Abstractive Text Summarization for Tamil Language Using m-T5

Summarization is the act of reducing a long form of text to a shorter version by lowering the length of the original text while retaining and keeping the summary’s core and the informational parts which conveys the content’s meaning. Summarizing lengthy papers by hand is difficult and prone to inaccuracy. The exponential rise of information and data has made automatic text summarization crucial. It leads to a summary generated by a summarizing system that enables users to understand the document’s contents. Abstractive and extractive summarization are the two categories into which summarization falls. We can see that there are not many automatic text summarizers available for Indian languages. Our main goal in this direction is to create an automatic abstractive text summarizer for Tamil using the m-T5 transformer model. We have also used manual dataset of Tamil Kid’s stories for training and testing.