25 March 2022 3 7K Report

I am trying to implement a VQA model in ecommerce, and would love to have a dataset that focuses on fashion (or any ecommerce type of goods). If there isn't an available one, is synthetically generating q&a pairs for a given image a good idea? If so, any idea how to approach such problem ?

More Seif Maged's questions See All
Similar questions and discussions