4M-21: An Any-to-Any Vision Model for Tens of Tasks and Modalities

Open in new window