Hi, thanks for sharing this great work!
I would like to know how I can train (or obtain) the VAE model only with your codebase.
Is there a configuration available for VAE-only training?
Or is there a pre-trained VAE model that can be used directly (e.g., for encoding single- or multi-view images into latents and decoding them back, without latent diffusion)?
Thanks a lot!
Hi, thanks for sharing this great work!
I would like to know how I can train (or obtain) the VAE model only with your codebase.
Is there a configuration available for VAE-only training?
Or is there a pre-trained VAE model that can be used directly (e.g., for encoding single- or multi-view images into latents and decoding them back, without latent diffusion)?
Thanks a lot!