 0.9565\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 8s 154ms/step - loss: 0.2214 - accuracy: 0.9284 - val_loss: 0.1712 - val_accuracy: 0.9565\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 8s 155ms/step - loss: 0.1783 - accuracy: 0.9497 - val_loss: 0.1644 - val_accuracy: 0.9478\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 8s 153ms/step - loss: 0.1371 - accuracy: 0.9652 - val_loss: 0.1499 - val_accuracy: 0.9826\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 8s 156ms/step - loss: 0.1381 - accuracy: 0.9594 - val_loss: 0.1329 - val_accuracy: 0.9652\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 8s 157ms/step - loss: 0.1076 - accuracy: 0.9632 - val_loss: 0.1134 - val_accuracy: 0.9478\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 8s 154ms/step - loss: 0.0927 - accuracy: 0.9691 - val_loss: 0.1055 - val_accuracy: 0.9826\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 8s 159ms/step - loss: 2.2121 - accuracy: 0.2213 - val_loss: 1.8476 - val_accuracy: 0.5304\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 8s 155ms/step - loss: 1.1694 - accuracy: 0.6889 - val_loss: 0.4806 - val_accuracy: 0.8696\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 8s 155ms/step - loss: 0.4594 - accuracy: 0.8599 - val_loss: 0.2702 - val_accuracy: 0.9130\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 8s 156ms/step - loss: 0.2897 - accuracy: 0.9043 - val_loss: 0.1688 - val_accuracy: 0.9478\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 8s 155ms/step - loss: 0.2121 - accuracy: 0.9372 - val_loss: 0.1387 - val_accuracy: 0.9478\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 8s 157ms/step - loss: 0.1569 - accuracy: 0.9507 - val_loss: 0.1163 - val_accuracy: 0.9739\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 8s 154ms/step - loss: 0.1357 - accuracy: 0.9565 - val_loss: 0.1209 - val_accuracy: 0.9652\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 8s 156ms/step - loss: 0.1025 - accuracy: 0.9691 - val_loss: 0.0940 - val_accuracy: 0.9826\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 8s 155ms/step - loss: 0.0862 - accuracy: 0.9778 - val_loss: 0.0687 - val_accuracy: 0.9826\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 8s 157ms/step - loss: 0.0676 - accuracy: 0.9807 - val_loss: 0.0779 - val_accuracy: 0.9826\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 8s 156ms/step - loss: 2.1816 - accuracy: 0.2406 - val_loss: 1.6786 - val_accuracy: 0.6087\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 8s 155ms/step - loss: 0.9683 - accuracy: 0.7527 - val_loss: 0.4973 - val_accuracy: 0.8783\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 8s 156ms/step - loss: 0.3807 - accuracy: 0.8792 - val_loss: 0.3214 - val_accuracy: 0.9304\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 8s 156ms/step - loss: 0.2701 - accuracy: 0.9179 - val_loss: 0.2181 - val_accuracy: 0.9565\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 8s 155ms/step - loss: 0.2015 - accuracy: 0.9498 - val_loss: 0.1911 - val_accuracy: 0.9565\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 8s 156ms/step - loss: 0.1604 - accuracy: 0.9449 - val_loss: 0.1721 - val_accuracy: 0.9652\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 8s 156ms/step - loss: 0.1391 - accuracy: 0.9556 - val_loss: 0.1851 - val_accuracy: 0.9478\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 8s 157ms/step - loss: 0.1165 - accuracy: 0.9604 - val_loss: 0.1378 - val_accuracy: 0.9826\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 8s 154ms/step - loss: 0.0930 - accuracy: 0.9700 - val_loss: 0.1374 - val_accuracy: 0.9652\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 8s 155ms/step - loss: 0.0835 - accuracy: 0.9749 - val_loss: 0.1339 - val_accuracy: 0.9739\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 8s 159ms/step - loss: 2.1977 - accuracy: 0.2908 - val_loss: 1.8468 - val_accuracy: 0.5565\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 8s 155ms/step - loss: 1.1146 - accuracy: 0.6812 - val_loss: 0.5520 - val_accuracy: 0.8870\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 8s 155ms/step - loss: 0.4522 - accuracy: 0.8638 - val_loss: 0.2242 - val_accuracy: 0.9652\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 8s 154ms/step - loss: 0.3267 - accuracy: 0.8908 - val_loss: 0.1534 - val_accuracy: 0.9652\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 8s 156ms/step - loss: 0.2248 - accuracy: 0.9343 - val_loss: 0.0974 - val_accuracy: 0.9913\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 8s 156ms/step - loss: 0.1707 - accuracy: 0.9430 - val_loss: 0.0562 - val_accuracy: 0.9913\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 8s 154ms/step - loss: 0.1387 - accuracy: 0.9556 - val_loss: 0.0389 - val_accuracy: 1.0000\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 8s 156ms/step - loss: 0.1277 - accuracy: 0.9594 - val_loss: 0.0530 - val_accuracy: 0.9913\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 8s 157ms/step - loss: 0.0967 - accuracy: 0.9710 - val_loss: 0.0437 - val_accuracy: 0.9913\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 8s 154ms/step - loss: 0.0779 - accuracy: 0.9652 - val_loss: 0.0370 - val_accuracy: 0.9913\n",
            "Epoch 1/5\n",
            "7/7 [==============================] - 2s 342ms/step - loss: 49.6951 - accuracy: 0.1093 - val_loss: 29.7665 - val_accuracy: 0.0696\n",
            "Epoch 2/5\n",
            "7/7 [==============================] - 2s 332ms/step - loss: 13.6137 - accuracy: 0.1122 - val_loss: 2.3038 - val_accuracy: 0.0957\n",
            "Epoch 3/5\n",
            "7/7 [==============================] - 2s 336ms/step - loss: 2.3082 - accuracy: 0.0938 - val_loss: 2.3008 - val_accuracy: 0.0957\n",
            "Epoch 4/5\n",
            "7/7 [==============================] - 2s 320ms/step - loss: 2.3060 - accuracy: 0.0928 - val_loss: 2.3005 - val_accuracy: 0.1130\n",
            "Epoch 5/5\n",
            "7/7 [==============================] - 2s 325ms/step - loss: 2.3028 - accuracy: 0.1064 - val_loss: 2.3016 - val_accuracy: 0.1130\n",
            "Epoch 1/5\n",
            "7/7 [==============================] - 2s 341ms/step - loss: 31.6649 - accuracy: 0.1006 - val_loss: 31.8897 - val_accuracy: 0.1217\n",
            "Epoch 2/5\n",
            "7/7 [==============================] - 2s 322ms/step - loss: 14.7158 - accuracy: 0.0938 - val_loss: 2.5532 - val_accuracy: 0.1130\n",
            "Epoch 3/5\n",
            "7/7 [==============================] - 2s 320ms/step - loss: 2.6101 - accuracy: 0.1093 - val_loss: 2.3376 - val_accuracy: 0.0696\n",
            "Epoch 4/5\n",
            "7/7 [==============================] - 2s 323ms/step - loss: 2.3242 - accuracy: 0.1054 - val_loss: 2.3113 - val_accuracy: 0.0696\n",
            "Epoch 5/5\n",
            "7/7 [==============================] - 2s 314ms/step - loss: 2.3061 - accuracy: 0.1064 - val_loss: 2.3075 - val_accuracy: 0.0696\n",
            "Epoch 1/5\n",
            "7/7 [==============================] - 2s 332ms/step - loss: 41.2231 - accuracy: 0.0860 - val_loss: 29.0817 - val_accuracy: 0.0348\n",
            "Epoch 2/5\n",
            "7/7 [==============================] - 2s 314ms/step - loss: 12.8098 - accuracy: 0.0870 - val_loss: 2.3062 - val_accuracy: 0.1043\n",
            "Epoch 3/5\n",
            "7/7 [==============================] - 2s 319ms/step - loss: 2.3784 - accuracy: 0.0899 - val_loss: 2.3035 - val_accuracy: 0.1043\n",
            "Epoch 4/5\n",
            "7/7 [==============================] - 2s 318ms/step - loss: 2.3097 - accuracy: 0.0957 - val_loss: 2.3013 - val_accuracy: 0.1043\n",
            "Epoch 5/5\n",
            "7/7 [==============================] - 2s 314ms/step - loss: 2.3057 - accuracy: 0.0918 - val_loss: 2.2997 - val_accuracy: 0.1043\n",
            "Epoch 1/5\n",
            "7/7 [==============================] - 2s 341ms/step - loss: 36.1698 - accuracy: 0.1130 - val_loss: 17.9537 - val_accuracy: 0.1130\n",
            "Epoch 2/5\n",
            "7/7 [==============================] - 2s 313ms/step - loss: 12.8490 - accuracy: 0.1082 - val_loss: 2.3744 - val_accuracy: 0.1391\n",
            "Epoch 3/5\n",
            "7/7 [==============================] - 2s 318ms/step - loss: 2.6049 - accuracy: 0.0763 - val_loss: 2.2968 - val_accuracy: 0.1217\n",
            "Epoch 4/5\n",
            "7/7 [==============================] - 2s 316ms/step - loss: 2.3141 - accuracy: 0.0986 - val_loss: 2.2975 - val_accuracy: 0.1217\n",
            "Epoch 5/5\n",
            "7/7 [==============================] - 2s 322ms/step - loss: 2.3066 - accuracy: 0.0976 - val_loss: 2.2995 - val_accuracy: 0.1217\n",
            "WARNING:tensorflow:5 out of the last 22 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd1962db70> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/5\n",
            "7/7 [==============================] - 2s 333ms/step - loss: 25.4080 - accuracy: 0.1121 - val_loss: 20.6726 - val_accuracy: 0.1913\n",
            "Epoch 2/5\n",
            "7/7 [==============================] - 2s 329ms/step - loss: 9.6751 - accuracy: 0.1024 - val_loss: 2.3117 - val_accuracy: 0.0783\n",
            "Epoch 3/5\n",
            "7/7 [==============================] - 2s 317ms/step - loss: 2.4062 - accuracy: 0.0889 - val_loss: 2.3113 - val_accuracy: 0.0783\n",
            "Epoch 4/5\n",
            "7/7 [==============================] - 2s 321ms/step - loss: 2.3132 - accuracy: 0.0937 - val_loss: 2.3105 - val_accuracy: 0.0870\n",
            "Epoch 5/5\n",
            "7/7 [==============================] - 2s 313ms/step - loss: 2.3069 - accuracy: 0.1053 - val_loss: 2.3097 - val_accuracy: 0.0870\n",
            "WARNING:tensorflow:6 out of the last 24 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbcfcb600d0> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/10\n",
            "21/21 [==============================] - 1s 34ms/step - loss: 2.3212 - accuracy: 0.0977 - val_loss: 2.3059 - val_accuracy: 0.0696\n",
            "Epoch 2/10\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 2.3048 - accuracy: 0.0977 - val_loss: 2.3041 - val_accuracy: 0.1043\n",
            "Epoch 3/10\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 2.3027 - accuracy: 0.1141 - val_loss: 2.3070 - val_accuracy: 0.1043\n",
            "Epoch 4/10\n",
            "21/21 [==============================] - 1s 26ms/step - loss: 2.3026 - accuracy: 0.1054 - val_loss: 2.3079 - val_accuracy: 0.1043\n",
            "Epoch 5/10\n",
            "21/21 [==============================] - 1s 27ms/step - loss: 2.3022 - accuracy: 0.1141 - val_loss: 2.3077 - val_accuracy: 0.1043\n",
            "Epoch 6/10\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 2.3022 - accuracy: 0.1141 - val_loss: 2.3064 - val_accuracy: 0.1043\n",
            "Epoch 7/10\n",
            "21/21 [==============================] - 1s 28ms/step - loss: 2.3025 - accuracy: 0.1141 - val_loss: 2.3054 - val_accuracy: 0.1043\n",
            "Epoch 8/10\n",
            "21/21 [==============================] - 1s 26ms/step - loss: 2.3020 - accuracy: 0.1141 - val_loss: 2.3078 - val_accuracy: 0.1043\n",
            "Epoch 9/10\n",
            "21/21 [==============================] - 1s 27ms/step - loss: 2.3010 - accuracy: 0.1141 - val_loss: 2.3068 - val_accuracy: 0.1043\n",
            "Epoch 10/10\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 2.3015 - accuracy: 0.1141 - val_loss: 2.3057 - val_accuracy: 0.1043\n",
            "WARNING:tensorflow:6 out of the last 11 calls to <function Model.make_predict_function.<locals>.predict_function at 0x7fbd035c3488> triggered tf.function retracing. Tracing is expensive and the excessive number of tracings could be due to (1) creating @tf.function repeatedly in a loop, (2) passing tensors with different shapes, (3) passing Python objects instead of tensors. For (1), please define your @tf.function outside of the loop. For (2), @tf.function has experimental_relax_shapes=True option that relaxes argument shapes that can avoid unnecessary retracing. For (3), please refer to https://www.tensorflow.org/tutorials/customization/performance#python_or_tensor_args and https://www.tensorflow.org/api_docs/python/tf/function for  more details.\n",
            "Epoch 1/10\n",
            "21/21 [==============================] - 1s 30ms/step - loss: 2.2616 - accuracy: 0.1257 - val_loss: 2.0096 - val_accuracy: 0.3391\n",
            "Epoch 2/10\n",
            "21/21 [==============================] - 1s 26ms/step - loss: 1.9665 - accuracy: 0.2708 - val_loss: 1.6229 - val_accuracy: 0.4435\n",
            "Epoch 3/10\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 1.8012 - accuracy: 0.3482 - val_loss: 1.1914 - val_accuracy: 0.6696\n",
            "Epoch 4/10\n",
            "21/21 [==============================] - 1s 28ms/step - loss: 1.7409 - accuracy: 0.3520 - val_loss: 1.2965 - val_accuracy: 0.7043\n",
            "Epoch 5/10\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 1.5825 - accuracy: 0.4178 - val_loss: 1.1196 - val_accuracy: 0.7391\n",
            "Epoch 6/10\n",
            "21/21 [==============================] - 1s 26ms/step - loss: 1.4747 - accuracy: 0.4468 - val_loss: 1.0859 - val_accuracy: 0.7304\n",
            "Epoch 7/10\n",
            "21/21 [==============================] - 1s 24ms/step - loss: 1.5370 - accuracy: 0.4304 - val_loss: 1.1351 - val_accuracy: 0.6261\n",
            "Epoch 8/10\n",
            "21/21 [==============================] - 1s 28ms/step - loss: 1.4364 - accuracy: 0.4652 - val_loss: 0.9485 - val_accuracy: 0.7304\n",
            "Epoch 9/10\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 1.3946 - accuracy: 0.4797 - val_loss: 0.9437 - val_accuracy: 0.7739\n",
            "Epoch 10/10\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 1.3798 - accuracy: 0.4971 - val_loss: 0.8902 - val_accuracy: 0.7217\n",
            "Epoch 1/10\n",
            "21/21 [==============================] - 1s 34ms/step - loss: 2.3207 - accuracy: 0.0995 - val_loss: 2.3065 - val_accuracy: 0.0783\n",
            "Epoch 2/10\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 2.2970 - accuracy: 0.1101 - val_loss: 2.2953 - val_accuracy: 0.1304\n",
            "Epoch 3/10\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 2.3054 - accuracy: 0.0841 - val_loss: 2.3003 - val_accuracy: 0.1043\n",
            "Epoch 4/10\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 2.3050 - accuracy: 0.1111 - val_loss: 2.3008 - val_accuracy: 0.1043\n",
            "Epoch 5/10\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 2.3032 - accuracy: 0.1053 - val_loss: 2.2982 - val_accuracy: 0.1130\n",
            "Epoch 6/10\n",
            "21/21 [==============================] - 1s 26ms/step - loss: 2.3020 - accuracy: 0.1053 - val_loss: 2.3001 - val_accuracy: 0.1043\n",
            "Epoch 7/10\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 2.3027 - accuracy: 0.1072 - val_loss: 2.3006 - val_accuracy: 0.1043\n",
            "Epoch 8/10\n",
            "21/21 [==============================] - 1s 24ms/step - loss: 2.3025 - accuracy: 0.1111 - val_loss: 2.2996 - val_accuracy: 0.1043\n",
            "Epoch 9/10\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 2.3025 - accuracy: 0.1111 - val_loss: 2.3011 - val_accuracy: 0.1043\n",
            "Epoch 10/10\n",
            "21/21 [==============================] - 1s 27ms/step - loss: 2.3029 - accuracy: 0.1111 - val_loss: 2.3015 - val_accuracy: 0.1043\n",
            "Epoch 1/10\n",
            "21/21 [==============================] - 1s 34ms/step - loss: 2.3024 - accuracy: 0.1014 - val_loss: 2.2810 - val_accuracy: 0.1478\n",
            "Epoch 2/10\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 2.1725 - accuracy: 0.1662 - val_loss: 2.1321 - val_accuracy: 0.2000\n",
            "Epoch 3/10\n",
            "21/21 [==============================] - 1s 28ms/step - loss: 2.1317 - accuracy: 0.1913 - val_loss: 1.9607 - val_accuracy: 0.2261\n",
            "Epoch 4/10\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 2.1107 - accuracy: 0.1575 - val_loss: 2.0186 - val_accuracy: 0.2348\n",
            "Epoch 5/10\n",
            "21/21 [==============================] - 1s 26ms/step - loss: 2.1248 - accuracy: 0.1401 - val_loss: 2.0192 - val_accuracy: 0.2261\n",
            "Epoch 6/10\n",
            "21/21 [==============================] - 1s 27ms/step - loss: 2.1255 - accuracy: 0.1585 - val_loss: 2.0145 - val_accuracy: 0.1739\n",
            "Epoch 7/10\n",
            "21/21 [==============================] - 1s 27ms/step - loss: 2.0973 - accuracy: 0.1507 - val_loss: 1.9219 - val_accuracy: 0.2435\n",
            "Epoch 8/10\n",
            "21/21 [==============================] - 1s 27ms/step - loss: 2.1107 - accuracy: 0.1758 - val_loss: 2.0184 - val_accuracy: 0.1913\n",
            "Epoch 9/10\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 2.1232 - accuracy: 0.1575 - val_loss: 2.1438 - val_accuracy: 0.0957\n",
            "Epoch 10/10\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 2.1264 - accuracy: 0.1353 - val_loss: 1.9593 - val_accuracy: 0.2261\n",
            "Epoch 1/10\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 2.3025 - accuracy: 0.1063 - val_loss: 2.1821 - val_accuracy: 0.1217\n",
            "Epoch 2/10\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 2.0404 - accuracy: 0.2319 - val_loss: 1.6573 - val_accuracy: 0.4261\n",
            "Epoch 3/10\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 1.6478 - accuracy: 0.3845 - val_loss: 1.1701 - val_accuracy: 0.5913\n",
            "Epoch 4/10\n",
            "21/21 [==============================] - 1s 28ms/step - loss: 1.3700 - accuracy: 0.5111 - val_loss: 1.1930 - val_accuracy: 0.6435\n",
            "Epoch 5/10\n",
            "21/21 [==============================] - 1s 27ms/step - loss: 1.3467 - accuracy: 0.5237 - val_loss: 0.8688 - val_accuracy: 0.7478\n",
            "Epoch 6/10\n",
            "21/21 [==============================] - 1s 27ms/step - loss: 1.2605 - accuracy: 0.5459 - val_loss: 0.7513 - val_accuracy: 0.8435\n",
            "Epoch 7/10\n",
            "21/21 [==============================] - 1s 26ms/step - loss: 1.1807 - accuracy: 0.5594 - val_loss: 0.7623 - val_accuracy: 0.8174\n",
            "Epoch 8/10\n",
            "21/21 [==============================] - 1s 29ms/step - loss: 1.1662 - accuracy: 0.5710 - val_loss: 0.6867 - val_accuracy: 0.8522\n",
            "Epoch 9/10\n",
            "21/21 [==============================] - 1s 25ms/step - loss: 0.9953 - accuracy: 0.6435 - val_loss: 0.6069 - val_accuracy: 0.8174\n",
            "Epoch 10/10\n",
            "21/21 [==============================] - 1s 26ms/step - loss: 1.0378 - accuracy: 0.6193 - val_loss: 0.6329 - val_accuracy: 0.8435\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 3s 141ms/step - loss: 2.3492 - accuracy: 0.1286 - val_loss: 2.2733 - val_accuracy: 0.1565\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 3s 135ms/step - loss: 2.1046 - accuracy: 0.2031 - val_loss: 1.6046 - val_accuracy: 0.4957\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 3s 135ms/step - loss: 1.7677 - accuracy: 0.3327 - val_loss: 1.3037 - val_accuracy: 0.4957\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 3s 136ms/step - loss: 1.7298 - accuracy: 0.3985 - val_loss: 2.0284 - val_accuracy: 0.2870\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 3s 139ms/step - loss: 2.3260 - accuracy: 0.1074 - val_loss: 2.3237 - val_accuracy: 0.0696\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 3s 134ms/step - loss: 2.3057 - accuracy: 0.1035 - val_loss: 2.3088 - val_accuracy: 0.1043\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 3s 135ms/step - loss: 2.3025 - accuracy: 0.1141 - val_loss: 2.3077 - val_accuracy: 0.1043\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 3s 135ms/step - loss: 2.3021 - accuracy: 0.1141 - val_loss: 2.3073 - val_accuracy: 0.1043\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 3s 139ms/step - loss: 2.3026 - accuracy: 0.1006 - val_loss: 2.3071 - val_accuracy: 0.1043\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 3s 135ms/step - loss: 2.3035 - accuracy: 0.1044 - val_loss: 2.3070 - val_accuracy: 0.1043\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 3s 134ms/step - loss: 2.3020 - accuracy: 0.1141 - val_loss: 2.3071 - val_accuracy: 0.1043\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 3s 134ms/step - loss: 2.3021 - accuracy: 0.1141 - val_loss: 2.3055 - val_accuracy: 0.1043\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 3s 135ms/step - loss: 2.3023 - accuracy: 0.1141 - val_loss: 2.3069 - val_accuracy: 0.1043\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 3s 136ms/step - loss: 2.3026 - accuracy: 0.1141 - val_loss: 2.3083 - val_accuracy: 0.1043\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 3s 136ms/step - loss: 2.3026 - accuracy: 0.1141 - val_loss: 2.3080 - val_accuracy: 0.1043\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 3s 143ms/step - loss: 2.3427 - accuracy: 0.1015 - val_loss: 2.3019 - val_accuracy: 0.0696\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 3s 135ms/step - loss: 2.3043 - accuracy: 0.1025 - val_loss: 2.3056 - val_accuracy: 0.1043\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 3s 136ms/step - loss: 2.3031 - accuracy: 0.1074 - val_loss: 2.3041 - val_accuracy: 0.1043\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 3s 135ms/step - loss: 2.3038 - accuracy: 0.1025 - val_loss: 2.3076 - val_accuracy: 0.1043\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 3s 137ms/step - loss: 2.3031 - accuracy: 0.1074 - val_loss: 2.3035 - val_accuracy: 0.1043\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 3s 135ms/step - loss: 2.3040 - accuracy: 0.0870 - val_loss: 2.3063 - val_accuracy: 0.0696\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 3s 134ms/step - loss: 2.3037 - accuracy: 0.1064 - val_loss: 2.3033 - val_accuracy: 0.1043\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 3s 135ms/step - loss: 2.3035 - accuracy: 0.0841 - val_loss: 2.3041 - val_accuracy: 0.1043\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 3s 137ms/step - loss: 2.3033 - accuracy: 0.1074 - val_loss: 2.3081 - val_accuracy: 0.1043\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 3s 134ms/step - loss: 2.3030 - accuracy: 0.1006 - val_loss: 2.3054 - val_accuracy: 0.0696\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 3s 136ms/step - loss: 2.3027 - accuracy: 0.0948 - val_loss: 2.3057 - val_accuracy: 0.1043\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 3s 139ms/step - loss: 2.3026 - accuracy: 0.1074 - val_loss: 2.3055 - val_accuracy: 0.1043\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 3s 139ms/step - loss: 2.3028 - accuracy: 0.1074 - val_loss: 2.3058 - val_accuracy: 0.1043\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 3s 137ms/step - loss: 2.3026 - accuracy: 0.1054 - val_loss: 2.3054 - val_accuracy: 0.0696\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 3s 139ms/step - loss: 2.3023 - accuracy: 0.1064 - val_loss: 2.3075 - val_accuracy: 0.0696\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 3s 143ms/step - loss: 2.3560 - accuracy: 0.1082 - val_loss: 2.3091 - val_accuracy: 0.0696\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 3s 136ms/step - loss: 2.3037 - accuracy: 0.1014 - val_loss: 2.3040 - val_accuracy: 0.0696\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 3s 136ms/step - loss: 2.3038 - accuracy: 0.1082 - val_loss: 2.2998 - val_accuracy: 0.1043\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 3s 139ms/step - loss: 2.3024 - accuracy: 0.0986 - val_loss: 2.2991 - val_accuracy: 0.1304\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 3s 137ms/step - loss: 2.3030 - accuracy: 0.1092 - val_loss: 2.3012 - val_accuracy: 0.1043\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 3s 134ms/step - loss: 2.3019 - accuracy: 0.1111 - val_loss: 2.2998 - val_accuracy: 0.1043\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 3s 134ms/step - loss: 2.3029 - accuracy: 0.1111 - val_loss: 2.2992 - val_accuracy: 0.1043\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 3s 137ms/step - loss: 2.3033 - accuracy: 0.1111 - val_loss: 2.3009 - val_accuracy: 0.1043\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 3s 134ms/step - loss: 2.3022 - accuracy: 0.1111 - val_loss: 2.2982 - val_accuracy: 0.1043\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 3s 135ms/step - loss: 2.3021 - accuracy: 0.1111 - val_loss: 2.2996 - val_accuracy: 0.1043\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 3s 136ms/step - loss: 2.3018 - accuracy: 0.1111 - val_loss: 2.2997 - val_accuracy: 0.1043\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 3s 138ms/step - loss: 2.3029 - accuracy: 0.1111 - val_loss: 2.3005 - val_accuracy: 0.1043\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 3s 135ms/step - loss: 2.3020 - accuracy: 0.1111 - val_loss: 2.3007 - val_accuracy: 0.1043\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 3s 136ms/step - loss: 2.3014 - accuracy: 0.1111 - val_loss: 2.2996 - val_accuracy: 0.1043\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 3s 135ms/step - loss: 2.3022 - accuracy: 0.1111 - val_loss: 2.2998 - val_accuracy: 0.1043\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 3s 141ms/step - loss: 2.4059 - accuracy: 0.1324 - val_loss: 2.2973 - val_accuracy: 0.1217\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 3s 135ms/step - loss: 2.1995 - accuracy: 0.1797 - val_loss: 1.9811 - val_accuracy: 0.3304\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 3s 136ms/step - loss: 2.1006 - accuracy: 0.1952 - val_loss: 1.9318 - val_accuracy: 0.2696\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 3s 140ms/step - loss: 2.0689 - accuracy: 0.1981 - val_loss: 1.8068 - val_accuracy: 0.2957\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 3s 134ms/step - loss: 1.9022 - accuracy: 0.2676 - val_loss: 1.6816 - val_accuracy: 0.3913\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 3s 137ms/step - loss: 1.8471 - accuracy: 0.2899 - val_loss: 1.7072 - val_accuracy: 0.3043\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 3s 136ms/step - loss: 1.8143 - accuracy: 0.3014 - val_loss: 1.8672 - val_accuracy: 0.2261\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 3s 137ms/step - loss: 1.7782 - accuracy: 0.3382 - val_loss: 1.6055 - val_accuracy: 0.3565\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 3s 135ms/step - loss: 1.7099 - accuracy: 0.3372 - val_loss: 1.5437 - val_accuracy: 0.4087\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 3s 135ms/step - loss: 1.7506 - accuracy: 0.3411 - val_loss: 1.6964 - val_accuracy: 0.2696\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 3s 136ms/step - loss: 1.7216 - accuracy: 0.3382 - val_loss: 1.6411 - val_accuracy: 0.3304\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 3s 136ms/step - loss: 1.6311 - accuracy: 0.4126 - val_loss: 1.2738 - val_accuracy: 0.5478\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 3s 136ms/step - loss: 1.6520 - accuracy: 0.4164 - val_loss: 1.4661 - val_accuracy: 0.5217\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 3s 135ms/step - loss: 1.5981 - accuracy: 0.4184 - val_loss: 1.3793 - val_accuracy: 0.5304\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 3s 137ms/step - loss: 1.6213 - accuracy: 0.4406 - val_loss: 1.0930 - val_accuracy: 0.7130\n",
            "Epoch 1/15\n",
            "21/21 [==============================] - 3s 140ms/step - loss: 2.3259 - accuracy: 0.1014 - val_loss: 2.3138 - val_accuracy: 0.0609\n",
            "Epoch 2/15\n",
            "21/21 [==============================] - 3s 136ms/step - loss: 2.3043 - accuracy: 0.1043 - val_loss: 2.3113 - val_accuracy: 0.0609\n",
            "Epoch 3/15\n",
            "21/21 [==============================] - 3s 138ms/step - loss: 2.3043 - accuracy: 0.1014 - val_loss: 2.3063 - val_accuracy: 0.1043\n",
            "Epoch 4/15\n",
            "21/21 [==============================] - 3s 138ms/step - loss: 2.3032 - accuracy: 0.1072 - val_loss: 2.3046 - val_accuracy: 0.1043\n",
            "Epoch 5/15\n",
            "21/21 [==============================] - 3s 136ms/step - loss: 2.3034 - accuracy: 0.1072 - val_loss: 2.3141 - val_accuracy: 0.1043\n",
            "Epoch 6/15\n",
            "21/21 [==============================] - 3s 135ms/step - loss: 2.3034 - accuracy: 0.1053 - val_loss: 2.3093 - val_accuracy: 0.0348\n",
            "Epoch 7/15\n",
            "21/21 [==============================] - 3s 140ms/step - loss: 2.3036 - accuracy: 0.0899 - val_loss: 2.3113 - val_accuracy: 0.0870\n",
            "Epoch 8/15\n",
            "21/21 [==============================] - 3s 143ms/step - loss: 2.3026 - accuracy: 0.1072 - val_loss: 2.3093 - val_accuracy: 0.1043\n",
            "Epoch 9/15\n",
            "21/21 [==============================] - 3s 145ms/step - loss: 2.3027 - accuracy: 0.1072 - val_loss: 2.3079 - val_accuracy: 0.1043\n",
            "Epoch 10/15\n",
            "21/21 [==============================] - 3s 142ms/step - loss: 2.3025 - accuracy: 0.0860 - val_loss: 2.3083 - val_accuracy: 0.1043\n",
            "Epoch 11/15\n",
            "21/21 [==============================] - 3s 138ms/step - loss: 2.3023 - accuracy: 0.1072 - val_loss: 2.3059 - val_accuracy: 0.1043\n",
            "Epoch 12/15\n",
            "21/21 [==============================] - 3s 137ms/step - loss: 2.3025 - accuracy: 0.1072 - val_loss: 2.3066 - val_accuracy: 0.1043\n",
            "Epoch 13/15\n",
            "21/21 [==============================] - 3s 139ms/step - loss: 2.3024 - accuracy: 0.1072 - val_loss: 2.3102 - val_accuracy: 0.1043\n",
            "Epoch 14/15\n",
            "21/21 [==============================] - 3s 146ms/step - loss: 2.3022 - accuracy: 0.1043 - val_loss: 2.3127 - val_accuracy: 0.0870\n",
            "Epoch 15/15\n",
            "21/21 [==============================] - 3s 144ms/step - loss: 2.3026 - accuracy: 0.1024 - val_loss: 2.3074 - val_accuracy: 0.1043\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 2s 36ms/step - loss: 1.6218 - accuracy: 0.4149 - val_loss: 0.8921 - val_accuracy: 0.7130\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 2s 31ms/step - loss: 0.5865 - accuracy: 0.8017 - val_loss: 0.4036 - val_accuracy: 0.8609\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 2s 31ms/step - loss: 0.4251 - accuracy: 0.8685 - val_loss: 0.2273 - val_accuracy: 0.9652\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 2s 33ms/step - loss: 0.3714 - accuracy: 0.8820 - val_loss: 0.2554 - val_accuracy: 0.9565\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 2s 32ms/step - loss: 0.3895 - accuracy: 0.8781 - val_loss: 0.2852 - val_accuracy: 0.9217\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 2s 31ms/step - loss: 0.3411 - accuracy: 0.8985 - val_loss: 0.3319 - val_accuracy: 0.9478\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 2s 31ms/step - loss: 0.3763 - accuracy: 0.8781 - val_loss: 0.1011 - val_accuracy: 0.9652\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 2s 32ms/step - loss: 0.4082 - accuracy: 0.8936 - val_loss: 0.2850 - val_accuracy: 0.9304\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 2s 30ms/step - loss: 0.3422 - accuracy: 0.8965 - val_loss: 0.1486 - val_accuracy: 0.9478\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 2s 31ms/step - loss: 0.3379 - accuracy: 0.8985 - val_loss: 0.1383 - val_accuracy: 0.9478\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 2s 34ms/step - loss: 1.8538 - accuracy: 0.3317 - val_loss: 0.7745 - val_accuracy: 0.7304\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 2s 32ms/step - loss: 0.7771 - accuracy: 0.7582 - val_loss: 0.3134 - val_accuracy: 0.9043\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 2s 32ms/step - loss: 0.5078 - accuracy: 0.8375 - val_loss: 0.3444 - val_accuracy: 0.8870\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 2s 30ms/step - loss: 0.4880 - accuracy: 0.8366 - val_loss: 0.2272 - val_accuracy: 0.9391\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 2s 32ms/step - loss: 0.4014 - accuracy: 0.8752 - val_loss: 0.2058 - val_accuracy: 0.9478\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 2s 31ms/step - loss: 0.4543 - accuracy: 0.8656 - val_loss: 0.1186 - val_accuracy: 0.9565\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 2s 31ms/step - loss: 0.3899 - accuracy: 0.8849 - val_loss: 0.2288 - val_accuracy: 0.9217\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 2s 31ms/step - loss: 0.2822 - accuracy: 0.9110 - val_loss: 0.1620 - val_accuracy: 0.9391\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 2s 30ms/step - loss: 0.2886 - accuracy: 0.9072 - val_loss: 0.1975 - val_accuracy: 0.9565\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 2s 31ms/step - loss: 0.3167 - accuracy: 0.9052 - val_loss: 0.2194 - val_accuracy: 0.9304\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 2s 35ms/step - loss: 1.7741 - accuracy: 0.3797 - val_loss: 0.7090 - val_accuracy: 0.8000\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 2s 32ms/step - loss: 0.7285 - accuracy: 0.7604 - val_loss: 0.3418 - val_accuracy: 0.8870\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 2s 30ms/step - loss: 0.4206 - accuracy: 0.8715 - val_loss: 0.3624 - val_accuracy: 0.8696\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 2s 31ms/step - loss: 0.3787 - accuracy: 0.8792 - val_loss: 0.1538 - val_accuracy: 0.9565\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 2s 31ms/step - loss: 0.3483 - accuracy: 0.8889 - val_loss: 0.3428 - val_accuracy: 0.9391\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 2s 32ms/step - loss: 0.2993 - accuracy: 0.9063 - val_loss: 0.2811 - val_accuracy: 0.9478\n",
            "Epoch 7/10\n",
            "52/52 [==============================] - 2s 30ms/step - loss: 0.3636 - accuracy: 0.8860 - val_loss: 0.3726 - val_accuracy: 0.8783\n",
            "Epoch 8/10\n",
            "52/52 [==============================] - 2s 30ms/step - loss: 0.3720 - accuracy: 0.9043 - val_loss: 0.5284 - val_accuracy: 0.9043\n",
            "Epoch 9/10\n",
            "52/52 [==============================] - 2s 31ms/step - loss: 0.2820 - accuracy: 0.9246 - val_loss: 0.1827 - val_accuracy: 0.9652\n",
            "Epoch 10/10\n",
            "52/52 [==============================] - 2s 30ms/step - loss: 0.3109 - accuracy: 0.9121 - val_loss: 0.1852 - val_accuracy: 0.9478\n",
            "Epoch 1/10\n",
            "52/52 [==============================] - 2s 33ms/step - loss: 2.1055 - accuracy: 0.2396 - val_loss: 1.0743 - val_accuracy: 0.6348\n",
            "Epoch 2/10\n",
            "52/52 [==============================] - 2s 30ms/step - loss: 0.9400 - accuracy: 0.6792 - val_loss: 0.4453 - val_accuracy: 0.8261\n",
            "Epoch 3/10\n",
            "52/52 [==============================] - 2s 32ms/step - loss: 0.6322 - accuracy: 0.7894 - val_loss: 0.4536 - val_accuracy: 0.8696\n",
            "Epoch 4/10\n",
            "52/52 [==============================] - 2s 31ms/step - loss: 0.5713 - accuracy: 0.8242 - val_loss: 0.2979 - val_accuracy: 0.9478\n",
            "Epoch 5/10\n",
            "52/52 [==============================] - 2s 31ms/step - loss: 0.3692 - accuracy: 0.8870 - val_loss: 0.2303 - val_accuracy: 0.9304\n",
            "Epoch 6/10\n",
            "52/52 [==============================] - 2s 33ms/step - loss: 0.3598 - accuracy: 0.8889 - val_loss: 0.2706 - val_accuracy: 0.9304\n",
            "Epoch 7/10\n",
            "33/52 [==================>...........] - ETA: 0s - loss: 0.3581 - accuracy: 0.8924Buffered data was truncated after reaching the output size limit."
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2i6O-uoWqMkB"
      },
      "source": [
        "### Finding the best hyperparameters:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XpcuKb6RlG6x",
        "outputId": "9acbae01-77ee-4a19-91fb-a7fd77dd6345",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "# View hyperparameters of best neural network\n",
        "print(\"Best parameters: {}\".format(cnn_drop_search.best_params_))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Best parameters: {'st2': 1, 'st1': 1, 'nf2': 256, 'nf1': 256, 'n_units': 100, 'lrt': 0.001, 'ks2': 3, 'ks1': 3, 'epochs': 15, 'drop_check': 1, 'batch_size': 20}\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "B_kQGWinqdss"
      },
      "source": [
        "### Evaluating Cross-Validation Performance:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vj5Ty2h5kiZX",
        "outputId": "a1c235e9-b8b3-44cd-b68e-64e16d772632",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "print(\"Best cross-validation accuracy: {:.2f}\".format(cnn_drop_search.best_score_))\n",
        "print(\"Best test accuracy: {:.2f}\".format(cnn_drop_search.score(X_test, y_test)))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Best cross-validation accuracy: 0.99\n",
            "Best test accuracy: 0.97\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "W_06Zoc1qVOL"
      },
      "source": [
        "### Fitting the best parameter:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "srmViJY1hlHm",
        "outputId": "e63cfab2-6029-4283-e36b-1d86d589d1bb",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 527
        }
      },
      "source": [
        "final_drop_model=cnn_drop_search.best_estimator_.fit(X_train, y_train,validation_split=0.1)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 1.6366 - accuracy: 0.4880 - val_loss: 0.5038 - val_accuracy: 0.8681\n",
            "Epoch 2/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.4007 - accuracy: 0.8855 - val_loss: 0.1937 - val_accuracy: 0.9583\n",
            "Epoch 3/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.2171 - accuracy: 0.9374 - val_loss: 0.2066 - val_accuracy: 0.9375\n",
            "Epoch 4/15\n",
            "65/65 [==============================] - 5s 80ms/step - loss: 0.1236 - accuracy: 0.9575 - val_loss: 0.1198 - val_accuracy: 0.9653\n",
            "Epoch 5/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.1164 - accuracy: 0.9606 - val_loss: 0.0857 - val_accuracy: 0.9722\n",
            "Epoch 6/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.1021 - accuracy: 0.9691 - val_loss: 0.0528 - val_accuracy: 0.9861\n",
            "Epoch 7/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0749 - accuracy: 0.9783 - val_loss: 0.0316 - val_accuracy: 0.9931\n",
            "Epoch 8/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0576 - accuracy: 0.9822 - val_loss: 0.0453 - val_accuracy: 0.9861\n",
            "Epoch 9/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0455 - accuracy: 0.9853 - val_loss: 0.0298 - val_accuracy: 0.9931\n",
            "Epoch 10/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0380 - accuracy: 0.9892 - val_loss: 0.0177 - val_accuracy: 0.9931\n",
            "Epoch 11/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0334 - accuracy: 0.9869 - val_loss: 0.0210 - val_accuracy: 0.9931\n",
            "Epoch 12/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0316 - accuracy: 0.9907 - val_loss: 0.0241 - val_accuracy: 0.9861\n",
            "Epoch 13/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0299 - accuracy: 0.9907 - val_loss: 0.0177 - val_accuracy: 0.9931\n",
            "Epoch 14/15\n",
            "65/65 [==============================] - 5s 80ms/step - loss: 0.0175 - accuracy: 0.9961 - val_loss: 0.0272 - val_accuracy: 0.9931\n",
            "Epoch 15/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0182 - accuracy: 0.9946 - val_loss: 0.0337 - val_accuracy: 0.9931\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "M1kyQhayCEJM"
      },
      "source": [
        "### View model summary"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QtMhDuMk5KYl"
      },
      "source": [
        "new_model = final_drop_model.model"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "aaV_6yCBYIeA",
        "outputId": "8cdc94f2-61b5-4744-bf81-971f85acf610",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 391
        }
      },
      "source": [
        "new_model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_502\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv2d_1004 (Conv2D)         (None, 8, 8, 256)         2560      \n",
            "_________________________________________________________________\n",
            "conv2d_1005 (Conv2D)         (None, 6, 6, 256)         590080    \n",
            "_________________________________________________________________\n",
            "max_pooling2d_367 (MaxPoolin (None, 3, 3, 256)         0         \n",
            "_________________________________________________________________\n",
            "flatten_367 (Flatten)        (None, 2304)              0         \n",
            "_________________________________________________________________\n",
            "dropout_367 (Dropout)        (None, 2304)              0         \n",
            "_________________________________________________________________\n",
            "dense_734 (Dense)            (None, 100)               230500    \n",
            "_________________________________________________________________\n",
            "dense_735 (Dense)            (None, 10)                1010      \n",
            "=================================================================\n",
            "Total params: 824,150\n",
            "Trainable params: 824,150\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VdpRgBjojnj1"
      },
      "source": [
        "### Evaluating Final Model Performance:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AhlXeJixY7MD",
        "outputId": "0e4be607-3ec9-4b0f-a6fe-21abf43fc9d1",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        }
      },
      "source": [
        "model_loss = pd.DataFrame(final_drop_model.history)\n",
        "model_loss.plot()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7fbcfcaba7b8>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 24
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pYz9MxxQEUeZ",
        "outputId": "f870ba85-013c-4e1c-c0bb-e7c41c23aa93",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "loss, acc = new_model.evaluate(X_test, y_test,verbose=0)\n",
        "print('\\nTesting loss: %.2f, acc: %.2f%%'%(loss, acc))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\n",
            "Testing loss: 0.04, acc: 0.99%\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QcjRjst1EVPz",
        "outputId": "df06fbcf-f51f-44d7-d779-167db5e4ed4b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "loss, acc = new_model.evaluate(X_train, y_train,verbose=0)\n",
        "print('\\nTraining loss: %.2f, acc: %.2f%%'%(loss, acc))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\n",
            "Training loss: 0.01, acc: 1.00%\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "deXN40lG51r5"
      },
      "source": [
        "### Visualize Training History"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FqB7P3RR2TuM",
        "outputId": "03797349-f7f2-4cb3-81bd-8c7ba426eda9",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        }
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "# Get training and test loss histories\n",
        "training_loss = final_drop_model.history[\"loss\"]\n",
        "test_loss = final_drop_model.history[\"val_loss\"]\n",
        "# Create count of the number of epochs\n",
        "epoch_count = range(1, len(training_loss) + 1)\n",
        "# Visualize loss history\n",
        "plt.plot(epoch_count, training_loss, \"r--\")\n",
        "plt.plot(epoch_count, test_loss, \"b-\")\n",
        "plt.legend([\"Training Loss\", \"Test Loss\"])\n",
        "plt.xlabel(\"Epoch\")\n",
        "plt.ylabel(\"Loss\")\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DpGgvNBY5d3-",
        "outputId": "b6b90471-74b9-4eed-f87e-583d114de9cd",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        }
      },
      "source": [
        "# Get training and test accuracy histories\n",
        "training_accuracy = final_drop_model.history[\"accuracy\"]\n",
        "test_accuracy = final_drop_model.history[\"val_accuracy\"]\n",
        "plt.plot(epoch_count, training_accuracy, \"r--\")\n",
        "plt.plot(epoch_count, test_accuracy, \"b-\")\n",
        "# Visualize accuracy history\n",
        "plt.legend([\"Training Accuracy\", \"Test Accuracy\"])\n",
        "plt.xlabel(\"Epoch\")\n",
        "plt.ylabel(\"Accuracy Score\")\n",
        "plt.show();\n",
        "\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MakfbDFuK9Q-"
      },
      "source": [
        "### Compute the mean and standard deviation of the best CNN model accuracy for 10 times in other way:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Y7XMsZwdLB8Z"
      },
      "source": [
        "#Create model with the best hyperparameters\n",
        "model=keras.models.Sequential()\n",
        "model.add(keras.layers.Conv2D(filters=400, kernel_size=5, strides=1, activation='elu', input_shape=(8,8,1), padding='same'))\n",
        "model.add(keras.layers.Conv2D(filters=32, kernel_size=5, strides=2, activation='relu'))\n",
        "model.add(keras.layers.MaxPool2D(pool_size=(2,2), strides=2, padding='same' ))\n",
        "model.add(keras.layers.Flatten())\n",
        "model.add(keras.layers.Dropout(0.5))\n",
        "model.add(keras.layers.Dense(100, activation='relu'))\n",
        "model.add(keras.layers.Dense(10, activation='softmax'))\n",
        "model.compile(optimizer=keras.optimizers.Adam(lr=0.001), loss='sparse_categorical_crossentropy', metrics=['accuracy'])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VQNbCKo5LGoK",
        "outputId": "6e67948f-f1b1-476a-9f13-5c6c15219b72",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "test_cnn_acc = []\n",
        "\n",
        "for i in range(1000):\n",
        "    #split test and train data\n",
        "    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)\n",
        "    \n",
        "    #Reshape inputs\n",
        "    X_train=X_train.reshape((X_train.shape[0],8,8,1))\n",
        "    X_test=X_test.reshape((X_test.shape[0],8,8,1))\n",
        "    \n",
        "    X_test = np.array(X_test)\n",
        "    y_test = np.array(y_test)\n",
        "\n",
        "    X_train = X_train / 255.0\n",
        "    X_test = X_test / 255.0\n",
        "    #Fit model on training data\n",
        "    cnn_drop_search.best_estimator_.fit(X_train, y_train,validation_split=0.1)\n",
        "    #history = model.fit(X_train, y_train, epochs=10,batch_size=20,verbose=2,validation_split=0.1)\n",
        "    \n",
        "    # Evaluate model on test data\n",
        "    acc = cnn_drop_search.best_estimator_.score(X_test, y_test)\n",
        "    #acc=history.score(X_test, y_test)\n",
        "    #acc=model.evaluate(X_test,y_test)\n",
        "    #acc = history.history['accuracy']\n",
        "    test_cnn_acc += [acc]\n",
        "\n",
        "print(\"Mean Test Accuracy:{:.2f}\".format(np.mean(test_cnn_acc)))\n",
        "print(\"standard deviation Test Accuracy:{:.2f}\".format(np.std(test_cnn_acc)))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 1.6106 - accuracy: 0.5151 - val_loss: 0.5725 - val_accuracy: 0.8333\n",
            "Epoch 2/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.3436 - accuracy: 0.8940 - val_loss: 0.1782 - val_accuracy: 0.9514\n",
            "Epoch 3/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.1978 - accuracy: 0.9312 - val_loss: 0.1512 - val_accuracy: 0.9514\n",
            "Epoch 4/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.1238 - accuracy: 0.9613 - val_loss: 0.0808 - val_accuracy: 0.9792\n",
            "Epoch 5/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.1176 - accuracy: 0.9644 - val_loss: 0.0763 - val_accuracy: 0.9861\n",
            "Epoch 6/15\n",
            "65/65 [==============================] - 5s 78ms/step - loss: 0.0665 - accuracy: 0.9737 - val_loss: 0.0545 - val_accuracy: 0.9792\n",
            "Epoch 7/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0777 - accuracy: 0.9714 - val_loss: 0.0509 - val_accuracy: 0.9792\n",
            "Epoch 8/15\n",
            "65/65 [==============================] - 5s 78ms/step - loss: 0.0476 - accuracy: 0.9853 - val_loss: 0.0306 - val_accuracy: 0.9931\n",
            "Epoch 9/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0315 - accuracy: 0.9876 - val_loss: 0.0257 - val_accuracy: 0.9931\n",
            "Epoch 10/15\n",
            "65/65 [==============================] - 5s 78ms/step - loss: 0.0402 - accuracy: 0.9845 - val_loss: 0.0664 - val_accuracy: 0.9722\n",
            "Epoch 11/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0214 - accuracy: 0.9946 - val_loss: 0.0231 - val_accuracy: 0.9931\n",
            "Epoch 12/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0208 - accuracy: 0.9938 - val_loss: 0.0135 - val_accuracy: 1.0000\n",
            "Epoch 13/15\n",
            "65/65 [==============================] - 5s 78ms/step - loss: 0.0239 - accuracy: 0.9938 - val_loss: 0.0279 - val_accuracy: 0.9931\n",
            "Epoch 14/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0358 - accuracy: 0.9845 - val_loss: 0.0171 - val_accuracy: 1.0000\n",
            "Epoch 15/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0197 - accuracy: 0.9946 - val_loss: 0.0244 - val_accuracy: 0.9931\n",
            "18/18 [==============================] - 0s 18ms/step - loss: 0.1301 - accuracy: 0.9833\n",
            "Epoch 1/15\n",
            "65/65 [==============================] - 5s 80ms/step - loss: 1.6130 - accuracy: 0.5012 - val_loss: 0.6467 - val_accuracy: 0.8125\n",
            "Epoch 2/15\n",
            "65/65 [==============================] - 5s 78ms/step - loss: 0.3573 - accuracy: 0.8840 - val_loss: 0.2438 - val_accuracy: 0.9375\n",
            "Epoch 3/15\n",
            "65/65 [==============================] - 5s 78ms/step - loss: 0.1781 - accuracy: 0.9513 - val_loss: 0.1785 - val_accuracy: 0.9514\n",
            "Epoch 4/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.1115 - accuracy: 0.9606 - val_loss: 0.1694 - val_accuracy: 0.9444\n",
            "Epoch 5/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0844 - accuracy: 0.9691 - val_loss: 0.1295 - val_accuracy: 0.9653\n",
            "Epoch 6/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0731 - accuracy: 0.9768 - val_loss: 0.0882 - val_accuracy: 0.9722\n",
            "Epoch 7/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0496 - accuracy: 0.9861 - val_loss: 0.1011 - val_accuracy: 0.9722\n",
            "Epoch 8/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0546 - accuracy: 0.9830 - val_loss: 0.1065 - val_accuracy: 0.9722\n",
            "Epoch 9/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0445 - accuracy: 0.9830 - val_loss: 0.0916 - val_accuracy: 0.9653\n",
            "Epoch 10/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0475 - accuracy: 0.9869 - val_loss: 0.1090 - val_accuracy: 0.9653\n",
            "Epoch 11/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0299 - accuracy: 0.9884 - val_loss: 0.1381 - val_accuracy: 0.9653\n",
            "Epoch 12/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0351 - accuracy: 0.9884 - val_loss: 0.1119 - val_accuracy: 0.9722\n",
            "Epoch 13/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0241 - accuracy: 0.9915 - val_loss: 0.0914 - val_accuracy: 0.9861\n",
            "Epoch 14/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0199 - accuracy: 0.9961 - val_loss: 0.1451 - val_accuracy: 0.9653\n",
            "Epoch 15/15\n",
            "65/65 [==============================] - 5s 78ms/step - loss: 0.0138 - accuracy: 0.9961 - val_loss: 0.1112 - val_accuracy: 0.9792\n",
            "18/18 [==============================] - 0s 18ms/step - loss: 0.1379 - accuracy: 0.9722\n",
            "Epoch 1/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 1.6570 - accuracy: 0.4919 - val_loss: 0.4400 - val_accuracy: 0.9236\n",
            "Epoch 2/15\n",
            "65/65 [==============================] - 5s 78ms/step - loss: 0.3632 - accuracy: 0.8902 - val_loss: 0.1573 - val_accuracy: 0.9583\n",
            "Epoch 3/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.2201 - accuracy: 0.9234 - val_loss: 0.0958 - val_accuracy: 0.9653\n",
            "Epoch 4/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.1355 - accuracy: 0.9590 - val_loss: 0.0869 - val_accuracy: 0.9792\n",
            "Epoch 5/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.1201 - accuracy: 0.9606 - val_loss: 0.0444 - val_accuracy: 0.9931\n",
            "Epoch 6/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0751 - accuracy: 0.9783 - val_loss: 0.1326 - val_accuracy: 0.9444\n",
            "Epoch 7/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0746 - accuracy: 0.9745 - val_loss: 0.0481 - val_accuracy: 0.9861\n",
            "Epoch 8/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0629 - accuracy: 0.9814 - val_loss: 0.0363 - val_accuracy: 0.9861\n",
            "Epoch 9/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0475 - accuracy: 0.9838 - val_loss: 0.0527 - val_accuracy: 0.9861\n",
            "Epoch 10/15\n",
            "65/65 [==============================] - 5s 80ms/step - loss: 0.0397 - accuracy: 0.9861 - val_loss: 0.0349 - val_accuracy: 0.9931\n",
            "Epoch 11/15\n",
            "65/65 [==============================] - 5s 80ms/step - loss: 0.0328 - accuracy: 0.9869 - val_loss: 0.0217 - val_accuracy: 0.9931\n",
            "Epoch 12/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0280 - accuracy: 0.9930 - val_loss: 0.1094 - val_accuracy: 0.9653\n",
            "Epoch 13/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0332 - accuracy: 0.9884 - val_loss: 0.0459 - val_accuracy: 0.9861\n",
            "Epoch 14/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0322 - accuracy: 0.9876 - val_loss: 0.0213 - val_accuracy: 0.9861\n",
            "Epoch 15/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0305 - accuracy: 0.9899 - val_loss: 0.0385 - val_accuracy: 0.9861\n",
            "18/18 [==============================] - 0s 18ms/step - loss: 0.0559 - accuracy: 0.9750\n",
            "Epoch 1/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 1.6327 - accuracy: 0.5143 - val_loss: 0.6150 - val_accuracy: 0.8403\n",
            "Epoch 2/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.3378 - accuracy: 0.9056 - val_loss: 0.2931 - val_accuracy: 0.9167\n",
            "Epoch 3/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.1822 - accuracy: 0.9397 - val_loss: 0.2106 - val_accuracy: 0.9236\n",
            "Epoch 4/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.1233 - accuracy: 0.9652 - val_loss: 0.2727 - val_accuracy: 0.8958\n",
            "Epoch 5/15\n",
            "65/65 [==============================] - 5s 80ms/step - loss: 0.0967 - accuracy: 0.9683 - val_loss: 0.2011 - val_accuracy: 0.9444\n",
            "Epoch 6/15\n",
            "65/65 [==============================] - 5s 80ms/step - loss: 0.0746 - accuracy: 0.9745 - val_loss: 0.2067 - val_accuracy: 0.9444\n",
            "Epoch 7/15\n",
            "65/65 [==============================] - 5s 80ms/step - loss: 0.0917 - accuracy: 0.9698 - val_loss: 0.1681 - val_accuracy: 0.9514\n",
            "Epoch 8/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0582 - accuracy: 0.9814 - val_loss: 0.1782 - val_accuracy: 0.9514\n",
            "Epoch 9/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0250 - accuracy: 0.9946 - val_loss: 0.1593 - val_accuracy: 0.9583\n",
            "Epoch 10/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0216 - accuracy: 0.9969 - val_loss: 0.2061 - val_accuracy: 0.9375\n",
            "Epoch 11/15\n",
            "65/65 [==============================] - 5s 80ms/step - loss: 0.0257 - accuracy: 0.9915 - val_loss: 0.1388 - val_accuracy: 0.9583\n",
            "Epoch 12/15\n",
            "65/65 [==============================] - 5s 79ms/step - loss: 0.0238 - accuracy: 0.9907 - val_loss: 0.1539 - val_accuracy: 0.9583\n",
            "Epoch 13/15\n",
            "65/65 [==============================] - 5s 80ms/step - loss: 0.0208 - accuracy: 0.9930 - val_loss: 0.1662 - val_accuracy: 0.9583\n",
            "Epoch 14/15\n",
            "65/65 [==============================] - 5s 80ms/step - loss: 0.0241 - accuracy: 0.9930 - val_loss: 0.1921 - val_accuracy: 0.9514\n",
            "Epoch 15/15\n",
            "65/65 [==============================] - 5s 80ms/step - loss: 0.0147 - accuracy: 0.9954 - val_loss: 0.1783 - val_accuracy: 0.9444\n",
            "18/18 [==============================] - 0s 18ms/step - loss: 0.0639 - accuracy: 0.9778\n",
            "Epoch 1/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 1.6717 - accuracy: 0.4571 - val_loss: 0.6568 - val_accuracy: 0.7708\n",
            "Epoch 2/15\n",
            "65/65 [==============================] - 5s 80ms/step - loss: 0.3347 - accuracy: 0.8987 - val_loss: 0.2797 - val_accuracy: 0.9167\n",
            "Epoch 3/15\n",
            "65/65 [==============================] - 5s 80ms/step - loss: 0.1907 - accuracy: 0.9335 - val_loss: 0.1979 - val_accuracy: 0.9375\n",
            "Epoch 4/15\n",
            "65/65 [==============================] - 5s 80ms/step - loss: 0.1118 - accuracy: 0.9637 - val_loss: 0.1746 - val_accuracy: 0.9444\n",
            "Epoch 5/15\n",
            "65/65 [==============================] - 5s 80ms/step - loss: 0.0851 - accuracy: 0.9768 - val_loss: 0.0938 - val_accuracy: 0.9653\n",
            "Epoch 6/15\n",
            "65/65 [==============================] - 5s 80ms/step - loss: 0.0532 - accuracy: 0.9830 - val_loss: 0.1165 - val_accuracy: 0.9583\n",
            "Epoch 7/15\n",
            "65/65 [==============================] - 5s 80ms/step - loss: 0.0502 - accuracy: 0.9799 - val_loss: 0.1379 - val_accuracy: 0.9653\n",
            "Epoch 8/15\n",
            "65/65 [==============================] - 5s 80ms/step - loss: 0.0373 - accuracy: 0.9899 - val_loss: 0.0983 - val_accuracy: 0.9792\n",
            "Epoch 9/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0368 - accuracy: 0.9884 - val_loss: 0.2574 - val_accuracy: 0.9514\n",
            "Epoch 10/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0402 - accuracy: 0.9838 - val_loss: 0.1340 - val_accuracy: 0.9653\n",
            "Epoch 11/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0286 - accuracy: 0.9907 - val_loss: 0.1497 - val_accuracy: 0.9583\n",
            "Epoch 12/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0145 - accuracy: 0.9977 - val_loss: 0.1358 - val_accuracy: 0.9583\n",
            "Epoch 13/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0179 - accuracy: 0.9954 - val_loss: 0.0794 - val_accuracy: 0.9792\n",
            "Epoch 14/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0260 - accuracy: 0.9907 - val_loss: 0.1530 - val_accuracy: 0.9514\n",
            "Epoch 15/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0163 - accuracy: 0.9954 - val_loss: 0.0893 - val_accuracy: 0.9792\n",
            "18/18 [==============================] - 0s 19ms/step - loss: 0.1035 - accuracy: 0.9722\n",
            "Epoch 1/15\n",
            "65/65 [==============================] - 5s 83ms/step - loss: 1.6553 - accuracy: 0.4756 - val_loss: 0.5057 - val_accuracy: 0.8472\n",
            "Epoch 2/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.3580 - accuracy: 0.8948 - val_loss: 0.1568 - val_accuracy: 0.9444\n",
            "Epoch 3/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.1848 - accuracy: 0.9474 - val_loss: 0.0754 - val_accuracy: 0.9931\n",
            "Epoch 4/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.1561 - accuracy: 0.9513 - val_loss: 0.0696 - val_accuracy: 0.9722\n",
            "Epoch 5/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.1101 - accuracy: 0.9637 - val_loss: 0.0421 - val_accuracy: 0.9931\n",
            "Epoch 6/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0999 - accuracy: 0.9698 - val_loss: 0.0683 - val_accuracy: 0.9722\n",
            "Epoch 7/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.0851 - accuracy: 0.9745 - val_loss: 0.0455 - val_accuracy: 0.9931\n",
            "Epoch 8/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0662 - accuracy: 0.9753 - val_loss: 0.0144 - val_accuracy: 1.0000\n",
            "Epoch 9/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0406 - accuracy: 0.9869 - val_loss: 0.0314 - val_accuracy: 0.9931\n",
            "Epoch 10/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0332 - accuracy: 0.9907 - val_loss: 0.0294 - val_accuracy: 0.9931\n",
            "Epoch 11/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0554 - accuracy: 0.9830 - val_loss: 0.0269 - val_accuracy: 0.9931\n",
            "Epoch 12/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0399 - accuracy: 0.9884 - val_loss: 0.0228 - val_accuracy: 1.0000\n",
            "Epoch 13/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0294 - accuracy: 0.9884 - val_loss: 0.0211 - val_accuracy: 0.9931\n",
            "Epoch 14/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0350 - accuracy: 0.9899 - val_loss: 0.0112 - val_accuracy: 0.9931\n",
            "Epoch 15/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0235 - accuracy: 0.9930 - val_loss: 0.0066 - val_accuracy: 1.0000\n",
            "18/18 [==============================] - 0s 19ms/step - loss: 0.0509 - accuracy: 0.9778\n",
            "Epoch 1/15\n",
            "65/65 [==============================] - 5s 83ms/step - loss: 1.6297 - accuracy: 0.5135 - val_loss: 0.5148 - val_accuracy: 0.8611\n",
            "Epoch 2/15\n",
            "65/65 [==============================] - 5s 80ms/step - loss: 0.2856 - accuracy: 0.9080 - val_loss: 0.1448 - val_accuracy: 0.9583\n",
            "Epoch 3/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.1610 - accuracy: 0.9505 - val_loss: 0.1159 - val_accuracy: 0.9583\n",
            "Epoch 4/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0914 - accuracy: 0.9737 - val_loss: 0.0996 - val_accuracy: 0.9583\n",
            "Epoch 5/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0936 - accuracy: 0.9698 - val_loss: 0.0452 - val_accuracy: 0.9861\n",
            "Epoch 6/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0555 - accuracy: 0.9807 - val_loss: 0.0337 - val_accuracy: 0.9931\n",
            "Epoch 7/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0504 - accuracy: 0.9838 - val_loss: 0.0574 - val_accuracy: 0.9722\n",
            "Epoch 8/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0394 - accuracy: 0.9884 - val_loss: 0.0295 - val_accuracy: 0.9931\n",
            "Epoch 9/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0371 - accuracy: 0.9876 - val_loss: 0.0254 - val_accuracy: 0.9931\n",
            "Epoch 10/15\n",
            "65/65 [==============================] - 5s 83ms/step - loss: 0.0306 - accuracy: 0.9907 - val_loss: 0.0438 - val_accuracy: 0.9792\n",
            "Epoch 11/15\n",
            "65/65 [==============================] - 5s 83ms/step - loss: 0.0286 - accuracy: 0.9907 - val_loss: 0.0247 - val_accuracy: 0.9861\n",
            "Epoch 12/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0293 - accuracy: 0.9915 - val_loss: 0.0185 - val_accuracy: 0.9931\n",
            "Epoch 13/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0199 - accuracy: 0.9938 - val_loss: 0.0133 - val_accuracy: 0.9931\n",
            "Epoch 14/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.0223 - accuracy: 0.9907 - val_loss: 0.0136 - val_accuracy: 1.0000\n",
            "Epoch 15/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.0404 - accuracy: 0.9853 - val_loss: 0.0220 - val_accuracy: 0.9931\n",
            "18/18 [==============================] - 0s 19ms/step - loss: 0.1718 - accuracy: 0.9583\n",
            "Epoch 1/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 1.6451 - accuracy: 0.4950 - val_loss: 0.4618 - val_accuracy: 0.8403\n",
            "Epoch 2/15\n",
            "65/65 [==============================] - 5s 80ms/step - loss: 0.3211 - accuracy: 0.9056 - val_loss: 0.2052 - val_accuracy: 0.9167\n",
            "Epoch 3/15\n",
            "65/65 [==============================] - 5s 80ms/step - loss: 0.1743 - accuracy: 0.9474 - val_loss: 0.1053 - val_accuracy: 0.9653\n",
            "Epoch 4/15\n",
            "65/65 [==============================] - 5s 80ms/step - loss: 0.1093 - accuracy: 0.9706 - val_loss: 0.0767 - val_accuracy: 0.9722\n",
            "Epoch 5/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0636 - accuracy: 0.9845 - val_loss: 0.0734 - val_accuracy: 0.9861\n",
            "Epoch 6/15\n",
            "65/65 [==============================] - 5s 84ms/step - loss: 0.0666 - accuracy: 0.9722 - val_loss: 0.1913 - val_accuracy: 0.9583\n",
            "Epoch 7/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0522 - accuracy: 0.9830 - val_loss: 0.0750 - val_accuracy: 0.9792\n",
            "Epoch 8/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0369 - accuracy: 0.9876 - val_loss: 0.1145 - val_accuracy: 0.9722\n",
            "Epoch 9/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0327 - accuracy: 0.9915 - val_loss: 0.0617 - val_accuracy: 0.9792\n",
            "Epoch 10/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.0340 - accuracy: 0.9892 - val_loss: 0.0631 - val_accuracy: 0.9792\n",
            "Epoch 11/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0295 - accuracy: 0.9915 - val_loss: 0.0503 - val_accuracy: 0.9792\n",
            "Epoch 12/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0249 - accuracy: 0.9899 - val_loss: 0.0569 - val_accuracy: 0.9861\n",
            "Epoch 13/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0469 - accuracy: 0.9853 - val_loss: 0.0622 - val_accuracy: 0.9931\n",
            "Epoch 14/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0214 - accuracy: 0.9923 - val_loss: 0.0679 - val_accuracy: 0.9583\n",
            "Epoch 15/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0269 - accuracy: 0.9907 - val_loss: 0.0777 - val_accuracy: 0.9861\n",
            "18/18 [==============================] - 0s 20ms/step - loss: 0.0808 - accuracy: 0.9778\n",
            "Epoch 1/15\n",
            "65/65 [==============================] - 5s 83ms/step - loss: 1.6706 - accuracy: 0.5081 - val_loss: 0.4437 - val_accuracy: 0.8403\n",
            "Epoch 2/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.3553 - accuracy: 0.8910 - val_loss: 0.2163 - val_accuracy: 0.9167\n",
            "Epoch 3/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.1817 - accuracy: 0.9428 - val_loss: 0.1013 - val_accuracy: 0.9722\n",
            "Epoch 4/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.1430 - accuracy: 0.9567 - val_loss: 0.0995 - val_accuracy: 0.9792\n",
            "Epoch 5/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.0950 - accuracy: 0.9706 - val_loss: 0.1042 - val_accuracy: 0.9583\n",
            "Epoch 6/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.0721 - accuracy: 0.9799 - val_loss: 0.0313 - val_accuracy: 0.9931\n",
            "Epoch 7/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.0445 - accuracy: 0.9892 - val_loss: 0.0334 - val_accuracy: 0.9861\n",
            "Epoch 8/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0481 - accuracy: 0.9830 - val_loss: 0.0322 - val_accuracy: 0.9861\n",
            "Epoch 9/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0435 - accuracy: 0.9838 - val_loss: 0.0655 - val_accuracy: 0.9861\n",
            "Epoch 10/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0466 - accuracy: 0.9822 - val_loss: 0.0266 - val_accuracy: 0.9931\n",
            "Epoch 11/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.0303 - accuracy: 0.9892 - val_loss: 0.0880 - val_accuracy: 0.9722\n",
            "Epoch 12/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.0415 - accuracy: 0.9853 - val_loss: 0.0337 - val_accuracy: 0.9931\n",
            "Epoch 13/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.0226 - accuracy: 0.9930 - val_loss: 0.0664 - val_accuracy: 0.9861\n",
            "Epoch 14/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.0209 - accuracy: 0.9930 - val_loss: 0.0420 - val_accuracy: 0.9861\n",
            "Epoch 15/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.0211 - accuracy: 0.9915 - val_loss: 0.0239 - val_accuracy: 0.9931\n",
            "18/18 [==============================] - 0s 19ms/step - loss: 0.0587 - accuracy: 0.9750\n",
            "Epoch 1/15\n",
            "65/65 [==============================] - 5s 84ms/step - loss: 1.6894 - accuracy: 0.4772 - val_loss: 0.5835 - val_accuracy: 0.8125\n",
            "Epoch 2/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.3610 - accuracy: 0.8809 - val_loss: 0.2036 - val_accuracy: 0.9306\n",
            "Epoch 3/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.1861 - accuracy: 0.9474 - val_loss: 0.1119 - val_accuracy: 0.9722\n",
            "Epoch 4/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.1158 - accuracy: 0.9667 - val_loss: 0.0871 - val_accuracy: 0.9792\n",
            "Epoch 5/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.0859 - accuracy: 0.9706 - val_loss: 0.0655 - val_accuracy: 0.9861\n",
            "Epoch 6/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.0913 - accuracy: 0.9729 - val_loss: 0.0867 - val_accuracy: 0.9792\n",
            "Epoch 7/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.0693 - accuracy: 0.9760 - val_loss: 0.0741 - val_accuracy: 0.9861\n",
            "Epoch 8/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.0488 - accuracy: 0.9814 - val_loss: 0.0485 - val_accuracy: 0.9861\n",
            "Epoch 9/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.0280 - accuracy: 0.9938 - val_loss: 0.0744 - val_accuracy: 0.9792\n",
            "Epoch 10/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.0393 - accuracy: 0.9869 - val_loss: 0.0426 - val_accuracy: 0.9931\n",
            "Epoch 11/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.0222 - accuracy: 0.9930 - val_loss: 0.0489 - val_accuracy: 0.9792\n",
            "Epoch 12/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.0303 - accuracy: 0.9876 - val_loss: 0.0527 - val_accuracy: 0.9792\n",
            "Epoch 13/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.0222 - accuracy: 0.9907 - val_loss: 0.0638 - val_accuracy: 0.9861\n",
            "Epoch 14/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.0347 - accuracy: 0.9869 - val_loss: 0.0544 - val_accuracy: 0.9861\n",
            "Epoch 15/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.0175 - accuracy: 0.9961 - val_loss: 0.0502 - val_accuracy: 0.9931\n",
            "18/18 [==============================] - 0s 19ms/step - loss: 0.0427 - accuracy: 0.9889\n",
            "Epoch 1/15\n",
            "65/65 [==============================] - 5s 83ms/step - loss: 1.6747 - accuracy: 0.4803 - val_loss: 0.4829 - val_accuracy: 0.8403\n",
            "Epoch 2/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.3096 - accuracy: 0.9087 - val_loss: 0.1519 - val_accuracy: 0.9514\n",
            "Epoch 3/15\n",
            "65/65 [==============================] - 5s 81ms/step - loss: 0.1748 - accuracy: 0.9397 - val_loss: 0.1050 - val_accuracy: 0.9722\n",
            "Epoch 4/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.1181 - accuracy: 0.9683 - val_loss: 0.0989 - val_accuracy: 0.9514\n",
            "Epoch 5/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.0977 - accuracy: 0.9644 - val_loss: 0.0810 - val_accuracy: 0.9722\n",
            "Epoch 6/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.0732 - accuracy: 0.9807 - val_loss: 0.0349 - val_accuracy: 0.9931\n",
            "Epoch 7/15\n",
            "65/65 [==============================] - 5s 84ms/step - loss: 0.0678 - accuracy: 0.9814 - val_loss: 0.0316 - val_accuracy: 0.9931\n",
            "Epoch 8/15\n",
            "65/65 [==============================] - 5s 84ms/step - loss: 0.0501 - accuracy: 0.9845 - val_loss: 0.0308 - val_accuracy: 0.9931\n",
            "Epoch 9/15\n",
            "65/65 [==============================] - 5s 84ms/step - loss: 0.0473 - accuracy: 0.9861 - val_loss: 0.0337 - val_accuracy: 0.9861\n",
            "Epoch 10/15\n",
            "65/65 [==============================] - 6s 85ms/step - loss: 0.0443 - accuracy: 0.9884 - val_loss: 0.0385 - val_accuracy: 0.9931\n",
            "Epoch 11/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.0541 - accuracy: 0.9783 - val_loss: 0.0383 - val_accuracy: 0.9792\n",
            "Epoch 12/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.0363 - accuracy: 0.9884 - val_loss: 0.0302 - val_accuracy: 0.9861\n",
            "Epoch 13/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.0302 - accuracy: 0.9884 - val_loss: 0.0111 - val_accuracy: 1.0000\n",
            "Epoch 14/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.0192 - accuracy: 0.9923 - val_loss: 0.0271 - val_accuracy: 0.9861\n",
            "Epoch 15/15\n",
            "65/65 [==============================] - 5s 82ms/step - loss: 0.0144 - accuracy: 0.9961 - val_loss: 0.0390 - val_accuracy: 0.9792\n",
            " 9/18 [==============>...............] - ETA: 0s - loss: 0.1121 - accuracy: 0.9722"
          ],
          "name": "stdout"
        }
      ]
    }
  ]
}