Alin is a photorealistic female character LoRA trained for Z-Image Turbo.
The LoRA is designed to preserve Alin's facial identity across a variety of facial expressions, camera angles, lighting conditions, outfits, and backgrounds.
Trigger word:
Alin
Recommended uses:
Photorealistic portraits
Facial expression generation
Close-up and head-and-shoulders photography
Lifestyle photography
Different camera angles and lighting conditions
The training dataset contains multiple facial expressions and variations intended to improve character consistency while retaining flexibility in prompts.
Recommended prompt example:
Alin, woman, photorealistic portrait, natural skin texture, soft natural lighting, close-up portraitFor stronger expression control, add expression descriptions such as:
happy, playful, angry, terrified, bored, determined, neutral, etc.This model is intended for fictional character generation and does not depict a real person.
Description
Continued training from the final checkpoint of v1.0 (epoch 16).
This version extends the previous training to further improve:
Alin facial identity consistency
Expression accuracy
Facial details across different camera angles
Generalization across lighting, clothing, and backgrounds
Trigger word: Alin
This is a continued-training version, not a training run from scratch.