~/miniconda3/envs/deep_mol/lib/python3.6/site-packages/torch/nn/modules/conv.py in forward(self, input)
318 def forward(self, input):
319 return F.conv2d(input, self.weight, self.bias, self.stride,
--> 320 self.padding, self.dilation, self.groups)
321
322
RuntimeError: Input type (torch.cuda.FloatTensor) and weight type (torch.FloatTensor) should be the same
Hi, I got this error message above when I try to use GPU, but I have already put my network to cuda.
So there is something wrong with the conv2d
?
Here is my network, I think there might be some bugs in my model
class Model(nn.Module):
def __init__(self, in_channel, num_channel, out_class):
super(QuickNAT, self).__init__()
self.in_channel = in_channel
self.num_channel = num_channel
self.out_class = out_class
def forward(self,x):
encoder = Encoder(self.in_channel, self.num_channel)
x5, x4, x3, x2, x1, idx1, idx2, idx3, idx4 = encoder(x)
decoder = Decoder(self.num_channel, self.out_class, idx1, idx2, idx3, idx4, x4,x3,x2,x1)
x = decoder(x5)
return x