Language Detection/Language Detection- 10k Samples/language_detection-10k.ipynb

{
 "cells": [
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Language Detection (of 10,000 Samples)\n",
    "\n",
    "Dataset Source: https://www.kaggle.com/datasets/basilb2s/language-detection"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Install Missing & Necessary Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "#%pip install --upgrade pip\n",
    "#%pip install --upgrade datasets transformers evaluate "
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Import Necessary Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "NOTE: Redirects are currently not supported in Windows or MacOs.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Git LFS initialized.\n"
     ]
    }
   ],
   "source": [
    "import sys, os\n",
    "os.environ['TOKENIZERS_PARALLELISM']='false'\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "import datasets\n",
    "from datasets import Dataset, DatasetDict\n",
    "\n",
    "import torch\n",
    "\n",
    "import transformers\n",
    "from transformers import AutoModelForSequenceClassification, AutoTokenizer, set_seed\n",
    "from transformers import Trainer, TrainingArguments, DataCollatorWithPadding\n",
    "\n",
    "import evaluate\n",
    "\n",
    "!git lfs install"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Display Versions of Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "        Python: 3.9.12\n",
      "         NumPy: 1.23.3\n",
      "        Pandas: 1.5.0\n",
      "      Datasets: 2.9.0\n",
      "  Transformers: 4.26.1\n",
      "         Torch: 1.12.1\n",
      "      Evaluate: 0.4.0\n"
     ]
    }
   ],
   "source": [
    "print(\"Python:\".rjust(15), sys.version[0:6])\n",
    "print(\"NumPy:\".rjust(15), np.__version__)\n",
    "print(\"Pandas:\".rjust(15), pd.__version__)\n",
    "print(\"Datasets:\".rjust(15), datasets.__version__)\n",
    "print(\"Transformers:\".rjust(15), transformers.__version__)\n",
    "print(\"Torch:\".rjust(15), torch.__version__)\n",
    "print(\"Evaluate:\".rjust(15), evaluate.__version__)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Load Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Int64Index: 10274 entries, 0 to 10336\n",
      "Data columns (total 2 columns):\n",
      " #   Column  Non-Null Count  Dtype \n",
      "---  ------  --------------  ----- \n",
      " 0   text    10274 non-null  object\n",
      " 1   labels  10274 non-null  object\n",
      "dtypes: object(2)\n",
      "memory usage: 498.8+ KB\n"
     ]
    }
   ],
   "source": [
    "parent_dir = \"/Users/briandunn/Desktop/Language Detection/10k-17 Languages\"\n",
    "\n",
    "df = pd.read_csv(os.path.join(parent_dir, \"Language Detection.csv\"))\n",
    "\n",
    "df = df[df[\"Language\"] != \"Hindi\"]\n",
    "df = df.dropna()\n",
    "\n",
    "df[\"Text\"] = df[\"Text\"].str.strip()\n",
    "df[\"Language\"] = df[\"Language\"].str.strip()\n",
    "\n",
    "# make sure that Swedish and Portuguese are spelt correctly for the model\n",
    "# Change 'Sweedish' to 'Swedish' & 'Portugeese' to 'Portuguese'\n",
    "df['Language'] = df['Language'].rename(\n",
    "        {\n",
    "            \"Sweedish\": \"Swedish\", \n",
    "            \"Portugeese\": \"Portuguese\"\n",
    "        }\n",
    "    )\n",
    "\n",
    "# Change Feature names\n",
    "df.rename(\n",
    "    columns={\n",
    "        \"Text\": \"text\", \n",
    "        \"Language\": \"labels\",\n",
    "        },\n",
    "    inplace=True\n",
    ")\n",
    "\n",
    "df.info()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Visualize: Length Text Length (in Words)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The maximum length of this dataset is approximately 309 words.\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABasAAAK9CAYAAAAqrzD0AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjYuMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy89olMNAAAACXBIWXMAAA9hAAAPYQGoP6dpAABLK0lEQVR4nO3deZyWZb0/8M/AsCgwg5DMyFHAHXEtSJmjuaJEk1nSYpmh2aa4Uib+XMAlIe0cPXZcMg2sXMpOWkku5IKdxA0PZVqkJuFJgTZAMAGZ5/dHL57TCCqjoxeD7/frdb+c+7qv57q/9zNXz8v5eHU9NZVKpRIAAAAAACioU+kCAAAAAABAWA0AAAAAQHHCagAAAAAAihNWAwAAAABQnLAaAAAAAIDihNUAAAAAABQnrAYAAAAAoDhhNQAAAAAAxQmrAQAAAAAoTlgNAMBrGjRoUI488sjSZWzwLrzwwmy11Vbp3Llzdtttt9LlvO3U1NTkuOOOK10GAMDblrAaAOBtZurUqampqcnDDz+81uv77rtvdtpppzd8n5/+9KeZOHHiGx7n7eKOO+7Il7/85ey5556ZMmVKzj///DX63HPPPampqVmno73cd999mThxYhYtWrRO/Y888sj07Nmz3e7f3tr6PAAAvHVqSxcAAMD6b86cOenUqW3rHH7605/m0ksvFVivo7vuuiudOnXK1Vdfna5du661zw477JDvfOc7rdpOO+209OzZM6effvqbUtd9992Xs88+O0ceeWR69+79ptzjrbShPQ8AwIZEWA0AwGvq1q1b6RLabNmyZenRo0fpMtbZwoULs9FGG71iUJ0kDQ0N+eQnP9mqbfLkyXnHO96xRjsAAHQ0tgEBAOA1vXzP6pUrV+bss8/Otttum+7du6dv377Za6+9Mn369CT/2Ari0ksvTZK1bk2xbNmyfPGLX8wWW2yRbt26Zfvtt8/Xvva1VCqVVvf9+9//nhNOOCHveMc70qtXr3zgAx/IH//4x9TU1LRasT1x4sTU1NTk8ccfzyc+8Ylssskm2WuvvZIkv/rVr3LkkUdmq622Svfu3dPY2JhPf/rT+ctf/tLqXqvH+N3vfpdPfvKTqa+vz6abbpozzzwzlUolzzzzTA455JDU1dWlsbEx//Zv/7ZO791LL72Uc889N1tvvXW6deuWQYMG5f/9v/+X5cuXV/vU1NRkypQpWbZsWfW9mjp16jqNvzaLFi3KSSedVH1/t9lmm3z1q19NS0tLkqRSqWS//fbLpptumoULF1Zft2LFiuy8887Zeuuts2zZskycODGnnHJKkmTLLbes1jZ37tzXXdtqDzzwQN773vemvr4+G2+8cfbZZ5/84he/aNVn9e/kySefrK6Erq+vz1FHHZUXXnihVd91mSvr+jw333xzdtppp3Tr1i077rhjbrvttjf8vAAAvDYrqwEA3qYWL16cP//5z2u0r1y58jVfO3HixEyaNCmf+cxnsvvuu2fJkiV5+OGH88gjj+TAAw/M5z//+Tz77LOZPn36GttWVCqVfOADH8jdd9+do48+Orvttltuv/32nHLKKfnjH/+Yiy66qNr3yCOPzPe///0cccQRGT58eGbMmJHm5uZXrOsjH/lItt1225x//vnV4Hv69On5/e9/n6OOOiqNjY157LHHcuWVV+axxx7L/fffv8b+zh/72Meyww47ZPLkyZk2bVrOO++89OnTJ9/4xjey//7756tf/WquvfbafOlLX8q73/3u7L333q/6Xn3mM5/JNddckw9/+MP54he/mAceeCCTJk3Kb37zm9x0001Jku985zu58sor8+CDD+aqq65Kkvzrv/7ra/4e1uaFF17IPvvskz/+8Y/5/Oc/nwEDBuS+++7Laaedlueeey4XX3xxampq8q1vfSu77LJLvvCFL+SHP/xhkmTChAl57LHHcs8996RHjx459NBD87vf/S7XX399LrroorzjHe9Ikmy66aavq7bV7rrrrowaNSpDhw7NhAkT0qlTp0yZMiX7779/fv7zn2f33Xdv1f+jH/1ottxyy0yaNCmPPPJIrrrqqvTr1y9f/epXq33WZa6sy/P893//d374wx/m2GOPTa9evXLJJZdk9OjRmTdvXvr27fuGnhsAgNdQAQDgbWXKlCmVJK967Ljjjq1eM3DgwMqYMWOq57vuumulubn5Ve8zduzYytr+dfPmm2+uJKmcd955rdo//OEPV2pqaipPPvlkpVKpVGbNmlVJUjnppJNa9TvyyCMrSSoTJkyotk2YMKGSpPLxj398jfu98MILa7Rdf/31lSSVe++9d40xPve5z1XbXnrppcrmm29eqampqUyePLna/re//a2y0UYbtXpP1mb27NmVJJXPfOYzrdq/9KUvVZJU7rrrrmrbmDFjKj169HjV8dZmxx13rOyzzz7V83PPPbfSo0ePyu9+97tW/caPH1/p3LlzZd68edW2b3zjG5Ukle9+97uV+++/v9K5c+c13u8LL7ywkqTy9NNPr1M9r/UcLS0tlW233bYycuTISktLS7X9hRdeqGy55ZaVAw88sNq2+nfy6U9/utUYH/rQhyp9+/atnrdlrrza8ySpdO3atToHK5VK5Ze//GUlSeXrX//6az47AABvjG1AAADepi699NJMnz59jWOXXXZ5zdf27t07jz32WJ544ok23/enP/1pOnfunBNOOKFV+xe/+MVUKpXceuutSVLdeuHYY49t1e/4449/xbG/8IUvrNG20UYbVX9+8cUX8+c//znDhw9PkjzyyCNr9P/MZz5T/blz584ZNmxYKpVKjj766Gp77969s/322+f3v//9K9aS/ONZk2TcuHGt2r/4xS8mSaZNm/aqr389brzxxrznPe/JJptskj//+c/VY8SIEVm1alXuvffeat/Pfe5zGTlyZI4//vgcccQR2XrrrXP++ee3e03/bPbs2XniiSfyiU98In/5y1+q9S1btiwHHHBA7r333up2Jau9/Pf6nve8J3/5y1+yZMmSJK9vrrySESNGZOutt66e77LLLqmrq3vN3zUAAG+cbUAAAN6mdt999wwbNmyN9tUh56s555xzcsghh2S77bbLTjvtlPe+97054ogj1ino/sMf/pD+/funV69erdp32GGH6vXV/+zUqVO23HLLVv222WabVxz75X2T5K9//WvOPvvs3HDDDa32Z07+sRXKyw0YMKDVeX19fbp3717dMuKf21++7/XLrX6Gl9fc2NiY3r17V5+1PT3xxBP51a9+9Ypbdbz8Pbj66quz9dZb54knnsh9993XKtx/M6z+Dxxjxox5xT6LFy/OJptsUj1/+e9k9bW//e1vqaure11z5ZW8/F6r7/e3v/2tzWMBANA2wmoAANps7733zlNPPZUf/ehHueOOO3LVVVfloosuyhVXXNFqZfJbbW1B60c/+tHcd999OeWUU7LbbrulZ8+eaWlpyXvf+941VvAm/1hNvS5tSdb4QshX8vJ9sd9MLS0tOfDAA/PlL395rde32267Vuf33HNP9cseH3300TQ1Nb3p9SXJhRdemN12222tfXr27Nnq/I2+/23xVt4LAIDWhNUAALwuffr0yVFHHZWjjjoqS5cuzd57752JEydWw+pXCmgHDhyYn/3sZ3n++edbra7+7W9/W72++p8tLS15+umns+2221b7Pfnkk+tc49/+9rfceeedOfvss3PWWWdV21/P9iWvx+pneOKJJ6orx5NkwYIFWbRoUfVZ29PWW2+dpUuXZsSIEa/Z97nnnsvxxx+fgw46KF27ds2XvvSljBw5slVd7R20r95io66ubp1qXBdtmStv5X84AACgbexZDQBAm718+4uePXtmm222qa7QTZIePXokSRYtWtSq7/ve976sWrUq//mf/9mq/aKLLkpNTU1GjRqVJBk5cmSS5LLLLmvV7+tf//o617l6lezLV8VefPHF6zzGG/G+971vrff793//9yRJc3Nzu9/zox/9aGbOnJnbb799jWuLFi3KSy+9VD3/7Gc/m5aWllx99dW58sorU1tbm6OPPrrV+/VKv8fXa+jQodl6663zta99LUuXLl3j+p/+9Kc2j9mWudLezwMAQPuxshoAgDYbMmRI9t133wwdOjR9+vTJww8/nB/84Ac57rjjqn2GDh2aJDnhhBMycuTIdO7cOYcddlgOPvjg7Lfffjn99NMzd+7c7Lrrrrnjjjvyox/9KCeddFJ15e3QoUMzevToXHzxxfnLX/6S4cOHZ8aMGfnd736XZN1WyNbV1WXvvffOBRdckJUrV+Zf/uVfcscdd+Tpp59+E96VNe26664ZM2ZMrrzyyixatCj77LNPHnzwwVxzzTX54Ac/mP3226/d73nKKafkxz/+cd7//vfnyCOPzNChQ7Ns2bI8+uij+cEPfpC5c+fmHe94R6ZMmZJp06Zl6tSp2XzzzZP8I9z95Cc/mcsvv7z6ZYWrf4+nn356DjvssHTp0iUHH3xwNfRdm5UrV+a8885bo71Pnz459thjc9VVV2XUqFHZcccdc9RRR+Vf/uVf8sc//jF333136urq8pOf/KRNz9yWufJ6ngcAgLeGsBoAgDY74YQT8uMf/zh33HFHli9fnoEDB+a8887LKaecUu1z6KGH5vjjj88NN9yQ7373u6lUKjnssMPSqVOn/PjHP85ZZ52V733ve5kyZUoGDRqUCy+8MF/84hdb3efb3/52Ghsbc/311+emm27KiBEj8r3vfS/bb799unfvvk61XnfddTn++ONz6aWXplKp5KCDDsqtt96a/v37t+t78kquuuqqbLXVVpk6dWpuuummNDY25rTTTsuECRPelPttvPHGmTFjRs4///zceOON+fa3v526urpst912Ofvss1NfX5///d//zcknn5yDDz641RcdHn744fmv//qvfPnLX86oUaOy5ZZb5t3vfnfOPffcXHHFFbntttuq2228Wri7YsWKnHnmmWu0b7311jn22GOz7777ZubMmTn33HPzn//5n1m6dGkaGxuzxx575POf//zreu51nSuv53kAAHhr1FR8UwgAAB3I7Nmz8853vjPf/e53c/jhh5cuh/WYuQIA0LHYsxoAgPXW3//+9zXaLr744nTq1Cl77713gYpYX5krAAAdn21AAABYb11wwQWZNWtW9ttvv9TW1ubWW2/Nrbfems997nPZYostSpfHesRcAQDo+GwDAgDAemv69Ok5++yz8/jjj2fp0qUZMGBAjjjiiJx++umprbXugv9jrgAAdHzCagAAAAAAirNnNQAAAAAAxQmrAQAAAAAorkNu3tbS0pJnn302vXr1Sk1NTelyAAAAAABYi0qlkueffz79+/dPp06vvna6Q4bVzz77rG/0BgAAAADoIJ555plsvvnmr9qnQ4bVvXr1SvKPB6yrqytcDQAAAAAAa7NkyZJsscUW1Uz31XTIsHr11h91dXXCagAAAACA9dy6bOfsCxYBAAAAAChOWA0AAAAAQHHCagAAAAAAihNWAwAAAABQnLAaAAAAAIDihNUAAAAAABQnrAYAAAAAoDhhNQAAAAAAxQmrAQAAAAAoTlgNAAAAAEBxwmoAAAAAAIoTVgMAAAAAUJywGgAAAACA4oTVAAAAAAAUJ6wGAAAAAKA4YTUAAAAAAMUJqwEAAAAAKE5YDQAAAABAccJqAAAAAACKE1YDAAAAAFCcsBoAAAAAgOKE1QAAAAAAFCesBgAAAACguDaF1RMnTkxNTU2rY/DgwdXrL774YsaOHZu+ffumZ8+eGT16dBYsWNBqjHnz5qW5uTkbb7xx+vXrl1NOOSUvvfRS+zwNAAAAAAAdUm1bX7DjjjvmZz/72f8NUPt/Q5x88smZNm1abrzxxtTX1+e4447LoYceml/84hdJklWrVqW5uTmNjY2577778txzz+VTn/pUunTpkvPPP78dHgcAAAAAgI6ozWF1bW1tGhsb12hfvHhxrr766lx33XXZf//9kyRTpkzJDjvskPvvvz/Dhw/PHXfckccffzw/+9nP0tDQkN122y3nnntuTj311EycODFdu3Z9408EAAAAAECH0+aw+oknnkj//v3TvXv3NDU1ZdKkSRkwYEBmzZqVlStXZsSIEdW+gwcPzoABAzJz5swMHz48M2fOzM4775yGhoZqn5EjR+aYY47JY489lne+851rvefy5cuzfPny6vmSJUvaWjZvskHjp7XLOHMnN7fLOAAAAABAx9KmPav32GOPTJ06Nbfddlsuv/zyPP3003nPe96T559/PvPnz0/Xrl3Tu3fvVq9paGjI/PnzkyTz589vFVSvvr762iuZNGlS6uvrq8cWW2zRlrIBAAAAAFjPtWll9ahRo6o/77LLLtljjz0ycODAfP/7389GG23U7sWtdtppp2XcuHHV8yVLlgisAQAAAAA2IG1aWf1yvXv3znbbbZcnn3wyjY2NWbFiRRYtWtSqz4IFC6p7XDc2NmbBggVrXF997ZV069YtdXV1rQ4AAAAAADYcbyisXrp0aZ566qlsttlmGTp0aLp06ZI777yzen3OnDmZN29empqakiRNTU159NFHs3Dhwmqf6dOnp66uLkOGDHkjpQAAAAAA0IG1aRuQL33pSzn44IMzcODAPPvss5kwYUI6d+6cj3/846mvr8/RRx+dcePGpU+fPqmrq8vxxx+fpqamDB8+PEly0EEHZciQITniiCNywQUXZP78+TnjjDMyduzYdOvW7U15QAAAAAAA1n9tCqv/93//Nx//+Mfzl7/8JZtuumn22muv3H///dl0002TJBdddFE6deqU0aNHZ/ny5Rk5cmQuu+yy6us7d+6cW265Jcccc0yamprSo0ePjBkzJuecc077PhUAAAAAAB1KTaVSqZQuoq2WLFmS+vr6LF682P7V64lB46e1yzhzJze3yzgAAAAAQHltyXLf0J7VAAAAAADQHoTVAAAAAAAUJ6wGAAAAAKA4YTUAAAAAAMUJqwEAAAAAKE5YDQAAAABAccJqAAAAAACKqy1dALwZBo2f1i7jzJ3c3C7jAAAAAACvzspqAAAAAACKE1YDAAAAAFCcsBoAAAAAgOKE1QAAAAAAFCesBgAAAACgOGE1AAAAAADFCasBAAAAAChOWA0AAAAAQHHCagAAAAAAihNWAwAAAABQnLAaAAAAAIDihNUAAAAAABQnrAYAAAAAoDhhNQAAAAAAxQmrAQAAAAAoTlgNAAAAAEBxwmoAAAAAAIqrLV0A/LNB46eVLgEAAAAAKMDKagAAAAAAihNWAwAAAABQnLAaAAAAAIDihNUAAAAAABQnrAYAAAAAoDhhNQAAAAAAxQmrAQAAAAAoTlgNAAAAAEBxwmoAAAAAAIoTVgMAAAAAUJywGgAAAACA4oTVAAAAAAAUJ6wGAAAAAKA4YTUAAAAAAMUJqwEAAAAAKE5YDQAAAABAccJqAAAAAACKE1YDAAAAAFCcsBoAAAAAgOJqSxcA67NB46e1yzhzJze3yzgAAAAAsKGyshoAAAAAgOKE1QAAAAAAFCesBgAAAACgOGE1AAAAAADFCasBAAAAAChOWA0AAAAAQHHCagAAAAAAihNWAwAAAABQnLAaAAAAAIDihNUAAAAAABQnrAYAAAAAoDhhNQAAAAAAxQmrAQAAAAAoTlgNAAAAAEBxwmoAAAAAAIoTVgMAAAAAUJywGgAAAACA4oTVAAAAAAAUJ6wGAAAAAKA4YTUAAAAAAMUJqwEAAAAAKE5YDQAAAABAccJqAAAAAACKE1YDAAAAAFCcsBoAAAAAgOKE1QAAAAAAFCesBgAAAACgOGE1AAAAAADFCasBAAAAAChOWA0AAAAAQHHCagAAAAAAihNWAwAAAABQnLAaAAAAAIDihNUAAAAAABQnrAYAAAAAoDhhNQAAAAAAxQmrAQAAAAAoTlgNAAAAAEBxwmoAAAAAAIoTVgMAAAAAUJywGgAAAACA4oTVAAAAAAAUJ6wGAAAAAKA4YTUAAAAAAMUJqwEAAAAAKE5YDQAAAABAccJqAAAAAACKE1YDAAAAAFCcsBoAAAAAgOKE1QAAAAAAFCesBgAAAACgOGE1AAAAAADFCasBAAAAAChOWA0AAAAAQHHCagAAAAAAihNWAwAAAABQnLAaAAAAAIDihNUAAAAAABQnrAYAAAAAoDhhNQAAAAAAxb2hsHry5MmpqanJSSedVG178cUXM3bs2PTt2zc9e/bM6NGjs2DBglavmzdvXpqbm7PxxhunX79+OeWUU/LSSy+9kVIAAAAAAOjAXndY/dBDD+Ub3/hGdtlll1btJ598cn7yk5/kxhtvzIwZM/Lss8/m0EMPrV5ftWpVmpubs2LFitx333255pprMnXq1Jx11lmv/ykAAAAAAOjQXldYvXTp0hx++OH55je/mU022aTavnjx4lx99dX593//9+y///4ZOnRopkyZkvvuuy/3339/kuSOO+7I448/nu9+97vZbbfdMmrUqJx77rm59NJLs2LFivZ5KgAAAAAAOpTXFVaPHTs2zc3NGTFiRKv2WbNmZeXKla3aBw8enAEDBmTmzJlJkpkzZ2bnnXdOQ0NDtc/IkSOzZMmSPPbYY2u93/Lly7NkyZJWBwAAAAAAG47atr7ghhtuyCOPPJKHHnpojWvz589P165d07t371btDQ0NmT9/frXPPwfVq6+vvrY2kyZNytlnn93WUgEAAAAA6CDatLL6mWeeyYknnphrr7023bt3f7NqWsNpp52WxYsXV49nnnnmLbs3AAAAAABvvjaF1bNmzcrChQvzrne9K7W1tamtrc2MGTNyySWXpLa2Ng0NDVmxYkUWLVrU6nULFixIY2NjkqSxsTELFixY4/rqa2vTrVu31NXVtToAAAAAANhwtCmsPuCAA/Loo49m9uzZ1WPYsGE5/PDDqz936dIld955Z/U1c+bMybx589LU1JQkaWpqyqOPPpqFCxdW+0yfPj11dXUZMmRIOz0WAAAAAAAdSZv2rO7Vq1d22mmnVm09evRI3759q+1HH310xo0blz59+qSuri7HH398mpqaMnz48CTJQQcdlCFDhuSII47IBRdckPnz5+eMM87I2LFj061bt3Z6LAAAAAAAOpI2f8Hia7nooovSqVOnjB49OsuXL8/IkSNz2WWXVa937tw5t9xyS4455pg0NTWlR48eGTNmTM4555z2LgUAAAAAgA6iplKpVEoX0VZLlixJfX19Fi9ebP/q9cSg8dNKl7Bemzu5uXQJAAAAAPCWa0uW26Y9qwEAAAAA4M0grAYAAAAAoDhhNQAAAAAAxQmrAQAAAAAoTlgNAAAAAEBxwmoAAAAAAIoTVgMAAAAAUJywGgAAAACA4oTVAAAAAAAUJ6wGAAAAAKA4YTUAAAAAAMUJqwEAAAAAKE5YDQAAAABAccJqAAAAAACKE1YDAAAAAFCcsBoAAAAAgOKE1QAAAAAAFCesBgAAAACgOGE1AAAAAADFCasBAAAAAChOWA0AAAAAQHHCagAAAAAAihNWAwAAAABQnLAaAAAAAIDihNUAAAAAABQnrAYAAAAAoDhhNQAAAAAAxQmrAQAAAAAoTlgNAAAAAEBxwmoAAAAAAIoTVgMAAAAAUJywGgAAAACA4oTVAAAAAAAUJ6wGAAAAAKA4YTUAAAAAAMUJqwEAAAAAKE5YDQAAAABAccJqAAAAAACKE1YDAAAAAFCcsBoAAAAAgOKE1QAAAAAAFCesBgAAAACgOGE1AAAAAADFCasBAAAAAChOWA0AAAAAQHHCagAAAAAAihNWAwAAAABQnLAaAAAAAIDihNUAAAAAABQnrAYAAAAAoDhhNQAAAAAAxdWWLgBYd4PGT2uXceZObm6XcQAAAACgvVhZDQAAAABAccJqAAAAAACKE1YDAAAAAFCcsBoAAAAAgOKE1QAAAAAAFCesBgAAAACgOGE1AAAAAADFCasBAAAAAChOWA0AAAAAQHHCagAAAAAAiqstXQC8HQwaP610CQAAAACwXrOyGgAAAACA4oTVAAAAAAAUJ6wGAAAAAKA4YTUAAAAAAMUJqwEAAAAAKE5YDQAAAABAccJqAAAAAACKE1YDAAAAAFCcsBoAAAAAgOJqSxcAdGyDxk9rl3HmTm5ul3EAAAAA6JisrAYAAAAAoDhhNQAAAAAAxQmrAQAAAAAoTlgNAAAAAEBxwmoAAAAAAIoTVgMAAAAAUJywGgAAAACA4oTVAAAAAAAUJ6wGAAAAAKC42tIFAPDaBo2f1i7jzJ3c3C7jAAAAALQ3K6sBAAAAAChOWA0AAAAAQHHCagAAAAAAihNWAwAAAABQnLAaAAAAAIDihNUAAAAAABQnrAYAAAAAoDhhNQAAAAAAxQmrAQAAAAAoTlgNAAAAAEBxwmoAAAAAAIqrLV0AAB3PoPHT2mWcuZOb22UcAAAAoOOzshoAAAAAgOKE1QAAAAAAFCesBgAAAACgOGE1AAAAAADFCasBAAAAAChOWA0AAAAAQHHCagAAAAAAiqstXQAAb51B46eVLgEAAABgraysBgAAAACgOGE1AAAAAADFCasBAAAAACjOntUAFNOee2jPndzcbmMBAAAAb702ray+/PLLs8suu6Suri51dXVpamrKrbfeWr3+4osvZuzYsenbt2969uyZ0aNHZ8GCBa3GmDdvXpqbm7PxxhunX79+OeWUU/LSSy+1z9MAAAAAANAhtSms3nzzzTN58uTMmjUrDz/8cPbff/8ccsgheeyxx5IkJ598cn7yk5/kxhtvzIwZM/Lss8/m0EMPrb5+1apVaW5uzooVK3LfffflmmuuydSpU3PWWWe171MBAAAAANChtGkbkIMPPrjV+Ve+8pVcfvnluf/++7P55pvn6quvznXXXZf9998/STJlypTssMMOuf/++zN8+PDccccdefzxx/Ozn/0sDQ0N2W233XLuuefm1FNPzcSJE9O1a9e13nf58uVZvnx59XzJkiVtfU4AAAAAANZjr/sLFletWpUbbrghy5YtS1NTU2bNmpWVK1dmxIgR1T6DBw/OgAEDMnPmzCTJzJkzs/POO6ehoaHaZ+TIkVmyZEl1dfbaTJo0KfX19dVjiy22eL1lAwAAAACwHmpzWP3oo4+mZ8+e6datW77whS/kpptuypAhQzJ//vx07do1vXv3btW/oaEh8+fPT5LMnz+/VVC9+vrqa6/ktNNOy+LFi6vHM88809ayAQAAAABYj7VpG5Ak2X777TN79uwsXrw4P/jBDzJmzJjMmDHjzaitqlu3bunWrdubeg8AAAAAAMppc1jdtWvXbLPNNkmSoUOH5qGHHsp//Md/5GMf+1hWrFiRRYsWtVpdvWDBgjQ2NiZJGhsb8+CDD7Yab8GCBdVrwFtj0PhppUtYQ3vVNHdyc7uMAwAAAMBb63XvWb1aS0tLli9fnqFDh6ZLly658847q9fmzJmTefPmpampKUnS1NSURx99NAsXLqz2mT59eurq6jJkyJA3WgoAAAAAAB1Um1ZWn3baaRk1alQGDBiQ559/Ptddd13uueee3H777amvr8/RRx+dcePGpU+fPqmrq8vxxx+fpqamDB8+PEly0EEHZciQITniiCNywQUXZP78+TnjjDMyduxY23wAAAAAALyNtSmsXrhwYT71qU/lueeeS319fXbZZZfcfvvtOfDAA5MkF110UTp16pTRo0dn+fLlGTlyZC677LLq6zt37pxbbrklxxxzTJqamtKjR4+MGTMm55xzTvs+FQAAAAAAHUpNpVKplC6irZYsWZL6+vosXrw4dXV1pcsh6+ceyLw9bah7Vvvf2GvbUH/3AAAA0JG1Jct9w3tWAwAAAADAGyWsBgAAAACgOGE1AAAAAADFtekLFgHeLuwRDQAAAPDWsrIaAAAAAIDihNUAAAAAABQnrAYAAAAAoDhhNQAAAAAAxQmrAQAAAAAoTlgNAAAAAEBxwmoAAAAAAIqrLV0AAKxPBo2f1i7jzJ3c3C7jAAAAwNuFldUAAAAAABQnrAYAAAAAoDhhNQAAAAAAxdmzGtigtNd+wwAAAAC8taysBgAAAACgOCurAdggWFUPAAAAHZuV1QAAAAAAFCesBgAAAACgOGE1AAAAAADFCasBAAAAAChOWA0AAAAAQHHCagAAAAAAihNWAwAAAABQnLAaAAAAAIDihNUAAAAAABQnrAYAAAAAoDhhNQAAAAAAxQmrAQAAAAAoTlgNAAAAAEBxwmoAAAAAAIoTVgMAAAAAUJywGgAAAACA4oTVAAAAAAAUJ6wGAAAAAKA4YTUAAAAAAMUJqwEAAAAAKE5YDQAAAABAccJqAAAAAACKE1YDAAAAAFCcsBoAAAAAgOKE1QAAAAAAFCesBgAAAACgOGE1AAAAAADFCasBAAAAAChOWA0AAAAAQHHCagAAAAAAihNWAwAAAABQnLAaAAAAAIDihNUAAAAAABQnrAYAAAAAoLja0gUAwIZo0Php7TLO3MnN7TIOAAAArO+E1QDwNiA8BwAAYH1nGxAAAAAAAIoTVgMAAAAAUJywGgAAAACA4oTVAAAAAAAUJ6wGAAAAAKA4YTUAAAAAAMUJqwEAAAAAKE5YDQAAAABAccJqAAAAAACKE1YDAAAAAFCcsBoAAAAAgOKE1QAAAAAAFCesBgAAAACguNrSBQAAr2zQ+GmlSwAAAIC3hJXVAAAAAAAUJ6wGAAAAAKA4YTUAAAAAAMUJqwEAAAAAKE5YDQAAAABAccJqAAAAAACKqy1dAADQcQwaP61dxpk7ubldxgEAAGDDYWU1AAAAAADFCasBAAAAAChOWA0AAAAAQHHCagAAAAAAihNWAwAAAABQnLAaAAAAAIDihNUAAAAAABQnrAYAAAAAoDhhNQAAAAAAxdWWLgAAePsZNH5au4wzd3Jzu4wDAABAeVZWAwAAAABQnLAaAAAAAIDihNUAAAAAABQnrAYAAAAAoDhhNQAAAAAAxQmrAQAAAAAoTlgNAAAAAEBxwmoAAAAAAIoTVgMAAAAAUJywGgAAAACA4oTVAAAAAAAUJ6wGAAAAAKA4YTUAAAAAAMUJqwEAAAAAKE5YDQAAAABAccJqAAAAAACKE1YDAAAAAFCcsBoAAAAAgOKE1QAAAAAAFCesBgAAAACgOGE1AAAAAADFtSmsnjRpUt797nenV69e6devXz74wQ9mzpw5rfq8+OKLGTt2bPr27ZuePXtm9OjRWbBgQas+8+bNS3NzczbeeOP069cvp5xySl566aU3/jQAAAAAAHRIbQqrZ8yYkbFjx+b+++/P9OnTs3Llyhx00EFZtmxZtc/JJ5+cn/zkJ7nxxhszY8aMPPvsszn00EOr11etWpXm5uasWLEi9913X6655ppMnTo1Z511Vvs9FQAAAAAAHUpNpVKpvN4X/+lPf0q/fv0yY8aM7L333lm8eHE23XTTXHfddfnwhz+cJPntb3+bHXbYITNnzszw4cNz66235v3vf3+effbZNDQ0JEmuuOKKnHrqqfnTn/6Url27vuZ9lyxZkvr6+ixevDh1dXWvt3za0aDx00qXAMDb0NzJzaVLAAAA4FW0Jct9Q3tWL168OEnSp0+fJMmsWbOycuXKjBgxotpn8ODBGTBgQGbOnJkkmTlzZnbeeedqUJ0kI0eOzJIlS/LYY4+t9T7Lly/PkiVLWh0AAAAAAGw4XndY3dLSkpNOOil77rlndtpppyTJ/Pnz07Vr1/Tu3btV34aGhsyfP7/a55+D6tXXV19bm0mTJqW+vr56bLHFFq+3bAAAAAAA1kOvO6weO3Zsfv3rX+eGG25oz3rW6rTTTsvixYurxzPPPPOm3xMAAAAAgLdO7et50XHHHZdbbrkl9957bzbffPNqe2NjY1asWJFFixa1Wl29YMGCNDY2Vvs8+OCDrcZbsGBB9dradOvWLd26dXs9pQIAAAAA0AG0aWV1pVLJcccdl5tuuil33XVXttxyy1bXhw4dmi5duuTOO++sts2ZMyfz5s1LU1NTkqSpqSmPPvpoFi5cWO0zffr01NXVZciQIW/kWQAAAAAA6KDatLJ67Nixue666/KjH/0ovXr1qu4xXV9fn4022ij19fU5+uijM27cuPTp0yd1dXU5/vjj09TUlOHDhydJDjrooAwZMiRHHHFELrjggsyfPz9nnHFGxo4da/U0AAAAAMDbVJvC6ssvvzxJsu+++7ZqnzJlSo488sgkyUUXXZROnTpl9OjRWb58eUaOHJnLLrus2rdz58655ZZbcswxx6SpqSk9evTImDFjcs4557yxJwEAAAAAoMOqqVQqldJFtNWSJUtSX1+fxYsXp66urnQ5JBk0flrpEgB4G5o7ubl0CQAAALyKtmS5bdqzGgAAAAAA3gzCagAAAAAAihNWAwAAAABQnLAaAAAAAIDihNUAAAAAABQnrAYAAAAAoDhhNQAAAAAAxQmrAQAAAAAoTlgNAAAAAEBxwmoAAAAAAIoTVgMAAAAAUFxt6QIAAF6vQeOntdtYcyc3t9tYAAAAtJ2V1QAAAAAAFCesBgAAAACgONuAvM215/99GgAAAADg9bKyGgAAAACA4oTVAAAAAAAUJ6wGAAAAAKA4YTUAAAAAAMUJqwEAAAAAKE5YDQAAAABAccJqAAAAAACKE1YDAAAAAFCcsBoAAAAAgOKE1QAAAAAAFCesBgAAAACgOGE1AAAAAADFCasBAAAAAChOWA0AAAAAQHHCagAAAAAAiqstXQAAwPpg0Php7TLO3MnN7TIOAADA242V1QAAAAAAFCesBgAAAACgOGE1AAAAAADFCasBAAAAAChOWA0AAAAAQHHCagAAAAAAihNWAwAAAABQnLAaAAAAAIDihNUAAAAAABQnrAYAAAAAoDhhNQAAAAAAxQmrAQAAAAAoTlgNAAAAAEBxtaULAABgTYPGT2uXceZObm6XcQAAAN5sVlYDAAAAAFCcsBoAAAAAgOKE1QAAAAAAFCesBgAAAACgOF+wCADQjtrrixEBAADebqysBgAAAACgOGE1AAAAAADFCasBAAAAAChOWA0AAAAAQHHCagAAAAAAihNWAwAAAABQnLAaAAAAAIDihNUAAAAAABQnrAYAAAAAoDhhNQAAAAAAxdWWLgAAgDfPoPHT2mWcuZOb22UcAACAV2JlNQAAAAAAxQmrAQAAAAAoTlgNAAAAAEBxwmoAAAAAAIoTVgMAAAAAUJywGgAAAACA4oTVAAAAAAAUJ6wGAAAAAKA4YTUAAAAAAMUJqwEAAAAAKE5YDQAAAABAccJqAAAAAACKqy1dAAAA679B46e1yzhzJze3yzgAAMCGx8pqAAAAAACKE1YDAAAAAFCcbUAAAOhwbEsCAAAbHiurAQAAAAAoTlgNAAAAAEBxwmoAAAAAAIoTVgMAAAAAUJywGgAAAACA4oTVAAAAAAAUJ6wGAAAAAKA4YTUAAAAAAMUJqwEAAAAAKE5YDQAAAABAcbWlCwAAgI5u0Php7TbW3MnN7TYWAAB0JFZWAwAAAABQnLAaAAAAAIDihNUAAAAAABQnrAYAAAAAoDhhNQAAAAAAxdWWLgAAgLePQeOnlS5hvdde79Hcyc3tMg4AALxVrKwGAAAAAKA4YTUAAAAAAMUJqwEAAAAAKE5YDQAAAABAcb5gEQCAty1f+AgAAOsPK6sBAAAAAChOWA0AAAAAQHHCagAAAAAAihNWAwAAAABQnLAaAAAAAIDihNUAAAAAABQnrAYAAAAAoLg2h9X33ntvDj744PTv3z81NTW5+eabW12vVCo566yzstlmm2WjjTbKiBEj8sQTT7Tq89e//jWHH3546urq0rt37xx99NFZunTpG3oQAAAAAAA6rjaH1cuWLcuuu+6aSy+9dK3XL7jgglxyySW54oor8sADD6RHjx4ZOXJkXnzxxWqfww8/PI899limT5+eW265Jffee28+97nPvf6nAAAAAACgQ6tt6wtGjRqVUaNGrfVapVLJxRdfnDPOOCOHHHJIkuTb3/52GhoacvPNN+ewww7Lb37zm9x222156KGHMmzYsCTJ17/+9bzvfe/L1772tfTv3/8NPA4AAAAAAB1Ru+5Z/fTTT2f+/PkZMWJEta2+vj577LFHZs6cmSSZOXNmevfuXQ2qk2TEiBHp1KlTHnjggbWOu3z58ixZsqTVAQAAAADAhqNdw+r58+cnSRoaGlq1NzQ0VK/Nnz8//fr1a3W9trY2ffr0qfZ5uUmTJqW+vr56bLHFFu1ZNgAAAAAAhbVrWP1mOe2007J48eLq8cwzz5QuCQAAAACAdtSuYXVjY2OSZMGCBa3aFyxYUL3W2NiYhQsXtrr+0ksv5a9//Wu1z8t169YtdXV1rQ4AAAAAADYc7RpWb7nllmlsbMydd95ZbVuyZEkeeOCBNDU1JUmampqyaNGizJo1q9rnrrvuSktLS/bYY4/2LAcAAAAAgA6itq0vWLp0aZ588snq+dNPP53Zs2enT58+GTBgQE466aScd9552XbbbbPlllvmzDPPTP/+/fPBD34wSbLDDjvkve99bz772c/miiuuyMqVK3PcccflsMMOS//+/dvtwQAAAAAA6DjaHFY//PDD2W+//arn48aNS5KMGTMmU6dOzZe//OUsW7Ysn/vc57Jo0aLstddeue2229K9e/fqa6699tocd9xxOeCAA9KpU6eMHj06l1xySTs8DgAAAAAAHVFNpVKplC6irZYsWZL6+vosXrzY/tVv0KDx00qXAADAm2Du5ObSJQAAQJuy3HbdsxoAAAAAAF4PYTUAAAAAAMW1ec9qAABg/dde273ZTgQAgLeKsBoAAHhFQm8AAN4qtgEBAAAAAKA4YTUAAAAAAMUJqwEAAAAAKE5YDQAAAABAccJqAAAAAACKE1YDAAAAAFCcsBoAAAAAgOJqSxcAAABs+AaNn9Yu48yd3Nwu4wAAsP6xshoAAAAAgOKE1QAAAAAAFCesBgAAAACgOGE1AAAAAADFCasBAAAAAChOWA0AAAAAQHHCagAAAAAAihNWAwAAAABQnLAaAAAAAIDihNUAAAAAABQnrAYAAAAAoDhhNQAAAAAAxQmrAQAAAAAoTlgNAAAAAEBxwmoAAAAAAIqrLV0AAADAuho0flq7jDN3cnO7jAMAQPuxshoAAAAAgOKE1QAAAAAAFGcbEAAAgNfJtiQAAO3HymoAAAAAAIqzshoAAHjbaa8V0QAAtB8rqwEAAAAAKE5YDQAAAABAccJqAAAAAACKE1YDAAAAAFCcL1gEAADYgLTXl0fOndzcLuMAAKwrK6sBAAAAAChOWA0AAAAAQHHCagAAAAAAihNWAwAAAABQnLAaAAAAAIDiaksXAAAA8HY3aPy00iUAABRnZTUAAAAAAMUJqwEAAAAAKE5YDQAAAABAccJqAAAAAACKE1YDAAAAAFCcsBoAAAAAgOKE1QAAAAAAFCesBgAAAACgOGE1AAAAAADFCasBAAAAACiutnQBAAAA8FYZNH5au4wzd3Jzu4wDAPwfK6sBAAAAAChOWA0AAAAAQHHCagAAAAAAihNWAwAAAABQnLAaAAAAAIDihNUAAAAAABRXW7oAAAAANlyDxk9rl3HmTm5ul3EAgPWXldUAAAAAABQnrAYAAAAAoDjbgAAAALDea6/tRACA9ZewGgAAgDUIhwGAt5ptQAAAAAAAKE5YDQAAAABAccJqAAAAAACKs2c1AAAAdHDttcf43MnN7TIOALweVlYDAAAAAFCcsBoAAAAAgOKE1QAAAAAAFCesBgAAAACgOGE1AAAAAADFCasBAAAAAChOWA0AAAAAQHG1pQsAAACAjmbQ+GntMs7cyc3tMs6GzHsN8PZhZTUAAAAAAMUJqwEAAAAAKM42IAAAAFBIe21xsT7akJ8NgDeHldUAAAAAABQnrAYAAAAAoDhhNQAAAAAAxQmrAQAAAAAoTlgNAAAAAEBxtaULAAAAAHizDRo/rV3GmTu5uV3GAWBNwmoAAAAgSfsFurw24TnAmmwDAgAAAABAccJqAAAAAACKE1YDAAAAAFCcsBoAAAAAgOJ8wSIAAADAOvIllABvHmE1AAAAAO2iPcP8uZOb220soGMQVgMAAAC8zVkxDqwP7FkNAAAAAEBxVlYDAAAAdFBWRPNGtdccsm0L7cHKagAAAAAAihNWAwAAAABQnG1AAAAAANhg2eYCOg5hNQAAAAC8hvVtf3DhORsi24AAAAAAAFCcldUAAAAArHfWt5XMvLr1bbuV9pw/VrG/daysBgAAAACgOCurAQAAAID1wvq4on59WzW+IRNWAwAAAEAHsz6GuvBGFd0G5NJLL82gQYPSvXv37LHHHnnwwQdLlgMAAAAAQCHFwurvfe97GTduXCZMmJBHHnkku+66a0aOHJmFCxeWKgkAAAAAgEKKhdX//u//ns9+9rM56qijMmTIkFxxxRXZeOON861vfatUSQAAAAAAFFJkz+oVK1Zk1qxZOe2006ptnTp1yogRIzJz5sw1+i9fvjzLly+vni9evDhJsmTJkje/2A1cy/IXSpcAAAAAABu8t2uWufq5K5XKa/YtElb/+c9/zqpVq9LQ0NCqvaGhIb/97W/X6D9p0qScffbZa7RvscUWb1qNAAAAAADtpf7i0hWU9fzzz6e+vv5V+xQJq9vqtNNOy7hx46rnLS0t+etf/5q+ffumpqamYGVlLFmyJFtssUWeeeaZ1NXVlS4H3jBzmg2NOc2GxpxmQ2NOs6Exp9kQmddsaN7Oc7pSqeT5559P//79X7NvkbD6He94Rzp37pwFCxa0al+wYEEaGxvX6N+tW7d069atVVvv3r3fzBI7hLq6urfd5GbDZk6zoTGn2dCY02xozGk2NOY0GyLzmg3N23VOv9aK6tWKfMFi165dM3To0Nx5553VtpaWltx5551pamoqURIAAAAAAAUV2wZk3LhxGTNmTIYNG5bdd989F198cZYtW5ajjjqqVEkAAAAAABRSLKz+2Mc+lj/96U8566yzMn/+/Oy222657bbb1vjSRdbUrVu3TJgwYY2tUaCjMqfZ0JjTbGjMaTY05jQbGnOaDZF5zYbGnF43NZVKpVK6CAAAAAAA3t6K7FkNAAAAAAD/TFgNAAAAAEBxwmoAAAAAAIoTVgMAAAAAUJywuoO59NJLM2jQoHTv3j177LFHHnzwwdIlwTq59957c/DBB6d///6pqanJzTff3Op6pVLJWWedlc022ywbbbRRRowYkSeeeKJMsbAOJk2alHe/+93p1atX+vXrlw9+8IOZM2dOqz4vvvhixo4dm759+6Znz54ZPXp0FixYUKhieHWXX355dtlll9TV1aWuri5NTU259dZbq9fNZzq6yZMnp6amJieddFK1zbymo5k4cWJqampaHYMHD65eN6fpiP74xz/mk5/8ZPr27ZuNNtooO++8cx5++OHqdX8r0pEMGjRojc/pmpqajB07NonP6XUhrO5Avve972XcuHGZMGFCHnnkkey6664ZOXJkFi5cWLo0eE3Lli3LrrvumksvvXSt1y+44IJccsklueKKK/LAAw+kR48eGTlyZF588cW3uFJYNzNmzMjYsWNz//33Z/r06Vm5cmUOOuigLFu2rNrn5JNPzk9+8pPceOONmTFjRp599tkceuihBauGV7b55ptn8uTJmTVrVh5++OHsv//+OeSQQ/LYY48lMZ/p2B566KF84xvfyC677NKq3bymI9pxxx3z3HPPVY///u//rl4zp+lo/va3v2XPPfdMly5dcuutt+bxxx/Pv/3bv2WTTTap9vG3Ih3JQw891Oozevr06UmSj3zkI0l8Tq+TCh3G7rvvXhk7dmz1fNWqVZX+/ftXJk2aVLAqaLsklZtuuql63tLSUmlsbKxceOGF1bZFixZVunXrVrn++usLVAhtt3DhwkqSyowZMyqVyj/mcJcuXSo33nhjtc9vfvObSpLKzJkzS5UJbbLJJptUrrrqKvOZDu3555+vbLvttpXp06dX9tlnn8qJJ55YqVR8TtMxTZgwobLrrruu9Zo5TUd06qmnVvbaa69XvO5vRTq6E088sbL11ltXWlpafE6vIyurO4gVK1Zk1qxZGTFiRLWtU6dOGTFiRGbOnFmwMnjjnn766cyfP7/V/K6vr88ee+xhftNhLF68OEnSp0+fJMmsWbOycuXKVvN68ODBGTBggHnNem/VqlW54YYbsmzZsjQ1NZnPdGhjx45Nc3Nzq/mb+Jym43riiSfSv3//bLXVVjn88MMzb968JOY0HdOPf/zjDBs2LB/5yEfSr1+/vPOd78w3v/nN6nV/K9KRrVixIt/97nfz6U9/OjU1NT6n15GwuoP485//nFWrVqWhoaFVe0NDQ+bPn1+oKmgfq+ew+U1H1dLSkpNOOil77rlndtpppyT/mNddu3ZN7969W/U1r1mfPfroo+nZs2e6deuWL3zhC7npppsyZMgQ85kO64YbbsgjjzySSZMmrXHNvKYj2mOPPTJ16tTcdtttufzyy/P000/nPe95T55//nlzmg7p97//fS6//PJsu+22uf3223PMMcfkhBNOyDXXXJPE34p0bDfffHMWLVqUI488Mol/91hXtaULAICObuzYsfn1r3/das9I6Ii23377zJ49O4sXL84PfvCDjBkzJjNmzChdFrwuzzzzTE488cRMnz493bt3L10OtItRo0ZVf95ll12yxx57ZODAgfn+97+fjTbaqGBl8Pq0tLRk2LBhOf/885Mk73znO/PrX/86V1xxRcaMGVO4Onhjrr766owaNSr9+/cvXUqHYmV1B/GOd7wjnTt3XuMbQhcsWJDGxsZCVUH7WD2HzW86ouOOOy633HJL7r777my++ebV9sbGxqxYsSKLFi1q1d+8Zn3WtWvXbLPNNhk6dGgmTZqUXXfdNf/xH/9hPtMhzZo1KwsXLsy73vWu1NbWpra2NjNmzMgll1yS2traNDQ0mNd0eL179852222XJ5980mc1HdJmm22WIUOGtGrbYYcdqtvb+FuRjuoPf/hDfvazn+Uzn/lMtc3n9LoRVncQXbt2zdChQ3PnnXdW21paWnLnnXemqampYGXwxm255ZZpbGxsNb+XLFmSBx54wPxmvVWpVHLcccflpptuyl133ZUtt9yy1fWhQ4emS5cureb1nDlzMm/ePPOaDqOlpSXLly83n+mQDjjggDz66KOZPXt29Rg2bFgOP/zw6s/mNR3d0qVL89RTT2WzzTbzWU2HtOeee2bOnDmt2n73u99l4MCBSfytSMc1ZcqU9OvXL83NzdU2n9PrxjYgHci4ceMyZsyYDBs2LLvvvnsuvvjiLFu2LEcddVTp0uA1LV26NE8++WT1/Omnn87s2bPTp0+fDBgwICeddFLOO++8bLvtttlyyy1z5plnpn///vngBz9Yrmh4FWPHjs11112XH/3oR+nVq1d1j7H6+vpstNFGqa+vz9FHH51x48alT58+qaury/HHH5+mpqYMHz68cPWwptNOOy2jRo3KgAED8vzzz+e6667LPffck9tvv918pkPq1atX9XsEVuvRo0f69u1bbTev6Wi+9KUv5eCDD87AgQPz7LPPZsKECencuXM+/vGP+6ymQzr55JPzr//6rzn//PPz0Y9+NA8++GCuvPLKXHnllUmSmpoafyvS4bS0tGTKlCkZM2ZMamv/L3r1Ob2OKnQoX//61ysDBgyodO3atbL77rtX7r///tIlwTq5++67K0nWOMaMGVOpVCqVlpaWyplnnllpaGiodOvWrXLAAQdU5syZU7ZoeBVrm89JKlOmTKn2+fvf/1459thjK5tsskll4403rnzoQx+qPPfcc+WKhlfx6U9/ujJw4MBK165dK5tuumnlgAMOqNxxxx3V6+YzG4J99tmncuKJJ1bPzWs6mo997GOVzTbbrNK1a9fKv/zLv1Q+9rGPVZ588snqdXOajugnP/lJZaeddqp069atMnjw4MqVV17Z6rq/Felobr/99kqStc5Tn9OvraZSqVTKxOQAAAAAAPAP9qwGAAAAAKA4YTUAAAAAAMUJqwEAAAAAKE5YDQAAAABAccJqAAAAAACKE1YDAAAAAFCcsBoAAAAAgOKE1QAAAAAAFCesBgCAl9l3331z0kknlS6jw5s7d25qamoye/bs0qUAANABCKsBAFivXHHFFenVq1deeumlatvSpUvTpUuX7Lvvvq363nPPPampqclTTz31Fle5/nj66afziU98Iv3790/37t2z+eab55BDDslvf/vb0qUBAECbCKsBAFiv7Lffflm6dGkefvjhatvPf/7zNDY25oEHHsiLL75Ybb/77rszYMCAbL311m2+T6VSaRWIr+9Wrly51rYDDzwwixcvzg9/+MPMmTMn3/ve97Lzzjtn0aJFb32RAADwBgirAQBYr2y//fbZbLPNcs8991Tb7rnnnhxyyCHZcsstc//997dq32+//ZIky5cvzwknnJB+/fqle/fu2WuvvfLQQw+16ltTU5Nbb701Q4cOTbdu3fLf//3fWbZsWT71qU+lZ8+e2WyzzfJv//Zva9R02WWXZdttt0337t3T0NCQD3/4w69Y/9SpU9O7d+/cfPPN1deMHDkyzzzzTKt+P/rRj/Kud70r3bt3z1ZbbZWzzz67VXheU1OTyy+/PB/4wAfSo0ePfOUrX1njXo899lieeuqpXHbZZRk+fHgGDhyYPffcM+edd16GDx9e7Xfqqadmu+22y8Ybb5ytttoqZ555Zqvwe+LEidltt93yrW99KwMGDEjPnj1z7LHHZtWqVbngggvS2NiYfv36rVHD6hpHjRqVjTbaKFtttVV+8IMfvOJ7kyS//vWvM2rUqPTs2TMNDQ054ogj8uc///lVXwMAwNuDsBoAgPXOfvvtl7vvvrt6fvfdd2fffffNPvvsU23/+9//ngceeKAaVn/5y1/Of/3Xf+Waa67JI488km222SYjR47MX//611Zjjx8/PpMnT85vfvOb7LLLLjnllFMyY8aM/OhHP8odd9yRe+65J4888ki1/8MPP5wTTjgh55xzTubMmZPbbrste++996vW/8ILL+QrX/lKvv3tb+cXv/hFFi1alMMOO6x6/ec//3k+9alP5cQTT8zjjz+eb3zjG5k6deoaYfDEiRPzoQ99KI8++mg+/elPr3GfTTfdNJ06dcoPfvCDrFq16hXr6dWrV6ZOnZrHH388//Ef/5FvfvObueiii1r1eeqpp3Lrrbfmtttuy/XXX5+rr746zc3N+d///d/MmDEjX/3qV3PGGWfkgQceaPW6M888M6NHj84vf/nLHH744TnssMPym9/8Zq11LFq0KPvvv3/e+c535uGHH85tt92WBQsW5KMf/eirvp8AALxNVAAAYD3zzW9+s9KjR4/KypUrK0uWLKnU1tZWFi5cWLnuuusqe++9d6VSqVTuvPPOSpLKH/7wh8rSpUsrXbp0qVx77bXVMVasWFHp379/5YILLqhUKpXK3XffXUlSufnmm6t9nn/++UrXrl0r3//+96ttf/nLXyobbbRR5cQTT6xUKpXKf/3Xf1Xq6uoqS5YsWafap0yZUklSuf/++6ttv/nNbypJKg888EClUqlUDjjggMr555/f6nXf+c53Kptttln1PEnlpJNOes37/ed//mdl4403rvTq1auy3377Vc4555zKU0899aqvufDCCytDhw6tnk+YMKGy8cYbt3rGkSNHVgYNGlRZtWpVtW377bevTJo0qVWNX/jCF1qNvccee1SOOeaYSqVSqTz99NOVJJX/+Z//qVQqlcq5555bOeigg1r1f+aZZypJKnPmzHnNZwUAYMNmZTUAAOudfffdN8uWLctDDz2Un//859luu+2y6aabZp999qnuW33PPfdkq622yoABA/LUU09l5cqV2XPPPatjdOnSJbvvvvsaq3yHDRtW/fmpp57KihUrsscee1Tb+vTpk+233756fuCBB2bgwIHZaqutcsQRR+Taa6/NCy+88Kr119bW5t3vfnf1fPDgwendu3e1ll/+8pc555xz0rNnz+rx2c9+Ns8991yrsf+51lcyduzYzJ8/P9dee22amppy4403Zscdd8z06dOrfb73ve9lzz33TGNjY3r27Jkzzjgj8+bNazXOoEGD0qtXr+p5Q0NDhgwZkk6dOrVqW7hwYavXNTU1rXH+Siurf/nLX+buu+9u9dyDBw9Okrf1l2QCAPAPwmoAANY722yzTTbffPPcfffdufvuu7PPPvskSfr3758tttgi9913X+6+++7sv//+bR67R48eberfq1evPPLII7n++uuz2Wab5ayzzsquu+76hr7AcOnSpTn77LMze/bs6vHoo4/miSeeSPfu3dtca69evXLwwQfnK1/5Sn75y1/mPe95T84777wkycyZM3P44Yfnfe97X2655Zb8z//8T04//fSsWLGi1RhdunRpdV5TU7PWtpaWltfzyEn+8dwHH3xwq+eePXt2nnjiidfcWgUAgA2fsBoAgPXSfvvtl3vuuSf33HNP9t1332r73nvvnVtvvTUPPvhgdb/qrbfeOl27ds0vfvGLar+VK1fmoYceypAhQ17xHltvvXW6dOnSah/mv/3tb/nd737Xql9tbW1GjBiRCy64IL/61a8yd+7c3HXXXa847ksvvZSHH364ej5nzpwsWrQoO+ywQ5LkXe96V+bMmZNtttlmjeOfVzK/HjU1NRk8eHCWLVuWJLnvvvsycODAnH766Rk2bFi23Xbb/OEPf3hD9/hn//yFl6vPVz/ny73rXe/KY489lkGDBq3x3G39jwgAAGx4aksXAAAAa7Pffvtl7NixWblyZXVldZLss88+Oe6447JixYpqWN2jR48cc8wxOeWUU9KnT58MGDAgF1xwQV544YUcffTRr3iPnj175uijj84pp5ySvn37pl+/fjn99NNbBca33HJLfv/732fvvffOJptskp/+9KdpaWlptVXIy3Xp0iXHH398LrnkktTW1ua4447L8OHDs/vuuydJzjrrrLz//e/PgAED8uEPfzidOnXKL3/5y/z617+uroheF7Nnz86ECRNyxBFHZMiQIenatWtmzJiRb33rWzn11FOTJNtuu23mzZuXG264Ie9+97szbdq03HTTTet8j9dy4403ZtiwYdlrr71y7bXX5sEHH8zVV1+91r5jx47NN7/5zXz84x/Pl7/85fTp0ydPPvlkbrjhhlx11VXp3Llzu9UFAEDHI6wGAGC9tN9+++Xvf/97Bg8enIaGhmr7Pvvsk+effz7bb799Nttss2r75MmT09LSkiOOOCLPP/98hg0blttvvz2bbLLJq97nwgsvrG5P0atXr3zxi1/M4sWLq9d79+6dH/7wh5k4cWJefPHFbLvttrn++uuz4447vuKYG2+8cU499dR84hOfyB//+Me85z3vaRXgjhw5MrfcckvOOeecfPWrX02XLl0yePDgfOYzn2nTe7T55ptn0KBBOfvsszN37tzU1NRUz08++eQkyQc+8IGcfPLJOe6447J8+fI0NzfnzDPPzMSJE9t0r1dy9tln54Ybbsixxx6bzTbbLNdff/0rrmbv379/fvGLX+TUU0/NQQcdlOXLl2fgwIF573vf+4ZXlAMA0PHVVCqVSukiAABgQzF16tScdNJJb2hP646ipqYmN910Uz74wQ+WLgUAgA2A5QsAAAAAABQnrAYAAAAAoDjbgAAAAAAAUJyV1QAAAAAAFCesBgAAAACgOGE1AAAAAADFCasBAAAAAChOWA0AAAAAQHHCagAAAAAAihNWAwAAAABQnLAaAAAAAIDi/j+M76P+2HbP0QAAAABJRU5ErkJggg==",
      "text/plain": [
       "<Figure size 1800x800 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df['Text Length'] = df['text'].str.split().str.len()\n",
    "\n",
    "max_len = df['Text Length'].max()\n",
    "print(f\"The maximum length of this dataset is approximately {max_len} words.\")\n",
    "\n",
    "n = 70\n",
    "df.hist('Text Length', grid=False, figsize=(18,8), range=(0, n), bins=n)\n",
    "plt.suptitle('')\n",
    "plt.title('Histogram of Text Length')\n",
    "plt.ylabel('')\n",
    "plt.xlabel('Words per Sample')\n",
    "plt.show()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Visualize: Length Text Length (in Words) By Language"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1800x800 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df.boxplot(\"Text Length\", by=\"labels\", grid=False, showfliers=False, color=\"Blue\", figsize=(18, 8))\n",
    "\n",
    "plt.suptitle('')\n",
    "plt.title('Description Length by Output Class')\n",
    "plt.xlabel('')\n",
    "plt.ylabel('Words per Complaint')\n",
    "plt.show()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Visualize: Number of Samples of Each Language"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1800x800 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "df['labels'].value_counts().plot(kind='bar', color=\"Blue\", figsize=(18, 8))\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Convert Labels Between String & Integer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "['Danish', 'Turkish', 'Malayalam', 'English', 'Russian', 'Spanish', 'Sweedish', 'German', 'French', 'Greek', 'Portugeese', 'Italian', 'Kannada', 'Arabic', 'Tamil', 'Dutch'] \n",
      " 16\n",
      "id2label: {0: 'Danish', 1: 'Turkish', 2: 'Malayalam', 3: 'English', 4: 'Russian', 5: 'Spanish', 6: 'Sweedish', 7: 'German', 8: 'French', 9: 'Greek', 10: 'Portugeese', 11: 'Italian', 12: 'Kannada', 13: 'Arabic', 14: 'Tamil', 15: 'Dutch'}\n",
      "label2id: {'Danish': 0, 'Turkish': 1, 'Malayalam': 2, 'English': 3, 'Russian': 4, 'Spanish': 5, 'Sweedish': 6, 'German': 7, 'French': 8, 'Greek': 9, 'Portugeese': 10, 'Italian': 11, 'Kannada': 12, 'Arabic': 13, 'Tamil': 14, 'Dutch': 15}\n"
     ]
    }
   ],
   "source": [
    "LABEL_NAMES = list(set(df['labels']))\n",
    "NUM_OF_LABELS = len(LABEL_NAMES)\n",
    "\n",
    "id2label = dict(enumerate(LABEL_NAMES))\n",
    "label2id = {v: k for k,v in id2label.items()}\n",
    "\n",
    "print(LABEL_NAMES, '\\n', NUM_OF_LABELS)\n",
    "print(\"id2label:\", id2label)\n",
    "print(\"label2id:\", label2id)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Final Preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "Int64Index: 10183 entries, 0 to 10336\n",
      "Data columns (total 2 columns):\n",
      " #   Column  Non-Null Count  Dtype \n",
      "---  ------  --------------  ----- \n",
      " 0   text    10183 non-null  object\n",
      " 1   labels  10183 non-null  int64 \n",
      "dtypes: int64(1), object(1)\n",
      "memory usage: 238.7+ KB\n"
     ]
    }
   ],
   "source": [
    "# Replace String label values with Integer label values\n",
    "df['labels'].replace(label2id, inplace=True)\n",
    "\n",
    "# Remove overly lengthy samples\n",
    "df = df[df[\"Text Length\"] < 71]\n",
    "\n",
    "# Remove unnecessary Feature\n",
    "df.drop(columns=[\"Text Length\"], inplace=True)\n",
    "df.info()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Convert DataFrame to Dataset & Split into Train/Test/Eval Datasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "training data shape: (8146, 3)\n",
      "testing data shape: (1018, 3)\n",
      "validation data shape: (1019, 3)\n"
     ]
    }
   ],
   "source": [
    "dataset = Dataset.from_pandas(df)\n",
    "\n",
    "train_testEval = dataset.train_test_split(train_size=0.80)\n",
    "test_eval = train_testEval['test'].train_test_split(train_size=0.50)\n",
    "\n",
    "ds = DatasetDict({\n",
    "    'train' : train_testEval['train'],\n",
    "    'test' : test_eval['train'],\n",
    "    'eval' : test_eval['test']\n",
    "})\n",
    "\n",
    "print('training data shape:', ds['train'].shape)\n",
    "print('testing data shape:', ds['test'].shape)\n",
    "print('validation data shape:', ds['eval'].shape)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Basic Values/Constants"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "MODEL_CKPT = \"distilbert-base-multilingual-cased\"\n",
    "              \n",
    "MODEL_NAME = MODEL_CKPT.split(\"/\")[-1] + \"-language_detection\"\n",
    "\n",
    "DEVICE = torch.device(\"cpu\")\n",
    "MAX_INPUT_LENGTH = int(70*1.30)\n",
    "\n",
    "LEARNING_RATE = 2e-5\n",
    "set_seed(42)\n",
    "\n",
    "BATCH_SIZE = 64\n",
    "NUM_OF_EPOCHS = 3\n",
    "\n",
    "METRIC_NAME = \"f1\"\n",
    "WEIGHT_DECAY = 0.01\n",
    "\n",
    "REPORTS_TO = \"tensorboard\"\n",
    "STRATEGY = \"epoch\""
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Tokenize Entire Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "51adb5f4da86487185c4f2ed49cd4997",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "A Jupyter Widget"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "1818d2bd3b2b4bc9b2fb0f9d11137fb6",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "A Jupyter Widget"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "76e817e1da574ac0b067f0b2f6c8f144",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "A Jupyter Widget"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "training data shape: (8146, 5)\n",
      "testing data shape: (1018, 5)\n",
      "validation data shape: (1019, 5)\n",
      "{'text': 'Mer komplicerad kodning av artiklar och mallar kan även göras i märkspråken html (för avancerad layout) och TeX (för matematiska formler och notskrift) samt programspråket Lua (för att skapa programmoduler som kan användas i avancerade mallar, och som körs på Mediawikiservern).', 'labels': 6, '__index_level_0__': 8367, 'input_ids': [101, 10734, 12240, 62631, 18599, 15328, 11269, 10170, 46118, 38255, 10177, 76437, 10354, 10905, 14127, 59805, 10107, 177, 181, 76743, 54609, 69379, 11062, 13222, 113, 10847, 41448, 18599, 71888, 114, 10177, 21452, 12674, 113, 10847, 43465, 106530, 10113, 12188, 10815, 10177, 10472, 65692, 114, 13611, 19316, 52302, 16343, 13650, 74564, 113, 10847, 10788, 31569, 11359, 12483, 11033, 42918, 10165, 10181, 10905, 91277, 177, 41448, 18752, 76437, 10354, 117, 10177, 10181, 68077, 10107, 10217, 13341, 92700, 12754, 38937, 114, 119, 102, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], 'attention_mask': [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]}\n"
     ]
    }
   ],
   "source": [
    "# Define the Tokenizer\n",
    "tokenizer = AutoTokenizer.from_pretrained(MODEL_CKPT)\n",
    "\n",
    "# Method to Tokenize each Sample in the Dataset\n",
    "def tokenize(batch):\n",
    "    return tokenizer(batch['text'], truncation=True, padding=True, max_length=MAX_INPUT_LENGTH)\n",
    "\n",
    "# Map the 'Tokenize' Method to All Samples in the Entire Dataset\n",
    "encoded_data = ds.map(tokenize, batched=True, batch_size=None)\n",
    "encoded_data = encoded_data\n",
    "\n",
    "print('training data shape:', encoded_data['train'].shape)\n",
    "print('testing data shape:', encoded_data['test'].shape)\n",
    "print('validation data shape:', encoded_data['eval'].shape)\n",
    "print(encoded_data['train'][24])"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Instantiate Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Some weights of the model checkpoint at distilbert-base-multilingual-cased were not used when initializing DistilBertForSequenceClassification: ['vocab_transform.weight', 'vocab_layer_norm.weight', 'vocab_projector.bias', 'vocab_projector.weight', 'vocab_layer_norm.bias', 'vocab_transform.bias']\n",
      "- This IS expected if you are initializing DistilBertForSequenceClassification from the checkpoint of a model trained on another task or with another architecture (e.g. initializing a BertForSequenceClassification model from a BertForPreTraining model).\n",
      "- This IS NOT expected if you are initializing DistilBertForSequenceClassification from the checkpoint of a model that you expect to be exactly identical (initializing a BertForSequenceClassification model from a BertForSequenceClassification model).\n",
      "Some weights of DistilBertForSequenceClassification were not initialized from the model checkpoint at distilbert-base-multilingual-cased and are newly initialized: ['classifier.bias', 'classifier.weight', 'pre_classifier.bias', 'pre_classifier.weight']\n",
      "You should probably TRAIN this model on a down-stream task to be able to use it for predictions and inference.\n"
     ]
    }
   ],
   "source": [
    "model = AutoModelForSequenceClassification.from_pretrained(MODEL_CKPT, \n",
    "                                                           num_labels=NUM_OF_LABELS,\n",
    "                                                           id2label=id2label,\n",
    "                                                           label2id=label2id\n",
    "                                                           ).to(DEVICE)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Define Function to Compute Metrics"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "def compute_metrics(p):\n",
    "    ### ------------------- Accuracy -------------------\n",
    "    \n",
    "    accuracy_metric = evaluate.load(\"accuracy\")\n",
    "    accuracy = accuracy_metric.compute(predictions=np.argmax(p.predictions, axis=1), references=p.label_ids)['accuracy']\n",
    "    \n",
    "    ### ------------------- F1 scores -------------------\n",
    "    \n",
    "    f1_score_metric = evaluate.load(\"f1\")\n",
    "    weighted_f1_score = f1_score_metric.compute(predictions=np.argmax(p.predictions, axis=1), references=p.label_ids, average='weighted')[\"f1\"]\n",
    "    micro_f1_score = f1_score_metric.compute(predictions=np.argmax(p.predictions, axis=1), references=p.label_ids, average='micro')['f1']\n",
    "    macro_f1_score = f1_score_metric.compute(predictions=np.argmax(p.predictions, axis=1), references=p.label_ids, average='macro')[\"f1\"]\n",
    "    \n",
    "    ### ------------------- recall -------------------\n",
    "    \n",
    "    recall_metric = evaluate.load(\"recall\")\n",
    "    weighted_recall = recall_metric.compute(predictions=np.argmax(p.predictions, axis=1), references=p.label_ids, average='weighted')[\"recall\"]\n",
    "    micro_recall = recall_metric.compute(predictions=np.argmax(p.predictions, axis=1), references=p.label_ids, average='micro')[\"recall\"]\n",
    "    macro_recall = recall_metric.compute(predictions=np.argmax(p.predictions, axis=1), references=p.label_ids, average='macro')[\"recall\"]\n",
    "    \n",
    "    ### ------------------- precision -------------------\n",
    "    \n",
    "    precision_metric = evaluate.load(\"precision\")\n",
    "    weighted_precision = precision_metric.compute(predictions=np.argmax(p.predictions, axis=1), references=p.label_ids, average='weighted')[\"precision\"]\n",
    "    micro_precision = precision_metric.compute(predictions=np.argmax(p.predictions, axis=1), references=p.label_ids, average='micro')[\"precision\"]\n",
    "    macro_precision = precision_metric.compute(predictions=np.argmax(p.predictions, axis=1), references=p.label_ids, average='macro')[\"precision\"]\n",
    "    \n",
    "    return {\"accuracy\" : accuracy, \n",
    "            \"Weighted F1\" : weighted_f1_score,\n",
    "            \"Micro F1\" : micro_f1_score,\n",
    "            \"Macro F1\" : macro_f1_score,\n",
    "            \"Weighted Recall\" : weighted_recall,\n",
    "            \"Micro Recall\" : micro_recall,\n",
    "            \"Macro Recall\" : macro_recall,\n",
    "            \"Weighted Precision\" : weighted_precision,\n",
    "            \"Micro Precision\" : micro_precision,\n",
    "            \"Macro Precision\" : macro_precision\n",
    "            }"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Define TrainingArguments"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "args = TrainingArguments(\n",
    "    output_dir=MODEL_NAME,\n",
    "    num_train_epochs=NUM_OF_EPOCHS,\n",
    "    per_device_train_batch_size=BATCH_SIZE,\n",
    "    per_device_eval_batch_size=BATCH_SIZE,\n",
    "    learning_rate=LEARNING_RATE,\n",
    "    weight_decay=WEIGHT_DECAY,\n",
    "    evaluation_strategy=STRATEGY,\n",
    "    save_strategy=STRATEGY,\n",
    "    logging_strategy=STRATEGY,\n",
    "    logging_first_step=True,\n",
    "    disable_tqdm=False,\n",
    "    report_to=REPORTS_TO,\n",
    "    load_best_model_at_end=True,\n",
    "    metric_for_best_model=\"eval_accuracy\",\n",
    "    hub_private_repo=True,\n",
    "    push_to_hub=True\n",
    ")"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Instantiate Data Collator"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "tokenized_datasets = encoded_data.remove_columns(ds[\"train\"].column_names)\n",
    "data_collator = DataCollatorWithPadding(tokenizer, padding=True)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Subclass Trainer to Correctly Compute Loss for Imbalanced Classes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "class CustomTrainer(Trainer):\n",
    "    def compute_loss(self, model, inputs, return_outputs=False):\n",
    "        labels = inputs.get(\"labels\")\n",
    "        outputs = model(**inputs)\n",
    "        logits = outputs.get(\"logits\")\n",
    "        loss_fct = torch.nn.CrossEntropyLoss(\n",
    "            weight=torch.tensor(\n",
    "                [1.0, 2.0, 3.0, 4.0, 5.0, 6.0, 7.0, 8.0, 9.0, 10.0, 11.0, 12.0, 13.0, 14.0, 15.0, 16.0]\n",
    "                )\n",
    "            )\n",
    "        loss = loss_fct(logits.view(-1, self.model.config.num_labels), labels.view(-1))\n",
    "        return (loss, outputs) if return_outputs else loss"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Instantiate Trainer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Cloning https://huggingface.co/DunnBC22/distilbert-base-multilingual-cased-language_detection into local empty directory.\n"
     ]
    }
   ],
   "source": [
    "trainer = CustomTrainer(\n",
    "    model=model,\n",
    "    args=args,\n",
    "    train_dataset=encoded_data[\"train\"],\n",
    "    eval_dataset=encoded_data[\"eval\"],\n",
    "    compute_metrics=compute_metrics,\n",
    "    tokenizer=tokenizer,\n",
    "    data_collator=data_collator,\n",
    ")"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Train Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "The following columns in the training set don't have a corresponding argument in `DistilBertForSequenceClassification.forward` and have been ignored: __index_level_0__, text. If __index_level_0__, text are not expected by `DistilBertForSequenceClassification.forward`,  you can safely ignore this message.\n",
      "/Users/briandunn/Documents/nlpnn/nlp_ml/lib/python3.9/site-packages/transformers/optimization.py:306: FutureWarning: This implementation of AdamW is deprecated and will be removed in a future version. Use the PyTorch implementation torch.optim.AdamW instead, or set `no_deprecation_warning=True` to disable this warning\n",
      "  warnings.warn(\n",
      "***** Running training *****\n",
      "  Num examples = 8146\n",
      "  Num Epochs = 3\n",
      "  Instantaneous batch size per device = 64\n",
      "  Total train batch size (w. parallel, distributed & accumulation) = 64\n",
      "  Gradient Accumulation steps = 1\n",
      "  Total optimization steps = 384\n",
      "  Number of trainable parameters = 135336976\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "9b6743859a634cf399d58f9a56fc30c7",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "A Jupyter Widget"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "You're using a DistilBertTokenizerFast tokenizer. Please note that with a fast tokenizer, using the `__call__` method is faster than using a method to encode the text followed by a call to the `pad` method to get a padded encoding.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'loss': 2.8224, 'learning_rate': 1.994791666666667e-05, 'epoch': 0.01}\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "The following columns in the evaluation set don't have a corresponding argument in `DistilBertForSequenceClassification.forward` and have been ignored: __index_level_0__, text. If __index_level_0__, text are not expected by `DistilBertForSequenceClassification.forward`,  you can safely ignore this message.\n",
      "***** Running Evaluation *****\n",
      "  Num examples = 1019\n",
      "  Batch size = 64\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'loss': 1.0783, 'learning_rate': 1.3333333333333333e-05, 'epoch': 1.0}\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "4d264a2afb2f4ca2907d9c6eb6f7c922",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "A Jupyter Widget"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Saving model checkpoint to distilbert-base-multilingual-cased-language_detection/checkpoint-128\n",
      "Configuration saved in distilbert-base-multilingual-cased-language_detection/checkpoint-128/config.json\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'eval_loss': 0.15443778038024902, 'eval_accuracy': 0.9823356231599607, 'eval_Weighted F1': 0.9818688279398573, 'eval_Micro F1': 0.9823356231599607, 'eval_Macro F1': 0.9805828778996699, 'eval_Weighted Recall': 0.9823356231599607, 'eval_Micro Recall': 0.9823356231599607, 'eval_Macro Recall': 0.9797963651863908, 'eval_Weighted Precision': 0.9846772481286562, 'eval_Micro Precision': 0.9823356231599607, 'eval_Macro Precision': 0.9852381510538253, 'eval_runtime': 210.5251, 'eval_samples_per_second': 4.84, 'eval_steps_per_second': 0.076, 'epoch': 1.0}\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Model weights saved in distilbert-base-multilingual-cased-language_detection/checkpoint-128/pytorch_model.bin\n",
      "tokenizer config file saved in distilbert-base-multilingual-cased-language_detection/checkpoint-128/tokenizer_config.json\n",
      "Special tokens file saved in distilbert-base-multilingual-cased-language_detection/checkpoint-128/special_tokens_map.json\n",
      "tokenizer config file saved in distilbert-base-multilingual-cased-language_detection/tokenizer_config.json\n",
      "Special tokens file saved in distilbert-base-multilingual-cased-language_detection/special_tokens_map.json\n",
      "Adding files tracked by Git LFS: ['.DS_Store']. This may take a bit of time if the files are large.\n",
      "The following columns in the evaluation set don't have a corresponding argument in `DistilBertForSequenceClassification.forward` and have been ignored: __index_level_0__, text. If __index_level_0__, text are not expected by `DistilBertForSequenceClassification.forward`,  you can safely ignore this message.\n",
      "***** Running Evaluation *****\n",
      "  Num examples = 1019\n",
      "  Batch size = 64\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'loss': 0.1189, 'learning_rate': 6.666666666666667e-06, 'epoch': 2.0}\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "82c1ae8db8034bc0ae179cb774b30a60",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "A Jupyter Widget"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Saving model checkpoint to distilbert-base-multilingual-cased-language_detection/checkpoint-256\n",
      "Configuration saved in distilbert-base-multilingual-cased-language_detection/checkpoint-256/config.json\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'eval_loss': 0.05953313037753105, 'eval_accuracy': 0.9970559371933267, 'eval_Weighted F1': 0.9970530462874927, 'eval_Micro F1': 0.9970559371933267, 'eval_Macro F1': 0.9977369718278565, 'eval_Weighted Recall': 0.9970559371933267, 'eval_Micro Recall': 0.9970559371933267, 'eval_Macro Recall': 0.9973849347051074, 'eval_Weighted Precision': 0.997074923908581, 'eval_Micro Precision': 0.9970559371933267, 'eval_Macro Precision': 0.9981114357461582, 'eval_runtime': 203.6131, 'eval_samples_per_second': 5.005, 'eval_steps_per_second': 0.079, 'epoch': 2.0}\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Model weights saved in distilbert-base-multilingual-cased-language_detection/checkpoint-256/pytorch_model.bin\n",
      "tokenizer config file saved in distilbert-base-multilingual-cased-language_detection/checkpoint-256/tokenizer_config.json\n",
      "Special tokens file saved in distilbert-base-multilingual-cased-language_detection/checkpoint-256/special_tokens_map.json\n",
      "tokenizer config file saved in distilbert-base-multilingual-cased-language_detection/tokenizer_config.json\n",
      "Special tokens file saved in distilbert-base-multilingual-cased-language_detection/special_tokens_map.json\n",
      "The following columns in the evaluation set don't have a corresponding argument in `DistilBertForSequenceClassification.forward` and have been ignored: __index_level_0__, text. If __index_level_0__, text are not expected by `DistilBertForSequenceClassification.forward`,  you can safely ignore this message.\n",
      "***** Running Evaluation *****\n",
      "  Num examples = 1019\n",
      "  Batch size = 64\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'loss': 0.0651, 'learning_rate': 0.0, 'epoch': 3.0}\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "3bf232a3817647999659478c1e1094c8",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "A Jupyter Widget"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Saving model checkpoint to distilbert-base-multilingual-cased-language_detection/checkpoint-384\n",
      "Configuration saved in distilbert-base-multilingual-cased-language_detection/checkpoint-384/config.json\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'eval_loss': 0.04728417843580246, 'eval_accuracy': 0.9970559371933267, 'eval_Weighted F1': 0.9970530462874927, 'eval_Micro F1': 0.9970559371933267, 'eval_Macro F1': 0.9977369718278565, 'eval_Weighted Recall': 0.9970559371933267, 'eval_Micro Recall': 0.9970559371933267, 'eval_Macro Recall': 0.9973849347051074, 'eval_Weighted Precision': 0.997074923908581, 'eval_Micro Precision': 0.9970559371933267, 'eval_Macro Precision': 0.9981114357461582, 'eval_runtime': 203.3366, 'eval_samples_per_second': 5.011, 'eval_steps_per_second': 0.079, 'epoch': 3.0}\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Model weights saved in distilbert-base-multilingual-cased-language_detection/checkpoint-384/pytorch_model.bin\n",
      "tokenizer config file saved in distilbert-base-multilingual-cased-language_detection/checkpoint-384/tokenizer_config.json\n",
      "Special tokens file saved in distilbert-base-multilingual-cased-language_detection/checkpoint-384/special_tokens_map.json\n",
      "tokenizer config file saved in distilbert-base-multilingual-cased-language_detection/tokenizer_config.json\n",
      "Special tokens file saved in distilbert-base-multilingual-cased-language_detection/special_tokens_map.json\n",
      "\n",
      "\n",
      "Training completed. Do not forget to share your model on huggingface.co/models =)\n",
      "\n",
      "\n",
      "Loading best model from distilbert-base-multilingual-cased-language_detection/checkpoint-256 (score: 0.9970559371933267).\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'train_runtime': 15770.3698, 'train_samples_per_second': 1.55, 'train_steps_per_second': 0.024, 'train_loss': 0.42528709520896274, 'epoch': 3.0}\n"
     ]
    }
   ],
   "source": [
    "train_results = trainer.train()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Evaluate Model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "The following columns in the evaluation set don't have a corresponding argument in `DistilBertForSequenceClassification.forward` and have been ignored: __index_level_0__, text. If __index_level_0__, text are not expected by `DistilBertForSequenceClassification.forward`,  you can safely ignore this message.\n",
      "***** Running Evaluation *****\n",
      "  Num examples = 1019\n",
      "  Batch size = 64\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "99562057f2424e6985737c903a4f5178",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "A Jupyter Widget"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'eval_loss': 0.05953313037753105, 'eval_accuracy': 0.9970559371933267, 'eval_Weighted F1': 0.9970530462874927, 'eval_Micro F1': 0.9970559371933267, 'eval_Macro F1': 0.9977369718278565, 'eval_Weighted Recall': 0.9970559371933267, 'eval_Micro Recall': 0.9970559371933267, 'eval_Macro Recall': 0.9973849347051074, 'eval_Weighted Precision': 0.997074923908581, 'eval_Micro Precision': 0.9970559371933267, 'eval_Macro Precision': 0.9981114357461582, 'eval_runtime': 206.7682, 'eval_samples_per_second': 4.928, 'eval_steps_per_second': 0.077, 'epoch': 3.0}\n"
     ]
    }
   ],
   "source": [
    "output_metrics = trainer.evaluate()\n",
    "\n",
    "print(output_metrics)"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Push Fine-Tuned Model to HuggingFace Hub (My Profile!!!)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Saving model checkpoint to distilbert-base-multilingual-cased-language_detection\n",
      "Configuration saved in distilbert-base-multilingual-cased-language_detection/config.json\n",
      "Model weights saved in distilbert-base-multilingual-cased-language_detection/pytorch_model.bin\n",
      "tokenizer config file saved in distilbert-base-multilingual-cased-language_detection/tokenizer_config.json\n",
      "Special tokens file saved in distilbert-base-multilingual-cased-language_detection/special_tokens_map.json\n",
      "Several commits (2) will be pushed upstream.\n",
      "The progress bars may be unreliable.\n"
     ]
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "098317ec53024aa4965bdac645d1ed9f",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "A Jupyter Widget"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "c4ae8eabd88d458eabedc7b6f9e0f465",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "A Jupyter Widget"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "ba882b526d454a9bb3adc12fca6952a5",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "A Jupyter Widget"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "remote: Scanning LFS files for validity...        \n",
      "remote: LFS file scan complete.        \n",
      "To https://huggingface.co/DunnBC22/distilbert-base-multilingual-cased-language_detection\n",
      "   4cbcd02..f9f839f  main -> main\n",
      "\n",
      "Dropping the following result as it does not have all the necessary fields:\n",
      "{'task': {'name': 'Text Classification', 'type': 'text-classification'}, 'metrics': [{'name': 'Accuracy', 'type': 'accuracy', 'value': 0.9970559371933267}]}\n",
      "To https://huggingface.co/DunnBC22/distilbert-base-multilingual-cased-language_detection\n",
      "   f9f839f..86be9b0  main -> main\n",
      "\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "'https://huggingface.co/DunnBC22/distilbert-base-multilingual-cased-language_detection/commit/f9f839f99b25ff49e3679e36452093d3085454a0'"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "trainer.push_to_hub(commit_message=\"All Dunn!!!\")"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Notes & Other Takeaways From This Projects\n",
    "****\n",
    "- The results exceeded my expectations!\n",
    "- I removed the Hindi language samples as they were far longer than the samples for the other 16 languages. This was done to keep the training time as reasonable as possible.\n",
    "****"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Citations\n",
    "\n",
    "- Model Checkpoint\n",
    "    - @article{Sanh2019DistilBERTAD,\n",
    "  title={DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter},\n",
    "  author={Victor Sanh and Lysandre Debut and Julien Chaumond and Thomas Wolf},\n",
    "  journal={ArXiv},\n",
    "  year={2019},\n",
    "  volume={abs/1910.01108}\n",
    "}"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "nlp_ch4",
   "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": "3.9.12"
  },
  "orig_nbformat": 4,
  "vscode": {
   "interpreter": {
    "hash": "a52fe47989fdc78fafbb981021cec52a6b82df6453830b9ffbd04250493e6cab"
   }
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}