course/videos/token_processing.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/iY2AZYdZAr0?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/iY2AZYdZAr0?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",
    "raw_datasets = load_dataset(\"conll2003\")\n",
    "raw_datasets = raw_datasets.remove_columns([\"chunk_tags\", \"id\", \"pos_tags\"])\n",
    "raw_datasets = raw_datasets.rename_column(\"ner_tags\", \"labels\")\n",
    "raw_datasets = raw_datasets.rename_column(\"tokens\", \"words\")\n",
    "raw_datasets[\"train\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "print(raw_datasets[\"train\"][0][\"words\"])\n",
    "print(raw_datasets[\"train\"][0][\"labels\"])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "label_names = raw_datasets[\"train\"].features[\"labels\"].feature.names\n",
    "label_names"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from transformers import AutoTokenizer\n",
    "\n",
    "model_checkpoint = \"bert-base-cased\"\n",
    "tokenizer = AutoTokenizer.from_pretrained(model_checkpoint)\n",
    "\n",
    "inputs = tokenizer(raw_datasets[\"train\"][0][\"words\"], is_split_into_words=True)\n",
    "inputs.tokens()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def shift_label(label):\n",
    "    # If the label is B-XXX we change it to I-XXX\n",
    "    if label % 2 == 1:\n",
    "        label += 1\n",
    "    return label\n",
    "\n",
    "def align_labels_with_tokens(labels, word_ids):\n",
    "    new_labels = []\n",
    "    current_word = None\n",
    "    for word_id in word_ids:\n",
    "        if word_id is None:\n",
    "            new_labels.append(-100)\n",
    "        elif word_id != current_word:\n",
    "            # Start of a new word!\n",
    "            current_word = word_id\n",
    "            new_labels.append(labels[word_id])\n",
    "        else:\n",
    "            new_labels.append(shift_label(labels[word_id]))\n",
    "\n",
    "    return new_labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def tokenize_and_align_labels(examples):\n",
    "    tokenized_inputs = tokenizer(examples[\"words\"], truncation=True, is_split_into_words=True)\n",
    "    new_labels = []\n",
    "    for i, labels in enumerate(examples[\"labels\"]):\n",
    "        word_ids = tokenized_inputs.word_ids(i)\n",
    "        new_labels.append(align_labels_with_tokens(labels, word_ids))\n",
    "\n",
    "    tokenized_inputs[\"labels\"] = new_labels\n",
    "    return tokenized_inputs\n",
    "\n",
    "tokenized_datasets = raw_datasets.map(tokenize_and_align_labels, batched=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from transformers import DataCollatorForTokenClassification\n",
    "\n",
    "data_collator = DataCollatorForTokenClassification(tokenizer=tokenizer)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "colab": {
   "name": "Data processing for Token Classification",
   "provenance": []
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}