wkgit88

mAP计算过程中预测结果异常

10

作者您好，我用自己数据集训练的模型mAP很低（同样数据集用您的yolov8接近您voc的mAP），打开detection-result发现输出结果中很多top和bottom值都是负数，比如person 0.0059 523 -114 545 -77，我对预测结果取绝对值后mAP没有改变，说明不是单纯的取反而是溢出了，请问这种情况改怎么修改呢？ 

RuntimeError: main thread is not in main loop

1

When I was training in BDD100k , error occurred. I changed the num_workers from 8 to 1 but this still happened: xception ignored in: Traceback (most recent call last): File...