Janessa – Fictional Mixed Latina Character for WAN
Details
Download Files
About this version
Model description
⚡ Trigger Word: Janessa
Who is Janessa?
Janessa is a completely fictional AI-created character. She represents a mixed ethnicity Latina — the daughter of one white/Caucasian parent and one Puerto Rican parent. Her features reflect this blended heritage: a warm Hispanic complexion, natural blue eyes with subtle hazel-brown flecks (a trait known as central heterochromia), and multi-toned blonde/brunette hair. While her default look has this mixed-tone hair color, you can freely prompt for alternate shades.
Her body shape defaults to busty with a thin waist, wide hips, and thick thighs. These proportions can be exaggerated with prompts, but if you want to stay true to her “default” figure, avoid adding size modifiers. Janessa is also commonly depicted wearing her signature drop earrings — if you prefer her without jewelry, you’ll need to prompt for specific earrings or include “earrings” in your negative prompt.
How she was created
Janessa was built from the ground up as an original AI person:
Base dataset: Generated using Google’s Nano Banana, guided by a detailed concept of her ethnicity and features.
Extended dataset: Additional variations were created in Flux Kontext MAX, using Nano Banana outputs as references to diversify angles, expressions, outfits, and scenarios.
Image cleanup: Every dataset image was edited in Photoshop to remove aberrations and artifacts.
Captioning: Dataset was captioned in BDTM using Qwen2.5-VL-7B-Instruct with a custom QUERY template to ensure structured, descriptive captions.
Training: LoRA trained in AI-Toolkit on RunPod for 3000 steps.
Usage Notes
Works seamlessly in WAN 2.1 and WAN 2.2, for both images and videos.
Default proportions: busty figure, thin waist, wide hips, thick thighs.
To keep default proportions → avoid body-size prompts.
To change earrings → prompt specific jewelry or put “earrings” in the negative prompt.
Default hair: blonde/brunette mix, but promptable to other shades.
Eyes: blue with central heterochromia (blue + brown flecks).
Recommended settings (for WAN workflows):
High Noise Strength: 0.6
Low Noise Strength: 0.9
Flexibility
This LoRA is highly flexible — Janessa can adapt to a wide range of aesthetics, outfits, and genres. She layers well with other LoRAs and style models, making her a versatile character for portraits, editorial fashion, glamour, lifestyle, lingerie, swimwear, or even cinematic video projects.
Please share your generations!