Hi,
In the technical report you mentioned you use DDIM inversion for video out painting. There are a few different papers to do DDIM inversion. Which way was tried?
Was this used - https://huggingface.co/docs/diffusers/v0.14.0/en/api/schedulers/ddim_inverse ?