免训练晶圆图缺陷分类：规则排序提示词让多模态模型达到93.8%准确率
Zero-Shot Multi-Class Wafer Map Classification with Rule-Ordered Prompts for Large Multimodal Models
晶圆缺陷检测不用再标注数据重新训练了，论文用提示词排序技巧让冻结多模态模型在WM811K上跑到81.3%，做工业质检的可以看看。
一篇arXiv论文提出免训练的六类晶圆图缺陷分类框架，将连通域高亮与形态、位置、排除感知提示词结合，调用冻结的多模态大模型完成分类。在Mixed-WM38数据集的6,866张晶圆图上，多区域提示达到92.9%宏平均和93.8%总体准确率。在WM811K数据集的9,000张图上，两项指标分别为79.1%和81.3%。论文还测试了8种规则排序，总体准确率在90.1-91.9%之间，低于选定排序的93.8%。
Zero-Shot Multi-Class Wafer Map Classification with Rule-Ordered Prompts for Large Multimodal Models
Wafer-map defect classification often requires labeled data and retraining when production conditions change. We present a training-free six-class framework that combines connected-component highlighting with morphology-, location-, and exclusion-aware prompts for a frozen large multimodal model. On Mixed-WM38 (6,866 maps), multi-region prompting achieves 92.9% macro and 93.8% overall accuracy. On WM811K (9,000 maps), it achieves 79.1% and 81.3%, respectively. Eight alternative rule orders achieve 90.1-91.9% overall accuracy, compared with 93.8% for the selected order.
