Papers › Self-training with Noisy Student improves ImageNet classification

Self-training with Noisy Student improves ImageNet classification

11 Nov 2019CVPR 2020 6arXiv:1911.04252archive 2025-07-28

Qizhe Xie, Minh-Thang Luong, Eduard Hovy, Quoc V. Le

We present Noisy Student Training, a semi-supervised learning approach that works well even when labeled data is abundant. Noisy Student Training achieves 88.4% top-1 accuracy on ImageNet, which is 2.0% better than the state-of-the-art model that requires 3.5B weakly labeled Instagram images. On robustness test sets, it improves ImageNet-A top-1 accuracy from 61.0% to 83.7%, reduces ImageNet-C mean corruption error from 45.7 to 28.3, and reduces ImageNet-P mean flip rate from 27.8 to 12.2. Noisy Student Training extends the idea of self-training and distillation with the use of equal-or-larger student models and noise added to the student during learning. On ImageNet, we first train an EfficientNet model on labeled images and use it as a teacher to generate pseudo labels for 300M unlabeled images. We then train a larger EfficientNet as a student model on the combination of labeled and pseudo labeled images. We iterate this process by putting back the student as the teacher. During the learning of the student, we inject noise such as dropout, stochastic depth, and data augmentation via RandAugment to the student so that the student generalizes better than the teacher. Models are available at https://github.com/tensorflow/tpu/tree/master/models/official/efficientnet. Code is available at https://github.com/google-research/noisystudent.

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

Syntology 24 code samples harvested from 2 repositories linked to this paper: 13 ran (of which 0 constructed an object rather than computing a result; 10 with no instrument failure: 2 honoured, 0 violated, 8 with no contract checked; 3 where Syntology's instrument failed) · 11 unverified. Where Syntology's instrument failed: 1 where our contract draft was wrong; 2 where the contract check could not run; failures of Syntology's instrument, not of the code.

By repository: official repository: 16 samples from 1 repository, 6 ran (6 with no instrument failure; 0 where Syntology's instrument failed); community (archive-listed): 8 samples from 1 repository, 7 ran (4 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 2 of its repositories that Syntology harvested for other papers:

Levels written between 2026-09-22 and 2026-09-25 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed. 1 of the 13 samples that ran has no recorded date.

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

google-research/noisystudent officialmentioned in papermentioned on GitHubtfApache-2.0Syntology: 16 samples, 6 ran (0 instrument-failed) · 10 unverified report
tensorflow/tpu officialmentioned in papertfApache-2.0Syntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Towards Understanding Why Label Smoothing Degrades Selective Classification and How to Fix It (arXiv:2403.14715; that paper's own run record: 9 ran (of which 2 constructed an object rather than computing a result; 6 with no instrument failure: 1 honoured, 0 violated, 5 with no contract checked; 3 where Syntology's instrument failed) · 1 unverified); nothing checks that they implement this paper's method. report
leondgarse/Keras_efficientnet_v2 mentioned on GitHubtfSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to EfficientNetV2: Smaller Models and Faster Training (arXiv:2104.00298; that paper's own run record: 55 ran (of which 14 constructed an object rather than computing a result; 43 with no instrument failure: 11 honoured, 0 violated, 32 with no contract checked; 12 where Syntology's instrument failed) · 24 unverified); nothing checks that they implement this paper's method. report
leondgarse/Keras_efficientnet_v2_test mentioned on GitHubtfApache-2.0Syntology: 8 samples, 7 ran (3 instrument-failed) · 1 unverified report
mhd-medfa/NoisyStudent-Based-Object-Recognition mentioned on GitHubtfMITSyntology: not harvested report
phamtrancsek12/offensive-identification mentioned on GitHubpytorchSyntology: not harvested report
stanleyjzheng/PyData mentioned on GitHubSyntology: not harvested report
tensorflow/tpu mentioned on GitHubtfSyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Towards Understanding Why Label Smoothing Degrades Selective Classification and How to Fix It (arXiv:2403.14715; that paper's own run record: 9 ran (of which 2 constructed an object rather than computing a result; 6 with no instrument failure: 1 honoured, 0 violated, 5 with no contract checked; 3 where Syntology's instrument failed) · 1 unverified); nothing checks that they implement this paper's method. report
thomasly/PaperTranslation mentioned on GitHubSyntology: not harvested report
xultaeculcis/coral-net mentioned on GitHubpytorchSyntology: not harvested report
yakhyo/EfficientNet-PyTorch 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

24 samples harvested: 13 ran (of which 0 constructed an object rather than computing a result; 10 with no instrument failure: 2 honoured, 0 violated, 8 with no contract checked; 3 where Syntology's instrument failed) · 11 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.

2ran · honoured contract
1ran · our draft was wrong
2ran; the contract check could not run
8ran
11unverified

Licence: 2 of the 24 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.

google-research/noisystudent official repository 16 samples, 6 ran (0 instrument-failed) · 10 unverified
efficientnet_params in 9 papers efficientnet_builder.py ran Apache-2.0 (permissive) · 00e16fdd1db60196 · report fixture: signature model_name:str/0/str · returned tuple · level recorded 2026-09-25
get_mean_std in 17 papers task_info.py ran Apache-2.0 (permissive) · ab69b327577ef8d4 · report fixture: signature task_name:str/0/str · returned tuple · level recorded 2026-09-25
get_num_eval_images task_info.py ran Apache-2.0 (permissive) · 5edab72c1f95b154 · report fixture: signature task_name:str/0/str · returned int · level recorded 2026-09-25
get_num_train_images task_info.py ran Apache-2.0 (permissive) · 8823cc7b394ad9d1 · report fixture: signature task_name:str/0/str · returned int · level recorded 2026-09-25
shuffle_split proc_svhn.py ran Apache-2.0 (permissive) · 0df948cabd4bdae3 · report fixture: signature data:dict/0/?, split:str/0/str · returned dict · level recorded 2026-09-25
solarize in 15 papers randaugment.py ran Apache-2.0 (permissive) · e40dd3a18c0fb8bc · report fixture: signature image:EagerTensor/3/int, threshold:int/0/int · returned EagerTensor · level recorded 2026-09-25
blend in 12 papers randaugment.py unverified Apache-2.0 (permissive) · 574c2e00bbe4667f · report
build_learning_rate utils.py unverified Apache-2.0 (permissive) · 53da5ea39ee2fa51 · report
build_optimizer in 83 papers utils.py unverified Apache-2.0 (permissive) · 7b88b9f2e3c23c21 · report
conv_kernel_initializer in 6 papers efficientnet_model.py unverified Apache-2.0 (permissive) · 4ae6bf83cccbb591 · report
cutout in 17 papers randaugment.py unverified Apache-2.0 (permissive) · d42d54a4353367b5 · report
dense_kernel_initializer in 5 papers efficientnet_model.py unverified Apache-2.0 (permissive) · e430f1fd246b1b51 · report
get_keep_image_idx filter_unlabel.py unverified Apache-2.0 (permissive) · aabbd55b6f0e8db6 · report
set_shapes predict_label.py unverified Apache-2.0 (permissive) · d8e0266f26c70d20 · report
small_image_crop preprocessing.py unverified Apache-2.0 (permissive) · 033941ec31eafc06 · report
stochastic_depth in 8 papers utils.py unverified Apache-2.0 (permissive) · c0942e1efe9598d2 · report
leondgarse/Keras_efficientnet_v2_test community (archive-listed) 8 samples, 7 ran (3 instrument-failed) · 1 unverified
batchnorm_with_activation keras_efficientnet_v2/efficientnet_v2.py ran · our draft was wrong Apache-2.0 (permissive) · e9ed9f7ced6189ee · report fixture: signature inputs:KerasTensor/0/float, activation:str/0/str, use_torch_eps:bool/0/bool, name:str/0/str · returned KerasTensor · level recorded 2026-09-22
conv2d_no_bias keras_efficientnet_v2/efficientnet_v2.py ran; the contract check could not run (RAISES) Apache-2.0 (permissive) · 96248e1c8a073860 · report fixture: signature inputs:EagerTensor/4/float, filters:int/0/int, kernel_size:int/0/int, strides:int/0/int, padding:str/0/str, use_torch_padding:bool/0/bool, name:str/0/str · returned EagerTensor · level recorded 2026-09-22
conv_kernel_initializer in 6 papers convert_effnetv2_model.py ran Apache-2.0 (permissive) · cafc76d8c85afb88 · report fixture: signature shape:tuple/1/int, dtype:DType/0/?, partition_info:NoneType/0/? · returned EagerTensor · level recorded 2026-09-24
dense_kernel_initializer in 5 papers convert_effnetv2_model.py ran Apache-2.0 (permissive) · 538d8c4191034b39 · report fixture: signature shape:tuple/1/int, dtype:DType/0/?, partition_info:NoneType/0/? · returned EagerTensor · level recorded 2026-09-24
round_filters in 47 papers convert_effnetv2_model.py ran · honoured contract Apache-2.0 (permissive) · 5640ec0612debaea · report fixture: signature filters:int/0/int, mconfig:MockMConfig/0/?, skip:bool/0/bool · returned int · level recorded, date not recorded
se_module keras_efficientnet_v2/efficientnet_v2.py ran; the contract check could not run (RAISES) fingerprinted Apache-2.0 (permissive) · b43c2e4c5402cf83 · report fixture: signature inputs:EagerTensor/4/float, se_ratio:int/0/int, name:str/0/str · returned EagerTensor · level recorded 2026-09-22
to_4d in 69 papers keras_efficientnet_v2/augment.py ran · honoured contract fingerprinted Apache-2.0 (permissive) here; the stored copy is pointer only · bd1c72a9d8556505 · report fixture: signature image:EagerTensor/2/float · returned EagerTensor · level recorded 2026-09-22
from_4d keras_efficientnet_v2/augment.py unverified Apache-2.0 (permissive) here; the stored copy is pointer only · b10fb8ac53c53a97 · 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 Noisy Student Training (graph method: Noisy Student Training). 1 brick; individually on a fixture: 0 ran · 1 unverified. The assembly has not been run end-to-end.

  1. reload_model_weights entry · level 0 · unverified · f734bb6fe9d256d2

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="1911.04252") returns these assemblies with each brick's role and wiring (how to connect).

Tasks

Data AugmentationGeneral ClassificationImage Classification

Results from the paper archive 2025-07-28

TaskDatasetModelMetricValueRank at snapshotLeaderboardReport
Image Classification ImageNet NoisyStudent (EfficientNet-L2) Hardware Burden 51800G #48 of 1060 Archive leaderboard report
Image Classification ImageNet NoisyStudent (EfficientNet-L2) Number of params 480M #48 of 1060 Archive leaderboard report
Image Classification ImageNet NoisyStudent (EfficientNet-L2) Top 1 Accuracy 88.4% #48 of 1060 Archive leaderboard report
Image Classification ImageNet NoisyStudent (EfficientNet-B7) GFLOPs 37 #118 of 1060 Archive leaderboard report
Image Classification ImageNet NoisyStudent (EfficientNet-B7) Number of params 66M #118 of 1060 Archive leaderboard report
Image Classification ImageNet NoisyStudent (EfficientNet-B7) Top 1 Accuracy 86.9% #118 of 1060 Archive leaderboard report
Image Classification ImageNet NoisyStudent (EfficientNet-B6) Number of params 43M #148 of 1060 Archive leaderboard report
Image Classification ImageNet NoisyStudent (EfficientNet-B6) Top 1 Accuracy 86.4% #148 of 1060 Archive leaderboard report
Image Classification ImageNet NoisyStudent (EfficientNet-B5) Number of params 30M #172 of 1060 Archive leaderboard report
Image Classification ImageNet NoisyStudent (EfficientNet-B5) Top 1 Accuracy 86.1% #172 of 1060 Archive leaderboard report
Image Classification ImageNet NoisyStudent (EfficientNet-B4) Number of params 19M #240 of 1060 Archive leaderboard report
Image Classification ImageNet NoisyStudent (EfficientNet-B4) Top 1 Accuracy 85.3% #240 of 1060 Archive leaderboard report
Image Classification ImageNet NoisyStudent (EfficientNet-B3) Number of params 12M #351 of 1060 Archive leaderboard report
Image Classification ImageNet NoisyStudent (EfficientNet-B3) Top 1 Accuracy 84.1% #351 of 1060 Archive leaderboard report
Image Classification ImageNet NoisyStudent (EfficientNet-B2) Number of params 9.2M #538 of 1060 Archive leaderboard report
Image Classification ImageNet NoisyStudent (EfficientNet-B2) Top 1 Accuracy 82.4% #538 of 1060 Archive leaderboard report
Image Classification ImageNet NoisyStudent (EfficientNet-B1) Number of params 7.8M #633 of 1060 Archive leaderboard report
Image Classification ImageNet NoisyStudent (EfficientNet-B1) Top 1 Accuracy 81.5% #633 of 1060 Archive leaderboard report
Image Classification ImageNet NoisyStudent (EfficientNet-B0) Number of params 5.3M #803 of 1060 Archive leaderboard report
Image Classification ImageNet NoisyStudent (EfficientNet-B0) Top 1 Accuracy 78.8% #803 of 1060 Archive leaderboard report
Image Classification ImageNet ReaL EfficientNet-L2 Accuracy 90.55% #16 of 57 Archive leaderboard report
Image Classification ImageNet ReaL EfficientNet-L2 Params 480M #16 of 57 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 ConvolutionAverage PoolingBatch NormalizationConvolutionDense ConnectionsDepthwise ConvolutionDepthwise Separable ConvolutionDropoutEfficientNetFixResInverted Residual BlockNoisy StudentPointwise ConvolutionRMSPropRandAugmentReLUSigmoid ActivationSqueeze-and-Excitation BlockStep DecayStochastic DepthTest

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