Contrastive CFG: Guiding Diffusion Sampling by Contrasting Positive and Negative Concepts

Jinho Chang, Changsun Lee, Hyungjin Chung, Jong Chul Ye

Keywords: Classifier-free guidance, Diffusion models, Guided sampling

TL;DR: We present a novel method to enhance negative guidance using contrastive loss, which resolves the downsides of previously used negative CFG.

Abstract: As Classifier-Free Guidance (CFG) has proven effective in conditional diffusion model sampling for improved condition alignment, many applications use a negated CFG term as a Negative Prompting (NP) to filter out unwanted features from samples. However, simply negating CFG guidance creates an inverted probability distribution, often distorting samples away from the marginal distribution. Inspired by recent advances in conditional diffusion models for inverse problems, here we present a novel method to achieve guidance toward the given condition using contrastive loss. Specifically, our guidance term aligns or repels the denoising direction based on the given condition through contrastive loss, achieving a similar guiding effect to traditional CFG for positive conditions while overcoming the limitations of existing negative guidance methods. Experimental results demonstrate that our approach effectively injects or removes the given concepts while maintaining sample quality across diverse scenarios, from simple class conditions to complex and overlapping text prompts.

Primary Area: generative models

Submission Number: 5624