Shengshu launches ViduS2: 720P real-time avatars and streaming video editing
生数科技 · wechat · 2026-09-15
Shengshu Technology released ViduS2, two real-time video models: ViduS2-Avatar generates interactive digital characters at 720P (up from 540P) with mid-stream reference updates, and ViduS2-Editing performs live style transfer, virtual try-on, character and background replacement on video streams. Both extend to spatial video for VR headsets. Key technique: Self-Replay Forcing (SRF), training on the model's own generated trajectories to curb streaming error accumulation, plus a lightweight single-step refiner and VLM-agent visual feedback. ViduS2 tops StreamAV-Bench on all 9 metrics and beats offline baselines on OpenVE/RefVIE (4.26 vs 3.92). Available now with an arXiv technical report.
Related event: Shengshu releases Vidu S2: real-time avatars and streaming video editing(9 posts)→
More from Multimodal
- Pika's pika-h3-flex Space Trends on Hugging Face — Pika-Labs · 2026-11-02
- awa-video-reel: free Claude Code skill auto-edits shorts and records website B-roll — TawohAwa · 2026-11-02
- Hobbyist 3D-scans Yokohama's abandoned Negishi racecourse grandstand with 3DGS — janusch_patas · 2026-11-02
- Drone-captured 400-ft guyed tower rebuilt as 3DGS via Metashape and LichtFeld Studio — janusch_patas · 2026-11-02
- TBC launches brain-cell-inspired AI video model Mischief at $0.005 per second — danfaggella · 2026-11-02
- Even a skeptic almost fell for an AI fruit cartoon reel — menhguin · 2026-11-02