Revisiting Classifier-Free Guidance Methods in Latent Diffusion Models
2026-08-17 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors looked at several ways to improve image generation from diffusion models without retraining them. These methods were originally tested on older models and focused only on image quality, ignoring how well the images matched the text prompts. When the authors tested these methods on newer transformer models using benchmarks that check image-text alignment, none consistently worked better than the standard Classifier-Free Guidance (CFG). Some methods showed slight improvements but often within measurement uncertainty, while others sometimes made results worse. Overall, CFG remained a reliable and simple method.
diffusion modelsClassifier-Free Guidanceinference-time enhancementU-Nettransformerimage-text alignmentcompositional alignmentattention perturbationmetricsimage generation
Authors
Artem Sergievskii, Artyom Turevich, Sergey Kastryulin
Abstract
Inference-time quality-enhancement methods are an effective and widely adopted means of improving diffusion models without expensive retraining. We study a family of training-free techniques conceptually rooted in Classifier-Free Guidance (CFG), most of which were originally proposed on older U-Net diffusion models and validated using metrics that assess image quality in isolation, without accounting for compositional alignment or semantic correspondence between the generated image and its associated text prompt. We re-evaluate eight such methods on two open-weight rectified-flow transformers under a fixed per-model protocol and three compositional-alignment benchmarks. No method consistently improves on CFG across the measured criteria. APG obtains several nominal best scores, but the corresponding gains often remain within the estimated evaluation uncertainty. Attention-perturbation methods provide isolated gains on SD3.5 Medium and more frequent degradations on FLUX.2 [klein] 4B Base, while CFG remains a competitive lower-cost baseline.