This project presents a hybrid text summarization system that combines extractive and abstractive approaches to improve the quality of generated summaries. It integrates the TextRank algorithm for selecting important sentences and the Flan-T5 transformer model for generating coherent and context-aware summaries. The system is implemented as a Flask-based web application that accepts DOCX files and produces concise summaries evaluated using ROUGE metrics. Experimental results show that the hybrid approach outperforms standalone methods in terms of accuracy and relevance. Additionally, MLOps tools like Prometheus are used for real-time monitoring of system performance, making the solution suitable for practical deployment.
Patil et al. (Thu,) studied this question.