MultiBanAbs: A Comprehensive Multi-Domain Bangla Abstractive Text Summarization Dataset
Researchers created MultiBanAbs, a large dataset of 54,620 Bangla articles and summaries from diverse sources like news, business reports, and blogs. This dataset captures varied writing styles. Evaluation showed that the BanglaT5-small model significantly outperformed other models like LSTM and mT5-small in generating high-quality abstractive summaries for Bengali text.