Guidance Free Image Editing via Explicit Conditioning
This addresses the problem of excessive inference time for researchers and practitioners using diffusion models in image editing, though it is incremental as it builds on existing guidance techniques.
The paper tackles the high computational cost of Classifier Free Guidance (CFG) in conditional diffusion models for image editing by introducing Explicit Conditioning (EC), which models noise to guide the process, resulting in significantly reduced computations while outperforming CFG in generating diverse high-quality images.
Current sampling mechanisms for conditional diffusion models rely mainly on Classifier Free Guidance (CFG) to generate high-quality images. However, CFG requires several denoising passes in each time step, e.g., up to three passes in image editing tasks, resulting in excessive computational costs. This paper introduces a novel conditioning technique to ease the computational burden of the well-established guidance techniques, thereby significantly improving the inference time of diffusion models. We present Explicit Conditioning (EC) of the noise distribution on the input modalities to achieve this. Intuitively, we model the noise to guide the conditional diffusion model during the diffusion process. We present evaluations on image editing tasks and demonstrate that EC outperforms CFG in generating diverse high-quality images with significantly reduced computations.