Vidu S2: Real-Time Interactive, Editable, and Spatial Video Generation
Vidu S2 claims real-time 720p avatar generation and live video-stream editing.
The paper describes two models: S2-Avatar for interactive digital characters and S2-Editing for real-time edits to incoming video. It says avatar references can be changed at any moment, with stronger instruction following, including actions like dancing. The editing model covers style rendering, clothing, character, and background replacement. The authors also test real-time spatial video generation and report that Vidu S2 beats their baselines. HF Daily Papers' note
The paper describes two models: S2-Avatar for interactive digital characters and S2-Editing for real-time edits to incoming video. It says avatar references can be changed at any moment, with stronger instruction following, including actions like dancing. The editing model covers style rendering, clothing, character, and background replacement. The authors also test real-time spatial video generation and report that Vidu S2 beats their baselines. HF Daily Papers' note
score 7