跟着小土堆的视频教学自己遇到的一些问题。
出现错误的原因:由于yolov5目前最新版本为v6.1,但我跑的是5.0版本,则运行detect.py时自动从github上下载的训练好的模型为最新版本v6.1。从而导致运行环境和模型版本不一致,从而报错。
一、AttributeError: Can‘t get attribute ‘SPPF‘ on <module ‘models.common‘ from ‘H:\\yolov5-5.0\\models\\
二、yolov5 ERROR: AttributeError: ‘Upsample‘ object has no attribute ‘recompute_scale_factor‘
三、yolov5中The size of tensor a (80) must match the size of tensor b (56) at non-singleton dimension 3
四、UserWarning: torch.meshgrid: in an upcoming release, it will be required to pass the indexing argument. (Triggered internally at C:\cb\pytorch_1000000000000\work\aten\src\ATen\native\TensorShape.cpp:2228.) return _VF.meshgrid(tensors, **kwargs) # type: ignore[attr-defined]
一、AttributeError: Can‘t get attribute ‘SPPF‘ on <module ‘models.common‘ from ‘H:\\yolov5-5.0\\models\\
解决方案:在新版本的models/common.py里面去找到这个SPPF的类,把它拷过来到你的models/common.py里面,这样你的代码就也有这个类了,还要引入一个warnings包就行了!还要引入一个warnings包就行了!
有的同学找不到SPPF这个类,那我现在直接粘贴在这里,你们只需要复制到你们的common.py里面即可,记得把import warnings放在上面去:
import warningsclass SPPF(nn.Module): # Spatial Pyramid Pooling - Fast (SPPF) layer for YOLOv5 by Glenn Jocher def __init__(self, c1, c2, k=5): # equivalent to SPP(k=(5, 9, 13)) super().__init__() c_ = c1 // 2 # hidden channels self.cv1 = Conv(c1, c_, 1, 1) self.cv2 = Conv(c_ * 4, c2, 1, 1) self.m = nn.MaxPool2d(kernel_size=k, stride=1, padding=k // 2) def forward(self, x): x = self.cv1(x) with warnings.catch_warnings(): warnings.simplefilter('ignore') # suppress torch 1.9.0 max_pool2d() warning y1 = self.m(x) y2 = self.m(y1) return self.cv2(torch.cat([x, y1, y2, self.m(y2)], 1))
二、yolov5 ERROR: AttributeError: ‘Upsample‘ object has no attribute ‘recompute_scale_factor‘
解决方案:
找到\torch\nn\modules\upsampling.py下的文件
import torch.nn.modules.upsampling,然后摁住ctrl+鼠标左键就会跳转到该文件下,或者摁提示报错的地方也可以跳转到该文件下之后将如下图所示154行注释掉就行了
三、yolov5中The size of tensor a (80) must match the size of tensor b (56) at non-singleton dimension 3
下载:
https://github.com/ultralytics/yolov5/releases/download/v5.0/yolov5s.pt
替换默认下载的yolov5s.pt,因为默认下载的是V6.1的
替换后,在运行 detect.py就OK了
四、UserWarning: torch.meshgrid: in an upcoming release, it will be required to pass the indexing argument. (Triggered internally at C:\cb\pytorch_1000000000000\work\aten\src\ATen\native\TensorShape.cpp:2228.)
return _VF.meshgrid(tensors, **kwargs) # type: ignore[attr-defined]
找到functional.py的第568行,
将 return _VF.meshgrid(tensors, **kwargs)改为 return _VF.meshgrid(tensors, **kwargs,indexing='ij')