Thank you for sharing your wonderful ideas!
But I have some question.
- I want to know the type of all edges used in the commonsense graph.
- Do you use the same commonsense graph for all training data? Or do you generate one common sense graph in advance and extract only common sense parts related to the objects in the image?
I will wait for your reply :)
Thank You!
Thank you for sharing your wonderful ideas!
But I have some question.
I will wait for your reply :)
Thank You!