LSTM.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "LSTM.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/srushtinandu/ML-Project/blob/main/LSTM.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "g7OtlcnyBS0a",
        "outputId": "3dabcf2d-91e7-4ad7-e6bc-233da8fbc337"
      },
      "source": [
        "import nltk\n",
        "nltk.download('stopwords')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[nltk_data] Downloading package stopwords to /root/nltk_data...\n",
            "[nltk_data]   Unzipping corpora/stopwords.zip.\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "True"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 1
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rAERqmoZiMwr"
      },
      "source": [
        "import pandas as pd\n",
        "import keras\n",
        "from keras.datasets import mnist\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dense, Dropout, LSTM, Embedding, SpatialDropout1D\n",
        "from keras.optimizers import Adam\n",
        "import seaborn as sns\n",
        "import matplotlib.pyplot as plt\n",
        "import re\n",
        "from nltk.corpus import stopwords\n",
        "from keras.preprocessing.text import Tokenizer\n",
        "from keras.preprocessing.sequence import pad_sequences\n",
        "from keras.callbacks import EarlyStopping"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8ID9kTabw_90",
        "outputId": "7ecc9cf2-7ec4-4710-ab19-b07c8c1ff6bf"
      },
      "source": [
        "url = 'https://raw.githubusercontent.com/srushtinandu/NLP-Project/main/data/train.tsv'\n",
        "df = pd.read_csv(url, sep = '\\t')\n",
        "df.columns=['ID', 'Label', 'Statement', 'subject(s)', 'speaker', 'speaker\\'s job', 'state info', 'party affiliation', 'barely true counts', 'false counts', 'half true counts', 'mostly true counts', 'pants on fire counts', 'context']\n",
        "df.info()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "RangeIndex: 10239 entries, 0 to 10238\n",
            "Data columns (total 14 columns):\n",
            " #   Column                Non-Null Count  Dtype  \n",
            "---  ------                --------------  -----  \n",
            " 0   ID                    10239 non-null  object \n",
            " 1   Label                 10239 non-null  object \n",
            " 2   Statement             10239 non-null  object \n",
            " 3   subject(s)            10237 non-null  object \n",
            " 4   speaker               10237 non-null  object \n",
            " 5   speaker's job         7342 non-null   object \n",
            " 6   state info            8031 non-null   object \n",
            " 7   party affiliation     10237 non-null  object \n",
            " 8   barely true counts    10237 non-null  float64\n",
            " 9   false counts          10237 non-null  float64\n",
            " 10  half true counts      10237 non-null  float64\n",
            " 11  mostly true counts    10237 non-null  float64\n",
            " 12  pants on fire counts  10237 non-null  float64\n",
            " 13  context               10137 non-null  object \n",
            "dtypes: float64(5), object(9)\n",
            "memory usage: 1.1+ MB\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        },
        "id": "SKZ6tNXnykK6",
        "outputId": "b6a89cb1-51cd-4970-bc62-579d0d025282"
      },
      "source": [
        "sns.countplot(x=df.Label)\n",
        "plt.xlabel('Labels')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NAPxLrkWN9mi"
      },
      "source": [
        "def to_label(label):\n",
        "  if label == 'mostly-true' or label == 'half-true':\n",
        "    return 'true'\n",
        "  elif label == 'pants-fire' or label == 'barely-true':\n",
        "    return 'false'\n",
        "  else:\n",
        "    return label\n",
        "\n",
        "df['Label'] = df.Label.apply(to_label)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sjYs0vPFfAw4"
      },
      "source": [
        "df = df[df['Statement'].notna()]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0Pacm8a7zoYp"
      },
      "source": [
        "df = df.reset_index(drop=True)\n",
        "NUMBERS_RE = re.compile('[^0-9]+')\n",
        "REPLACE_BY_SPACE_RE = re.compile('[/(){}\\[\\]\\|@,;]')\n",
        "BAD_SYMBOLS_RE = re.compile('[^0-9a-z #+_]')\n",
        "STOPWORDS = set(stopwords.words('english'))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "d5bmF0am0MU_"
      },
      "source": [
        "def clean_text(text):\n",
        "    \"\"\"\n",
        "        text: a string\n",
        "        \n",
        "        return: modified initial string\n",
        "    \"\"\"\n",
        "    result = re.match(r\"[0-9]+\", text)\n",
        "    if result != None:\n",
        "      text = NUMBERS_RE.sub('X', text)\n",
        "\n",
        "    text = text.replace('X', '')\n",
        "    text = text.lower() # lowercase text\n",
        "    text = REPLACE_BY_SPACE_RE.sub(' ', text) # replace REPLACE_BY_SPACE_RE symbols by space in text. substitute the matched string in REPLACE_BY_SPACE_RE with space.\n",
        "    text = BAD_SYMBOLS_RE.sub('', text) # remove symbols which are in BAD_SYMBOLS_RE from text. substitute the matched string in BAD_SYMBOLS_RE with nothing. \n",
        "    #text = re.sub(r'\\W+', '', text)\n",
        "    text = ' '.join(word for word in text.split() if word not in STOPWORDS) # remove stopwors from text\n",
        "    return text\n",
        "\n",
        "df['Statement'] = df['Statement'].apply(clean_text)\n",
        "df['Statement'] = df['Statement'].str.replace('\\d+', '')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "JBvu4bZ-01Zw",
        "outputId": "7a7744e8-5df5-4157-dcba-f6d6fd905b05"
      },
      "source": [
        "# The maximum number of words to be used. (most frequent)\n",
        "MAX_NB_WORDS = 10000\n",
        "# Max number of words in each complaint.\n",
        "MAX_SEQUENCE_LENGTH = 50\n",
        "# This is fixed.\n",
        "EMBEDDING_DIM = 100\n",
        "tokenizer = Tokenizer(num_words=MAX_NB_WORDS, filters='!\"#$%&()*+,-./:;<=>?@[\\]^_`{|}~', lower=True)\n",
        "tokenizer.fit_on_texts(df['Statement'].values)\n",
        "word_index = tokenizer.word_index\n",
        "print('Found %s unique tokens.' % len(word_index))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 12281 unique tokens.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "eIA8_scg2qnp",
        "outputId": "8b14bbc3-4c7a-4f95-cb90-30a9a6fabfe0"
      },
      "source": [
        "X_train = tokenizer.texts_to_sequences(df['Statement'].values)\n",
        "X_train = pad_sequences(X_train, maxlen=MAX_SEQUENCE_LENGTH)\n",
        "print('Shape of data tensor:', X_train.shape)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Shape of data tensor: (10239, 50)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "eiaTLsJIA3DS",
        "outputId": "466deb34-87f9-45f1-c69b-a002a8b44a8b"
      },
      "source": [
        "y_train = pd.get_dummies(df['Label']).values\n",
        "print('Shape of label tensor:', y_train.shape)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Shape of label tensor: (10239, 2)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZxQ3cr68D984",
        "outputId": "7462db10-a010-4015-92a1-25c3a0619843"
      },
      "source": [
        "url = 'https://raw.githubusercontent.com/srushtinandu/NLP-Project/main/data/valid.tsv'\n",
        "valid_data = pd.read_csv(url, sep = '\\t')\n",
        "valid_data.columns=['ID', 'Label', 'Statement', 'subject(s)', 'speaker', 'speaker\\'s job', 'state info', 'party affiliation', 'barely true counts', 'false counts', 'half true counts', 'mostly true counts', 'pants on fire counts', 'context']\n",
        "valid_data.info()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<class 'pandas.core.frame.DataFrame'>\n",
            "RangeIndex: 1283 entries, 0 to 1282\n",
            "Data columns (total 14 columns):\n",
            " #   Column                Non-Null Count  Dtype \n",
            "---  ------                --------------  ----- \n",
            " 0   ID                    1283 non-null   object\n",
            " 1   Label                 1283 non-null   object\n",
            " 2   Statement             1283 non-null   object\n",
            " 3   subject(s)            1283 non-null   object\n",
            " 4   speaker               1283 non-null   object\n",
            " 5   speaker's job         938 non-null    object\n",
            " 6   state info            1004 non-null   object\n",
            " 7   party affiliation     1283 non-null   object\n",
            " 8   barely true counts    1283 non-null   int64 \n",
            " 9   false counts          1283 non-null   int64 \n",
            " 10  half true counts      1283 non-null   int64 \n",
            " 11  mostly true counts    1283 non-null   int64 \n",
            " 12  pants on fire counts  1283 non-null   int64 \n",
            " 13  context               1271 non-null   object\n",
            "dtypes: int64(5), object(9)\n",
            "memory usage: 140.5+ KB\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ElON2zfxh_8A"
      },
      "source": [
        "valid_data = valid_data[valid_data['Statement'].notna()]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "CJ6FhCxqEPwZ",
        "outputId": "a7940302-de89-47bc-b701-5fe42228b825"
      },
      "source": [
        "valid_data['Label'] = valid_data.Label.apply(to_label)\n",
        "valid_data = valid_data.reset_index(drop=True)\n",
        "valid_data['Statement'] = valid_data['Statement'].apply(clean_text)\n",
        "valid_data['Statement'] = valid_data['Statement'].str.replace('\\d+', '')\n",
        "tokenizer.fit_on_texts(valid_data['Statement'].values)\n",
        "word_index = tokenizer.word_index\n",
        "print('Found %s unique tokens.' % len(word_index))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 13002 unique tokens.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "x-tVNUuWF7pR",
        "outputId": "9eca080a-fb6d-43f5-be64-b9e77d30e04f"
      },
      "source": [
        "X_valid = tokenizer.texts_to_sequences(valid_data['Statement'].values)\n",
        "X_valid = pad_sequences(X_valid, maxlen=MAX_SEQUENCE_LENGTH)\n",
        "print('Shape of data tensor:', X_valid.shape)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Shape of data tensor: (1283, 50)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ywl_t52JGNS6",
        "outputId": "0db8974d-672b-44b8-ef5a-5704c314f11e"
      },
      "source": [
        "y_valid = pd.get_dummies(valid_data['Label']).values\n",
        "print('Shape of label tensor:', y_valid.shape)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Shape of label tensor: (1283, 2)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "rU73Kx-XGcE9",
        "outputId": "081259e0-47ae-4f80-8360-42674adf8b30"
      },
      "source": [
        "model = Sequential()\n",
        "model.add(Embedding(MAX_NB_WORDS, EMBEDDING_DIM, input_length=X_train.shape[1]))\n",
        "model.add(SpatialDropout1D(0.2))\n",
        "model.add(LSTM(128, dropout=0.3, recurrent_dropout=0.2))\n",
        "model.add(Dense(64, activation='relu'))\n",
        "model.add(Dropout(0.3))\n",
        "model.add(Dense(2, activation='relu'))\n",
        "model.compile(loss='binary_crossentropy', optimizer='Adam', metrics=['binary_accuracy'])\n",
        "\n",
        "epochs = 7\n",
        "batch_size = 32\n",
        "\n",
        "history = model.fit(X_train, y_train, epochs=epochs, batch_size=batch_size,validation_data=(X_valid, y_valid), callbacks=[EarlyStopping(monitor='val_loss', patience=3, min_delta=0.0001)])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "WARNING:tensorflow:Layer lstm will not use cuDNN kernel since it doesn't meet the cuDNN kernel criteria. It will use generic GPU kernel as fallback when running on GPU\n",
            "Epoch 1/7\n",
            "320/320 [==============================] - 65s 186ms/step - loss: 0.8647 - binary_accuracy: 0.5265 - val_loss: 0.6929 - val_binary_accuracy: 0.5187\n",
            "Epoch 2/7\n",
            "320/320 [==============================] - 60s 189ms/step - loss: 0.6569 - binary_accuracy: 0.6254 - val_loss: 0.7122 - val_binary_accuracy: 0.5269\n",
            "Epoch 3/7\n",
            "320/320 [==============================] - 61s 191ms/step - loss: 0.5743 - binary_accuracy: 0.7331 - val_loss: 0.7929 - val_binary_accuracy: 0.5121\n",
            "Epoch 4/7\n",
            "320/320 [==============================] - 61s 190ms/step - loss: 0.5033 - binary_accuracy: 0.7717 - val_loss: 1.6794 - val_binary_accuracy: 0.5156\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "NiH6N6NmYM54",
        "outputId": "413a41f2-62c6-48d5-a2a7-394fb72384ec"
      },
      "source": [
        "model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "embedding (Embedding)        (None, 50, 100)           1000000   \n",
            "_________________________________________________________________\n",
            "spatial_dropout1d (SpatialDr (None, 50, 100)           0         \n",
            "_________________________________________________________________\n",
            "lstm (LSTM)                  (None, 128)               117248    \n",
            "_________________________________________________________________\n",
            "dense (Dense)                (None, 64)                8256      \n",
            "_________________________________________________________________\n",
            "dropout (Dropout)            (None, 64)                0         \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 2)                 130       \n",
            "=================================================================\n",
            "Total params: 1,125,634\n",
            "Trainable params: 1,125,634\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hPfqOCifGoie"
      },
      "source": [
        "url = 'https://raw.githubusercontent.com/srushtinandu/NLP-Project/main/data/test.tsv'\n",
        "test_data = pd.read_csv(url, sep = '\\t')\n",
        "test_data.columns=['ID', 'Label', 'Statement', 'subject(s)', 'speaker', 'speaker\\'s job', 'state info', 'party affiliation', 'barely true counts', 'false counts', 'half true counts', 'mostly true counts', 'pants on fire counts', 'context']"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1sRqdi6LrpfA"
      },
      "source": [
        "test_data = test_data[test_data['Statement'].notna()]"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xAwVTbbLHILS",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "efafec59-a52b-4b7f-aff6-86c8af1cb000"
      },
      "source": [
        "test_data['Label'] = test_data.Label.apply(to_label)\n",
        "test_data = test_data.reset_index(drop=True)\n",
        "test_data['Statement'] = test_data['Statement'].apply(clean_text)\n",
        "test_data['Statement'] = test_data['Statement'].str.replace('\\d+', '')\n",
        "tokenizer.fit_on_texts(test_data['Statement'].values)\n",
        "word_index = tokenizer.word_index\n",
        "print('Found %s unique tokens.' % len(word_index))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 13699 unique tokens.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VAb37Q7yHXgm",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "dde6e7b0-53d2-4bf3-e37f-4d6a63372ca1"
      },
      "source": [
        "X_test = tokenizer.texts_to_sequences(test_data['Statement'].values)\n",
        "X_test = pad_sequences(X_test, maxlen=MAX_SEQUENCE_LENGTH)\n",
        "print('Shape of data tensor:', X_test.shape)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Shape of data tensor: (1266, 50)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YMe75BrPHfuN",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "db2ff489-965c-4e40-eb41-aa43d4d5cd2f"
      },
      "source": [
        "y_test = pd.get_dummies(test_data['Label']).values\n",
        "print('Shape of label tensor:', y_test.shape)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Shape of label tensor: (1266, 2)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0T2tlmQtJd0O",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "f18db093-62ea-4c4a-e5e0-f2853387451d"
      },
      "source": [
        "accr = model.evaluate(X_test,y_test)\n",
        "print('Test set\\n  Loss: {:0.3f}\\n  Accuracy: {:0.3f}'.format(accr[0],accr[1]))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "40/40 [==============================] - 1s 13ms/step - loss: 1.4731 - binary_accuracy: 0.5340\n",
            "Test set\n",
            "  Loss: 1.473\n",
            "  Accuracy: 0.534\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gafFyokqtr2Z",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 281
        },
        "outputId": "f1e15993-7e2a-41d1-f39d-e8a59d08a02b"
      },
      "source": [
        "plt.title('Loss')\n",
        "plt.plot(history.history['loss'], label='train')\n",
        "plt.plot(history.history['val_loss'], label='valid')\n",
        "plt.legend()\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WcZ2-1nYxv7I",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "c0d07d1c-da32-4384-a1e8-45dde5e08196"
      },
      "source": [
        "for key in history.history.keys():\n",
        "    print(key)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "loss\n",
            "binary_accuracy\n",
            "val_loss\n",
            "val_binary_accuracy\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-IJKApWvt1vc",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 281
        },
        "outputId": "845ef8e4-528b-49da-e363-72ff566bf11a"
      },
      "source": [
        "plt.title('Accuracy')\n",
        "plt.plot(history.history['binary_accuracy'], label='train')\n",
        "plt.plot(history.history['val_binary_accuracy'], label='valid')\n",
        "plt.legend()\n",
        "plt.show();"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}