This is a photorealistic character LoRA of Lulu02, trained on the Z-Image base model. It perfectly captures her unique pink-and-purple ombré hair and facial features while remaining highly flexible for different outfits, styles, and environments.Trigger word: <code>Lulu02</code> Recommended weight: <code>0.8 - 1.0</code>

Lulu01

Initial release of Lulu02 character LoRA. Trained on Z-Image base model using 1024x1024 resolution. Excellent facial consistency and flexibility for changing outfits, environments, and poses.

Lulu02_epoch_6.safetensors

Lulu02_epoch_7.safetensors

Lulu02_epoch_8.safetensors

Lulu02_epoch_4.safetensors

lulu02_zimage_lora.safetensors

### Meet Lulu – High-Quality SDXL Character LoRALulu is a versatile, photorealistic character model optimized for SDXL. She features a highly distinctive look centered around her vibrant pink-and-purple ombré wavy hair and captivating green-grey eyes. This model was meticulously trained using a cleaned-up dataset with precise captioning. Because all facial and hair features were intentionally left out of the text tokens during training, her identity is incredibly stable, while her clothing, poses, and backgrounds remain 100% flexible. She transitions effortlessly from high-fashion studio shoots to gritty cyberpunk streets, fitness gear, or seasonal outfits.---### 🚀 How to use &amp; Generation TipsTo get the absolute best, ultra-realistic photographic results, please follow these recommendations:* Trigger Word: <code>lulu</code>* LoRA Weight: <code>0.8 - 1.0</code> (Start at <code>1.0</code> for maximum likeness, dial down to <code>0.8</code> if you want more base model flexibility)* Recommended Base Models: For true photorealism, use specialized checkpoints like RealVisXL, Juggernaut XL, AlbedoBase XL, or CyberRealistic XL.* Sampler: <code>DPM++ SDE Karras</code> (Essential for skin texture development)* Steps / CFG: <code>30-35 steps</code> / <code>CFG Scale 5.5 - 6.5</code>* Must-Have Tool: Turn on ADetailer (FaceDetailer) for medium or full-body shots to restore micro-pores, eyelashes, and eye reflections.---### 💡 Example PromptPositive: <code>lulu, 1girl, solo, close up portrait, raw skin texture, detailed eyes, looking at camera, soft studio lighting, sharp focus, highly detailed, photorealistic</code>Negative:<code>3d render, plastic skin, CGI, illustration, drawing, anime, deformed, bad eyes, blurry</code>

Lulu01.1

This is a photorealistic character LoRA of Lulu02, trained on the Z-Image base model. It perfectly captures her unique pink-and-purple ombré hair and facial features while remaining highly flexible for different outfits, styles, and environments.

Files with SHA256: b47463180e08ffd35f803a580b58f28dcad8efe10a1b17c7e23c95745e265151