Toward Generalizable Prompt Learning via Multi-Reg

Following 12 feeds