Advanced BERT Question Answering System

This repository presents an advanced Python implementation of a question-answering system leveraging the power of the BERT (Bidirectional Encoder Representations from Transformers) model, specifically bert-large-uncased-whole-word-masking-finetuned-squad, for providing context-based answers to user queries.

Key Features

• BERT Integration: Harnesses a pre-trained BERT model from Hugging Face's comprehensive transformers library.
• Contextual Understanding: Employs Whole Word Masking for a nuanced understanding of the context provided by the text.
• Dynamic Interaction: Users can interact with the system in real-time by posing questions about the loaded article content.
• End-of-Session Control: Implements a user-friendly mechanism to terminate the question-answering loop with a simple 'quit' command.

How It Works

• The script initializes a question-answering pipeline using the pipeline function from the transformers library.
• It reads the text content of an article file specified by the user, which serves as the context for answering questions.
• Through a function, it processes user-input questions and retrieves answers from the pre-trained BERT model based on the context of the loaded article.
• The interactive loop allows continuous querying until the user decides to exit.

Installation & Setup

1. Ensure you have Python 3.x installed.

2. Install the transformers library using pip:

   pip install transformers

Clone this repository to your local machine.

1. Place the text file with the article you want to analyze in the root directory of the project, named article_content.txt.

Usage

1. Execute the script from the command line. You will be prompted to enter a question. The script will then display the answer derived from the context of the article content. To end the session, type 'quit' when prompted for a question.

   python simpleai.py

Advanced Configuration

1. To use a different model or tokenizer, modify the pipeline instantiation in the script. The transformers library offers a wide array of pre-trained models that can be leveraged for various languages and domains.

Contributions

1. Contributions to this project are welcome. Please fork the repository, make your changes, and submit a pull request.