在使用yolov7-pose训练时候和yolov7训练时候唯一的区别是:
1..yaml配置文件
# train and val data as 1) directory: path/images/, 2) file: path/images.txt, or 3) list: [path1/images/, path2/images/] train: datasets/train2017.txt # 128 images val: datasets/val2017.txt # 128 images # number of classes nc: 1 # class names names: [ 'person' ]
2.标注格式:
3.数据集包如下:
4.train.py参数里面区别(解决corrupted):
使用编译器运行时候:parser.add_argument('--kpt-label', default='True',action='store_true', help='use keypoint labels for training')一定要设置成默认值为True,很重要!!!!
使用命令行运行则:python train.py --data data/coco_kpts_128.yaml --cfg cfg/yolov7-w6-pose.yaml --weights yolov5s.pt --kpt-label