提出SSR-Merge方法，无需训练即可高效融合多个LoRA，避免参数冲突。
SSR-Merge: Subspace Signal Routing for Training-Free LoRA Merging in Diffusion Models
- 通过子空间信号路由，不合并参数而分流信号，解决干扰问题。
- 在12个扩散模型任务上平均性能提升6.3%，超越现有最优方法。
- 支持流式更新，内存占用降低78%，适合快速迭代与资源受限场景。
低秩适配（LoRA）融合可高效整合多个训练好的LoRA所携带的生成能力，用于扩散模型。然而，现有融合方法常因共享参数空间中的严重参数干扰，导致破坏性冲突。为此，我们提出子空间信号路由（SSR），通过路由内部信号而非合并参数空间来化解干扰。具体而言，SSR首先沿秩维度拼接候选LoRA构建统一子空间；接着利用逆相关矩阵对空间内混杂信号进行去相关；最后通过方向引导矩阵将净化后的信号导向各自的任务专属子空间。我们提供了严格的理论分析，证明SSR与普通最小二乘法（OLS）解一致，保证数学最优性。利用充分统计量的可加性，设计了流式算法，实现即时更新，显著降低内存开销和计算时间。大量实验验证，SSR显著优于当前最先进方法，同时保持相近效率。代码已公开于https://github.com/nagara214/SSR-Merge。
原文摘要 · Abstract (English)
Low-Rank Adaptation (LoRA) merging can efficiently combine diverse generative capabilities from multiple trained LoRAs for a diffusion model. However, existing LoRA merging techniques often suffer from severe parameter interference, causing destructive collisions in the shared parameter space. To address this, we propose Subspace Signal Routing (SSR), which resolves interference by routing internal signals instead of performing parameter-space merge. Specifically, SSR first constructs a unified subspace by concatenating candidate LoRAs along the rank dimension. Next, SSR employs an inverse correlation matrix to decorrelate mixed signals within this space. Finally, a directional guide matrix steers these purified signals into their respective task-specific subspaces. We provide a rigorous theoretical analysis proving that SSR aligns with the Ordinary Least Squares (OLS) solution, thereby ensuring mathematical optimality. We utilize the additivity of sufficient statistics to design a streaming algorithm. This enables on-the-fly updates that significantly reduce memory overhead and computation time. Extensive experiments validate that SSR significantly outperforms state-of-the-art methods while maintaining comparable efficiency. Code is available at https://github.com/nagara214/SSR-Merge.
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。