Papers › Attention Sparsity is Input-Stable: Training-Free Sparse Attention for Video...

Attention Sparsity is Input-Stable: Training-Free Sparse Attention for Video Generation via Offline Sparsity Profiling and Online QK Co-Clustering

19 Mar 2026ICMLarXiv:2603.18636added by Syntology

Jiayi Luo, Jiayu Chen, Jiankun Wang, Cong Wang, Hanxin Zhu, Qingyun Sun, Chen Gao, Zhibo Chen, Jianxin Li

Title, abstract, authors and date from arXiv's metadata (CC0); this paper is not in the Papers with Code archive (frozen 2025-07-28).

Diffusion Transformers (DiTs) achieve strong video generation quality but suffer from high inference cost due to dense 3D attention, motivating sparse attention techniques for improving efficiency. However, existing training-free sparse attention methods for video generation still face two unresolved limitations: ignoring layer heterogeneity in attention pruning and ignoring query-key coupling in block partitioning, which hinder a better quality-speedup trade-off. In this work, we uncover a critical insight: attention sparsity is an intrinsic layer-wise property, with only minor variation across different inputs. Motivated by this observation, we propose SVOO, a training-free sparse attention framework for fast video generation via offline layer-wise sparsity profiling and online bidirectional co-clustering. Specifically, SVOO adopts a two-stage paradigm: (i) offline layer-wise sensitivity profiling to derive intrinsic per-layer pruning levels, and (ii) online block-wise sparse attention via a bidirectional co-clustering algorithm. Extensive experiments on seven widely used video generation models demonstrate that SVOO achieves a superior quality-speedup trade-off over state-of-the-art methods, delivering up to 1.93x speedup while maintaining a PSNR of up to 29 dB on Wan2.1. Code is available at: https://github.com/Mutual-Luo/SVOO.

PaperPDFCode Syntology ran

In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.

Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, on Syntology's MCP service (how to connect):

Code

Syntology 11 code samples harvested from 1 repository linked to this paper: 8 ran (of which 0 constructed an object rather than computing a result; 8 with no instrument failure: 0 honoured, 0 violated, 8 with no contract checked; 0 where Syntology's instrument failed) · 3 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code; here there are none.

By repository: found in paper text by Syntology: 11 samples from 1 repository, 8 ran (8 with no instrument failure; 0 where Syntology's instrument failed). The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

Levels written on 2026-09-25 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

Mutual-Luo/SVOO found in paper text by SyntologySyntology: 11 samples, 8 ran (0 instrument-failed) · 3 unverified report

Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

11 samples harvested: 8 ran (of which 0 constructed an object rather than computing a result; 8 with no instrument failure: 0 honoured, 0 violated, 8 with no contract checked; 0 where Syntology's instrument failed) · 3 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page last changed ; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

8ran
3unverified

Licence: 0 of the 11 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from Mutual-Luo/SVOO. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.

Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of six reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity.

Mutual-Luo/SVOO found in paper text by Syntology 11 samples, 8 ran (0 instrument-failed) · 3 unverified
generate_temporal_head_mask_mod svoo/models/hunyuan10/utils.py ran Apache-2.0 (permissive) · 680b148aa522eec6 · report fixture: signature context_length:int/0/int, prompt_length:int/0/int, num_frames:int/0/int, token_per_frame:int/0/int, mul:int/0/int · returned function · level recorded 2026-09-25
generate_temporal_head_mask_mod svoo/models/wan/utils.py ran Apache-2.0 (permissive) · f7ed76a6faefd4f2 · report fixture: signature context_length:int/0/int, prompt_length:int/0/int, num_frames:int/0/int, token_per_frame:int/0/int, mul:int/0/int · returned function · level recorded 2026-09-25
get_attention_mask in 11 papers svoo/models/hunyuan10/utils.py ran Apache-2.0 (permissive) · 1e76f50cbc64301f · report fixture: signature mask_name:str/0/str, sample_mse_max_row:int/0/int, context_length:int/0/int, num_frame:int/0/int, frame_size:int/0/int, device:str/0/str · returned Tensor · level recorded 2026-09-25
hunyuan_token_reorder_to_frame_major svoo/models/hunyuan10/placement.py ran Apache-2.0 (permissive) · 0543cdd0b9d08934 · report fixture: signature tensor:Tensor/4/float, fix_len:int/0/int, reorder_len:int/0/int, reorder_num_frame:int/0/int, frame_size:int/0/int · returned Tensor · level recorded 2026-09-25
hunyuan_token_reorder_to_token_major svoo/models/hunyuan10/placement.py ran Apache-2.0 (permissive) · 0847ff97c9658511 · report fixture: signature tensor:Tensor/4/float, fix_len:int/0/int, reorder_len:int/0/int, reorder_num_frame:int/0/int, frame_size:int/0/int · returned Tensor · level recorded 2026-09-25
sparsity_to_width svoo/models/wan/utils.py ran fingerprinted Apache-2.0 (permissive) · 793abbf309d92d09 · report fixture: signature sparsity:float/0/float, context_length:int/0/int, num_frame:int/0/int, frame_size:int/0/int · returned float · level recorded 2026-09-25
wan_token_reorder_to_frame_major svoo/models/wan/placement.py ran Apache-2.0 (permissive) · 2cbbc31b189869b2 · report fixture: signature tensor:Tensor/4/float, fix_len:int/0/int, reorder_len:int/0/int, reorder_num_frame:int/0/int, frame_size:int/0/int · returned Tensor · level recorded 2026-09-25
wan_token_reorder_to_token_major svoo/models/wan/placement.py ran Apache-2.0 (permissive) · 272c97b0d22d2805 · report fixture: signature tensor:Tensor/4/float, fix_len:int/0/int, reorder_len:int/0/int, reorder_num_frame:int/0/int, frame_size:int/0/int · returned Tensor · level recorded 2026-09-25
prepare_flashinfer_attention svoo/models/wan/attention.py unverified Apache-2.0 (permissive) · 2191dca04d00a3c3 · report
ref_hunyuan_sparse_head_placement svoo/models/hunyuan10/placement.py unverified Apache-2.0 (permissive) · c5941af3e2811f0d · report
ref_wan_sparse_head_placement svoo/models/wan/placement.py unverified Apache-2.0 (permissive) · 3c0438a386efef9e · report

Results from the paper

The Papers with Code archive ends with its 2025-07-28 snapshot. This paper's arXiv identifier, 2603.18636, was issued in March 2026, after that date, so the archive has no leaderboard rows for it.

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections