course/videos/building_tokenizer.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This notebook regroups the code sample of the video below, which is a part of the [Hugging Face course](https://huggingface.co/course)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "cellView": "form"
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<iframe width=\"560\" height=\"315\" src=\"https://www.youtube.com/embed/MR8tZm5ViWU?rel=0&amp;controls=0&amp;showinfo=0\" frameborder=\"0\" allowfullscreen></iframe>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "execution_count": null,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#@title\n",
    "from IPython.display import HTML\n",
    "\n",
    "HTML('<iframe width=\"560\" height=\"315\" src=\"https://www.youtube.com/embed/MR8tZm5ViWU?rel=0&amp;controls=0&amp;showinfo=0\" frameborder=\"0\" allowfullscreen></iframe>')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Install the Transformers and Datasets libraries to run this notebook."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "! pip install datasets transformers[sentencepiece]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from datasets import load_dataset\n",
    "\n",
    "dataset = load_dataset(\"wikitext\", name=\"wikitext-2-raw-v1\", split=\"train\")\n",
    "\n",
    "\n",
    "def get_training_corpus():\n",
    "    for i in range(0, len(dataset), 1000):\n",
    "        yield dataset[i : i + 1000][\"text\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tokenizers import Tokenizer, models, normalizers, pre_tokenizers, trainers, processors, decoders"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "tokenizer = Tokenizer(models.WordPiece(unk_token=\"[UNK]\"))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "tokenizer.normalizer = normalizers.Sequence(\n",
    "    [\n",
    "        normalizers.Replace(Regex(r\"[\\p{Other}&&[^\\n\\t\\r]]\"), \"\"),\n",
    "        normalizers.Replace(Regex(r\"[\\s]\"), \" \"),\n",
    "        normalizers.Lowercase(),\n",
    "        normalizers.NFD(), normalizers.StripAccents()]\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "tokenizer.pre_tokenizer = pre_tokenizers.Whitespace()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "special_tokens = [\"[UNK]\", \"[PAD]\", \"[CLS]\", \"[SEP]\", \"[MASK]\"]\n",
    "trainer = trainers.WordPieceTrainer(vocab_size=25000, special_tokens=special_tokens)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "tokenizer.train_from_iterator(get_training_corpus(), trainer=trainer)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "cls_token_id = tokenizer.token_to_id(\"[CLS]\")\n",
    "sep_token_id = tokenizer.token_to_id(\"[SEP]\")\n",
    "tokenizer.post_processor = processors.TemplateProcessing(\n",
    "    single=f\"[CLS]:0 $A:0 [SEP]:0\",\n",
    "    pair=f\"[CLS]:0 $A:0 [SEP]:0 $B:1 [SEP]:1\",\n",
    "    special_tokens=[(\"[CLS]\", cls_token_id), (\"[SEP]\", sep_token_id)],\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "tokenizer.decoder = decoders.WordPiece(prefix=\"##\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "colab": {
   "name": "Building a new tokenizer",
   "provenance": []
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}