YOLO-PEFT让轻量微调适配实时目标检测器，自动规划可执行的参数高效微调方案。
YOLO-PEFT: Parameter-Efficient Fine-Tuning on YOLO Family
- 基于检测器图结构智能规划适配器位置，确保符合硬件与模型约束。
- 在YOLO11s和YOLO12s上实现mAP50-95提升至0.7138和0.7307，优于全参数微调。
- 支持拒绝不合理配置，适合需要可解释性与部署保障的工业级检测场景。
从语言模型迁移而来的通用参数高效微调（PEFT）方法在实时检测器上可能失效，因其异构算子与检测特有组件带来布局约束，这些约束在标准Transformer中并不存在。本文提出YOLO-PEFT，一个结构感知框架，将适配器放置建模为可审计的约束规划问题。给定检测器图、微调请求与资源预算，该框架分配算子与语义角色，评估算子有效性、检测语义、图接口及部署谓词，并为每个被排除模块记录原因码，最终输出预算内目标模块计划或提前拒绝训练。在官方VOC07+12训练集到VOC07测试集协议下，规划选定的RS-LoRA在YOLO11s和YOLO12s上分别达到0.7138和0.7307 mAP50-95，优于全参数微调的0.6428和0.6662。在RT-DETR-L上，所有七种评估的LoRA配置均触发预设灾难阈值，支持在覆盖范围内做出拒绝转全微调的校准决策。控制实验显示，LoRA使峰值训练内存降低43.9%，但训练时间延长1.72倍。在已评估的检测器族、放置策略与校准覆盖范围内，YOLO-PEFT以可检查的规划替代手动试错，同时保留验证过的训练-保存-合并-导出路径；对未见检测器架构的拒绝仍属开放验证问题。
原文摘要 · Abstract (English)
Generic parameter-efficient fine-tuning (PEFT) methods transferred from language models can fail silently on real-time detectors, whose heterogeneous operators and detection-specific components impose placement constraints absent from regular Transformer stacks. We propose YOLO-PEFT, a structure-aware framework that formulates adapter placement as an auditable constraint-planning problem. Given a detector graph, a PEFT request, and a resource budget, YOLO-PEFT assigns operator and semantic roles, evaluates explicit operator-validity, detector-semantic, graph-interface, and deployment predicates, records a reason code for each excluded module, and either emits a budgeted target-module plan or returns Refuse before training. Under the official VOC07+12 trainval-to-VOC07 test protocol, planner-selected RS-LoRA reaches 0.7138 and 0.7307 mAP50-95 on YOLO11s and YOLO12s, respectively, compared with 0.6428 and 0.6662 for Full-SFT. On RT-DETR-L, all seven evaluated LoRA-family configurations cross the predefined catastrophic threshold, supporting a calibrated Refuse-to-Full-SFT decision within the evaluated coverage. A controlled YOLO11 audit further shows that LoRA reduces peak training memory by 43.9 percent, although training takes 1.72 times longer. Within the evaluated detector families, placement policies, and calibration coverage, YOLO-PEFT replaces manual target-module trial and error with explicit, inspectable planning while preserving verified train-save-merge-export paths; refusal on unseen detector architectures remains an open validation problem. Project Page: github.com/Tencent/YOLO-Master
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。