DiTango: Cost-Effective Parallel Diffusion Generation with Selective Attention State Reuse
Researchers introduce DiTango, a parallel framework for Diffusion Transformers that selectively reuses attention states to reduce communication overhead in multi-node environments. DiTango achieves up to 1.9x end-to-end and 3.2x attention speedup, with near-linear scaling, while maintaining generation quality comparable to state-of-the-art methods. The framework uses an anchor-guided state selection planner and a runtime for efficient state-centric operations.
Why it matters: DiTango addresses a key scalability bottleneck in diffusion model inference, enabling faster and more cost-effective high-resolution content generation in distributed settings.
Full story at: arXiv Computer Vision ↗