GSConv+VoVGSCSP(Slimneck报错修正)

参考有温度的AI提供的Slimneck代码进行的报错修改：

链接：YOLOv5改进之YOLOv5+GSConv+Slim Neck_vovgscsp-CSDN博客

报错：

TypeError: conv2d() received an invalid combination of arguments - got (Tensor, Parameter, NoneType, tuple, tuple, tuple, int), but expected one of:
 * (Tensor input, Tensor weight, Tensor bias, tuple of ints stride, tuple of ints padding, tuple of ints dilation, int groups)
      didn't match because some of the arguments have invalid types: (Tensor, Parameter, NoneType, tuple, tuple, tuple, int)
 * (Tensor input, Tensor weight, Tensor bias, tuple of ints stride, str padding, tuple of ints dilation, int groups)
      didn't match because some of the arguments have invalid types: (Tensor, Parameter, NoneType, tuple, tuple, tuple, int)

修正：

将GSConv部分的代码替换如下，其他保持不变。

class GSConv(nn.Module):
    def __init__(self, c1, c2, k=1, s=1, g=1, act=True):
        super().__init__()
        c_ = c2 // 2
        self.cv1 = Conv(c1, c_, k, s, None, g, 1,act)
        self.cv2 = Conv(c_, c_, 5, 1, None, c_,1, act)

    def forward(self, x):
        x1 = self.cv1(x)
        x2 = torch.cat((x1, self.cv2(x1)), 1)
        # shuffle
        b, n, h, w = x2.data.size()
        b_n = b * n // 2
        y = x2.reshape(b_n, 2, h * w)
        y = y.permute(1, 0, 2)
        y = y.reshape(2, -1, n // 2, h, w)
        return torch.cat((y[0], y[1]), 1)