mlsanityData to 14 Oct 2026

VIVAT: Virtuous Improving VAE Training through Artifact Mitigation

arXiv 2506.07863 ↗First posted 9 Jun 2025cs.CV cs.LG cs.MMcited by 3

Not on the map. Why?

Abstract (arXiv)

Variational Autoencoders (VAEs) remain a cornerstone of generative computer vision, yet their training is often plagued by artifacts that degrade reconstruction and generation quality. This paper introduces VIVAT, a systematic approach to mitigating common artifacts in KL-VAE training without requiring radical architectural changes. We present a detailed taxonomy of five prevalent artifacts - color shift, grid patterns, blur, corner and droplet artifacts - and analyze their root causes. Through straightforward modifications, including adjustments to loss weights, padding strategies, and the integration of Spatially Conditional Normalization, we demonstrate significant improvements in VAE performance. Our method achieves state-of-the-art results in image reconstruction metrics (PSNR and SSIM) across multiple benchmarks and enhances text-to-image generation quality, as evidenced by superior CLIP scores. By preserving the simplicity of the KL-VAE framework while addressing its practical challenges, VIVAT offers actionable insights for researchers and practitioners aiming to optimize VAE training.

Key facts. Cited by 3 papers in the map (data through 14 Oct 2026). Builds on 40 of its 54 references that are in the map. First posted on arXiv 9 Jun 2025.

Builds on

40 of 54 references.

References · 40

  1. 2017Attention Is All You Needcited by 60,361
  2. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  3. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,744
  4. 2014Microsoft COCO: Common Objects in Contextcited by 17,608
  5. 2018The Unreasonable Effectiveness of Deep Features as a Perceptual Metriccited by 11,177
  6. 2022Hierarchical Text-Conditional Image Generation with CLIP Latentscited by 6,585
  7. 2018A Style-Based Generator Architecture for Generative Adversarial Networkscited by 6,013
  8. 2017Neural Discrete Representation Learningcited by 4,712
  9. 2023SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesiscited by 4,356
  10. 2024Scaling Rectified Flow Transformers for High-Resolution Image Synthesiscited by 4,181
  11. 2022LAION-5B: An open large-scale dataset for training next generation image-text modelscited by 4,152
  12. 2020Taming Transformers for High-Resolution Image Synthesiscited by 3,253
28 more
  1. 2016Perceptual Losses for Real-Time Style Transfer and Super-Resolutioncited by 2,694
  2. 2021CLIPScore: A Reference-free Evaluation Metric for Image Captioningcited by 2,424
  3. 2024CogVideoX: Text-to-Video Diffusion Models with An Expert Transformercited by 2,261
  4. 2024HunyuanVideo: A Systematic Framework For Large Video Generative Modelscited by 1,660
  5. 2018Group Normalizationcited by 1,524
  6. 2019Generating Diverse High-Fidelity Images with VQ-VAE-2cited by 1,387
  7. 2019An Introduction to Variational Autoencoderscited by 1,000
  8. 2025Cosmos World Foundation Model Platform for Physical AIcited by 864
  9. 2021Vector-quantized Image Modeling with Improved VQGANcited by 658
  10. 2024Movie Gen: A Cast of Media Foundation Modelscited by 644
  11. 2020NVAE: A Deep Hierarchical Variational Autoencodercited by 637
  12. 2015Autoencoding beyond pixels using a learned similarity metriccited by 407
  13. 2017Wasserstein Auto-Encoderscited by 373
  14. 2020Very Deep VAEs Generalize Autoregressive Models and Can Outperform Them on Imagescited by 264
  15. 2016Deep Unsupervised Clustering with Gaussian Mixture Variational Autoencoderscited by 220
  16. 2019Understanding and Improving Layer Normalizationcited by 217
  17. 2024Deep Compression Autoencoder for Efficient High-Resolution Diffusion Modelscited by 217
  18. 2020Focal Frequency Loss for Image Reconstruction and Synthesiscited by 206
  19. 2024Fast Timing-Conditioned Latent Audio Diffusioncited by 181
  20. 2022MoVQ: Modulating Quantized Vectors for High-Fidelity Image Generationcited by 125
  21. 2022DiffuseVAE: Efficient, Controllable and High-Fidelity Generation from Low-Dimensional Latentscited by 104
  22. 2015Attribute2Image: Conditional Image Generation from Visual Attributescited by 93
  23. 2025EQ-VAE: Equivariance Regularized Latent Space for Improved Generative Image Modelingcited by 89
  24. 2020Simple and Effective VAE Training with Calibrated Decoderscited by 57
  25. 2021Consistency Regularization for Variational Auto-Encoderscited by 51
  26. 2024Exploring the Role of Large Language Models in Prompt Encoding for Diffusion Modelscited by 45
  27. 2018Variational Autoencoder with Implicit Optimal Priorscited by 33
  28. 2024Kandinsky 3: Text-to-Image Synthesis for Multifunctional Generative Frameworkcited by 25

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 1; 2026: 2.

3 papers.

  1. 2025ART-VITON: Measurement-Guided Latent Diffusion for Artifact-Free Virtual Try-Oncited by 0
  2. 2026TransNormal-2: Geometry-Grounded Rectified Flow with Edge-Aware Decoding for Precise Normal Estimationcited by 0
  3. 2026Kandinsky 6.0 Video: Foundation Models for Synchronized Video and Audio Generationcited by 0