✨ Shaevra Woman – Realistic Nude Female Character | SDXL LoRA

Shaevra Woman is a realistic SDXL character LoRA built for photoreal portraits, upper-body renders, and full-body nude generations. Designed around a blonde female character with blue eyes, natural skin, and a realistic overall look, it performs well in both clean studio-style prompts and more scenic natural settings.

This LoRA is focused on photoreal consistency, natural skin, and strong results across both simple prompts and more detailed scene setups.

🧾 Training Info

Trained on 39 curated images using SDXL base 1.0 at 1024 resolution with a captioned realistic-character workflow focused on photoreal consistency, natural skin, and strong overall results.

The final run used:

• 12 repeats

• 12 epochs

• 5616 total steps

🎯 Best Use Cases

• Photoreal portraits

• Upper-body nude renders

• Full-body nude renders

• Realistic skin and anatomy

• Studio and natural-light scenes

• Scenic outdoor / naturist-style prompts

⚙️ Recommended Settings

• Base model: SDXL 1.0 / realistic SDXL checkpoints

• LoRA weight: 0.7–0.9

• Sweet spot: 0.8

• Sampler: DPM++ 2M Karras

• Steps: 25–35

• CFG: 4.5–6

• Best resolution: 1024x1024 or 1024x1280

🔑 Trigger Word

shaevra woman

🧠 Prompting Tips

This LoRA works best with realistic SDXL checkpoints and simple, direct prompting.

For a more neutral base, use:

• stable-diffusion-xl-base-1.0

For more polished photoreal results, it also works well with:

• RealVisXL V5.0

• Juggernaut XL

Prompting advice

• Keep the trigger in the prompt

• Use simple SDXL-style prompting

• For nude or full-body generations, include relevant anatomy terms when needed

• For face-focused generations, keep the prompt cleaner and avoid unnecessary body tags

Useful anatomy terms when relevant:

• nude

• breasts

• nipples

• pussy

• ass

• pubic hair

• feet

• soles

🏷️ Most Frequent Tags in the Training Captions

shaevra woman, nude, tattoo, blonde hair, full body, barefoot, breasts, feet, nipples, blue eyes

🖼️ Example Prompt

Prompt
shaevra woman, close-up portrait, looking directly at viewer, blonde hair, blue eyes, natural skin, soft lips, subtle expression, soft natural light, warm tones, elegant softly blurred background, photo

Negative
deformed face, asymmetrical eyes, bad anatomy, bad hands, bad feet, extra fingers, extra toes, blurry

Example settings
SDXL base 1.0, LoRA weight 0.8, DPM++ 2M Karras, 36 steps, CFG 5, 1024x1024

✅ Works Well With

• stable-diffusion-xl-base-1.0

• RealVisXL V5.0

• Juggernaut XL

📌 Notes

• Extreme angles may reduce consistency

• Tattoo visibility may vary depending on prompt, framing, and checkpoint

• Realistic SDXL checkpoints generally give the best results

• Simple prompting usually works better than overly cluttered prompt soup

💬 Feedback Welcome

If you try this LoRA, please post your generations and settings.

Feedback on the following is especially helpful for future improvements:

• realism

• image quality

• prompt responsiveness

• anatomy consistency

• tattoo handling

• flexibility

• overall results

I’d love to see what you create with it — please share your generations and let me know what looks strongest, what feels most successful, and what could be improved in future versions.