arXiv Computer Vision

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

arXiv Computer Vision
Oct 16

Sparse Attention for Video Generation Acceleration via Growing Sparsity and Reduced Search

The paper introduces Growing Sparsity and Reduced Search (GSRS), a training‑free framework that accelerates Diffusion Transformers for video generation by progressively increasing sparsity and reusing early attention blocks to limit search space. GSRS uses Top‑p selection for adaptive sparsity and a precise evaluation scheme that computes exact attention weights with negligible overhead. Experiments on HunyuanVideo and Wan 2.1 show 1.55×–2.21× speedups while preserving generation quality comparable to full attention.

By Yun Dai, Yuyun Chen, Jiarui Wen, Huiping Zhuang, Cen Chen, Ziqian Zeng
arXiv Computer Vision
Nov 9

TriTD: Tri-Partite Trajectory-Distribution Distillation for Real-Time Autoregressive Video Generation

TriTD introduces a tri-partite distillation framework for autoregressive video diffusion, addressing three error sources—base model error, train‑inference distribution mismatch, and upstream error propagation. It combines Next‑Node Trajectory Distillation for efficient ODE initialization, Chunk‑Graded Noising to reduce noise mismatch and eliminate redundant KV‑cache updates, and Orthogonal‑Parallel DMD to enforce distribution matching while suppressing off‑axis drift. Experiments demonstrate that TriTD outperforms existing baselines in visual quality and achieves real‑time throughput of 26.8 FPS on a single H100 GPU.

By Jia Li, Xurui Peng, Haowei Zhu, Tianyu Zhao, Jiexi Wang, Youwei Zheng, Yuxi Ren, Xinglong Wu, Xing Wang, Hayden Kwok-Hay So
arXiv Computer Vision
Oct 19

Sink-Sparse Hybrid Attention for Efficient Inference in T2I/T2V Diffusion Transformers

Sink‑Sparse Hybrid Attention is a train‑free method that replaces dense self‑attention in Diffusion Transformers with reusable static sink masks and a lightweight pooled‑KV branch, keeping model weights and sampling schedules unchanged. It identifies reusable high‑mass structures from a small offline calibration set and applies them across future generations. Experiments on Wan2.1‑T2V‑1.3B and Z‑Image‑Turbo show 1.54× and 1.57× end‑to‑end speedups while preserving generation quality.

By Bingxin Xing, Yu Zhang, Jiatong Ding, Lei Lu, Ping Wang
arXiv Computer Vision
Oct 21

PISA: Piecewise Sparse Attention Is Wiser for Efficient Diffusion Transformers

PISA introduces a training‑free Piecewise Sparse Attention mechanism that preserves full attention span while reducing complexity. It keeps exact computation for critical blocks and approximates non‑critical ones using block‑wise Taylor expansion, avoiding the quality loss seen in conventional sparse attention. Experiments show significant speedups—up to 2.57× on Hunyuan‑Video—while maintaining or improving visual quality across video and image generation tasks.

By Haopeng Li, Shitong Shao, Wenliang Zhong, Zikai Zhou, Lichen Bai, Hui Xiong, Zeke Xie
arXiv AI
Oct 1

In-Flight KV Cache with Clean Anchors for Faster Autoregressive Video Diffusion

arXiv:2609.32540v2 Announce Type: replace-cross Abstract: Few-step autoregressive video diffusion generates a long video by splitting the video into temporal chunks and generating chunk-by-chunk, eac...

By Yikai Wang, Xiao Han, Mengmeng Xu, Juan Camilo Perez, Yiannis Douratsos, Sen He, Zijian Zhou, Fei Zhang, Zhaochong An, Juan-Manuel Perez-Rua, Chen Change Loy, Tao Xiang