Smittie's Bernini Video Edit Workflow

Description

With this workflow you can edit videos using Bernini (1.3B or 14B) and Wan2.2 with better performance and much less memory usage than usual. However, it is quite tricky to use, so read all notes in order!

At least 16 GB of VRAM and 32 GB RAM recommended, but with more quantized versions of the Bernini GGUF models you can go even lower. This workflow was tested with 24 GB of VRAM and 64 GB of RAM.

#1 Video has removed her "Kick me" note at her back.

#2 Video has removed the picture in the background, via a tight video snippet. One can see the outlines of that edit.

#3 Video has removed the picture in the background, via a slightly bigger video snippet. The outline is now much cleaner, but still slightly visible.

Install Instructions

1. Install ffmpeg

2. Update ComfyUI to a version that contains the "Bernini Conditioning" node

3. Install all missing custom nodes via the ComfyUI Manager

4. Download all needed models

Acknowledgement

Most credits goes to the Bernini, Lightning, Wan, ffmpeg, ComfyUI and CivitAI team, as well as kijai, Kosinkadink, yolain, Dr.Lt.Data, robinjhuang. Thank you guys!