GenghisHan0911/hari-zit-lora

Hari

Prompt

-

Prompt

-

Prompt

-

Prompt

-

Prompt

-

Prompt

-

Model description

📌 Model Description

This is a LoRA (Low-Rank Adaptation) model fine-tuned on the Z-Image Turbo (ZIT) architecture. It is exclusively designed to generate consistent facial features, styling, and identity for the virtual AI influencer, "Hari (하리)".

• Base Model: Z-Image Turbo (ZIT) - Single-Stream Diffusion Transformer
• Task: Text-to-Image Generation (Character Consistency)
• Resolution: Highly optimized for 1024x1024

⚙️ Usage Guidelines

[WARNING] Since this LoRA is based on the ZIT architecture, it is NOT compatible with SDXL, SD 1.5, or SD 2.1. Please ensure you are using a ZIT-compatible pipeline (e.g., ComfyUI with ZIT nodes).

• Trigger Word(s): `[hari]`
• Recommended Weight: `[1.0]` (Adjust based on desired likeness and prompt flexibility. Higher weights may over-saturate the image.)

Trigger words

You should use hari to trigger the image generation.

Download model

Download them in the Files & versions tab.