The following project aims to fine-tune and implement the Portuguese Vocabulary Pretrained T5 model on abstractive text summarization tasks in Brazilian Portuguese
By finetuning Google's T5 model for abstractive text summarization over the XSum Dataset, as well as its Portuguese trained counterpart PT-T5, this project explores the abstractive summarization field with Deep Learning for both English and Portuguese languages.
Please refer to the adjacent PDF paper for further explanation on the project's development and results.
This repository was developed as the final project of the IA376's graduate course, taught by Professors Rodrigo Nogueira and Roberto Lotufo at the University of Campinas (UNICAMP).