Closed OswaldoBornemann closed 7 months ago
I think the DAE[0] is related to the diffusion of shape latent, and the DAE[1] is related to the diffusion of latent point, which should be the local prior. So why the comment in the code which points to the DAE[1] is global prior?
DAE[0]
shape latent
DAE[1]
latent point
global prior
https://github.com/nv-tlabs/LION/blob/b16212120e90742c62dbb8151560a56bbff8e53b/trainers/train_2prior.py#L422
good catch! You are right. I fix the comment.
I think the
DAE[0]
is related to the diffusion ofshape latent
, and theDAE[1]
is related to the diffusion oflatent point
, which should be the local prior. So why the comment in the code which points to theDAE[1]
isglobal prior
?https://github.com/nv-tlabs/LION/blob/b16212120e90742c62dbb8151560a56bbff8e53b/trainers/train_2prior.py#L422