Actor-agnostic Multi-label Action Recognition with Multi-modal Query

Open in new window