无需训练数据，跨模型迁移扩散模型的生成知识。
Delta Sampling: Data-Free Knowledge Transfer Across Diffusion Models
- 利用模型适配前后的预测差异（delta）指导新模型去噪。
- 在不同Stable Diffusion版本间迁移风格、概念和结构效果，提升稳定。
- 纯推理阶段操作，可直接替换原有适配组件，即插即用。
Stable Diffusion等扩散模型推动了开源生态的发展，包括全微调检查点和参数高效适配器（如LoRA、LyCORIS、ControlNet）。然而，这些适配组件与特定基础模型紧密耦合，当基础模型升级（如从SD 1.x到2.x）时，因参数和架构变化难以复用。本文提出Delta Sampling（DS），一种无需原始训练数据即可在不同架构的基础模型间实现知识迁移的新方法。DS在推理阶段运行，通过利用适配前后模型预测的差值（delta）来引导新模型的去噪过程。我们在多个Stable Diffusion版本上评估了DS，结果表明其在不同采样策略下均能一致提升生成效果（如视觉风格、语义概念、结构控制）。这证明了DS是一种高效、即插即用的扩散图像生成知识迁移机制。代码已开源。
原文摘要 · Abstract (English)
Diffusion models like Stable Diffusion (SD) drive a vibrant open-source ecosystem including fully fine-tuned checkpoints and parameter-efficient adapters such as LoRA, LyCORIS, and ControlNet. However, these adaptation components are tightly coupled to a specific base model, making them difficult to reuse when the base model is upgraded (e.g., from SD 1.x to 2.x) due to substantial changes in model parameters and architecture. In this work, we propose Delta Sampling (DS), a novel method that enables knowledge transfer across base models with different architectures, without requiring access to the original training data. DS operates entirely at inference time by leveraging the delta: the difference in model predictions before and after the adaptation of a base model. This delta is then used to guide the denoising process of a new base model. We evaluate DS across various SD versions, demonstrating that DS achieves consistent improvements in creating desired effects (e.g., visual styles, semantic concepts, and structures) under different sampling strategies. These results highlight DS as an effective, plug-and-play mechanism for knowledge transfer in diffusion-based image synthesis. Code:~ https://github.com/Zhidong-Gao/DeltaSampling
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。