Experiments.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import yaml\n",
    "import tqdm\n",
    "import time\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import sys\n",
    "import itertools\n",
    "from collections import defaultdict\n",
    "sys.path.append(\"./src/\")\n",
    "from preprocess.imdb_reviews import PreprocessIMDB\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "from torch.utils.data import TensorDataset, DataLoader"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "albumentations.core.composition.Compose"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import albumentations as A\n",
    "transform = A.Compose([\n",
    "     A.RandomCrop(width=256, height=256),\n",
    "     A.HorizontalFlip(p=0.5),\n",
    "     A.RandomBrightnessContrast(p=0.2),\n",
    " ])\n",
    "type(transform)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Random"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package stopwords to\n",
      "[nltk_data]     C:\\Users\\shanmukh\\AppData\\Roaming\\nltk_data...\n",
      "[nltk_data]   Package stopwords is already up-to-date!\n",
      "Traceback (most recent call last):\n",
      "  File \"d:\\Learning\\NLP\\Projects\\scratch_nlp\\src\\main.py\", line 62, in <module>\n",
      "    main()\n",
      "  File \"d:\\Learning\\NLP\\Projects\\scratch_nlp\\src\\main.py\", line 58, in main\n",
      "    algo.run()\n",
      "  File \"d:\\Learning\\NLP\\Projects\\scratch_nlp\\src\\core\\bow\\bow.py\", line 16, in run\n",
      "    self.fit_transform()\n",
      "  File \"d:\\Learning\\NLP\\Projects\\scratch_nlp\\src\\core\\bow\\bow.py\", line 30, in fit_transform\n",
      "    self.fit(text_ls, y)\n",
      "  File \"d:\\Learning\\NLP\\Projects\\scratch_nlp\\src\\core\\bow\\bow.py\", line 23, in fit\n",
      "    for text in text_ls:\n",
      "TypeError: 'NoneType' object is not iterable\n"
     ]
    }
   ],
   "source": [
    "!python src/main.py --config_path \"D:\\Learning\\NLP\\Projects\\scratch_nlp\\configs\\bert.yaml\" --algo \"BERT\" --log_folder \"D:\\Learning\\NLP\\Projects\\scratch_nlp\\output\\bert\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "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>0</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>A</th>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>B</th>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   0\n",
       "A  2\n",
       "B  4"
      ]
     },
     "execution_count": 1,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "\n",
    "dict_ = {\n",
    "    \"A\": 2,\n",
    "    \"B\": 4\n",
    "}\n",
    "df = pd.DataFrame(dict_, index=[0]).T\n",
    "df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(600, 9834)"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import numpy as np\n",
    "\n",
    "arr = np.load(\"./output/bow/Text Vector.npy\")\n",
    "arr.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['e:\\\\anaconda\\\\envs\\\\nlp_env',\n",
       " 'C:\\\\Program Files\\\\dotnet\\\\',\n",
       " 'C:\\\\Program Files\\\\Git\\\\cmd',\n",
       " 'E:\\\\anaconda\\\\Scripts',\n",
       " 'E:\\\\anaconda\\\\Library',\n",
       " 'E:\\\\anaconda\\\\Library\\\\bin',\n",
       " 'E:\\\\anaconda\\\\Library\\\\mingw-w64\\\\bin',\n",
       " 'C:\\\\Users\\\\shanmukh\\\\AppData\\\\Local\\\\Programs\\\\Python\\\\Python38-32\\\\Scripts\\\\',\n",
       " 'C:\\\\Users\\\\shanmukh\\\\AppData\\\\Local\\\\Programs\\\\Python\\\\Python38-32\\\\',\n",
       " 'C:\\\\Users\\\\shanmukh\\\\AppData\\\\Local\\\\Microsoft\\\\WindowsApps',\n",
       " 'C:\\\\Users\\\\shanmukh\\\\AppData\\\\Local\\\\GitHubDesktop\\\\bin',\n",
       " 'C:\\\\Program Files\\\\mingw-w64\\\\x86_64-8.1.0-posix-seh-rt_v6-rev0\\\\mingw64\\\\bin',\n",
       " 'C:\\\\Users\\\\shanmukh\\\\AppData\\\\Local\\\\Microsoft\\\\WindowsApps',\n",
       " 'D:\\\\heroku\\\\bin',\n",
       " 'E:\\\\vscode\\\\Microsoft VS Code\\\\bin',\n",
       " 'E:\\\\PyCharm\\\\PyCharm Community Edition 2021.2.3\\\\bin',\n",
       " '',\n",
       " 'E:\\\\anaconda',\n",
       " 'E:\\\\anaconda\\\\Library\\\\bin',\n",
       " 'E:\\\\anaconda\\\\Library\\\\bin',\n",
       " 'E:\\\\apache-jena-4.7.0\\\\bat',\n",
       " 'C:\\\\Program Files\\\\Java\\\\jdk-19\\\\bin',\n",
       " 'C:\\\\Program Files\\\\Common Files\\\\Oracle\\\\Java\\\\javapath',\n",
       " 'C:\\\\Users\\\\shanmukh\\\\Downloads\\\\spark-3.3.2-bin-hadoop3\\\\spark-3.3.2-bin-hadoop3',\n",
       " '',\n",
       " 'C:\\\\Program Files\\\\dotnet\\\\',\n",
       " 'C:\\\\Program Files\\\\Git\\\\cmd',\n",
       " 'E:\\\\anaconda\\\\Scripts',\n",
       " 'E:\\\\anaconda\\\\Library',\n",
       " 'E:\\\\anaconda\\\\Library\\\\bin',\n",
       " 'E:\\\\anaconda\\\\Library\\\\mingw-w64\\\\bin',\n",
       " 'C:\\\\Users\\\\shanmukh\\\\AppData\\\\Local\\\\Programs\\\\Python\\\\Python38-32\\\\Scripts\\\\',\n",
       " 'C:\\\\Users\\\\shanmukh\\\\AppData\\\\Local\\\\Programs\\\\Python\\\\Python38-32\\\\',\n",
       " 'C:\\\\Users\\\\shanmukh\\\\AppData\\\\Local\\\\Microsoft\\\\WindowsApps',\n",
       " 'C:\\\\Users\\\\shanmukh\\\\AppData\\\\Local\\\\GitHubDesktop\\\\bin',\n",
       " 'C:\\\\Program Files\\\\mingw-w64\\\\x86_64-8.1.0-posix-seh-rt_v6-rev0\\\\mingw64\\\\bin',\n",
       " 'C:\\\\Users\\\\shanmukh\\\\AppData\\\\Local\\\\Microsoft\\\\WindowsApps',\n",
       " 'D:\\\\heroku\\\\bin',\n",
       " 'E:\\\\vscode\\\\Microsoft VS Code\\\\bin',\n",
       " 'E:\\\\PyCharm\\\\PyCharm Community Edition 2021.2.3\\\\bin',\n",
       " '',\n",
       " 'E:\\\\anaconda',\n",
       " 'E:\\\\anaconda\\\\Library\\\\bin',\n",
       " 'E:\\\\anaconda\\\\Library\\\\bin',\n",
       " 'E:\\\\apache-jena-4.7.0\\\\bat',\n",
       " 'C:\\\\Program Files\\\\Java\\\\jdk-19\\\\bin',\n",
       " 'C:\\\\Program Files\\\\Common Files\\\\Oracle\\\\Java\\\\javapath',\n",
       " 'C:\\\\Users\\\\shanmukh\\\\Downloads\\\\spark-3.3.2-bin-hadoop3\\\\spark-3.3.2-bin-hadoop3',\n",
       " '']"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import os\n",
    "\n",
    "x = os.environ['PATH']\n",
    "x.split(\";\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "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>Word</th>\n",
       "      <th>Frequency</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>better whatev person love one</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>gate treasur hunter oh excus</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>atmospher place jump scare also</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>deliv like joke either riley</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>forsaken desert giant forest full</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                Word  Frequency\n",
       "0      better whatev person love one          1\n",
       "1       gate treasur hunter oh excus          1\n",
       "2    atmospher place jump scare also          1\n",
       "3       deliv like joke either riley          1\n",
       "4  forsaken desert giant forest full          1"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "\n",
    "df = pd.read_csv(r\"D:\\Learning\\NLP\\Projects\\scratch_nlp\\output\\ngram\\Vocab Frequency.csv\")\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "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>Word</th>\n",
       "      <th>Frequency</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>el día de la bestia</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>anoth episod like one brak</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>atmospher place jump scare also</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>deliv like joke either riley</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>forsaken desert giant forest full</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                Word  Frequency\n",
       "0                el día de la bestia          2\n",
       "1         anoth episod like one brak          2\n",
       "2    atmospher place jump scare also          1\n",
       "3       deliv like joke either riley          1\n",
       "4  forsaken desert giant forest full          1"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.sort_values(by=\"Frequency\", ascending=False, ignore_index=True, inplace=True)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['Word_1', 'Word_2', 'Word_3', 'Word_4', 'Word_5']\n"
     ]
    },
    {
     "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>Word</th>\n",
       "      <th>Frequency</th>\n",
       "      <th>Word_1</th>\n",
       "      <th>Word_2</th>\n",
       "      <th>Word_3</th>\n",
       "      <th>Word_4</th>\n",
       "      <th>Word_5</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>el día de la bestia</td>\n",
       "      <td>2</td>\n",
       "      <td>el</td>\n",
       "      <td>día</td>\n",
       "      <td>de</td>\n",
       "      <td>la</td>\n",
       "      <td>bestia</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>anoth episod like one brak</td>\n",
       "      <td>2</td>\n",
       "      <td>anoth</td>\n",
       "      <td>episod</td>\n",
       "      <td>like</td>\n",
       "      <td>one</td>\n",
       "      <td>brak</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>atmospher place jump scare also</td>\n",
       "      <td>1</td>\n",
       "      <td>atmospher</td>\n",
       "      <td>place</td>\n",
       "      <td>jump</td>\n",
       "      <td>scare</td>\n",
       "      <td>also</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>deliv like joke either riley</td>\n",
       "      <td>1</td>\n",
       "      <td>deliv</td>\n",
       "      <td>like</td>\n",
       "      <td>joke</td>\n",
       "      <td>either</td>\n",
       "      <td>riley</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>forsaken desert giant forest full</td>\n",
       "      <td>1</td>\n",
       "      <td>forsaken</td>\n",
       "      <td>desert</td>\n",
       "      <td>giant</td>\n",
       "      <td>forest</td>\n",
       "      <td>full</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                Word  Frequency     Word_1  Word_2 Word_3  \\\n",
       "0                el día de la bestia          2         el     día     de   \n",
       "1         anoth episod like one brak          2      anoth  episod   like   \n",
       "2    atmospher place jump scare also          1  atmospher   place   jump   \n",
       "3       deliv like joke either riley          1      deliv    like   joke   \n",
       "4  forsaken desert giant forest full          1   forsaken  desert  giant   \n",
       "\n",
       "   Word_4  Word_5  \n",
       "0      la  bestia  \n",
       "1     one    brak  \n",
       "2   scare    also  \n",
       "3  either   riley  \n",
       "4  forest    full  "
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "n = 5\n",
    "word_cols = [f\"Word_{i+1}\" for i in range(n)]\n",
    "print(word_cols)\n",
    "\n",
    "df[word_cols] = df[\"Word\"].str.split(\" \", expand=True)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "new_df = df.iloc[:10]\n",
    "\n",
    "def percentage_growth(l):\n",
    "    s = 0\n",
    "    res = [0]\n",
    "    for i in range(len(l)-1):\n",
    "        s += l.iloc[i]\n",
    "        res.append(s / sum(l))\n",
    "    return res\n",
    "\n",
    "cmap = plt.get_cmap(\"rainbow\")\n",
    "\n",
    "for i, col in enumerate(word_cols):\n",
    "    radius=i+2\n",
    "    width=1\n",
    "    frame = new_df.groupby(word_cols[:i+1])[\"Frequency\"].sum()\n",
    "    colors = cmap(percentage_growth(frame))\n",
    "    labels = [x[-1] if isinstance(x, tuple) else x for x in frame.index.to_numpy()]\n",
    "    plt.pie(frame, labels=labels, colors=colors, radius=radius, wedgeprops=dict(width=width, edgecolor='w'), textprops=dict(size=20), labeldistance=0.8 + i/60)    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import nltk\n",
    "nltk.download(\"treebank\")\n",
    "data = list(nltk.corpus.treebank.tagged_sents())\n",
    "data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.plotly.v1+json": {
       "config": {
        "plotlyServerURL": "https://plot.ly"
       },
       "data": [
        {
         "hovertemplate": "x=%{x}<br>y=%{y}<br>text=%{text}<extra></extra>",
         "legendgroup": "",
         "marker": {
          "color": "#636efa",
          "symbol": "circle"
         },
         "mode": "markers+text",
         "name": "",
         "orientation": "v",
         "showlegend": false,
         "text": [
          ".",
          "EX",
          "VBZ",
          "RP",
          "VBN",
          "NNS",
          "VBG",
          "IN",
          "NNP",
          "CC",
          "VB",
          "WDT",
          "RBR",
          "PRP$",
          "WP",
          ",",
          "VBP",
          "CD",
          "TO",
          "MD",
          "JJ",
          "JJR",
          "RB",
          "PRP",
          "VBD",
          "POS",
          "NN",
          "JJS",
          "DT",
          "``",
          "-NONE-",
          "''",
          ":"
         ],
         "textposition": "bottom right",
         "type": "scatter",
         "x": [
          1.417842149734497,
          1.2521332502365112,
          1.521378517150879,
          1.0054429769515991,
          1.6221271753311157,
          1.6223433017730713,
          1.6188766956329346,
          1.584828495979309,
          1.6231746673583984,
          1.8485268354415894,
          1.613977074623108,
          1.896132469177246,
          1.3310317993164062,
          1.8603366613388062,
          2.1881279945373535,
          1.1122279167175293,
          1.609326958656311,
          1.6221072673797607,
          1.0108835697174072,
          1.57920503616333,
          1.620813250541687,
          1.8424478769302368,
          1.622568130493164,
          1.604093313217163,
          1.621645450592041,
          1.9139959812164307,
          1.6229921579360962,
          1.5940717458724976,
          1.6028313636779785,
          1.562522530555725,
          1.6229703426361084,
          2.1944563388824463,
          2.2573070526123047
         ],
         "xaxis": "x",
         "y": [
          2.6740496158599854,
          1.5386372804641724,
          2.0951223373413086,
          2.2090647220611572,
          2.068803310394287,
          2.068403720855713,
          2.0701708793640137,
          2.0795867443084717,
          2.068553924560547,
          1.7648667097091675,
          2.072434186935425,
          2.44883131980896,
          1.7520054578781128,
          2.0009896755218506,
          1.7721447944641113,
          2.4553325176239014,
          2.072134017944336,
          2.0691885948181152,
          1.8830689191818237,
          2.4559733867645264,
          2.065619945526123,
          2.6745285987854004,
          2.067958354949951,
          2.073824405670166,
          2.0683000087738037,
          1.498351812362671,
          2.069370985031128,
          2.0768625736236572,
          2.074612617492676,
          1.4339810609817505,
          2.069317579269409,
          2.360992670059204,
          2.028695821762085
         ],
         "yaxis": "y"
        }
       ],
       "layout": {
        "legend": {
         "tracegroupgap": 0
        },
        "margin": {
         "t": 60
        },
        "template": {
         "data": {
          "bar": [
           {
            "error_x": {
             "color": "#2a3f5f"
            },
            "error_y": {
             "color": "#2a3f5f"
            },
            "marker": {
             "line": {
              "color": "#E5ECF6",
              "width": 0.5
             },
             "pattern": {
              "fillmode": "overlay",
              "size": 10,
              "solidity": 0.2
             }
            },
            "type": "bar"
           }
          ],
          "barpolar": [
           {
            "marker": {
             "line": {
              "color": "#E5ECF6",
              "width": 0.5
             },
             "pattern": {
              "fillmode": "overlay",
              "size": 10,
              "solidity": 0.2
             }
            },
            "type": "barpolar"
           }
          ],
          "carpet": [
           {
            "aaxis": {
             "endlinecolor": "#2a3f5f",
             "gridcolor": "white",
             "linecolor": "white",
             "minorgridcolor": "white",
             "startlinecolor": "#2a3f5f"
            },
            "baxis": {
             "endlinecolor": "#2a3f5f",
             "gridcolor": "white",
             "linecolor": "white",
             "minorgridcolor": "white",
             "startlinecolor": "#2a3f5f"
            },
            "type": "carpet"
           }
          ],
          "choropleth": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "type": "choropleth"
           }
          ],
          "contour": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "contour"
           }
          ],
          "contourcarpet": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "type": "contourcarpet"
           }
          ],
          "heatmap": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "heatmap"
           }
          ],
          "heatmapgl": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "heatmapgl"
           }
          ],
          "histogram": [
           {
            "marker": {
             "pattern": {
              "fillmode": "overlay",
              "size": 10,
              "solidity": 0.2
             }
            },
            "type": "histogram"
           }
          ],
          "histogram2d": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "histogram2d"
           }
          ],
          "histogram2dcontour": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "histogram2dcontour"
           }
          ],
          "mesh3d": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "type": "mesh3d"
           }
          ],
          "parcoords": [
           {
            "line": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "parcoords"
           }
          ],
          "pie": [
           {
            "automargin": true,
            "type": "pie"
           }
          ],
          "scatter": [
           {
            "fillpattern": {
             "fillmode": "overlay",
             "size": 10,
             "solidity": 0.2
            },
            "type": "scatter"
           }
          ],
          "scatter3d": [
           {
            "line": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatter3d"
           }
          ],
          "scattercarpet": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattercarpet"
           }
          ],
          "scattergeo": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattergeo"
           }
          ],
          "scattergl": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattergl"
           }
          ],
          "scattermapbox": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattermapbox"
           }
          ],
          "scatterpolar": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatterpolar"
           }
          ],
          "scatterpolargl": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatterpolargl"
           }
          ],
          "scatterternary": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatterternary"
           }
          ],
          "surface": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "surface"
           }
          ],
          "table": [
           {
            "cells": {
             "fill": {
              "color": "#EBF0F8"
             },
             "line": {
              "color": "white"
             }
            },
            "header": {
             "fill": {
              "color": "#C8D4E3"
             },
             "line": {
              "color": "white"
             }
            },
            "type": "table"
           }
          ]
         },
         "layout": {
          "annotationdefaults": {
           "arrowcolor": "#2a3f5f",
           "arrowhead": 0,
           "arrowwidth": 1
          },
          "autotypenumbers": "strict",
          "coloraxis": {
           "colorbar": {
            "outlinewidth": 0,
            "ticks": ""
           }
          },
          "colorscale": {
           "diverging": [
            [
             0,
             "#8e0152"
            ],
            [
             0.1,
             "#c51b7d"
            ],
            [
             0.2,
             "#de77ae"
            ],
            [
             0.3,
             "#f1b6da"
            ],
            [
             0.4,
             "#fde0ef"
            ],
            [
             0.5,
             "#f7f7f7"
            ],
            [
             0.6,
             "#e6f5d0"
            ],
            [
             0.7,
             "#b8e186"
            ],
            [
             0.8,
             "#7fbc41"
            ],
            [
             0.9,
             "#4d9221"
            ],
            [
             1,
             "#276419"
            ]
           ],
           "sequential": [
            [
             0,
             "#0d0887"
            ],
            [
             0.1111111111111111,
             "#46039f"
            ],
            [
             0.2222222222222222,
             "#7201a8"
            ],
            [
             0.3333333333333333,
             "#9c179e"
            ],
            [
             0.4444444444444444,
             "#bd3786"
            ],
            [
             0.5555555555555556,
             "#d8576b"
            ],
            [
             0.6666666666666666,
             "#ed7953"
            ],
            [
             0.7777777777777778,
             "#fb9f3a"
            ],
            [
             0.8888888888888888,
             "#fdca26"
            ],
            [
             1,
             "#f0f921"
            ]
           ],
           "sequentialminus": [
            [
             0,
             "#0d0887"
            ],
            [
             0.1111111111111111,
             "#46039f"
            ],
            [
             0.2222222222222222,
             "#7201a8"
            ],
            [
             0.3333333333333333,
             "#9c179e"
            ],
            [
             0.4444444444444444,
             "#bd3786"
            ],
            [
             0.5555555555555556,
             "#d8576b"
            ],
            [
             0.6666666666666666,
             "#ed7953"
            ],
            [
             0.7777777777777778,
             "#fb9f3a"
            ],
            [
             0.8888888888888888,
             "#fdca26"
            ],
            [
             1,
             "#f0f921"
            ]
           ]
          },
          "colorway": [
           "#636efa",
           "#EF553B",
           "#00cc96",
           "#ab63fa",
           "#FFA15A",
           "#19d3f3",
           "#FF6692",
           "#B6E880",
           "#FF97FF",
           "#FECB52"
          ],
          "font": {
           "color": "#2a3f5f"
          },
          "geo": {
           "bgcolor": "white",
           "lakecolor": "white",
           "landcolor": "#E5ECF6",
           "showlakes": true,
           "showland": true,
           "subunitcolor": "white"
          },
          "hoverlabel": {
           "align": "left"
          },
          "hovermode": "closest",
          "mapbox": {
           "style": "light"
          },
          "paper_bgcolor": "white",
          "plot_bgcolor": "#E5ECF6",
          "polar": {
           "angularaxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           },
           "bgcolor": "#E5ECF6",
           "radialaxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           }
          },
          "scene": {
           "xaxis": {
            "backgroundcolor": "#E5ECF6",
            "gridcolor": "white",
            "gridwidth": 2,
            "linecolor": "white",
            "showbackground": true,
            "ticks": "",
            "zerolinecolor": "white"
           },
           "yaxis": {
            "backgroundcolor": "#E5ECF6",
            "gridcolor": "white",
            "gridwidth": 2,
            "linecolor": "white",
            "showbackground": true,
            "ticks": "",
            "zerolinecolor": "white"
           },
           "zaxis": {
            "backgroundcolor": "#E5ECF6",
            "gridcolor": "white",
            "gridwidth": 2,
            "linecolor": "white",
            "showbackground": true,
            "ticks": "",
            "zerolinecolor": "white"
           }
          },
          "shapedefaults": {
           "line": {
            "color": "#2a3f5f"
           }
          },
          "ternary": {
           "aaxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           },
           "baxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           },
           "bgcolor": "#E5ECF6",
           "caxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           }
          },
          "title": {
           "x": 0.05
          },
          "xaxis": {
           "automargin": true,
           "gridcolor": "white",
           "linecolor": "white",
           "ticks": "",
           "title": {
            "standoff": 15
           },
           "zerolinecolor": "white",
           "zerolinewidth": 2
          },
          "yaxis": {
           "automargin": true,
           "gridcolor": "white",
           "linecolor": "white",
           "ticks": "",
           "title": {
            "standoff": 15
           },
           "zerolinecolor": "white",
           "zerolinewidth": 2
          }
         }
        },
        "xaxis": {
         "anchor": "y",
         "domain": [
          0,
          1
         ],
         "title": {
          "text": "x"
         }
        },
        "yaxis": {
         "anchor": "x",
         "domain": [
          0,
          1
         ],
         "title": {
          "text": "y"
         }
        }
       }
      }
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import plotly.express as px\n",
    "from sklearn.manifold import TSNE\n",
    "\n",
    "arr_df = pd.read_csv(\"./output/hmm/Emission Matrix.csv\")\n",
    "cols = list(arr_df.columns)\n",
    "vocab = cols[1:]\n",
    "pos = list(arr_df[cols[0]])\n",
    "\n",
    "tsne = TSNE(n_components=2)\n",
    "arr_tsne = tsne.fit_transform(np.array(arr_df[vocab]))\n",
    "\n",
    "fig = px.scatter(x=arr_tsne[:,0], y=arr_tsne[:,1],text=pos)\n",
    "fig.update_traces(textposition=\"bottom right\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "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>Unnamed: 0</th>\n",
       "      <th>appeal</th>\n",
       "      <th>off</th>\n",
       "      <th>62.1</th>\n",
       "      <th>blow</th>\n",
       "      <th>President</th>\n",
       "      <th>damn</th>\n",
       "      <th>professor</th>\n",
       "      <th>adds</th>\n",
       "      <th>Community</th>\n",
       "      <th>...</th>\n",
       "      <th>Otero</th>\n",
       "      <th>require</th>\n",
       "      <th>paltry</th>\n",
       "      <th>impede</th>\n",
       "      <th>Primerica</th>\n",
       "      <th>electric</th>\n",
       "      <th>lease</th>\n",
       "      <th>Fees</th>\n",
       "      <th>endless</th>\n",
       "      <th>physical</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>NNPS</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>RB</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.001335</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.00089</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>VBD</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>RP</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.110429</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>JJR</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 10966 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "  Unnamed: 0  appeal       off  62.1  blow  President     damn  professor  \\\n",
       "0       NNPS     0.0  0.000000   0.0   0.0        0.0  0.00000        0.0   \n",
       "1         RB     0.0  0.001335   0.0   0.0        0.0  0.00089        0.0   \n",
       "2        VBD     0.0  0.000000   0.0   0.0        0.0  0.00000        0.0   \n",
       "3         RP     0.0  0.110429   0.0   0.0        0.0  0.00000        0.0   \n",
       "4        JJR     0.0  0.000000   0.0   0.0        0.0  0.00000        0.0   \n",
       "\n",
       "   adds  Community  ...  Otero  require  paltry  impede  Primerica  electric  \\\n",
       "0   0.0        0.0  ...    0.0      0.0     0.0     0.0        0.0       0.0   \n",
       "1   0.0        0.0  ...    0.0      0.0     0.0     0.0        0.0       0.0   \n",
       "2   0.0        0.0  ...    0.0      0.0     0.0     0.0        0.0       0.0   \n",
       "3   0.0        0.0  ...    0.0      0.0     0.0     0.0        0.0       0.0   \n",
       "4   0.0        0.0  ...    0.0      0.0     0.0     0.0        0.0       0.0   \n",
       "\n",
       "   lease  Fees  endless  physical  \n",
       "0    0.0   0.0      0.0       0.0  \n",
       "1    0.0   0.0      0.0       0.0  \n",
       "2    0.0   0.0      0.0       0.0  \n",
       "3    0.0   0.0      0.0       0.0  \n",
       "4    0.0   0.0      0.0       0.0  \n",
       "\n",
       "[5 rows x 10966 columns]"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "arr_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([1, 2, 3, 4]), array([1, 1, 1, 1]))"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import numpy as np\n",
    "\n",
    "arr = np.array([2,1,3,4])\n",
    "np.unique(arr, return_counts=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# Create a DataFrame\n",
    "df = pd.DataFrame({'A': [1, 2, 3], 'B': [4, 5, 6]})\n",
    "\n",
    "# Create a figure and axes\n",
    "fig, ax = plt.subplots()\n",
    "\n",
    "# Hide axes\n",
    "ax.axis('off')\n",
    "\n",
    "# Create table\n",
    "table = ax.table(cellText=df.values, colLabels=df.columns, loc='center')\n",
    "\n",
    "# Adjust layout\n",
    "table.auto_set_font_size(False)\n",
    "table.set_fontsize(10)\n",
    "\n",
    "# Show the plot\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Create Sample Data Folder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import shutil\n",
    "import json\n",
    "\n",
    "import pandas as pd\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "### eng2tel\n",
    "\n",
    "with open(\"data\\eng2tel\\english_telugu_data.txt\", \"r\", encoding=\"utf-8\") as f:\n",
    "    lines = f.readlines()\n",
    "\n",
    "with open(\"sample_data\\eng2tel\\english_telugu_data.txt\", \"w\", encoding=\"utf-8\") as wf:\n",
    "    wf.writelines(lines[:1000])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(\"data\\eng2tel\\english_telugu_data.txt\", \"r\", encoding=\"utf-8\") as f:\n",
    "    lines = f.readlines()\n",
    "\n",
    "with open(\"sample_data\\eng2tel\\english_telugu_data.txt\", \"w\", encoding=\"utf-8\") as wf:\n",
    "    wf.writelines(lines[:1000])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "### flickr_8k\n",
    "\n",
    "with open(r\"data\\flickr_8k\\captions.txt\", \"r\") as f:\n",
    "    lines = f.readlines()\n",
    "\n",
    "rand_ids = np.random.choice(np.arange(len(lines)-1)+1, 96, replace=False)\n",
    "new_lines = [lines[0]]\n",
    "\n",
    "for id_ in rand_ids:\n",
    "    line = lines[id_]\n",
    "    new_lines.append(line)\n",
    "    im_name = line.split(\",\")[0]\n",
    "    src = os.path.join(r\"data\\flickr_8k\\Images\", im_name)\n",
    "    dst = os.path.join(r\"sample_data\\flickr_8k\\Images\", im_name)\n",
    "    shutil.copy(src, dst)\n",
    "\n",
    "with open(r\"sample_data\\flickr_8k\\captions.txt\", \"w\") as wf:\n",
    "    wf.writelines(new_lines)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "### genius_lyrics\n",
    "\n",
    "df = pd.read_csv(r\"data\\genius_lyrics\\english_lyrics.csv\")\n",
    "rand_ids = np.random.choice(np.arange(len(df)-1)+1, 256)\n",
    "new_df = df.iloc[rand_ids]\n",
    "new_df.to_csv(r\"sample_data\\genius_lyrics\\english_lyrics.csv\", index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "### wiki_en\n",
    "\n",
    "df = pd.read_csv(r\"data\\wiki_en\\wiki_en.csv\")\n",
    "rand_ids = np.random.choice(np.arange(len(df)-1)+1, 256)\n",
    "new_df = df.iloc[rand_ids]\n",
    "new_df.to_csv(r\"sample_data\\wiki_en\\wiki_en.csv\", index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "### squad\n",
    "\n",
    "with open(r\"data\\squad\\squad_v1.json\") as f:\n",
    "    squad_dict = json.load(f)\n",
    "\n",
    "rand_ids = np.random.choice(len(squad_dict[\"data\"]), 8, replace=False)\n",
    "\n",
    "new_dict = {\"version\": squad_dict[\"version\"]}\n",
    "new_dict[\"data\"] = []\n",
    "for id_ in rand_ids:\n",
    "    new_dict[\"data\"].append(squad_dict[\"data\"][id_])\n",
    "\n",
    "with open(r\"sample_data\\squad\\squad_v1.json\", \"w\") as wf:\n",
    "    json.dump(new_dict, wf)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Word2Vec"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "config_path = \"./configs/word2vec.yaml\"\n",
    "with open(config_path, \"r\") as stream:\n",
    "    config_dict = yaml.safe_load(stream)\n",
    "\n",
    "cbow = Word2VecDataset(config_dict)\n",
    "l_cxt, r_cxt, l_lbl, r_lbl = cbow.make_pairs()\n",
    "\n",
    "print(cbow.vocab_freq)\n",
    "print(l_cxt)\n",
    "print(r_cxt)\n",
    "print(l_lbl)\n",
    "print(r_lbl)\n",
    "\n",
    "train_left_loader, train_right_loader, val_left_loader, val_right_loader = create_dataloader(l_cxt, r_cxt, l_lbl, r_lbl, 0.2, 8, 2023)\n",
    "\n",
    "# for batch in iter(train_left_loader):\n",
    "#     l_c, l_l = batch\n",
    "#     print(l_c)\n",
    "#     print(l_l)\n",
    "#     break\n",
    "\n",
    "# for batch in iter(train_right_loader):\n",
    "#     r_c, r_l = batch\n",
    "#     print(r_c)\n",
    "#     print(r_l)\n",
    "#     break"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-----------Epoch 1/10-----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: 100%|██████████| 206/206 [00:26<00:00,  7.86it/s]\n",
      "Validation: 100%|██████████| 51/51 [00:19<00:00,  2.56it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Loss : 5.19072961807251 - Val Loss : 2.880483865737915\n",
      "Validation Loss improved from inf to 2.880483865737915\n",
      "-----------Epoch 2/10-----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: 100%|██████████| 206/206 [00:19<00:00, 10.36it/s]\n",
      "Validation: 100%|██████████| 51/51 [00:17<00:00,  3.00it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Loss : 2.8874399662017822 - Val Loss : 2.194394826889038\n",
      "Validation Loss improved from 2.880483865737915 to 2.194394826889038\n",
      "-----------Epoch 3/10-----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: 100%|██████████| 206/206 [00:17<00:00, 11.60it/s]\n",
      "Validation: 100%|██████████| 51/51 [00:17<00:00,  2.95it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Loss : 2.2812020778656006 - Val Loss : 1.9462478160858154\n",
      "Validation Loss improved from 2.194394826889038 to 1.9462478160858154\n",
      "-----------Epoch 4/10-----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: 100%|██████████| 206/206 [00:17<00:00, 11.60it/s]\n",
      "Validation: 100%|██████████| 51/51 [00:16<00:00,  3.05it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Loss : 1.9103882312774658 - Val Loss : 1.7780832052230835\n",
      "Validation Loss improved from 1.9462478160858154 to 1.7780832052230835\n",
      "-----------Epoch 5/10-----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: 100%|██████████| 206/206 [00:19<00:00, 10.31it/s]\n",
      "Validation: 100%|██████████| 51/51 [00:15<00:00,  3.28it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Loss : 1.6663691997528076 - Val Loss : 1.6299480199813843\n",
      "Validation Loss improved from 1.7780832052230835 to 1.6299480199813843\n",
      "-----------Epoch 6/10-----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: 100%|██████████| 206/206 [00:16<00:00, 12.51it/s]\n",
      "Validation: 100%|██████████| 51/51 [00:14<00:00,  3.46it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Loss : 1.4816888570785522 - Val Loss : 1.539669156074524\n",
      "Validation Loss improved from 1.6299480199813843 to 1.539669156074524\n",
      "-----------Epoch 7/10-----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: 100%|██████████| 206/206 [00:15<00:00, 13.27it/s]\n",
      "Validation: 100%|██████████| 51/51 [00:14<00:00,  3.44it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Loss : 1.3658875226974487 - Val Loss : 1.4765939712524414\n",
      "Validation Loss improved from 1.539669156074524 to 1.4765939712524414\n",
      "-----------Epoch 8/10-----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: 100%|██████████| 206/206 [00:14<00:00, 14.01it/s]\n",
      "Validation: 100%|██████████| 51/51 [00:14<00:00,  3.60it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Loss : 1.2907533645629883 - Val Loss : 1.4481770992279053\n",
      "Validation Loss improved from 1.4765939712524414 to 1.4481770992279053\n",
      "-----------Epoch 9/10-----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: 100%|██████████| 206/206 [00:15<00:00, 13.37it/s]\n",
      "Validation: 100%|██████████| 51/51 [00:14<00:00,  3.52it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Loss : 1.2339324951171875 - Val Loss : 1.4516034126281738\n",
      "Validation loss didn't improve from 1.4481770992279053\n",
      "-----------Epoch 10/10-----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: 100%|██████████| 206/206 [00:16<00:00, 12.78it/s]\n",
      "Validation: 100%|██████████| 51/51 [00:14<00:00,  3.47it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Loss : 1.1893771886825562 - Val Loss : 1.4282432794570923\n",
      "Validation Loss improved from 1.4481770992279053 to 1.4282432794570923\n",
      "Training completed in 0h 5m 55s\n",
      "Best Val RMSE: 1.4282432794570923\n"
     ]
    }
   ],
   "source": [
    "model = CBOW(config_dict)\n",
    "\n",
    "train_loader = (train_left_loader, train_right_loader)\n",
    "val_loader = (val_left_loader, val_right_loader)\n",
    "optim = torch.optim.Adam(model.parameters(), lr=config_dict[\"train\"][\"lr\"])\n",
    "history = fit_model(model, optim, train_loader, val_loader, config_dict)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\anaconda\\envs\\nlp_env\\lib\\site-packages\\joblib\\externals\\loky\\backend\\context.py:136: UserWarning: Could not find the number of physical cores for the following reason:\n",
      "[WinError 2] The system cannot find the file specified\n",
      "Returning the number of logical cores instead. You can silence this warning by setting LOKY_MAX_CPU_COUNT to the number of cores you want to use.\n",
      "  warnings.warn(\n",
      "  File \"e:\\anaconda\\envs\\nlp_env\\lib\\site-packages\\joblib\\externals\\loky\\backend\\context.py\", line 257, in _count_physical_cores\n",
      "    cpu_info = subprocess.run(\n",
      "  File \"e:\\anaconda\\envs\\nlp_env\\lib\\subprocess.py\", line 501, in run\n",
      "    with Popen(*popenargs, **kwargs) as process:\n",
      "  File \"e:\\anaconda\\envs\\nlp_env\\lib\\subprocess.py\", line 947, in __init__\n",
      "    self._execute_child(args, executable, preexec_fn, close_fds,\n",
      "  File \"e:\\anaconda\\envs\\nlp_env\\lib\\subprocess.py\", line 1416, in _execute_child\n",
      "    hp, ht, pid, tid = _winapi.CreateProcess(executable, args,\n"
     ]
    },
    {
     "data": {
      "application/vnd.plotly.v1+json": {
       "config": {
        "plotlyServerURL": "https://plot.ly"
       },
       "data": [
        {
         "hovertemplate": "X=%{x}<br>Y=%{y}<br>Z=%{z}<br>Word=%{text}<extra></extra>",
         "legendgroup": "",
         "marker": {
          "color": "#636efa",
          "symbol": "circle"
         },
         "mode": "markers+text",
         "name": "",
         "scene": "scene",
         "showlegend": false,
         "text": [
          "br",
          "movi",
          "film",
          "tom",
          "time",
          "like",
          "end",
          "charact",
          "one",
          "seen",
          "much",
          "go",
          "see",
          "back",
          "know",
          "well",
          "good",
          "think",
          "great",
          "hous",
          "come",
          "get",
          "cecilia",
          "goe",
          "hope",
          "cut",
          "look",
          "play",
          "two",
          "littl",
          "<UNK>"
         ],
         "type": "scatter3d",
         "x": [
          108.84769439697266,
          41.907989501953125,
          -26.60728645324707,
          -34.34479904174805,
          -138.39060974121094,
          76.93354797363281,
          117.14276123046875,
          18.68572235107422,
          52.014129638671875,
          91.37653350830078,
          -30.865123748779297,
          -7.416396141052246,
          -62.26999282836914,
          54.28514862060547,
          -89.03902435302734,
          -23.826425552368164,
          -98.29833984375,
          23.955326080322266,
          99.63731384277344,
          -112.65848541259766,
          47.895503997802734,
          18.824796676635742,
          -34.687076568603516,
          39.32181930541992,
          -115.32209014892578,
          -85.29039001464844,
          122.88488006591797,
          -50.484317779541016,
          -24.0411376953125,
          -13.677151679992676,
          39.51049041748047
         ],
         "y": [
          27.484676361083984,
          3.79726505279541,
          -83.88947296142578,
          -27.206012725830078,
          43.81248474121094,
          96.99076080322266,
          16.286916732788086,
          16.82794189453125,
          -45.0168571472168,
          44.30453872680664,
          -60.943031311035156,
          141.3827362060547,
          9.633867263793945,
          -34.673301696777344,
          92.85338592529297,
          30.257152557373047,
          -91.2662582397461,
          -128.75579833984375,
          -75.89686584472656,
          -59.27724838256836,
          -117.12508392333984,
          78.8982162475586,
          -101.43305969238281,
          102.29987335205078,
          -2.0167365074157715,
          55.56837463378906,
          -49.61564636230469,
          97.60891723632812,
          0.6844167709350586,
          55.043792724609375,
          -50.643192291259766
         ],
         "z": [
          -11.879032135009766,
          42.62489318847656,
          88.1482162475586,
          17.84119415283203,
          -45.8897590637207,
          28.79369354248047,
          79.55487060546875,
          -54.795345306396484,
          117.26560974121094,
          -97.58773040771484,
          -107.56600189208984,
          14.39186954498291,
          -53.31477737426758,
          -119.97496795654297,
          16.916479110717773,
          -134.13845825195312,
          35.50104904174805,
          30.7963809967041,
          40.448421478271484,
          -55.1917839050293,
          -69.4092025756836,
          97.5378189086914,
          -29.81121063232422,
          -56.91859436035156,
          34.05947494506836,
          97.5815658569336,
          -50.45220184326172,
          -71.5807113647461,
          104.26795959472656,
          12.972654342651367,
          -18.06292724609375
         ]
        }
       ],
       "layout": {
        "legend": {
         "tracegroupgap": 0
        },
        "margin": {
         "t": 60
        },
        "scene": {
         "domain": {
          "x": [
           0,
           1
          ],
          "y": [
           0,
           1
          ]
         },
         "xaxis": {
          "title": {
           "text": "X"
          }
         },
         "yaxis": {
          "title": {
           "text": "Y"
          }
         },
         "zaxis": {
          "title": {
           "text": "Z"
          }
         }
        },
        "template": {
         "data": {
          "bar": [
           {
            "error_x": {
             "color": "#2a3f5f"
            },
            "error_y": {
             "color": "#2a3f5f"
            },
            "marker": {
             "line": {
              "color": "#E5ECF6",
              "width": 0.5
             },
             "pattern": {
              "fillmode": "overlay",
              "size": 10,
              "solidity": 0.2
             }
            },
            "type": "bar"
           }
          ],
          "barpolar": [
           {
            "marker": {
             "line": {
              "color": "#E5ECF6",
              "width": 0.5
             },
             "pattern": {
              "fillmode": "overlay",
              "size": 10,
              "solidity": 0.2
             }
            },
            "type": "barpolar"
           }
          ],
          "carpet": [
           {
            "aaxis": {
             "endlinecolor": "#2a3f5f",
             "gridcolor": "white",
             "linecolor": "white",
             "minorgridcolor": "white",
             "startlinecolor": "#2a3f5f"
            },
            "baxis": {
             "endlinecolor": "#2a3f5f",
             "gridcolor": "white",
             "linecolor": "white",
             "minorgridcolor": "white",
             "startlinecolor": "#2a3f5f"
            },
            "type": "carpet"
           }
          ],
          "choropleth": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "type": "choropleth"
           }
          ],
          "contour": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "contour"
           }
          ],
          "contourcarpet": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "type": "contourcarpet"
           }
          ],
          "heatmap": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "heatmap"
           }
          ],
          "heatmapgl": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "heatmapgl"
           }
          ],
          "histogram": [
           {
            "marker": {
             "pattern": {
              "fillmode": "overlay",
              "size": 10,
              "solidity": 0.2
             }
            },
            "type": "histogram"
           }
          ],
          "histogram2d": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "histogram2d"
           }
          ],
          "histogram2dcontour": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "histogram2dcontour"
           }
          ],
          "mesh3d": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "type": "mesh3d"
           }
          ],
          "parcoords": [
           {
            "line": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "parcoords"
           }
          ],
          "pie": [
           {
            "automargin": true,
            "type": "pie"
           }
          ],
          "scatter": [
           {
            "fillpattern": {
             "fillmode": "overlay",
             "size": 10,
             "solidity": 0.2
            },
            "type": "scatter"
           }
          ],
          "scatter3d": [
           {
            "line": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatter3d"
           }
          ],
          "scattercarpet": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattercarpet"
           }
          ],
          "scattergeo": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattergeo"
           }
          ],
          "scattergl": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattergl"
           }
          ],
          "scattermapbox": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scattermapbox"
           }
          ],
          "scatterpolar": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatterpolar"
           }
          ],
          "scatterpolargl": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatterpolargl"
           }
          ],
          "scatterternary": [
           {
            "marker": {
             "colorbar": {
              "outlinewidth": 0,
              "ticks": ""
             }
            },
            "type": "scatterternary"
           }
          ],
          "surface": [
           {
            "colorbar": {
             "outlinewidth": 0,
             "ticks": ""
            },
            "colorscale": [
             [
              0,
              "#0d0887"
             ],
             [
              0.1111111111111111,
              "#46039f"
             ],
             [
              0.2222222222222222,
              "#7201a8"
             ],
             [
              0.3333333333333333,
              "#9c179e"
             ],
             [
              0.4444444444444444,
              "#bd3786"
             ],
             [
              0.5555555555555556,
              "#d8576b"
             ],
             [
              0.6666666666666666,
              "#ed7953"
             ],
             [
              0.7777777777777778,
              "#fb9f3a"
             ],
             [
              0.8888888888888888,
              "#fdca26"
             ],
             [
              1,
              "#f0f921"
             ]
            ],
            "type": "surface"
           }
          ],
          "table": [
           {
            "cells": {
             "fill": {
              "color": "#EBF0F8"
             },
             "line": {
              "color": "white"
             }
            },
            "header": {
             "fill": {
              "color": "#C8D4E3"
             },
             "line": {
              "color": "white"
             }
            },
            "type": "table"
           }
          ]
         },
         "layout": {
          "annotationdefaults": {
           "arrowcolor": "#2a3f5f",
           "arrowhead": 0,
           "arrowwidth": 1
          },
          "autotypenumbers": "strict",
          "coloraxis": {
           "colorbar": {
            "outlinewidth": 0,
            "ticks": ""
           }
          },
          "colorscale": {
           "diverging": [
            [
             0,
             "#8e0152"
            ],
            [
             0.1,
             "#c51b7d"
            ],
            [
             0.2,
             "#de77ae"
            ],
            [
             0.3,
             "#f1b6da"
            ],
            [
             0.4,
             "#fde0ef"
            ],
            [
             0.5,
             "#f7f7f7"
            ],
            [
             0.6,
             "#e6f5d0"
            ],
            [
             0.7,
             "#b8e186"
            ],
            [
             0.8,
             "#7fbc41"
            ],
            [
             0.9,
             "#4d9221"
            ],
            [
             1,
             "#276419"
            ]
           ],
           "sequential": [
            [
             0,
             "#0d0887"
            ],
            [
             0.1111111111111111,
             "#46039f"
            ],
            [
             0.2222222222222222,
             "#7201a8"
            ],
            [
             0.3333333333333333,
             "#9c179e"
            ],
            [
             0.4444444444444444,
             "#bd3786"
            ],
            [
             0.5555555555555556,
             "#d8576b"
            ],
            [
             0.6666666666666666,
             "#ed7953"
            ],
            [
             0.7777777777777778,
             "#fb9f3a"
            ],
            [
             0.8888888888888888,
             "#fdca26"
            ],
            [
             1,
             "#f0f921"
            ]
           ],
           "sequentialminus": [
            [
             0,
             "#0d0887"
            ],
            [
             0.1111111111111111,
             "#46039f"
            ],
            [
             0.2222222222222222,
             "#7201a8"
            ],
            [
             0.3333333333333333,
             "#9c179e"
            ],
            [
             0.4444444444444444,
             "#bd3786"
            ],
            [
             0.5555555555555556,
             "#d8576b"
            ],
            [
             0.6666666666666666,
             "#ed7953"
            ],
            [
             0.7777777777777778,
             "#fb9f3a"
            ],
            [
             0.8888888888888888,
             "#fdca26"
            ],
            [
             1,
             "#f0f921"
            ]
           ]
          },
          "colorway": [
           "#636efa",
           "#EF553B",
           "#00cc96",
           "#ab63fa",
           "#FFA15A",
           "#19d3f3",
           "#FF6692",
           "#B6E880",
           "#FF97FF",
           "#FECB52"
          ],
          "font": {
           "color": "#2a3f5f"
          },
          "geo": {
           "bgcolor": "white",
           "lakecolor": "white",
           "landcolor": "#E5ECF6",
           "showlakes": true,
           "showland": true,
           "subunitcolor": "white"
          },
          "hoverlabel": {
           "align": "left"
          },
          "hovermode": "closest",
          "mapbox": {
           "style": "light"
          },
          "paper_bgcolor": "white",
          "plot_bgcolor": "#E5ECF6",
          "polar": {
           "angularaxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           },
           "bgcolor": "#E5ECF6",
           "radialaxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           }
          },
          "scene": {
           "xaxis": {
            "backgroundcolor": "#E5ECF6",
            "gridcolor": "white",
            "gridwidth": 2,
            "linecolor": "white",
            "showbackground": true,
            "ticks": "",
            "zerolinecolor": "white"
           },
           "yaxis": {
            "backgroundcolor": "#E5ECF6",
            "gridcolor": "white",
            "gridwidth": 2,
            "linecolor": "white",
            "showbackground": true,
            "ticks": "",
            "zerolinecolor": "white"
           },
           "zaxis": {
            "backgroundcolor": "#E5ECF6",
            "gridcolor": "white",
            "gridwidth": 2,
            "linecolor": "white",
            "showbackground": true,
            "ticks": "",
            "zerolinecolor": "white"
           }
          },
          "shapedefaults": {
           "line": {
            "color": "#2a3f5f"
           }
          },
          "ternary": {
           "aaxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           },
           "baxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           },
           "bgcolor": "#E5ECF6",
           "caxis": {
            "gridcolor": "white",
            "linecolor": "white",
            "ticks": ""
           }
          },
          "title": {
           "x": 0.05
          },
          "xaxis": {
           "automargin": true,
           "gridcolor": "white",
           "linecolor": "white",
           "ticks": "",
           "title": {
            "standoff": 15
           },
           "zerolinecolor": "white",
           "zerolinewidth": 2
          },
          "yaxis": {
           "automargin": true,
           "gridcolor": "white",
           "linecolor": "white",
           "ticks": "",
           "title": {
            "standoff": 15
           },
           "zerolinecolor": "white",
           "zerolinewidth": 2
          }
         }
        }
       }
      }
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn.manifold import TSNE\n",
    "import plotly.express as px\n",
    "\n",
    "model.load_state_dict(torch.load(\"./output/word2vec/best_model.pt\", weights_only=True))\n",
    "\n",
    "embeds = model.cxt_embedding.weight.detach().numpy()\n",
    "vocab = list(cbow.vocab_freq.keys())\n",
    "\n",
    "tsne = TSNE(n_components=3)\n",
    "embeds_tsne = tsne.fit_transform(embeds)\n",
    "\n",
    "tsne_df = pd.DataFrame.from_dict({\n",
    "    \"X\": embeds_tsne[:, 0],\n",
    "    \"Y\": embeds_tsne[:, 1],\n",
    "    \"Z\": embeds_tsne[:, 2], \n",
    "    \"Word\": vocab\n",
    "})\n",
    "px.scatter_3d(tsne_df, x=\"X\", y=\"Y\", z=\"Z\", text=\"Word\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "vocab_dict = {0: 4, 1: 6, 2: 3, 3: 2, 4: 2}\n",
    "\n",
    "tree = HuffmanBTree(vocab_dict)\n",
    "print(tree.word_code)\n",
    "print(tree.word_path)\n",
    "print(tree.left_dict)\n",
    "print(tree.right_dict)\n",
    "\n",
    "# {1: [1, 1], 0: [1, 0], 3: [0, 1, 1], 4: [0, 1, 0], 2: [0, 0]}\n",
    "# {1: [8, 7], 0: [8, 7], 3: [8, 6, 5], 4: [8, 6, 5], 2: [8, 6]}\n",
    "# ([[8], [8, 7], [], [6, 5], [6]], [[7], [], [8, 6], [8], [8, 5]])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "freq_df = pd.read_csv(\"./output/bow/Vocab Frequency.csv\")\n",
    "freq_df.sort_values(by=\"Frequency\", ascending=False, ignore_index=True, inplace=True)\n",
    "freq_df = freq_df.iloc[:1000]\n",
    "num_vocab = 5\n",
    "freq_df = freq_df.iloc[np.random.choice(len(freq_df), num_vocab)]\n",
    "\n",
    "vocab = list(freq_df[\"Word\"])\n",
    "freq = list(freq_df[\"Frequency\"])\n",
    "vocab_idx = [i for i in range(len(vocab))]\n",
    "vocab_dict = {k:v for k,v in zip(vocab_idx, freq)}\n",
    "\n",
    "freq_df.head(num_vocab)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [],
   "source": [
    "class Word:\n",
    "    def __init__(self, word, freq):\n",
    "        self.word = word\n",
    "        self.freq = freq\n",
    "\n",
    "class Tree:\n",
    "    def __init__(self, word, freq, left, right):\n",
    "        self.word = word\n",
    "        self.freq = freq\n",
    "        self.right = right\n",
    "        self.left = left\n",
    "\n",
    "def construct_tree(vocab, freq):\n",
    "    word_list = []\n",
    "    for w,f in zip(vocab, freq):\n",
    "        word_list.append(Word(w,f))\n",
    "\n",
    "    count = 0\n",
    "    while len(word_list) > 1:\n",
    "        left = word_list[-1]\n",
    "        right = word_list[-2]\n",
    "\n",
    "        word = f\"W{count}\"\n",
    "        freq = left.freq + right.freq\n",
    "\n",
    "        print(f\"Root: ({word}, {freq}), Left: ({left.word}, {left.freq}), Right: ({right.word}, {right.freq})\")\n",
    "\n",
    "        tree = Tree(word, freq, left, right)\n",
    "        word_list = word_list[:-2]\n",
    "        \n",
    "        word_list.append(tree)\n",
    "        word_list = sorted(word_list, key=lambda a:a.freq, reverse=False)\n",
    "\n",
    "        count += 1\n",
    "\n",
    "    return tree\n",
    "\n",
    "def encode_tree(tree, path_dict, path):\n",
    "\n",
    "    if isinstance(tree, Word):\n",
    "        path_dict[tree.word] = path\n",
    "    \n",
    "    if isinstance(tree.left, Word):\n",
    "        path_dict[tree.left.word] = path + [0]\n",
    "    else:\n",
    "        encode_tree(tree.left, path_dict, path+[0])\n",
    "\n",
    "    if isinstance(tree.right, Word):\n",
    "        path_dict[tree.right.word] = path + [1]\n",
    "    else:\n",
    "        encode_tree(tree.right, path_dict, path+[1])\n",
    "\n",
    "    if len(path_dict) == num_vocab:\n",
    "        return path_dict"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "tree = construct_tree(vocab, freq)\n",
    "\n",
    "path_dict = {}\n",
    "path = []\n",
    "path_dict = encode_tree(tree, path_dict, path)\n",
    "path_dict"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### GloVe"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import yaml\n",
    "import os\n",
    "import sys\n",
    "sys.path.append(\"./src/\")\n",
    "from src.core.word2vec.dataset import Word2VecDataset\n",
    "\n",
    "from collections import defaultdict\n",
    "import tqdm\n",
    "import time\n",
    "\n",
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "from torch.utils.data import TensorDataset, DataLoader\n",
    "\n",
    "from sklearn.model_selection import train_test_split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Extracting from: neg: 100%|██████████| 5/5 [00:00<00:00, 107.32it/s]\n",
      "Extracting from: pos: 100%|██████████| 5/5 [00:00<00:00, 140.87it/s]\n",
      "Extracting from: unsup: 100%|██████████| 5/5 [00:00<00:00, 107.52it/s]\n"
     ]
    }
   ],
   "source": [
    "config_path = \"./configs/glove.yaml\"\n",
    "with open(config_path, \"r\") as stream:\n",
    "    config_dict = yaml.safe_load(stream)\n",
    "\n",
    "cbow_ds = Word2VecDataset(config_dict)\n",
    "\n",
    "text_ls = cbow_ds.text_ls\n",
    "word2id = cbow_ds.word2id\n",
    "id2word = cbow_ds.id2word\n",
    "num_vocab = len(word2id)\n",
    "context = config_dict[\"dataset\"][\"context\"]\n",
    "\n",
    "val_split = config_dict[\"dataset\"][\"val_split\"]\n",
    "batch_size = config_dict[\"dataset\"][\"batch_size\"]\n",
    "seed = config_dict[\"dataset\"][\"batch_size\"]\n",
    "\n",
    "xmax = config_dict[\"train\"][\"x_max\"]\n",
    "alpha = config_dict[\"train\"][\"alpha\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Extracting from: neg: 100%|██████████| 5/5 [00:00<00:00, 273.71it/s]\n",
      "Extracting from: pos: 100%|██████████| 5/5 [00:00<00:00, 364.00it/s]\n",
      "Extracting from: unsup: 100%|██████████| 5/5 [00:00<00:00, 395.03it/s]\n",
      "Training: 100%|██████████| 127/127 [00:04<00:00, 25.92it/s]\n",
      "Validation: 100%|██████████| 31/31 [00:03<00:00,  8.30it/s]\n",
      "Training: 100%|██████████| 127/127 [00:04<00:00, 30.36it/s]\n",
      "Validation: 100%|██████████| 31/31 [00:03<00:00,  8.28it/s]\n",
      "Training: 100%|██████████| 127/127 [00:03<00:00, 32.75it/s]\n",
      "Validation: 100%|██████████| 31/31 [00:03<00:00,  8.71it/s]\n",
      "e:\\anaconda\\envs\\nlp_env\\lib\\site-packages\\joblib\\externals\\loky\\backend\\context.py:136: UserWarning: Could not find the number of physical cores for the following reason:\n",
      "[WinError 2] The system cannot find the file specified\n",
      "Returning the number of logical cores instead. You can silence this warning by setting LOKY_MAX_CPU_COUNT to the number of cores you want to use.\n",
      "  warnings.warn(\n",
      "  File \"e:\\anaconda\\envs\\nlp_env\\lib\\site-packages\\joblib\\externals\\loky\\backend\\context.py\", line 257, in _count_physical_cores\n",
      "    cpu_info = subprocess.run(\n",
      "  File \"e:\\anaconda\\envs\\nlp_env\\lib\\subprocess.py\", line 501, in run\n",
      "    with Popen(*popenargs, **kwargs) as process:\n",
      "  File \"e:\\anaconda\\envs\\nlp_env\\lib\\subprocess.py\", line 947, in __init__\n",
      "    self._execute_child(args, executable, preexec_fn, close_fds,\n",
      "  File \"e:\\anaconda\\envs\\nlp_env\\lib\\subprocess.py\", line 1416, in _execute_child\n",
      "    hp, ht, pid, tid = _winapi.CreateProcess(executable, args,\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1500x1500 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from core.glove.glove import GloVe\n",
    "\n",
    "glove_model = GloVe(config_dict)\n",
    "glove_model.run()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "dict_keys(['br', 'film', 'time', 'one', 'movie', 'like', 'airport', 'could', 'two', 'much', 'see', 'never', 'good', 'love', 'bolt', 'day', 'scene', 'think', 'plane', 'three', 'little', 'scenes', 'cagney', 'great', 'seen', 'plot', 'many', 'acting', 'oscar', 'people', 'brooks', 'even', 'make', 'future', '77', 'air', 'help', 'films', 'made', 'well', 'high', 'chasers', 'etting', 'story', 'man', 'audience', 'making', 'dialogue', 'would', 'better', 'stars', 'rich', 'james', 'also', 'right', 'disaster', 'say', 'far', 'best', 'maybe', 'either', 'something', 'look', 'cut', 'looks', 'leading', 'actors', 'watch', 'take', 'movies', 'life', 'without', 'allen', 'characters', 'bromwell', 'teachers', 'years', 'lead', 'homeless', 'streets', 'bet', 'mel', 'warren', 'opening', 'comedy', 'new', '747', 'valuable', 'takes', 'different', 'second', 'classic', 'although', 'actually', 'liked', 'character', 'dull', 'minutes', 'george', 'back', '<UNK>'])"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "glove_model.glove_ds.word2id.keys()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([63, 128])"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from preprocess.utils import preprocess_text\n",
    "\n",
    "operations = config_dict[\"preprocess\"][\"operations\"]\n",
    "sentence = preprocess_text(text_ls[0], operations)\n",
    "word_ls = sentence.split()\n",
    "vocab = glove_model.glove_ds.word2id.keys()\n",
    "word_ls = [i if i in vocab else \"<UNK>\" for i in word_ls]\n",
    "word_ids = [glove_model.glove_ds.word2id[word] for word in word_ls]\n",
    "word_ids = torch.Tensor(word_ids).to(torch.long)\n",
    "word_embeds = glove_model.model.ctr_embedding(word_ids)\n",
    "word_embeds.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [],
   "source": [
    "cooccur_mat = np.zeros((num_vocab, num_vocab))\n",
    "\n",
    "for text in text_ls:\n",
    "    words = text.split()\n",
    "\n",
    "    if len(words) < 1 + context:\n",
    "        continue\n",
    "\n",
    "    for i in range(context, len(words)-context):\n",
    "        id_i = word2id[words[i]]\n",
    "        for j in range(i-context, i+context):\n",
    "            id_j = word2id[words[j]]\n",
    "            dist = np.abs(j-i)\n",
    "            if dist != 0: cooccur_mat[id_i][id_j] += 1/dist\n",
    "\n",
    "X_ctr, X_cxt = np.indices((num_vocab, num_vocab))\n",
    "X_ctr, X_cxt = X_ctr.flatten(), X_cxt.flatten()\n",
    "X_count = cooccur_mat.flatten()\n",
    "# fig, ax = plt.subplots(1, 1, figsize=(15,15))\n",
    "# sns.heatmap(cooccur_mat, xticklabels=word2id.keys(), yticklabels=word2id.keys(), ax=ax)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# glove_model = GloVeModel(config_dict)\n",
    "# optim = torch.optim.Adam(glove_model.parameters(), lr=config_dict[\"train\"][\"lr\"])\n",
    "\n",
    "# glove_trainer = GloVeTrainer(glove_model, optim, config_dict)\n",
    "# glove_trainer.fit(train_loader, val_loader)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### RNN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import yaml\n",
    "import os\n",
    "import sys\n",
    "sys.path.append(\"./src/\")\n",
    "from preprocess.imdb_reviews import PreprocessIMDB\n",
    "from src.core.word2vec.dataset import Word2VecDataset\n",
    "\n",
    "from collections import defaultdict\n",
    "import tqdm\n",
    "import time\n",
    "\n",
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "from torch.utils.data import TensorDataset, DataLoader\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.preprocessing import OneHotEncoder\n",
    "from sklearn.metrics import accuracy_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Extracting from: neg: 100%|██████████| 64/64 [00:00<00:00, 272.08it/s]\n",
      "Extracting from: pos: 100%|██████████| 64/64 [00:00<00:00, 238.62it/s]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(array([0., 1.]), array([128, 128]))\n",
      "102 100\n"
     ]
    }
   ],
   "source": [
    "### Dataset\n",
    "config_path = \"./configs/rnn.yaml\"\n",
    "with open(config_path, \"r\") as stream:\n",
    "    config_dict = yaml.safe_load(stream)\n",
    "\n",
    "cbow_ds = Word2VecDataset(config_dict)\n",
    "\n",
    "text_ls = cbow_ds.text_ls\n",
    "label_ls = np.array(cbow_ds.preproc_cls.label_ls)\n",
    "le = OneHotEncoder()\n",
    "label_ls = le.fit_transform(label_ls.reshape(-1, 1)).toarray()\n",
    "word2id = cbow_ds.word2id\n",
    "id2word = cbow_ds.id2word\n",
    "\n",
    "word2id[\"<PAD>\"] = len(word2id)\n",
    "id2word[len(id2word)] = \"<PAD>\"\n",
    "\n",
    "num_vocab = len(word2id)\n",
    "\n",
    "print(np.unique(label_ls, return_counts=True))\n",
    "\n",
    "seq_len = config_dict[\"dataset\"][\"seq_len\"]\n",
    "print(num_vocab, seq_len)\n",
    "\n",
    "text_id_ls = []\n",
    "for text in text_ls:\n",
    "    ls = []\n",
    "    for word in text.split()[:seq_len]:\n",
    "        ls.append(word2id[word])\n",
    "    if len(ls) < seq_len:\n",
    "        num_pad = seq_len - len(ls)\n",
    "        ls.extend([word2id[\"<PAD>\"]]*num_pad)\n",
    "    text_id_ls.append(ls)\n",
    "\n",
    "text_id_ls, label_ls = np.array(text_id_ls), np.array(label_ls)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([8, 100])\n",
      "torch.Size([8, 2])\n"
     ]
    }
   ],
   "source": [
    "### Data Loader\n",
    "\n",
    "val_split = config_dict[\"dataset\"][\"val_split\"]\n",
    "batch_size = config_dict[\"dataset\"][\"batch_size\"]\n",
    "seed = config_dict[\"dataset\"][\"batch_size\"]\n",
    "\n",
    "def create_dataloader(X, y, val_split, batch_size, seed):\n",
    "    train_x, val_x, train_y, val_y = train_test_split(X, y, test_size=val_split, random_state=seed)\n",
    "\n",
    "    train_ds = TensorDataset(torch.Tensor(train_x), torch.Tensor(train_y))\n",
    "    train_loader = DataLoader(train_ds, batch_size=batch_size, shuffle=True, drop_last=True, num_workers=1, pin_memory=True)\n",
    "\n",
    "    val_ds = TensorDataset(torch.Tensor(val_x), torch.Tensor(val_y))\n",
    "    val_loader = DataLoader(val_ds, batch_size=batch_size, shuffle=False, drop_last=True, num_workers=1, pin_memory=True)\n",
    "    return train_loader, val_loader\n",
    "\n",
    "train_loader, val_loader = create_dataloader(text_id_ls, label_ls, val_split, batch_size, seed)\n",
    "\n",
    "X, y = next(iter(train_loader))\n",
    "print(X.shape)\n",
    "print(y.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "### RNN Cell\n",
    "\n",
    "class RNNCell(nn.Module):\n",
    "    def __init__(self, h_dim, inp_x_dim, out_x_dim):\n",
    "        super(RNNCell, self).__init__()\n",
    "        \n",
    "        self.hh_dense = nn.Linear(h_dim, h_dim)\n",
    "        self.hx_dense = nn.Linear(inp_x_dim, h_dim)\n",
    "\n",
    "        self.xh_dense = nn.Linear(h_dim, out_x_dim)\n",
    "\n",
    "    def forward(self, ht_1, xt):\n",
    "        ht = nn.Tanh()(self.hh_dense(ht_1) + self.hx_dense(xt))\n",
    "        yt = self.xh_dense(ht)\n",
    "        return ht, yt\n",
    "\n",
    "class RNNModel(nn.Module):\n",
    "    def __init__(self, config_dict):\n",
    "        super(RNNModel, self).__init__()\n",
    "\n",
    "        self.seq_len = config_dict[\"dataset\"][\"seq_len\"]\n",
    "        self.num_layers = config_dict[\"model\"][\"num_layers\"]\n",
    "        self.h_dims = config_dict[\"model\"][\"h_dim\"]\n",
    "        self.x_dims = config_dict[\"model\"][\"x_dim\"]\n",
    "        self.clf_dims = config_dict[\"model\"][\"clf_dim\"]\n",
    "        self.num_classes = config_dict[\"dataset\"][\"num_classes\"]\n",
    "\n",
    "        num_vocab = 2 + config_dict[\"dataset\"][\"num_vocab\"]\n",
    "        embed_dim = config_dict[\"model\"][\"embed_dim\"]\n",
    "        self.embed_layer = nn.Embedding(num_vocab, embed_dim)\n",
    "\n",
    "        self.rnn_cells = []\n",
    "        for i in range(self.num_layers):\n",
    "            h_dim = self.h_dims[i]\n",
    "            inp_x_dim, out_x_dim = self.x_dims[i], self.x_dims[i+1]\n",
    "            self.rnn_cells.append(RNNCell(h_dim, inp_x_dim, out_x_dim))\n",
    "\n",
    "    def forward(self, X):\n",
    "        x_embed = self.embed_layer(X.to(torch.long))\n",
    "        self.num_samples = x_embed.size(0)\n",
    "        \n",
    "        yt_1s = x_embed\n",
    "        hts = self.init_hidden()\n",
    "        for i in range(self.num_layers):\n",
    "            yts = []\n",
    "            rnn_cell = self.rnn_cells[i]\n",
    "\n",
    "            ht = hts[i]\n",
    "            for j in range(self.seq_len):\n",
    "                ht, yt = rnn_cell(ht, yt_1s[:, j, :])\n",
    "                yts.append(yt)\n",
    "\n",
    "            yt_1s = torch.transpose(torch.stack(yts), 0, 1)\n",
    "\n",
    "        out = yt_1s[:, -1, :]\n",
    "        for i in range(len(self.clf_dims)-1):\n",
    "            out = nn.Linear(self.clf_dims[i], self.clf_dims[i+1])(out)\n",
    "            out = nn.ReLU()(out)\n",
    "\n",
    "        out = nn.Linear(self.clf_dims[-1], self.num_classes)(out)\n",
    "\n",
    "        return out\n",
    "\n",
    "    def init_hidden(self):\n",
    "        hts = [nn.init.kaiming_uniform(torch.empty(self.num_samples, dim)) for dim in self.h_dims]\n",
    "\n",
    "        return hts"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([8, 2])"
      ]
     },
     "execution_count": 40,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "rnn_model = RNNModel(config_dict)\n",
    "out = rnn_model(X)\n",
    "out.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "loss_fn = nn.CrossEntropyLoss(reduction=\"sum\")\n",
    "\n",
    "def train_one_epoch(model, optimizer, data_loader, epoch):\n",
    "    model.train()\n",
    "    total_loss, num_instances = 0, 0\n",
    "    y_true, y_pred = [], []\n",
    "\n",
    "    print(f\"-----------Epoch {epoch}/{config_dict['train']['epochs']}-----------\")\n",
    "    pbar = tqdm.tqdm(enumerate(data_loader), total=len(data_loader), desc=\"Training\")\n",
    "\n",
    "    for batch_id, (X, y) in pbar:\n",
    "        y_hat = model(X)\n",
    "\n",
    "        loss = loss_fn(y, y_hat)\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "        optimizer.zero_grad()\n",
    "\n",
    "        total_loss += loss\n",
    "        num_instances += y.size(0)\n",
    "\n",
    "        y_true.append(y.cpu().detach().numpy().argmax(axis=1))\n",
    "        y_pred.append(y_hat.cpu().detach().numpy().argmax(axis=1))\n",
    "\n",
    "    train_loss = total_loss/num_instances\n",
    "\n",
    "    y_true = np.concatenate(y_true, axis=0)\n",
    "    y_pred = np.concatenate(y_pred, axis=0)\n",
    "    train_acc = accuracy_score(y_true, y_pred)\n",
    "\n",
    "    return train_loss, train_acc\n",
    "        \n",
    "@torch.no_grad()\n",
    "def val_one_epoch(model, data_loader):\n",
    "    model.eval()\n",
    "    total_loss, num_instances = 0, 0\n",
    "    y_true, y_pred = [], []\n",
    "\n",
    "    pbar = tqdm.tqdm(enumerate(data_loader), total=len(data_loader), desc=\"Validation\")\n",
    "\n",
    "    for batch_id, (X, y) in pbar:\n",
    "        y_hat = model(X)\n",
    "\n",
    "        loss = loss_fn(y, y_hat)\n",
    "\n",
    "        total_loss += loss\n",
    "        num_instances += y.size(0)\n",
    "\n",
    "        y_true.append(y.cpu().detach().numpy().argmax(axis=1))\n",
    "        y_pred.append(y_hat.cpu().detach().numpy().argmax(axis=1))\n",
    "\n",
    "    val_loss = total_loss/num_instances\n",
    "    \n",
    "    y_true = np.concatenate(y_true, axis=0)\n",
    "    y_pred = np.concatenate(y_pred, axis=0)\n",
    "    val_acc = accuracy_score(y_true, y_pred)\n",
    "\n",
    "    return val_loss, val_acc\n",
    "\n",
    "def fit_model(model, optimizer, train_loader, val_loader):\n",
    "    num_epochs = config_dict[\"train\"][\"epochs\"]\n",
    "    output_folder = config_dict[\"paths\"][\"output_folder\"]\n",
    "\n",
    "    best_val_acc = -np.inf\n",
    "    history = defaultdict(list)\n",
    "\n",
    "    start = time.time()\n",
    "    for epoch in range(1, num_epochs+1):\n",
    "        train_loss, train_acc = train_one_epoch(model, optimizer, train_loader, epoch)\n",
    "        val_loss, val_acc = val_one_epoch(model, val_loader)\n",
    "\n",
    "        print(f\"Train Loss : {train_loss} - Train Acc: {train_acc} - Val Loss : {val_loss} - Val Acc: {val_acc}\")\n",
    "            \n",
    "        history[\"train_loss\"].append(float(train_loss.detach().numpy()))\n",
    "        history[\"val_loss\"].append(float(val_loss.detach().numpy()))\n",
    "        history[\"train_acc\"].append(train_acc)\n",
    "        history[\"val_acc\"].append(val_acc)\n",
    "\n",
    "        if val_acc >= best_val_acc:\n",
    "            print(f\"Validation Accuracy improved from {best_val_acc} to {val_acc}\")\n",
    "            best_val_acc = val_acc\n",
    "            torch.save(model.state_dict(), os.path.join(output_folder, \"best_model.pt\"))\n",
    "        else:\n",
    "            print(f\"Validation Accuracy didn't improve from {best_val_acc}\")\n",
    "\n",
    "    end = time.time()\n",
    "    time_taken = end-start\n",
    "    print('Training completed in {:.0f}h {:.0f}m {:.0f}s'.format(\n",
    "        time_taken // 3600, (time_taken % 3600) // 60, (time_taken % 3600) % 60))\n",
    "    print(f\"Best Val Accuracy: {best_val_acc}\")\n",
    "\n",
    "    return history"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "-----------Epoch 1/10-----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: 100%|██████████| 6/6 [00:04<00:00,  1.44it/s]\n",
      "Validation: 100%|██████████| 1/1 [00:02<00:00,  2.71s/it]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Loss : 0.007845318876206875 - Train Acc: 0.5208333333333334 - Val Loss : 0.09338304400444031 - Val Acc: 0.5\n",
      "Validation Accuracy improved from -inf to 0.5\n",
      "-----------Epoch 2/10-----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: 100%|██████████| 6/6 [00:04<00:00,  1.25it/s]\n",
      "Validation: 100%|██████████| 1/1 [00:03<00:00,  3.35s/it]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Loss : 0.06674794107675552 - Train Acc: 0.5833333333333334 - Val Loss : -0.10954995453357697 - Val Acc: 0.5\n",
      "Validation Accuracy improved from 0.5 to 0.5\n",
      "-----------Epoch 3/10-----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: 100%|██████████| 6/6 [00:04<00:00,  1.50it/s]\n",
      "Validation: 100%|██████████| 1/1 [00:03<00:00,  3.42s/it]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Loss : 0.0038952704053372145 - Train Acc: 0.5208333333333334 - Val Loss : -0.20257733762264252 - Val Acc: 0.5\n",
      "Validation Accuracy improved from 0.5 to 0.5\n",
      "-----------Epoch 4/10-----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: 100%|██████████| 6/6 [00:03<00:00,  1.53it/s]\n",
      "Validation: 100%|██████████| 1/1 [00:03<00:00,  3.24s/it]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Loss : -0.033416640013456345 - Train Acc: 0.5833333333333334 - Val Loss : 0.3872898519039154 - Val Acc: 0.5\n",
      "Validation Accuracy improved from 0.5 to 0.5\n",
      "-----------Epoch 5/10-----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: 100%|██████████| 6/6 [00:04<00:00,  1.46it/s]\n",
      "Validation: 100%|██████████| 1/1 [00:04<00:00,  4.52s/it]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Loss : 0.15104039013385773 - Train Acc: 0.5208333333333334 - Val Loss : -0.4174274802207947 - Val Acc: 0.5\n",
      "Validation Accuracy improved from 0.5 to 0.5\n",
      "-----------Epoch 6/10-----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: 100%|██████████| 6/6 [00:04<00:00,  1.47it/s]\n",
      "Validation: 100%|██████████| 1/1 [00:07<00:00,  7.09s/it]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Loss : 0.1364857703447342 - Train Acc: 0.4375 - Val Loss : 0.27591562271118164 - Val Acc: 0.5\n",
      "Validation Accuracy improved from 0.5 to 0.5\n",
      "-----------Epoch 7/10-----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: 100%|██████████| 6/6 [00:11<00:00,  1.94s/it]\n",
      "Validation: 100%|██████████| 1/1 [00:06<00:00,  6.90s/it]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Loss : 0.025014957413077354 - Train Acc: 0.4583333333333333 - Val Loss : -0.24408555030822754 - Val Acc: 0.5\n",
      "Validation Accuracy improved from 0.5 to 0.5\n",
      "-----------Epoch 8/10-----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: 100%|██████████| 6/6 [00:09<00:00,  1.55s/it]\n",
      "Validation: 100%|██████████| 1/1 [00:08<00:00,  8.43s/it]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Loss : -0.03482829034328461 - Train Acc: 0.5208333333333334 - Val Loss : 0.36403200030326843 - Val Acc: 0.5\n",
      "Validation Accuracy improved from 0.5 to 0.5\n",
      "-----------Epoch 9/10-----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: 100%|██████████| 6/6 [00:10<00:00,  1.80s/it]\n",
      "Validation: 100%|██████████| 1/1 [00:07<00:00,  7.64s/it]\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Loss : 0.15296544134616852 - Train Acc: 0.5416666666666666 - Val Loss : 0.08933401107788086 - Val Acc: 0.5\n",
      "Validation Accuracy improved from 0.5 to 0.5\n",
      "-----------Epoch 10/10-----------\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training: 100%|██████████| 6/6 [00:11<00:00,  1.89s/it]\n",
      "Validation: 100%|██████████| 1/1 [00:07<00:00,  7.85s/it]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Train Loss : -0.006540011614561081 - Train Acc: 0.625 - Val Loss : -0.18438105285167694 - Val Acc: 0.5\n",
      "Validation Accuracy improved from 0.5 to 0.5\n",
      "Training completed in 0h 2m 6s\n",
      "Best Val Accuracy: 0.5\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "optimizer = torch.optim.Adam(rnn_model.parameters(), lr=1e-3)\n",
    "history = fit_model(rnn_model, optimizer, train_loader, val_loader)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "from sklearn.datasets import make_classification\n",
    "from sklearn.metrics import confusion_matrix, ConfusionMatrixDisplay\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.svm import SVC\n",
    "X, y = make_classification(random_state=0)\n",
    "X_train, X_test, y_train, y_test = train_test_split(X, y,\n",
    "                                                    random_state=0)\n",
    "clf = SVC(random_state=0)\n",
    "clf.fit(X_train, y_train)\n",
    "predictions = clf.predict(X_test)\n",
    "cm = confusion_matrix(y_test, predictions, labels=clf.classes_)\n",
    "disp = ConfusionMatrixDisplay(confusion_matrix=cm,\n",
    "                              display_labels=clf.classes_)\n",
    "fig = disp.plot()\n",
    "fig.figure_"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "disp = ConfusionMatrixDisplay.from_predictions(y_test, predictions, labels=clf.classes_)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### LSTM"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import yaml\n",
    "import itertools\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from collections import Counter\n",
    "\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "from torch.utils.data import Dataset, DataLoader\n",
    "from torch.cuda import amp\n",
    "import torch.optim as optim\n",
    "from torch.optim import lr_scheduler as LR\n",
    "import timm\n",
    "\n",
    "import albumentations as A\n",
    "from albumentations.pytorch import ToTensorV2\n",
    "\n",
    "import sys\n",
    "sys.path.append(\"./src/\")\n",
    "from preprocess.utils import preprocess_text"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "config_path = \"./configs/lstm.yaml\"\n",
    "with open(config_path, \"r\") as stream:\n",
    "    config_dict = yaml.safe_load(stream)\n",
    "\n",
    "#############\n",
    "### flickr.py\n",
    "#############\n",
    "\n",
    "### extract_data()\n",
    "im_folder = config_dict[\"paths\"][\"image_folder\"]\n",
    "cap_file = config_dict[\"paths\"][\"captions_file\"]\n",
    "operations = config_dict[\"preprocess\"][\"operations\"]\n",
    "\n",
    "with open(cap_file, \"r\") as f:\n",
    "    lines = np.array(f.readlines()[1:])\n",
    "\n",
    "num_train = config_dict[\"dataset\"][\"train_samples\"]\n",
    "num_test = config_dict[\"dataset\"][\"test_samples\"]\n",
    "rand_ids = np.random.choice(len(df), num_train+num_test)\n",
    "\n",
    "paths, captions = zip(*(s.split(\",\") for s in lines[rand_ids]))\n",
    "paths = [os.path.join(im_folder, i) for i in paths]\n",
    "df = pd.DataFrame.from_dict({\n",
    "    \"Path\": paths,\n",
    "    \"Caption\": captions\n",
    "})\n",
    "df[\"Caption\"] = df[\"Caption\"].map(lambda x: preprocess_text(x, operations))\n",
    "\n",
    "train_df = df.iloc[rand_ids[:num_train]]\n",
    "test_df = df.iloc[num_train:]\n",
    "\n",
    "df.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "### get_vocab()\n",
    "#uniq_words = list(set(itertools.chain.from_iterable(map(str.split, train_df[\"Caption\"]))))\n",
    "num_vocab = config_dict[\"dataset\"][\"num_vocab\"]\n",
    "all_words = []\n",
    "\n",
    "for text in train_df[\"Caption\"]:\n",
    "    all_words += text.split()\n",
    "\n",
    "topk_vocab_freq = Counter(all_words).most_common(num_vocab-4)\n",
    "vocab = [\"<START>\", \"<END>\", \"<PAD>\", \"<UNK>\"] + [i[0] for i in topk_vocab_freq]\n",
    "word2id = {w:i for i,w in enumerate(vocab)}\n",
    "id2word = {v:k for k,v in word2id.items()}\n",
    "\n",
    "### word_to_tokens()\n",
    "seq_len = config_dict[\"dataset\"][\"seq_len\"]\n",
    "\n",
    "tmp_df = test_df\n",
    "tokens = np.zeros((len(tmp_df), seq_len))\n",
    "\n",
    "for i, text in enumerate(tmp_df[\"Caption\"]):\n",
    "    words = [\"<START>\"] + text.split()[:seq_len] \n",
    "    if len(words) < seq_len - 1:\n",
    "        words += [\"<PAD>\"]*(seq_len-1-len(words))\n",
    "    words += [\"<END>\"]\n",
    "\n",
    "    for j, w in enumerate(words):\n",
    "        if w in vocab:\n",
    "            tokens[i, j] = word2id[w]   \n",
    "        else:\n",
    "            tokens[i, j] = word2id[\"<UNK>\"]   \n",
    "\n",
    "### image_transforms()  \n",
    "im_w, im_h = config_dict[\"preprocess\"][\"image_dim\"]\n",
    "train_transforms = A.Compose([\n",
    "                            A.Resize(im_w, im_h),\n",
    "                            A.HorizontalFlip(p=0.5),\n",
    "                            A.RandomBrightnessContrast(p=0.2),\n",
    "                            A.Normalize(\n",
    "                                    mean=[0.485, 0.456, 0.406], \n",
    "                                    std=[0.229, 0.224, 0.225], \n",
    "                                    max_pixel_value=255.0, \n",
    "                                    p=1.0 \n",
    "                                ),\n",
    "                            ToTensorV2()], p=1.0)\n",
    "\n",
    "test_transforms = A.Compose([\n",
    "                            A.Resize(im_w, im_h),\n",
    "                            A.Normalize(\n",
    "                                    mean=[0.485, 0.456, 0.406], \n",
    "                                    std=[0.229, 0.224, 0.225], \n",
    "                                    max_pixel_value=255.0, \n",
    "                                    p=1.0 \n",
    "                                ),\n",
    "                            ToTensorV2()], p=1.0)\n",
    "\n",
    "### get_data()\n",
    "# returns (train_paths, train_tokens), (test_paths, test_tokens), (train_transforms, test_transforms)\n",
    "\n",
    "\n",
    "### def create_dataloader(train_data, test_data, transforms)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "metadata": {},
   "outputs": [],
   "source": [
    "### LSTM Cell\n",
    "import torch.nn as nn\n",
    "import timm\n",
    "\n",
    "class LSTMCell(nn.Module):\n",
    "    def __init__(self, h_dim, inp_x_dim, out_x_dim):\n",
    "        super(LSTMCell, self).__init__()\n",
    "\n",
    "        self.wf_dense = nn.Linear(h_dim, h_dim)\n",
    "        self.uf_dense = nn.Linear(inp_x_dim, h_dim)\n",
    "\n",
    "        self.wi_dense = nn.Linear(h_dim, h_dim)\n",
    "        self.ui_dense = nn.Linear(inp_x_dim, h_dim)\n",
    "\n",
    "        self.wo_dense = nn.Linear(h_dim, h_dim)\n",
    "        self.uo_dense = nn.Linear(inp_x_dim, h_dim)\n",
    "\n",
    "        self.wc_dense = nn.Linear(h_dim, h_dim)\n",
    "        self.uc_dense = nn.Linear(inp_x_dim, h_dim)\n",
    "\n",
    "        self.xh_dense = nn.Linear(h_dim, out_x_dim)\n",
    "\n",
    "    def forward(self, ht_1, ct_1, xt):\n",
    "        ft = nn.Sigmoid()(self.wf_dense(ht_1) + self.uf_dense(xt))\n",
    "        it = nn.Sigmoid()(self.wi_dense(ht_1) + self.ui_dense(xt))\n",
    "        ot = nn.Sigmoid()(self.wo_dense(ht_1) + self.uo_dense(xt))\n",
    "\n",
    "        ct_ = nn.Tanh()(self.wc_dense(ht_1) + self.uc_dense(xt))\n",
    "        ct = ft*ct_1 + it*ct_\n",
    "        ht = ot*nn.Tanh()(ct)\n",
    "\n",
    "        yt = self.xh_dense(ht)\n",
    "\n",
    "        return ht, ct, yt\n",
    "\n",
    "\n",
    "### LSTM Model\n",
    "class LSTMModel(nn.Module):\n",
    "    def __init__(self, config_dict):\n",
    "        super(LSTMModel, self).__init__()\n",
    "\n",
    "        self.seq_len = config_dict[\"dataset\"][\"seq_len\"]\n",
    "        self.num_layers = config_dict[\"model\"][\"num_layers\"]\n",
    "        self.h_dims = config_dict[\"model\"][\"h_dim\"]\n",
    "        self.x_dims = config_dict[\"model\"][\"x_dim\"]\n",
    "        image_dim = config_dict[\"preprocess\"][\"image_dim\"]\n",
    "\n",
    "        num_vocab = config_dict[\"dataset\"][\"num_vocab\"]\n",
    "        embed_dim = config_dict[\"model\"][\"embed_dim\"]\n",
    "        self.embed_layer = nn.Embedding(num_vocab, embed_dim)\n",
    "        \n",
    "        image_backbone = config_dict[\"model\"][\"image_backbone\"]\n",
    "        self.im_head = timm.create_model(image_backbone, pretrained=True, num_classes=0)\n",
    "        im_feat_dim = self.im_head(torch.rand(tuple([1] + image_dim))).data.shape[-1]\n",
    "        self.im_dense = nn.Linear(im_feat_dim, embed_dim)\n",
    "\n",
    "        self.word_classifier = nn.Linear(self.x_dims[-1], num_vocab)\n",
    "\n",
    "        self.lstm_cells = []\n",
    "        for i in range(self.num_layers):\n",
    "            h_dim = self.h_dims[i]\n",
    "            inp_x_dim, out_x_dim = self.x_dims[i], self.x_dims[i+1]\n",
    "            self.lstm_cells.append(LSTMCell(h_dim, inp_x_dim, out_x_dim))\n",
    "\n",
    "    def forward(self, images, tokens=None):\n",
    "        self.num_samples = images.size(0)\n",
    "\n",
    "        hts = self.init_hidden()\n",
    "        cts = self.init_hidden()\n",
    "\n",
    "        im_feat = self.im_head(images)\n",
    "        yt = self.im_dense(im_feat)\n",
    "\n",
    "        for i in range(self.num_layers):\n",
    "            ht, ct = hts[i], cts[i]\n",
    "            ht, ct, yt = self.lstm_cells[i](ht, ct, yt)\n",
    "            hts[i], cts[i] = ht, ct\n",
    "\n",
    "        if tokens is not None:\n",
    "            x_embed = self.embed_layer(tokens.to(torch.long))\n",
    "            yt = x_embed[:, 0, :]\n",
    "        else:\n",
    "            start_token = torch.zeros(1, 1).to(torch.long)\n",
    "            yt = self.embed_layer(start_token)[:, 0, :]\n",
    "\n",
    "        pts = []\n",
    "        for j in range(self.seq_len):\n",
    "            for i in range(self.num_layers):\n",
    "                ht, ct = hts[i], cts[i]\n",
    "                ht, ct, yt = self.lstm_cells[i](ht, ct, yt)\n",
    "                hts[i], cts[i] = ht, ct\n",
    "                \n",
    "            yt = self.word_classifier(yt)\n",
    "            pt = nn.Softmax()(yt)[:, None, :]\n",
    "            pts.append(pt)\n",
    "            \n",
    "            if j >=  self.seq_len - 1:\n",
    "                break\n",
    "            if tokens is not None:\n",
    "                yt = x_embed[:, j+1, :]\n",
    "            else:\n",
    "                yt = yt.argmax(axis=1)\n",
    "                yt = self.embed_layer(yt.to(torch.long))\n",
    "\n",
    "        return torch.concat(pts, dim=1)\n",
    "\n",
    "\n",
    "    def init_hidden(self):\n",
    "        hts = [nn.init.kaiming_uniform_(torch.empty(self.num_samples, dim)) for dim in self.h_dims]\n",
    "\n",
    "        return hts"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\anaconda\\envs\\nlp_env\\lib\\site-packages\\tqdm\\auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
      "  from .autonotebook import tqdm as notebook_tqdm\n",
      "Training:   0%|          | 0/6 [00:00<?, ?it/s]e:\\anaconda\\envs\\nlp_env\\lib\\site-packages\\torch\\nn\\modules\\module.py:1736: UserWarning: Implicit dimension choice for softmax has been deprecated. Change the call to include dim=X as an argument.\n",
      "  return self._call_impl(*args, **kwargs)\n",
      "e:\\anaconda\\envs\\nlp_env\\lib\\site-packages\\torch\\nn\\_reduction.py:51: UserWarning: size_average and reduce args will be deprecated, please use reduction='mean' instead.\n",
      "  warnings.warn(warning.format(ret))\n",
      "Training: 100%|██████████| 6/6 [00:07<00:00,  1.26s/it]\n",
      "Validation: 100%|██████████| 1/1 [00:05<00:00,  5.66s/it]\n",
      "Training: 100%|██████████| 6/6 [00:05<00:00,  1.04it/s]\n",
      "Validation: 100%|██████████| 1/1 [00:05<00:00,  5.24s/it]\n",
      "Training: 100%|██████████| 6/6 [00:09<00:00,  1.64s/it]\n",
      "Validation: 100%|██████████| 1/1 [00:05<00:00,  5.49s/it]\n"
     ]
    }
   ],
   "source": [
    "import yaml\n",
    "import sys\n",
    "sys.path.append(\"./src/\")\n",
    "from src.core.lstm import lstm\n",
    "\n",
    "config_path = \"./configs/lstm.yaml\"\n",
    "with open(config_path, \"r\") as stream:\n",
    "    config_dict = yaml.safe_load(stream)\n",
    "\n",
    "algo = lstm.LSTM(config_dict)\n",
    "algo.run()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "41.210533142089844"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "history = algo.history\n",
    "\n",
    "history[\"train_Perplexity\"][0].item()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Inference:   0%|          | 0/2 [00:00<?, ?it/s]e:\\anaconda\\envs\\nlp_env\\lib\\site-packages\\torch\\nn\\modules\\module.py:1736: UserWarning: Implicit dimension choice for softmax has been deprecated. Change the call to include dim=X as an argument.\n",
      "  return self._call_impl(*args, **kwargs)\n",
      "Inference: 100%|██████████| 2/2 [00:37<00:00, 18.73s/it]\n"
     ]
    }
   ],
   "source": [
    "test_captions, test_pred_captions = algo.run_infer()\n",
    "\n",
    "for true, pred in zip(test_captions, test_pred_captions):\n",
    "    print(true, pred)\n",
    "\n",
    "test_tokens, test_pred_tokens = algo.run_infer()\n",
    "test_pred_tokens = test_pred_tokens.argmax(axis=-1).astype(\"int\")\n",
    "test_tokens = test_tokens.astype(\"int\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['two <UNK> <UNK> <UNK> <UNK> <UNK> <UNK> <UNK> <UNK> <UNK> ',\n",
       " '<UNK> <UNK> dog carrying <UNK> mouth <UNK> ',\n",
       " 'black <UNK> dog <UNK> <UNK> course ',\n",
       " '<UNK> <UNK> wearing green <UNK> ',\n",
       " 'black dog <UNK> <UNK> <UNK> <UNK> <UNK> fence ',\n",
       " 'young man <UNK> <UNK> mouth wearing black jacket ',\n",
       " '<UNK> dog <UNK> <UNK> <UNK> <UNK> ',\n",
       " 'two <UNK> <UNK> <UNK> <UNK> case ']"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import numpy as np\n",
    "\n",
    "func = lambda x: algo.flickr_ds.id2word[x]\n",
    "vect_func = np.vectorize(func)\n",
    "\n",
    "test_words = vect_func(test_tokens)\n",
    "test_pred_words = vect_func(test_pred_tokens)\n",
    "\n",
    "test_captions = []\n",
    "for words in test_words:\n",
    "    txt = \"\"\n",
    "    for word in words:\n",
    "        if word not in  [\"<START>\", \"<END>\", \"<PAD>\"]:\n",
    "            txt += f\"{word} \"\n",
    "    test_captions.append(txt)\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "images, tokens = next(iter(train_loader))\n",
    "\n",
    "model = LSTMModel(config_dict)\n",
    "out = model(images)\n",
    "print(out.shape, tokens.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "metadata": {},
   "outputs": [],
   "source": [
    "import torch.nn.functional as F\n",
    "\n",
    "flat_out = torch.flatten(out, end_dim=1)\n",
    "flat_tokens = torch.flatten(tokens)\n",
    "flat_tokens = F.one_hot(flat_tokens.to(torch.long), num_classes=config_dict[\"dataset\"][\"num_vocab\"])\n",
    "\n",
    "loss_fn = nn.CrossEntropyLoss(reduce=\"sum\")\n",
    "loss_fn(flat_out, flat_tokens.to(torch.float))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([32, 1, 100])"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "inp = torch.ones((32, 100))\n",
    "inp = inp[:, None, :]\n",
    "inp.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import timm\n",
    "import torch\n",
    "\n",
    "im_model = timm.create_model(\"mobilenetv3_small_050\", pretrained=True, num_classes=0)\n",
    "inp = torch.randn((1, 3, 64, 64))\n",
    "im_model(inp).data.shape[-1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "for layer in im_model.children():\n",
    "    if hasattr(layer, \"out_features\"):\n",
    "        print(layer.out_features)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "6\n",
      "[1 5 3 0 4 2]\n"
     ]
    }
   ],
   "source": [
    "### METEOR\n",
    "\n",
    "import numpy as np\n",
    "\n",
    "refs = np.array([0, 1, 2, 3, 0, 4])#, 7]\n",
    "preds = np.array([3, 0, 4, 2, 0, 1])#, 6, 8]\n",
    "\n",
    "m = 0\n",
    "mask = np.ones(preds.shape)\n",
    "match_ids = -1*np.ones(refs.shape).astype(\"int\")\n",
    "for i, w_r in enumerate(refs):\n",
    "    for j, w_p in enumerate(preds):\n",
    "        if w_r == w_p and mask[j]:\n",
    "            m += 1\n",
    "            mask[j] = 0\n",
    "            match_ids[i] = j\n",
    "            break\n",
    "\n",
    "print(m)\n",
    "print(match_ids)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0.5 0.7142857142857143 0.588235294117647\n"
     ]
    }
   ],
   "source": [
    "### ROUGE-N\n",
    "\n",
    "import numpy as np\n",
    "\n",
    "refs = np.array([0, 1, 2, 3, 0, 4, 7])\n",
    "preds = np.array([3, 0, 4, 2, 0, 1, 6, 8, 1, 1])\n",
    "n = 1\n",
    "\n",
    "refs_n = [\" \".join([str(refs[j+k]) for k in range(n)]) for j in range(len(refs)-n+1)]\n",
    "preds_n = [\" \".join([str(preds[j+k]) for k in range(n)]) for j in range(len(preds)-n+1)]\n",
    "\n",
    "n_unq_ngrams = len(set(refs_n) & set(preds_n))\n",
    "precision = n_unq_ngrams/len(preds_n)\n",
    "recall = n_unq_ngrams/len(refs_n)\n",
    "fscore = 2*precision*recall/(precision + recall)\n",
    "\n",
    "print(precision, recall, fscore)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0 2\n",
      "1 1\n",
      "2 1\n",
      "3 1\n",
      "4 1\n",
      "7 1\n"
     ]
    }
   ],
   "source": [
    "from collections import Counter\n",
    "\n",
    "ref_cnt_dict = Counter(refs)\n",
    "pred_cnt_dict = Counter(refs)\n",
    "\n",
    "sum = 0\n",
    "for k in pred_cnt_dict.keys():\n",
    "    cnt_k = min(pred_cnt_dict[k], ref_cnt_dict.get(k, 0))\n",
    "    print(k, cnt_k)\n",
    "    sum += cnt_k"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "5.0"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### ROUGE-L\n",
    "\n",
    "refs = np.array([0, 1, 2, 3, 0, 4, 7])\n",
    "preds = np.array([3, 0, 4, 2, 3, 4, 6, 7])\n",
    "\n",
    "m = len(refs)\n",
    "n = len(preds)\n",
    "\n",
    "lcs_mat = np.zeros((m+1, n+1))\n",
    "\n",
    "for i in range(1, m+1):\n",
    "    for j in range(1, n+1):\n",
    "        if refs[i-1] == preds[j-1]:\n",
    "            lcs_mat[i, j] = lcs_mat[i-1][j-1] + 1\n",
    "        else:\n",
    "            lcs_mat[i, j] = max(lcs_mat[i-1][j], lcs_mat[i][j-1])\n",
    "\n",
    "float(lcs_mat[m][n])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['0 1', '1 2', '2 3', '3 4', '4 5', '5 6']"
      ]
     },
     "execution_count": 70,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### ROUGE-S\n",
    "\n",
    "pred = np.array([0, 1, 2, 3, 4, 5, 6])\n",
    "n = 4\n",
    "\n",
    "# for i in range(len(preds)-1):\n",
    "#     for k in range(i+1, min(len(preds)-1, i+n+1)):\n",
    "#         print(pred[i], pred[k])\n",
    "\n",
    "import itertools\n",
    "ls = [[\" \".join([str(pred[i]), str(pred[k])]) for k in range(i+1, min(len(preds)-1, i+n+1))] for i in range(len(preds)-1)]\n",
    "ls = list(itertools.chain.from_iterable(ls))\n",
    "\n",
    "[\" \".join([str(pred[j]), str(pred[j+1])]) for j in range(len(pred)-1)]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### GRU"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package universal_tagset to\n",
      "[nltk_data]     C:\\Users\\shanmukh\\AppData\\Roaming\\nltk_data...\n",
      "[nltk_data]   Unzipping taggers\\universal_tagset.zip.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "True"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import nltk\n",
    "\n",
    "# nltk.download(\"brown\")\n",
    "# nltk.download(\"conll2000\")\n",
    "nltk.download(\"universal_tagset\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "import yaml\n",
    "import sys\n",
    "sys.path.append(\"./src/\")\n",
    "from preprocess.utils import preprocess_text\n",
    "from sklearn.preprocessing import OneHotEncoder\n",
    "\n",
    "from src.core.gru import gru\n",
    "\n",
    "from nltk.corpus import brown\n",
    "from nltk.corpus import treebank\n",
    "from nltk.corpus import conll2000\n",
    "\n",
    "from collections import Counter\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training:   0%|          | 0/12 [00:00<?, ?it/s]e:\\anaconda\\envs\\nlp_env\\lib\\site-packages\\torch\\nn\\modules\\module.py:1736: UserWarning: Implicit dimension choice for softmax has been deprecated. Change the call to include dim=X as an argument.\n",
      "  return self._call_impl(*args, **kwargs)\n",
      "Training: 100%|██████████| 12/12 [00:03<00:00,  3.24it/s]\n",
      "Validation:   0%|          | 0/3 [00:00<?, ?it/s]e:\\anaconda\\envs\\nlp_env\\lib\\site-packages\\torch\\nn\\modules\\module.py:1736: UserWarning: Implicit dimension choice for softmax has been deprecated. Change the call to include dim=X as an argument.\n",
      "  return self._call_impl(*args, **kwargs)\n",
      "Validation: 100%|██████████| 3/3 [00:04<00:00,  1.47s/it]\n",
      "Training:   0%|          | 0/12 [00:00<?, ?it/s]e:\\anaconda\\envs\\nlp_env\\lib\\site-packages\\torch\\nn\\modules\\module.py:1736: UserWarning: Implicit dimension choice for softmax has been deprecated. Change the call to include dim=X as an argument.\n",
      "  return self._call_impl(*args, **kwargs)\n",
      "Training: 100%|██████████| 12/12 [00:04<00:00,  2.48it/s]\n",
      "Validation:   0%|          | 0/3 [00:00<?, ?it/s]e:\\anaconda\\envs\\nlp_env\\lib\\site-packages\\torch\\nn\\modules\\module.py:1736: UserWarning: Implicit dimension choice for softmax has been deprecated. Change the call to include dim=X as an argument.\n",
      "  return self._call_impl(*args, **kwargs)\n",
      "Validation: 100%|██████████| 3/3 [00:05<00:00,  1.75s/it]\n"
     ]
    }
   ],
   "source": [
    "config_path = \"./configs/gru.yaml\"\n",
    "with open(config_path, \"r\") as stream:\n",
    "    config_dict = yaml.safe_load(stream)\n",
    "\n",
    "algo = gru.GRU(config_dict)\n",
    "algo.run()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "loader = algo.run_infer()\n",
    "\n",
    "iter_loader = iter(loader)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([4, 32])\n"
     ]
    }
   ],
   "source": [
    "try:\n",
    "    iter_loader = next(iter_loader)\n",
    "    print(iter_loader[0].shape)\n",
    "except:\n",
    "    print(\"Done\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[('Pierre', 'NOUN'),\n",
       " ('Vinken', 'NOUN'),\n",
       " (',', '.'),\n",
       " ('61', 'NUM'),\n",
       " ('years', 'NOUN'),\n",
       " ('old', 'ADJ'),\n",
       " (',', '.'),\n",
       " ('will', 'VERB'),\n",
       " ('join', 'VERB'),\n",
       " ('the', 'DET'),\n",
       " ('board', 'NOUN'),\n",
       " ('as', 'ADP'),\n",
       " ('a', 'DET'),\n",
       " ('nonexecutive', 'ADJ'),\n",
       " ('director', 'NOUN'),\n",
       " ('Nov.', 'NOUN'),\n",
       " ('29', 'NUM'),\n",
       " ('.', '.')]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "treebank_corpus = treebank.tagged_sents(tagset='universal')\n",
    "brown_corpus = brown.tagged_sents(tagset='universal')\n",
    "conll_corpus = conll2000.tagged_sents(tagset='universal')\n",
    "\n",
    "corpus = treebank_corpus\n",
    "corpus[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "operations = config_dict[\"preprocess\"][\"operations\"]\n",
    "num_vocab = config_dict[\"dataset\"][\"num_vocab\"]\n",
    "seq_len = config_dict[\"dataset\"][\"seq_len\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "X = [[i[0] for i in sent] for sent in corpus]\n",
    "y = [[i[1] for i in sent] for sent in corpus]\n",
    "\n",
    "X = [preprocess_text(\" \".join(sent), operations).split() for sent in X]\n",
    "\n",
    "all_words = [word for sent in X for word in sent]\n",
    "topk_vocab_freq = Counter(all_words).most_common(num_vocab-2)\n",
    "vocabX = [\"<PAD>\", \"<UNK>\"] + [word[0] for word in topk_vocab_freq]\n",
    "unq_pos = [\"<PAD>\", \"<UNK>\"] + list(set([word_pos for sent_pos in y for word_pos in sent_pos]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "word2idX = {w:i for i, w in enumerate(vocabX)}\n",
    "id2wordX = {v:k for k,v in word2idX.items()}\n",
    "\n",
    "posEnc = {w:i for i, w in enumerate(unq_pos)}\n",
    "posDec = {v:k for k,v in posEnc.items()}\n",
    "\n",
    "lencoder = OneHotEncoder()\n",
    "lencoder.fit(np.arange(len(unq_pos)).reshape(-1,1))\n",
    "\n",
    "tokenX = np.zeros((len(X), seq_len))\n",
    "Y = np.zeros((len(y), seq_len, len(unq_pos)))\n",
    "for i, (sent, sent_pos) in enumerate(zip(X, y)):\n",
    "    labels = np.zeros((seq_len))\n",
    "    sent = sent[:seq_len]\n",
    "    sent_pos = sent_pos[:seq_len]\n",
    "    if len(sent) < seq_len:\n",
    "        sent = sent + [\"<PAD>\"]*(seq_len - len(sent))\n",
    "        sent_pos = sent_pos + [\"<PAD>\"]*(seq_len - len(sent))\n",
    "    for j, (word, word_pos) in enumerate(zip(sent, sent_pos)):\n",
    "        bool_ = word in vocabX\n",
    "        tokenX[i][j] = word2idX[word] if bool_ else word2idX[\"<UNK>\"]\n",
    "        labels[j] = posEnc[word_pos] if bool_ else posEnc[\"<UNK>\"]\n",
    "        \n",
    "    Y[i] = lencoder.transform(labels.reshape(-1,1)).toarray()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'A': {'precision': 0.5,\n",
       "  'recall': 1.0,\n",
       "  'f1-score': 0.6666666666666666,\n",
       "  'support': 1.0},\n",
       " 'B': {'precision': 0.0, 'recall': 0.0, 'f1-score': 0.0, 'support': 1.0},\n",
       " 'C': {'precision': 1.0,\n",
       "  'recall': 0.6666666666666666,\n",
       "  'f1-score': 0.8,\n",
       "  'support': 3.0},\n",
       " 'accuracy': 0.6,\n",
       " 'macro avg': {'precision': 0.5,\n",
       "  'recall': 0.5555555555555555,\n",
       "  'f1-score': 0.48888888888888893,\n",
       "  'support': 5.0},\n",
       " 'weighted avg': {'precision': 0.7,\n",
       "  'recall': 0.6,\n",
       "  'f1-score': 0.6133333333333334,\n",
       "  'support': 5.0}}"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.metrics import classification_report\n",
    "y_true = [0, 1, 2, 2, 2]\n",
    "y_pred = [0, 0, 2, 2, 1]\n",
    "\n",
    "report = classification_report(y_true, y_pred, target_names=[\"A\", \"B\", \"C\"], output_dict=True)\n",
    "report"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(18, 2)"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import numpy as np\n",
    "\n",
    "x = np.ones((3, 4, 3))\n",
    "x.reshape((-1, 2)).shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[np.str_('amsterdam'), np.str_('paris'), np.str_('tokyo')]"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn.preprocessing import OneHotEncoder\n",
    "\n",
    "le = OneHotEncoder()\n",
    "a = np.array([\"paris\", \"paris\", \"tokyo\", \"amsterdam\"]).reshape(-1,1)\n",
    "le.fit(a)\n",
    "list(le.categories_[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Seq2Seq"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "import yaml\n",
    "import pandas as pd\n",
    "import sys\n",
    "sys.path.append(\"./src/\")\n",
    "\n",
    "from src.preprocess import eng2tel\n",
    "from src.core.seq2seq import seq2seq\n",
    "\n",
    "from collections import Counter\n",
    "import numpy as np\n",
    "\n",
    "import re\n",
    "import unicodedata"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [],
   "source": [
    "config_path = \"./configs/seq2seq.yaml\"\n",
    "with open(config_path, \"r\") as stream:\n",
    "    config_dict = yaml.safe_load(stream)\n",
    "\n",
    "# seq2seq_ds = eng2tel.PreprocessSeq2Seq(config_dict)\n",
    "# X, Y = seq2seq_ds.get_data(seq2seq_ds.df)\n",
    "\n",
    "algo = seq2seq.Seq2Seq(config_dict)\n",
    "train_loader, val_loader = algo.run()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([4, 16]) torch.Size([4, 16])\n"
     ]
    }
   ],
   "source": [
    "src, tgt = next(iter(train_loader))\n",
    "print(src.shape, tgt.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = Seq2SeqModel(config_dict)\n",
    "lr = config_dict[\"train\"][\"lr\"]\n",
    "optim = torch.optim.Adam(model.parameters(), lr=lr)\n",
    "trainer = Seq2SeqTrainer(model, optim, config_dict)\n",
    "trainer.fit(train_loader, val_loader)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([4, 16, 256]) torch.Size([4, 256])\n"
     ]
    }
   ],
   "source": [
    "encoder = Seq2SeqEncoder(config_dict)\n",
    "yts, ht = encoder(src)\n",
    "\n",
    "print(yts.shape, ht.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([4, 16, 256])\n",
      "torch.Size([4, 1, 16])\n",
      "torch.Size([4, 1, 16])\n",
      "torch.Size([4, 1, 16])\n",
      "torch.Size([4, 1, 16])\n",
      "torch.Size([4, 1, 16])\n",
      "torch.Size([4, 1, 16])\n",
      "torch.Size([4, 1, 16])\n",
      "torch.Size([4, 1, 16])\n",
      "torch.Size([4, 1, 16])\n",
      "torch.Size([4, 1, 16])\n",
      "torch.Size([4, 1, 16])\n",
      "torch.Size([4, 1, 16])\n",
      "torch.Size([4, 1, 16])\n",
      "torch.Size([4, 1, 16])\n",
      "torch.Size([4, 1, 16])\n",
      "torch.Size([4, 1, 16])\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\anaconda\\envs\\nlp_env\\lib\\site-packages\\torch\\nn\\modules\\module.py:1736: UserWarning: Implicit dimension choice for softmax has been deprecated. Change the call to include dim=X as an argument.\n",
      "  return self._call_impl(*args, **kwargs)\n"
     ]
    }
   ],
   "source": [
    "model = Seq2SeqModel(config_dict)\n",
    "\n",
    "probs, weights = model(src, tgt)\n",
    "print(probs.shape)\n",
    "for w in weights:\n",
    "    print(w.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([4, 126])"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x = torch.randn((4, 16, 126))\n",
    "y = list(torch.transpose(x, 1, 0))\n",
    "y[0].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1 5\n",
      "2 4\n",
      "3 3\n",
      "4 2\n",
      "5 1\n"
     ]
    }
   ],
   "source": [
    "ls = [1, 2, 3, 4, 5]\n",
    "\n",
    "for i in range(len(ls)):\n",
    "    i1, i2 = i, len(ls)-i-1\n",
    "    print(ls[i1], ls[i2])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "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>English</th>\n",
       "      <th>Telugu</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>His legs are long.</td>\n",
       "      <td>అతని కాళ్ళు పొడవుగా ఉన్నాయి.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Who taught Tom how to speak French?</td>\n",
       "      <td>టామ్ ఫ్రెంచ్ మాట్లాడటం ఎలా నేర్పించారు?</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>I swim in the sea every day.</td>\n",
       "      <td>నేను ప్రతి రోజు సముద్రంలో ఈత కొడతాను.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Tom popped into the supermarket on his way hom...</td>\n",
       "      <td>టామ్ కొంచెం పాలు కొనడానికి ఇంటికి వెళ్ళేటప్పుడ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Smoke filled the room.</td>\n",
       "      <td>పొగ గదిని నింపింది.</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                                             English  \\\n",
       "0                                 His legs are long.   \n",
       "1                Who taught Tom how to speak French?   \n",
       "2                       I swim in the sea every day.   \n",
       "3  Tom popped into the supermarket on his way hom...   \n",
       "4                             Smoke filled the room.   \n",
       "\n",
       "                                              Telugu  \n",
       "0                       అతని కాళ్ళు పొడవుగా ఉన్నాయి.  \n",
       "1            టామ్ ఫ్రెంచ్ మాట్లాడటం ఎలా నేర్పించారు?  \n",
       "2              నేను ప్రతి రోజు సముద్రంలో ఈత కొడతాను.  \n",
       "3  టామ్ కొంచెం పాలు కొనడానికి ఇంటికి వెళ్ళేటప్పుడ...  \n",
       "4                                పొగ గదిని నింపింది.  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fpath = config_dict[\"paths\"][\"input_file\"]\n",
    "\n",
    "with open(fpath, \"r\", encoding=\"utf-8\") as f:\n",
    "    lines = f.readlines()\n",
    "\n",
    "eng = [i.split(\"++++$++++\")[0] for i in lines]\n",
    "tel = [i.split(\"++++$++++\")[1].rstrip() for i in lines]\n",
    "\n",
    "df = pd.DataFrame.from_dict({\n",
    "    \"English\": eng,\n",
    "    \"Telugu\": tel\n",
    "})\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [],
   "source": [
    "def preprocess_en(text):\n",
    "    text = text.lower().strip()\n",
    "    text = re.sub(r'([?.!,¿_])',r' \\1 ',text)\n",
    "    text = re.sub(r'[\" \"]+', \" \",text)\n",
    "    text = re.sub(r\"[^a-zA-Z?.!,¿_]+\",\" \",text)\n",
    "    text = text.strip()\n",
    "    return text\n",
    "\n",
    "\n",
    "def preprocess_te(text):\n",
    "    text = ''.join(\n",
    "        c for c in unicodedata.normalize('NFD', text)\n",
    "        if unicodedata.category(c) != 'BN'\n",
    "    )\n",
    "    text = re.sub(r'([?.!,¿_])',r' \\1 ',text)\n",
    "    text = re.sub(r'[\" \"]+', \" \",text)\n",
    "    text = text.strip()\n",
    "    return text"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([1, 2, 3])"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x = np.array([1,2,3, 4, 5])\n",
    "x[:-2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from collections import Counter\n",
    "\n",
    "Counter(\" \".join(preprocess_te(i) for i in tel))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "Counter(\" \".join(preprocess_en(i) for i in eng))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'పొగ గదిని నింపింది.'"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def unicodeToAscii(s):\n",
    "    return ''.join(\n",
    "        c for c in unicodedata.normalize('NFD', s)\n",
    "        if unicodedata.category(c) != 'BN'\n",
    "    )\n",
    "\n",
    "def normalizeString(s):\n",
    "    s = unicodeToAscii(s.lower().strip())\n",
    "    # s = re.sub(r\"([.!?])\", r\" \\1\", s)\n",
    "    # s = re.sub(r\"[^a-zA-Z!?]+\", r\" \", s)\n",
    "    return s.strip()\n",
    "\n",
    "normalizeString(tel[4])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "టామ్ ఫ్రెంచ్ మాట్లాడటం ఎలా నేర్పించారు?\n"
     ]
    }
   ],
   "source": [
    "from indicnlp.normalize.indic_normalize import IndicNormalizerFactory\n",
    "\n",
    "factory=IndicNormalizerFactory()\n",
    "normalizer=factory.get_normalizer(\"te\",remove_nuktas=False)\n",
    "output_text=normalizer.normalize(tel[1])\n",
    "print(output_text)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'అతని కాళ్ళు పొడవుగా ఉన్నాయి.'"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### Seq2Seq\n",
    "tel[0].encode(\"utf-8\").decode()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Transformer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import yaml\n",
    "import pandas as pd\n",
    "import sys\n",
    "sys.path.append(\"./src/\")\n",
    "from tqdm import tqdm\n",
    "from collections import defaultdict\n",
    "import numpy as np\n",
    "\n",
    "from src.preprocess.utils import preprocess_text, BytePairEncoding\n",
    "from src.core.transformer.transformer import Transformer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "config_path = \"./configs/transformer.yaml\"\n",
    "with open(config_path, \"r\") as stream:\n",
    "    config_dict = yaml.safe_load(stream)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "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>id</th>\n",
       "      <th>lyrics</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5FDh0G5oQ2UD9JxWiDCT1x</td>\n",
       "      <td>paper doll come try it on step out of that bla...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>6XK6Zw6JkFsHXzAcMWNiIr</td>\n",
       "      <td>on my way on my way on my way on my way i woul...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>7GX5flRQZVHRAGd6B4TmDO</td>\n",
       "      <td>are you alright i m alright i m quite alright ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>73q3FpQVXWk5eSUnyo83E4</td>\n",
       "      <td>it s tearin up my heart when i m with you but ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1IdTIzkTK6CcpKKgMUGbg7</td>\n",
       "      <td>some of us we re hardly ever here the rest of ...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                       id                                             lyrics\n",
       "0  5FDh0G5oQ2UD9JxWiDCT1x  paper doll come try it on step out of that bla...\n",
       "1  6XK6Zw6JkFsHXzAcMWNiIr  on my way on my way on my way on my way i woul...\n",
       "2  7GX5flRQZVHRAGd6B4TmDO  are you alright i m alright i m quite alright ...\n",
       "3  73q3FpQVXWk5eSUnyo83E4  it s tearin up my heart when i m with you but ...\n",
       "4  1IdTIzkTK6CcpKKgMUGbg7  some of us we re hardly ever here the rest of ..."
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "input_path = config_dict[\"paths\"][\"input_file\"]\n",
    "\n",
    "df = pd.read_csv(input_path, nrows=100)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "bpe = BytePairEncoding(config_dict)\n",
    "\n",
    "train_text_ls, test_text_ls = list(df[\"lyrics\"].iloc[:50]), list(df[\"lyrics\"].iloc[50:])\n",
    "train_words = bpe.fit(train_text_ls)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['you</w>',\n",
       " 'say</w>',\n",
       " 'that</w>',\n",
       " 'we</w>',\n",
       " 've</w>',\n",
       " 'got</w>',\n",
       " 'no thing</w>',\n",
       " 'in</w>',\n",
       " 'c om mo n</w>',\n",
       " 'no</w>']"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "test_words = bpe.transform(test_text_ls)\n",
    "test_words[:10]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "import re\n",
    "\n",
    "vocab = list(bpe.vocab_freq.keys())\n",
    "\n",
    "words = bpe.preprocess(test_text_ls, \"test\")\n",
    "\n",
    "for i, word in enumerate(words):\n",
    "    merge = True\n",
    "\n",
    "    while merge:\n",
    "        tokens = word.split()\n",
    "        merge_count = 0\n",
    "\n",
    "        for j in range(len(tokens)-1):\n",
    "            pair_ = (tokens[j], tokens[j+1])\n",
    "            best_chars = re.escape(\" \".join(pair_))\n",
    "            replace = re.compile(r'(?<!\\S)' + best_chars + r'(?!\\S)')\n",
    "\n",
    "            if \"\".join(pair_) in vocab:\n",
    "                word = replace.sub(\"\".join(pair_), word)\n",
    "                merge_count += 1\n",
    "                break\n",
    "        \n",
    "        if merge_count == 0:\n",
    "            words[i] = word\n",
    "            merge=False"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "algo = Transformer(config_dict)\n",
    "tokens = algo.run()\n",
    "print(tokens.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "### BPE https://arxiv.org/pdf/1508.07909v5\n",
    "\n",
    "# Get Stats\n",
    "from collections import Counter\n",
    "\n",
    "num_vocab = config_dict[\"dataset\"][\"num_vocab\"]\n",
    "operations = config_dict[\"preprocess\"][\"operations\"]\n",
    "\n",
    "## Preprocessing\n",
    "text_ls = [preprocess_text(text, operations) for text in df[\"lyrics\"]]\n",
    "text_lens = [len(text.split()) for text in text_ls]\n",
    "corpus = \" \".join(text_ls)\n",
    "words = corpus.split()\n",
    "words = [\" \".join(list(w))+ \" </w>\" for w in words]\n",
    "vocab_freq = Counter(list(corpus))\n",
    "del vocab_freq[\" \"]\n",
    "vocab_freq[\"</w>\"] = len(words)\n",
    "\n",
    "## Get Stats\n",
    "def get_stats(words):\n",
    "    words_freq = Counter(words)\n",
    "    pair_dict = defaultdict(int)\n",
    "    for word, freq in words_freq.items():\n",
    "        chars = word.split()\n",
    "        for i in range(len(chars)-1):\n",
    "            pair_dict[(chars[i], chars[i+1])] += freq\n",
    "    return pair_dict\n",
    "\n",
    "## Merge Vocab\n",
    "def merge_chars(words):\n",
    "    pair_dict = get_stats(words)\n",
    "    best_pair = max(pair_dict, key=pair_dict.get)\n",
    "    best_pair_count = pair_dict[best_pair]\n",
    "    \n",
    "    vocab_freq[\"\".join(best_pair)] = best_pair_count\n",
    "    vocab_freq[best_pair[0]] -= best_pair_count\n",
    "    vocab_freq[best_pair[1]] -= best_pair_count\n",
    "\n",
    "    if vocab_freq[best_pair[0]] == 0: del vocab_freq[best_pair[0]]\n",
    "    if vocab_freq[best_pair[1]] == 0: del vocab_freq[best_pair[1]]\n",
    "    \n",
    "    best_chars = re.escape(\" \".join(best_pair))\n",
    "    replace = re.compile(r'(?<!\\S)' + best_chars + r'(?!\\S)')\n",
    "\n",
    "    for i, word in enumerate(words):\n",
    "        words[i] = replace.sub(\"\".join(best_pair), word)\n",
    "    \n",
    "    return words\n",
    "\n",
    "## Loop until num_vocab reached\n",
    "num_merges = num_vocab - len(vocab_freq)\n",
    "for _ in range(num_merges):\n",
    "    words = merge_chars(words)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[218.,   1.,  61., ...,   0.,   0.,   0.],\n",
       "       [ 75.,  73., 161., ...,   0.,   0.,   0.],\n",
       "       [192.,  54., 206., ...,   0.,   0.,   0.],\n",
       "       ...,\n",
       "       [134.,  54.,  18., ...,   0.,   0.,   0.],\n",
       "       [ 54.,  17.,   6., ...,   0.,   0.,   0.],\n",
       "       [253., 212., 216., ...,   0.,   0.,   0.]])"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "seq_len = config_dict[\"dataset\"][\"seq_len\"]\n",
    "vocab = [\"<PAD>\"] + list(vocab_freq.keys())\n",
    "\n",
    "word2id = {w:i for i, w in enumerate(vocab)}\n",
    "id2word = {v:k for k,v in word2id.items()}\n",
    "\n",
    "text_tokens = np.zeros((len(text_ls), seq_len))\n",
    "\n",
    "count = 0\n",
    "for i, text_len in enumerate(text_lens):\n",
    "    text = \" \".join(words[count: count+text_len]).split()\n",
    "    count += text_len\n",
    "    text = text[:seq_len]\n",
    "\n",
    "    if len(text) < seq_len:\n",
    "        text = text + [\"<PAD>\"]*(seq_len - len(text))\n",
    "\n",
    "    text_tokens[i] = np.array([word2id[ch] for ch in text])\n",
    "\n",
    "text_tokens"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "wa nt wan ed\n",
      "wan wa nt ed\n",
      "wa nt ed wan\n",
      "ab cdef\n"
     ]
    }
   ],
   "source": [
    "import re\n",
    "\n",
    "words = [\"wa nt wa n ed\", \"wa n wa nt ed\", \"wa nt ed wa n\", \"ab cdef\"]\n",
    "best_pair = (\"wa\", \"n\")\n",
    "\n",
    "best_chars = re.escape(\" \".join(best_pair))\n",
    "replace = re.compile(r'(?<!\\S)' + best_chars + r'(?!\\S)')\n",
    "\n",
    "new_words = []\n",
    "for word in words:\n",
    "    print(replace.sub(\"\".join(best_pair), word))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'you say that we ve got nothing in common no common ground to start from and we re '"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x = sentences[0].split(\"</w>\")\n",
    "x = [i.replace(\" \", \"\") for i in x]\n",
    "\" \".join(x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'you say that we ve got n o thing in c om m on no c om m on g r oun d to st ar t f ro m and we re'"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "\n",
    "df = pd.read_csv(r\"output\\transformer\\Test Predictions.csv\")\n",
    "sentences = list(df[\"Sentence\"])\n",
    "\n",
    "new_sentences = [s.replace(\"</w>\", \"\") for s in sentences]\n",
    "new_sentences[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import torch\n",
    "\n",
    "seq_len = 10\n",
    "d_model = 100\n",
    "a = torch.pow(10000, torch.arange(0, seq_len, 2)/d_model)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "d_model = config_dict[\"model\"][\"d_model\"]\n",
    "seq_len = config_dict[\"dataset\"][\"seq_len\"]\n",
    "\n",
    "pe = torch.zeros(seq_len, d_model)\n",
    "pos = torch.arange(seq_len).unsqueeze(1)\n",
    "denom = torch.pow(10000, torch.arange(0, d_model, 2)/d_model)\n",
    "pe[:, 0::2] = torch.sin(pos/denom)\n",
    "pe[:, 1::2] = torch.cos(pos/denom)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "torch.Size([64, 32, 128])"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import torch\n",
    "import torch.nn as nn\n",
    "import math\n",
    "\n",
    "batch_size = 64\n",
    "seq_len = 32\n",
    "d_model = 128\n",
    "num_heads = 8\n",
    "d_k = d_model//num_heads\n",
    "Q = torch.randn(batch_size, seq_len, d_model)\n",
    "K = torch.randn(batch_size, seq_len, d_model)\n",
    "V = torch.randn(batch_size, seq_len, d_model)\n",
    "\n",
    "Wq = nn.Linear(d_model, d_model)\n",
    "Wk = nn.Linear(d_model, d_model)\n",
    "Wv = nn.Linear(d_model, d_model)\n",
    "\n",
    "Q = Wq(Q)\n",
    "K = Wk(K)\n",
    "V = Wv(V)\n",
    "\n",
    "Q_ls = Q.view(batch_size, seq_len, num_heads, d_k).transpose(1, 2)\n",
    "K_ls = K.view(batch_size, seq_len, num_heads, d_k).transpose(1, 2)\n",
    "V_ls = V.view(batch_size, seq_len, num_heads, d_k).transpose(1, 2)\n",
    "# torch.split(V, split_size_or_sections=d_k, dim=-1)\n",
    "\n",
    "matmul = torch.matmul(Q_ls, K_ls.transpose(-1, -2))\n",
    "mask_ids = torch.triu_indices(seq_len, seq_len, offset=1)\n",
    "matmul[:, :, mask_ids[0], mask_ids[1]] = -1e9\n",
    "scale = matmul/math.sqrt(d_k)\n",
    "softmax = nn.Softmax(dim=-1)(scale)\n",
    "attn_qkv = torch.matmul(softmax, V_ls)\n",
    "\n",
    "attn_qkv_concat = attn_qkv.transpose(1, 2).reshape(batch_size, seq_len, d_model)\n",
    "attn_qkv_concat.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[ 1.1576, -0.1198, -1.6164],\n",
       "        [ 0.1751, -0.1444, -1.3969],\n",
       "        [ 0.2944, -0.8525,  0.5381]])"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x = torch.randn(3,3)\n",
    "x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor([[ 1.1576e+00, -1.0000e+09, -1.0000e+09],\n",
       "        [ 1.7509e-01, -1.4436e-01, -1.0000e+09],\n",
       "        [ 2.9442e-01, -8.5255e-01,  5.3811e-01]])"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ids = torch.triu_indices(3,3, offset=1)\n",
    "x[ids[0], ids[1]] = -1e9\n",
    "x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "mask_ids = torch.triu_indices(seq_len, seq_len)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### BERT"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "### Convert parquet to CSV file with 25000 rows\n",
    "\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "path = \"./data/wiki_en/3000000_to_3250000.parquet\"\n",
    "df = pd.read_parquet(path, engine='pyarrow')\n",
    "\n",
    "rand_ids = np.random.choice(250000, 25000, replace=False)\n",
    "new_df = df.iloc[rand_ids][[\"title\", \"text\"]].reset_index(drop=True)\n",
    "\n",
    "new_df.to_csv(\"./data/wiki_en.csv\", index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "import yaml\n",
    "import re\n",
    "import pandas as pd\n",
    "import sys\n",
    "sys.path.append(\"./src/\")\n",
    "from tqdm import tqdm\n",
    "from collections import defaultdict, Counter\n",
    "import numpy as np\n",
    "import torch\n",
    "from torch.utils.data import Dataset, DataLoader\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "from src.preprocess.utils import preprocess_text, WordPiece\n",
    "from src.core.bert.dataset_pretrain import BERTPretrainDataset, PreprocessBERTPretrain, create_dataloader_pretrain\n",
    "from src.core.bert.dataset_finetune import PreprocessBERTFinetune, create_data_loader_finetune\n",
    "from core.transformer.model import EncoderLayer, PositionalEncoding\n",
    "from src.core.bert.model import BERTPretrainModel, BERTFinetuneModel\n",
    "from src.core.bert.pretrain import BERTPretrainTrainer\n",
    "from src.core.bert.bert import BERT"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "config_path = \"./configs/bert.yaml\"\n",
    "with open(config_path, \"r\") as stream:\n",
    "    config_dict = yaml.safe_load(stream)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "bert = BERT(config_dict)\n",
    "bert.run()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<All keys matched successfully>"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model = bert.trainer_pretrain.model\n",
    "layer = model.encoder_layers[0]\n",
    "layer.load_state_dict(model.encoder_layers[0].state_dict())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Pretraining"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(256, 128) (256,)\n",
      "torch.Size([32, 128]) torch.Size([32, 128]) tensor([0., 0., 1., 0., 0., 1., 0., 0., 0., 0., 0., 0., 1., 1., 0., 0., 0., 1.,\n",
      "        1., 1., 1., 1., 1., 0., 0., 0., 1., 1., 0., 1., 0., 1.])\n"
     ]
    }
   ],
   "source": [
    "preprocess_bert = PreprocessBERTPretrain(config_dict)\n",
    "text_tokens, nsp_labels = preprocess_bert.get_data()\n",
    "word2id = preprocess_bert.word2id\n",
    "print(text_tokens.shape, nsp_labels.shape)\n",
    "\n",
    "# bert_ds = BERTDataset(torch.tensor(text_tokens), torch.tensor(nsp_labels), config_dict, word2id)\n",
    "# text_token, lbl_mask, nsp_label = bert_ds.__getitem__(0)\n",
    "# print(text_token.shape, lbl_mask.shape, nsp_label)\n",
    "\n",
    "train_loader, val_loader, test_loader = create_dataloader_pretrain(text_tokens, nsp_labels, config_dict, word2id)\n",
    "text_token_bs, lbl_mask_bs, nsp_label_bs = next(iter(train_loader))\n",
    "print(text_token_bs.shape, lbl_mask_bs.shape, nsp_label_bs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = BERTPretrainModel(config_dict)\n",
    "lr = config_dict[\"train\"][\"lr\"]\n",
    "optim = torch.optim.Adam(model.parameters(), lr=lr)\n",
    "\n",
    "trainer = BERTPretrainTrainer(model, optim, config_dict)\n",
    "history = trainer.fit(train_loader, val_loader)\n",
    "\n",
    "y_tokens_true, y_nsp_true, y_tokens_pred, y_nsp_pred = trainer.predict(test_loader)\n",
    "print(y_tokens_true.shape, y_nsp_true.shape, y_tokens_pred.shape, y_nsp_pred.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "e:\\anaconda\\envs\\nlp_env\\lib\\site-packages\\torch\\nn\\_reduction.py:51: UserWarning: size_average and reduce args will be deprecated, please use reduction='mean' instead.\n",
      "  warnings.warn(warning.format(ret))\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "tensor(0.6936, grad_fn=<BinaryCrossEntropyBackward0>) tensor(5.5453, grad_fn=<DivBackward1>)\n"
     ]
    }
   ],
   "source": [
    "model = BERTPretrainModel(config_dict)\n",
    "\n",
    "output, nsp_output = model(text_token_bs)\n",
    "\n",
    "nsp_loss_fn = nn.BCELoss()\n",
    "nsp_loss = nsp_loss_fn(nsp_output.squeeze(), nsp_label_bs)\n",
    "\n",
    "mask_flatten = torch.flatten(lbl_mask_bs).bool()\n",
    "mask_flatten = torch.stack([mask_flatten]*config_dict[\"dataset\"][\"num_vocab\"], dim=1)\n",
    "\n",
    "y_pred = torch.flatten(output, end_dim=1)\n",
    "y_true = torch.flatten(text_token_bs)\n",
    "y_true = F.one_hot(\n",
    "            y_true,\n",
    "            num_classes = config_dict[\"dataset\"][\"num_vocab\"]\n",
    "        ).to(torch.float)\n",
    "\n",
    "y_pred_mask = torch.masked_select(y_pred, mask_flatten).reshape(-1, y_pred.shape[1])\n",
    "y_true_mask = torch.masked_select(y_true, mask_flatten).reshape(-1, y_true.shape[1])\n",
    "\n",
    "ce_loss_fn = nn.CrossEntropyLoss(reduce=\"sum\")\n",
    "ce_loss = ce_loss_fn(y_pred_mask, y_true_mask)\n",
    "\n",
    "print(nsp_loss, ce_loss)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "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>title</th>\n",
       "      <th>text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Emotional Girl</td>\n",
       "      <td>\"Emotional Girl\" is a song co-written and reco...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Railway Place, Melbourne</td>\n",
       "      <td>Railway Place is a residential street in the w...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Maingkaing</td>\n",
       "      <td>Maingkaing is a village in Homalin Township, H...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Euphoria (Enrique Iglesias album)</td>\n",
       "      <td>Euphoria is the ninth studio album by Spanish ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>MC Bogy</td>\n",
       "      <td>MC Bogy (1979 in Berlin), born Moritz Christop...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                               title  \\\n",
       "0                     Emotional Girl   \n",
       "1           Railway Place, Melbourne   \n",
       "2                         Maingkaing   \n",
       "3  Euphoria (Enrique Iglesias album)   \n",
       "4                            MC Bogy   \n",
       "\n",
       "                                                text  \n",
       "0  \"Emotional Girl\" is a song co-written and reco...  \n",
       "1  Railway Place is a residential street in the w...  \n",
       "2  Maingkaing is a village in Homalin Township, H...  \n",
       "3  Euphoria is the ninth studio album by Spanish ...  \n",
       "4  MC Bogy (1979 in Berlin), born Moritz Christop...  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "input_path = config_dict[\"paths\"][\"input_file\"]\n",
    "\n",
    "df = pd.read_csv(input_path)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "operations = config_dict[\"preprocess\"][\"operations\"]\n",
    "\n",
    "text_ls = list(df[\"text\"].iloc[:96])\n",
    "text_ls = [preprocess_text(doc, operations) for doc in text_ls]\n",
    "text_ls = [i for i in text_ls if len(text_ls)]\n",
    "text_lens = [len(text.split()) for text in text_ls]\n",
    "\n",
    "wordpiece = WordPiece(config_dict)\n",
    "corpus = wordpiece.fit(text_ls)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.preprocessing import LabelEncoder\n",
    "\n",
    "seq_len = config_dict[\"dataset\"][\"seq_len\"]\n",
    "num_vocab = config_dict[\"dataset\"][\"num_vocab\"]\n",
    "\n",
    "vocab = [\"<PAD>\", \"<UNK>\", \"<CLS>\", \"<SEP>\", \"<MASK>\"] + list(wordpiece.vocab_freq.keys())\n",
    "word2id = {w:i for i, w in enumerate(vocab)}\n",
    "id2word = {v:k for k,v in word2id.items()}\n",
    "\n",
    "half_seq_len = seq_len//2 - 1\n",
    "text_tokens_a = np.zeros((len(text_ls), half_seq_len+1))\n",
    "text_tokens_b = np.zeros((len(text_ls), half_seq_len+1))\n",
    "count = 0\n",
    "new_text_ls = []\n",
    "for i, text_len in enumerate(text_lens):\n",
    "    tokens_ls = corpus[count: count+text_len]\n",
    "    tokens = [i for ls in tokens_ls for i in ls]\n",
    "    text = \" \".join(tokens).split()\n",
    "    count += text_len\n",
    "\n",
    "    text = text[:seq_len]\n",
    "\n",
    "    if len(text) < seq_len:\n",
    "        text = text + [\"<PAD>\"]*(seq_len - len(text))\n",
    "\n",
    "    text_tokens_a[i] = np.array([word2id[ch] for ch in [\"<CLS>\"] + text[:half_seq_len]])\n",
    "    text_tokens_b[i] = np.array([word2id[ch] for ch in [\"<SEP>\"] + text[half_seq_len:2*half_seq_len]])\n",
    "\n",
    "reorder_tokens_b = np.random.choice(len(text_ls), len(text_ls), replace=False)\n",
    "text_tokens_a = np.concatenate([text_tokens_a, text_tokens_a], axis=0)\n",
    "text_tokens_b = np.concatenate([text_tokens_b, text_tokens_b[reorder_tokens_b]], axis=0)\n",
    "text_tokens = np.concatenate([text_tokens_a, text_tokens_b], axis=-1)\n",
    "nsp_labels = np.array([\"IsNext\"]*len(text_ls) + [\"NotNext\"]*len(text_ls))\n",
    "lencoder = LabelEncoder()\n",
    "nsp_labels = lencoder.fit_transform(nsp_labels)\n",
    "\n",
    "pred_prob = config_dict[\"preprocess\"][\"replace_token\"][\"prediction\"]\n",
    "pred_mask = config_dict[\"preprocess\"][\"replace_token\"][\"mask\"]\n",
    "pred_random = config_dict[\"preprocess\"][\"replace_token\"][\"random\"]\n",
    "num_pred_tokens = int(pred_prob*seq_len)\n",
    "num_mask_tokens = int(num_pred_tokens*pred_mask)\n",
    "num_rand_tokens = int(num_pred_tokens*pred_random)\n",
    "\n",
    "label_mask_a, label_mask_b = [], []\n",
    "for i in range(2*len(text_ls)):\n",
    "    label_mask_a.append(np.random.choice(half_seq_len, num_pred_tokens).tolist())\n",
    "    label_mask_b.append(np.random.choice(half_seq_len, num_pred_tokens).tolist())\n",
    "label_mask_a, label_mask_b = 1 + np.array(label_mask_a), 1 + seq_len//2 + np.array(label_mask_b)\n",
    "label_mask_ids = np.concatenate([label_mask_a, label_mask_b], axis=-1)\n",
    "label_mask = np.zeros_like(text_tokens)\n",
    "rows = np.stack([np.arange(2*len(text_ls))]*2*num_pred_tokens, axis=-1).flatten()\n",
    "label_mask[rows, label_mask_ids.flatten()] = 1\n",
    "\n",
    "rand_tokens = np.random.choice(np.arange(5, num_vocab), num_rand_tokens*2*len(text_ls))\n",
    "\n",
    "rows = np.stack([np.arange(2*len(text_ls))]*num_mask_tokens, axis=-1).flatten()\n",
    "text_tokens[rows, label_mask_ids[:, :num_mask_tokens].flatten()] = word2id[\"<MASK>\"]\n",
    "rows = np.stack([np.arange(2*len(text_ls))]*num_rand_tokens, axis=-1).flatten()\n",
    "text_tokens[rows, label_mask_ids[:, num_mask_tokens:num_mask_tokens+num_rand_tokens].flatten()] = rand_tokens\n",
    "\n",
    "## text_tokens, label_mask, nsp_labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "### BERT Dataset\n",
    "\n",
    "import torch\n",
    "\n",
    "num_vocab = config_dict[\"dataset\"][\"num_vocab\"]\n",
    "seq_len = config_dict[\"dataset\"][\"seq_len\"]\n",
    "half_seq_len = seq_len//2 - 1\n",
    "\n",
    "pred_prob = config_dict[\"preprocess\"][\"replace_token\"][\"prediction\"]\n",
    "pred_mask = config_dict[\"preprocess\"][\"replace_token\"][\"mask\"]\n",
    "pred_random = config_dict[\"preprocess\"][\"replace_token\"][\"random\"]\n",
    "num_extra_tokens = config_dict[\"dataset\"][\"num_extra_tokens\"]\n",
    "\n",
    "num_pred_tokens_half = int(pred_prob*half_seq_len)\n",
    "num_mask_tokens = int(2*num_pred_tokens_half*pred_mask)\n",
    "num_rand_tokens = int(2*num_pred_tokens_half*pred_random)\n",
    "\n",
    "sample_text_tokens = torch.tensor(text_tokens[0]).to(torch.int64)\n",
    "\n",
    "lbl_mask_ids_a = 1 + torch.randperm(half_seq_len)[:num_pred_tokens_half]\n",
    "lbl_mask_ids_b = 1 + seq_len//2 + torch.randperm(half_seq_len)[:num_pred_tokens_half]\n",
    "lbl_mask_ids = torch.concat([lbl_mask_ids_a, lbl_mask_ids_b], axis=0)\n",
    "label_mask = torch.zeros_like(sample_text_tokens)\n",
    "label_mask[lbl_mask_ids] = 1\n",
    "\n",
    "rand_tokens = torch.randperm(num_vocab-num_extra_tokens)[:num_rand_tokens] + num_extra_tokens\n",
    "sample_text_tokens[lbl_mask_ids[:num_mask_tokens]] = word2id[\"<MASK>\"]\n",
    "sample_text_tokens[lbl_mask_ids[num_mask_tokens:num_mask_tokens+num_rand_tokens]] = rand_tokens"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "### Word Piece Tokenization\n",
    "\n",
    "words = \" \".join(text_ls).split()\n",
    "corpus = []\n",
    "\n",
    "## Build Vocabulary and Corpus\n",
    "word_freq = Counter(words)\n",
    "vocab_freq = defaultdict()\n",
    "for word in words:\n",
    "    chars = []\n",
    "    for i, ch in enumerate(word):\n",
    "        if i != 0: ch = f\"##{ch}\"\n",
    "        vocab_freq[ch] = vocab_freq.get(ch, 0) + 1\n",
    "        chars.append(ch)\n",
    "    corpus.append(chars)\n",
    "\n",
    "### Build Pairs\n",
    "def get_stats(corpus):\n",
    "    pair_freq = defaultdict(int)\n",
    "    for corp in corpus:\n",
    "        if len(corp) == 1:\n",
    "            continue\n",
    "        for i in range(len(corp)-1):\n",
    "            pair_freq[(corp[i], corp[i+1])] += 1\n",
    "    return pair_freq\n",
    "\n",
    "## Get Likelihood\n",
    "def get_likelihood(pair, vocab_freq, pair_freq):\n",
    "    p12 = pair_freq[pair]\n",
    "    p1, p2 = vocab_freq[pair[0]], vocab_freq[pair[1]]\n",
    "    lkhd = p12/(p1*p2)\n",
    "\n",
    "    return lkhd\n",
    "\n",
    "## Combine Tokens\n",
    "def combine(pair):\n",
    "    token1, token2 = pair\n",
    "    return token1 + token2[2:] if token2.startswith(\"##\") else token1 + token2\n",
    "\n",
    "### Build Vocab\n",
    "def build_vocab(corpus):\n",
    "    pair_freq = get_stats(corpus)\n",
    "    best_pair = max(pair_freq.keys(), key=lambda x: get_likelihood(x, vocab_freq, pair_freq))\n",
    "    new_ch = combine(best_pair)\n",
    "    best_pair_count = pair_freq[best_pair]\n",
    "\n",
    "    for i, corp in enumerate(corpus):\n",
    "        if len(corp) == 1:\n",
    "            continue\n",
    "        j = 0\n",
    "        while j < len(corp) - 1:\n",
    "            if (corp[j], corp[j+1]) == best_pair:\n",
    "                corp = corp[:j] + [new_ch] + corp[j+2:]\n",
    "            else:\n",
    "                j += 1\n",
    "        corpus[i] = corp\n",
    "\n",
    "    vocab_freq[new_ch] = best_pair_count\n",
    "    vocab_freq[best_pair[0]] -= best_pair_count\n",
    "    vocab_freq[best_pair[1]] -= best_pair_count\n",
    "\n",
    "    if vocab_freq[best_pair[0]] == 0: del vocab_freq[best_pair[0]]\n",
    "    if best_pair[0] != best_pair[1]: \n",
    "        if vocab_freq[best_pair[1]] == 0: del vocab_freq[best_pair[1]]\n",
    "    \n",
    "    return corpus\n",
    "\n",
    "## Merge Characters\n",
    "def merge_chars(word, vocab_freq):\n",
    "    j = 0\n",
    "    while j < len(word) - 1:\n",
    "        ch1, ch2 = word[j], word[j+1]\n",
    "        new_ch = combine((ch1, ch2))\n",
    "        if new_ch in vocab_freq:\n",
    "            word = word[:j] + [new_ch] + word[j+2:]\n",
    "        else:\n",
    "            j+=1\n",
    "    return word\n",
    "\n",
    "while len(vocab_freq) > 256:\n",
    "    corpus = build_vocab(corpus)\n",
    "    print(f\"Vocab: {len(vocab_freq)}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 226,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['emotional', 'regional']"
      ]
     },
     "execution_count": 226,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### Merging Tokens after Inference\n",
    "\n",
    "seq = corpus[0] + corpus[99]\n",
    "\n",
    "start_id = 0\n",
    "words = []\n",
    "for i, ch in enumerate(seq):\n",
    "    if \"##\" != ch[:2] and i != 0:\n",
    "        tokens = seq[start_id:i]\n",
    "        word = \"\".join([w if i==0 else w[2:] for i,w in enumerate(tokens)])\n",
    "        words.append(word)\n",
    "        start_id = i\n",
    "final_word = \"\".join([w if i==0 else w[2:] for i,w in enumerate(seq[start_id:])])\n",
    "words.append(final_word)\n",
    "words"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "### BERT Pretraining Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "### BERT Pretraining Training"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Finetuning"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([32, 128])\n",
      "tensor([ 0., 45.,  9., 42.,  0., 36., 31.,  0., 46.,  8.])\n",
      "tensor([ 6., 59., 34., 57., 11., 39., 44., 33., 55., 18.])\n"
     ]
    }
   ],
   "source": [
    "wordpiece = bert.bert_pretrain_ds.wordpiece\n",
    "word2id = bert.bert_pretrain_ds.word2id\n",
    "\n",
    "bert_finetune_ds = PreprocessBERTFinetune(config_dict, wordpiece, word2id)\n",
    "tokens, start_ids, end_ids, topics = bert_finetune_ds.get_data()\n",
    "\n",
    "train_loader, val_loader, test_loader = create_data_loader_finetune(tokens, start_ids, end_ids, topics)\n",
    "train_tokens, train_start_ids, train_end_ids = next(iter(train_loader))\n",
    "print(train_tokens.shape)\n",
    "print(train_start_ids[:10])\n",
    "print(train_end_ids[:10])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "### BERT Finetune Loss Function\n",
    "\n",
    "bert_finetune_model = BERTFinetuneModel(config_dict)\n",
    "\n",
    "start_probs, end_probs = bert_finetune_model(train_tokens.to(torch.long))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "tensor(4.8106, grad_fn=<NllLossBackward0>)"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import torch.nn as nn\n",
    "\n",
    "loss_fn = nn.NLLLoss()\n",
    "loss_fn(start_probs, train_start_ids.to(torch.long))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [],
   "source": [
    "from torch.autograd import Variable \n",
    "\n",
    "d_model = config_dict[\"model\"][\"d_model\"]\n",
    "seq_len = config_dict[\"dataset\"][\"seq_len\"]\n",
    "num_samples = enc_output.size(0)\n",
    "\n",
    "start = nn.init.xavier_uniform_(torch.empty(d_model, 1))\n",
    "start = Variable(start, requires_grad=True)\n",
    "\n",
    "end = nn.init.xavier_uniform_(torch.empty(d_model, 1))\n",
    "end = Variable(end, requires_grad=True)\n",
    "\n",
    "cxt_enc_output = enc_output[:, seq_len//2:, :]\n",
    "start_muls = torch.matmul(cxt_enc_output, start).squeeze()\n",
    "start_probs = nn.LogSoftmax(dim=1)(start_muls)\n",
    "end_muls = torch.matmul(cxt_enc_output, end).squeeze()\n",
    "end_probs = nn.LogSoftmax(dim=1)(end_muls)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_max_end_index(tensor, start_indices):\n",
    "    num_rows, num_cols = tensor.shape\n",
    "    mask = torch.arange(num_cols).expand(num_rows, num_cols) <= start_indices.unsqueeze(1)\n",
    "    masked_tensor = tensor.clone().detach().requires_grad_(False)\n",
    "    masked_tensor[mask] = -1e9\n",
    "    _, max_indices = torch.max(masked_tensor, dim=1)\n",
    "    \n",
    "    return max_indices \n",
    "\n",
    "start_ids_pred = torch.argmax(start_probs, dim=1)\n",
    "start_max_muls = start_muls[torch.arange(num_samples), start_ids_pred]\n",
    "sum_muls = start_max_muls.view(-1, 1) + end_muls\n",
    "end_ids_pred = get_max_end_index(sum_muls, start_ids_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "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>Topic</th>\n",
       "      <th>Context</th>\n",
       "      <th>Question</th>\n",
       "      <th>Answer Start ID</th>\n",
       "      <th>Num Words</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Super_Bowl_50</td>\n",
       "      <td>Super Bowl 50 was an American football game to...</td>\n",
       "      <td>Which NFL team represented the AFC at Super Bo...</td>\n",
       "      <td>28</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Super_Bowl_50</td>\n",
       "      <td>Super Bowl 50 was an American football game to...</td>\n",
       "      <td>Which NFL team represented the NFC at Super Bo...</td>\n",
       "      <td>37</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Super_Bowl_50</td>\n",
       "      <td>Super Bowl 50 was an American football game to...</td>\n",
       "      <td>Which NFL team won Super Bowl 50?</td>\n",
       "      <td>28</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Super_Bowl_50</td>\n",
       "      <td>Super Bowl 50 was an American football game to...</td>\n",
       "      <td>What was the theme of Super Bowl 50?</td>\n",
       "      <td>79</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Super_Bowl_50</td>\n",
       "      <td>Super Bowl 50 was an American football game to...</td>\n",
       "      <td>What is the AFC short for?</td>\n",
       "      <td>23</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           Topic                                            Context  \\\n",
       "0  Super_Bowl_50  Super Bowl 50 was an American football game to...   \n",
       "1  Super_Bowl_50  Super Bowl 50 was an American football game to...   \n",
       "2  Super_Bowl_50  Super Bowl 50 was an American football game to...   \n",
       "3  Super_Bowl_50  Super Bowl 50 was an American football game to...   \n",
       "4  Super_Bowl_50  Super Bowl 50 was an American football game to...   \n",
       "\n",
       "                                            Question  Answer Start ID  \\\n",
       "0  Which NFL team represented the AFC at Super Bo...               28   \n",
       "1  Which NFL team represented the NFC at Super Bo...               37   \n",
       "2                  Which NFL team won Super Bowl 50?               28   \n",
       "3               What was the theme of Super Bowl 50?               79   \n",
       "4                         What is the AFC short for?               23   \n",
       "\n",
       "   Num Words  \n",
       "0          2  \n",
       "1          2  \n",
       "2          2  \n",
       "3          2  \n",
       "4          3  "
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### Extract Data\n",
    "import json\n",
    "\n",
    "json_path = config_dict[\"finetune\"][\"paths\"][\"input_file\"]\n",
    "num_topics = config_dict[\"finetune\"][\"dataset\"][\"num_topics\"]\n",
    "\n",
    "with open(json_path, 'r') as f:\n",
    "    data = json.load(f)\n",
    "\n",
    "data = data[\"data\"]\n",
    "topics, contexts, ques, ans_texts,   = [], [], [], []\n",
    "ans_start_ids, num_words, valid_ids = [], [], []\n",
    "\n",
    "for topic in data[:num_topics]:\n",
    "    for para in topic[\"paragraphs\"]:\n",
    "        for que in para[\"qas\"]:\n",
    "            topics.append(topic[\"title\"])\n",
    "            contexts.append(para[\"context\"])\n",
    "            ques.append(que[\"question\"])\n",
    "            ans_texts.append(que[\"answers\"][0][\"text\"])\n",
    "\n",
    "for j, (ans, context) in enumerate(zip(ans_texts, contexts)):\n",
    "    num_word = len(ans.split())\n",
    "    tokens = context.split()\n",
    "    \n",
    "    for i in range(len(tokens) - num_word):\n",
    "        is_ans = \" \".join(tokens[i:i+num_word])\n",
    "        if is_ans == ans:\n",
    "            ans_start_ids.append(i)\n",
    "            num_words.append(num_word)\n",
    "            is_present = True\n",
    "            valid_ids.append(j)\n",
    "            break\n",
    "valid_ids = np.array(valid_ids)\n",
    "\n",
    "df = pd.DataFrame.from_dict({\n",
    "    \"Topic\": np.array(topics)[valid_ids],\n",
    "    \"Context\": np.array(contexts)[valid_ids],\n",
    "    \"Question\": np.array(ques)[valid_ids],\n",
    "    \"Answer Start ID\": ans_start_ids,\n",
    "    \"Num Words\": num_words\n",
    "})\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "3432it [00:01, 2387.82it/s]"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(594, 128)\n",
      "[64, 104, 67, 64, 118, 64, 118, 64, 118, 64]\n",
      "[73, 108, 74, 67, 127, 67, 127, 67, 127, 67]\n",
      "['Super_Bowl_50', 'Super_Bowl_50', 'Super_Bowl_50', 'Super_Bowl_50', 'Super_Bowl_50', 'Super_Bowl_50', 'Super_Bowl_50', 'Super_Bowl_50', 'Super_Bowl_50', 'Super_Bowl_50']\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "### Get Data\n",
    "\n",
    "operations = config_dict[\"finetune\"][\"preprocess\"][\"operations\"]\n",
    "seq_len = config_dict[\"dataset\"][\"seq_len\"]\n",
    "\n",
    "# token_text.replace(\"''\", '\"').replace(\"``\", '\"')\n",
    "df[\"Context\"] = df[\"Context\"].map(lambda x: preprocess_text(x, operations))\n",
    "df[\"Question\"] = df[\"Question\"].map(lambda x: preprocess_text(x, operations))\n",
    "num_samples = len(df)\n",
    "\n",
    "wordpiece = bert.bert_pretrain_ds.wordpiece\n",
    "word2id = bert.bert_pretrain_ds.word2id\n",
    "id2word = bert.bert_pretrain_ds.id2word\n",
    "vocab = word2id.keys()\n",
    "\n",
    "half_seq_len = seq_len//2 - 1\n",
    "\n",
    "ques_tokens, cxt_tokens = [], []\n",
    "start_ids, end_ids = [], []\n",
    "topics = []\n",
    "\n",
    "cxt_lens = [len(text.split()) for text in df[\"Context\"]]\n",
    "ques_lens = [len(text.split()) for text in df[\"Question\"]]\n",
    "\n",
    "cxt_corpus = wordpiece.transform(df[\"Context\"])\n",
    "ques_corpus = wordpiece.transform(df[\"Question\"])\n",
    "\n",
    "cxt_count, ques_count = 0, 0\n",
    "for id, (cxt_len, ques_len) in tqdm(enumerate(zip(cxt_lens, ques_lens))):\n",
    "    cxt_token = cxt_corpus[cxt_count: cxt_count+cxt_len]\n",
    "    cxt_token = [i for ls in cxt_token for i in ls]\n",
    "    cxt_token = cxt_token[:half_seq_len]\n",
    "    if len(cxt_token) < half_seq_len:\n",
    "        cxt_token = cxt_token + [\"<PAD>\"]*(half_seq_len - len(cxt_token))\n",
    "\n",
    "    ques_token = ques_corpus[ques_count: ques_count+ques_len]\n",
    "    ques_token = [i for ls in ques_token for i in ls]\n",
    "    ques_token = ques_token[:half_seq_len]\n",
    "    if len(ques_token) < half_seq_len:\n",
    "        ques_token = ques_token + [\"<PAD>\"]*(half_seq_len - len(ques_token))\n",
    "    \n",
    "    start_word_id = df.iloc[id][\"Answer Start ID\"]\n",
    "    num_word = df.iloc[id][\"Num Words\"]\n",
    "    start_id = len([i for ls in cxt_corpus[cxt_count:cxt_count+start_word_id] for i in ls])\n",
    "    end_id = len([i for ls in cxt_corpus[cxt_count:cxt_count+start_word_id+num_word] for i in ls])\n",
    "\n",
    "    if end_id <= half_seq_len:\n",
    "        cxt_tokens.append([word2id[ch] if ch in vocab else word2id[\"<UNK>\"] for ch in [\"<CLS>\"] + cxt_token])\n",
    "        ques_tokens.append([word2id[ch] if ch in vocab else word2id[\"<UNK>\"] for ch in [\"<SEP>\"] + ques_token])\n",
    "        start_ids.append(half_seq_len + 1 + start_id) \n",
    "        end_ids.append(half_seq_len + 1 + end_id)\n",
    "        topics.append(df.iloc[id][\"Topic\"])\n",
    "    \n",
    "    cxt_count += cxt_len\n",
    "    ques_count += ques_len\n",
    "\n",
    "tokens = np.concatenate([ques_tokens, cxt_tokens], axis=-1)\n",
    "\n",
    "print(tokens.shape)\n",
    "print(start_ids[:10])\n",
    "print(end_ids[:10])\n",
    "print(topics[:10])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### GPT"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import yaml\n",
    "import glob\n",
    "import re\n",
    "import time\n",
    "import pandas as pd\n",
    "import sys\n",
    "sys.path.append(\"./src/\")\n",
    "import tqdm\n",
    "from collections import defaultdict, Counter\n",
    "import numpy as np\n",
    "import torch\n",
    "import torch.nn as nn\n",
    "import torch.nn.functional as F\n",
    "from torch.utils.data import Dataset, DataLoader\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "from src.preprocess.utils import preprocess_text, BytePairEncoding\n",
    "from src.core.gpt.dataset import PreprocessGPT, create_dataloader\n",
    "from src.core.gpt.model import GPTModel, GPTTrainer\n",
    "from src.core.gpt.gpt import GPT"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Training:   0%|          | 0/15 [00:00<?, ?it/s]e:\\anaconda\\envs\\nlp_env\\lib\\site-packages\\torch\\nn\\_reduction.py:51: UserWarning: size_average and reduce args will be deprecated, please use reduction='mean' instead.\n",
      "  warnings.warn(warning.format(ret))\n",
      "Training: 100%|██████████| 15/15 [00:05<00:00,  2.81it/s]\n",
      "Validation: 100%|██████████| 3/3 [00:03<00:00,  1.02s/it]\n",
      "Training: 100%|██████████| 15/15 [00:03<00:00,  4.50it/s]\n",
      "Validation: 100%|██████████| 3/3 [00:02<00:00,  1.15it/s]\n",
      "Training: 100%|██████████| 15/15 [00:03<00:00,  4.70it/s]\n",
      "Validation: 100%|██████████| 3/3 [00:02<00:00,  1.13it/s]\n",
      "Training: 100%|██████████| 15/15 [00:03<00:00,  4.27it/s]\n",
      "Validation: 100%|██████████| 3/3 [00:02<00:00,  1.01it/s]\n",
      "Training: 100%|██████████| 15/15 [00:05<00:00,  2.87it/s]\n",
      "Validation: 100%|██████████| 3/3 [00:03<00:00,  1.14s/it]\n",
      "Training: 100%|██████████| 15/15 [00:04<00:00,  3.21it/s]\n",
      "Validation: 100%|██████████| 3/3 [00:03<00:00,  1.27s/it]\n",
      "Training: 100%|██████████| 15/15 [00:06<00:00,  2.28it/s]\n",
      "Validation: 100%|██████████| 3/3 [00:03<00:00,  1.26s/it]\n",
      "Training: 100%|██████████| 15/15 [00:04<00:00,  3.47it/s]\n",
      "Validation: 100%|██████████| 3/3 [00:03<00:00,  1.32s/it]\n",
      "Training: 100%|██████████| 15/15 [00:08<00:00,  1.77it/s]\n",
      "Validation: 100%|██████████| 3/3 [00:03<00:00,  1.14s/it]\n",
      "Training: 100%|██████████| 15/15 [00:04<00:00,  3.65it/s]\n",
      "Validation: 100%|██████████| 3/3 [00:02<00:00,  1.02it/s]\n",
      "Training: 100%|██████████| 15/15 [00:03<00:00,  3.84it/s]\n",
      "Validation: 100%|██████████| 3/3 [00:03<00:00,  1.00s/it]\n",
      "Training: 100%|██████████| 15/15 [00:03<00:00,  4.14it/s]\n",
      "Validation: 100%|██████████| 3/3 [00:03<00:00,  1.05s/it]\n",
      "Training: 100%|██████████| 15/15 [00:03<00:00,  4.04it/s]\n",
      "Validation: 100%|██████████| 3/3 [00:04<00:00,  1.53s/it]\n",
      "Training: 100%|██████████| 15/15 [00:04<00:00,  3.68it/s]\n",
      "Validation: 100%|██████████| 3/3 [00:02<00:00,  1.02it/s]\n",
      "Training: 100%|██████████| 15/15 [00:05<00:00,  2.90it/s]\n",
      "Validation: 100%|██████████| 3/3 [00:04<00:00,  1.57s/it]\n",
      "Training: 100%|██████████| 15/15 [00:06<00:00,  2.17it/s]\n",
      "Validation: 100%|██████████| 3/3 [00:03<00:00,  1.18s/it]\n",
      "Training: 100%|██████████| 15/15 [00:03<00:00,  3.85it/s]\n",
      "Validation: 100%|██████████| 3/3 [00:04<00:00,  1.38s/it]\n",
      "Training: 100%|██████████| 15/15 [00:04<00:00,  3.31it/s]\n",
      "Validation: 100%|██████████| 3/3 [00:03<00:00,  1.05s/it]\n",
      "Training: 100%|██████████| 15/15 [00:04<00:00,  3.54it/s]\n",
      "Validation: 100%|██████████| 3/3 [00:03<00:00,  1.08s/it]\n",
      "Training: 100%|██████████| 15/15 [00:04<00:00,  3.21it/s]\n",
      "Validation: 100%|██████████| 3/3 [00:03<00:00,  1.23s/it]\n",
      "e:\\anaconda\\envs\\nlp_env\\lib\\site-packages\\joblib\\externals\\loky\\backend\\context.py:136: UserWarning: Could not find the number of physical cores for the following reason:\n",
      "[WinError 2] The system cannot find the file specified\n",
      "Returning the number of logical cores instead. You can silence this warning by setting LOKY_MAX_CPU_COUNT to the number of cores you want to use.\n",
      "  warnings.warn(\n",
      "  File \"e:\\anaconda\\envs\\nlp_env\\lib\\site-packages\\joblib\\externals\\loky\\backend\\context.py\", line 257, in _count_physical_cores\n",
      "    cpu_info = subprocess.run(\n",
      "  File \"e:\\anaconda\\envs\\nlp_env\\lib\\subprocess.py\", line 501, in run\n",
      "    with Popen(*popenargs, **kwargs) as process:\n",
      "  File \"e:\\anaconda\\envs\\nlp_env\\lib\\subprocess.py\", line 947, in __init__\n",
      "    self._execute_child(args, executable, preexec_fn, close_fds,\n",
      "  File \"e:\\anaconda\\envs\\nlp_env\\lib\\subprocess.py\", line 1416, in _execute_child\n",
      "    hp, ht, pid, tid = _winapi.CreateProcess(executable, args,\n",
      "Generation: 100%|██████████| 2/2 [00:03<00:00,  2.00s/it]\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1800x3600 with 15 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "config_path = \"./configs/gpt.yaml\"\n",
    "with open(config_path, \"r\") as stream:\n",
    "    config_dict = yaml.safe_load(stream)\n",
    "\n",
    "gpt = GPT(config_dict)\n",
    "gpt.run()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "253"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(gpt.gpt_ds.word2id.keys())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "251"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(gpt.gpt_ds.bpe.vocab_freq.keys())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(256, 64)"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "gpt.model.embed_layer.weight.detach().numpy().shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "torch.Size([32, 65]) torch.Size([20, 96])\n"
     ]
    }
   ],
   "source": [
    "gpt_ds = PreprocessGPT(config_dict)\n",
    "train_text = gpt_ds.get_data()\n",
    "test_text = gpt_ds.get_test_data()\n",
    "\n",
    "val_split = config_dict[\"dataset\"][\"val_split\"]\n",
    "train_loader, val_loader = create_dataloader(train_text, \"train\", val_split, batch_size=32, seed=2024)\n",
    "test_loader = create_dataloader(test_text, \"test\", batch_size=32, seed=2024)\n",
    "\n",
    "train_X = next(iter(train_loader))\n",
    "test_X = next(iter(test_loader))\n",
    "\n",
    "print(train_X[0].shape, test_X[0].shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "gpt = GPTModel(config_dict)\n",
    "tokens = train_X[0].to(torch.long)\n",
    "output = gpt(tokens[:, :-1])\n",
    "print(output.shape)\n",
    "\n",
    "test_tokens_pred = gpt.generate(test_X[0].to(torch.long))\n",
    "print(test_tokens_pred.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 32/32 [00:02<00:00, 13.06it/s]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "torch.Size([20, 96])"
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "seq_len = config_dict[\"dataset\"][\"seq_len\"]\n",
    "test_tokens = test_X[0].to(torch.long)\n",
    "\n",
    "num_predict_tokens = config_dict[\"test\"][\"predict_tokens\"]\n",
    "seq_len = config_dict[\"dataset\"][\"seq_len\"]\n",
    "\n",
    "test_tokens_pred = torch.zeros_like(test_tokens)\n",
    "test_tokens_pred[:, :seq_len] = test_tokens[:, :seq_len]\n",
    "\n",
    "for i in tqdm(range(num_predict_tokens)):\n",
    "    inputs = test_tokens_pred[:, i:i+seq_len].to(torch.long)\n",
    "    outputs = gpt(inputs)\n",
    "    new_token = torch.argmax(outputs[:, -1, :], dim=-1)\n",
    "    test_tokens_pred[:, seq_len+i] = new_token.squeeze()\n",
    "test_tokens_pred.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "300"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "### Extract Data\n",
    "\n",
    "main_folder = config_dict[\"paths\"][\"input_folder\"]\n",
    "random_lines = config_dict[\"dataset\"][\"random_lines\"]\n",
    "num_setns_per_doc = config_dict[\"dataset\"][\"num_sents_per_doc\"]\n",
    "\n",
    "text_ls = []\n",
    "\n",
    "for path in glob.glob(f\"{main_folder}\\*.txt\"):\n",
    "    with open(path, \"r\") as f:\n",
    "        lines = np.array(f.readlines())\n",
    "    if random_lines:\n",
    "        ids = np.random.choice(len(lines), num_setns_per_doc, replace=False)\n",
    "    else:\n",
    "        ids = np.arange(num_setns_per_doc)\n",
    "    \n",
    "    lines = lines[ids]\n",
    "    text_ls.extend(lines)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Verify Config"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import yaml\n",
    "import logging\n",
    "\n",
    "from src.utils import ValidateConfig"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "BOW\n",
      "NGRAM\n",
      "TFIDF\n",
      "HMM\n",
      "WORD2VEC\n",
      "GLOVE\n",
      "RNN\n",
      "LSTM\n",
      "SEQ2SEQ\n",
      "GRU\n",
      "TRANSFORMER\n",
      "BERT\n",
      "GPT\n"
     ]
    }
   ],
   "source": [
    "algos = [\n",
    "    \"BOW\", \"NGRAM\", \"TFIDF\", \"HMM\", \"WORD2VEC\", \"GLOVE\", \"RNN\", \"LSTM\", \"SEQ2SEQ\", \"GRU\", \"TRANSFORMER\", \"BERT\", \"GPT\"\n",
    "]\n",
    "\n",
    "for algo in algos:\n",
    "    print(algo)\n",
    "    config_path = f\"./configs/{algo.lower()}.yaml\"\n",
    "    with open(config_path, \"r\") as stream:\n",
    "        config_dict = yaml.safe_load(stream)\n",
    "\n",
    "    cfgval = ValidateConfig(config_dict, algo)\n",
    "    cfgval.run_verify()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "nlp_env",
   "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": "3.12.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}