I am trying to run PyTorch with a cpu. I instantiate the device with
device = torch.device( "cuda" if torch.cuda.is_available() else "cpu" , index = 0)
and the output of print(device) is device: cpu:0.
But when I try to send tensors to my device with .to(device), I get the error
RuntimeError: Device index must not be negative
Is there another way to manually set the index of a device?