mlsanityData to 14 Oct 2026

TFG: Unified Training-Free Guidance for Diffusion Models

arXiv 2409.15761 ↗First posted 24 Sep 2024cs.LG cs.AIcited by 107

On the map: Diffusion models → Diffusion foundations and sampling

Abstract (arXiv)

Given an unconditional diffusion model and a predictor for a target property of interest (e.g., a classifier), the goal of training-free guidance is to generate samples with desirable target properties without additional training. Existing methods, though effective in various individual applications, often lack theoretical grounding and rigorous testing on extensive benchmarks. As a result, they could even fail on simple tasks, and applying them to a new problem becomes unavoidably difficult. This paper introduces a novel algorithmic framework encompassing existing methods as special cases, unifying the study of training-free guidance into the analysis of an algorithm-agnostic design space. Via theoretical and empirical investigation, we propose an efficient and effective hyper-parameter searching strategy that can be readily applied to any downstream task. We systematically benchmark across 7 diffusion models on 16 tasks with 40 targets, and improve performance by 8.5% on average. Our framework and benchmark offer a solid foundation for conditional generation in a training-free manner.

Key facts. Cited by 107 papers in the map (data through 14 Oct 2026). Placed in Diffusion foundations and sampling, one of 14 directions in Diffusion models. Builds on 58 of its 82 references that are in the map. First posted on arXiv 24 Sep 2024.

Builds on

58 of 82 references.

Closest ancestors · 15

Same direction, newest first

  1. 2023Manifold Preserving Guided Diffusion28 Nov
  2. 2023Conditional Generation from Unconditional Diffusion Models using Denoiser Representations2 Jun
  3. 2023FreeDoM: Training-Free Energy-Guided Conditional Diffusion Model17 Mar
  4. 2023Reduce, Reuse, Recycle: Compositional Generation with Energy-Based Diffusion Models and MCMC22 Feb
  5. 2023Universal Guidance for Diffusion Models14 Feb
  6. 2022Classifier-Free Diffusion Guidance26 Jul
  7. 2021High-Resolution Image Synthesis with Latent Diffusion Models20 Dec
  8. 2021Conditional Image Generation with Score-Based Diffusion Models26 Nov
  9. 2021Cascaded Diffusion Models for High Fidelity Image Generation30 May
  10. 2021Diffusion Models Beat GANs on Image Synthesis11 May
  11. 2021Improved Denoising Diffusion Probabilistic Models18 Feb
  12. 2021Maximum Likelihood Training of Score-Based Diffusion Models22 Jan
3 more
  1. 2020Score-Based Generative Modeling through Stochastic Differential Equations26 Nov
  2. 2020Denoising Diffusion Implicit Models6 Oct
  3. 2020Denoising Diffusion Probabilistic Models19 Jun

Other references · 43

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 30,186
  3. 2015U-Net: Convolutional Networks for Biomedical Image Segmentationcited by 19,713
  4. 2018The Unreasonable Effectiveness of Deep Features as a Perceptual Metriccited by 11,177
  5. 2014ImageNet Large Scale Visual Recognition Challengecited by 8,498
  6. 2015Deep Unsupervised Learning using Nonequilibrium Thermodynamicscited by 7,058
  7. 2020Training data-efficient image transformers & distillation through attentioncited by 4,586
  8. 2022LAION-5B: An open large-scale dataset for training next generation image-text modelscited by 4,152
  9. 2019Generative Modeling by Estimating Gradients of the Data Distributioncited by 4,101
  10. 2021GLIDE: Towards Photorealistic Image Generation and Editing with Text-Guided Diffusion Modelscited by 3,614
  11. 2017Progressive Growing of GANs for Improved Quality, Stability, and Variationcited by 2,988
  12. 2016Perceptual Losses for Real-Time Style Transfer and Super-Resolutioncited by 2,694
31 more
  1. 2022RePaint: Inpainting using Denoising Diffusion Probabilistic Modelscited by 1,430
  2. 2020DiffWave: A Versatile Diffusion Model for Audio Synthesiscited by 1,354
  3. 2022Scaling Autoregressive Models for Content-Rich Text-to-Image Generationcited by 1,295
  4. 2018Demystifying MMD GANscited by 1,293
  5. 2021E(n) Equivariant Graph Neural Networkscited by 965
  6. 2022Denoising Diffusion Restoration Modelscited by 963
  7. 2021Diffusion Probabilistic Models for 3D Point Cloud Generationcited by 735
  8. 2022Equivariant Diffusion for Molecule Generation in 3Dcited by 677
  9. 2023AudioLDM: Text-to-Audio Generation with Latent Diffusion Modelscited by 599
  10. 2022Zero-Shot Image Restoration Using Denoising Diffusion Null-Space Modelcited by 583
  11. 2022GeoDiff: a Geometric Diffusion Model for Molecular Conformation Generationcited by 516
  12. 2019Deep Learning for Image Super-resolution: A Surveycited by 369
  13. 2021Score-based diffusion models for accelerated MRIcited by 311
  14. 2022OpenOOD: Benchmarking Generalized Out-of-Distribution Detectioncited by 268
  15. 2018Fr'echet Audio Distance: A Metric for Evaluating Music Enhancement Algorithmscited by 225
  16. 2023Geometric Latent Diffusion Models for 3D Molecule Generationcited by 186
  17. 2015Large-scale Classification of Fine-Art Paintings: Learning The Right Metric on The Right Featurecited by 182
  18. 2023Contrastive Energy Prediction for Exact Energy-Guided Diffusion Sampling in Offline Reinforcement Learningcited by 149
  19. 2014Part-based R-CNNs for Fine-grained Category Detectioncited by 109
  20. 2021A Conditional Point Diffusion-Refinement Paradigm for 3D Point Cloud Completioncited by 100
  21. 2022Equivariant Energy-Guided SDE for Inverse Molecular Designcited by 73
  22. 2022Solving Audio Inverse Problems with a Diffusion Modelcited by 62
  23. 2024Motion Guidance: Diffusion-Based Image Editing with Differentiable Motion Estimatorscited by 39
  24. 2024Diffusion Models for Audio Restorationcited by 32
  25. 2023Controllable Music Production with Diffusion Models and Guidance Gradientscited by 29
  26. 2023Control3Diff: Learning Controllable 3D Diffusion Models from Single-view Imagescited by 27
  27. 2022Freeze then Train: Towards Provable Representation Learning under Spurious Correlations and Feature Noisecited by 22
  28. 2023A framework for conditional diffusion modelling with applications in motif scaffolding for protein designcited by 22
  29. 2023Steered Diffusion: A Generalized Framework for Plug-and-Play Conditional Image Synthesiscited by 18
  30. 2024Generalizing Alignment Paradigm of Text-to-Image Generation with Preferences through f-divergence Minimizationcited by 15
  31. 2023VRDMG: Vocal Restoration via Diffusion Posterior Sampling with Multiple Guidancecited by 9

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 6; 2025: 47; 2026: 54.

107 papers.

Sort:

By area:

  1. 2025From Reflection to Perfection: Scaling Inference-Time Optimization for Text-to-Image Diffusion Models via Reflection Tuningcited by 56
  2. 2025Inference-time Scaling of Diffusion Models through Classical Searchcited by 49
  3. 2025On the Guidance of Flow Matchingcited by 45
  4. 2026Inference-time Physics Alignment of Video Generative Models with Latent World Modelscited by 37
  5. 2025Taming Video Diffusion Prior with Scene-Grounding Guidance for 3D Gaussian Splatting from Sparse Inputscited by 29
  6. 2026Diamond Maps: Efficient Reward Alignment via Stochastic Flow Mapscited by 29
  7. 2024DyMO: Training-Free Diffusion Model Alignment with Dynamic Multi-Objective Schedulingcited by 27
  8. 2025Training-Free Guidance Beyond Differentiability: Scalable Path Steering with Tree Search in Diffusion and Flow Modelscited by 24
  9. 2024On Statistical Rates of Conditional Diffusion Transformers: Approximation, Estimation and Minimax Optimalitycited by 22
  10. 2025Variational Control for Guidance in Diffusion Modelscited by 22
  11. 2025Training-Free Constrained Generation With Stable Diffusion Modelscited by 20
  12. 2025Composition and Control with Distilled Energy Diffusion Models and Sequential Monte Carlocited by 20
38 more
  1. 2025Solving Inverse Problems via Diffusion-Based Priors: An Approximation-Free Ensemble Sampling Approachcited by 18
  2. 2025GLASS Flows: Transition Sampling for Alignment of Flow and Diffusion Modelscited by 17
  3. 2025Feedback Guidance of Diffusion Modelscited by 16
  4. 2025Psi-Sampler: Initial Particle Sampling for SMC-Based Inference-Time Reward Alignment in Score Modelscited by 15
  5. 2025CoDe: Blockwise Control for Denoising Diffusion Modelscited by 14
  6. 2025LinEAS: End-to-end Learning of Activation Steering with a Distributional Losscited by 14
  7. 2025SPARKE: Scalable Prompt-Aware Diversity and Novelty Guidance in Diffusion Models via RKE Scorecited by 14
  8. 2025Source-Guided Flow Matchingcited by 13
  9. 2025Data-regularized Reinforcement Learning for Diffusion Models at Scalecited by 13
  10. 2024Decoupling Training-Free Guided Diffusion by ADMMcited by 11
  11. 2025PhySense: Sensor Placement Optimization for Accurate Physics Sensingcited by 11
  12. 2025CREPE: Controlling Diffusion with Replica Exchangecited by 11
  13. 2026Training-Free Adaptation of Diffusion Models via Doob's h-Transformcited by 11
  14. 2025ORIGEN: Zero-Shot 3D Orientation Grounding in Text-to-Image Generationcited by 10
  15. 2025Neuro-Symbolic Generative Diffusion Models for Physically Grounded, Robust, and Safe Generationcited by 10
  16. 2024Gradient-Free Classifier Guidance for Diffusion Model Samplingcited by 8
  17. 2025Diffusion Blend: Inference-Time Multi-Preference Alignment for Diffusion Modelscited by 7
  18. 2026C2FG: Control Classifier-Free Guidance via Score Discrepancy Analysiscited by 7
  19. 2025Product of Experts for Visual Generationcited by 6
  20. 2026Proprio: Latent Self-Scoring and Inference-Time Refinement for Physically Plausible Video Generationcited by 6
  21. 2025TFG-Flow: Training-free Guidance in Multimodal Generative Flowcited by 5
  22. 2025ANYPORTAL: Zero-Shot Consistent Video Background Replacementcited by 5
  23. 2025GOOD: Training-Free Guided Diffusion Sampling for Out-of-Distribution Detectioncited by 5
  24. 2026Calibrated Test-Time Guidance for Bayesian Inferencecited by 5
  25. 2024Flow Matching Posterior Sampling: A Training-free Conditional Generation for Flow Matchingcited by 4
  26. 2025Zero-Shot Cyclic Peptide Design via Composable Geometric Constraintscited by 4
  27. 2025TITAN-Guide: Taming Inference-Time AligNment for Guided Text-to-Video Diffusion Modelscited by 4
  28. 2025Training-Free Reward-Guided Image Editing via Trajectory Optimal Controlcited by 4
  29. 2026HyperAlign: Hypernetwork for Efficient Test-Time Alignment of Diffusion Modelscited by 4
  30. 2026Training-Free Refinement of Flow Matching with Divergence-based Samplingcited by 4
  31. 2025Noise is All You Need: Solving Linear Inverse Problems by Noise Combination Sampling with Diffusion Modelscited by 3
  32. 2026MMD Guidance: Training-Free Distribution Adaptation for Diffusion Models via Maximum Mean Discrepancy Guidancecited by 3
  33. 2026Alignment of Diffusion Model and Flow Matching for Text-to-Image Generationcited by 3
  34. 2026Entropy Aware Reward Guidance for Diffusion Language Model Alignmentcited by 3
  35. 2026DMGD: Train-Free Dataset Distillation with Semantic-Distribution Matching in Diffusion Modelscited by 3
  36. 2025Geometric Consistency Refinement for Single Image Novel View Synthesis via Test-Time Adaptation of Diffusion Modelscited by 2
  37. 2025Diffusion-based Evolutionary Optimization for 3D Multi-Objective Molecular Generationcited by 2
  38. 2025Diffusion Models with Double Guidance: Generate with aggregated datasetscited by 2

50 most cited of 107 All 107 (JSON)

Similar, not cited

Close in wording, no citation either way

  1. 2023Bridging the Gap: Addressing Discrepancies in Diffusion Model Training for Classifier-Free Guidance
  2. 2024No Training, No Problem: Rethinking Classifier-Free Guidance for Diffusion Models
  3. 2025Towards a unified framework for guided diffusion models
  4. 2024Rectified Diffusion Guidance for Conditional Generation
  5. 2023Elucidating The Design Space of Classifier-Guided Diffusion Generation
  6. 2025Stochastic Self-Guidance for Training-Free Enhancement of Diffusion Models
  7. 2025Improving Classifier-Free Guidance in Masked Diffusion: Low-Dim Theoretical Insights with High-Dim Impact
  8. 2025Token Perturbation Guidance for Diffusion Models