Any reason for not using vae.std to generate RGB latent?

[Jerrypiglet]

Hi there, I notice the pipeline of Marigold is different from the original SD pipeline in that, when encoding RGB/depth into latents, instead of doing something like vae.encode(pixel_values).latent_dist.sample(), the implementation uses function def encode_rgb where only the mean of the vae distribution is used as latent (see here), without adding std * sample as in here. Is there a reason for this, to remove the randomness of the vae encoding?

[markkua]

Hi, as we are trying to do depth estimation, which is supposed to be deterministic unlike generation tasks. We intentionally did this to remove undesired randomness.