Papers › Multi-Scale Vision Longformer: A New Vision Transformer for High-Resolution Image Encoding

Multi-Scale Vision Longformer: A New Vision Transformer for High-Resolution Image Encoding

29 Mar 2021ICCV 2021 10arXiv:2103.15358archive 2025-07-28

Pengchuan Zhang, Xiyang Dai, Jianwei Yang, Bin Xiao, Lu Yuan, Lei Zhang, Jianfeng Gao

This paper presents a new Vision Transformer (ViT) architecture Multi-Scale Vision Longformer, which significantly enhances the ViT of for encoding high-resolution images using two techniques. The first is the multi-scale model structure, which provides image encodings at multiple scales with manageable computational cost. The second is the attention mechanism of vision Longformer, which is a variant of Longformer, originally developed for natural language processing, and achieves a linear complexity w.r.t. the number of input tokens. A comprehensive empirical study shows that the new ViT significantly outperforms several strong baselines, including the existing ViT models and their ResNet counterparts, and the Pyramid Vision Transformer from a concurrent work, on a range of vision tasks, including image classification, object detection, and segmentation. The models and source code are released at https://github.com/microsoft/vision-longformer.

PaperPDFConference PDFCodeCode 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 10 code samples harvested from 2 repositories linked to this paper: 8 ran (of which 0 constructed an object rather than computing a result; 5 with no instrument failure: 1 honoured, 2 violated, 2 with no contract checked; 3 where Syntology's instrument failed) · 2 unverified. Where Syntology's instrument failed: 3 where our contract draft was wrong; failures of Syntology's instrument, not of the code.

By repository: official repository: 3 samples from 1 repository, 3 ran (1 with no instrument failure; 2 where Syntology's instrument failed); community (archive-listed): 7 samples from 1 repository, 5 ran (4 with no instrument failure; 1 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.

No sample is linked to this paper from 1 of its repositories that Syntology harvested for other papers:

Levels written between 2026-09-22 and 2026-09-24 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed. 2 of the 8 samples that ran have no recorded date.

microsoft/vision-longformer officialmentioned in papermentioned on GitHubpytorchSyntology: 3 samples, 3 ran (2 instrument-failed) · 0 unverified report
microsoft/VisionLongformerForObjectDetection mentioned on GitHubpytorchMITSyntology: 7 samples, 5 ran (1 instrument-failed) · 2 unverified report
microsoft/esvit pytorchSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to CvT: Introducing Convolutions to Vision Transformers (arXiv:2103.15808; that paper's own run record: 39 ran (of which 19 constructed an object rather than computing a result; 36 with no instrument failure: 2 honoured, 0 violated, 34 with no contract checked; 3 where Syntology's instrument failed) · 8 unverified) and Efficient Self-supervised Vision Transformers for Representation Learning (arXiv:2106.09785; that paper's own run record: 2 ran (of which 0 constructed an object rather than computing a result; 2 with no instrument failure: 0 honoured, 0 violated, 2 with no contract checked; 0 where Syntology's instrument failed) · 4 unverified); nothing checks that they implement this paper's method. report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. 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.

A paper named beside a repository with no sample linked to this paper is shown with that paper's own run record, not this paper's: “ran” means executed on a synthesized input, not that the code is correct, and “where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Papers are listed in arXiv-id order, at most three per repository.

Code Syntology ran Syntology

10 samples harvested: 8 ran (of which 0 constructed an object rather than computing a result; 5 with no instrument failure: 1 honoured, 2 violated, 2 with no contract checked; 3 where Syntology's instrument failed) · 2 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page has not changed since , the first build that kept a record date for it; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

1ran · honoured contract
2ran · violated contract
3ran · our draft was wrong
2ran
2unverified

Licence: 2 of the 10 samples are pointer only, meaning Syntology does not serve that copy's text. For 2 of them, this paper's copy is under a permissive licence and cleared, but the stored copy of the same code that Syntology serves is not; the licence label's tooltip gives Syntology's recorded reason. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from 2 repositories linked to this paper, official or community; the list below is grouped by repository, each group headed by its name and what it is to the paper. “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. A census tag appears when Syntology's census ran implementations with that name on one shared input per bucket: it gives how many were compared, from how many papers, and how many distinct outputs they produced, and names this sample's bucket when this sample was one of them.

microsoft/vision-longformer official repository 3 samples, 3 ran (2 instrument-failed) · 0 unverified
find_modules in 6 papers src/models/msvit.py ran · our draft was wrong MIT (permissive) · 4e291c3ece372165 · report fixture: signature nn_module:callable/0/?, type:callable/0/? · returned list · level recorded 2026-09-22
get_module_device in 7 papers src/models/msvit.py ran · our draft was wrong MIT (permissive) · f8ef5ca2539fcee5 · report fixture: signature module:callable/0/? · returned device · level recorded 2026-09-22
init_ in 12 papers src/models/msvit.py ran · honoured contract fingerprinted MIT (permissive) · 3444251df88e6169 · report fixture: signature tensor:Tensor/2/float · returned Tensor · level recorded 2026-09-22
microsoft/VisionLongformerForObjectDetection community (archive-listed) 7 samples, 5 ran (1 instrument-failed) · 2 unverified
default in 430 papers msvit/performer.py ran · violated contract fingerprinted MIT (permissive) here; the stored copy is pointer only · 60fff7c3c400d7ff · report fixture: signature val:int/0/int, d:int/0/int · returned int · level recorded, date not recorded
exists in 528 papers
census: 5 implementations from 103 papers, 1 distinct output on one shared input this sample: bucket 1 of 1, arg 1: rank 1, kind int, dtype int64
msvit/performer.py ran · violated contract MIT (permissive) here; the stored copy is pointer only · aa5486a3650902d8 · report fixture: signature val:Tensor/1/int · returned bool · level recorded, date not recorded
resize_pos_embed_1d converter.py ran MIT (permissive) · 0b417de020bf073b · report fixture: signature posemb:Tensor/3/float, shape_new:tuple/1/int · returned Tensor · level recorded 2026-09-24
resize_pos_embed_2d converter.py ran fingerprinted MIT (permissive) · 3d16b011ef55503e · report fixture: signature posemb:Tensor/2/float, shape_new:tuple/1/int · returned Tensor · level recorded 2026-09-24
softmax_kernel in 7 papers msvit/performer.py ran · our draft was wrong MIT (permissive) · 757aee54843af070 · report fixture: signature data:Tensor/4/float, projection_matrix:Tensor/2/float, is_query:bool/0/bool, normalize_data:bool/0/bool, eps:float/0/float, device:NoneType/0/? · returned Tensor · level recorded 2026-09-22
mask_invalid_locations msvit/slidingchunk_2d.py unverified MIT (permissive) · ca8f2c57ed27ae92 · report
slidingchunk_2dautograd msvit/slidingchunk_2d.py unverified MIT (permissive) · 56b6ba9dcdd328f4 · report

Tasks

Image ClassificationInstance SegmentationObject Detectionimage-classificationobject-detection

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Image Classification ImageNet ViL-Medium-D GFLOPs 8.7 #442 of 1060 Archive leaderboard report
Image Classification ImageNet ViL-Medium-D Number of params 39.7M #442 of 1060 Archive leaderboard report
Image Classification ImageNet ViL-Medium-D Top 1 Accuracy 83.3% #442 of 1060 Archive leaderboard report
Image Classification ImageNet ViL-Base-D GFLOPs 13.4 #460 of 1060 Archive leaderboard report
Image Classification ImageNet ViL-Base-D Number of params 55.7M #460 of 1060 Archive leaderboard report
Image Classification ImageNet ViL-Base-D Top 1 Accuracy 83.2% #460 of 1060 Archive leaderboard report
Image Classification ImageNet ViL-Medium-W Number of params 39.8M #487 of 1060 Archive leaderboard report
Image Classification ImageNet ViL-Medium-W Top 1 Accuracy 82.9% #487 of 1060 Archive leaderboard report
Image Classification ImageNet ViL-Small GFLOPs 4.86 #584 of 1060 Archive leaderboard report
Image Classification ImageNet ViL-Small Number of params 24.6M #584 of 1060 Archive leaderboard report
Image Classification ImageNet ViL-Small Top 1 Accuracy 82% #584 of 1060 Archive leaderboard report
Image Classification ImageNet ViL-Base-W GFLOPs 6.74 #600 of 1060 Archive leaderboard report
Image Classification ImageNet ViL-Base-W Number of params 79M #600 of 1060 Archive leaderboard report
Image Classification ImageNet ViL-Base-W Top 1 Accuracy 81.9% #600 of 1060 Archive leaderboard report
Image Classification ImageNet ViL-Tiny-RPB GFLOPs 1.3 #904 of 1060 Archive leaderboard report
Image Classification ImageNet ViL-Tiny-RPB Number of params 6.7M #904 of 1060 Archive leaderboard report
Image Classification ImageNet ViL-Tiny-RPB Top 1 Accuracy 76.7% #904 of 1060 Archive leaderboard report
Instance Segmentation COCO minival Mask R-CNN (ViL Base, multi-scale, 3x lr) AP75 49.9 #47 of 93 Archive leaderboard report
Instance Segmentation COCO minival Mask R-CNN (ViL Base, multi-scale, 3x lr) mask AP 45.7 #47 of 93 Archive leaderboard report
Instance Segmentation COCO minival Mask R-CNN (ViL Base, 1x lr) AP50 67.2 #48 of 93 Archive leaderboard report
Instance Segmentation COCO minival Mask R-CNN (ViL Base, 1x lr) AP75 49.3 #48 of 93 Archive leaderboard report
Instance Segmentation COCO minival Mask R-CNN (ViL Base, 1x lr) mask AP 45.1 #48 of 93 Archive leaderboard report
Object Detection COCO minival RetinaNet (ViL-Base, multi-scale, 3x) AP75 47.6 #124 of 220 Archive leaderboard report
Object Detection COCO minival RetinaNet (ViL-Base, multi-scale, 3x) APL 58.1 #124 of 220 Archive leaderboard report
Object Detection COCO minival RetinaNet (ViL-Base, multi-scale, 3x) APM 48 #124 of 220 Archive leaderboard report
Object Detection COCO minival RetinaNet (ViL-Base, multi-scale, 3x) APS 29.9 #124 of 220 Archive leaderboard report
Object Detection COCO minival RetinaNet (ViL-Base, multi-scale, 3x) box AP 44.7 #124 of 220 Archive leaderboard report
Object Detection COCO minival RetinaNet (ViL-Base) AP50 65.5 #132 of 220 Archive leaderboard report
Object Detection COCO minival RetinaNet (ViL-Base) AP75 47.1 #132 of 220 Archive leaderboard report
Object Detection COCO minival RetinaNet (ViL-Base) APL 58.3 #132 of 220 Archive leaderboard report
Object Detection COCO minival RetinaNet (ViL-Base) APM 47.9 #132 of 220 Archive leaderboard report
Object Detection COCO minival RetinaNet (ViL-Base) APS 28.9 #132 of 220 Archive leaderboard report
Object Detection COCO minival RetinaNet (ViL-Base) box AP 44.3 #132 of 220 Archive leaderboard report

Ranks are positions in the archive's leaderboards as they stood at the 2025-07-28 snapshot. Results published since then are not among these rows, so a rank here is not a current standing.

Methods

1x1 ConvolutionAbsolute Position EncodingsAdamAdamWAttentionAttention DropoutAverage PoolingBPEBatch NormalizationBottleneck Residual BlockConvolutionDense ConnectionsDropoutGlobal Average PoolingKaiming InitializationLabel SmoothingLayer NormalizationLinear LayerLinear Warmup With Linear DecayLongformerMax PoolingMulti-Head AttentionPosition-Wise Feed-Forward LayerReLUResidual BlockResidual ConnectionSoftmaxTransformerVision TransformerWeight DecayWordPiece

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