Generative Vision Atlas

Representation option

Deep-compression latent

A latent pushed to much higher spatial compression ratios (up to 128x) than standard VAEs, via residual autoencoding and staged high-resolution adaptation.

Introduced by

Used by (1)

DC-AE

Alternatives on this axis

Continuous tokens, Frozen encoder + trained decoder, Hybrid semantic + detail codebook, Multi-scale (next-scale) tokens, Pixels, Semantic / foundation latent, VAE latent

← All Representation options