Pytorch Summary - sensor on GPU

Method of use to() Can be Tensor stay CPU and GPU（ Need hardware ⽀ Support ） Move with each other .

if torch.cuda.is_available(): # return True or False # When GPU Available as True, And vice versa False
device = torch.device("cuda") # GPU
y = torch.ones_like(x, device=device) # Create directly ⼀ It's in GPU Upper Tensor
x = x.to(device) # Equivalent to .to("cuda")
z = x + y
print(z)
print(z.to("cpu", torch.double)) # to() You can also 更 Change data type