Regarding generation of Relation Conversation.

[KanghoonYoon]

To generate the dataset with predicates, which are represented as predicate, we need to parse the predicate from long sentences (i.e., captions). How did you parse these? did human manually parse, or use a rule-based parser?

[Weiyun1025]

Sorry for the late response. We parse it with a rule-based parser.

You can refer to this script or this script for more details about the parser.