【目标检测-YOLO】YOLOv5-6.1v实践过程

1. 下载源码：
wget https://github.com/ultralytics/yolov5/archive/refs/tags/v6.1.zip

unzip v6.1.zip && rm v6.1.zip

2. 添加 yolov5-6.1/data/VOCTVlogo2022.yaml。这里我是YOLO v5 训练自定义数据集_星魂非梦的博客-CSDN博客 复制的。

3. 下载预训练权重:

wget https://github.com/ultralytics/yolov5/releases/download/v6.1/yolov5s6.pt

4. 修改超参数文件：hyp.scratch-TVlogo2022.yaml。下面是我修改后的。

lr0: 0.01  # initial learning rate (SGD=1E-2, Adam=1E-3)
lrf: 0.1  # final OneCycleLR learning rate (lr0 * lrf)
momentum: 0.937  # SGD momentum/Adam beta1
weight_decay: 0.0005  # optimizer weight decay 5e-4
warmup_epochs: 3.0  # warmup epochs (fractions ok)
warmup_momentum: 0.8  # warmup initial momentum
warmup_bias_lr: 0.1  # warmup initial bias lr
box: 0.05  # box loss gain
cls: 0.5  # cls loss gain
cls_pw: 1.0  # cls BCELoss positive_weight
obj: 0.7  # obj