← sources

You do need character swap LoRA for MiniMax H3

source: reddit (reddit:StableDiffusion)published 2026-09-29 06:29ingested 2026-09-30 02:41

retained text

... if you want to use the hybrid model.

Ref2VA has certain quality issues and struggles with mid‑distance subjects and colors, it's a known issue the Minimax team has acknowledged.  [Hybrid](https://huggingface.co/smhfacct/Minimax-H3-fl2va-ref2va-hybrid-models) gives you FL2VA's quality with Ref2VA's flexibility. But some capabilities take a hit, video referencing is one. The [character swap](https://huggingface.co/akatz-ai/MiniMax-H3-Character-Swap-LoRA) LoRA brings it back and makes it even better. Even Ref2VA strongly benefits from it according to my tests.

At 88% denoise plus LoRA, the accuracy is even better but it may not be suitable on all scenarios. It depends on the video.

I'm including more tests in the comments.

document fetched 2026-09-30 00:35 · canonical https://reddit.com/r/StableDiffusion/comments/1wt2pd3/you_do_need_character_swap_lora_for_minimax_h3