pytorch 0.2 version cuda8.0
the code:
inputs, targets = inputs.to(device), targets.to(device)
show error:
AttributeError: ‘torch.FloatTensor’ object has no attribute ‘to’
How do I modify the code?
pytorch 0.2 version cuda8.0
the code:
inputs, targets = inputs.to(device), targets.to(device)
show error:
AttributeError: ‘torch.FloatTensor’ object has no attribute ‘to’
How do I modify the code?
PyTorch 0.2
is really old by now and the .to()
operation was introduced in newer version. I would recommend to update to the latest version. You’ll find the install instructions here.
If that’s not possible for whatever reason, you could use .cuda()
and .cpu()
instead.