mlsanityData to 14 Oct 2026

Fake it till you make it: Learning transferable representations from synthetic ImageNet clones

arXiv 2212.08420 ↗First posted 16 Dec 2022cs.CV cs.LGcited by 172

On the map: Diffusion models → Image generation, text-to-image and editing

Abstract (arXiv)

Recent image generation models such as Stable Diffusion have exhibited an impressive ability to generate fairly realistic images starting from a simple text prompt. Could such models render real images obsolete for training image prediction models? In this paper, we answer part of this provocative question by investigating the need for real images when training models for ImageNet classification. Provided only with the class names that have been used to build the dataset, we explore the ability of Stable Diffusion to generate synthetic clones of ImageNet and measure how useful these are for training classification models from scratch. We show that with minimal and class-agnostic prompt engineering, ImageNet clones are able to close a large part of the gap between models produced by synthetic images and models trained with real images, for the several standard classification benchmarks that we consider in this study. More importantly, we show that models trained on synthetic images exhibit strong generalization properties and perform on par with models trained on real data for transfer. Project page: https://europe.naverlabs.com/imagenet-sd/

Key facts. Cited by 172 papers in the map (data through 14 Oct 2026). Placed in Image generation, text-to-image and editing, one of 14 directions in Diffusion models. Builds on 68 of its 116 references that are in the map. First posted on arXiv 16 Dec 2022.

Builds on

68 of 116 references.

Closest ancestors · 1

Same direction, newest first

  1. 2022Diffusion Art or Digital Forgery? Investigating Data Replication in Diffusion Models7 Dec

Other references · 67

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2020Denoising Diffusion Probabilistic Modelscited by 21,076
  3. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,305
  4. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,863
  5. 2015Distilling the Knowledge in a Neural Networkcited by 10,324
  6. 2014ImageNet Large Scale Visual Recognition Challengecited by 8,498
  7. 2021Emerging Properties in Self-Supervised Vision Transformerscited by 6,968
  8. 2022Classifier-Free Diffusion Guidancecited by 5,847
  9. 2021Zero-Shot Text-to-Image Generationcited by 4,288
  10. 2014Going Deeper with Convolutionscited by 4,241
  11. 2022LAION-5B: An open large-scale dataset for training next generation image-text modelscited by 4,152
  12. 2021Scaling Up Visual and Vision-Language Representation Learning With Noisy Text Supervisioncited by 3,864
38 more
  1. 2018Large Scale GAN Training for High Fidelity Natural Image Synthesiscited by 2,626
  2. 2019Optuna: A Next-generation Hyperparameter Optimization Frameworkcited by 2,501
  3. 2013Describing Textures in the Wildcited by 2,125
  4. 2020The Many Faces of Robustness: A Critical Analysis of Out-of-Distribution Generalizationcited by 1,850
  5. 2013Fine-Grained Visual Classification of Aircraftcited by 1,726
  6. 2019Do ImageNet Classifiers Generalize to ImageNet?cited by 1,515
  7. 2017EuroSAT: A Novel Dataset and Deep Learning Benchmark for Land Use and Land Cover Classificationcited by 1,496
  8. 2019Natural Adversarial Examplescited by 1,408
  9. 2018Continual Lifelong Learning with Neural Networks: A Reviewcited by 1,330
  10. 2019Contrastive Multiview Codingcited by 1,301
  11. 2019AMASS: Archive of Motion Capture as Surface Shapescited by 1,236
  12. 2015FlowNet: Learning Optical Flow with Convolutional Networkscited by 1,059
  13. 2019Learning Robust Global Representations by Penalizing Local Predictive Powercited by 1,013
  14. 2016Playing for Data: Ground Truth from Computer Gamescited by 818
  15. 2018Do Better ImageNet Models Transfer Better?cited by 565
  16. 2017Zero-Shot Learning -- A Comprehensive Evaluation of the Good, the Bad and the Uglycited by 538
  17. 2020Dataset Condensation with Gradient Matchingcited by 516
  18. 2022Dataset Distillation by Matching Training Trajectoriescited by 442
  19. 2019Dreaming to Distill: Data-free Knowledge Transfer via DeepInversioncited by 423
  20. 2014Understanding Deep Image Representations by Inverting Themcited by 403
  21. 2021ResNet strikes back: An improved training procedure in timmcited by 368
  22. 2017Learning from Synthetic Humanscited by 359
  23. 2021Multimodal datasets: misogyny, pornography, and malignant stereotypescited by 327
  24. 2022Is synthetic data from generative models ready for image recognition?cited by 320
  25. 2022Easily Accessible Text-to-Image Generation Amplifies Demographic Stereotypes at Large Scalecited by 262
  26. 2019On the "steerability" of generative adversarial networkscited by 248
  27. 2019Data-Free Learning of Student Networkscited by 204
  28. 2016What makes ImageNet good for transfer learning?cited by 192
  29. 2020Learning Diverse and Discriminative Representations via the Principle of Maximal Coding Rate Reductioncited by 142
  30. 2020Robust and Generalizable Visual Representation Learning via Random Convolutionscited by 142
  31. 2020Towards Accountability for Machine Learning Datasets: Practices from Software Engineering and Infrastructurecited by 138
  32. 2019Classification Accuracy Score for Conditional Generative Modelscited by 136
  33. 2018Learning Semantic Segmentation from Synthetic Data: A Geometrically Guided Input-Output Adaptation Approachcited by 119
  34. 2020Image Representations Learned With Unsupervised Pre-Training Contain Human-like Biasescited by 107
  35. 2021Generative Models as a Data Source for Multiview Representation Learningcited by 104
  36. 2021Pre-training without Natural Imagescited by 100
  37. 2022BigDatasetGAN: Synthesizing ImageNet with Pixel-wise Annotationscited by 99
  38. 2020Why Do Better Loss Functions Lead to Less Transferable Features?cited by 73

50 most cited of 67 All 67 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2022: 1; 2023: 47; 2024: 56; 2025: 50; 2026: 18.

172 papers.

Sort:

By area:

  1. 2023DreamSim: Learning New Dimensions of Human Visual Similarity using Synthetic Datacited by 337
  2. 2023Self-Consuming Generative Models Go MADcited by 250
  3. 2023StableRep: Synthetic Images from Text-to-Image Models Make Strong Visual Representation Learnerscited by 177
  4. 2023Dataset Diffusion: Diffusion-based Synthetic Dataset Generation for Pixel-Level Semantic Segmentationcited by 112
  5. 2023Scaling Laws of Synthetic Images for Model Training ... for Nowcited by 108
  6. 2023Synthetic Experience Replaycited by 101
  7. 2023Efficient Dataset Distillation via Minimax Diffusioncited by 83
  8. 2023Dream the Impossible: Outlier Imagination with Diffusion Modelscited by 77
  9. 2023Towards Unified Text-based Person Retrieval: A Large-scale Multi-Attribute and Language Search Benchmarkcited by 75
  10. 2024No "Zero-Shot" Without Exponential Data: Pretraining Concept Frequency Determines Multimodal Model Performancecited by 75
  11. 2023Learning Vision from Models Rivals Learning Vision from Datacited by 71
  12. 2024SynthCLIP: Are We Ready for a Fully Synthetic CLIP Training?cited by 64
38 more
  1. 2023FreeMask: Synthetic Images with Dense Annotations Make Stronger Segmentation Modelscited by 61
  2. 2023Training on Thin Air: Improve Image Classification with Generated Datacited by 50
  3. 2023Generative Models: What Do They Know? Do They Know Things? Let's Find Out!cited by 50
  4. 2023DGInStyle: Domain-Generalizable Semantic Segmentation with Image Diffusion Models and Stylized Semantic Controlcited by 45
  5. 2023StyleGAN knows Normal, Depth, Albedo, and Morecited by 44
  6. 2024Is Synthetic Data all We Need? Benchmarking the Robustness of Models Trained with Synthetic Imagescited by 41
  7. 2023Farm3D: Learning Articulated 3D Animals by Distilling 2D Diffusioncited by 38
  8. 2024TIPS: Text-Image Pretraining with Spatial awarenesscited by 37
  9. 2023Real-Fake: Effective Training Data Synthesis Through Distribution Matchingcited by 36
  10. 2024A Vision Check-up for Language Modelscited by 36
  11. 2023Fill-Up: Balancing Long-Tailed Data with Generative Modelscited by 34
  12. 2023Collaborating Foundation Models for Domain Generalized Semantic Segmentationcited by 34
  13. 2024Enhance Image Classification via Inter-Class Image Mixup with Diffusion Modelcited by 34
  14. 2023Feedback-guided Data Synthesis for Imbalanced Classificationcited by 33
  15. 2023Do text-free diffusion models learn discriminative visual representations?cited by 33
  16. 2024SubjectDrive: Scaling Generative Data in Autonomous Driving via Subject Controlcited by 32
  17. 2023Class-Incremental Learning using Diffusion Model for Distillation and Replaycited by 31
  18. 2023PUG: Photorealistic and Semantically Controllable Synthetic Data for Representation Learningcited by 31
  19. 2024Video Diffusion Models: A Surveycited by 30
  20. 2024DataDream: Few-shot Guided Dataset Generationcited by 29
  21. 2024UNIC: Universal Classification Models via Multi-teacher Distillationcited by 27
  22. 2023Reliability in Semantic Segmentation: Can We Use Synthetic Data?cited by 24
  23. 2024DreamDA: Generative Data Augmentation with Diffusion Modelscited by 24
  24. 2024Would Deep Generative Models Amplify Bias in Future Models?cited by 24
  25. 2024Dataset Enhancement with Instance-Level Augmentationscited by 23
  26. 2023Improving Adversarial Transferability by Stable Diffusioncited by 22
  27. 2023Diversify, Don't Fine-Tune: Scaling Up Visual Recognition Training with Synthetic Imagescited by 20
  28. 2024The Unmet Promise of Synthetic Training Images: Using Retrieved Real Images Performs Bettercited by 19
  29. 2023Not Only Generative Art: Stable Diffusion for Content-Style Disentanglement in Art Analysiscited by 18
  30. 2022Not Just Pretty Pictures: Toward Interventional Data Augmentation Using Text-to-Image Generatorscited by 17
  31. 2024ODGEN: Domain-specific Object Detection Data Generation with Diffusion Modelscited by 17
  32. 2024ViPer: Visual Personalization of Generative Models via Individual Preference Learningcited by 17
  33. 2025Improving the Scaling Laws of Synthetic Data with Deliberate Practicecited by 17
  34. 2023HiPA: Enabling One-Step Text-to-Image Diffusion Models via High-Frequency-Promoting Adaptationcited by 16
  35. 2024FedBiP: Heterogeneous One-Shot Federated Learning with Personalized Latent Diffusion Modelscited by 16
  36. 2024FreeTumor: Advance Tumor Segmentation via Large-Scale Tumor Synthesiscited by 15
  37. 2024SELMA: Learning and Merging Skill-Specific Text-to-Image Experts with Auto-Generated Datacited by 14
  38. 2024GLASS: Guided Latent Slot Diffusion for Object-Centric Learningcited by 14

50 most cited of 172 All 172 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2022Evaluating a Synthetic Image Dataset Generated with Stable Diffusion
  2. 2023Synthetic Data from Diffusion Models Improves ImageNet Classification
  3. 2023Diversity and Diffusion: Observations on Synthetic Image Distributions with Stable Diffusion
  4. 2025Augmented Conditioning Is Enough For Effective Training Image Generation
  5. 2025Do We Need All the Synthetic Data? Targeted Image Augmentation via Diffusion Models
  6. 2026Representation-Conditioned Diffusion Models for Guided Training Data Generation
  7. 2024On Improved Conditioning Mechanisms and Pre-training Strategies for Diffusion Models
  8. 2024Diffusion Curriculum: Synthetic-to-Real Data Curriculum via Image-Guided Diffusion