Troubleshooting Character Swaps in Minimax H3 Video Generation
GivePLZ-DoritosChip · reddit · 2026-08-13
A user details their struggles with performing character face swaps using Minimax H3 (ref2v model). Despite using multiple reference images, LLM-assisted prompting, and video references, the model consistently renders the original scene without changing the faces. The post asks the community for the correct prompting setup to achieve a successful character replacement.
Related event: MiniMax H3 Video Face-Swap Fails on Hands(3 posts)→
More from Multimodal
- XGRIDS builds world's largest 3DGS town: 23.5 km² and 6.2 billion Gaussian Splats — bilawalsidhu · 2026-10-16
- IMMERSION: AI film compilation links 12 shorts with one continuous camera move — AI_Andrew · 2026-10-16
- A song with vocals, lyrics and 3D head animation created entirely by Opus 5.5 — repligate · 2026-10-16
- Alibaba open-sources CommerceGen-Studio for commercial-grade e-commerce images — alibabagroup · 2026-10-16
- EVA-Cap: Fudan and OPPO align video captioning with atomic audiovisual events via GRPO — Fudan-University · 2026-10-16
- SCOPED: skill-decoupled expert distillation beats mixed-capability RL across 8 benchmarks — Yuqi Xiong · 2026-10-16