This LoRA model is trained on a curated dataset of high-resolution images featuring European marble and porcelain sculptures from the Baroque, Rococo, and Neoclassical periods. The model captures the refined anatomy, theatrical drapery, and emotive gestures of mythological and allegorical subjects, rendered in matte marble and glossy ceramic finishes.

It excels at generating:

• Full-body classical nudes (male and female)

• Mythological scenes (e.g. Venus, Bacchus, Diana, Cupid)

• Allegorical poses (e.g. Virtue defeating Vice, The Three Graces)

• Material-specific realism — soft matte marble, glazed porcelain sheen, and detailed sculptural folds

The training set includes multiple viewing angles and compositions: single figures, intertwined groups, dynamic gestures, seated and reclining poses, with emphasis on anatomical fidelity and historical styling.