多模态模型

“Unified-IO 2: Scaling Autoregressive Multimodal Models with Vision, Language, Audio, and Action” Unified-IO 2号称是第一个能够理解和生成图像、文本、音频和动...