AnimateLCM-SVD-xt

AnimateLCM-SVD-xt is a consistency accelerated Stable Video Diffusion following the strategies proposed in AnimateLCM.

It can generate good image animations in 2~8 steps without requiring the classifier-free guidance (cfg).

Important Notes:

Default euler or heun might not generate good results, I have rewritten a noise scheduler to make it work. Need some extra time to integrate it into diffusers or ComfyUI.

For more details, please refer to https://github.com/G-U-N/AnimateLCM

Links

Contact: [email protected]

[Paper]

[Code]

[Demos]