aitech-eks-pub/cw/10_CRF.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "![Logo 1](https://git.wmi.amu.edu.pl/AITech/Szablon/raw/branch/master/Logotyp_AITech1.jpg)\n",
    "<div class=\"alert alert-block alert-info\">\n",
    "<h1> Ekstrakcja informacji </h1>\n",
    "<h2> 10. <i>CRF</i>  [ćwiczenia]</h2> \n",
    "<h3> Jakub Pokrywka (2021)</h3>\n",
    "</div>\n",
    "\n",
    "![Logo 2](https://git.wmi.amu.edu.pl/AITech/Szablon/raw/branch/master/Logotyp_AITech2.jpg)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Podejście softmax z embeddingami na przykładzie NER"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "scrolled": true
   },
   "source": [
    "https://pytorch-crf.readthedocs.io/en/stable/"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "https://www.aclweb.org/anthology/W03-0419.pdf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import gensim\n",
    "import torch\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "from datasets import load_dataset\n",
    "from torchtext.vocab import Vocab\n",
    "from collections import Counter\n",
    "\n",
    "from sklearn.datasets import fetch_20newsgroups\n",
    "# https://scikit-learn.org/0.19/datasets/twenty_newsgroups.html\n",
    "\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "from sklearn.metrics import accuracy_score\n",
    "\n",
    "from tqdm.notebook import tqdm\n",
    "\n",
    "import torch\n",
    "from torchcrf import CRF"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "dataset = load_dataset(\"conll2003\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def build_vocab(dataset):\n",
    "    counter = Counter()\n",
    "    for document in dataset:\n",
    "        counter.update(document)\n",
    "    return Vocab(counter, specials=['<unk>', '<pad>', '<bos>', '<eos>'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "vocab = build_vocab(dataset['train']['tokens'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "len(vocab.itos)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "vocab['on']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def data_process(dt):\n",
    "    return [ torch.tensor([vocab['<bos>']] +[vocab[token]  for token in  document ] + [vocab['<eos>']], dtype = torch.long) for document in dt]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def labels_process(dt):\n",
    "    return [ torch.tensor([0] + document + [0], dtype = torch.long) for document in dt]\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_tokens_ids = data_process(dataset['train']['tokens'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_tokens_ids = data_process(dataset['test']['tokens'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "validation_tokens_ids =  data_process(dataset['validation']['tokens'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "train_labels = labels_process(dataset['train']['ner_tags'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "validation_labels = labels_process(dataset['validation']['ner_tags'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "test_labels = labels_process(dataset['test']['ner_tags'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "train_tokens_ids[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def get_scores(y_true, y_pred):\n",
    "    acc_score = 0\n",
    "    tp = 0\n",
    "    fp = 0\n",
    "    selected_items = 0\n",
    "    relevant_items = 0 \n",
    "\n",
    "    for p,t in zip(y_pred, y_true):\n",
    "        if p == t:\n",
    "            acc_score +=1\n",
    "\n",
    "        if p > 0 and p == t:\n",
    "            tp +=1\n",
    "\n",
    "        if p > 0:\n",
    "            selected_items += 1\n",
    "\n",
    "        if t > 0 :\n",
    "            relevant_items +=1\n",
    "\n",
    "            \n",
    "            \n",
    "    if selected_items == 0:\n",
    "        precision = 1.0\n",
    "    else:\n",
    "        precision = tp / selected_items\n",
    "        \n",
    "            \n",
    "    if relevant_items == 0:\n",
    "        recall = 1.0\n",
    "    else:\n",
    "        recall = tp / relevant_items\n",
    "    \n",
    "    \n",
    "    if precision + recall == 0.0 :\n",
    "        f1 = 0.0\n",
    "    else:\n",
    "        f1 = 2* precision * recall  / (precision + recall)\n",
    "\n",
    "    return precision, recall, f1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "num_tags = max([max(x) for x in dataset['train']['ner_tags'] ]) + 1 "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "class FF(torch.nn.Module):\n",
    "\n",
    "    def __init__(self,):\n",
    "        super(FF, self).__init__()\n",
    "        self.emb = torch.nn.Embedding(23627,200)\n",
    "        self.fc1 = torch.nn.Linear(200,num_tags)\n",
    "       \n",
    "\n",
    "    def forward(self, x):\n",
    "        x = self.emb(x)\n",
    "        x = self.fc1(x)\n",
    "        return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "ff = FF()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "crf = CRF(num_tags)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "params = list(ff.parameters()) + list(crf.parameters())\n",
    "\n",
    "optimizer = torch.optim.Adam(params)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "def eval_model(dataset_tokens, dataset_labels):\n",
    "    Y_true = []\n",
    "    Y_pred = []\n",
    "    ff.eval()\n",
    "    crf.eval()\n",
    "    for i in tqdm(range(len(dataset_labels))):\n",
    "        batch_tokens = dataset_tokens[i]\n",
    "        tags = list(dataset_labels[i].numpy())\n",
    "        emissions = ff(batch_tokens).unsqueeze(1)\n",
    "        Y_pred += crf.decode(emissions)[0]\n",
    "        Y_true += tags\n",
    "\n",
    "    return get_scores(Y_true, Y_pred)\n",
    "        "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "NUM_EPOCHS = 4"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "for i in range(NUM_EPOCHS):\n",
    "    ff.train()\n",
    "    crf.train()\n",
    "    for i in tqdm(range(len(train_labels))):\n",
    "        batch_tokens = train_tokens_ids[i]\n",
    "        tags = train_labels[i].unsqueeze(1)\n",
    "        emissions = ff(batch_tokens).unsqueeze(1)\n",
    "\n",
    "        optimizer.zero_grad()\n",
    "        loss  = -crf(emissions,tags)\n",
    "        loss.backward()\n",
    "        optimizer.step()\n",
    "        \n",
    "    ff.eval()\n",
    "    crf.eval()\n",
    "    print(eval_model(validation_tokens_ids, validation_labels))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "eval_model(validation_tokens_ids, validation_labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "eval_model(test_tokens_ids, test_labels)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "len(train_tokens_ids)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Zadanie domowe\n",
    "\n",
    "- sklonować repozytorium https://git.wmi.amu.edu.pl/kubapok/en-ner-conll-2003\n",
    "- stworzyć klasyfikator bazujący na sieci neuronowej feed forward w pytorchu  + CRF (można bazować na tym jupyterze lub nie).\n",
    "- sieć feedforward powinna obejmować aktualne słowo, poprzednie i następne + dodatkowe cechy (np. długość wyrazu, czy wyraz zaczyna się od wielkiej litery, stemmming słowa, czy zawiera cyfrę)\n",
    "- stworzyć predykcje w plikach dev-0/out.tsv oraz test-A/out.tsv\n",
    "- wynik fscore sprawdzony za pomocą narzędzia geval (patrz poprzednie zadanie) powinien wynosić conajmniej 0.65\n",
    "- proszę umieścić predykcję oraz skrypty generujące (w postaci tekstowej a nie jupyter) w repo, a w MS TEAMS umieścić link do swojego repo\n",
    "termin 15.06, 60 punktów, za najlepszy wynik- 100 punktów\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "author": "Jakub Pokrywka",
  "email": "kubapok@wmi.amu.edu.pl",
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "lang": "pl",
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.3"
  },
  "subtitle": "10.CRF[ćwiczenia]",
  "title": "Ekstrakcja informacji",
  "year": "2021"
 },
 "nbformat": 4,
 "nbformat_minor": 4
}