北京航空航天大学学报2026，Vol.52Issue(9)：3066-3074,9.DOI:10.13700/j.bh.1001-5965.2025.0411
面向无人机平台的CNN-Transformer混合模型训练后量化方法
A post-training quantization method for CNN-Transformer hybrid models on UAV platforms
赵呈功 1胡纪远 2杨星宇 3陈磊 4姜宏旭5
作者信息
- 1. 北京航空航天大学杭州创新研究院,杭州 310056||北京航空航天大学 软件学院,北京 100191
- 2. 北京航空航天大学杭州创新研究院,杭州 310056||北京航空航天大学 计算机学院,北京 100191
- 3. 内蒙古大学 计算机学院(软件学院),呼和浩特 010021
- 4. 杭州萧山国际机场飞行区中心,杭州 311241
- 5. 北京航空航天大学 计算机学院,北京 100191
- 折叠
摘要
Abstract
In recent years,drones have played increasingly important roles in various low-altitude production and service scenarios.Due to the limited storage and computing resources of edge computing platforms,the deployment performance of vision models becomes a bottleneck for various algorithm applications.The long-tail distribution of activation values in convolutional neural network(CNN)-Transformer hybrid models causes substantial quantization accuracy deterioration,an efficient post-training quantization method is proposed,which this study addresses by suggesting an effective post-training quantization technique.This method uses activation noise compensation(ANC)and adaptive difficulty migration to suppress the impact of outliers on quantization accuracy and improve inference efficiency.According to experimental results,the proposed method achieves less than 1%accuracy loss on typical models under 8-bit quantization,and it improves inference time after quantization by at most 200%.In summary,the proposed method significantly enhances vision model inference performance on edge devices and supports model deployment on low-altitude drone platforms.关键词
训练后量化/模型压缩/推理加速/异常值感知/边缘计算Key words
post-training quantization/model compression/inference acceleration/outlier-aware/edge computing分类
航空航天引用本文复制引用
赵呈功,胡纪远,杨星宇,陈磊,姜宏旭..面向无人机平台的CNN-Transformer混合模型训练后量化方法[J].北京航空航天大学学报,2026,52(9):3066-3074,9.
