xingyi-li / 3d-cinemagraphy Goto Github PK

Excellent Projects！！
May I ask if you can provide the reproduction code of the results in the project page，
in this way, I can avoid manually annotating the target regions (masks) and desired movement directions (hints)
Thank you for your reply！

Hi @xingyi-li,

Thanks for the impressive work and the released code.

I have two questions regarding the codes:

• 1. Is it possible to release the training code of the proposed method?
• 2. For the baseline method[19: Animating pictures with eulerian motion fields]. It seems the author didn't release the code as well, how to get the corresponding results of [19]?

Best regards and thanks in advance,

Hi, thanks for this amazing work and I really can't wait to try this out!! Is there an estimated date for the code release? Thanks a lot :)

Hi, thanks for your wonderful work. I wondered if I can just inference from only img? not use labelme to specify the hint and mask.
Or, beacuse I use the no-GUI Linux, it can not direct use labelme on Linux, is there any other way to label?
Thanks!

Hi,

I recentrly tried to run your code but I'm facing an issue when it comes to load weights for SpaceTimeAnimationModel
The line torch.load(ckpts/model_150000.pth) (l.147 in lib/model/inpaint/models.py) shows only the following keys :

['netG', 'netD', 'optimG', 'optimD', 'schedG', 'schedD']

But there is no netG network defined in this class ? Any clues ?
Thanks a lot for your time & work !