Tutorial on Text Classification (NLP) using ULMFiT and fastai Library in Python.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Implementation in Python "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "!pip install torch_nightly -f https://download.pytorch.org/whl/nightly/cu92/torch_nightly.html\n",
    "!pip install fastai"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# import libraries\n",
    "import fastai\n",
    "from fastai import *\n",
    "from fastai.text import * \n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "from functools import partial\n",
    "import io\n",
    "import os"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.datasets import fetch_20newsgroups\n",
    "dataset = fetch_20newsgroups(shuffle=True, random_state=1, remove=('headers', 'footers', 'quotes'))\n",
    "documents = dataset.data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.DataFrame({'label':dataset.target, 'text':dataset.data})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = df[df['label'].isin([1,10])]\n",
    "df = df.reset_index(drop = True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['label'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df['text'] = df['text'].str.replace(\"[^a-zA-Z]\", \" \")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import nltk\n",
    "nltk.download('stopwords')\n",
    "\n",
    "from nltk.corpus import stopwords \n",
    "stop_words = stopwords.words('english')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# tokenization \n",
    "tokenized_doc = df['text'].apply(lambda x: x.split())\n",
    "\n",
    "# remove stop-words \n",
    "tokenized_doc = tokenized_doc.apply(lambda x: [item for item in x if item not in stop_words])\n",
    "\n",
    "# de-tokenization \n",
    "detokenized_doc = [] \n",
    "for i in range(len(df)): \n",
    "    t = ' '.join(tokenized_doc[i]) \n",
    "    detokenized_doc.append(t) \n",
    "\n",
    "df['text'] = detokenized_doc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "\n",
    "# split data into training and validation set\n",
    "df_trn, df_val = train_test_split(df, stratify = df['label'], test_size = 0.4, random_state = 12)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_trn.shape, df_val.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Language model data\n",
    "data_lm = TextLMDataBunch.from_df(train_df = df_trn, valid_df = df_val, path = \"\")\n",
    "\n",
    "# Classifier model data\n",
    "data_clas = TextClasDataBunch.from_df(path = \"\", train_df = df_trn, valid_df = df_val, vocab=data_lm.train_ds.vocab, bs=32)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "learn = language_model_learner(data_lm, pretrained_model=URLs.WT103, drop_mult=0.7)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# train the learner object with learning rate = 1e-2\n",
    "learn.fit_one_cycle(1, 1e-2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "learn.save_encoder('ft_enc')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "learn = text_classifier_learner(data_clas, drop_mult=0.7)\n",
    "learn.load_encoder('ft_enc')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "learn.fit_one_cycle(1, 1e-2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# get predictions\n",
    "preds, targets = learn.get_preds()\n",
    "\n",
    "predictions = np.argmax(preds, axis = 1)\n",
    "pd.crosstab(predictions, targets)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}