Search papers, labs, and topics across Lattice.
The paper introduces DuSPiT, a novel Dual-branch Sub-Patch Pixel Diffusion Transformer that separates global structural reasoning from local appearance modeling in image generation. By employing a compact base branch for efficient global reasoning alongside a high-capacity pixel branch organized into subpatch groups, DuSPiT enhances the preservation of detailed appearance while maintaining effective communication between branches through cross-attention. Experimental results demonstrate that DuSPiT achieves superior image quality with richer details and a more favorable quality-efficiency trade-off compared to existing pixel-space diffusion transformers.
DuSPiT generates images with unprecedented detail and structure by decoupling global reasoning from local appearance in a dual-branch architecture.
Diffusion Transformers achieve strong image generation performance, but most operate in compressed latent spaces. Pixel-space diffusion avoids this information loss, yet existing approaches map each raw image patch to a single token, forcing one representation to handle both global communication and fine-grained details. We address this issue by proposing a new architecture, \textbf{DuSPiT}, a \textbf{Du}al-branch \textbf{S}ub\textbf{P}atch \textbf{Pi}xel \textbf{T}ransformer. This model separates global structural reasoning from local appearance modeling. DuSPiT uses a compact base branch for efficient global reasoning and a parallel, high-capacity pixel branch, organized into subpatch groups, to preserve detailed appearance, with the two branches interacting through cross-attention. Our results show that DuSPiT generates images with richer details and stronger fine-grained structures, while also achieving a better quality--efficiency trade-off than prior pixel-space diffusion transformers.
