Text Summarization Projects

This repository contains two text summarization projects built using natural language processing (NLP) models. The projects are implemented using Flask for the backend and HTML/CSS/JavaScript for the frontend.

---

Projects Overview

1. Ensemble Summarization

This project uses an ensemble of two pre-trained models, BART and GPT-2, to generate summaries of input text. The ensemble method combines the outputs of both models and selects the most frequent summary as the final result.

• Models Used:

  • BART (Bidirectional and Auto-Regressive Transformers): A transformer-based model fine-tuned for summarization tasks.
  • GPT-2 (Generative Pre-trained Transformer 2): A transformer-based model capable of generating coherent and contextually relevant text.

• Features:

  • Input text is summarized using both models.
  • The final summary is selected based on the most frequent output.
  • Interactive web interface for input and output.

---

2. T5 Summarization

This project uses the T5 (Text-To-Text Transfer Transformer) model for text summarization. T5 is a versatile transformer model that treats all NLP tasks as a text-to-text problem, making it highly effective for summarization.

• Model Used:

  • T5 (Text-To-Text Transfer Transformer): A transformer-based model fine-tuned for summarization tasks.

• Features:

  • Input text is summarized using the T5 model.
  • Interactive web interface for input and output.

---

Technologies Used

• Backend:

  • Python
  • Flask (Web framework)
  • Hugging Face Transformers (for pre-trained models)
  • PyTorch (for model inference)

• Frontend:

  • HTML
  • CSS
  • JavaScript

• Pre-trained Models:

  • BART (facebook/bart-large-cnn)
  • GPT-2 (gpt2)
  • T5 (t5-base)

---

Application Interface

T5 Summarization

Ensemble Summarization