发表机构
Institute of Advanced Intelligence and Computing (IAIC), A*STAR; Nanyang Technological University（先进智能与计算研究所（IAIC），新加坡科技研究局; 南洋理工大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本研究提出无标签的DiD方法，通过对齐检测器接口张量，将Softmax注意力ViT主干转换为线性注意力主干，在DOTA-v1.5上性能优异且耗时短，能大幅降低检测的推理延迟与内存占用。
AI 中文摘要
线性注意力因能降低全局token混合的成本，成为高分辨率目标检测中颇具吸引力的机制，但将已训练检测器的Softmax注意力ViT主干转换为线性注意力主干并非简单的直接替换。直接替换注意力算子会导致严重的性能下降，而通用无标签蒸馏虽对分类任务有效，却常在检测任务上失效。我们认为核心挑战在于“检测器接口保留”：转换后的主干必须重现固定下游检测器所需的精确特征张量，而非仅模仿内部Softmax隐藏状态。为解决这一问题，我们提出检测器接口蒸馏（Detector-Interface Distillation, DiD），这是一种无标签转换方法，仅通过对齐面向检测器的接口张量与冻结的Softmax教师模型的对应张量，来训练线性注意力主干。在DOTA-v1.5数据集上，DiD的性能显著优于现有基线，且与经过全监督训练的线性模型相当。该适配过程在4块GPU上仅需约87分钟即可完成，线性化后的主干将推理延迟降低约62%，峰值内存减少约49%。我们希望这些发现为社区提供一种简单的无标签路径，将已训练的Softmax检测器复用为高效的线性检测器，并鼓励未来的架构转换工作采用感知接口的目标函数。
英文摘要
While linear attention is a compelling mechanism for high-resolution object detection due to its reduced cost for global token mixing, converting the Softmax-attention ViT backbone of a trained detector into a linear-attention one is not a trivial drop-in replacement. Directly swapping the attention operator leads to severe performance degradation, and generic label-free distillation, though effective for classification, often fails on detection tasks. We argue that the central challenge is \textit{detector-interface preservation}: the converted backbone must reproduce the exact feature tensors expected by the fixed downstream detector, rather than merely imitating internal Softmax hidden states. To address this, we introduce Detector-Interface Distillation (DiD), a label-free conversion method that exclusively trains the linear-attention backbone by aligning detector-facing interface tensors with those of a frozen Softmax teacher. On DOTA-v1.5, DiD substantially outperforms established baselines and matches supervised, fully trained linear models. Adaptation completes in roughly 87 minutes on 4 GPUs, and the linearized backbone cuts inference latency by ~62% and peak memory by ~49%. We hope our findings offer the community a simple, label-free route to reusing trained Softmax detectors as efficient linear ones, and encourage interface-aware objectives in future architecture-conversion work.
