We read every piece of feedback, and take your input very seriously.
To see all available qualifiers, see our documentation.
Classifies action-object pairs given a video sequence.