Supervised_glove_100k.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "executionInfo": {
     "elapsed": 33317,
     "status": "ok",
     "timestamp": 1622242632908,
     "user": {
      "displayName": "Ai Zhong",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14Gh2MWN-q_WG3Xb3YELMkck5bLErFAYBIGqZ_8lQ=s64",
      "userId": "11851498678274176197"
     },
     "user_tz": 420
    },
    "id": "PkZPXM8NrnSg"
   },
   "outputs": [],
   "source": [
    "import os\n",
    "import numpy as np\n",
    "import pickle\n",
    "\n",
    "## if you do not have glove:\n",
    "# !wget http://nlp.stanford.edu/data/glove.6B.zip\n",
    "# !unzip glove*.zip\n",
    "\n",
    "GLOVE_DIR = 'drive/MyDrive/Milestone2/glove.6B.100d.txt'\n",
    "glove_dict = {}\n",
    "with open(GLOVE_DIR) as f:\n",
    "    for line in f:\n",
    "        word, *vector = line.split()\n",
    "        glove_dict[word]=np.array(vector).astype(float)\n",
    "\n",
    "re_tokenized = pickle.load(open('sample_data/re_tokenized_lemma.pkl','rb'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "executionInfo": {
     "elapsed": 6,
     "status": "ok",
     "timestamp": 1622242632909,
     "user": {
      "displayName": "Ai Zhong",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14Gh2MWN-q_WG3Xb3YELMkck5bLErFAYBIGqZ_8lQ=s64",
      "userId": "11851498678274176197"
     },
     "user_tz": 420
    },
    "id": "d7frfiHPrzdA"
   },
   "outputs": [],
   "source": [
    "def get_average_glove_vec(list_of_tokens):\n",
    "    avg_vec=[]\n",
    "    for token in list_of_tokens:\n",
    "    if token in glove_dict:\n",
    "        avg_vec.append(glove_dict[token])\n",
    "\n",
    "    if len(avg_vec)>0:\n",
    "    avg_vec = np.mean(avg_vec, axis=0)\n",
    "    else:\n",
    "    avg_vec = np.zeros(100)\n",
    "\n",
    "    return avg_vec"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 17160,
     "status": "ok",
     "timestamp": 1622242650065,
     "user": {
      "displayName": "Ai Zhong",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14Gh2MWN-q_WG3Xb3YELMkck5bLErFAYBIGqZ_8lQ=s64",
      "userId": "11851498678274176197"
     },
     "user_tz": 420
    },
    "id": "5vJF4m56rzfs",
    "outputId": "b5e8f495-f5b4-4658-cee1-3129f0b586c6"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 416768/416768 [00:14<00:00, 28299.14it/s]\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "from tqdm import tqdm\n",
    "\n",
    "df_original = pd.read_csv(\"sample_data/WikiLarge_Train.csv\")\n",
    "df_vecs = []\n",
    "\n",
    "for ls in tqdm(re_tokenized):\n",
    "    avg_vec = get_average_glove_vec(ls)\n",
    "    df_vecs.append(avg_vec)\n",
    "\n",
    "df_original['glove_avg'] = df_vecs\n",
    "\n",
    "df = df_original.sample(100000, random_state=42)\n",
    "X = df['glove_avg']\n",
    "y = df['label']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "executionInfo": {
     "elapsed": 1034,
     "status": "ok",
     "timestamp": 1622242651093,
     "user": {
      "displayName": "Ai Zhong",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14Gh2MWN-q_WG3Xb3YELMkck5bLErFAYBIGqZ_8lQ=s64",
      "userId": "11851498678274176197"
     },
     "user_tz": 420
    },
    "id": "wTFYVcj8rzik"
   },
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "Train_X, Test_X, Train_Y, Test_Y = train_test_split(X,y,test_size=0.2, random_state=42)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "executionInfo": {
     "elapsed": 6,
     "status": "ok",
     "timestamp": 1622242651093,
     "user": {
      "displayName": "Ai Zhong",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14Gh2MWN-q_WG3Xb3YELMkck5bLErFAYBIGqZ_8lQ=s64",
      "userId": "11851498678274176197"
     },
     "user_tz": 420
    },
    "id": "vxdXrLNmYTwH"
   },
   "outputs": [],
   "source": [
    "# # save for later complex NN\n",
    "# train_indexes = Train_X.index\n",
    "# test_indexes = Test_X.index\n",
    "Train_X = np.vstack(Train_X)\n",
    "Test_X = np.vstack(Test_X)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "executionInfo": {
     "elapsed": 433,
     "status": "ok",
     "timestamp": 1622242651522,
     "user": {
      "displayName": "Ai Zhong",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14Gh2MWN-q_WG3Xb3YELMkck5bLErFAYBIGqZ_8lQ=s64",
      "userId": "11851498678274176197"
     },
     "user_tz": 420
    },
    "id": "_TUISuFrsFrH"
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "kYWHduFisJfL"
   },
   "source": [
    "# Random Forest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 1076015,
     "status": "ok",
     "timestamp": 1621577045775,
     "user": {
      "displayName": "Ai Zhong",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14Gh2MWN-q_WG3Xb3YELMkck5bLErFAYBIGqZ_8lQ=s64",
      "userId": "11851498678274176197"
     },
     "user_tz": 420
    },
    "id": "s2Vz5FEGsLKg",
    "outputId": "6104f682-c5c1-4dd4-abe8-c7532213392f"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(0.68005, 0.6875335709751452)"
      ]
     },
     "execution_count": 9,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.metrics import accuracy_score, f1_score\n",
    "\n",
    "rf = RandomForestClassifier(bootstrap= True,\n",
    " max_depth=70,\n",
    " max_features='auto',\n",
    " min_samples_leaf=4,\n",
    " min_samples_split=10,\n",
    " n_estimators=800)\n",
    "\n",
    "rf.fit(Train_X, Train_Y)\n",
    "rf_preds = rf.predict(Test_X)\n",
    "rf_acc = accuracy_score(Test_Y, rf_preds)\n",
    "rf_f1 = f1_score(Test_Y, rf_preds)\n",
    "rf_acc, rf_f1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "J96Zh8UQsQOL"
   },
   "outputs": [],
   "source": [
    "pickle.dump(rf, open('drive/MyDrive/Milestone2/trained_models/rf_word2vec.pkl', 'wb'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "executionInfo": {
     "elapsed": 7836,
     "status": "ok",
     "timestamp": 1622242668586,
     "user": {
      "displayName": "Ai Zhong",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14Gh2MWN-q_WG3Xb3YELMkck5bLErFAYBIGqZ_8lQ=s64",
      "userId": "11851498678274176197"
     },
     "user_tz": 420
    },
    "id": "W1JIffdrkLPi"
   },
   "outputs": [],
   "source": [
    "reloaded_rf = pickle.load(open('drive/MyDrive/Milestone2/trained_models/rf_word2vec.pkl', 'rb'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 5097,
     "status": "ok",
     "timestamp": 1622242673680,
     "user": {
      "displayName": "Ai Zhong",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14Gh2MWN-q_WG3Xb3YELMkck5bLErFAYBIGqZ_8lQ=s64",
      "userId": "11851498678274176197"
     },
     "user_tz": 420
    },
    "id": "guJ1rkTFkRmT",
    "outputId": "143ce756-2eb0-4f11-dbf3-6cf7c743a80b"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(68.00500000000001, 68.75335709751452)"
      ]
     },
     "execution_count": 8,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import accuracy_score, f1_score\n",
    "\n",
    "rf_preds = reloaded_rf.predict(Test_X)\n",
    "rf_acc = accuracy_score(Test_Y, rf_preds)*100\n",
    "rf_f1 = f1_score(Test_Y, rf_preds)*100\n",
    "rf_acc, rf_f1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 17800,
     "status": "ok",
     "timestamp": 1622242693614,
     "user": {
      "displayName": "Ai Zhong",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14Gh2MWN-q_WG3Xb3YELMkck5bLErFAYBIGqZ_8lQ=s64",
      "userId": "11851498678274176197"
     },
     "user_tz": 420
    },
    "id": "_Ci1CahxkRo5",
    "outputId": "8e89d1a2-614d-43cb-a99e-bbf1133d4188"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9695875"
      ]
     },
     "execution_count": 9,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "reloaded_rf.score(Train_X, Train_Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 410
    },
    "executionInfo": {
     "elapsed": 5748,
     "status": "ok",
     "timestamp": 1622242699349,
     "user": {
      "displayName": "Ai Zhong",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14Gh2MWN-q_WG3Xb3YELMkck5bLErFAYBIGqZ_8lQ=s64",
      "userId": "11851498678274176197"
     },
     "user_tz": 420
    },
    "id": "b8p0pgzVlTCP",
    "outputId": "bf43cd45-4bf6-43b5-f575-738a0844fa07"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_auc_score, roc_curve, auc\n",
    "\n",
    "y_score = reloaded_rf.predict_proba(Test_X)\n",
    "fpr, tpr, thresholds = roc_curve(Test_Y, y_score[:,1], pos_label=1)\n",
    "area = auc(fpr, tpr)\n",
    "\n",
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "sns.set_theme()\n",
    "\n",
    "plt.figure(figsize=(8,6))\n",
    "lw = 2\n",
    "plt.plot(fpr, tpr, color='darkorange',\n",
    "         lw=lw, label='ROC curve (area = %0.2f)' % area)\n",
    "plt.plot([0, 1], [0, 1], color='navy', linestyle='--')\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.05])\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.title('ROC score RandomForest with glove features')\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.savefig('ROC_RandomForest_glove.jpg')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "executionInfo": {
     "elapsed": 582,
     "status": "ok",
     "timestamp": 1622242749469,
     "user": {
      "displayName": "Ai Zhong",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14Gh2MWN-q_WG3Xb3YELMkck5bLErFAYBIGqZ_8lQ=s64",
      "userId": "11851498678274176197"
     },
     "user_tz": 420
    },
    "id": "5CH5jfVakRsd"
   },
   "outputs": [],
   "source": [
    "roc_df = pd.DataFrame([fpr, tpr,]).T\n",
    "roc_df.columns = ['rf_fpr', 'rf_tpr']\n",
    "roc_df['feat'] = 'glove'\n",
    "roc_df.to_csv('roc_glove_100k.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "8_P4XEsMM0ZJ"
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "iXuXtLBDM0by"
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "0rkPMbZKsRAo"
   },
   "source": [
    "# Skipping SVM with large sample data"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "D0pjpGcetN9m"
   },
   "source": [
    "# MLP\n",
    "\n",
    "### For Glove vectors do not need to do feature scaling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "uMyJqzwgtNHA"
   },
   "outputs": [],
   "source": [
    "from sklearn.neural_network import MLPClassifier\n",
    "# from sklearn.metrics import f1_score, accuracy_score\n",
    "\n",
    "mlp = MLPClassifier(hidden_layer_sizes=(150), random_state=1, max_iter=300).fit(Train_X, Train_Y)\n",
    "mlp_predict = mlp.predict(Test_X)\n",
    "mlp_acc = accuracy_score(Test_Y, mlp_predict)\n",
    "mlp_f1 = f1_score(Test_Y, mlp_predict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "F-JfgW1DtTDj"
   },
   "outputs": [],
   "source": [
    "pickle.dump(mlp, open('drive/MyDrive/Milestone2/trained_models/mlp_word2vec.pkl', 'wb'))\n",
    "reloaded_mlp = pickle.load(open('drive/MyDrive/Milestone2/trained_models/mlp_word2vec.pkl', 'rb'))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 120,
     "status": "ok",
     "timestamp": 1622098091106,
     "user": {
      "displayName": "Ai Zhong",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14Gh2MWN-q_WG3Xb3YELMkck5bLErFAYBIGqZ_8lQ=s64",
      "userId": "11851498678274176197"
     },
     "user_tz": 420
    },
    "id": "eJ-Ff81wh99R",
    "outputId": "f605af4e-2933-4b7a-ef36-3bb881fc36ab"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(64.05999999999999, 62.81043046357616)"
      ]
     },
     "execution_count": 13,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "mlp_predict = reloaded_mlp.predict(Test_X)\n",
    "mlp_acc = accuracy_score(Test_Y, mlp_predict)*100\n",
    "mlp_f1 = f1_score(Test_Y, mlp_predict)*100\n",
    "mlp_acc, mlp_f1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "executionInfo": {
     "elapsed": 318,
     "status": "ok",
     "timestamp": 1622098111519,
     "user": {
      "displayName": "Ai Zhong",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14Gh2MWN-q_WG3Xb3YELMkck5bLErFAYBIGqZ_8lQ=s64",
      "userId": "11851498678274176197"
     },
     "user_tz": 420
    },
    "id": "YLE5bd1XlHMT",
    "outputId": "85945370-edc4-437f-a5e7-7309296646bf"
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7930125"
      ]
     },
     "execution_count": 14,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "reloaded_mlp.score(Train_X, Train_Y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 410
    },
    "executionInfo": {
     "elapsed": 688,
     "status": "ok",
     "timestamp": 1622098341397,
     "user": {
      "displayName": "Ai Zhong",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14Gh2MWN-q_WG3Xb3YELMkck5bLErFAYBIGqZ_8lQ=s64",
      "userId": "11851498678274176197"
     },
     "user_tz": 420
    },
    "id": "JD3MibNolvmm",
    "outputId": "168982d5-7852-4425-b0fa-a4873c8a39c1"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "tags": []
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.metrics import roc_auc_score, roc_curve, auc\n",
    "\n",
    "y_score = reloaded_mlp.predict_proba(Test_X)\n",
    "fpr, tpr, thresholds = roc_curve(Test_Y, y_score[:,1], pos_label=1)\n",
    "area = auc(fpr, tpr)\n",
    "\n",
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "sns.set_theme()\n",
    "\n",
    "plt.figure(figsize=(8,6))\n",
    "lw = 2\n",
    "plt.plot(fpr, tpr, color='darkorange',\n",
    "         lw=lw, label='ROC curve (area = %0.2f)' % area)\n",
    "plt.plot([0, 1], [0, 1], color='navy', linestyle='--')\n",
    "plt.xlim([0.0, 1.0])\n",
    "plt.ylim([0.0, 1.05])\n",
    "plt.xlabel('False Positive Rate')\n",
    "plt.ylabel('True Positive Rate')\n",
    "plt.title('ROC score MLP with glove features')\n",
    "plt.legend(loc=\"lower right\")\n",
    "plt.savefig('ROC_MLP_glove.jpg')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "1Q6maMGAlvpv"
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "zfBlQLBrlvsw"
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "xrq3N_rRlvvV"
   },
   "source": [
    "# Summary"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "RkzKL2Ce2BpP"
   },
   "outputs": [],
   "source": [
    "score_df = pd.DataFrame([[mlp_acc, rf_acc],[mlp_f1, rf_f1]]).T\n",
    "score_df.columns=['accuracy', 'f1 score']\n",
    "score_df['model'] = ['MLP','RandomForest']\n",
    "score_df.to_csv('scores_10w_glove.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 111
    },
    "executionInfo": {
     "elapsed": 119,
     "status": "ok",
     "timestamp": 1622098158684,
     "user": {
      "displayName": "Ai Zhong",
      "photoUrl": "https://lh3.googleusercontent.com/a-/AOh14Gh2MWN-q_WG3Xb3YELMkck5bLErFAYBIGqZ_8lQ=s64",
      "userId": "11851498678274176197"
     },
     "user_tz": 420
    },
    "id": "bmnhiAUwlS-0",
    "outputId": "4021a99c-3b37-4d8b-f08a-9b8c72c76d53"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<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>accuracy</th>\n",
       "      <th>f1 score</th>\n",
       "      <th>model</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>64.060</td>\n",
       "      <td>62.810430</td>\n",
       "      <td>MLP</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>68.005</td>\n",
       "      <td>68.753357</td>\n",
       "      <td>RandomForest</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   accuracy   f1 score         model\n",
       "0    64.060  62.810430           MLP\n",
       "1    68.005  68.753357  RandomForest"
      ]
     },
     "execution_count": 16,
     "metadata": {
      "tags": []
     },
     "output_type": "execute_result"
    }
   ],
   "source": [
    "score_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "hav2OJxulgCc"
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "colab": {
   "authorship_tag": "ABX9TyPFN1s6pIcFQX6Xf/TkiuFm",
   "mount_file_id": "1VJIa6RjgjQWrS3St8tdN02DHS7VF4D25",
   "name": "Milestone2_glove_100k.ipynb",
   "provenance": [],
   "toc_visible": true
  },
  "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": "3.6.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}