
CV
望长安于日下
静以修身
展开
-
安装yolox
https://github.com/Megvii-BaseDetection/YOLOXhttps://zhuanlan.zhihu.com/p/392570215安装cocoapimkdir cocoapi && cd cocoapigit clone --filter=blob:none -q https://github.com/cocodataset/cocoapi.git .pip3 install cythonpython3 setup.py install --u原创 2021-12-05 21:50:06 · 609 阅读 · 0 评论 -
标签平滑和样本类别不均衡处理方式
文章目录1. 论文1.1 Label Smoothing1.2 Focal Loss2. 代码实现2.1 Label Smoothing loss function2.2 labels mooth+类别不均衡loss1. 论文1.1 Label Smoothinghttps://arxiv.org/abs/1906.02629 When Does Label Smoothing Help?https://arxiv.org/abs/1701.06548 Regularizing Neural Ne原创 2021-07-16 14:39:38 · 925 阅读 · 1 评论 -
Pytorch实现EfficientNet
1. 概述1.1 论文下载EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks百度云下载,提取码1.2 论文解读参考:https://blog.youkuaiyun.com/qq_37541097/article/details/1144340462.性能对比在这里插入图片描述3. 网络backboneModelinput_sizewidth_coefficientdepth_coeffici原创 2021-06-26 12:33:56 · 2731 阅读 · 0 评论 -
CV算法整理
1.vggNetEfficientNet原创 2021-06-26 12:11:14 · 590 阅读 · 0 评论