每日AIGC最新进展(83)：华为提出频驱动的通用肖像视频编辑RASA、香港科技大学提出长视频生成Multi-Agent、北京大学提出高效视频到音频生成模型TA-V2A

Diffusion Models专栏文章汇总：入门与实战

目录

RASA: Replace Anyone, Say Anything -- A Training-Free Framework for Audio-Driven and Universal Portrait Video Editing

Long-Video Audio Synthesis with Multi-Agent Collaboration

TA-V2A: Textually Assisted Video-to-Audio Generation

RASA: Replace Anyone, Say Anything -- A Training-Free Framework for Audio-Driven and Universal Portrait Video Editing