This model is trained on a dataset of 1000 thematic images, to provide a consistent generalized style without overfitting.
