Input type and weight type should be the same

[Ca-ressemble-a-du-fake]

Hi,

While running inference I keep getting IMS-Toucan/TrainingInterfaces/Text_to_Spectrogram/PortaSpeech/Glow.py", line 123, in forward z = F.conv2d(x, weight) RuntimeError: Input type (torch.cuda.FloatTensor) and weight type (torch.FloatTensor) should be the same

It is solved by running on CPU but it can also be solved more elegantly by bringing the model weight to the same device than the data (adding .to(x.device) on line 123 in forward method of Glow.py) :
weight = weight.view(self.n_split, self.n_split, 1, 1).to(x.device)

Should I write a pull request or you want to correct it in a different way ?

[Flux9665]

good catch, thanks for letting me know. Feel free to make a PR, otherwise I'll fix it some time in the next few days.

[Ca-ressemble-a-du-fake]

Ok gonna do it.