little question about the data

Sorry, I don’t understand very well. Why are the LR and HR images the same size? If I want to use your data to train a super-resolution network, I should enter a small LR and a large HR. How can I get a small LR in the real world?