Hyper-SD: Trajectory Segmented Consistency Model for Efficient Image Synthesis

TL;DR

Hyper-SD introduces a novel trajectory segmented consistency distillation framework that significantly improves efficient image synthesis by preserving diffusion model trajectories and incorporating human feedback, achieving state-of-the-art results with fewer inference steps.

Contribution

The paper proposes Hyper-SD, a new method combining trajectory preservation and reformulation with human feedback, enabling near-lossless step compression in diffusion models.

Findings

Achieves state-of-the-art performance with 1 to 8 inference steps.

Surpasses existing models in CLIP and Aes scores at low steps.

Demonstrates effectiveness on SDXL and SD1.5 datasets.

Abstract

Recently, a series of diffusion-aware distillation algorithms have emerged to alleviate the computational overhead associated with the multi-step inference process of Diffusion Models (DMs). Current distillation techniques often dichotomize into two distinct aspects: i) ODE Trajectory Preservation; and ii) ODE Trajectory Reformulation. However, these approaches suffer from severe performance degradation or domain shifts. To address these limitations, we propose Hyper-SD, a novel framework that synergistically amalgamates the advantages of ODE Trajectory Preservation and Reformulation, while maintaining near-lossless performance during step compression. Firstly, we introduce Trajectory Segmented Consistency Distillation to progressively perform consistent distillation within pre-defined time-step segments, which facilitates the preservation of the original ODE trajectory from a higher-order perspective. Secondly, we incorporate human feedback learning to boost the performance of the model in a low-step regime and mitigate the performance loss incurred by the distillation process. Thirdly, we integrate score distillation to further improve the low-step generation capability of the model and offer the first attempt to leverage a unified LoRA to support the inference process at all steps. Extensive experiments and user studies demonstrate that Hyper-SD achieves SOTA performance from 1 to 8 inference steps for both SDXL and SD1.5. For example, Hyper-SDXL surpasses SDXL-Lightning by +0.68 in CLIP Score and +0.51 in Aes Score in the 1-step inference.