Data to 7 Oct 2026

Differentiable Augmentation for Data-Efficient GAN Training

arXiv 2006.10738 ↗First posted 18 Jun 2020cs.CV cs.GR cs.LGcited by 361

Not on the map. Why?

Cited by 59 in Diffusion models, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

The performance of generative adversarial networks (GANs) heavily deteriorates given a limited amount of training data. This is mainly because the discriminator is memorizing the exact training set. To combat it, we propose Differentiable Augmentation (DiffAugment), a simple method that improves the data efficiency of GANs by imposing various types of differentiable augmentations on both real and fake samples. Previous attempts to directly augment the training data manipulate the distribution of real images, yielding little benefit; DiffAugment enables us to adopt the differentiable augmentation for the generated samples, effectively stabilizes training, and leads to better convergence. Experiments demonstrate consistent gains of our method over a variety of GAN architectures and loss functions for both unconditional and class-conditional generation. With DiffAugment, we achieve a state-of-the-art FID of 6.80 with an IS of 100.8 on ImageNet 128x128 and 2-4x reductions of FID given 1,000 images on FFHQ and LSUN. Furthermore, with only 20% training data, we can match the top performance on CIFAR-10 and CIFAR-100. Finally, our method can generate high-fidelity images using only 100 images without pre-training, while being on par with existing transfer learning algorithms. Code is available at https://github.com/mit-han-lab/data-efficient-gans.

Key facts. Cited by 361 papers in the map (data through 7 Oct 2026). First posted on arXiv 18 Jun 2020.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 98; 2022: 85; 2023: 82; 2024: 48; 2025: 28; 2026: 20.

361 papers.

Sort:

By area:

2022Generative Adversarial Networkscited by 5,603
2022Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flowcited by 3,550
2022Multi-Concept Customization of Text-to-Image Diffusioncited by 1,099
2023Extracting Training Data from Diffusion Modelscited by 723
2021Tackling the Generative Learning Trilemma with Denoising Diffusion GANscited by 515
2023Scaling up GANs for Text-to-Image Synthesiscited by 492
2022StyleGAN-XL: Scaling StyleGAN to Large Diverse Datasetscited by 489
2022Dataset Distillation by Matching Training Trajectoriescited by 437
2021Dataset Condensation with Distribution Matchingcited by 362
2023Ablating Concepts in Text-to-Image Diffusion Modelscited by 312
2021Dataset Condensation with Differentiable Siamese Augmentationcited by 290
2023Effective Data Augmentation With Diffusion Modelscited by 282

38 more

2025Diffusion Transformers with Representation Autoencoderscited by 275
2021Deep Generative Modelling: A Comparative Review of VAEs, GANs, Normalizing Flows, Energy-Based and Autoregressive Modelscited by 251
2021Score-Based Generative Modeling with Critically-Damped Langevin Diffusioncited by 230
2021Large Scale Image Completion via Co-Modulated Generative Adversarial Networkscited by 219
2021A Survey of Human-in-the-loop for Machine Learningcited by 214
2023StyleGAN-T: Unlocking the Power of GANs for Fast Large-Scale Text-to-Image Synthesiscited by 212
2024Score identity Distillation: Exponentially Fast Distillation of Pretrained Diffusion Models for One-Step Generationcited by 211
2021Few-shot Image Generation via Cross-domain Correspondencecited by 206
2021Projected GANs Converge Fastercited by 204
2022Diffusion-GAN: Training GANs with Diffusioncited by 203
2022Generating Videos with Dynamics-aware Implicit Generative Adversarial Networkscited by 186
2022Scaling Up Dataset Distillation to ImageNet-1K with Constant Memorycited by 180
2022SinNeRF: Training Neural Radiance Fields on Complex Scenes from a Single Imagecited by 165
2021StyleGAN-NADA: CLIP-Guided Domain Adaptation of Image Generatorscited by 164
2021TransGAN: Two Pure Transformers Can Make One Strong GAN, and That Can Scale Upcited by 162
2021A Good Image Generator Is What You Need for High-Resolution Video Synthesiscited by 162
2021StyleSwin: Transformer-based GAN for High-resolution Image Generationcited by 152
2023Diverse Data Augmentation with Diffusions for Effective Test-time Prompt Tuningcited by 150
2022StyleGAN-Human: A Data-Centric Odyssey of Human Generationcited by 143
2021Towards Faster and Stabilized GAN Training for High-fidelity Few-shot Image Synthesiscited by 140
2023Patch Diffusion: Faster and More Data-Efficient Training of Diffusion Modelscited by 131
2021Unconstrained Scene Generation with Locally Conditioned Radiance Fieldscited by 130
2023Generalizing Dataset Distillation via Deep Generative Priorcited by 125
2021Instance-Conditioned GANcited by 119
2022Wavelet Diffusion Models are fast and scalable Image Generatorscited by 116
2021Robust Learning Meets Generative Models: Can Proxy Distributions Improve Adversarial Robustness?cited by 113
2022Generating Long Videos of Dynamic Scenescited by 113
2022Efficient Dataset Distillation Using Random Feature Approximationcited by 111
2021ViTGAN: Training GANs with Vision Transformerscited by 110
2023AnomalyDiffusion: Few-Shot Anomaly Image Generation with Diffusion Modelcited by 110
2021Regularizing Generative Adversarial Networks under Limited Datacited by 105
2021Recent advances and clinical applications of deep learning in medical image analysiscited by 104
2023DREAM: Efficient Dataset Distillation by Representative Matchingcited by 93
2023Minimizing Trajectory Curvature of ODE-based Generative Modelscited by 87
2021Ensembling Off-the-shelf Models for GAN Trainingcited by 84
2024Distilling Diffusion Models into Conditional GANscited by 78
2021Defect-GAN: High-Fidelity Defect Synthesis for Automated Defect Inspectioncited by 77
2022Synthesizing Informative Training Samples with GANcited by 77

50 most cited of 361 All 361 (JSON)