H3mini R2V - Character Sheet vs Reference Image
retained text
Hi, I've recently started the journey into H3minimax, mainly for r2v. I'm using some workflows found online, basically just a sample video and a formatted prompt to use, replacing the subject details with similar appropriate descriptions. I'm finding the consistency in the replacements very hit-or-miss, and not to the same quality of the example outputs, or other similar work I see online. So far, I have been using 1 reference image , as per the workflow guide, but I'm wondering if using an additional reference image, or a character sheet would improve my results? Has anyone experienced this and found a solution?
document fetched 2026-10-14 00:32 · canonical https://reddit.com/r/StableDiffusion/comments/1x4mm4o/h3mini_r2v_character_sheet_vs_reference_image