Ref2va Usage Help: Achieving Stable Person Replacement in Video
PensionNew1814 · reddit · 2026-08-16
A Reddit user is seeking help with the video generation model Ref2va, specifically trying to perform person replacement using a reference image within a video. Despite knowing the basic tags like <subjects> and <pictures>, the user struggles to get consistent results, often seeing no change or only partial motion transfer. The post asks for advice on missing techniques or prompting strategies to achieve stable subject replacement.
Related event: Users Seek Workflow for Reference Image in MiniMax H3(4 posts)→
More from Multimodal
- FrankenSim builds a physics-based virtual orchestra, piano and drum demos live — doodlestein · 2026-10-16
- vLLM officially supports Muse Spark Open: multimodal, 1M context, agentic — vllm_project · 2026-10-16
- InfoVis: 9B unified model tops open-source slide generation and editing, trained on 13M records — Jiahao Wang · 2026-10-16
- RepForcing: feature-space supervision improves few-step autoregressive video distillation — zju · 2026-10-16
- SCOPE: semantic span-aware GRPO objective improves T2I prompt rewriting — university-of-rochester · 2026-10-16
- Orbia benchmark asks: do generated videos form a consistent 3D world? — tencent · 2026-10-16