Fizgig 6.0 - RefMod Training, Editing and Exploring
retained text
Fizgig now makes RefMods. Plain encode like the pack's extractor, or tuned against H3 itself. (This gives you the same likeness on shots like your photos, better generalisation on ones they never showed, cleaner/sharper results). Works for ref2va and fl2va, clips as motion, style presets, and a RefMod Studio tab to A/B a mod against the base before ComfyUI. Fully compatable with existing ComfyUI nodes Free, open source. [https://github.com/shootthesound/Fizgig/releases/tag/v6.0.1](https://github.com/shootthesound/Fizgig/releases/tag/v6.0.1)
document fetched 2026-09-17 00:32 · canonical https://reddit.com/r/StableDiffusion/comments/1wi4y9c/fizgig_60_refmod_training_editing_and_exploring