NLA_CourseProject_VQA Paper Link: https://arxiv.org/pdf/1505.00468.pdf This repository is for VQA implementation and enhancement using Stacked Attention. Currently the code is optimised to work on multiple gpu's for training and takes 7 hours for 30 epochs.