Papers › FastViT: A Fast Hybrid Vision Transformer using Structural Reparameterization

FastViT: A Fast Hybrid Vision Transformer using Structural Reparameterization

24 Mar 2023ICCV 2023 1arXiv:2303.14189archive 2025-07-28

Pavan Kumar Anasosalu Vasu, James Gabriel, Jeff Zhu, Oncel Tuzel, Anurag Ranjan

The recent amalgamation of transformer and convolutional designs has led to steady improvements in accuracy and efficiency of the models. In this work, we introduce FastViT, a hybrid vision transformer architecture that obtains the state-of-the-art latency-accuracy trade-off. To this end, we introduce a novel token mixing operator, RepMixer, a building block of FastViT, that uses structural reparameterization to lower the memory access cost by removing skip-connections in the network. We further apply train-time overparametrization and large kernel convolutions to boost accuracy and empirically show that these choices have minimal effect on latency. We show that - our model is 3.5x faster than CMT, a recent state-of-the-art hybrid transformer architecture, 4.9x faster than EfficientNet, and 1.9x faster than ConvNeXt on a mobile device for the same accuracy on the ImageNet dataset. At similar latency, our model obtains 4.2% better Top-1 accuracy on ImageNet than MobileOne. Our model consistently outperforms competing architectures across several tasks -- image classification, detection, segmentation and 3D mesh regression with significant improvement in latency on both a mobile device and a desktop GPU. Furthermore, our model is highly robust to out-of-distribution samples and corruptions, improving over competing robust models. Code and models are available at https://github.com/apple/ml-fastvit.

PaperPDFConference PDFCodeCode Syntology ranMethod as code

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

No sample from the archive's official repositories (apple/ml-fastvit, rwightman/pytorch-image-models) is linked to this paper; the 5 samples below come from 1 community repository. Syntology harvested code from rwightman/pytorch-image-models for another paper; apple/ml-fastvit was not harvested.

rwightman/pytorch-image-models: Syntology's graph links these samples to AdamP: Slowing Down the Slowdown for Momentum Optimizers on Scale-invariant Weights (arXiv:2006.08217; that paper's own run record: 4 ran (of which 1 constructed an object rather than computing a result; 2 with no instrument failure: 1 honoured, 0 violated, 1 with no contract checked; 2 where Syntology's instrument failed) · 0 unverified); nothing checks that they implement this paper's method.

Syntology 5 code samples harvested from 1 repository linked to this paper: 1 ran (of which 1 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 0 where Syntology's instrument failed) · 4 unverified; the one sample that ran constructed an object rather than computing a result. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code; here there are none.

By repository: community (archive-listed): 5 samples from 1 repository, 1 ran (1 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.

No sample is linked to this paper from 2 of its repositories that Syntology harvested for other papers:

Levels written on 2026-09-22 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

apple/ml-fastvit officialmentioned in papermentioned on GitHubpytorchNOASSERTIONSyntology: not harvested report
rwightman/pytorch-image-models officialmentioned in papermentioned on GitHubpytorchSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to AdamP: Slowing Down the Slowdown for Momentum Optimizers on Scale-invariant Weights (arXiv:2006.08217; that paper's own run record: 4 ran (of which 1 constructed an object rather than computing a result; 2 with no instrument failure: 1 honoured, 0 violated, 1 with no contract checked; 2 where Syntology's instrument failed) · 0 unverified); nothing checks that they implement this paper's method. report
balala8/FastViT_pytorch mentioned on GitHubpytorchSyntology: 5 samples, 1 ran (0 instrument-failed) · 4 unverified report
MindCode-4/code-1 mindsporeSyntology: not harvested report
leondgarse/keras_cv_attention_models tfSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to RepViT: Revisiting Mobile CNN From ViT Perspective (arXiv:2307.09283; that paper's own run record: 6 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 0 violated, 3 with no contract checked; 3 where Syntology's instrument failed) · 6 unverified); nothing checks that they implement this paper's method. report
gitlab.com/birder/birder pytorchSyntology: not harvested 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

5 samples harvested: 1 ran (of which 1 constructed an object rather than computing a result; 1 with no instrument failure: 0 honoured, 0 violated, 1 with no contract checked; 0 where Syntology's instrument failed) · 4 unverified; the one sample that ran constructed an object rather than computing a result. “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
4unverified

Licence: 5 of the 5 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 balala8/FastViT_pytorch. “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”. For 1 sample that ran, the graph records a signature this page does not show: for a class the fixture built without calling, the harness called the object with inputs it made up itself, so the recorded signature is the harness's probe, not the fixture's arguments. “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.

balala8/FastViT_pytorch community (archive-listed) 5 samples, 1 ran (0 instrument-failed) · 4 unverified
RepBlock model.py ran fingerprinted no licence file found · pointer only · 619ff60ae2e6b453 · report fixture: returned RepBlock · level recorded 2026-09-22
AttnTokenMixer model.py unverified no licence file found · pointer only · 8a8b8dbae01c1b35 · report
FastStage model.py unverified no licence file found · pointer only · 5208ceab2ca87ffd · report
FastViTBlock model.py unverified no licence file found · pointer only · f22147add9162e32 · report
FastVit model.py unverified no licence file found · pointer only · 753ec9398525102f · report

Method as code Syntology

What Syntology's graph records as this paper's claimed method, lifted into code: each assembly names its bricks (functions and classes), each with its own role and its own verification level. An assembly is not a check that the code implements the method correctly or reproduces the paper's results. Levels are each brick's own verification level in Syntology's graph (level 0: no recorded run), listed brick by brick and never combined into one level; “ran” means the brick executed on a synthesized input, not that its output is correct, and the instrument figure counts failures of Syntology's instrument, not of the code. Where a model read the code against the paper's claim, each assembly's label verdict is one model's judgement, kept in this page's JSON twin; it is not calibrated, so no card prints it.

1 assembly on this paper: 1 harvested, 0 generated.

harvested Assembly for FastViT (graph method: FastViT). 6 bricks; individually on a fixture: 2 ran (of which 1 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. The assembly has not been run end-to-end.

  1. FastVit entry · level 0 · unverified · 753ec9398525102f
  2. FastStage model · level 0 · unverified · 5208ceab2ca87ffd
  3. RepBlock model · level 1 · ran · 619ff60ae2e6b453
  4. FastViTBlock model · level 0 · unverified · f22147add9162e32
  5. AttnTokenMixer model · level 0 · unverified · 8a8b8dbae01c1b35
  6. Attention model · level 1 · ran · a17528cd8209371a

Each brick ends with its code_sha256, Syntology's identity for that exact code; click one to copy the get_code call that returns it. For agents, on Syntology's MCP service: assembly(arxiv_id="2303.14189") returns these assemblies with each brick's role and wiring (how to connect).

Tasks

3D Hand Pose EstimationImage ClassificationSemantic Segmentationimage-classification

1 archive task tag without a task page not shown.

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
3D Hand Pose Estimation FreiHAND FastViT-MA36 PA-F@15mm 0.981 #17 of 33 Archive leaderboard report
3D Hand Pose Estimation FreiHAND FastViT-MA36 PA-F@5mm 0.722 #17 of 33 Archive leaderboard report
3D Hand Pose Estimation FreiHAND FastViT-MA36 PA-MPJPE 6.6 #17 of 33 Archive leaderboard report
3D Hand Pose Estimation FreiHAND FastViT-MA36 PA-MPVPE 6.7 #17 of 33 Archive leaderboard report
Image Classification ImageNet FastViT-MA36 Top 1 Accuracy 84.9% #276 of 1060 Archive leaderboard report
Image Classification ImageNet FastViT-SA36 Top 1 Accuracy 84.5% #307 of 1060 Archive leaderboard report
Image Classification ImageNet FastViT-SA24 Top 1 Accuracy 82.6% #521 of 1060 Archive leaderboard report
Image Classification ImageNet FastViT-SA12 Top 1 Accuracy 80.6% #693 of 1060 Archive leaderboard report
Image Classification ImageNet FastViT-S12 Top 1 Accuracy 79.8% #737 of 1060 Archive leaderboard report
Image Classification ImageNet FastViT-T12 Top 1 Accuracy 79.1% #780 of 1060 Archive leaderboard report
Image Classification ImageNet FastViT-T8 Top 1 Accuracy 75.6% #942 of 1060 Archive leaderboard report
Semantic Segmentation ADE20K FastViT-MA36 Mean IoU (class) 44.6 #232 of 235 Archive leaderboard report
Semantic Segmentation ADE20K FastViT-SA36 Mean IoU (class) 42.9 #233 of 235 Archive leaderboard report
Semantic Segmentation ADE20K FastViT-SA24 Mean IoU (class) 41 #234 of 235 Archive leaderboard report
Semantic Segmentation ADE20K FastViT-SA12 Mean IoU (class) 38 #235 of 235 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 ConvolutionAttentionAverage PoolingBatch NormalizationConvNeXtConvolutionDense ConnectionsDepthwise ConvolutionDepthwise Separable ConvolutionDropoutInverted Residual BlockLayer NormalizationLinear LayerMulti-Head AttentionPointwise ConvolutionRMSPropReLUResidual ConnectionSigmoid ActivationSoftmaxSqueeze-and-Excitation BlockVision Transformer

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