AI Paper Atlas

Research map / Video / Video generation

QuantSparse: Comprehensively Compressing Video Diffusion Transformer with Model Quantization and Attention Sparsification

Weilun Feng, Chuanguang Yang, Haotong Qin, Mingqiang Wu, Yuqi Li, Xiangqi Li, Zhulin An, Libo Huang, Yulun Zhang, Michele Magno, Yongjun Xu

ICLR 2026 · Video › Video generation · approach: Efficient video diffusion models

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add global attention distillation (ablation)QuantSparse · Wan2.1-14B W4A8PSNR: 14.35 -> 16.01Improved
add local salient distillation (ablation)QuantSparse · Wan2.1-14B W4A8PSNR: 14.35 -> 16.82Improved
combine global and local distillation (MSAD) (ablation)QuantSparse · Wan2.1-14B W4A8PSNR: 14.35 -> 18.72Improved
select top-k salient queries instead of random queries for local distillation (ablation)QuantSparse · Wan2.1-14B W4A8PSNR: 15.49 -> 16.82Improved
add first-order residual reparameterization (ablation)QuantSparse · Wan2.1-14B W4A8PSNR: 14.16 -> 17.08Improved
use second-order residual instead of first-order (ablation)QuantSparse · Wan2.1-14B W4A8VQA: 70.82 -> 89.73Improved
add SVD projection to second-order residual (SSAR) (ablation)QuantSparse · Wan2.1-14B W4A8VQA: 89.73 -> 91.98Improved
shorten cache refresh interval from 5 to 3 (ablation)QuantSparse · Wan2.1-14B W4A8PSNR: 18.22 -> 18.86Mixed
lengthen cache refresh interval from 5 to 6 (ablation)QuantSparse · Wan2.1-14B W4A8PSNR: 18.22 -> 17.72Hurt
naively combine quantization with sparse attention (Q-VDiT+SVG)Q-VDiT · HunyuanVideo W4A8 25% densityPSNR: 16.85 -> 16.71No clear effect
combine quantization and sparsification (attention shift) (ablation)QuaRot + SVG · Wan2.1-1.3B W4A8attention shift (MSE): 0.216 and 0.134 separately -> 0.685 combinedHurt
lower attention density from 15% to 10% (ablation)QuantSparse · Wan2.1-14B W4A8PSNR: 18.22 -> 17.73Mixed

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy