Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Motion-aware Autoregressive Generation #49

Open
ZyuanW opened this issue Jan 17, 2025 · 3 comments
Open

Motion-aware Autoregressive Generation #49

ZyuanW opened this issue Jan 17, 2025 · 3 comments

Comments

@ZyuanW
Copy link

ZyuanW commented Jan 17, 2025

Hi, thanks for your work!
May i ask where in the code the sampling of M historical frames as input is implemented? Could you kindly point me to the relevant part of the codebase or provide a brief explanation?

@Jieqianyu
Copy link
Collaborator

Thank you for your interest! For more details, you can refer to the build_clips function in this file.

@ZyuanW
Copy link
Author

ZyuanW commented Jan 19, 2025

Thank you for your interest! For more details, you can refer to the build_clips function in this file.

and may i ask what's the main different between base dreamer and dreamforge? is the baseDreamer an autoregressive model? btw,what is the relationship between 'current video clip and "Motion frames"?

@Jieqianyu
Copy link
Collaborator

Thank you for your interest! The main difference between Base Dreamer and DreamForge is that the former is a single-frame version, while the latter is a video version. In DreamForge, motion frames are sampled from the previous clip. For more details, you can refer to the Arxiv paper.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants