Papers › High-Performance Large-Scale Image Recognition Without Normalization

High-Performance Large-Scale Image Recognition Without Normalization

11 Feb 2021arXiv:2102.06171archive 2025-07-28

Andrew Brock, Soham De, Samuel L. Smith, Karen Simonyan

Batch normalization is a key component of most image classification models, but it has many undesirable properties stemming from its dependence on the batch size and interactions between examples. Although recent work has succeeded in training deep ResNets without normalization layers, these models do not match the test accuracies of the best batch-normalized networks, and are often unstable for large learning rates or strong data augmentations. In this work, we develop an adaptive gradient clipping technique which overcomes these instabilities, and design a significantly improved class of Normalizer-Free ResNets. Our smaller models match the test accuracy of an EfficientNet-B7 on ImageNet while being up to 8.7x faster to train, and our largest models attain a new state-of-the-art top-1 accuracy of 86.5%. In addition, Normalizer-Free models attain significantly better performance than their batch-normalized counterparts when finetuning on ImageNet after large-scale pre-training on a dataset of 300 million labeled images, with our best models obtaining an accuracy of 89.2%. Our code is available at https://github.com/deepmind/ deepmind-research/tree/master/nfnets

PaperPDFCodeCode 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

No sample from the archive's official repository (deepmind/deepmind-research) is linked to this paper; the 3 samples below come from 1 community repository. Syntology harvested code from deepmind/deepmind-research for 4 other papers.

deepmind/deepmind-research: Syntology's graph links these samples to Disentangling by Subspace Diffusion (arXiv:2006.12982; that paper's own run record: 6 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; 6 where Syntology's instrument failed) · 3 unverified), Characterizing signal propagation to close the performance gap in unnormalized ResNets (arXiv:2101.08692; that paper's own run record: 1 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; 1 where Syntology's instrument failed) · 0 unverified), Perceiver: General Perception with Iterative Attention (arXiv:2103.03206; that paper's own run record: 41 ran (of which 20 constructed an object rather than computing a result; 30 with no instrument failure: 0 honoured, 2 violated, 28 with no contract checked; 11 where Syntology's instrument failed) · 14 unverified) and 1 more (every one in this page's JSON twin); nothing checks that they implement this paper's method.

Syntology 3 code samples harvested from 1 repository linked to this paper: 3 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; 3 where Syntology's instrument failed) · 0 unverified. Where Syntology's instrument failed: 3 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: community (archive-listed): 3 samples from 1 repository, 3 ran (0 with no instrument failure; 3 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:

No date is recorded for any of the 3 samples that ran: their levels were written before Syntology's graph kept the date.

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

deepmind/deepmind-research officialmentioned in papertfSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Disentangling by Subspace Diffusion (arXiv:2006.12982; that paper's own run record: 6 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; 6 where Syntology's instrument failed) · 3 unverified), Characterizing signal propagation to close the performance gap in unnormalized ResNets (arXiv:2101.08692; that paper's own run record: 1 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; 1 where Syntology's instrument failed) · 0 unverified), Perceiver: General Perception with Iterative Attention (arXiv:2103.03206; that paper's own run record: 41 ran (of which 20 constructed an object rather than computing a result; 30 with no instrument failure: 0 honoured, 2 violated, 28 with no contract checked; 11 where Syntology's instrument failed) · 14 unverified) and 1 more (every one in this page's JSON twin); nothing checks that they implement this paper's method. report
deepmind/deepmind-research officialjaxSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Disentangling by Subspace Diffusion (arXiv:2006.12982; that paper's own run record: 6 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; 6 where Syntology's instrument failed) · 3 unverified), Characterizing signal propagation to close the performance gap in unnormalized ResNets (arXiv:2101.08692; that paper's own run record: 1 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; 1 where Syntology's instrument failed) · 0 unverified), Perceiver: General Perception with Iterative Attention (arXiv:2103.03206; that paper's own run record: 41 ran (of which 20 constructed an object rather than computing a result; 30 with no instrument failure: 0 honoured, 2 violated, 28 with no contract checked; 11 where Syntology's instrument failed) · 14 unverified) and 1 more (every one in this page's JSON twin); nothing checks that they implement this paper's method. report
ahsanmemon/NFNets-Tensorflow mentioned on GitHubtfSyntology: not harvested report
ayulockin/Explore-NFNet mentioned on GitHubtfSyntology: not harvested report
benjs/nfnets_pytorch mentioned on GitHubpytorchSyntology: not harvested report
callmefish/NFnet mentioned on GitHubpytorchSyntology: 3 samples, 3 ran (3 instrument-failed) · 0 unverified report
hoangthang1607/nfnets-Tensorflow-2 mentioned on GitHubtfSyntology: not harvested report
nmichlo/experiment-normalized-activations mentioned on GitHubpytorchSyntology: not harvested report
psh01087/simple-nfnet mentioned on GitHubpytorchSyntology: not harvested report
rwightman/efficientdet-pytorch mentioned on GitHubpytorchSyntology: not harvested report
rwightman/pytorch-image-models mentioned 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
sayakpaul/Adaptive-Gradient-Clipping mentioned on GitHubtfSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift (arXiv:1502.03167; that paper's own run record: 19 ran (of which 5 constructed an object rather than computing a result; 14 with no instrument failure: 0 honoured, 0 violated, 14 with no contract checked; 5 where Syntology's instrument failed) · 2 unverified); nothing checks that they implement this paper's method. report
sense-x/tokenmix mentioned on GitHubpytorchSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to TokenMix: Rethinking Image Mixing for Data Augmentation in Vision Transformers (arXiv:2207.08409; that paper's own run record: 0 ran · 1 unverified); nothing checks that they implement this paper's method. report
smilingwolf/sw-cv-modelzoo mentioned on GitHubtfSyntology: not harvested report
vballoli/nfnets-pytorch mentioned on GitHubpytorchSyntology: not harvested report
ypeleg/nfnets-keras mentioned on GitHubtfSyntology: not harvested report
MS-Mind/MS-Code-06 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
mindspore-ai/models mindsporeSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to AdaBin: Improving Binary Neural Networks with Adaptive Binary Sets (arXiv:2208.08084; that paper's own run record: 4 ran (of which 0 constructed an object rather than computing a result; 4 with no instrument failure: 0 honoured, 0 violated, 4 with no contract checked; 0 where Syntology's instrument failed) · 1 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

3 samples harvested: 3 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; 3 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.

3ran; the contract check could not run

Licence: 3 of the 3 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 callmefish/NFnet. “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.

callmefish/NFnet community (archive-listed) 3 samples, 3 ran (3 instrument-failed) · 0 unverified
reduce_fn in 5 papers utils.py ran; the contract check could not run (RAISES) · metamorphic tier: invariant no licence file found · pointer only · 72ab490c68e07338 · report fixture: signature x:ArrayImpl/1/float, mode:str/0/str · returned ArrayImpl · level recorded, date not recorded
softmax_cross_entropy in 11 papers utils.py ran; the contract check could not run (RAISES) · metamorphic tier: invariant no licence file found · pointer only · 1f05ed2c304f566d · report fixture: signature logits:ArrayImpl/2/float, labels:ArrayImpl/2/float, reduction:str/0/str · returned ArrayImpl · level recorded, date not recorded
topk_correct utils.py ran; the contract check could not run (RAISES) · metamorphic tier: deterministic no licence file found · pointer only · 07d4886e953a2f31 · report fixture: signature logits:ArrayImpl/2/float, labels:ArrayImpl/1/int, mask:ArrayImpl/1/float, prefix:str/0/str, topk:tuple/1/int · returned dict · level recorded, date not recorded

Tasks

Image ClassificationVocal Bursts Intensity Predictionimage-classification

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Image Classification ImageNet NFNet-F4+ GFLOPs 367 #23 of 1060 Archive leaderboard report
Image Classification ImageNet NFNet-F4+ Number of params 527M #23 of 1060 Archive leaderboard report
Image Classification ImageNet NFNet-F4+ Top 1 Accuracy 89.2% #23 of 1060 Archive leaderboard report
Image Classification ImageNet NFNet-F6 w/ SAM GFLOPs 377.28 #141 of 1060 Archive leaderboard report
Image Classification ImageNet NFNet-F6 w/ SAM Number of params 438.4M #141 of 1060 Archive leaderboard report
Image Classification ImageNet NFNet-F6 w/ SAM Top 1 Accuracy 86.5% #141 of 1060 Archive leaderboard report
Image Classification ImageNet NFNet-F5 w/ SAM GFLOPs 289.76 #163 of 1060 Archive leaderboard report
Image Classification ImageNet NFNet-F5 w/ SAM Number of params 377.2M #163 of 1060 Archive leaderboard report
Image Classification ImageNet NFNet-F5 w/ SAM Top 1 Accuracy 86.3% #163 of 1060 Archive leaderboard report
Image Classification ImageNet NFNet-F5 GFLOPs 289.76 #184 of 1060 Archive leaderboard report
Image Classification ImageNet NFNet-F5 Number of params 377.2M #184 of 1060 Archive leaderboard report
Image Classification ImageNet NFNet-F5 Top 1 Accuracy 86.0% #184 of 1060 Archive leaderboard report
Image Classification ImageNet NFNet-F4 GFLOPs 215.24 #188 of 1060 Archive leaderboard report
Image Classification ImageNet NFNet-F4 Number of params 316.1M #188 of 1060 Archive leaderboard report
Image Classification ImageNet NFNet-F4 Top 1 Accuracy 85.9% #188 of 1060 Archive leaderboard report
Image Classification ImageNet NFNet-F3 GFLOPs 114.76 #211 of 1060 Archive leaderboard report
Image Classification ImageNet NFNet-F3 Number of params 254.9M #211 of 1060 Archive leaderboard report
Image Classification ImageNet NFNet-F3 Top 1 Accuracy 85.7% #211 of 1060 Archive leaderboard report
Image Classification ImageNet NFNet-F2 GFLOPs 62.59 #260 of 1060 Archive leaderboard report
Image Classification ImageNet NFNet-F2 Number of params 193.8M #260 of 1060 Archive leaderboard report
Image Classification ImageNet NFNet-F2 Top 1 Accuracy 85.1% #260 of 1060 Archive leaderboard report
Image Classification ImageNet NFNet-F1 GFLOPs 35.54 #301 of 1060 Archive leaderboard report
Image Classification ImageNet NFNet-F1 Number of params 132.6M #301 of 1060 Archive leaderboard report
Image Classification ImageNet NFNet-F1 Top 1 Accuracy 84.7% #301 of 1060 Archive leaderboard report
Image Classification ImageNet NFNet-F0 GFLOPs 12.38 #416 of 1060 Archive leaderboard report
Image Classification ImageNet NFNet-F0 Number of params 71.5M #416 of 1060 Archive leaderboard report
Image Classification ImageNet NFNet-F0 Top 1 Accuracy 83.6% #416 of 1060 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

Gradient Clipping

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