本质子空间中的模型合并
Model Merging in the Essential Subspace
对需要把多个任务微调模型（如多个下游视觉任务的 LoRA/全参微调权重）零训练合成一个多任务模型的场景，提供基于 PCA 特征偏移子空间投影 + 低秩分解 + 极化缩放的可复用合并流程，可直接用于视觉基础模型的多任务部署。
原文依据 Each task's parameter update matrix is projected onto its respective essential subspace for low-rank decomposition before merging.
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论模型合并旨在将源自同一预训练检查点的多个任务特定微调模型整合为单一多任务模型，且无需额外训练。尽管已有大量研究，任务干扰仍是主要障碍，常常削弱合并模型的性能。本文提出 ESM（本质子空间合并），一个用于有效模型合并的稳健框架。我们首先对参数更新所引发的特征偏移进行主成分分析（PCA）。所得主方向张成一个本质子空间，该子空间对特征表示起主导影响。在合并之前，每个任务的参数更新矩阵被投影到其各自的本质子空间上以进行低秩分解。该方法在保留核心任务特定功能的同时缓解了任务间干扰。此外，我们引入多级极化缩放策略，放大包含关键知识的参数并抑制冗余参数，防止本质知识在融合过程中被淹没。在多个任务集与模型规模上的大量实验表明，我们的方法在多任务模型合并中取得了最优性能。
英文原文摘要 1249 字符
Model merging aims to integrate multiple task-specific fine-tuned models derived from a shared pre-trained checkpoint into a single multi-task model without additional training. Despite extensive research, task interference remains a major obstacle that often undermines the performance of merged models. In this paper, we propose ESM (Essential Subspace Merging) , a robust framework for effective model merging. We begin by performing Principal Component Analysis (PCA) on feature shifts induced by parameter updates. The resulting principal directions span an essential subspace that dominantly influences feature representations. Each task's parameter update matrix is projected onto its respective essential subspace for low-rank decomposition before merging. This methodology mitigates inter-task interference while preserving core task-specific functionality. Furthermore, we introduce a multi-level polarized scaling strategy that amplifies parameters containing critical knowledge and suppresses redundant ones, preventing essential knowledge from being overwhelmed during fusion. Extensive experiments across multiple task sets and model scales demonstrate that our method achieves state-of-the-art performance in multi-task model merging.