Generative Vision Atlas

Architecture option

UNet backbone

A convolutional encoder-decoder with skip connections; the dominant diffusion backbone before diffusion transformers.

Used by (1)

LDM / Stable Diffusion

Alternatives on this axis

DiT with diffusion head (DiT^DH), Diffusion Transformer (DiT), MM-DiT (dual-stream joint attention), Single-stream (unified-weight) DiT

← All Architecture options