Adaptive Sampling Scheduler

2025-09-15

AI Summary

Existing consistency distillation methods rely on predefined or random timestep schedules, lacking generality and adaptability—thereby limiting sampling flexibility across diverse distillation frameworks and compromising stability under high classifier-free guidance (CFG) scales. This work proposes a universal, adaptive sampling scheduler: it dynamically selects critical timesteps based on timestep importance, designs an optimization trajectory via alternating forward denoising and backward noising guidance, and incorporates smooth clipping and color balancing to enhance generation robustness at high CFG weights. The scheduler requires no framework-specific customization and is plug-and-play. Experiments demonstrate significant improvements in both generation quality and inference efficiency across multiple consistency distillation architectures. Notably, it maintains stable, high-fidelity outputs under extreme CFG conditions (e.g., CFG ≥ 10), thereby extending the practical applicability of diffusion models to demanding generative tasks.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Planning, Routing, and Scheduling: Planning with Language Models
• Machine Learning: Calibration & Uncertainty Quantification

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Consistent distillation methods have evolved into effective techniques that significantly accelerate the sampling process of diffusion models. Although existing methods have achieved remarkable results, the selection of target timesteps during distillation mainly relies on deterministic or stochastic strategies, which often require sampling schedulers to be designed specifically for different distillation processes. Moreover, this pattern severely limits flexibility, thereby restricting the full sampling potential of diffusion models in practical applications. To overcome these limitations, this paper proposes an adaptive sampling scheduler that is applicable to various consistency distillation frameworks. The scheduler introduces three innovative strategies: (i) dynamic target timestep selection, which adapts to different consistency distillation frameworks by selecting timesteps based on their computed importance; (ii) Optimized alternating sampling along the solution trajectory by guiding forward denoising and backward noise addition based on the proposed time step importance, enabling more effective exploration of the solution space to enhance generation performance; and (iii) Utilization of smoothing clipping and color balancing techniques to achieve stable and high-quality generation results at high guidance scales, thereby expanding the applicability of consistency distillation models in complex generation scenarios. We validated the effectiveness and flexibility of the adaptive sampling scheduler across various consistency distillation methods through comprehensive experimental evaluations. Experimental results consistently demonstrated significant improvements in generative performance, highlighting the strong adaptability achieved by our method.

Problem

• Adaptive scheduler for flexible diffusion model distillation
• Dynamic timestep selection to enhance sampling efficiency
• Stabilizing high-guidance generation in complex scenarios

Innovation

• Dynamic target timestep selection
• Optimized alternating sampling trajectory
• Smoothing clipping and color balancing