AttributeError: ‘NoneType‘ object has no attribute ‘fill_‘

AttributeError: ‘NoneType’ object has no attribute ‘fill_’

前言：在使用pytorch搭建网络的过程中，出现AttributeError: ‘NoneType’ object has no attribute 'fill_'的错误，记录下

一、搭建网络后，开始测试网络输出

model = mobilenet_v1()
input = nn.torch.randn(1,3,224,224)
out = model(input)
print(out.shape)

二、出现以下错误

Traceback (most recent call last):
  File "mobilenet_v1.py", line 142, in <module>
    model = mobilenet_v1()
  File "mobilenet_v1.py", line 96, in __init__
    self.init_param()
  File "mobilenet_v1.py", line 104, in init_param
    nn.init.constant_(m.bias, 0)
  File "J:\WorkSoft\envs\yolov5_test\lib\site-packages\torch\nn\init.py", line 186, in constant_
    return _no_grad_fill_(tensor, val)
  File "J:\WorkSoft\envs\yolov5_test\lib\site-packages\torch\nn\init.py", line 59, in _no_grad_fill_
    return tensor.fill_(val)
AttributeError: 'NoneType' object has no attribute 'fill_'

三、在源码中找到出错的地方

 model = mobilenet_v1()  ->  self.init_param() ->  nn.init.constant_(m.bias, 0)

对应self.param() 函数如下

def init_param(self):
    for m in self.modules():
       if isinstance(m, nn.Conv2d):
           nn.init.kaiming_normal_(m.weight)
           nn.init.constant_(m.bias,0)
       elif isinstance(m, nn.Linear) or isinstance(m, nn.BatchNorm2d):
           nn.init.constant_(m.weight,1)
           nn.init.constant_(m.bias, 0)
      

在Conv2d、Linear 和 BacthNorm2d 中设置了bias = True 才能赋值，否则需要注释bias 的赋值。