百度上线蒸汽机2.0视频生成大模型，实现多人有声视频一体化

Core Insights - Baidu's MuseSteamer audio-video integration model has been upgraded to enable multi-person audio-video generation [1] - The MuseSteamer is an I2V model for Chinese audio-video integration, utilizing multi-modal latent space planning technology to autonomously coordinate multiple roles, emotions, and interaction logic [1] - This series of large models has been applied in various scenarios including Baidu's search and marketing [1]