Target-Driven Distillation: Consistency Distillation with Target Timestep Selection and Decoupled Guidance
This work addresses image quality issues in generative tasks for researchers and practitioners using consistency distillation, though it appears incremental as it builds on existing methods.
The paper tackles the problem of blur and detail loss in consistency distillation methods for diffusion models by introducing Target-Driven Distillation (TDD), which improves training efficiency and flexibility, achieving state-of-the-art performance in few-step generation.
Consistency distillation methods have demonstrated significant success in accelerating generative tasks of diffusion models. However, since previous consistency distillation methods use simple and straightforward strategies in selecting target timesteps, they usually struggle with blurs and detail losses in generated images. To address these limitations, we introduce Target-Driven Distillation (TDD), which (1) adopts a delicate selection strategy of target timesteps, increasing the training efficiency; (2) utilizes decoupled guidances during training, making TDD open to post-tuning on guidance scale during inference periods; (3) can be optionally equipped with non-equidistant sampling and x0 clipping, enabling a more flexible and accurate way for image sampling. Experiments verify that TDD achieves state-of-the-art performance in few-step generation, offering a better choice among consistency distillation models.