答疑很快 https://github.com/floodsung/Deep-Learning-Papers-Reading-Roadmap
迁移学习 https://github.com/jindongwang
先将labelImg标记的多个xml文件转化成一个文件然后转成csv格式,地址:https://www.cnblogs.com/lky-learning/p/11640180.html。 然后按比例随机划分数据集csv2voc/coco/labelme,需要用到的程序地址为:https://github.com/spytensor/prepare_detection_dataset