search

dvlab-research / Video-P2P

Video-P2P: Video Editing with Cross-attention Control
https://video-p2p.github.io/
386 stars 25 forks source link

Qs about the paper #16

Closed zws98 closed 4 months ago

zws98 commented 4 months ago

Thanks for your excellent job! I'm wondering the difference between initialized unconditional embedding and optimized unconditional embedding. The latter is optimized by ddim inversion, but there is no explanation for the generation of initialized one.

ShaoTengLiu commented 4 months ago

Thanks for your interest. The initialized embedding refers to the text embedding for an empty prompt.