Abstract: Current image compression models often require separate models for each quality level, making them resource-intensive in terms of both training and storage. To address these limitations, we ...