CrossDistill: Balancing Quality and Diversity via Trajectory-Level Hybrid Few-Step Distillation

2026-09-13

AI Summary

本文提出CrossDistill方法，通过在不同噪声阶段应用不同的目标函数来平衡生成模型的质量与多样性问题。

Abstract

Few-step distillation accelerates diffusion models but must balance diversity and fidelity: trajectory-based distillation preserves mode coverage, while distribution matching sharpens samples but can reduce diversity. We show that this tension can be exploited in a noise-regime-dependent way: high-noise steps largely determine global modes, whereas low-noise steps refine local details. We propose CrossDistill, a trajectory-level hybrid distillation framework that splits the sampling trajectory at a crossover point, applies a trajectory-preserving objective on the high-noise interval and a distribution-matching objective on the low-noise interval, and couples the two stages through the crossover state. In contrast to loss-level mixing, and complementarily to training-time two-stage recipes, CrossDistill explicitly assigns complementary objectives along the noise axis, so that global branching is preserved before local statistics are sharpened. CrossDistill is a noise-level scheduling policy: PCM and DMD are plug-in instantiations, while the noise partition, crossover coupling, and objective ordering are the key design elements. Experiments on text-to-video diffusion models and qualitative image-to-video results show that CrossDistill expands the few-step quality-diversity frontier, retaining seed-level variation while achieving competitive visual fidelity.

Problem

• few-step distillation
• diversity and fidelity
• diffusion models

Innovation

• CrossDistill
• trajectory-level hybrid distillation
• noise-regime-dependent
• mode coverage preservation
• local detail refinement