</> Code with Papers by Syntology

Datasets › ImageNet

ImageNet

Introduced in ImageNet: A large-scale hierarchical image database1 Jan 2009 archive 2025-07-28

The ImageNet dataset contains 14,197,122 annotated images according to the WordNet hierarchy. Since 2010 the dataset is used in the ImageNet Large Scale Visual Recognition Challenge (ILSVRC), a benchmark in image classification and object detection. The publicly released dataset contains a set of manually annotated training images. A set of test images is also released, with the manual annotations withheld. ILSVRC annotations fall into one of two categories: (1) image-level annotation of a binary label for the presence or absence of an object class in the image, e.g., “there are cars in this image” but “there are no tigers,” and (2) object-level annotation of a tight bounding box and class label around an object instance in the image, e.g., “there is a screwdriver centered at position (20,25) with width of 50 pixels and height of 30 pixels”. The ImageNet project does not own the copyright of the images, therefore only thumbnails and URLs of images are provided.

  • Total number of non-empty WordNet synsets: 21841
  • Total number of images: 14197122
  • Number of images with bounding box annotations: 1,034,908
  • Number of synsets with SIFT features: 1000
  • Number of images with SIFT features: 1.2 million

Source: ImageNet Large Scale Visual Recognition Challenge Image Source: https://cs.stanford.edu/people/karpathy/cnnembed/

HomepageBenchmarksPapers JSON

Benchmarks archive 2025-07-28

All 52 leaderboards whose dataset resolves to this page shown (sort by any header). "First row" is the archive's own first row at snapshot, in the archive's row order; nothing here re-ranks and metric direction is not asserted.

First row (archive order)PaperCode
Image Classification ImageNet CoCa (finetuned) Top 1 Accuracy 91.0% CoCa: Contrastive Captioners are Image-Text Foundation Models mlfoundations/open_clip +5 1,060 Compare
Self-Supervised Image Classification ImageNet DINOv2+reg (ViT-g/14) Top 1 Accuracy 87.1 Vision Transformers Need Registers rwightman/pytorch-image-models +5 144 Compare
Neural Architecture Search ImageNet DeepMAD-50M Top-1 Error Rate 16.1 DeepMAD: Mathematical Architecture Design for Deep... alibaba/lightweight-neural-architecture-search 135 Compare
Semi-Supervised Image Classification ImageNet - 10% labeled data DHO (ViT-Large) Top 1 Accuracy 85.9% Simple Semi-supervised Knowledge Distillation from... erjui/DHO 75 Compare
Self-Supervised Image Classification ImageNet (finetuned) DINOv2 (ViT-g/14, 448) Top 1 Accuracy 88.9% DINOv2: Learning Robust Visual Features without Supervision huggingface/transformers +25 65 Compare
Semi-Supervised Image Classification ImageNet - 1% labeled data DHO (ViT-Large) Top 1 Accuracy 84.6% Simple Semi-supervised Knowledge Distillation from... erjui/DHO 65 Compare
Knowledge Distillation ImageNet ScaleKD (T:BEiT-L S:ViT-B/14) Top-1 accuracy % 86.43 ScaleKD: Strong Vision Transformers Could Be Excellent Teachers deep-optimization/scalekd 52 Compare
Image Classification ImageNet V2 Model soups (BASIC-L) Top 1 Accuracy 84.63 Model soups: averaging weights of multiple fine-tuned... mlfoundations/model-soups +5 33 Compare
Quantization ImageNet FQ-ViT (ViT-L) Top-1 Accuracy (%) 85.03 FQ-ViT: Post-Training Quantization for Fully Quantized... megvii-research/FQ-ViT 27 Compare
Zero-Shot Transfer Image Classification ImageNet M2-Encoder Param 10B M2-Encoder: Advancing Bilingual Image-Text Understanding... alipay/Ant-Multi-Modal-Framework 23 Compare
Data Augmentation ImageNet DeiT-B (+MixPro) Accuracy (%) 82.9 MixPro: Data Augmentation with MaskMix and Progressive... fistyee/mixpro 17 Compare
Network Pruning ImageNet ResNet50-2.3 GFLOPs Accuracy 78.79 Pruning Filters for Efficient ConvNets PaddlePaddle/PaddleOCR +20 16 Compare
Image Reconstruction ImageNet MGVQ (16x16x8) FID 0.49 MGVQ: Could VQ-VAE Beat VAE? A Generalizable Tokenizer... MKJia/MGVQ 15 Compare
Prompt Engineering ImageNet PromptKD Harmonic mean 77.62 PromptKD: Unsupervised Prompt Distillation for... zhengli97/promptkd 15 Compare
Contrastive Learning imagenet-1k ResNet50 ImageNet Top-1 Accuracy 73.6 Matrix Information Theory for Self-Supervised Learning yifanzhang-pro/matrix-ssl +2 14 Compare
Zero-Shot Transfer Image Classification ImageNet V2 BASIC (Lion) Accuracy (Private) 81.2 — — 13 Compare
Image Clustering ImageNet TURTLE (CLIP + DINOv2) Accuracy 72.9 Let Go of Your Labels with Unsupervised Transfer mlbio-epfl/turtle 12 Compare
Model Compression ImageNet ADLIK-MO-ResNet50+W4A4 Top-1 77.878 Learned Step Size Quantization zhutmost/lsq-net +8 12 Compare
Zero-Shot Composed Image Retrieval (ZS-CIR) ImageNet iSEARLE-XL (CLIP L/14) Average Recall 24.46 iSEARLE: Improving Textual Inversion for Zero-Shot... miccunifi/searle +1 11 Compare
Sparse Learning ImageNet Resnet-50: 80% Sparse Top-1 Accuracy 77.1 Rigging the Lottery: Making All Tickets Winners google-research/rigl +10 9 Compare
Unsupervised Image Classification ImageNet TURTLE (CLIP + DINOv2) Accuracy (%) 72.9 Let Go of Your Labels with Unsupervised Transfer mlbio-epfl/turtle 9 Compare
Feature Upsampling ImageNet JAFAR ADCC 73.3 JAFAR: Jack up Any Feature at Any Resolution PaulCouairon/JAFAR 8 Compare
Few-Shot Image Classification ImageNet - 1-shot ViT-MoE-15B (Every-2) Top 1 Accuracy 68.66 Scaling Vision with Sparse Mixture of Experts google-research/vmoe 8 Compare
Few-Shot Image Classification ImageNet - 5-shot ViT-MoE-15B (Every-2) Top 1 Accuracy 82.78 Scaling Vision with Sparse Mixture of Experts google-research/vmoe 8 Compare
Prompt Engineering ImageNet V2 HPT++ Top-1 accuracy % 65.31 HPT++: Hierarchically Prompting Vision-Language Models... vill-lab/2024-aaai-hpt +1 8 Compare
Few-Shot Image Classification ImageNet - 10-shot MAWS (ViT-6.5B) Top 1 Accuracy 84.6 The effectiveness of MAE pre-pretraining for... facebookresearch/maws 7 Compare
Image Super-Resolution ImageNet DAVI FID 36.27 Diffusion Prior-Based Amortized Variational Inference... mlvlab/davi +1 6 Compare
JPEG Decompression ImageNet Palette (QF: 20) FID-5K 4.3 Palette: Image-to-Image Diffusion Models Janspiry/Palette-Image-to-Image-Diffusion-Models +4 6 Compare
Weakly-Supervised Object Localization ImageNet Stable diffusion GT-known localization accuracy 75.0 Generative Prompt Model for Weakly Supervised Object Localization callsys/genpromp 6 Compare
Few-Shot Image Classification ImageNet - 0-Shot DebiasPL (ResNet50) Accuracy 68.3% Debiased Learning from Naturally Imbalanced Pseudo-Labels frank-xwang/debiased-pseudo-labeling 5 Compare
Image Inpainting ImageNet WavePaint FID 3.21 WavePaint: Resource-efficient Token-mixer for... pranavphoenix/WavePaint 5 Compare
Adversarial Robustness ImageNet ResNet-50 (SGD, Cosine) Accuracy 77.4 Are Transformers More Robust Than CNNs? ytongbai/ViTs-vs-CNNs 4 Compare
Image Classification with Differential Privacy ImageNet NFResnet-50 Top 1 Accuracy 39.2 TAN Without a Burn: Scaling Laws of DP-SGD facebookresearch/tan 4 Compare
Image Colorization ImageNet DDRM Consistency 260.4 Zero-Shot Image Restoration Using Denoising Diffusion... wyhuai/ddnm +3 4 Compare
Weakly Supervised Object Detection ImageNet PCL-OB-G-Ens + FRCNN MAP 19.6 PCL: Proposal Cluster Learning for Weakly Supervised... ppengtang/pcl.pytorch +3 4 Compare
Adversarial Defense ImageNet ResNet101 Accuracy 99.8% NOMARO: Defending against Adversarial Attacks by... as791/NOMARO_defense 3 Compare
Image Deblurring ImageNet DDNM FID 1.15 Zero-Shot Image Restoration Using Denoising Diffusion... wyhuai/ddnm +3 3 Compare
Semi-Supervised Image Classification ImageNet - 0.2% labeled data DebiasPL (ResNet-50) ImageNet Top-1 Accuracy 69.6% Debiased Learning from Naturally Imbalanced Pseudo-Labels frank-xwang/debiased-pseudo-labeling 3 Compare
Color Image Denoising ImageNet sigma100 DMID-p LPIPS 0.156 Stimulating Diffusion Model for Image Denoising via... li-tong-621/dmid 2 Compare
Color Image Denoising ImageNet sigma150 DMID-p LPIPS 0.259 Stimulating Diffusion Model for Image Denoising via... li-tong-621/dmid 2 Compare
Color Image Denoising ImageNet sigma200 DMID-p LPIPS 0.259 Stimulating Diffusion Model for Image Denoising via... li-tong-621/dmid 2 Compare
Color Image Denoising ImageNet sigma250 DMID-p LPIPS 0.289 Stimulating Diffusion Model for Image Denoising via... li-tong-621/dmid 2 Compare
Color Image Denoising ImageNet sigma50 DMID-p LPIPS 0.087 Stimulating Diffusion Model for Image Denoising via... li-tong-621/dmid 2 Compare
Medical Image Classification ImageNet DaViT-T GFLOPs 4.5 DaViT: Dual Attention Vision Transformers rwightman/pytorch-image-models +3 2 Compare
Zero-Shot Learning ImageNet ZLaP Top 1 Accuracy 72.1 Label Propagation for Zero-shot Classification with... vladan-stojnic/zlap 2 Compare
Image Classification imagenet-1k BinaryViT Top 1 Accuracy 70.6 BinaryViT: Pushing Binary Vision Transformers Towards... phuoc-hoan-le/binaryvit 1 Compare
Image Clustering imagenet-1k TAC ARI 0.435 Image Clustering with External Guidance xlearning-scu/2024-icml-tac 1 Compare
Image Segmentation ImageNet MobileOne-S0 GFLOPs 0.275 MobileOne: An Improved One millisecond Mobile Backbone rwightman/pytorch-image-models +9 1 Compare
Transductive Zero-Shot Classification ImageNet ZLaP Top 1 Accuracy 72.7 Label Propagation for Zero-shot Classification with... vladan-stojnic/zlap 1 Compare
Visual Question Answering (VQA) ImageNet BLIP-2 OPT ClipMatch@1 57.10 Open-ended VQA benchmarking of Vision-Language models by... lmb-freiburg/ovqa 1 Compare
Classification imagenet-1k no rows — — 0 Compare
Semi-Supervised Image Classification ImageNet no rows — — 0 Compare

Papers archive 2025-07-28

30 shown of 630 papers with a leaderboard row on this dataset's benchmarks, newest first. The archive's own "papers using this dataset" list was never published, so this is the benchmark-backed subset; the archive's count for this dataset is 15,430. The Syntology column is from Syntology's graph (read 2026-09-24), stated per sample; it is not part of any archive number.

DateSamples run Syntology
MGVQ: Could VQ-VAE Beat VAE? A Generalizable Tokenizer with Multi-group Quantization 1 2 14 Jul 2025 not harvested
JAFAR: Jack up Any Feature at Any Resolution 1 1 10 Jun 2025 not harvested
SST: Self-training with Self-adaptive Thresholding for Semi-supervised Learning 0 10 31 May 2025 not harvested
MMRL++: Parameter-Efficient and Interaction-Aware Representation Learning for Vision-Language Models 1 1 15 May 2025 not harvested
Simple Semi-supervised Knowledge Distillation from Vision-Language Models via 𝐃ual-𝐇ead 𝐎ptimization 1 4 12 May 2025 not harvested
GigaTok: Scaling Visual Tokenizers to 3 Billion Parameters for Autoregressive Image Generation 1 1 11 Apr 2025 ran 10 of 15 samples (5 unverified; 15 pointer-only for licence)
Enhanced OoD Detection through Cross-Modal Alignment of Multi-Modal Representations 1 1 24 Mar 2025 not harvested
Exploring a Principled Framework for Deep Subspace Clustering 1 1 21 Mar 2025 ran 4 of 4 samples (0 unverified; 4 pointer-only for licence)
MMRL: Multi-Modal Representation Learning for Vision-Language Models 1 2 11 Mar 2025 ran 0 of 1 samples (1 unverified)
Compress image to patches for Vision Transformer 1 1 14 Feb 2025 not harvested
Polynomial, trigonometric, and tropical activations 1 1 3 Feb 2025 not harvested
SpectralKD: A Unified Framework for Interpreting and Distilling Vision Transformers via Spectral Analysis 1 3 26 Dec 2024 not harvested
Mamba2D: A Natively Multi-Dimensional State-Space Model for Vision Tasks 1 1 20 Dec 2024 not harvested
Preventing Local Pitfalls in Vector Quantization via Optimal Transport 1 2 19 Dec 2024 not harvested
Taming Scalable Visual Tokenizer for Autoregressive Image Generation 2 1 3 Dec 2024 ran 9 of 13 samples (4 unverified; 3 pointer-only for licence)
On the Performance Analysis of Momentum Method: A Frequency Domain Perspective 1 2 29 Nov 2024 ran 1 of 2 samples (1 unverified)
Multimodal Autoregressive Pre-training of Large Vision Encoders 1 6 21 Nov 2024 ran 0 of 3 samples (3 unverified; 3 pointer-only for licence)
ScaleKD: Strong Vision Transformers Could Be Excellent Teachers 1 5 11 Nov 2024 not harvested
Performance of Gaussian Mixture Model Classifiers on Embedded Feature Spaces 1 1 17 Oct 2024 not harvested
Stabilize the Latent Space for Image Autoregressive Modeling: A Unified Perspective 1 1 16 Oct 2024 ran 3 of 5 samples (2 unverified)
DeBiFormer: Vision Transformer with Deformable Agent Bi-level Routing Attention 1 3 11 Oct 2024 not harvested
SynCo: Synthetic Hard Negatives in Contrastive Learning for Better Unsupervised Visual Representations 1 4 3 Oct 2024 not harvested
HVT: A Comprehensive Vision Framework for Learning in Non-Euclidean Space 1 3 25 Sep 2024 not harvested
MaskBit: Embedding-free Image Generation via Bit Tokens 1 1 24 Sep 2024 ran 8 of 9 samples (1 unverified)
Kolmogorov-Arnold Transformer 1 3 16 Sep 2024 ran 3 of 4 samples (1 unverified)
Open-MAGVIT2: An Open-Source Project Toward Democratizing Auto-regressive Visual Generation 2 1 6 Sep 2024 ran 3 of 3 samples (0 unverified; 1 pointer-only for licence)
HPT++: Hierarchically Prompting Vision-Language Models with Multi-Granularity Knowledge Generation and Improved Structure Modeling 2 2 27 Aug 2024 not harvested
CAS-ViT: Convolutional Additive Self-attention Vision Transformers for Efficient Mobile Applications 2 4 7 Aug 2024 ran 4 of 6 samples (2 unverified)
Unsupervised Representation Learning by Balanced Self Attention Matching 1 2 4 Aug 2024 not harvested
Diffusion Prior-Based Amortized Variational Inference for Noisy Inverse Problems 2 1 23 Jul 2024 ran 21 of 27 samples (6 unverified; 5 pointer-only for licence)

The full list of 630 is in the JSON twin.

Dataset loaders archive 2025-07-28

huggingface.co/datasets/coallaoh/ImageNet-ABhuggingface/datasetstfpytorchjax
huggingface.co/datasets/slegroux/tiny-imagenet-200-cleanhuggingface/datasetstfpytorchjax
huggingface.co/datasets/evanarlian/imagenet_1k_resized_256huggingface/datasetstfpytorchjax
huggingface.co/datasets/imagenet-1khuggingface/datasetstfpytorchjax
huggingface.co/datasets/Maysee/tiny-imagenethuggingface/datasetstfpytorchjax
huggingface.co/datasets/zh-plus/tiny-imagenethuggingface/datasetstfpytorchjax
pytorch.org/vision/stable/generated/torchvision.datasets.ImageNet.htmlpytorch/visionpytorch
docs.voxel51.com/user_guide/dataset_zoo/datasets.html#imagenet-2012voxel51/fiftyonetfpytorch
docs.activeloop.ai/datasets/imagenet-datasetactiveloopai/Hubtfpytorch
www.tensorflow.org/datasets/catalog/imagenet2012tensorflow/datasetstfjax
open-mmlab/mmclassificationopen-mmlab/mmclassificationpytorch

11 loaders as listed in the archive; links are outbound and not re-checked here.

Tasks archive 2025-07-28

Image ClassificationClassificationImage GenerationZero-Shot LearningVisual Question Answering (VQA)Few-Shot Image ClassificationImage Super-ResolutionColor Image DenoisingFew-Shot LearningSemi-Supervised Image ClassificationImage ClusteringImage ReconstructionNeural Architecture SearchTransductive Zero-Shot ClassificationWeakly Supervised Object DetectionImage InpaintingBinarizationZero-Shot Transfer Image ClassificationPrompt EngineeringObject RecognitionImage SegmentationAdversarial DefenseQuantizationMedical Image ClassificationZero-Shot Composed Image Retrieval (ZS-CIR)Knowledge DistillationImage DeblurringWeakly-Supervised Object LocalizationUnsupervised Image ClassificationAdversarial RobustnessNetwork PruningImage Compressed SensingData AugmentationContrastive LearningModel CompressionSparse LearningSelf-Supervised Image ClassificationClassification with Binary Neural NetworkImage ColorizationJPEG DecompressionImage Classification with Differential PrivacyZero-Shot Transfer Image Classification (CN)Feature Upsampling

License archive 2025-07-28

Custom (research, non-commercial)

Modalities archive 2025-07-28

Images

Languages archive 2025-07-28

Chinese

Variants archive 2025-07-28

  • imagenet-1k
  • ImageNet sigma50
  • ImageNet sigma250
  • ImageNet sigma200
  • ImageNet sigma150
  • ImageNet sigma100
  • ImageNet100
  • ImageNet (linear)
  • ImageNet (finetuned)
  • ImageNet - 5-shot
  • ImageNet - 5 labeled data per class
  • ImageNet - 2 labeled data per class
  • ImageNet - 1-shot
  • ImageNet - 1 labeled data per class
  • ImageNet - 10-shot
  • ImageNet - 0.2% labeled data
  • ImageNetV2
  • ImageNet V2
  • ImageNet - 10% labeled data
  • ImageNet - 1% labeled data
  • ImageNet - 0-Shot
  • ImageNet

22 variant names, as the archive lists them.

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