mlsanityData to 7 Oct 2026

Laminating Representation Autoencoders for Efficient Diffusion

arXiv 2602.04873 ↗First posted 4 Feb 2026cs.CVcited by 6

On the map: Diffusion models → Diffusion foundations and sampling

Abstract (arXiv)

Recent work has shown that diffusion models can generate high-quality images by operating directly on SSL patch features rather than pixel-space latents. However, the dense patch grids from encoders like DINOv2 contain significant redundancy, making diffusion needlessly expensive. We introduce FlatDINO, a variational autoencoder that compresses this representation into a one-dimensional sequence of just 32 continuous tokens -an 8x reduction in sequence length and 48x compression in total dimensionality. On ImageNet 256x256, a DiT-XL trained on FlatDINO latents achieves a gFID of 1.80 with classifier-free guidance while requiring 8x fewer FLOPs per forward pass and up to 4.5x fewer FLOPs per training step compared to diffusion on uncompressed DINOv2 features. These are preliminary results and this work is in progress.

Key facts. Cited by 6 papers in the map (data through 7 Oct 2026). Placed in Diffusion foundations and sampling, one of 14 directions in Diffusion models. Builds on 41 of its 61 references that are in the map. First posted on arXiv 4 Feb 2026.

Builds on

41 of 61 references.

Closest ancestors · 15

Same direction, newest first

  1. 2025PixelDiT: Pixel Diffusion Transformers for Image Generation25 Nov
  2. 2025Representation Entanglement for Generation: Training Diffusion Transformers Is Much Easier Than You Think2 Jul
  3. 2025REPA-E: Unlocking VAE for End-to-End Tuning with Latent Diffusion Transformers14 Apr
  4. 2025Reconstruction vs. Generation: Taming Optimization Dilemma in Latent Diffusion Models2 Jan
  5. 2024Simpler Diffusion (SiD2): 1.5 FID on ImageNet512 with pixel-space diffusion25 Oct
  6. 2024Representation Alignment for Generation: Training Diffusion Transformers Is Easier Than You Think9 Oct
  7. 2024Applying Guidance in a Limited Interval Improves Sample and Distribution Quality in Diffusion Models11 Apr
  8. 2024Scaling Rectified Flow Transformers for High-Resolution Image Synthesis5 Mar
  9. 2023SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesis4 Jul
  10. 2023Fast Training of Diffusion Models with Masked Transformers15 Jun
  11. 2022Scalable Diffusion Models with Transformers19 Dec
  12. 2021High-Resolution Image Synthesis with Latent Diffusion Models20 Dec
3 more
  1. 2021Diffusion Models Beat GANs on Image Synthesis11 May
  2. 2020Score-Based Generative Modeling through Stochastic Differential Equations26 Nov
  3. 2020Denoising Diffusion Implicit Models6 Oct

Other references · 26

  1. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 29,915
  2. 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 11,979
  3. 2018The Unreasonable Effectiveness of Deep Features as a Perceptual Metriccited by 11,029
  4. 2023DINOv2: Learning Robust Visual Features without Supervisioncited by 7,764
  5. 2019Momentum Contrast for Unsupervised Visual Representation Learningcited by 7,482
  6. 2015Deep Unsupervised Learning using Nonequilibrium Thermodynamicscited by 7,005
  7. 2021Emerging Properties in Self-Supervised Vision Transformerscited by 6,893
  8. 2017Neural Discrete Representation Learningcited by 4,650
  9. 2019Generative Modeling by Estimating Gradients of the Data Distributioncited by 4,067
  10. 2020Taming Transformers for High-Resolution Image Synthesiscited by 3,229
  11. 2023Diffusion Policy: Visuomotor Policy Learning via Action Diffusioncited by 2,730
  12. 2013Describing Textures in the Wildcited by 2,110
14 more
  1. 2020GLU Variants Improve Transformercited by 1,724
  2. 2017Deep Image Priorcited by 1,097
  3. 2024Visual Autoregressive Modeling: Scalable Image Generation via Next-Scale Predictioncited by 1,005
  4. 2023Vision Transformers Need Registerscited by 844
  5. 2024MiniCPM: Unveiling the Potential of Small Language Models with Scalable Training Strategiescited by 694
  6. 2024Autoregressive Image Generation without Vector Quantizationcited by 644
  7. 2024An Image is Worth 32 Tokens for Reconstruction and Generationcited by 295
  8. 2022Splicing ViT Features for Semantic Appearance Transfercited by 209
  9. 2025FlexTok: Resampling Images into 1D Token Sequences of Flexible Lengthcited by 105
  10. 2025PixelFlow: Pixel-Space Generative Models with Flowcited by 91
  11. 2025What matters for Representation Alignment: Global Information or Spatial Structure?cited by 76
  12. 2025Beyond Next-Token: Next-X Prediction for Autoregressive Visual Generationcited by 71
  13. 2025Democratizing Text-to-Image Masked Generative Models with Compact Text-Aware One-Dimensional Tokenscited by 48
  14. 2024A Review on Discriminative Self-supervised Learning Methods in Computer Visioncited by 6

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 6.

6 papers.

  1. 2026Improved Baselines with Representation Autoencoderscited by 24
  2. 2026Making Reconstruction FID Predictive of Diffusion Generation FIDcited by 12
  3. 2026What Matters for Diffusion-Friendly Latent Manifold? Prior-Aligned Autoencoders for Latent Diffusioncited by 7
  4. 2026Beyond Gaussian Bottlenecks: Topologically Aligned Encoding of Vision-Transformer Feature Spacescited by 1
  5. 2026SRC-Flow: Compact Semantic Representations Enable Normalizing Flows for Image Generationcited by 0
  6. 2026FuseReg: Regularizing Layer Fusion Mitigates the Reconstruction-Generation Gap in Representation Autoencoderscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025DC-AE 1.5: Accelerating Diffusion Model Convergence with Structured Latent Space
  2. 2025Diffusion Transformers with Representation Autoencoders
  3. 2025Latent Diffusion Models with Masked AutoEncoders
  4. 2024PaGoDA: Progressive Growing of a One-Step Generator from a Low-Resolution Diffusion Teacher
  5. 2026Pooling Representation Autoencoders for Efficient Diffusion
  6. 2023Denoising Diffusion Autoencoders are Unified Self-supervised Learners
  7. 2022DiffuseVAE: Efficient, Controllable and High-Fidelity Generation from Low-Dimensional Latents
  8. 2026DC-SAE: Deep Compression Semantic Autoencoder for Faster Diffusion Convergence