Abstract: Despite its fruitful applications in remote sensing, image super-resolution (SR) is troublesome to train and deploy as it handles different resolution magnifications with separate models.