A strange bug encountered during training was the first propagation of the previous item in Variance_Attention.

[Wei-JL]

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)

[Damonjzh]

I also meet this bug, did you solve it? Thanks!

[cccly056]

You need to check if any arguments in conv2 are set to False, when they should be 0