LIFT paper on video generation under large viewpoint changes
The paper’s title expands LIFT to “Layout-In-Future” and names on-policy self-distillation as its method.
TLDR
A post shares the LIFT paper page and a video link. The paper’s title describes video generation under large viewpoint changes via on-policy self-distillation.
Combined views
2.5K
1 Source, first seen ago
