A mid-century storybook illustration style LoRA trained on a custom 120-image dataset. Produces portraits and full scenes with oversized round doe eyes, thick black eyeliner, long dramatic lashes, flat color blocking, thick black outlines, and visible canvas grain/brush texture — a retro Keane-inspired painterly look, not a flat digital/anime style.

Trigger word: whimsicalartstyle

Good for: character portraits, fashion-plate style couples, full-body figures, whimsical scenes with simple props (jewelry, florals, city or nature backdrops).

Known limitations: the model can occasionally render a heart-shaped mark or symbol unprompted (a training data artifact) and can struggle with animal companions in frame, sometimes fusing human and animal features into a hybrid rather than rendering them as separate figures. Suggest using the negative prompt below, especially for scenes involving animals.

Suggested negative prompt:
smooth anime shading, large almond eyes, glossy digital rendering, airbrushed gradients, photorealistic skin, oversaturated, plastic, blurry, low quality, distorted, watermark, text, heavy eyelashes, defined eyelids, heart, heart shape, heart symbol, human face on animal, second face, extra head, anthropomorphic animal, human-animal hybrid, merged figures