💖Asian Egirl XL – Realistic uwu Egirl Tattoo Model SDXL (NSFW) v1.0💖

by Nynx_AI

Hello :3 💖

This is my first experimental release of the Asian Egirl uwu XL Checkpoint + LoRA!

Trained on ~50 carefully picked images, this version is still a v1 test run — so please be kind 💕. - i uploaded the LoRA aswell - Check it out ! :)

The model aims to capture the tattooed kawaii Asian egirl / influencer style:

✨ pastel / pink hair vibes

✨ uwu makeup & expressions

✨ sleeve + chest tattoos

✨ cute but hot outfits (lingerie, crop tops, bikinis, cosplay)

Trained on epiCRealism XL v5 Ultimate (SDXL finetune) with ~50 curated images. Focused on tattoos, uwu makeup & egirl vibe.

📌 Example Prompts

Prompt 1 (Portrait / Selfie)

asian egirl uwu, masterpiece, ultra-detailed, close-up portrait of a kawaii tattooed asian egirl influencer, pastel pink ponytail, sleeve tattoos visible, glossy lips, bright luminous almond eyes with catchlight, cute uwu makeup, playful expression, cinematic soft light, Sony A7R IV, HDR sharp detail

Prompt 2 (Full Body / Fashion)

asian egirl uwu, masterpiece, ultra-detailed, full body street photo of a kawaii tattoo model, pastel pink twin tails, sleeve and chest tattoos, glossy lips, uwu eyeliner makeup, wearing black crop top and pleated skirt, edgy influencer pose, neon-lit city street at night, cinematic HDR photography

Negative Prompt

(worst quality, low quality, normal quality:2), bad anatomy, deformed, disfigured, extra limbs, missing fingers, fused fingers, mangled limbs, cropped body parts, blurry, pixelated, text, watermark, signature, logo, ugly, (crooked eyes, shadowed eyes, dark eyes), flat face, 3d render, cartoon, anime, cgi, doll skin

📌 Recommended Settings

• Sampler: DPM++ 2M SDE Karras

• Steps: 28

• CFG Scale: 3 – 6.5 (sweet spot around 5.5)

• Resolution: 1024×1024 (or 832×1216 portrait / 1216×832 landscape)

⚠️ Notes:

- Faces are quite good, but some artifacts / limbs / hands may happen, so please use solid negative prompts.

- Portraits work best right now! Full body can sometimes need retries.

- v2 with more training data (300+ curated images) is already in the works 🚀

If you like it, please leave a ♥ or comment — feedback helps me improve future versions a lot!

Credits are always appreciated if you use the model in your work ✨

Stay tuned, more styles + concepts coming soon 👀