New Style Transfer LoRA for MiniMax H3 Restyles Entire Videos From One Reference Image
PurzBeats · x · 2026-09-13
Developer Alissonerdx released an open-source style transfer LoRA for MiniMax H3 (ref2va) that restyles an entire video from a single reference image. It ships on Hugging Face with a ComfyUI workflow, Rank 64, under Apache-2.0.
Two usage modes:
- Drop in a style image and point at it with <Picture 1>;
- Skip the image and just name a style the base model already knows.
- Don't combine both: the model follows the text and ignores the image.
How it works & caveats:
- The source video conditions on guide latents, locking motion frame-for-frame; the style image rides the reference slot, enabling transfer of looks you can't name.
- A reference containing a prominent person can leak their identity into the video.
The repo also includes a minimaxh3lmsv1.0r64 sharpening LoRA that boosts photorealistic sharpness, with workflow files and multiple before/after example videos.
More from Multimodal
- Element Forge: open-source editable AI VFX for real footage in ComfyUI — zitroffell · 2026-10-30
- Claymation body horror on 8GB: three wrong settings, not the GPU, made local renders slow — roro9- · 2026-10-30
- AI-generated Haribo Starmix bean bag product shot looks like a real catalog page — Delahuntagram · 2026-10-30
- MiniMax LoRA creator ships open-source ComfyUI alternative to Higgsfield's Influencer Studio — infroy28 · 2026-10-30
- Scenario open-sources 3D OS: 80 agent skills taking sketches to game-ready assets — Scobleizer · 2026-10-30
- Open-source ComfyUI app reshoots footage keeping camera and framing, swapping scenes — 3dkvisual · 2026-10-30