SOURCE-LINKED INTELLIGENCE

SparkDiffusion: Mitigating the High-Sparsity Trap --- A Unified Framework for up to $265\times$ Single-GPU Acceleration of Visual Generation

arXiv · AI, language, vision and robotics · article · Sep 19, 2026 · UTC

Video diffusion transformers are expensive because attention dominates long spatiotemporal token sequences. We identify the \emph{high-sparsity trap}: at extreme attention sparsity, step-local training losses keep decreasing while terminal generation quality stagnates or degrades. The trap is one of supervision: the dominant terminal errors originate in the high-noise structure-generation stage, and terminal-aligned training corrects terminal errors that substantially extended step-local training cannot. This yields a simple staging principle: \emph{first adapt the sparse architecture into a c

Read original source ↗ Open in workspace

recordType

paper

region

Global