Papers › Restormer: Efficient Transformer for High-Resolution Image Restoration

Restormer: Efficient Transformer for High-Resolution Image Restoration

18 Nov 2021CVPR 2022 1arXiv:2111.09881archive 2025-07-28

Syed Waqas Zamir, Aditya Arora, Salman Khan, Munawar Hayat, Fahad Shahbaz Khan, Ming-Hsuan Yang

Since convolutional neural networks (CNNs) perform well at learning generalizable image priors from large-scale data, these models have been extensively applied to image restoration and related tasks. Recently, another class of neural architectures, Transformers, have shown significant performance gains on natural language and high-level vision tasks. While the Transformer model mitigates the shortcomings of CNNs (i.e., limited receptive field and inadaptability to input content), its computational complexity grows quadratically with the spatial resolution, therefore making it infeasible to apply to most image restoration tasks involving high-resolution images. In this work, we propose an efficient Transformer model by making several key designs in the building blocks (multi-head attention and feed-forward network) such that it can capture long-range pixel interactions, while still remaining applicable to large images. Our model, named Restoration Transformer (Restormer), achieves state-of-the-art results on several image restoration tasks, including image deraining, single-image motion deblurring, defocus deblurring (single-image and dual-pixel data), and image denoising (Gaussian grayscale/color denoising, and real image denoising). The source code and pre-trained models are available at https://github.com/swz30/Restormer.

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 4 code samples harvested from 2 repositories linked to this paper: 4 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 4 where Syntology's instrument failed) · 0 unverified. Where Syntology's instrument failed: 2 where our contract draft was wrong; 2 where the contract check could not run; failures of Syntology's instrument, not of the code. Every run on this page is a failure of Syntology's instrument, not of the code.

By repository: official repository: 2 samples from 1 repository, 2 ran (0 with no instrument failure; 2 where Syntology's instrument failed); community (archive-listed): 2 samples from 1 repository, 2 ran (0 with no instrument failure; 2 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 6 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. 2 of the 4 samples that ran have no recorded date.

13 repositories listed; official and paper-mentioned ones first.

swz30/restormer officialmentioned in papermentioned on GitHubpytorchMITSyntology: 2 samples, 2 ran (2 instrument-failed) · 0 unverified report
GarrickZ2/Image-Denoising mentioned on GitHubpytorchSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Learning to Generate Realistic Noisy Images via Pixel-level Noise-aware Adversarial Training (arXiv:2204.02844; that paper's own run record: 14 ran (of which 0 constructed an object rather than computing a result; 13 with no instrument failure: 0 honoured, 0 violated, 13 with no contract checked; 1 where Syntology's instrument failed) · 4 unverified); nothing checks that they implement this paper's method. report
HDCVLab/MC-Blur-Dataset mentioned on GitHubpytorchSyntology: not harvested report
MKFMIKU/VIDM mentioned on GitHubpytorchSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to VIDM: Video Implicit Diffusion Models (arXiv:2212.00235; that paper's own run record: 2 ran (of which 0 constructed an object rather than computing a result; 1 with no instrument failure: 1 honoured, 0 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 0 unverified); nothing checks that they implement this paper's method. report
gymoon10/Instance-Segmentation-with-SpatialEmbedding-CA mentioned on GitHubpytorchSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Instance Segmentation by Jointly Optimizing Spatial Embeddings and Clustering Bandwidth (arXiv:1906.11109; 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, 2 violated, 0 with no contract checked; 0 where Syntology's instrument failed) · 0 unverified) and BiSeNet V2: Bilateral Network with Guided Aggregation for Real-time Semantic Segmentation (arXiv:2004.02147; that paper's own run record: 7 ran (of which 0 constructed an object rather than computing a result; 6 with no instrument failure: 2 honoured, 3 violated, 1 with no contract checked; 1 where Syntology's instrument failed) · 3 unverified); nothing checks that they implement this paper's method. report
leftthomas/restormer mentioned on GitHubpytorchSyntology: 2 samples, 2 ran (2 instrument-failed) · 0 unverified report
prakashSidd18/blind_augmentation mentioned on GitHubpytorchSyntology: not harvested report
stephen0808/dnlut mentioned on GitHubpytorchSyntology: not harvested report
swz30/CycleISP mentioned on GitHubpytorchSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Multi-Stage Progressive Image Restoration (arXiv:2102.02808; that paper's own run record: 18 ran (of which 13 constructed an object rather than computing a result; 17 with no instrument failure: 0 honoured, 0 violated, 17 with no contract checked; 1 where Syntology's instrument failed) · 8 unverified); nothing checks that they implement this paper's method. report
swz30/MIRNet mentioned on GitHubpytorchNOASSERTIONSyntology: not harvested report
swz30/MPRNet mentioned on GitHubpytorchSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Multi-Stage Progressive Image Restoration (arXiv:2102.02808; that paper's own run record: 18 ran (of which 13 constructed an object rather than computing a result; 17 with no instrument failure: 0 honoured, 0 violated, 17 with no contract checked; 1 where Syntology's instrument failed) · 8 unverified); nothing checks that they implement this paper's method. report
swz30/mirnetv2 mentioned on GitHubpytorchSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Multi-Stage Progressive Image Restoration (arXiv:2102.02808; that paper's own run record: 18 ran (of which 13 constructed an object rather than computing a result; 17 with no instrument failure: 0 honoured, 0 violated, 17 with no contract checked; 1 where Syntology's instrument failed) · 8 unverified); nothing checks that they implement this paper's method. report
txyugood/Restormer_Paddle mentioned on GitHubpaddleSyntology: 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

4 samples harvested: 4 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 4 where Syntology's instrument failed) · 0 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Every run on this page is a failure 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.

2ran · our draft was wrong
2ran; the contract check could not run

Licence: 4 of the 4 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 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. 2 samples whose link to this paper records no repository are listed under one of this paper's own repositories, because identical code was harvested from it for another 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.

swz30/restormer official repository 2 samples, 2 ran (2 instrument-failed) · 0 unverified
to_3d in 66 papers
census: 3 implementations from 20 papers, 1 distinct output on one shared input this sample: bucket 1 of 1, arg 1: rank 4, kind float, dtype float32
basicsr/models/archs/restormer_arch.py identical code Syntology harvested from this repository for another paper; this paper's link records no repository, path or licence of its own ran; the contract check could not run (RAISES) · metamorphic tier: invariant fingerprinted licence of this copy not recorded · 82a15cc1e46f7e4d · report fixture: signature x:Tensor/4/float · returned Tensor · level recorded, date not recorded
to_4d in 69 papers basicsr/models/archs/restormer_arch.py identical code Syntology harvested from this repository for another paper; this paper's link records no repository, path or licence of its own ran; the contract check could not run (RAISES) · metamorphic tier: invariant licence of this copy not recorded · b20f2a5df739a59e · report fixture: signature x:Tensor/3/float, h:int/0/int, w:int/0/int · returned Tensor · level recorded, date not recorded
leftthomas/restormer community (archive-listed) 2 samples, 2 ran (2 instrument-failed) · 0 unverified
init_args in 12 papers utils.py ran · our draft was wrong no licence file found · pointer only · 4ea9fc528e09fe19 · report fixture: signature args:Args/0/? · returned Config · level recorded 2026-09-22
rgb_to_y utils.py ran · our draft was wrong fingerprinted no licence file found · pointer only · f66118f3a493e2fd · report fixture: signature x:Tensor/4/float · returned Tensor · level recorded 2026-09-22

Tasks

Color Image DenoisingDeblurringDenoisingGrayscale Image DenoisingImage DeblurringImage Defocus DeblurringImage DenoisingImage RestorationRain RemovalSingle Image DerainingSingle Image DesnowingSpectral ReconstructionVideo derainingVocal Bursts Intensity Prediction

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Color Image Denoising Kodak24 sigma50 Restormer PSNR 30.01 #4 of 9 Archive leaderboard report
Color Image Denoising Urban100 sigma50 Restormer PSNR 30.02 #5 of 9 Archive leaderboard report
Color Image Denoising urban100 sigma15 Restormer Average PSNR 35.13 #4 of 8 Archive leaderboard report
Deblurring GoPro Restormer PSNR 32.92 #29 of 56 Archive leaderboard report
Deblurring GoPro Restormer SSIM 0.961 #29 of 56 Archive leaderboard report
Deblurring HIDE (trained on GOPRO) Restormer PSNR (sRGB) 31.22 #14 of 26 Archive leaderboard report
Deblurring HIDE (trained on GOPRO) Restormer Params (M) 26.13 #14 of 26 Archive leaderboard report
Deblurring HIDE (trained on GOPRO) Restormer SSIM (sRGB) 0.942 #14 of 26 Archive leaderboard report
Deblurring MSU BASED Restormer local ERQAv2.0 0.73875 #10 of 11 Archive leaderboard report
Deblurring MSU BASED Restormer local LPIPS 0.08251 #10 of 11 Archive leaderboard report
Deblurring MSU BASED Restormer local PSNR 31.12341 #10 of 11 Archive leaderboard report
Deblurring MSU BASED Restormer local SSIM 0.94217 #10 of 11 Archive leaderboard report
Deblurring MSU BASED Restormer local Subjective 0.1231 #10 of 11 Archive leaderboard report
Deblurring MSU BASED Restormer local VMAF 65.25911 #10 of 11 Archive leaderboard report
Deblurring MSU BASED Restormer ERQAv2.0 0.74776 #11 of 11 Archive leaderboard report
Deblurring MSU BASED Restormer LPIPS 0.08239 #11 of 11 Archive leaderboard report
Deblurring MSU BASED Restormer PSNR 31.76111 #11 of 11 Archive leaderboard report
Deblurring MSU BASED Restormer SSIM 0.94632 #11 of 11 Archive leaderboard report
Deblurring MSU BASED Restormer Subjective 0.1175 #11 of 11 Archive leaderboard report
Deblurring MSU BASED Restormer VMAF 66.3964 #11 of 11 Archive leaderboard report
Deblurring RSBlur Restormer Average PSNR 33.69 #8 of 12 Archive leaderboard report
Deblurring RealBlur-J (trained on GoPro) Restormer PSNR (sRGB) 28.96 #6 of 15 Archive leaderboard report
Deblurring RealBlur-J (trained on GoPro) Restormer SSIM (sRGB) 0.879 #6 of 15 Archive leaderboard report
Deblurring RealBlur-R (trained on GoPro) Restormer PSNR (sRGB) 36.19 #4 of 19 Archive leaderboard report
Deblurring RealBlur-R (trained on GoPro) Restormer SSIM (sRGB) 0.957 #4 of 19 Archive leaderboard report
Grayscale Image Denoising BSD68 sigma15 Restormer PSNR 31.96 #4 of 16 Archive leaderboard report
Grayscale Image Denoising Urban100 sigma15 Restormer PSNR 33.79 #1 of 7 Archive leaderboard report
Grayscale Image Denoising Urban100 sigma25 Restormer PSNR 31.46 #3 of 10 Archive leaderboard report
Grayscale Image Denoising Urban100 sigma50 Restormer PSNR 28.29 #4 of 10 Archive leaderboard report
Image Deblurring GoPro Restormer PSNR 32.92 #31 of 55 Archive leaderboard report
Image Deblurring GoPro Restormer Params (M) 26.13 #31 of 55 Archive leaderboard report
Image Deblurring GoPro Restormer SSIM 0.961 #31 of 55 Archive leaderboard report
Image Denoising DND Restormer PSNR (sRGB) 40.03 #4 of 16 Archive leaderboard report
Image Denoising DND Restormer SSIM (sRGB) 0.956 #4 of 16 Archive leaderboard report
Image Denoising SIDD Restormer PSNR (sRGB) 40.02 #6 of 22 Archive leaderboard report
Image Denoising SIDD Restormer SSIM (sRGB) 0.960 #6 of 22 Archive leaderboard report
Image Restoration CDD-11 Restormer Average PSNR (dB) 26.99 #3 of 14 Archive leaderboard report
Image Restoration CDD-11 Restormer SSIM 0.8646 #3 of 14 Archive leaderboard report
Single Image Deraining Rain100H Restormer PSNR 31.46 #6 of 19 Archive leaderboard report
Single Image Deraining Rain100H Restormer SSIM 0.904 #6 of 19 Archive leaderboard report
Single Image Deraining Rain100L Restormer PSNR 38.99 #7 of 19 Archive leaderboard report
Single Image Deraining Rain100L Restormer SSIM 0.978 #7 of 19 Archive leaderboard report
Single Image Deraining Test100 Restormer PSNR 32.00 #1 of 12 Archive leaderboard report
Single Image Deraining Test100 Restormer SSIM 0.923 #1 of 12 Archive leaderboard report
Single Image Deraining Test1200 Restormer PSNR 33.19 #5 of 14 Archive leaderboard report
Single Image Deraining Test1200 Restormer SSIM 0.926 #5 of 14 Archive leaderboard report
Single Image Deraining Test2800 Restormer PSNR 34.18 #2 of 12 Archive leaderboard report
Single Image Deraining Test2800 Restormer SSIM 0.944 #2 of 12 Archive leaderboard report
Single Image Desnowing CSD Restormer Average PSNR (dB) 35.43 #3 of 6 Archive leaderboard report
Spectral Reconstruction ARAD-1K Restormer MRAE 0.1833 #4 of 11 Archive leaderboard report
Spectral Reconstruction ARAD-1K Restormer PSNR 33.40 #4 of 11 Archive leaderboard report
Spectral Reconstruction ARAD-1K Restormer RMSE 0.0274 #4 of 11 Archive leaderboard report
Video deraining VRDS Restormer PSNR 29.59 #3 of 8 Archive leaderboard report
Video deraining VRDS Restormer SSIM 0.9206 #3 of 8 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

Absolute Position EncodingsAdamAttentionBPEDense ConnectionsDropoutLabel SmoothingLayer NormalizationLinear LayerMulti-Head AttentionPosition-Wise Feed-Forward LayerResidual ConnectionSoftmaxTransformer

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