跑YOLOv5遇到的问题

一、RuntimeError: a view of a leaf Variable that requires grad is being used in an in-place operation

解决办法：在第148行下添加一行with torch.no_grad():

参考文章：RuntimeError: a view of a leaf Variable that requires grad is being used in an in-place operation_一起变胖吧的博客-CSDN博客

二、RuntimeError: No such operator torchvision::nms

解决办法：pytorch和torchvision的版本不匹配，查找相对应的版本并重装

参考文章：RuntimeError: No such operator torchvision::nms问题解决方法_StrongerTang的博客-CSDN博客

arm64下安装pytorch,torchvision,torchaudio以及它们之间的版本对应关系_SimonLiu009的博客-CSDN博客_arm64 pytorch

三、TypeError: can't convert cuda:0 device type tensor to numpy. Use Tensor.

解决办法：根据提示在tensor.py中第678行添加.cpu()

参考文章：Pycharm中运行yolov5-pytorch出现错误“TypeError: can‘t convert cuda:0 device type tensor to numpy....“_明德zhuang的博客-CSDN博客四、AttributeError: 'Hardswish' object has no attribute 'inplace'

解决办法：根据提示定位到activation.py 中第475行，将input后的删掉

 参考文章：yolov5报错：AttributeError: ‘Hardswish‘ object has no attribute ‘inplace‘_长沙有肥鱼的博客-CSDN博客

五、AttributeError: ‘Upsample‘ object has no attribute ‘recompute_scale_factor‘

解决办法：在detect.py运行过程中报错，根据提示定位到upsampling.py中的153行，将代码改成如下

参考文章： AttributeError: ‘Upsample‘ object has no attribute ‘recompute_scale_factor‘_阳光不锈@的博客-CSDN博客

六、UnicodeDecodeError: 'utf-8' codec can't decode byte 0xd3 in position 357: invalid continuation byte

解决办法：在…/yolov5-master/train.py中把with open(opt.hyp) as f:改为with open(opt.hyp, encoding='gb2312') as f:
 

参考文章：

YOLOv5中‘utf-8‘ codec can‘t decode编码问题-CSDN博客