Papers › Memorize When Needed: Decoupled Memory Control for Spatially Consistent Long-Horizon...

Memorize When Needed: Decoupled Memory Control for Spatially Consistent Long-Horizon Video Generation

20 Apr 2026arXiv:2604.18215added by Syntology

Yanjun Guo, Zhengqiang Zhang, Pengfei Wang, Xinyue Liang, Zhiyuan Ma, Lei Zhang

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

Spatially consistent long-horizon video generation aims to maintain temporal and spatial consistency along predefined camera trajectories. Existing methods mostly entangle memory modeling with video generation, leading to inconsistent content during scene revisits and diminished generative capacity when exploring novel regions, even trained on extensive annotated data. To address these limitations, we propose a decoupled framework that separates memory conditioning from generation. Our approach significantly reduces training costs while simultaneously enhancing spatial consistency and preserving the generative capacity for novel scene exploration. Specifically, we employ a lightweight, independent memory branch to learn precise spatial consistency from historical observation. We first introduce a hybrid memory representation to capture complementary temporal and spatial cues from generated frames, then leverage a per-frame cross-attention mechanism to ensure each frame is conditioned exclusively on the most spatially relevant historical information, which is injected into the generative model to ensure spatial consistency. When generating new scenes, a camera-aware gating mechanism is proposed to mediate the interaction between memory and generation modules, enabling memory conditioning only when meaningful historical references exist. Compared with the existing method, our method is highly data-efficient, yet the experiments demonstrate that our approach achieves state-of-the-art performance in terms of both visual quality and spatial consistency.

PaperPDFCode Syntology ran

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

For agents, Syntology's MCP tool lists every function and class Syntology harvested from this paper and whether it ran (how to connect): get_harvested_code_for_paper(arxiv_id="2604.18215")

Code

Syntology Ran 3 of 19 code samples harvested from 1 repository linked to this paper; 16 have no recorded run. Of those that ran: 1 ran · violated contract; 1 ran · our draft was wrong; 1 ran · fixture could not drive it.

By repository: found in paper text by Syntology: 19 samples from 1 repository, 3 ran. The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

iguoyanjun/Memorize-When-Needed found in paper text by Syntology 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

19 samples harvested; 3 ran; 0 honoured the contract we drafted; 16 have no recorded run. Read from Syntology's graph 2026-09-24; that is when this build read the record, not when the samples ran.

1ran · violated contract
1ran · our draft was wrong
1ran · fixture could not drive it
16unverified

Licence: 19 of the 19 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 iguoyanjun/Memorize-When-Needed. “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. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself; “our draft was wrong” and “fixture could not drive it” are failures of Syntology's instrument, not of the code.

Each sample ends with its code_sha256, Syntology's identity for that exact code. An agent fetches the stored sample with Syntology's MCP tool get_code(code_sha256="…") (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 from a repository marked official are listed first. 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 four 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; or the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar). 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.

retrieve_timesteps iguoyanjun/Memorize-When-Needed/pipeline.py found in paper text by Syntology ran · our draft was wrong no licence file found · pointer only · 22b1f260da28f6a5 · report
rope_params iguoyanjun/Memorize-When-Needed/models/controlnets/wan_transformer_dual.py found in paper text by Syntology ran · violated contract fingerprinted no licence file found · pointer only · dc160d63c5846f4a · report
sinusoidal_embedding_1d iguoyanjun/Memorize-When-Needed/models/controlnets/wan_transformer_dual.py found in paper text by Syntology ran · fixture could not drive it fingerprinted no licence file found · pointer only · e34a97853e924d88 · report
attention iguoyanjun/Memorize-When-Needed/models/wan_modules/attention_utils.py found in paper text by Syntology unverified no licence file found · pointer only · c0487d5566533bde · report
build_controlnet_token_mask iguoyanjun/Memorize-When-Needed/pipeline.py found in paper text by Syntology unverified no licence file found · pointer only · 8b0ccead2e4ce95f · report
build_reference_frames iguoyanjun/Memorize-When-Needed/infer.py found in paper text by Syntology unverified no licence file found · pointer only · 798aecf07ae276dd · report
color_transfer iguoyanjun/Memorize-When-Needed/models/utils/utils.py found in paper text by Syntology unverified no licence file found · pointer only · 06338ab44c6754a4 · report
filter_kwargs iguoyanjun/Memorize-When-Needed/infer.py found in paper text by Syntology unverified no licence file found · pointer only · 92f121bb438f3774 · report
filter_kwargs iguoyanjun/Memorize-When-Needed/models/utils/utils.py found in paper text by Syntology unverified no licence file found · pointer only · fbc3e873361f20d5 · report
flash_attention iguoyanjun/Memorize-When-Needed/models/wan_modules/attention_utils.py found in paper text by Syntology unverified no licence file found · pointer only · 7ea44a4680b76539 · report
get_sampling_sigmas iguoyanjun/Memorize-When-Needed/models/utils/fm_solvers.py found in paper text by Syntology unverified no licence file found · pointer only · 1532dbb1d28a2c99 · report
get_width_and_height_from_image_and_base_resolution iguoyanjun/Memorize-When-Needed/models/utils/utils.py found in paper text by Syntology unverified no licence file found · pointer only · 406364af353a424a · report
mat3_to_quat_wxyz iguoyanjun/Memorize-When-Needed/dataload/utils.py found in paper text by Syntology unverified no licence file found · pointer only · 5485127438111d03 · report
orthonormalize_R iguoyanjun/Memorize-When-Needed/dataload/utils.py found in paper text by Syntology unverified no licence file found · pointer only · 3493f2b0b7ba5d3b · report
quat_wxyz_to_mat3 iguoyanjun/Memorize-When-Needed/dataload/utils.py found in paper text by Syntology unverified no licence file found · pointer only · 2001736e8a6a9441 · report
ray_condition iguoyanjun/Memorize-When-Needed/models/controlnets/wan_controlnet_camera.py found in paper text by Syntology unverified no licence file found · pointer only · 6e301fe216cd0f77 · report
resize_mask iguoyanjun/Memorize-When-Needed/pipeline.py found in paper text by Syntology unverified no licence file found · pointer only · ddf3a7a391972394 · report
retrieve_timesteps iguoyanjun/Memorize-When-Needed/models/utils/fm_solvers.py found in paper text by Syntology unverified no licence file found · pointer only · 3eeae6ab9b822c6c · report
zero_module iguoyanjun/Memorize-When-Needed/models/controlnets/wan_controlnet_camera.py found in paper text by Syntology unverified no licence file found · pointer only · d81381b327c76edd · report

Results from the paper

The Papers with Code archive ends with its 2025-07-28 snapshot. This paper's arXiv identifier, 2604.18215, was issued in April 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