Research map / Image generation / Fast diffusion sampling
FreeDoM: Training-Free Energy-Guided Conditional Diffusion Model
Jiwen Yu, Yinhuai Wang, Chen Zhao, Bernard Ghanem, Jian Zhang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| FreeDoM instead of TediGAN | text-to-face generation · Text to faces | FID: 71.71 -> 55.91 | Improved |
| FreeDoM instead of TediGAN | sketch-to-face generation · Sketches to faces | FID: 91.11 -> 70.97 | Improved |
| FreeDoM instead of TediGAN | segmentation-to-face generation · Segmentation maps to faces | FID: 52.77 -> 53.08 | No clear effect |
| FreeDoM instead of TediGAN | segmentation-to-face generation · Segmentation maps to faces | condition distance: 2037.2 -> 1696.1 | Improved |
| FreeDoM instead of TediGAN | sketch-to-face generation · Sketches to faces | condition distance: 48.61 -> 33.29 | Improved |
| FreeDoM instead of TediGAN | text-to-face generation · Text to faces | condition distance: 12.31 -> 10.83 | Improved |
| FreeDoM instead of UGD | style-guided 512x512 generation · Style-guided generation | sampling time: about 40 minutes -> about 84 seconds | Improved |
| add time-travel strategy (ablation) | FreeDoM on ImageNet · ImageNet text 'orange' | qualitative condition match | Improved |
| apply time-travel only in semantic stage (ablation) | FreeDoM · ImageNet | sampling cost | Improved |
| increase energy-function learning rate rho_t from 0 (ablation) | FreeDoM · ImageNet and face ID | control strength | Improved |
Later papers that compared with this method
19 later accepted papers report FreeDoM as a baseline in their results tables; 17 of them report a better number on most of the results they share with it.
- Efficient Weighted Sampling via Score-based Generative Models (CVPR 2026 · better on 3 of 3 shared results)
- Spectral Guidance for Flexible and Efficient Control of Diffusion Models (ICML 2026 · better on 1 of 1 shared result)
- DICT: Data Injection and Contrastive Trajectory Refinement for Conditional Image Generation with Diffusion Models (ECCV 2026 · better on 1 of 1 shared result)
- NoiseTilt: Noise-Tilted Reverse Kernels for Diffusion Reward Alignment (ECCV 2026 · better on 1 of 1 shared result)
- DiffRGD: An Inference-Time Diffusion Guidance Through Riemannian Gradient Descent (ECCV 2026 · better on 1 of 1 shared result)
- Z-Magic: Zero-shot Multiple Attributes Guided Image Creator (CVPR 2025 · better on 8 of 8 shared results)
- ORIGEN: Zero-Shot 3D Orientation Grounding in Text-to-Image Generation (NeurIPS 2025 · better on 4 of 5 shared results)
- NoiseCtrl: A Sampling-Algorithm-Agnostic Conditional Generation Method for Diffusion Models (CVPR 2025 · better on 3 of 4 shared results)
11 more not listed.
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- TediGAN — TediGAN: Text-Guided Diverse Face Image Generation and Manipulation (CVPR 2021) · better on 5 of 6 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- No Training, No Problem: Rethinking Classifier-Free Guidance for Diffusion Models (ICLR 2025 · similarity 0.90)
- TFG: Unified Training-Free Guidance for Diffusion Models (NeurIPS 2024 · similarity 0.89)
- Random Conditioning for Diffusion Model Compression with Distillation (CVPR 2025 · similarity 0.89)
- Understanding and Improving Training-free Loss-based Diffusion Guidance (NeurIPS 2024 · similarity 0.88)
- Data-free Distillation of Diffusion Models with Bootstrapping (ICML 2024 · similarity 0.88)
- SaRA: High-Efficient Diffusion Model Fine-tuning with Progressive Sparse Low-Rank Adaptation (ICLR 2025 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
