EXPLORE CONNECTIONS
Patch-as-Decodable-Token: Towards Unified Multi-Modal Vision Tasks in MLLMs
Follow the relationships that help you find your next source.
Connections are not prepared for this record yet, or its source metadata has changed. The original record remains available while background snapshots are rebuilt.