Research map / Image generation / Fast diffusion sampling
Overshoot and Shrinkage in Classifier-Free Guidance: From Theory to Practice
Krunoslav Lehman Pavasovic, Jakob Verbeek, Giulio Biroli, Marc Mezard
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 |
|---|---|---|---|
| add power-law nonlinearity (alpha=0.9) to standard CFG | standard CFG · EDM2-S (class-cond, ImageNet-1K 512) | FID: 2.29 -> 1.93 | Improved |
| add power-law nonlinearity to standard CFG | standard CFG · DiT/XL-2 (class-cond, ImageNet-1K 256) | FID: 2.27 -> 2.05 | Improved |
| add power-law nonlinearity to standard CFG | standard CFG · Diff. MMDiT (T2IM, CC12M) | FID: 8.58 -> 8.11 | Improved |
| add power-law nonlinearity to standard CFG | standard CFG · FM MMDiT (T2IM, COCO) | FID: 5.20 -> 4.81 | Improved |
| combine power-law with CADS | CADS · EDM2-S (class-cond, ImageNet-1K 512) | FID: 1.60 -> 1.52 | Improved |
| combine power-law with limited-interval guidance | Limited interval · EDM2-S (class-cond, ImageNet-1K 512) | FID: 1.87 -> 1.73 | Improved |
| combine power-law with limited-interval guidance | Limited interval · EDM2-S (class-cond, ImageNet-1K 512) | precision: 0.760 -> 0.752 | Hurt |
| add power-law nonlinearity to standard CFG | standard CFG · FM MMDiT (T2IM, COCO) | precision: 0.629 -> 0.621 | Hurt |
| add power-law nonlinearity to standard CFG | standard CFG · Diff. MMDiT (T2IM, CC12M) | recall: 0.569 -> 0.553 | Hurt |
| increase alpha of power-law CFG (ablation) | power-law CFG · EDM2-S ImageNet-1K 512 | FID | Improved |
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.
- CADS — CADS: Unleashing the Diversity of Diffusion Models through Condition-Annealed Sampling (ICLR 2024) · 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
- Rectified Diffusion Guidance for Conditional Generation (CVPR 2025 · similarity 0.93)
- Eliminating Oversaturation and Artifacts of High Guidance Scales in Diffusion Models (ICLR 2025 · similarity 0.92)
- CFG++: Manifold-constrained Classifier Free Guidance for Diffusion Models (ICLR 2025 · similarity 0.92)
- Adaptive Guidance: Training-free Acceleration of Conditional Diffusion Models (AAAI 2025 · similarity 0.91)
- No Training, No Problem: Rethinking Classifier-Free Guidance for Diffusion Models (ICLR 2025 · similarity 0.91)
- Rethinking the Spatial Inconsistency in Classifier-Free Diffusion Guidance (CVPR 2024 · similarity 0.91)
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.
