Mixture of Distributions Matters: Dynamic Sparse Attention for Efficient Video Diffusion Transformers

Documenti analoghi

• RoPeSLR: 3D RoPE-driven Sparse-LowRank Attention for Efficient Diffusion Transformers
  di: Liu, Yuxi, et al.
  Pubblicazione: (2026)
• LVSA: Training-Free Sparse Attention for Long Video Diffusion
  di: Glorian, Gael, et al.
  Pubblicazione: (2026)
• Sparse Forcing: Native Trainable Sparse Attention for Real-time Autoregressive Diffusion Video Generation
  di: Xu, Boxun, et al.
  Pubblicazione: (2026)
• Sparse-vDiT: Unleashing the Power of Sparse Attention to Accelerate Video Diffusion Transformers
  di: Chen, Pengtao, et al.
  Pubblicazione: (2025)
• Sparse VideoGen: Accelerating Video Diffusion Transformers with Spatial-Temporal Sparsity
  di: Xi, Haocheng, et al.
  Pubblicazione: (2025)