mlsanityData to 7 Oct 2026

PixelDiT: Pixel Diffusion Transformers for Image Generation

arXiv 2511.20645 ↗First posted 25 Nov 2025cs.CVcited by 75

On the map: Diffusion models → Diffusion foundations and sampling

Abstract (arXiv)

Latent-space modeling has been the standard for Diffusion Transformers (DiTs). However, it relies on a two-stage pipeline where the pretrained autoencoder introduces lossy reconstruction, leading to error accumulation while hindering joint optimization. To address these issues, we propose PixelDiT, a single-stage, end-to-end model that eliminates the need for the autoencoder and learns the diffusion process directly in the pixel space. PixelDiT adopts a fully transformer-based architecture shaped by a dual-level design: a patch-level DiT that captures global semantics and a pixel-level DiT that refines texture details, enabling efficient training of a pixel-space diffusion model while preserving fine details. PixelDiT achieves 1.61 FID on ImageNet 256 and 1.81 FID on ImageNet 512, surpassing existing pixel generative models. We further extend PixelDiT to text-to-image generation and pretrain it at the 10242resolution in pixel space. It achieves 0.74 on GenEval and 83.5 on DPG-bench, approaching the best latent diffusion models. Code: https://github.com/NVlabs/PixelDiT

Key facts. Cited by 75 papers in the map (data through 7 Oct 2026). Placed in Diffusion foundations and sampling, one of 14 directions in Diffusion models. Builds on 43 of its 52 references that are in the map. First posted on arXiv 25 Nov 2025.

Builds on

43 of 52 references.

Closest ancestors · 26

Same direction, newest first

  1. 2025Latent Diffusion Model without Variational Autoencoder17 Oct
  2. 2025Diffusion Transformers with Representation Autoencoders13 Oct
  3. 2025PixNerd: Pixel Neural Field Diffusion31 Jul
  4. 2025REPA-E: Unlocking VAE for End-to-End Tuning with Latent Diffusion Transformers14 Apr
  5. 2025Masked Autoencoders Are Effective Tokenizers for Diffusion Models5 Feb
  6. 2025Reconstruction vs. Generation: Taming Optimization Dilemma in Latent Diffusion Models2 Jan
  7. 2024Simpler Diffusion (SiD2): 1.5 FID on ImageNet512 with pixel-space diffusion25 Oct
  8. 2024Deep Compression Autoencoder for Efficient High-Resolution Diffusion Models14 Oct
  9. 2024Representation Alignment for Generation: Training Diffusion Transformers Is Easier Than You Think9 Oct
  10. 2024Lumina-Next: Making Lumina-T2X Stronger and Faster with Next-DiT5 Jun
  11. 2024Applying Guidance in a Limited Interval Improves Sample and Distribution Quality in Diffusion Models11 Apr
  12. 2024PixArt-Sigma: Weak-to-Strong Training of Diffusion Transformer for 4K Text-to-Image Generation7 Mar
14 more
  1. 2024Scaling Rectified Flow Transformers for High-Resolution Image Synthesis5 Mar
  2. 2024SiT: Exploring Flow and Diffusion-based Generative Models with Scalable Interpolant Transformers16 Jan
  3. 2023Analyzing and Improving the Training Dynamics of Diffusion Models5 Dec
  4. 2023PixArt-α: Fast Training of Diffusion Transformer for Photorealistic Text-to-Image Synthesis30 Sep
  5. 2023SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesis4 Jul
  6. 2023Fast Training of Diffusion Models with Masked Transformers15 Jun
  7. 2023Understanding Diffusion Objectives as the ELBO with Simple Data Augmentation1 Mar
  8. 2023Simple diffusion: End-to-end diffusion for high resolution images26 Jan
  9. 2022Scalable Diffusion Models with Transformers19 Dec
  10. 2022DPM-Solver++: Fast Solver for Guided Sampling of Diffusion Probabilistic Models2 Nov
  11. 2022All are Worth Words: A ViT Backbone for Diffusion Models25 Sep
  12. 2021High-Resolution Image Synthesis with Latent Diffusion Models20 Dec
  13. 2021Cascaded Diffusion Models for High Fidelity Image Generation30 May
  14. 2021Diffusion Models Beat GANs on Image Synthesis11 May

Other references · 17

  1. 2014ImageNet Large Scale Visual Recognition Challengecited by 8,441
  2. 2023DINOv2: Learning Robust Visual Features without Supervisioncited by 7,764
  3. 2022Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flowcited by 3,550
  4. 2024Gemma 2: Improving Open Language Models at a Practical Sizecited by 1,928
  5. 2023GenEval: An Object-Focused Framework for Evaluating Text-to-Image Alignmentcited by 946
  6. 2022StyleGAN-XL: Scaling StyleGAN to Large Diverse Datasetscited by 489
  7. 2024ELLA: Equip Diffusion Models with LLM for Enhanced Semantic Alignmentcited by 455
  8. 2024Hunyuan-DiT: A Powerful Multi-Resolution Diffusion Transformer with Fine-Grained Chinese Understandingcited by 294
  9. 2024Playground v2.5: Three Insights towards Enhancing Aesthetic Quality in Text-to-Image Generationcited by 249
  10. 2024FlowEdit: Inversion-Free Text-Based Editing Using Pre-Trained Flow Modelscited by 214
  11. 2022Scalable Adaptive Computation for Iterative Generationcited by 190
  12. 2024Playground v3: Improving Text-to-Image Alignment with Deep-Fusion Large Language Modelscited by 113
5 more
  1. 2025PixelFlow: Pixel-Space Generative Models with Flowcited by 91
  2. 2024JetFormer: An Autoregressive Generative Model of Raw Images and Textcited by 44
  3. 2025Fractal Generative Modelscited by 44
  4. 2025FARMER: Flow AutoRegressive Transformer over Pixelscited by 17
  5. 2025ZipIR: Latent Pyramid Diffusion Transformer for High-Resolution Image Restorationcited by 5

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2025: 1; 2026: 74.

75 papers.

Sort:

By area:

  1. 2026Generative Modeling via Driftingcited by 98
  2. 2026One-step Latent-free Image Generation with Pixel Mean Flowscited by 49
  3. 2026SenseNova-U1: Unifying Multimodal Understanding and Generation with NEO-unify Architecturecited by 27
  4. 2026Tuna-2: Pixel Embeddings Beat Vision Encoders for Multimodal Understanding and Generationcited by 17
  5. 2026PixelGen: Improving Pixel Diffusion with Perceptual Supervisioncited by 10
  6. 2026Image Generation with a Sphere Encodercited by 8
  7. 2026L2P: Unlocking Latent Potential for Pixel Generationcited by 7
  8. 2026HyperDiT: Hyper-Connected Transformers for High-Fidelity Pixel-Space Diffusioncited by 7
  9. 2026Laminating Representation Autoencoders for Efficient Diffusioncited by 6
  10. 2026End-to-End Training for Unified Tokenization and Latent Denoisingcited by 6
  11. 2026PiD: Fast and High-Resolution Latent Decoding with Pixel Diffusioncited by 6
  12. 2026PixelU: A U-Shaped Transformer for Efficient End-to-End Pixel Diffusioncited by 6
38 more
  1. 2026Spectrally-Guided Diffusion Noise Schedulescited by 5
  2. 2026Asymmetric Flow Modelscited by 4
  3. 2026FrequencyBooster: Full-Frequency Modeling for High-Fidelity Pixel Diffusioncited by 4
  4. 2026Chimera: Designing and Chinchilla-Scaling Hybrid Visual Diffusion Transformerscited by 3
  5. 2026CMuon: Accelerating and Stabilizing Diffusion Transformer Training via Chunked Momentum Orthogonalizationcited by 3
  6. 2026An Empirical Study of Training Pixel-Space Text-to-Image Diffusion Modelscited by 3
  7. 2026SenseNova-U1.5: Towards Native Unified Visual Intelligencecited by 3
  8. 2026V-Co: A Closer Look at Visual Representation Alignment via Co-Denoisingcited by 2
  9. 2026FREPix: Frequency-Heterogeneous Flow Matching for Pixel-Space Image Generationcited by 2
  10. 2026Registers Matter for Pixel-Space Diffusion Transformerscited by 2
  11. 2026RiT: Vanilla Diffusion Transformers Suffice in Representation Spacecited by 2
  12. 2026CrossFlow: One-Step Generation Across Latent and Pixel Spacescited by 2
  13. 2026PixGS: Pixel-Space Diffusion for Direct 3D Gaussian Splat Generationcited by 2
  14. 2026Three-Body Scattering for Generative Modelingcited by 2
  15. 2026WaiT for the Signal: Simple Frequency-Aware Flow-Matchingcited by 2
  16. 2026A Frozen Pixel-Space Diffusion Model Can Guide Itself with Its Own Samplescited by 2
  17. 2026GenFirst: Generation Before Reconstruction for Stable End-to-End Latent Generative Modelingcited by 2
  18. 2025RePack then Refine: Efficient Diffusion Transformer with Vision Foundation Modelcited by 1
  19. 2026NNiT: Width-Agnostic Neural Network Generation with Structurally Aligned Weight Spacescited by 1
  20. 2026WiT: Waypoint Diffusion Transformers for Alleviating Trajectory Conflict in Pixel-Space Image Generationcited by 1
  21. 2026WavTTS: Towards High-Quality Zero-Shot TTS via Direct Raw Waveform Modelingcited by 1
  22. 2026BareWave: Waveform-Native Flow-Matching Text-to-Speechcited by 1
  23. 2026GEAR: Guided End-to-End AutoRegression for Image Synthesiscited by 1
  24. 2026SPEED: One-Step Pixel Diffusion for High-quality Video Frame Interpolationcited by 1
  25. 2026DuSPiT: Dual-Branch Sub-Patch Pixel Diffusion Transformercited by 1
  26. 2026Generative Video Compression with Adaptive Score Distillationcited by 1
  27. 2026PixelControl: Fine-Grained Condition Fidelity in Text-to-Image Diffusioncited by 1
  28. 2026PXDepth: Pixel-Space Modeling for Structure Preserving Monocular Depth Estimationcited by 1
  29. 2026Pixel-Space Diffusion via Observation Operatorscited by 1
  30. 2026Balancing Frequencies and Pixels in Flow Matchingcited by 1
  31. 2026Rethinking Vector Field Learning for Generative Segmentationcited by 0
  32. 2026UniCSG: Unified High-Fidelity Content-Constrained Style-Driven Generation via Staged Semantic and Frequency Disentanglementcited by 0
  33. 2026BitLM: Unlocking Multi-Token Language Generation with Bitwise Continuous Diffusioncited by 0
  34. 2026RaPD: Resolution-Agnostic Pixel Diffusion via Semantics-Enriched Implicit Representationscited by 0
  35. 2026Mat'ern Noise for Triangulation-Agnostic Flow Matching on Meshescited by 0
  36. 2026PixIE: Prompted Pixel-Space Low-Light Image Enhancementcited by 0
  37. 2026Vision-driven Preference Synthesis for Mitigating Hallucinations in VLMscited by 0
  38. 2026Scaling Storm-Resolving Atmospheric AI Simulation to the Entire Planetcited by 0

50 most cited of 75 All 75 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2025DeCo: Frequency-Decoupled Pixel Diffusion for End-to-End Image Generation
  2. 2025DiP: Taming Diffusion Models in Pixel Space
  3. 2024Pixel-Space Post-Training of Latent Diffusion Models
  4. 2025There is No VAE: End-to-End Pixel-Space Generative Modeling via Self-Supervised Pre-training
  5. 2026PixelART: Image-to-Layer Decomposition without Latents or Text-to-Image Pretraining
  6. 2026Iris-3B: Going Beyond the Latent with Pixel-Space Diffusion Training, Conversion and Fine-Tuning
  7. 2026HRDiT: Training-Free High-Resolution Image Generation with Off-the-Shelf Diffusion Transformer Models
  8. 2024Diffusion Models Need Visual Priors for Image Generation