PyTorch: while loading batched data using Dataloader, how to transfer the data to GPU automatically

Question

If we use a combination of the Dataset and Dataloader classes (as shown below), I have to explicitly load the data onto the GPU using