yolov5--Can't get attribute 'SPPF' on

测试yolov5时报错 ：

Can't get attribute 'SPPF' on <module 'models.common

解决方法:

到tags的版本V6.0的models 目录下的common.py目录复制类SPPF到报错的类文件中

同时倒入头文件 warnings

如下： 

import warnings

class 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))

还报错：

The size of tensor a (80) must match the size of tensor b (56) at non-singleton dimension

下载对应的v5.0版本的.pt文件，地址如下：

https://github.com/ultralytics/yolov5/releases/download/v5.0/yolov5s.pt

因为默认下载的不是v5.0的，所以要重新下载.pt文件。

再在运行 detect.py

Ok