mlsanityData to 7 Oct 2026

Analysis of Classifier-Free Guidance Weight Schedulers

arXiv 2404.13040 ↗First posted 19 Apr 2024cs.CV cs.LGcited by 64

On the map: Diffusion models → Diffusion foundations and sampling

Abstract (arXiv)

Classifier-Free Guidance (CFG) enhances the quality and condition adherence of text-to-image diffusion models. It operates by combining the conditional and unconditional predictions using a fixed weight. However, recent works vary the weights throughout the diffusion process, reporting superior results but without providing any rationale or analysis. By conducting comprehensive experiments, this paper provides insights into CFG weight schedulers. Our findings suggest that simple, monotonically increasing weight schedulers consistently lead to improved performances, requiring merely a single line of code. In addition, more complex parametrized schedulers can be optimized for further improvement, but do not generalize across different models and tasks.

Key facts. Cited by 64 papers in the map (data through 7 Oct 2026). Placed in Diffusion foundations and sampling, one of 14 directions in Diffusion models. Builds on 36 of its 44 references that are in the map. First posted on arXiv 19 Apr 2024.

Builds on

36 of 44 references.

Closest ancestors · 13

Same direction, newest first

  1. 2024Applying Guidance in a Limited Interval Improves Sample and Distribution Quality in Diffusion Models11 Apr
  2. 2023Adaptive Guidance: Training-free Acceleration of Conditional Diffusion Models19 Dec
  3. 2023SDXL: Improving Latent Diffusion Models for High-Resolution Image Synthesis4 Jul
  4. 2023MDTv2: Masked Diffusion Transformer is a Strong Image Synthesizer25 Mar
  5. 2022DPM-Solver++: Fast Solver for Guided Sampling of Diffusion Probabilistic Models2 Nov
  6. 2022Classifier-Free Diffusion Guidance26 Jul
  7. 2022Entropy-driven Sampling and Training Scheme for Conditional Diffusion Generation23 Jun
  8. 2022DPM-Solver: A Fast ODE Solver for Diffusion Probabilistic Model Sampling in Around 10 Steps2 Jun
  9. 2021High-Resolution Image Synthesis with Latent Diffusion Models20 Dec
  10. 2021Diffusion Models Beat GANs on Image Synthesis11 May
  11. 2021Improved Denoising Diffusion Probabilistic Models18 Feb
  12. 2020Denoising Diffusion Implicit Models6 Oct
1 more
  1. 2020Denoising Diffusion Probabilistic Models19 Jun

Other references · 23

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
  2. 2014Microsoft COCO: Common Objects in Contextcited by 17,498
  3. 2023DINOv2: Learning Robust Visual Features without Supervisioncited by 7,764
  4. 2022Hierarchical Text-Conditional Image Generation with CLIP Latentscited by 6,539
  5. 2022Photorealistic Text-to-Image Diffusion Models with Deep Language Understandingcited by 6,322
  6. 2022LAION-5B: An open large-scale dataset for training next generation image-text modelscited by 4,131
  7. 2021GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Modelscited by 3,595
  8. 2022Flow Straight and Fast: Learning to Generate and Transfer Data with Rectified Flowcited by 3,550
  9. 2022DreamBooth: Fine Tuning Text-to-Image Diffusion Models for Subject-Driven Generationcited by 3,286
  10. 2018Large Scale GAN Training for High Fidelity Natural Image Synthesiscited by 2,609
  11. 2023Stable Video Diffusion: Scaling Latent Video Diffusion Models to Large Datasetscited by 2,526
  12. 2021Blended Diffusion for Text-driven Editing of Natural Imagescited by 924
11 more
  1. 2022eDiff-I: Text-to-Image Diffusion Models with an Ensemble of Expert Denoiserscited by 921
  2. 2016Generative Adversarial Text to Image Synthesiscited by 735
  3. 2023Muse: Text-To-Image Generation via Masked Generative Transformerscited by 640
  4. 2022Executing your Commands via Motion Diffusion in Latent Spacecited by 567
  5. 2023Scaling up GANs for Text-to-Image Synthesiscited by 492
  6. 2023VideoFusion: Decomposed Diffusion Models for High-Quality Video Generationcited by 330
  7. 2023Your Diffusion Model is Secretly a Zero-Shot Classifiercited by 301
  8. 2018Adversarial Audio Synthesiscited by 222
  9. 2015Generating Images from Captions with Attentioncited by 135
  10. 2022Grad-StyleSpeech: Any-speaker Adaptive Text-to-Speech Synthesis with Diffusion Modelscited by 34
  11. 2022SCAM! Transferring humans between images with Semantic Cross Attention Modulationcited by 10

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 9; 2025: 34; 2026: 21.

64 papers.

Sort:

By area:

  1. 2024Guiding a Diffusion Model with a Bad Version of Itselfcited by 286
  2. 2024PeRFlow: Piecewise Rectified Flow as Universal Plug-and-Play Acceleratorcited by 95
  3. 2025PixelFlow: Pixel-Space Generative Models with Flowcited by 91
  4. 2025CFG-Zero*: Improved Classifier-Free Guidance for Flow Matching Modelscited by 47
  5. 2024Dynamic Negative Guidance of Diffusion Modelscited by 28
  6. 2025OmniSync: Towards Universal Lip Synchronization via Diffusion Transformerscited by 26
  7. 2024Shielded Diffusion: Generating Novel and Diverse Images using Sparse Repellencycited by 21
  8. 2024Classifier-Free Guidance inside the Attraction Basin May Cause Memorizationcited by 21
  9. 2024LightningDrag: Lightning Fast and Accurate Drag-based Image Editing Emerging from Videoscited by 18
  10. 2025Token-Shuffle: Towards High-Resolution Image Generation with Autoregressive Modelscited by 18
  11. 2025Feedback Guidance of Diffusion Modelscited by 16
  12. 2024The Silent Assistant: NoiseQuery as Implicit Guidance for Goal-Driven Image Generationcited by 14
38 more
  1. 2025Rectified-CFG++ for Flow Based Modelscited by 14
  2. 2025Effective and Efficient Masked Image Generation Modelscited by 13
  3. 2025MixerMDM: Learnable Composition of Human Motion Diffusion Modelscited by 13
  4. 2025Guidance in the Frequency Domain Enables High-Fidelity Sampling at Low CFG Scalescited by 13
  5. 2025Normalized Attention Guidance: Universal Negative Guidance for Diffusion Modelscited by 12
  6. 2025Learn to Guide Your Diffusion Modelcited by 12
  7. 2025Guiding a Diffusion Transformer with the Internal Dynamics of Itselfcited by 12
  8. 2025Entropy Rectifying Guidance for Diffusion and Flow Modelscited by 11
  9. 2025Improving Classifier-Free Guidance in Masked Diffusion: Low-Dim Theoretical Insights with High-Dim Impactcited by 11
  10. 2025Dynamic Classifier-Free Diffusion Guidance via Online Feedbackcited by 11
  11. 2024Gradient-Free Classifier Guidance for Diffusion Model Samplingcited by 8
  12. 2025REG: Rectified Gradient Guidance for Conditional Diffusion Modelscited by 7
  13. 2025Efficient Distillation of Classifier-Free Guidance using Adapterscited by 5
  14. 2025Mitigating Diffusion Model Hallucinations with Dynamic Guidancecited by 5
  15. 2025Conditional Diffusion Models with Classifier-Free Gibbs-like Guidancecited by 4
  16. 2025Pyramidal Patchification Flow for Visual Generationcited by 4
  17. 2025Selective Classifier-free Guidance for Zero-shot Text-to-speechcited by 4
  18. 2026C2FG: Control Classifier-Free Guidance via Score Discrepancy Analysiscited by 4
  19. 2025Modular Diffusion Policy Training: Decoupling and Recombining Guidance and Diffusion for Offline RLcited by 3
  20. 2025Training-Free Synthetic Data Generation with Dual IP-Adapter Guidancecited by 3
  21. 2026Emergence of Distortions in High-Dimensional Guided Diffusion Modelscited by 3
  22. 2026CFG-Ctrl: Control-Based Classifier-Free Diffusion Guidancecited by 3
  23. 2026Diffusion Reinforcement Learning via Centered Reward Distillationcited by 3
  24. 2024Guiding a diffusion model using sliding windowscited by 2
  25. 2025HiGS: History-Guided Sampling for Plug-and-Play Enhancement of Diffusion Modelscited by 2
  26. 2026SceneNAT: Masked Generative Modeling for Language-Guided Indoor Scene Synthesiscited by 2
  27. 2026Dimension-Free Multimodal Sampling via Preconditioned Annealed Langevin Dynamicscited by 2
  28. 2026Are we really tilting? The mechanics of reward guidance in flow and diffusion modelscited by 2
  29. 2025Towards Generalized and Training-Free Text-Guided Semantic Manipulationcited by 1
  30. 2025Contrastive Conditional-Unconditional Alignment for Long-tailed Diffusion Modelcited by 1
  31. 2025MAMBO-G: Magnitude-Aware Mitigation for Boosted Guidancecited by 1
  32. 2025MIRO: MultI-Reward cOnditioned pretraining improves T2I quality and efficiencycited by 1
  33. 2025RSTR: Reducing SpatioTemporal Redundancy in Diffusion Transformerscited by 1
  34. 2026GenMed: A Pairwise Generative Reformulation of Medical Diagnostic Taskscited by 1
  35. 2026ModaFlow: Modality-Aware Flow Matching for High-Fidelity Virtual Try-Oncited by 1
  36. 2025FastFace: Tuning Identity Preservation in Distilled Diffusion via Guidance and Attentioncited by 0
  37. 2025PromptLoop: Plug-and-Play Prompt Refinement via Latent Feedback for Diffusion Model Alignmentcited by 0
  38. 2025GDiffuSE: Diffusion-based speech enhancement with noise model guidancecited by 0

50 most cited of 64 All 64 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2026Information-Theoretic Classifier-Free Guidance with Adaptive Schedule Optimization
  2. 2025Classifier-free Guidance with Adaptive Scaling
  3. 2024Classifier-Free Guidance is a Predictor-Corrector
  4. 2026Analytic Distribution of Classifier-Free Guidance for Schedule Design
  5. 2025Diffusion Models without Classifier-free Guidance
  6. 2026Revisiting Classifier-Free Guidance Methods in Latent Diffusion Models
  7. 2025Towards a Golden Classifier-Free Guidance Path via Foresight Fixed Point Iterations
  8. 2025Towards Understanding the Mechanisms of Classifier-Free Guidance