/Arabic-Book-Summary-Generator

Primary LanguageJupyter Notebook

Arabic Book Summary Generator

Overview

This project is an Arabic Book Summary Generator that utilizes GPT-2 (Aragpt2) to generate concise summaries of Arabic books based on their descriptions. It also includes a Gradio interface that allows users to interact with the model and generate summaries in real-time.

Features

  • Scrapes book data and summaries from the Arab Books website.
  • Cleans and preprocesses Arabic text for training.
  • Trains a GPT-2 (Aragpt2) model on the scraped data to generate book summaries.
  • Provides an interactive Gradio interface for easy user interaction.
  • Supports generating summaries for new book descriptions in Arabic.

Project Workflow

  1. Data Collection:

    • Web scraping Arabic book data from Arab Books.
    • Cleaning and preprocessing the text to prepare it for training.

  2. Model Training:

    • Tokenizing and preparing the data for GPT-2 (Aragpt2) training.
    • Fine-tuning the GPT-2 model on the cleaned Arabic book summaries.

  3. Summary Generation:

    • Using the fine-tuned model to generate summaries based on user-provided book descriptions.

  4. Deployment:

    • Deploying a Gradio interface that allows users to input book descriptions and get automatically generated summaries.