toxic_comments_classification.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "uu1CkkGatnsq"
   },
   "source": [
    "# **Toxic Comments Classification**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "JCmk_f5-tnsv"
   },
   "source": [
    "### Import Required Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "id": "EJ_93CLctnsw"
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "from matplotlib import pyplot as plt\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras.layers import TextVectorization\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Embedding,Bidirectional, LSTM, Dropout, Dense\n",
    "from tensorflow.keras.metrics import Precision, Recall, CategoricalAccuracy\n",
    "\n",
    "import pickle\n",
    "\n",
    "import warnings\n",
    "warnings.filterwarnings('ignore')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "BVjdxSzCtnsy"
   },
   "source": [
    "### Loading Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "id": "Ovj3l6ATtnsz"
   },
   "outputs": [],
   "source": [
    "df = pd.read_csv('./Datasets/train.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 337
    },
    "id": "ILrW8o2Ztnsz",
    "outputId": "31eaec0d-d3dd-4b96-86a7-e8948739bfaa"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "\n",
       "  <div id=\"df-761a750b-6e92-4c4d-908e-abf2bfeab7dc\">\n",
       "    <div class=\"colab-df-container\">\n",
       "      <div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>comment_text</th>\n",
       "      <th>toxic</th>\n",
       "      <th>severe_toxic</th>\n",
       "      <th>obscene</th>\n",
       "      <th>threat</th>\n",
       "      <th>insult</th>\n",
       "      <th>identity_hate</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0000997932d777bf</td>\n",
       "      <td>Explanation\\nWhy the edits made under my usern...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>000103f0d9cfb60f</td>\n",
       "      <td>D'aww! He matches this background colour I'm s...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>000113f07ec002fd</td>\n",
       "      <td>Hey man, I'm really not trying to edit war. It...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0001b41b1c6bb37e</td>\n",
       "      <td>\"\\nMore\\nI can't make any real suggestions on ...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0001d958c54c6e35</td>\n",
       "      <td>You, sir, are my hero. Any chance you remember...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>\n",
       "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-761a750b-6e92-4c4d-908e-abf2bfeab7dc')\"\n",
       "              title=\"Convert this dataframe to an interactive table.\"\n",
       "              style=\"display:none;\">\n",
       "        \n",
       "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
       "       width=\"24px\">\n",
       "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
       "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
       "  </svg>\n",
       "      </button>\n",
       "      \n",
       "  <style>\n",
       "    .colab-df-container {\n",
       "      display:flex;\n",
       "      flex-wrap:wrap;\n",
       "      gap: 12px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert {\n",
       "      background-color: #E8F0FE;\n",
       "      border: none;\n",
       "      border-radius: 50%;\n",
       "      cursor: pointer;\n",
       "      display: none;\n",
       "      fill: #1967D2;\n",
       "      height: 32px;\n",
       "      padding: 0 0 0 0;\n",
       "      width: 32px;\n",
       "    }\n",
       "\n",
       "    .colab-df-convert:hover {\n",
       "      background-color: #E2EBFA;\n",
       "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
       "      fill: #174EA6;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert {\n",
       "      background-color: #3B4455;\n",
       "      fill: #D2E3FC;\n",
       "    }\n",
       "\n",
       "    [theme=dark] .colab-df-convert:hover {\n",
       "      background-color: #434B5C;\n",
       "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
       "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
       "      fill: #FFFFFF;\n",
       "    }\n",
       "  </style>\n",
       "\n",
       "      <script>\n",
       "        const buttonEl =\n",
       "          document.querySelector('#df-761a750b-6e92-4c4d-908e-abf2bfeab7dc button.colab-df-convert');\n",
       "        buttonEl.style.display =\n",
       "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
       "\n",
       "        async function convertToInteractive(key) {\n",
       "          const element = document.querySelector('#df-761a750b-6e92-4c4d-908e-abf2bfeab7dc');\n",
       "          const dataTable =\n",
       "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
       "                                                     [key], {});\n",
       "          if (!dataTable) return;\n",
       "\n",
       "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
       "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
       "            + ' to learn more about interactive tables.';\n",
       "          element.innerHTML = '';\n",
       "          dataTable['output_type'] = 'display_data';\n",
       "          await google.colab.output.renderOutput(dataTable, element);\n",
       "          const docLink = document.createElement('div');\n",
       "          docLink.innerHTML = docLinkHtml;\n",
       "          element.appendChild(docLink);\n",
       "        }\n",
       "      </script>\n",
       "    </div>\n",
       "  </div>\n",
       "  "
      ],
      "text/plain": [
       "                 id                                       comment_text  toxic  \\\n",
       "0  0000997932d777bf  Explanation\\nWhy the edits made under my usern...      0   \n",
       "1  000103f0d9cfb60f  D'aww! He matches this background colour I'm s...      0   \n",
       "2  000113f07ec002fd  Hey man, I'm really not trying to edit war. It...      0   \n",
       "3  0001b41b1c6bb37e  \"\\nMore\\nI can't make any real suggestions on ...      0   \n",
       "4  0001d958c54c6e35  You, sir, are my hero. Any chance you remember...      0   \n",
       "\n",
       "   severe_toxic  obscene  threat  insult  identity_hate  \n",
       "0             0        0       0       0              0  \n",
       "1             0        0       0       0              0  \n",
       "2             0        0       0       0              0  \n",
       "3             0        0       0       0              0  \n",
       "4             0        0       0       0              0  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# lets check the glimpse of first five rows of train dataset\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "2DOu-0yytns1",
    "outputId": "87dbc35f-db6f-4c19-ea25-c8d7f876f2ad"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(159571, 8)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# shape of train dataframe\n",
    "df.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data Preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "id": "txbgNvcatns2"
   },
   "outputs": [],
   "source": [
    "# Separate Independent and Dependent Variables\n",
    "X = df['comment_text']\n",
    "y = df.loc[:, 'toxic':].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "fty90eV6tns3",
    "outputId": "88a96b77-70eb-4479-dd66-8b0cfc004dc6"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5000"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# lets check the comment maximum length \n",
    "df['comment_text'].str.len().max()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "id": "pdIxvzBdtns4"
   },
   "outputs": [],
   "source": [
    "MAX_FEATURES = 200000 # number of words in the vocab"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "id": "pryLpOTYtns5"
   },
   "outputs": [],
   "source": [
    "# Text Vectorization is the process of converting text into a numerical representation.\n",
    "# It transforms text into a more suitable form so that ML or DL algorithms can perform better.\n",
    "vectorizer = TextVectorization(max_tokens=MAX_FEATURES, \n",
    "                               output_sequence_length=1800,  # Maximum comment (text) size in words\n",
    "                               output_mode='int')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "id": "-voV-Pm5tns5"
   },
   "outputs": [],
   "source": [
    "vectorizer.adapt(X.values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "id": "VNsyQN5ktns6"
   },
   "outputs": [],
   "source": [
    "# apply vectorizer on X\n",
    "vectorized_text = vectorizer(X.values)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "id": "OqQZ2C0Ltns6"
   },
   "outputs": [],
   "source": [
    "# prepare tensorflow dataset\n",
    "# it converts the data stored in Pandas Data Frame into data stored in TensorFlow Data Set.\n",
    "dataset = tf.data.Dataset.from_tensor_slices((vectorized_text, y))\n",
    "# cache keeps the images in memory after they're loaded off disk during the first epoch.\n",
    "dataset = dataset.cache()\n",
    "# For true randomness, we set the shuffle buffer to the full dataset size.\n",
    "dataset = dataset.shuffle(160000)\n",
    "# Batch after shuffling to get unique batches at each epoch\n",
    "dataset = dataset.batch(16)\n",
    "# prefetch overlaps data preprocessing and model execution while training.\n",
    "dataset = dataset.prefetch(8) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "id": "u-XRXCJxtns7"
   },
   "outputs": [],
   "source": [
    "# train(70%), validation(20%) and test(10%) split\n",
    "train = dataset.take(int(len(dataset)*.7))\n",
    "val = dataset.skip(int(len(dataset)*.7)).take(int(len(dataset)*.2))\n",
    "test = dataset.skip(int(len(dataset)*.9)).take(int(len(dataset)*.1))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model Building"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "id": "Ktqe4lVxtns8"
   },
   "outputs": [],
   "source": [
    "# Formation of CNN model\n",
    "# Sequential layer\n",
    "model = Sequential()\n",
    "# Create the embedding layer \n",
    "model.add(Embedding(MAX_FEATURES+1, 32))\n",
    "# Bidirectional LSTM Layer\n",
    "model.add(Bidirectional(LSTM(32, activation='tanh')))\n",
    "# Feature extractor Fully connected layers\n",
    "model.add(Dense(128, activation='relu'))\n",
    "model.add(Dense(256, activation='relu'))\n",
    "model.add(Dense(128, activation='relu'))\n",
    "# Final layer/ output layer \n",
    "model.add(Dense(6, activation='sigmoid'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "id": "_IK0W1Tdtns9"
   },
   "outputs": [],
   "source": [
    "# Compile the Model\n",
    "model.compile(loss='BinaryCrossentropy', optimizer='Adam')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "RszCisN_tns9",
    "outputId": "601c9889-9372-48c2-945f-d1032e4ad920"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"sequential\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " embedding (Embedding)       (None, None, 32)          6400032   \n",
      "                                                                 \n",
      " bidirectional (Bidirectiona  (None, 64)               16640     \n",
      " l)                                                              \n",
      "                                                                 \n",
      " dense (Dense)               (None, 128)               8320      \n",
      "                                                                 \n",
      " dense_1 (Dense)             (None, 256)               33024     \n",
      "                                                                 \n",
      " dense_2 (Dense)             (None, 128)               32896     \n",
      "                                                                 \n",
      " dense_3 (Dense)             (None, 6)                 774       \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 6,491,686\n",
      "Trainable params: 6,491,686\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "# model summary\n",
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Model Training"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "bM62WKfWtns-",
    "outputId": "1e4718e3-0b76-4ff8-a6e0-357b303136fb"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/9\n",
      "6981/6981 [==============================] - 617s 87ms/step - loss: 0.0620 - val_loss: 0.0453\n",
      "Epoch 2/9\n",
      "6981/6981 [==============================] - 618s 89ms/step - loss: 0.0457 - val_loss: 0.0394\n",
      "Epoch 3/9\n",
      "6981/6981 [==============================] - 615s 88ms/step - loss: 0.0402 - val_loss: 0.0352\n",
      "Epoch 4/9\n",
      "6981/6981 [==============================] - 615s 88ms/step - loss: 0.0358 - val_loss: 0.0326\n",
      "Epoch 5/9\n",
      "6981/6981 [==============================] - 634s 91ms/step - loss: 0.0323 - val_loss: 0.0281\n",
      "Epoch 6/9\n",
      "6981/6981 [==============================] - 629s 90ms/step - loss: 0.0289 - val_loss: 0.0247\n",
      "Epoch 7/9\n",
      "6981/6981 [==============================] - 619s 89ms/step - loss: 0.0255 - val_loss: 0.0214\n",
      "Epoch 8/9\n",
      "6981/6981 [==============================] - 622s 89ms/step - loss: 0.0232 - val_loss: 0.0192\n",
      "Epoch 9/9\n",
      "6981/6981 [==============================] - 633s 91ms/step - loss: 0.0204 - val_loss: 0.0179\n"
     ]
    }
   ],
   "source": [
    "# fit model for only 9 epochs\n",
    "history = model.fit(train, epochs=9, validation_data=val)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 283
    },
    "id": "r8ak_1Mdtns_",
    "outputId": "e40baa77-dd91-43fb-d6e0-48bd46db44fe"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Figure size 576x360 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,5))\n",
    "pd.DataFrame(history.history).plot()\n",
    "plt.xlabel('Epochs')\n",
    "plt.ylabel('Loss')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "id": "ldblnI8etntA"
   },
   "outputs": [],
   "source": [
    "# save the model for future use\n",
    "model.save('toxic_comments_model.h5')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# load the model\n",
    "model = tf.keras.models.load_model('toxic_comments_model.h5')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "# use text vectorization on raw data to test the model\n",
    "input_text = vectorizer(np.expand_dims('I am going to hit you.',0))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[[0.9829285  0.36757064 0.771432   0.41207135 0.6273679  0.17438579]]\n"
     ]
    }
   ],
   "source": [
    "# model prediction on raw text\n",
    "res = model.predict(input_text)\n",
    "print(res)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "# lets check the precision, recall and accuracy on test dataset \n",
    "pre = Precision()\n",
    "re = Recall()\n",
    "acc = CategoricalAccuracy()\n",
    "\n",
    "for batch in test.as_numpy_iterator(): \n",
    "    # Unpack the batch \n",
    "    X_true, y_true = batch\n",
    "    # Make a prediction \n",
    "    yhat = model.predict(X_true)\n",
    "    \n",
    "    # Flatten the true values\n",
    "    y_true = y_true.flatten()\n",
    "    # Flatten the predictions\n",
    "    yhat = yhat.flatten()\n",
    "    \n",
    "    pre.update_state(y_true, yhat)\n",
    "    re.update_state(y_true, yhat)\n",
    "    acc.update_state(y_true, yhat)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Precision: 0.9402751922607422, Recall:0.8658220767974854, Accuracy:0.5195586681365967\n"
     ]
    }
   ],
   "source": [
    "print(f'Precision: {pre.result().numpy()}, Recall:{re.result().numpy()}, Accuracy:{acc.result().numpy()}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "def score_comment(comment):\n",
    "    vectorized_comment = vectorizer([comment])\n",
    "    results = model.predict(vectorized_comment)\n",
    "    \n",
    "    text = ''\n",
    "    for idx, col in enumerate(df.columns[2:]):\n",
    "        text += '{}: {}\\n'.format(col, results[0][idx]>0.5)\n",
    "    \n",
    "    return text"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "toxic: True\n",
      "severe_toxic: False\n",
      "obscene: False\n",
      "threat: True\n",
      "insult: False\n",
      "identity_hate: False\n",
      "\n"
     ]
    }
   ],
   "source": [
    "print(score_comment('I am going to kill you.'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "# save the labels for further use in deployment\n",
    "labels = ['toxic', 'severe_toxic', 'obscene', 'threat', 'insult', 'identity_hate']\n",
    "pickle.dump(labels,open('labels.pkl','wb'))"
   ]
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "name": "toxic_comments_classification.ipynb",
   "provenance": []
  },
  "gpuClass": "standard",
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}