mlsanityData to 7 Oct 2026

NVAE: A Deep Hierarchical Variational Autoencoder

arXiv 2007.03898 ↗First posted 8 Jul 2020stat.ML cs.CV cs.LGcited by 635

Not on the map. Why?

Cited by 149 in Diffusion models, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

Normalizing flows, autoregressive models, variational autoencoders (VAEs), and deep energy-based models are among competing likelihood-based frameworks for deep generative learning. Among them, VAEs have the advantage of fast and tractable sampling and easy-to-access encoding networks. However, they are currently outperformed by other models such as normalizing flows and autoregressive models. While the majority of the research in VAEs is focused on the statistical challenges, we explore the orthogonal direction of carefully designing neural architectures for hierarchical VAEs. We propose Nouveau VAE (NVAE), a deep hierarchical VAE built for image generation using depth-wise separable convolutions and batch normalization. NVAE is equipped with a residual parameterization of Normal distributions and its training is stabilized by spectral regularization. We show that NVAE achieves state-of-the-art results among non-autoregressive likelihood-based models on the MNIST, CIFAR-10, CelebA 64, and CelebA HQ datasets and it provides a strong baseline on FFHQ. For example, on CIFAR-10, NVAE pushes the state-of-the-art from 2.98 to 2.91 bits per dimension, and it produces high-quality images on CelebA HQ. To the best of our knowledge, NVAE is the first successful VAE applied to natural images as large as 256×256 pixels. The source code is available at https://github.com/NVlabs/NVAE .

Key facts. Cited by 635 papers in the map (data through 7 Oct 2026). First posted on arXiv 8 Jul 2020.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 121; 2022: 128; 2023: 146; 2024: 113; 2025: 89; 2026: 38.

635 papers.

Sort:

By area:

  1. 2021High-Resolution Image Synthesis with Latent Diffusion Modelscited by 18,694
  2. 2021Diffusion Models Beat GANs on Image Synthesiscited by 8,185
  3. 2022Hierarchical Text-Conditional Image Generation with CLIP Latentscited by 6,539
  4. 2022Flow Matching for Generative Modelingcited by 5,426
  5. 2021Improved Denoising Diffusion Probabilistic Modelscited by 3,736
  6. 2021Image Super-Resolution via Iterative Refinementcited by 1,506
  7. 2021Palette: Image-to-Image Diffusion Modelscited by 1,296
  8. 2023Mastering Diverse Domains through World Modelscited by 1,193
  9. 2021Variational Diffusion Modelscited by 1,145
  10. 2022Diffusion Models: A Comprehensive Survey of Methods and Applicationscited by 1,107
  11. 2022MaskGIT: Masked Generative Image Transformercited by 1,082
  12. 2021Maximum Likelihood Training of Score-Based Diffusion Modelscited by 670
38 more
  1. 2021Vector-quantized Image Modeling with Improved VQGANcited by 656
  2. 2021VideoGPT: Video Generation using VQ-VAE and Transformerscited by 651
  3. 2021Score-based Generative Modeling in Latent Spacecited by 634
  4. 2021ILVR: Conditioning Method for Denoising Diffusion Probabilistic Modelscited by 634
  5. 2023NExT-GPT: Any-to-Any Multimodal LLMcited by 610
  6. 2022LION: Latent Point Diffusion Models for 3D Shape Generationcited by 552
  7. 2021Tackling the Generative Learning Trilemma with Denoising Diffusion GANscited by 515
  8. 2022Make-A-Scene: Scene-Based Text-to-Image Generation with Human Priorscited by 487
  9. 2022Diffusion Models for Adversarial Purificationcited by 469
  10. 2021Diffusion Autoencoders: Toward a Meaningful and Decodable Representationcited by 447
  11. 2023Improved Techniques for Training Consistency Modelscited by 430
  12. 2022PhysDiff: Physics-Guided Human Motion Diffusion Modelcited by 365
  13. 2021Knowledge Distillation in Iterative Generative Models for Improved Sampling Speedcited by 344
  14. 2022Analytic-DPM: an Analytic Estimate of the Optimal Reverse Variance in Diffusion Probabilistic Modelscited by 331
  15. 2022Rodin: A Generative Model for Sculpting 3D Digital Avatars Using Diffusioncited by 309
  16. 2022Perception Prioritized Training of Diffusion Modelscited by 301
  17. 2021Deep Generative Modelling: A Comparative Review of VAEs, GANs, Normalizing Flows, Energy-Based and Autoregressive Modelscited by 251
  18. 2021Score-Based Generative Modeling with Critically-Damped Langevin Diffusioncited by 230
  19. 2023A Comprehensive Survey of AI-Generated Content (AIGC): A History of Generative AI from GAN to ChatGPTcited by 221
  20. 2021Likelihood Training of Schr"odinger Bridge using Forward-Backward SDEs Theorycited by 216
  21. 2023Implicit Diffusion Models for Continuous Super-Resolutioncited by 216
  22. 2024Score identity Distillation: Exponentially Fast Distillation of Pretrained Diffusion Models for One-Step Generationcited by 211
  23. 2021Autoregressive Diffusion Modelscited by 208
  24. 2021Few-shot Image Generation via Cross-domain Correspondencecited by 206
  25. 2023Exposing flaws of generative model evaluation metrics and their unfair treatment of diffusion modelscited by 199
  26. 2022Generative Modelling With Inverse Heat Dissipationcited by 164
  27. 2023Structured World Models from Human Videoscited by 144
  28. 2024Lumina-T2X: Transforming Text into Any Modality, Resolution, and Duration via Flow-based Large Diffusion Transformerscited by 137
  29. 2022MoVQ: Modulating Quantized Vectors for High-Fidelity Image Generationcited by 125
  30. 2021ImageBART: Bidirectional Context with Multinomial Diffusion for Autoregressive Image Synthesiscited by 122
  31. 2023Machine Learning for Synthetic Data Generation: A Reviewcited by 118
  32. 2022Wavelet Diffusion Models are fast and scalable Image Generatorscited by 116
  33. 2022Generating Long Videos of Dynamic Scenescited by 113
  34. 2021NeRF-VAE: A Geometry Aware 3D Scene Generative Modelcited by 111
  35. 2021GENESIS-V2: Inferring Unordered Object Representations without Iterative Refinementcited by 109
  36. 2023DiffiT: Diffusion Vision Transformers for Image Generationcited by 104
  37. 2022DiffuseVAE: Efficient, Controllable and High-Fidelity Generation from Low-Dimensional Latentscited by 103
  38. 2022GFlowNets and variational inferencecited by 103

50 most cited of 635 All 635 (JSON)