Frame to Frame Controlnet sampling?

volotat / SD-CN-Animation

This script allows to automate video stylization task using StableDiffusion and ControlNet.

MIT License

807 stars 61 forks source link

Frame to Frame Controlnet sampling? #179

Closed Duemellon closed 10 months ago

Duemellon commented 11 months ago

if I was to enable ControlNet & have it use OpenPose preprocessor it will ask for an example image. Is there a way to get each frame of a V2V (vid2vid) to feed the preprocessor instead?

Or, depthmap, or canny, etc.

What about using the Batch function?

Duemellon commented 11 months ago

It was clarified that if you don't include a reference it will sample each frame as it comes up in the v2v. Is there a way to have a frame-by-frame in the CN for a t2v?