facebookresearch / jepa

PyTorch code and models for V-JEPA self-supervised learning from video.
Other
2.68k stars 254 forks source link

Question: no release of decoder checkpoint? #37

Open LumenYoung opened 8 months ago

LumenYoung commented 8 months ago

Hi, thanks for the great work.

I checked the downloaded checkpoint and find that there is no decoder model checkpoint mentioned in the paper that was used for mapping the feature back to the .

I would like to ask if you have any plan to release that checkpoint? Cause it is also the reason why jepa is powerful for the further researching in the downstream area.

Thanks!

betterze commented 8 months ago

I have the same request. Could you share the diffusion model checkpoint? thx a lot.