GROOT-2: Weakly Supervised Multi-Modal Instruction Following Agents

Open in new window