RNA_RNAInteractionEmbedding.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "RNA-RNAInteractionEmbedding.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "UVgmpeXRjoXq"
      },
      "source": [
        "from google.colab import auth\n",
        "import pandas as pd\n",
        "import gspread\n",
        "from oauth2client.client import GoogleCredentials\n",
        "\n",
        "from keras import layers\n",
        "import numpy as np\n",
        "from keras.models import Model\n",
        "import pandas as pd\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Embedding"
      ],
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kykqczf6f0RL"
      },
      "source": [
        "auth.authenticate_user()\n",
        "gc = gspread.authorize(GoogleCredentials.get_application_default())\n",
        "\"We took our dataset from: SPLASH: in vivo mapping of eukaryotic RNA interactomes\"\n",
        "worksheet = gc.open('shuffled_final_splash_df').sheet1\n",
        "rows = worksheet.get_all_values()\n",
        "df = pd.DataFrame(rows)\n",
        "df.columns = df.iloc[0]\n",
        "df = df.iloc[1:]"
      ],
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-b0uCr6of4Kt"
      },
      "source": [
        "def EncodeSequences(dataframe):\n",
        "    \"\"\"\n",
        "    Using Squares of Pure numbers as codes for each nucleotide(Other Encoding \n",
        "    methods such as one hot encoding, and encoding based on number of hydrogen bonds of each nucleotide\n",
        "    have been tested but integer encoding reached higher accuracy.)\n",
        "    :param dataframe: contains [two RNA sequences] + [one-hot label (whether two sequences have interaction or not)]\n",
        "    :return: Two encoded RNA sequences \n",
        "    \"\"\"\n",
        "    seq1_ = list(dataframe['seq1'])\n",
        "    seq2_ = list(dataframe['seq2'])\n",
        "    seq1s = []\n",
        "    seq2s = []\n",
        "    for s in seq1_:\n",
        "      # =========== Encode Sequence 1  ==========\n",
        "      seq1 = [529 if x == 'A' else x for x in s]\n",
        "      seq1 = [361 if x == 'U' else x for x in seq1]\n",
        "      seq1 = [121 if x == 'C' else x for x in seq1]\n",
        "      seq1 = [169 if x == 'G' else x for x in seq1]\n",
        "      seq1 = [0 if x == 'N' else x for x in seq1]\n",
        "\n",
        "      seq1s.append(seq1)\n",
        "    for s in seq2_:\n",
        "      # ============ Encode Sequence 2 ==========\n",
        "      seq2 = [529 if x == 'A' else x for x in s]\n",
        "      seq2 = [361 if x == 'U' else x for x in seq2]\n",
        "      seq2 = [121 if x == 'C' else x for x in seq2]\n",
        "      seq2 = [169 if x == 'G' else x for x in seq2]\n",
        "      seq2 = [0 if x == 'N' else x for x in seq2]\n",
        "      seq2s.append(seq2)\n",
        "\n",
        "    seq1s = np.array(seq1s)\n",
        "    seq2s = np.array(seq2s)\n",
        "\n",
        "    return seq1s, seq2s"
      ],
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "seegEF0rmVEV"
      },
      "source": [
        "def model(trainSeq1, trainSeq2, trainLabel, testSeq1, testSeq2, testLabel):\n",
        "    \"\"\"\n",
        "    Multi-input model which learn features of the RNA transcripts separately, \n",
        "    then concatenate feature vectors of RNA sequences and fed the new feature \n",
        "    vector to three fully connected layers.\n",
        "    :param trainSeq1, trainSeq2, trainLabel, testSeq1, testSeq2, testLabel\n",
        "    :return: \n",
        "\n",
        "    \"\"\"\n",
        "\n",
        "    input_1 = layers.Input(\n",
        "        shape=( np.shape(trainSeq1)[1]))\n",
        "\n",
        "    emb1 = layers.Embedding(530 , 10, input_length = np.shape(trainSeq1)[1])(input_1)         \n",
        "    emb1 = layers.Reshape((np.shape(trainSeq1)[1], 10))(emb1)\n",
        "\n",
        "    cnn1 = layers.Conv1D(128, kernel_size=3, activation = 'relu')(emb1)\n",
        "    cnn1 = layers.Conv1D(64, kernel_size=3, activation = 'relu')(cnn1)\n",
        "    cnn1 = layers.MaxPooling1D(pool_size=2)(cnn1)\n",
        "    cnn1 = layers.Conv1D(32, kernel_size=3, activation = 'relu')(cnn1)\n",
        "    cnn1 = layers.MaxPooling1D(pool_size=2)(cnn1)\n",
        "    cnn1 = layers.Conv1D(32, kernel_size=3, activation = 'relu')(cnn1)\n",
        "    cnn1 = layers.MaxPooling1D(pool_size=2)(cnn1)\n",
        "\n",
        "    cnn1 = layers.Flatten()(cnn1)\n",
        "\n",
        "\n",
        "    input_2 = layers.Input(\n",
        "        shape=( np.shape(trainSeq1)[1]))\n",
        "    \n",
        "    emb2 = Embedding(530 , 10, input_length = np.shape(trainSeq1)[1])(input_2)\n",
        "    emb2 = layers.Reshape((np.shape(trainSeq1)[1], 10))(emb2)\n",
        "\n",
        "    cnn2 = layers.Conv1D(filters=128, kernel_size=3, activation = 'relu')(emb2)\n",
        "    cnn2 = layers.Conv1D(filters=64, kernel_size=3, activation = 'relu')(cnn2)\n",
        "    cnn2 = layers.MaxPooling1D(pool_size=2)(cnn2)\n",
        "    cnn2 = layers.Conv1D(filters=32, kernel_size=3, activation = 'relu')(cnn2)\n",
        "    cnn2 = layers.MaxPooling1D(pool_size=2)(cnn2)\n",
        "    cnn2 = layers.Conv1D(filters=32, kernel_size=3, activation = 'relu')(cnn2)\n",
        "    cnn2 = layers.MaxPooling1D(pool_size=2)(cnn2)\n",
        "    cnn2 = layers.Flatten()(cnn2)\n",
        "\n",
        "\n",
        "    merged_vector = layers.Add()([cnn1, cnn2])\n",
        "\n",
        "    fc = layers.Dense(128, activation='sigmoid')(merged_vector)\n",
        "    fc = layers.Dense(32, activation='relu')(fc)\n",
        " \n",
        "    out_put = layers.Dense(2, activation='softmax')(fc)             \n",
        "\n",
        "\n",
        "    model = Model(inputs = [input_1, input_2], outputs = out_put)\n",
        "\n",
        "    model.compile(optimizer='adam', loss='binary_crossentropy', metrics=['accuracy'])\n",
        "    \n",
        "    history = model.fit([trainSeq1, trainSeq2], trainLabel, validation_data = ([testSeq1, testSeq2], testLabel), batch_size= 64, epochs=20, verbose=1)\n",
        "    model.summary()\n",
        "    score = model.evaluate([testSeq1, testSeq2], testLabel, verbose=1)\n",
        "\n",
        "    return score[1], history"
      ],
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vu1E3FlKp4Q8",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "eabbcfb8-dbc2-4cf9-9f60-7cda913d31a6"
      },
      "source": [
        "from sklearn.model_selection import KFold\n",
        "from matplotlib import pyplot as plt\n",
        "\"10 Fold Cross Validation\"\n",
        "kfold = KFold(10, True, 1)\n",
        "acc = 0\n",
        "cnt = 0\n",
        "for train, test in kfold.split(df):\n",
        "  train_df = df.iloc[train, :]\n",
        "  test_df = df.iloc[test, :]\n",
        "\n",
        "  seq1_train, seq2_train = EncodeSequences(train_df)\n",
        "  seq1_test, seq2_test = EncodeSequences(test_df)\n",
        "\n",
        "  label_train = np.array(train_df[['tag1', 'tag2']].values.astype(int).tolist())\n",
        "  label_test = np.array(test_df[['tag1', 'tag2']].values.astype(int).tolist())\n",
        "  score, history = model(seq1_train, seq2_train, label_train, seq1_test, seq2_test, label_test)\n",
        "  acc += score\n",
        "  cnt += 1\n",
        "  if (cnt == 10):\n",
        "    #  \"Accuracy\"\n",
        "    plt.plot(history.history['accuracy'])\n",
        "    plt.plot(history.history['val_accuracy'])\n",
        "    plt.title('model accuracy')\n",
        "    plt.ylabel('accuracy')\n",
        "    plt.xlabel('epoch')\n",
        "    plt.legend(['train', 'validation'], loc='upper left')\n",
        "    plt.show()\n",
        "    # \"Loss\"\n",
        "    plt.plot(history.history['loss'])\n",
        "    plt.plot(history.history['val_loss'])\n",
        "    plt.title('model loss')\n",
        "    plt.ylabel('loss')\n",
        "    plt.xlabel('epoch')\n",
        "    plt.legend(['train', 'validation'], loc='upper left')\n",
        "    plt.show()\n",
        "\n",
        "acc /= 10\n",
        "print(acc)"
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/20\n",
            "11/11 [==============================] - 3s 180ms/step - loss: 0.7047 - accuracy: 0.4937 - val_loss: 0.6895 - val_accuracy: 0.5641\n",
            "Epoch 2/20\n",
            "11/11 [==============================] - 2s 155ms/step - loss: 0.6965 - accuracy: 0.5268 - val_loss: 0.6935 - val_accuracy: 0.4359\n",
            "Epoch 3/20\n",
            "11/11 [==============================] - 2s 150ms/step - loss: 0.6987 - accuracy: 0.5147 - val_loss: 0.7003 - val_accuracy: 0.4359\n",
            "Epoch 4/20\n",
            "11/11 [==============================] - 2s 154ms/step - loss: 0.6927 - accuracy: 0.5066 - val_loss: 0.6898 - val_accuracy: 0.5641\n",
            "Epoch 5/20\n",
            "11/11 [==============================] - 2s 155ms/step - loss: 0.6939 - accuracy: 0.4873 - val_loss: 0.6989 - val_accuracy: 0.4359\n",
            "Epoch 6/20\n",
            "11/11 [==============================] - 2s 157ms/step - loss: 0.6944 - accuracy: 0.4917 - val_loss: 0.6940 - val_accuracy: 0.4359\n",
            "Epoch 7/20\n",
            "11/11 [==============================] - 2s 155ms/step - loss: 0.6911 - accuracy: 0.5059 - val_loss: 0.6844 - val_accuracy: 0.6538\n",
            "Epoch 8/20\n",
            "11/11 [==============================] - 2s 156ms/step - loss: 0.6714 - accuracy: 0.6615 - val_loss: 0.6038 - val_accuracy: 0.6923\n",
            "Epoch 9/20\n",
            "11/11 [==============================] - 2s 156ms/step - loss: 0.5885 - accuracy: 0.7032 - val_loss: 0.5172 - val_accuracy: 0.7308\n",
            "Epoch 10/20\n",
            "11/11 [==============================] - 2s 156ms/step - loss: 0.5214 - accuracy: 0.7504 - val_loss: 0.4692 - val_accuracy: 0.7821\n",
            "Epoch 11/20\n",
            "11/11 [==============================] - 2s 154ms/step - loss: 0.4597 - accuracy: 0.7776 - val_loss: 0.4394 - val_accuracy: 0.7692\n",
            "Epoch 12/20\n",
            "11/11 [==============================] - 2s 157ms/step - loss: 0.4160 - accuracy: 0.8192 - val_loss: 0.3860 - val_accuracy: 0.8077\n",
            "Epoch 13/20\n",
            "11/11 [==============================] - 2s 156ms/step - loss: 0.3506 - accuracy: 0.8435 - val_loss: 0.3034 - val_accuracy: 0.8846\n",
            "Epoch 14/20\n",
            "11/11 [==============================] - 2s 154ms/step - loss: 0.2596 - accuracy: 0.9099 - val_loss: 0.2465 - val_accuracy: 0.9231\n",
            "Epoch 15/20\n",
            "11/11 [==============================] - 2s 148ms/step - loss: 0.2057 - accuracy: 0.9390 - val_loss: 0.1924 - val_accuracy: 0.9359\n",
            "Epoch 16/20\n",
            "11/11 [==============================] - 2s 153ms/step - loss: 0.1501 - accuracy: 0.9575 - val_loss: 0.1579 - val_accuracy: 0.9615\n",
            "Epoch 17/20\n",
            "11/11 [==============================] - 2s 154ms/step - loss: 0.0936 - accuracy: 0.9728 - val_loss: 0.1409 - val_accuracy: 0.9359\n",
            "Epoch 18/20\n",
            "11/11 [==============================] - 2s 155ms/step - loss: 0.0785 - accuracy: 0.9772 - val_loss: 0.1138 - val_accuracy: 0.9359\n",
            "Epoch 19/20\n",
            "11/11 [==============================] - 2s 152ms/step - loss: 0.0383 - accuracy: 0.9919 - val_loss: 0.1414 - val_accuracy: 0.9487\n",
            "Epoch 20/20\n",
            "11/11 [==============================] - 2s 152ms/step - loss: 0.0497 - accuracy: 0.9870 - val_loss: 0.1010 - val_accuracy: 0.9487\n",
            "Model: \"model_22\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "input_45 (InputLayer)           [(None, 200)]        0                                            \n",
            "__________________________________________________________________________________________________\n",
            "input_46 (InputLayer)           [(None, 200)]        0                                            \n",
            "__________________________________________________________________________________________________\n",
            "embedding_44 (Embedding)        (None, 200, 10)      5300        input_45[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "embedding_45 (Embedding)        (None, 200, 10)      5300        input_46[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "reshape_44 (Reshape)            (None, 200, 10)      0           embedding_44[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "reshape_45 (Reshape)            (None, 200, 10)      0           embedding_45[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_176 (Conv1D)             (None, 198, 128)     3968        reshape_44[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_180 (Conv1D)             (None, 198, 128)     3968        reshape_45[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_177 (Conv1D)             (None, 196, 64)      24640       conv1d_176[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_181 (Conv1D)             (None, 196, 64)      24640       conv1d_180[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_132 (MaxPooling1D (None, 98, 64)       0           conv1d_177[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_135 (MaxPooling1D (None, 98, 64)       0           conv1d_181[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_178 (Conv1D)             (None, 96, 32)       6176        max_pooling1d_132[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_182 (Conv1D)             (None, 96, 32)       6176        max_pooling1d_135[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_133 (MaxPooling1D (None, 48, 32)       0           conv1d_178[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_136 (MaxPooling1D (None, 48, 32)       0           conv1d_182[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_179 (Conv1D)             (None, 46, 32)       3104        max_pooling1d_133[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_183 (Conv1D)             (None, 46, 32)       3104        max_pooling1d_136[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_134 (MaxPooling1D (None, 23, 32)       0           conv1d_179[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_137 (MaxPooling1D (None, 23, 32)       0           conv1d_183[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "flatten_44 (Flatten)            (None, 736)          0           max_pooling1d_134[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "flatten_45 (Flatten)            (None, 736)          0           max_pooling1d_137[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "add_22 (Add)                    (None, 736)          0           flatten_44[0][0]                 \n",
            "                                                                 flatten_45[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "dense_66 (Dense)                (None, 128)          94336       add_22[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "dense_67 (Dense)                (None, 32)           4128        dense_66[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "dense_68 (Dense)                (None, 2)            66          dense_67[0][0]                   \n",
            "==================================================================================================\n",
            "Total params: 184,906\n",
            "Trainable params: 184,906\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "3/3 [==============================] - 0s 18ms/step - loss: 0.1010 - accuracy: 0.9487\n",
            "Epoch 1/20\n",
            "11/11 [==============================] - 3s 188ms/step - loss: 0.7026 - accuracy: 0.5099 - val_loss: 0.6967 - val_accuracy: 0.4744\n",
            "Epoch 2/20\n",
            "11/11 [==============================] - 2s 155ms/step - loss: 0.6951 - accuracy: 0.5093 - val_loss: 0.6909 - val_accuracy: 0.5256\n",
            "Epoch 3/20\n",
            "11/11 [==============================] - 2s 156ms/step - loss: 0.6913 - accuracy: 0.5084 - val_loss: 0.6887 - val_accuracy: 0.4744\n",
            "Epoch 4/20\n",
            "11/11 [==============================] - 2s 152ms/step - loss: 0.6856 - accuracy: 0.5035 - val_loss: 0.6707 - val_accuracy: 0.5641\n",
            "Epoch 5/20\n",
            "11/11 [==============================] - 2s 157ms/step - loss: 0.6500 - accuracy: 0.6424 - val_loss: 0.5849 - val_accuracy: 0.8205\n",
            "Epoch 6/20\n",
            "11/11 [==============================] - 2s 158ms/step - loss: 0.5406 - accuracy: 0.8207 - val_loss: 0.4373 - val_accuracy: 0.8718\n",
            "Epoch 7/20\n",
            "11/11 [==============================] - 2s 152ms/step - loss: 0.3752 - accuracy: 0.8785 - val_loss: 0.3061 - val_accuracy: 0.8974\n",
            "Epoch 8/20\n",
            "11/11 [==============================] - 2s 154ms/step - loss: 0.2542 - accuracy: 0.9015 - val_loss: 0.2553 - val_accuracy: 0.8974\n",
            "Epoch 9/20\n",
            "11/11 [==============================] - 2s 157ms/step - loss: 0.1392 - accuracy: 0.9650 - val_loss: 0.2219 - val_accuracy: 0.8846\n",
            "Epoch 10/20\n",
            "11/11 [==============================] - 2s 155ms/step - loss: 0.0724 - accuracy: 0.9781 - val_loss: 0.2187 - val_accuracy: 0.8974\n",
            "Epoch 11/20\n",
            "11/11 [==============================] - 2s 158ms/step - loss: 0.0378 - accuracy: 0.9976 - val_loss: 0.1830 - val_accuracy: 0.9231\n",
            "Epoch 12/20\n",
            "11/11 [==============================] - 2s 155ms/step - loss: 0.0140 - accuracy: 1.0000 - val_loss: 0.1944 - val_accuracy: 0.8974\n",
            "Epoch 13/20\n",
            "11/11 [==============================] - 2s 155ms/step - loss: 0.0086 - accuracy: 1.0000 - val_loss: 0.1965 - val_accuracy: 0.8974\n",
            "Epoch 14/20\n",
            "11/11 [==============================] - 2s 156ms/step - loss: 0.0044 - accuracy: 1.0000 - val_loss: 0.1482 - val_accuracy: 0.9487\n",
            "Epoch 15/20\n",
            "11/11 [==============================] - 2s 158ms/step - loss: 0.0029 - accuracy: 1.0000 - val_loss: 0.1779 - val_accuracy: 0.9231\n",
            "Epoch 16/20\n",
            "11/11 [==============================] - 2s 152ms/step - loss: 0.0022 - accuracy: 1.0000 - val_loss: 0.1662 - val_accuracy: 0.9359\n",
            "Epoch 17/20\n",
            "11/11 [==============================] - 2s 156ms/step - loss: 0.0016 - accuracy: 1.0000 - val_loss: 0.1815 - val_accuracy: 0.9231\n",
            "Epoch 18/20\n",
            "11/11 [==============================] - 2s 155ms/step - loss: 0.0013 - accuracy: 1.0000 - val_loss: 0.1767 - val_accuracy: 0.9231\n",
            "Epoch 19/20\n",
            "11/11 [==============================] - 2s 150ms/step - loss: 0.0012 - accuracy: 1.0000 - val_loss: 0.1848 - val_accuracy: 0.9231\n",
            "Epoch 20/20\n",
            "11/11 [==============================] - 2s 158ms/step - loss: 9.6686e-04 - accuracy: 1.0000 - val_loss: 0.1842 - val_accuracy: 0.9231\n",
            "Model: \"model_23\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "input_47 (InputLayer)           [(None, 200)]        0                                            \n",
            "__________________________________________________________________________________________________\n",
            "input_48 (InputLayer)           [(None, 200)]        0                                            \n",
            "__________________________________________________________________________________________________\n",
            "embedding_46 (Embedding)        (None, 200, 10)      5300        input_47[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "embedding_47 (Embedding)        (None, 200, 10)      5300        input_48[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "reshape_46 (Reshape)            (None, 200, 10)      0           embedding_46[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "reshape_47 (Reshape)            (None, 200, 10)      0           embedding_47[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_184 (Conv1D)             (None, 198, 128)     3968        reshape_46[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_188 (Conv1D)             (None, 198, 128)     3968        reshape_47[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_185 (Conv1D)             (None, 196, 64)      24640       conv1d_184[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_189 (Conv1D)             (None, 196, 64)      24640       conv1d_188[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_138 (MaxPooling1D (None, 98, 64)       0           conv1d_185[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_141 (MaxPooling1D (None, 98, 64)       0           conv1d_189[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_186 (Conv1D)             (None, 96, 32)       6176        max_pooling1d_138[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_190 (Conv1D)             (None, 96, 32)       6176        max_pooling1d_141[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_139 (MaxPooling1D (None, 48, 32)       0           conv1d_186[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_142 (MaxPooling1D (None, 48, 32)       0           conv1d_190[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_187 (Conv1D)             (None, 46, 32)       3104        max_pooling1d_139[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_191 (Conv1D)             (None, 46, 32)       3104        max_pooling1d_142[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_140 (MaxPooling1D (None, 23, 32)       0           conv1d_187[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_143 (MaxPooling1D (None, 23, 32)       0           conv1d_191[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "flatten_46 (Flatten)            (None, 736)          0           max_pooling1d_140[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "flatten_47 (Flatten)            (None, 736)          0           max_pooling1d_143[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "add_23 (Add)                    (None, 736)          0           flatten_46[0][0]                 \n",
            "                                                                 flatten_47[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "dense_69 (Dense)                (None, 128)          94336       add_23[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "dense_70 (Dense)                (None, 32)           4128        dense_69[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "dense_71 (Dense)                (None, 2)            66          dense_70[0][0]                   \n",
            "==================================================================================================\n",
            "Total params: 184,906\n",
            "Trainable params: 184,906\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "3/3 [==============================] - 0s 13ms/step - loss: 0.1842 - accuracy: 0.9231\n",
            "Epoch 1/20\n",
            "11/11 [==============================] - 4s 227ms/step - loss: 0.7140 - accuracy: 0.4940 - val_loss: 0.6950 - val_accuracy: 0.5000\n",
            "Epoch 2/20\n",
            "11/11 [==============================] - 2s 158ms/step - loss: 0.6943 - accuracy: 0.4591 - val_loss: 0.6934 - val_accuracy: 0.5000\n",
            "Epoch 3/20\n",
            "11/11 [==============================] - 2s 156ms/step - loss: 0.6940 - accuracy: 0.4762 - val_loss: 0.6928 - val_accuracy: 0.5000\n",
            "Epoch 4/20\n",
            "11/11 [==============================] - 2s 159ms/step - loss: 0.6933 - accuracy: 0.4984 - val_loss: 0.6915 - val_accuracy: 0.5000\n",
            "Epoch 5/20\n",
            "11/11 [==============================] - 2s 152ms/step - loss: 0.6895 - accuracy: 0.5638 - val_loss: 0.6817 - val_accuracy: 0.7051\n",
            "Epoch 6/20\n",
            "11/11 [==============================] - 2s 151ms/step - loss: 0.6685 - accuracy: 0.7643 - val_loss: 0.6344 - val_accuracy: 0.6154\n",
            "Epoch 7/20\n",
            "11/11 [==============================] - 2s 155ms/step - loss: 0.5946 - accuracy: 0.7044 - val_loss: 0.5596 - val_accuracy: 0.7179\n",
            "Epoch 8/20\n",
            "11/11 [==============================] - 2s 158ms/step - loss: 0.5076 - accuracy: 0.7694 - val_loss: 0.5020 - val_accuracy: 0.7308\n",
            "Epoch 9/20\n",
            "11/11 [==============================] - 2s 159ms/step - loss: 0.3782 - accuracy: 0.8564 - val_loss: 0.3572 - val_accuracy: 0.8590\n",
            "Epoch 10/20\n",
            "11/11 [==============================] - 2s 157ms/step - loss: 0.3005 - accuracy: 0.8698 - val_loss: 0.2914 - val_accuracy: 0.8846\n",
            "Epoch 11/20\n",
            "11/11 [==============================] - 2s 161ms/step - loss: 0.2141 - accuracy: 0.9400 - val_loss: 0.2409 - val_accuracy: 0.8846\n",
            "Epoch 12/20\n",
            "11/11 [==============================] - 2s 160ms/step - loss: 0.1545 - accuracy: 0.9619 - val_loss: 0.1504 - val_accuracy: 0.9744\n",
            "Epoch 13/20\n",
            "11/11 [==============================] - 2s 160ms/step - loss: 0.0765 - accuracy: 0.9874 - val_loss: 0.1492 - val_accuracy: 0.9359\n",
            "Epoch 14/20\n",
            "11/11 [==============================] - 2s 160ms/step - loss: 0.0374 - accuracy: 0.9945 - val_loss: 0.1297 - val_accuracy: 0.9615\n",
            "Epoch 15/20\n",
            "11/11 [==============================] - 2s 154ms/step - loss: 0.0145 - accuracy: 1.0000 - val_loss: 0.1176 - val_accuracy: 0.9615\n",
            "Epoch 16/20\n",
            "11/11 [==============================] - 2s 155ms/step - loss: 0.0079 - accuracy: 1.0000 - val_loss: 0.1213 - val_accuracy: 0.9615\n",
            "Epoch 17/20\n",
            "11/11 [==============================] - 2s 157ms/step - loss: 0.0050 - accuracy: 1.0000 - val_loss: 0.1268 - val_accuracy: 0.9615\n",
            "Epoch 18/20\n",
            "11/11 [==============================] - 2s 158ms/step - loss: 0.0035 - accuracy: 1.0000 - val_loss: 0.1287 - val_accuracy: 0.9615\n",
            "Epoch 19/20\n",
            "11/11 [==============================] - 2s 161ms/step - loss: 0.0024 - accuracy: 1.0000 - val_loss: 0.1314 - val_accuracy: 0.9615\n",
            "Epoch 20/20\n",
            "11/11 [==============================] - 2s 157ms/step - loss: 0.0019 - accuracy: 1.0000 - val_loss: 0.1337 - val_accuracy: 0.9615\n",
            "Model: \"model_24\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "input_49 (InputLayer)           [(None, 200)]        0                                            \n",
            "__________________________________________________________________________________________________\n",
            "input_50 (InputLayer)           [(None, 200)]        0                                            \n",
            "__________________________________________________________________________________________________\n",
            "embedding_48 (Embedding)        (None, 200, 10)      5300        input_49[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "embedding_49 (Embedding)        (None, 200, 10)      5300        input_50[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "reshape_48 (Reshape)            (None, 200, 10)      0           embedding_48[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "reshape_49 (Reshape)            (None, 200, 10)      0           embedding_49[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_192 (Conv1D)             (None, 198, 128)     3968        reshape_48[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_196 (Conv1D)             (None, 198, 128)     3968        reshape_49[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_193 (Conv1D)             (None, 196, 64)      24640       conv1d_192[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_197 (Conv1D)             (None, 196, 64)      24640       conv1d_196[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_144 (MaxPooling1D (None, 98, 64)       0           conv1d_193[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_147 (MaxPooling1D (None, 98, 64)       0           conv1d_197[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_194 (Conv1D)             (None, 96, 32)       6176        max_pooling1d_144[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_198 (Conv1D)             (None, 96, 32)       6176        max_pooling1d_147[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_145 (MaxPooling1D (None, 48, 32)       0           conv1d_194[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_148 (MaxPooling1D (None, 48, 32)       0           conv1d_198[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_195 (Conv1D)             (None, 46, 32)       3104        max_pooling1d_145[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_199 (Conv1D)             (None, 46, 32)       3104        max_pooling1d_148[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_146 (MaxPooling1D (None, 23, 32)       0           conv1d_195[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_149 (MaxPooling1D (None, 23, 32)       0           conv1d_199[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "flatten_48 (Flatten)            (None, 736)          0           max_pooling1d_146[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "flatten_49 (Flatten)            (None, 736)          0           max_pooling1d_149[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "add_24 (Add)                    (None, 736)          0           flatten_48[0][0]                 \n",
            "                                                                 flatten_49[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "dense_72 (Dense)                (None, 128)          94336       add_24[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "dense_73 (Dense)                (None, 32)           4128        dense_72[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "dense_74 (Dense)                (None, 2)            66          dense_73[0][0]                   \n",
            "==================================================================================================\n",
            "Total params: 184,906\n",
            "Trainable params: 184,906\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "3/3 [==============================] - 0s 17ms/step - loss: 0.1337 - accuracy: 0.9615\n",
            "Epoch 1/20\n",
            "11/11 [==============================] - 3s 188ms/step - loss: 0.6959 - accuracy: 0.4860 - val_loss: 0.6917 - val_accuracy: 0.5641\n",
            "Epoch 2/20\n",
            "11/11 [==============================] - 2s 162ms/step - loss: 0.6938 - accuracy: 0.4855 - val_loss: 0.6919 - val_accuracy: 0.4359\n",
            "Epoch 3/20\n",
            "11/11 [==============================] - 2s 160ms/step - loss: 0.6897 - accuracy: 0.5319 - val_loss: 0.6846 - val_accuracy: 0.5256\n",
            "Epoch 4/20\n",
            "11/11 [==============================] - 2s 158ms/step - loss: 0.6745 - accuracy: 0.6901 - val_loss: 0.6399 - val_accuracy: 0.7821\n",
            "Epoch 5/20\n",
            "11/11 [==============================] - 2s 161ms/step - loss: 0.6085 - accuracy: 0.7541 - val_loss: 0.5551 - val_accuracy: 0.7051\n",
            "Epoch 6/20\n",
            "11/11 [==============================] - 2s 159ms/step - loss: 0.5145 - accuracy: 0.7965 - val_loss: 0.3890 - val_accuracy: 0.8333\n",
            "Epoch 7/20\n",
            "11/11 [==============================] - 2s 158ms/step - loss: 0.3690 - accuracy: 0.8629 - val_loss: 0.3528 - val_accuracy: 0.8077\n",
            "Epoch 8/20\n",
            "11/11 [==============================] - 2s 162ms/step - loss: 0.2429 - accuracy: 0.9203 - val_loss: 0.2806 - val_accuracy: 0.8974\n",
            "Epoch 9/20\n",
            "11/11 [==============================] - 2s 160ms/step - loss: 0.1350 - accuracy: 0.9677 - val_loss: 0.2602 - val_accuracy: 0.8974\n",
            "Epoch 10/20\n",
            "11/11 [==============================] - 2s 161ms/step - loss: 0.0820 - accuracy: 0.9841 - val_loss: 0.2360 - val_accuracy: 0.9103\n",
            "Epoch 11/20\n",
            "11/11 [==============================] - 2s 161ms/step - loss: 0.0318 - accuracy: 0.9983 - val_loss: 0.2290 - val_accuracy: 0.9103\n",
            "Epoch 12/20\n",
            "11/11 [==============================] - 2s 160ms/step - loss: 0.0223 - accuracy: 0.9983 - val_loss: 0.2202 - val_accuracy: 0.9103\n",
            "Epoch 13/20\n",
            "11/11 [==============================] - 2s 160ms/step - loss: 0.0128 - accuracy: 1.0000 - val_loss: 0.2244 - val_accuracy: 0.9103\n",
            "Epoch 14/20\n",
            "11/11 [==============================] - 2s 161ms/step - loss: 0.0079 - accuracy: 1.0000 - val_loss: 0.2397 - val_accuracy: 0.9103\n",
            "Epoch 15/20\n",
            "11/11 [==============================] - 2s 164ms/step - loss: 0.0047 - accuracy: 1.0000 - val_loss: 0.2371 - val_accuracy: 0.9103\n",
            "Epoch 16/20\n",
            "11/11 [==============================] - 2s 159ms/step - loss: 0.0035 - accuracy: 1.0000 - val_loss: 0.2425 - val_accuracy: 0.9103\n",
            "Epoch 17/20\n",
            "11/11 [==============================] - 2s 154ms/step - loss: 0.0030 - accuracy: 1.0000 - val_loss: 0.2466 - val_accuracy: 0.9103\n",
            "Epoch 18/20\n",
            "11/11 [==============================] - 2s 161ms/step - loss: 0.0024 - accuracy: 1.0000 - val_loss: 0.2517 - val_accuracy: 0.9103\n",
            "Epoch 19/20\n",
            "11/11 [==============================] - 2s 158ms/step - loss: 0.0022 - accuracy: 1.0000 - val_loss: 0.2576 - val_accuracy: 0.9103\n",
            "Epoch 20/20\n",
            "11/11 [==============================] - 2s 155ms/step - loss: 0.0019 - accuracy: 1.0000 - val_loss: 0.2621 - val_accuracy: 0.9103\n",
            "Model: \"model_25\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "input_51 (InputLayer)           [(None, 200)]        0                                            \n",
            "__________________________________________________________________________________________________\n",
            "input_52 (InputLayer)           [(None, 200)]        0                                            \n",
            "__________________________________________________________________________________________________\n",
            "embedding_50 (Embedding)        (None, 200, 10)      5300        input_51[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "embedding_51 (Embedding)        (None, 200, 10)      5300        input_52[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "reshape_50 (Reshape)            (None, 200, 10)      0           embedding_50[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "reshape_51 (Reshape)            (None, 200, 10)      0           embedding_51[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_200 (Conv1D)             (None, 198, 128)     3968        reshape_50[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_204 (Conv1D)             (None, 198, 128)     3968        reshape_51[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_201 (Conv1D)             (None, 196, 64)      24640       conv1d_200[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_205 (Conv1D)             (None, 196, 64)      24640       conv1d_204[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_150 (MaxPooling1D (None, 98, 64)       0           conv1d_201[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_153 (MaxPooling1D (None, 98, 64)       0           conv1d_205[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_202 (Conv1D)             (None, 96, 32)       6176        max_pooling1d_150[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_206 (Conv1D)             (None, 96, 32)       6176        max_pooling1d_153[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_151 (MaxPooling1D (None, 48, 32)       0           conv1d_202[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_154 (MaxPooling1D (None, 48, 32)       0           conv1d_206[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_203 (Conv1D)             (None, 46, 32)       3104        max_pooling1d_151[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_207 (Conv1D)             (None, 46, 32)       3104        max_pooling1d_154[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_152 (MaxPooling1D (None, 23, 32)       0           conv1d_203[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_155 (MaxPooling1D (None, 23, 32)       0           conv1d_207[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "flatten_50 (Flatten)            (None, 736)          0           max_pooling1d_152[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "flatten_51 (Flatten)            (None, 736)          0           max_pooling1d_155[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "add_25 (Add)                    (None, 736)          0           flatten_50[0][0]                 \n",
            "                                                                 flatten_51[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "dense_75 (Dense)                (None, 128)          94336       add_25[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "dense_76 (Dense)                (None, 32)           4128        dense_75[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "dense_77 (Dense)                (None, 2)            66          dense_76[0][0]                   \n",
            "==================================================================================================\n",
            "Total params: 184,906\n",
            "Trainable params: 184,906\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "3/3 [==============================] - 0s 17ms/step - loss: 0.2621 - accuracy: 0.9103\n",
            "Epoch 1/20\n",
            "11/11 [==============================] - 3s 191ms/step - loss: 0.7280 - accuracy: 0.5058 - val_loss: 0.7008 - val_accuracy: 0.4545\n",
            "Epoch 2/20\n",
            "11/11 [==============================] - 2s 161ms/step - loss: 0.6948 - accuracy: 0.5124 - val_loss: 0.6953 - val_accuracy: 0.4545\n",
            "Epoch 3/20\n",
            "11/11 [==============================] - 2s 156ms/step - loss: 0.6971 - accuracy: 0.5083 - val_loss: 0.6936 - val_accuracy: 0.6494\n",
            "Epoch 4/20\n",
            "11/11 [==============================] - 2s 159ms/step - loss: 0.6964 - accuracy: 0.5661 - val_loss: 0.6952 - val_accuracy: 0.4545\n",
            "Epoch 5/20\n",
            "11/11 [==============================] - 2s 159ms/step - loss: 0.6909 - accuracy: 0.5533 - val_loss: 0.6865 - val_accuracy: 0.5455\n",
            "Epoch 6/20\n",
            "11/11 [==============================] - 2s 160ms/step - loss: 0.6863 - accuracy: 0.5329 - val_loss: 0.6982 - val_accuracy: 0.5065\n",
            "Epoch 7/20\n",
            "11/11 [==============================] - 2s 162ms/step - loss: 0.6637 - accuracy: 0.6151 - val_loss: 0.6597 - val_accuracy: 0.6234\n",
            "Epoch 8/20\n",
            "11/11 [==============================] - 2s 155ms/step - loss: 0.5787 - accuracy: 0.7374 - val_loss: 0.5657 - val_accuracy: 0.8182\n",
            "Epoch 9/20\n",
            "11/11 [==============================] - 2s 153ms/step - loss: 0.4287 - accuracy: 0.8823 - val_loss: 0.4709 - val_accuracy: 0.8442\n",
            "Epoch 10/20\n",
            "11/11 [==============================] - 2s 160ms/step - loss: 0.2639 - accuracy: 0.9152 - val_loss: 0.4179 - val_accuracy: 0.8831\n",
            "Epoch 11/20\n",
            "11/11 [==============================] - 2s 161ms/step - loss: 0.1620 - accuracy: 0.9494 - val_loss: 0.3355 - val_accuracy: 0.8831\n",
            "Epoch 12/20\n",
            "11/11 [==============================] - 2s 159ms/step - loss: 0.0810 - accuracy: 0.9775 - val_loss: 0.3421 - val_accuracy: 0.8831\n",
            "Epoch 13/20\n",
            "11/11 [==============================] - 2s 160ms/step - loss: 0.0302 - accuracy: 0.9983 - val_loss: 0.2942 - val_accuracy: 0.9221\n",
            "Epoch 14/20\n",
            "11/11 [==============================] - 2s 162ms/step - loss: 0.0210 - accuracy: 0.9995 - val_loss: 0.2677 - val_accuracy: 0.9091\n",
            "Epoch 15/20\n",
            "11/11 [==============================] - 2s 163ms/step - loss: 0.0143 - accuracy: 0.9987 - val_loss: 0.2819 - val_accuracy: 0.9351\n",
            "Epoch 16/20\n",
            "11/11 [==============================] - 2s 163ms/step - loss: 0.0065 - accuracy: 1.0000 - val_loss: 0.2754 - val_accuracy: 0.9351\n",
            "Epoch 17/20\n",
            "11/11 [==============================] - 2s 159ms/step - loss: 0.0044 - accuracy: 1.0000 - val_loss: 0.2535 - val_accuracy: 0.9481\n",
            "Epoch 18/20\n",
            "11/11 [==============================] - 2s 160ms/step - loss: 0.0023 - accuracy: 1.0000 - val_loss: 0.2789 - val_accuracy: 0.9481\n",
            "Epoch 19/20\n",
            "11/11 [==============================] - 2s 155ms/step - loss: 0.0018 - accuracy: 1.0000 - val_loss: 0.2740 - val_accuracy: 0.9481\n",
            "Epoch 20/20\n",
            "11/11 [==============================] - 2s 158ms/step - loss: 0.0015 - accuracy: 1.0000 - val_loss: 0.2833 - val_accuracy: 0.9481\n",
            "Model: \"model_26\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "input_53 (InputLayer)           [(None, 200)]        0                                            \n",
            "__________________________________________________________________________________________________\n",
            "input_54 (InputLayer)           [(None, 200)]        0                                            \n",
            "__________________________________________________________________________________________________\n",
            "embedding_52 (Embedding)        (None, 200, 10)      5300        input_53[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "embedding_53 (Embedding)        (None, 200, 10)      5300        input_54[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "reshape_52 (Reshape)            (None, 200, 10)      0           embedding_52[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "reshape_53 (Reshape)            (None, 200, 10)      0           embedding_53[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_208 (Conv1D)             (None, 198, 128)     3968        reshape_52[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_212 (Conv1D)             (None, 198, 128)     3968        reshape_53[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_209 (Conv1D)             (None, 196, 64)      24640       conv1d_208[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_213 (Conv1D)             (None, 196, 64)      24640       conv1d_212[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_156 (MaxPooling1D (None, 98, 64)       0           conv1d_209[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_159 (MaxPooling1D (None, 98, 64)       0           conv1d_213[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_210 (Conv1D)             (None, 96, 32)       6176        max_pooling1d_156[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_214 (Conv1D)             (None, 96, 32)       6176        max_pooling1d_159[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_157 (MaxPooling1D (None, 48, 32)       0           conv1d_210[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_160 (MaxPooling1D (None, 48, 32)       0           conv1d_214[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_211 (Conv1D)             (None, 46, 32)       3104        max_pooling1d_157[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_215 (Conv1D)             (None, 46, 32)       3104        max_pooling1d_160[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_158 (MaxPooling1D (None, 23, 32)       0           conv1d_211[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_161 (MaxPooling1D (None, 23, 32)       0           conv1d_215[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "flatten_52 (Flatten)            (None, 736)          0           max_pooling1d_158[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "flatten_53 (Flatten)            (None, 736)          0           max_pooling1d_161[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "add_26 (Add)                    (None, 736)          0           flatten_52[0][0]                 \n",
            "                                                                 flatten_53[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "dense_78 (Dense)                (None, 128)          94336       add_26[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "dense_79 (Dense)                (None, 32)           4128        dense_78[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "dense_80 (Dense)                (None, 2)            66          dense_79[0][0]                   \n",
            "==================================================================================================\n",
            "Total params: 184,906\n",
            "Trainable params: 184,906\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "3/3 [==============================] - 0s 20ms/step - loss: 0.2833 - accuracy: 0.9481\n",
            "Epoch 1/20\n",
            "11/11 [==============================] - 3s 194ms/step - loss: 0.6995 - accuracy: 0.4498 - val_loss: 0.6931 - val_accuracy: 0.5714\n",
            "Epoch 2/20\n",
            "11/11 [==============================] - 2s 163ms/step - loss: 0.6936 - accuracy: 0.5206 - val_loss: 0.7002 - val_accuracy: 0.4286\n",
            "Epoch 3/20\n",
            "11/11 [==============================] - 2s 162ms/step - loss: 0.6929 - accuracy: 0.5564 - val_loss: 0.6830 - val_accuracy: 0.5714\n",
            "Epoch 4/20\n",
            "11/11 [==============================] - 2s 156ms/step - loss: 0.6880 - accuracy: 0.5431 - val_loss: 0.6813 - val_accuracy: 0.4286\n",
            "Epoch 5/20\n",
            "11/11 [==============================] - 2s 159ms/step - loss: 0.6627 - accuracy: 0.5745 - val_loss: 0.6164 - val_accuracy: 0.6364\n",
            "Epoch 6/20\n",
            "11/11 [==============================] - 2s 162ms/step - loss: 0.5866 - accuracy: 0.7381 - val_loss: 0.4932 - val_accuracy: 0.8571\n",
            "Epoch 7/20\n",
            "11/11 [==============================] - 2s 161ms/step - loss: 0.4301 - accuracy: 0.8495 - val_loss: 0.3612 - val_accuracy: 0.8831\n",
            "Epoch 8/20\n",
            "11/11 [==============================] - 2s 160ms/step - loss: 0.2849 - accuracy: 0.8965 - val_loss: 0.2420 - val_accuracy: 0.9351\n",
            "Epoch 9/20\n",
            "11/11 [==============================] - 2s 163ms/step - loss: 0.2132 - accuracy: 0.9256 - val_loss: 0.1539 - val_accuracy: 0.9481\n",
            "Epoch 10/20\n",
            "11/11 [==============================] - 2s 156ms/step - loss: 0.1321 - accuracy: 0.9551 - val_loss: 0.0945 - val_accuracy: 1.0000\n",
            "Epoch 11/20\n",
            "11/11 [==============================] - 2s 164ms/step - loss: 0.0621 - accuracy: 0.9911 - val_loss: 0.0704 - val_accuracy: 0.9870\n",
            "Epoch 12/20\n",
            "11/11 [==============================] - 2s 167ms/step - loss: 0.0290 - accuracy: 1.0000 - val_loss: 0.0675 - val_accuracy: 1.0000\n",
            "Epoch 13/20\n",
            "11/11 [==============================] - 2s 163ms/step - loss: 0.0157 - accuracy: 1.0000 - val_loss: 0.0622 - val_accuracy: 1.0000\n",
            "Epoch 14/20\n",
            "11/11 [==============================] - 2s 161ms/step - loss: 0.0089 - accuracy: 1.0000 - val_loss: 0.0291 - val_accuracy: 1.0000\n",
            "Epoch 15/20\n",
            "11/11 [==============================] - 2s 162ms/step - loss: 0.0059 - accuracy: 1.0000 - val_loss: 0.0393 - val_accuracy: 1.0000\n",
            "Epoch 16/20\n",
            "11/11 [==============================] - 2s 159ms/step - loss: 0.0038 - accuracy: 1.0000 - val_loss: 0.0244 - val_accuracy: 1.0000\n",
            "Epoch 17/20\n",
            "11/11 [==============================] - 2s 147ms/step - loss: 0.0029 - accuracy: 1.0000 - val_loss: 0.0302 - val_accuracy: 1.0000\n",
            "Epoch 18/20\n",
            "11/11 [==============================] - 2s 163ms/step - loss: 0.0024 - accuracy: 1.0000 - val_loss: 0.0262 - val_accuracy: 1.0000\n",
            "Epoch 19/20\n",
            "11/11 [==============================] - 2s 161ms/step - loss: 0.0019 - accuracy: 1.0000 - val_loss: 0.0250 - val_accuracy: 1.0000\n",
            "Epoch 20/20\n",
            "11/11 [==============================] - 2s 164ms/step - loss: 0.0016 - accuracy: 1.0000 - val_loss: 0.0234 - val_accuracy: 1.0000\n",
            "Model: \"model_27\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "input_55 (InputLayer)           [(None, 200)]        0                                            \n",
            "__________________________________________________________________________________________________\n",
            "input_56 (InputLayer)           [(None, 200)]        0                                            \n",
            "__________________________________________________________________________________________________\n",
            "embedding_54 (Embedding)        (None, 200, 10)      5300        input_55[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "embedding_55 (Embedding)        (None, 200, 10)      5300        input_56[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "reshape_54 (Reshape)            (None, 200, 10)      0           embedding_54[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "reshape_55 (Reshape)            (None, 200, 10)      0           embedding_55[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_216 (Conv1D)             (None, 198, 128)     3968        reshape_54[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_220 (Conv1D)             (None, 198, 128)     3968        reshape_55[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_217 (Conv1D)             (None, 196, 64)      24640       conv1d_216[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_221 (Conv1D)             (None, 196, 64)      24640       conv1d_220[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_162 (MaxPooling1D (None, 98, 64)       0           conv1d_217[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_165 (MaxPooling1D (None, 98, 64)       0           conv1d_221[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_218 (Conv1D)             (None, 96, 32)       6176        max_pooling1d_162[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_222 (Conv1D)             (None, 96, 32)       6176        max_pooling1d_165[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_163 (MaxPooling1D (None, 48, 32)       0           conv1d_218[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_166 (MaxPooling1D (None, 48, 32)       0           conv1d_222[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_219 (Conv1D)             (None, 46, 32)       3104        max_pooling1d_163[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_223 (Conv1D)             (None, 46, 32)       3104        max_pooling1d_166[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_164 (MaxPooling1D (None, 23, 32)       0           conv1d_219[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_167 (MaxPooling1D (None, 23, 32)       0           conv1d_223[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "flatten_54 (Flatten)            (None, 736)          0           max_pooling1d_164[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "flatten_55 (Flatten)            (None, 736)          0           max_pooling1d_167[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "add_27 (Add)                    (None, 736)          0           flatten_54[0][0]                 \n",
            "                                                                 flatten_55[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "dense_81 (Dense)                (None, 128)          94336       add_27[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "dense_82 (Dense)                (None, 32)           4128        dense_81[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "dense_83 (Dense)                (None, 2)            66          dense_82[0][0]                   \n",
            "==================================================================================================\n",
            "Total params: 184,906\n",
            "Trainable params: 184,906\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "3/3 [==============================] - 0s 18ms/step - loss: 0.0234 - accuracy: 1.0000\n",
            "Epoch 1/20\n",
            "11/11 [==============================] - 4s 189ms/step - loss: 0.7360 - accuracy: 0.5053 - val_loss: 0.6937 - val_accuracy: 0.6104\n",
            "Epoch 2/20\n",
            "11/11 [==============================] - 2s 163ms/step - loss: 0.6960 - accuracy: 0.5301 - val_loss: 0.7030 - val_accuracy: 0.3896\n",
            "Epoch 3/20\n",
            "11/11 [==============================] - 2s 159ms/step - loss: 0.6933 - accuracy: 0.5161 - val_loss: 0.6972 - val_accuracy: 0.3896\n",
            "Epoch 4/20\n",
            "11/11 [==============================] - 2s 160ms/step - loss: 0.6919 - accuracy: 0.5255 - val_loss: 0.6999 - val_accuracy: 0.3896\n",
            "Epoch 5/20\n",
            "11/11 [==============================] - 2s 160ms/step - loss: 0.6934 - accuracy: 0.4969 - val_loss: 0.6950 - val_accuracy: 0.3896\n",
            "Epoch 6/20\n",
            "11/11 [==============================] - 2s 162ms/step - loss: 0.6903 - accuracy: 0.5023 - val_loss: 0.7022 - val_accuracy: 0.3896\n",
            "Epoch 7/20\n",
            "11/11 [==============================] - 2s 154ms/step - loss: 0.6796 - accuracy: 0.5546 - val_loss: 0.6700 - val_accuracy: 0.8442\n",
            "Epoch 8/20\n",
            "11/11 [==============================] - 2s 164ms/step - loss: 0.6460 - accuracy: 0.8312 - val_loss: 0.6280 - val_accuracy: 0.7143\n",
            "Epoch 9/20\n",
            "11/11 [==============================] - 2s 154ms/step - loss: 0.5344 - accuracy: 0.8356 - val_loss: 0.5811 - val_accuracy: 0.6623\n",
            "Epoch 10/20\n",
            "11/11 [==============================] - 2s 163ms/step - loss: 0.3944 - accuracy: 0.8569 - val_loss: 0.4336 - val_accuracy: 0.7922\n",
            "Epoch 11/20\n",
            "11/11 [==============================] - 2s 165ms/step - loss: 0.2514 - accuracy: 0.9158 - val_loss: 0.3051 - val_accuracy: 0.8701\n",
            "Epoch 12/20\n",
            "11/11 [==============================] - 2s 164ms/step - loss: 0.1532 - accuracy: 0.9465 - val_loss: 0.2706 - val_accuracy: 0.8831\n",
            "Epoch 13/20\n",
            "11/11 [==============================] - 2s 165ms/step - loss: 0.0760 - accuracy: 0.9879 - val_loss: 0.1680 - val_accuracy: 0.9221\n",
            "Epoch 14/20\n",
            "11/11 [==============================] - 2s 164ms/step - loss: 0.0358 - accuracy: 1.0000 - val_loss: 0.1307 - val_accuracy: 0.9481\n",
            "Epoch 15/20\n",
            "11/11 [==============================] - 2s 161ms/step - loss: 0.0181 - accuracy: 1.0000 - val_loss: 0.1338 - val_accuracy: 0.9481\n",
            "Epoch 16/20\n",
            "11/11 [==============================] - 2s 151ms/step - loss: 0.0093 - accuracy: 1.0000 - val_loss: 0.1797 - val_accuracy: 0.9351\n",
            "Epoch 17/20\n",
            "11/11 [==============================] - 2s 162ms/step - loss: 0.0065 - accuracy: 1.0000 - val_loss: 0.1156 - val_accuracy: 0.9481\n",
            "Epoch 18/20\n",
            "11/11 [==============================] - 2s 160ms/step - loss: 0.0049 - accuracy: 1.0000 - val_loss: 0.1410 - val_accuracy: 0.9481\n",
            "Epoch 19/20\n",
            "11/11 [==============================] - 2s 160ms/step - loss: 0.0039 - accuracy: 1.0000 - val_loss: 0.1216 - val_accuracy: 0.9481\n",
            "Epoch 20/20\n",
            "11/11 [==============================] - 2s 164ms/step - loss: 0.0033 - accuracy: 1.0000 - val_loss: 0.1359 - val_accuracy: 0.9481\n",
            "Model: \"model_28\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "input_57 (InputLayer)           [(None, 200)]        0                                            \n",
            "__________________________________________________________________________________________________\n",
            "input_58 (InputLayer)           [(None, 200)]        0                                            \n",
            "__________________________________________________________________________________________________\n",
            "embedding_56 (Embedding)        (None, 200, 10)      5300        input_57[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "embedding_57 (Embedding)        (None, 200, 10)      5300        input_58[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "reshape_56 (Reshape)            (None, 200, 10)      0           embedding_56[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "reshape_57 (Reshape)            (None, 200, 10)      0           embedding_57[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_224 (Conv1D)             (None, 198, 128)     3968        reshape_56[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_228 (Conv1D)             (None, 198, 128)     3968        reshape_57[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_225 (Conv1D)             (None, 196, 64)      24640       conv1d_224[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_229 (Conv1D)             (None, 196, 64)      24640       conv1d_228[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_168 (MaxPooling1D (None, 98, 64)       0           conv1d_225[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_171 (MaxPooling1D (None, 98, 64)       0           conv1d_229[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_226 (Conv1D)             (None, 96, 32)       6176        max_pooling1d_168[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_230 (Conv1D)             (None, 96, 32)       6176        max_pooling1d_171[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_169 (MaxPooling1D (None, 48, 32)       0           conv1d_226[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_172 (MaxPooling1D (None, 48, 32)       0           conv1d_230[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_227 (Conv1D)             (None, 46, 32)       3104        max_pooling1d_169[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_231 (Conv1D)             (None, 46, 32)       3104        max_pooling1d_172[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_170 (MaxPooling1D (None, 23, 32)       0           conv1d_227[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_173 (MaxPooling1D (None, 23, 32)       0           conv1d_231[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "flatten_56 (Flatten)            (None, 736)          0           max_pooling1d_170[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "flatten_57 (Flatten)            (None, 736)          0           max_pooling1d_173[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "add_28 (Add)                    (None, 736)          0           flatten_56[0][0]                 \n",
            "                                                                 flatten_57[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "dense_84 (Dense)                (None, 128)          94336       add_28[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "dense_85 (Dense)                (None, 32)           4128        dense_84[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "dense_86 (Dense)                (None, 2)            66          dense_85[0][0]                   \n",
            "==================================================================================================\n",
            "Total params: 184,906\n",
            "Trainable params: 184,906\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "3/3 [==============================] - 0s 19ms/step - loss: 0.1359 - accuracy: 0.9481\n",
            "Epoch 1/20\n",
            "11/11 [==============================] - 3s 185ms/step - loss: 0.7003 - accuracy: 0.4880 - val_loss: 0.6915 - val_accuracy: 0.5714\n",
            "Epoch 2/20\n",
            "11/11 [==============================] - 2s 165ms/step - loss: 0.6945 - accuracy: 0.4849 - val_loss: 0.7096 - val_accuracy: 0.4286\n",
            "Epoch 3/20\n",
            "11/11 [==============================] - 2s 161ms/step - loss: 0.6950 - accuracy: 0.5109 - val_loss: 0.6902 - val_accuracy: 0.5714\n",
            "Epoch 4/20\n",
            "11/11 [==============================] - 2s 165ms/step - loss: 0.6931 - accuracy: 0.5057 - val_loss: 0.6942 - val_accuracy: 0.4286\n",
            "Epoch 5/20\n",
            "11/11 [==============================] - 2s 162ms/step - loss: 0.6909 - accuracy: 0.5451 - val_loss: 0.6883 - val_accuracy: 0.7403\n",
            "Epoch 6/20\n",
            "11/11 [==============================] - 2s 161ms/step - loss: 0.6873 - accuracy: 0.5902 - val_loss: 0.6723 - val_accuracy: 0.6883\n",
            "Epoch 7/20\n",
            "11/11 [==============================] - 2s 160ms/step - loss: 0.6536 - accuracy: 0.8195 - val_loss: 0.6059 - val_accuracy: 0.6104\n",
            "Epoch 8/20\n",
            "11/11 [==============================] - 2s 160ms/step - loss: 0.5286 - accuracy: 0.8070 - val_loss: 0.3660 - val_accuracy: 0.9091\n",
            "Epoch 9/20\n",
            "11/11 [==============================] - 2s 162ms/step - loss: 0.3364 - accuracy: 0.8912 - val_loss: 0.2033 - val_accuracy: 0.9481\n",
            "Epoch 10/20\n",
            "11/11 [==============================] - 2s 163ms/step - loss: 0.2081 - accuracy: 0.9270 - val_loss: 0.1426 - val_accuracy: 0.9610\n",
            "Epoch 11/20\n",
            "11/11 [==============================] - 2s 154ms/step - loss: 0.1104 - accuracy: 0.9666 - val_loss: 0.1160 - val_accuracy: 0.9740\n",
            "Epoch 12/20\n",
            "11/11 [==============================] - 2s 156ms/step - loss: 0.0671 - accuracy: 0.9859 - val_loss: 0.0989 - val_accuracy: 0.9610\n",
            "Epoch 13/20\n",
            "11/11 [==============================] - 2s 163ms/step - loss: 0.0498 - accuracy: 0.9925 - val_loss: 0.0766 - val_accuracy: 0.9740\n",
            "Epoch 14/20\n",
            "11/11 [==============================] - 2s 153ms/step - loss: 0.0258 - accuracy: 0.9959 - val_loss: 0.1288 - val_accuracy: 0.9610\n",
            "Epoch 15/20\n",
            "11/11 [==============================] - 2s 163ms/step - loss: 0.0160 - accuracy: 0.9989 - val_loss: 0.0766 - val_accuracy: 0.9740\n",
            "Epoch 16/20\n",
            "11/11 [==============================] - 2s 164ms/step - loss: 0.0093 - accuracy: 0.9991 - val_loss: 0.1094 - val_accuracy: 0.9481\n",
            "Epoch 17/20\n",
            "11/11 [==============================] - 2s 157ms/step - loss: 0.0062 - accuracy: 1.0000 - val_loss: 0.0787 - val_accuracy: 0.9740\n",
            "Epoch 18/20\n",
            "11/11 [==============================] - 2s 167ms/step - loss: 0.0037 - accuracy: 1.0000 - val_loss: 0.0785 - val_accuracy: 0.9740\n",
            "Epoch 19/20\n",
            "11/11 [==============================] - 2s 158ms/step - loss: 0.0027 - accuracy: 1.0000 - val_loss: 0.0771 - val_accuracy: 0.9740\n",
            "Epoch 20/20\n",
            "11/11 [==============================] - 2s 163ms/step - loss: 0.0021 - accuracy: 1.0000 - val_loss: 0.0787 - val_accuracy: 0.9740\n",
            "Model: \"model_29\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "input_59 (InputLayer)           [(None, 200)]        0                                            \n",
            "__________________________________________________________________________________________________\n",
            "input_60 (InputLayer)           [(None, 200)]        0                                            \n",
            "__________________________________________________________________________________________________\n",
            "embedding_58 (Embedding)        (None, 200, 10)      5300        input_59[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "embedding_59 (Embedding)        (None, 200, 10)      5300        input_60[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "reshape_58 (Reshape)            (None, 200, 10)      0           embedding_58[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "reshape_59 (Reshape)            (None, 200, 10)      0           embedding_59[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_232 (Conv1D)             (None, 198, 128)     3968        reshape_58[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_236 (Conv1D)             (None, 198, 128)     3968        reshape_59[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_233 (Conv1D)             (None, 196, 64)      24640       conv1d_232[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_237 (Conv1D)             (None, 196, 64)      24640       conv1d_236[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_174 (MaxPooling1D (None, 98, 64)       0           conv1d_233[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_177 (MaxPooling1D (None, 98, 64)       0           conv1d_237[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_234 (Conv1D)             (None, 96, 32)       6176        max_pooling1d_174[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_238 (Conv1D)             (None, 96, 32)       6176        max_pooling1d_177[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_175 (MaxPooling1D (None, 48, 32)       0           conv1d_234[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_178 (MaxPooling1D (None, 48, 32)       0           conv1d_238[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_235 (Conv1D)             (None, 46, 32)       3104        max_pooling1d_175[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_239 (Conv1D)             (None, 46, 32)       3104        max_pooling1d_178[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_176 (MaxPooling1D (None, 23, 32)       0           conv1d_235[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_179 (MaxPooling1D (None, 23, 32)       0           conv1d_239[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "flatten_58 (Flatten)            (None, 736)          0           max_pooling1d_176[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "flatten_59 (Flatten)            (None, 736)          0           max_pooling1d_179[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "add_29 (Add)                    (None, 736)          0           flatten_58[0][0]                 \n",
            "                                                                 flatten_59[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "dense_87 (Dense)                (None, 128)          94336       add_29[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "dense_88 (Dense)                (None, 32)           4128        dense_87[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "dense_89 (Dense)                (None, 2)            66          dense_88[0][0]                   \n",
            "==================================================================================================\n",
            "Total params: 184,906\n",
            "Trainable params: 184,906\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "3/3 [==============================] - 0s 18ms/step - loss: 0.0787 - accuracy: 0.9740\n",
            "Epoch 1/20\n",
            "11/11 [==============================] - 3s 191ms/step - loss: 0.6947 - accuracy: 0.5167 - val_loss: 0.6944 - val_accuracy: 0.5065\n",
            "Epoch 2/20\n",
            "11/11 [==============================] - 2s 167ms/step - loss: 0.6950 - accuracy: 0.5026 - val_loss: 0.6925 - val_accuracy: 0.4935\n",
            "Epoch 3/20\n",
            "11/11 [==============================] - 2s 164ms/step - loss: 0.6924 - accuracy: 0.5039 - val_loss: 0.6886 - val_accuracy: 0.6364\n",
            "Epoch 4/20\n",
            "11/11 [==============================] - 2s 165ms/step - loss: 0.6888 - accuracy: 0.4968 - val_loss: 0.6787 - val_accuracy: 0.5195\n",
            "Epoch 5/20\n",
            "11/11 [==============================] - 2s 164ms/step - loss: 0.6678 - accuracy: 0.6495 - val_loss: 0.6403 - val_accuracy: 0.7143\n",
            "Epoch 6/20\n",
            "11/11 [==============================] - 2s 159ms/step - loss: 0.5989 - accuracy: 0.7886 - val_loss: 0.5553 - val_accuracy: 0.7273\n",
            "Epoch 7/20\n",
            "11/11 [==============================] - 2s 164ms/step - loss: 0.4719 - accuracy: 0.8126 - val_loss: 0.4288 - val_accuracy: 0.8571\n",
            "Epoch 8/20\n",
            "11/11 [==============================] - 2s 159ms/step - loss: 0.3366 - accuracy: 0.8797 - val_loss: 0.2952 - val_accuracy: 0.8961\n",
            "Epoch 9/20\n",
            "11/11 [==============================] - 2s 162ms/step - loss: 0.1985 - accuracy: 0.9407 - val_loss: 0.2017 - val_accuracy: 0.9221\n",
            "Epoch 10/20\n",
            "11/11 [==============================] - 2s 165ms/step - loss: 0.1304 - accuracy: 0.9722 - val_loss: 0.1302 - val_accuracy: 0.9610\n",
            "Epoch 11/20\n",
            "11/11 [==============================] - 2s 166ms/step - loss: 0.0556 - accuracy: 0.9891 - val_loss: 0.1903 - val_accuracy: 0.9091\n",
            "Epoch 12/20\n",
            "11/11 [==============================] - 2s 166ms/step - loss: 0.0480 - accuracy: 0.9901 - val_loss: 0.0996 - val_accuracy: 0.9740\n",
            "Epoch 13/20\n",
            "11/11 [==============================] - 2s 164ms/step - loss: 0.0181 - accuracy: 1.0000 - val_loss: 0.1134 - val_accuracy: 0.9610\n",
            "Epoch 14/20\n",
            "11/11 [==============================] - 2s 165ms/step - loss: 0.0106 - accuracy: 1.0000 - val_loss: 0.0632 - val_accuracy: 0.9870\n",
            "Epoch 15/20\n",
            "11/11 [==============================] - 2s 168ms/step - loss: 0.0059 - accuracy: 1.0000 - val_loss: 0.0947 - val_accuracy: 0.9740\n",
            "Epoch 16/20\n",
            "11/11 [==============================] - 2s 163ms/step - loss: 0.0046 - accuracy: 1.0000 - val_loss: 0.0804 - val_accuracy: 0.9740\n",
            "Epoch 17/20\n",
            "11/11 [==============================] - 2s 164ms/step - loss: 0.0035 - accuracy: 1.0000 - val_loss: 0.0929 - val_accuracy: 0.9740\n",
            "Epoch 18/20\n",
            "11/11 [==============================] - 2s 161ms/step - loss: 0.0028 - accuracy: 1.0000 - val_loss: 0.0919 - val_accuracy: 0.9740\n",
            "Epoch 19/20\n",
            "11/11 [==============================] - 2s 159ms/step - loss: 0.0025 - accuracy: 1.0000 - val_loss: 0.0956 - val_accuracy: 0.9740\n",
            "Epoch 20/20\n",
            "11/11 [==============================] - 2s 163ms/step - loss: 0.0022 - accuracy: 1.0000 - val_loss: 0.0949 - val_accuracy: 0.9740\n",
            "Model: \"model_30\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "input_61 (InputLayer)           [(None, 200)]        0                                            \n",
            "__________________________________________________________________________________________________\n",
            "input_62 (InputLayer)           [(None, 200)]        0                                            \n",
            "__________________________________________________________________________________________________\n",
            "embedding_60 (Embedding)        (None, 200, 10)      5300        input_61[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "embedding_61 (Embedding)        (None, 200, 10)      5300        input_62[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "reshape_60 (Reshape)            (None, 200, 10)      0           embedding_60[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "reshape_61 (Reshape)            (None, 200, 10)      0           embedding_61[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_240 (Conv1D)             (None, 198, 128)     3968        reshape_60[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_244 (Conv1D)             (None, 198, 128)     3968        reshape_61[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_241 (Conv1D)             (None, 196, 64)      24640       conv1d_240[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_245 (Conv1D)             (None, 196, 64)      24640       conv1d_244[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_180 (MaxPooling1D (None, 98, 64)       0           conv1d_241[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_183 (MaxPooling1D (None, 98, 64)       0           conv1d_245[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_242 (Conv1D)             (None, 96, 32)       6176        max_pooling1d_180[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_246 (Conv1D)             (None, 96, 32)       6176        max_pooling1d_183[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_181 (MaxPooling1D (None, 48, 32)       0           conv1d_242[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_184 (MaxPooling1D (None, 48, 32)       0           conv1d_246[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_243 (Conv1D)             (None, 46, 32)       3104        max_pooling1d_181[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_247 (Conv1D)             (None, 46, 32)       3104        max_pooling1d_184[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_182 (MaxPooling1D (None, 23, 32)       0           conv1d_243[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_185 (MaxPooling1D (None, 23, 32)       0           conv1d_247[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "flatten_60 (Flatten)            (None, 736)          0           max_pooling1d_182[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "flatten_61 (Flatten)            (None, 736)          0           max_pooling1d_185[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "add_30 (Add)                    (None, 736)          0           flatten_60[0][0]                 \n",
            "                                                                 flatten_61[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "dense_90 (Dense)                (None, 128)          94336       add_30[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "dense_91 (Dense)                (None, 32)           4128        dense_90[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "dense_92 (Dense)                (None, 2)            66          dense_91[0][0]                   \n",
            "==================================================================================================\n",
            "Total params: 184,906\n",
            "Trainable params: 184,906\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "3/3 [==============================] - 0s 19ms/step - loss: 0.0949 - accuracy: 0.9740\n",
            "Epoch 1/20\n",
            "11/11 [==============================] - 3s 197ms/step - loss: 0.7022 - accuracy: 0.4874 - val_loss: 0.7002 - val_accuracy: 0.4416\n",
            "Epoch 2/20\n",
            "11/11 [==============================] - 2s 165ms/step - loss: 0.6940 - accuracy: 0.5071 - val_loss: 0.6918 - val_accuracy: 0.5584\n",
            "Epoch 3/20\n",
            "11/11 [==============================] - 2s 166ms/step - loss: 0.6927 - accuracy: 0.5053 - val_loss: 0.6959 - val_accuracy: 0.4416\n",
            "Epoch 4/20\n",
            "11/11 [==============================] - 2s 160ms/step - loss: 0.6900 - accuracy: 0.5227 - val_loss: 0.6855 - val_accuracy: 0.7922\n",
            "Epoch 5/20\n",
            "11/11 [==============================] - 2s 156ms/step - loss: 0.6800 - accuracy: 0.7915 - val_loss: 0.6601 - val_accuracy: 0.8442\n",
            "Epoch 6/20\n",
            "11/11 [==============================] - 2s 165ms/step - loss: 0.6292 - accuracy: 0.8590 - val_loss: 0.5635 - val_accuracy: 0.8182\n",
            "Epoch 7/20\n",
            "11/11 [==============================] - 2s 165ms/step - loss: 0.4848 - accuracy: 0.8794 - val_loss: 0.4489 - val_accuracy: 0.8312\n",
            "Epoch 8/20\n",
            "11/11 [==============================] - 2s 166ms/step - loss: 0.3112 - accuracy: 0.9072 - val_loss: 0.3304 - val_accuracy: 0.8701\n",
            "Epoch 9/20\n",
            "11/11 [==============================] - 2s 167ms/step - loss: 0.1797 - accuracy: 0.9516 - val_loss: 0.2409 - val_accuracy: 0.9351\n",
            "Epoch 10/20\n",
            "11/11 [==============================] - 2s 166ms/step - loss: 0.0910 - accuracy: 0.9822 - val_loss: 0.2168 - val_accuracy: 0.9481\n",
            "Epoch 11/20\n",
            "11/11 [==============================] - 2s 168ms/step - loss: 0.0633 - accuracy: 0.9869 - val_loss: 0.1953 - val_accuracy: 0.9481\n",
            "Epoch 12/20\n",
            "11/11 [==============================] - 2s 171ms/step - loss: 0.0272 - accuracy: 0.9968 - val_loss: 0.1449 - val_accuracy: 0.9610\n",
            "Epoch 13/20\n",
            "11/11 [==============================] - 2s 156ms/step - loss: 0.0127 - accuracy: 1.0000 - val_loss: 0.1454 - val_accuracy: 0.9481\n",
            "Epoch 14/20\n",
            "11/11 [==============================] - 2s 164ms/step - loss: 0.0072 - accuracy: 1.0000 - val_loss: 0.1498 - val_accuracy: 0.9610\n",
            "Epoch 15/20\n",
            "11/11 [==============================] - 2s 166ms/step - loss: 0.0043 - accuracy: 1.0000 - val_loss: 0.1588 - val_accuracy: 0.9610\n",
            "Epoch 16/20\n",
            "11/11 [==============================] - 2s 164ms/step - loss: 0.0028 - accuracy: 1.0000 - val_loss: 0.1659 - val_accuracy: 0.9610\n",
            "Epoch 17/20\n",
            "11/11 [==============================] - 2s 166ms/step - loss: 0.0019 - accuracy: 1.0000 - val_loss: 0.1750 - val_accuracy: 0.9610\n",
            "Epoch 18/20\n",
            "11/11 [==============================] - 2s 165ms/step - loss: 0.0013 - accuracy: 1.0000 - val_loss: 0.1807 - val_accuracy: 0.9610\n",
            "Epoch 19/20\n",
            "11/11 [==============================] - 2s 167ms/step - loss: 9.4100e-04 - accuracy: 1.0000 - val_loss: 0.1868 - val_accuracy: 0.9610\n",
            "Epoch 20/20\n",
            "11/11 [==============================] - 2s 160ms/step - loss: 7.7137e-04 - accuracy: 1.0000 - val_loss: 0.1919 - val_accuracy: 0.9610\n",
            "Model: \"model_31\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "input_63 (InputLayer)           [(None, 200)]        0                                            \n",
            "__________________________________________________________________________________________________\n",
            "input_64 (InputLayer)           [(None, 200)]        0                                            \n",
            "__________________________________________________________________________________________________\n",
            "embedding_62 (Embedding)        (None, 200, 10)      5300        input_63[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "embedding_63 (Embedding)        (None, 200, 10)      5300        input_64[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "reshape_62 (Reshape)            (None, 200, 10)      0           embedding_62[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "reshape_63 (Reshape)            (None, 200, 10)      0           embedding_63[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_248 (Conv1D)             (None, 198, 128)     3968        reshape_62[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_252 (Conv1D)             (None, 198, 128)     3968        reshape_63[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_249 (Conv1D)             (None, 196, 64)      24640       conv1d_248[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_253 (Conv1D)             (None, 196, 64)      24640       conv1d_252[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_186 (MaxPooling1D (None, 98, 64)       0           conv1d_249[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_189 (MaxPooling1D (None, 98, 64)       0           conv1d_253[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_250 (Conv1D)             (None, 96, 32)       6176        max_pooling1d_186[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_254 (Conv1D)             (None, 96, 32)       6176        max_pooling1d_189[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_187 (MaxPooling1D (None, 48, 32)       0           conv1d_250[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_190 (MaxPooling1D (None, 48, 32)       0           conv1d_254[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_251 (Conv1D)             (None, 46, 32)       3104        max_pooling1d_187[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "conv1d_255 (Conv1D)             (None, 46, 32)       3104        max_pooling1d_190[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_188 (MaxPooling1D (None, 23, 32)       0           conv1d_251[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "max_pooling1d_191 (MaxPooling1D (None, 23, 32)       0           conv1d_255[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "flatten_62 (Flatten)            (None, 736)          0           max_pooling1d_188[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "flatten_63 (Flatten)            (None, 736)          0           max_pooling1d_191[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "add_31 (Add)                    (None, 736)          0           flatten_62[0][0]                 \n",
            "                                                                 flatten_63[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "dense_93 (Dense)                (None, 128)          94336       add_31[0][0]                     \n",
            "__________________________________________________________________________________________________\n",
            "dense_94 (Dense)                (None, 32)           4128        dense_93[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "dense_95 (Dense)                (None, 2)            66          dense_94[0][0]                   \n",
            "==================================================================================================\n",
            "Total params: 184,906\n",
            "Trainable params: 184,906\n",
            "Non-trainable params: 0\n",
            "__________________________________________________________________________________________________\n",
            "3/3 [==============================] - 0s 20ms/step - loss: 0.1919 - accuracy: 0.9610\n"
          ],
          "name": "stdout"
        },
        {
          "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"
          }
        },
        {
          "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"
          }
        },
        {
          "output_type": "stream",
          "text": [
            "0.954878443479538\n"
          ],
          "name": "stdout"
        }
      ]
    }
  ]
}