ValueError: in_channels must be divisible by groups Additional No response @zai1318hello, I see that you are encountering aValueErrorstating that 'in_channels must be divisible by groups'. This error occurs when the number of input channels (in_channels) to the convolutional layers in your mode...
Hi, I want to train the deeplabv3 based on the ResNext model. But when I try to run the script, it informs that the errors like: I check the code find groups=64 by default. could you help me? Thanks ### Besides, I am also wondering one small problem. It seems the performance...
cin,cout=4,5groups=2x=torch.randn(1,cin,24,24)weight=torch.randn(cout,cin//groups,3,3)# Unclear error messageF.conv2d(x,weight,groups=groups)>RuntimeError:std::exception# Proper error messagenn.Conv2d(cin,cout,kernel_size=3,groups=groups)>ValueError:out_channelsmustbedivisiblebygroups ...