🌿✨ Z-image Soft Depth Anime LoRA

This LoRA is trained on a Z-image base and is a style of LoRA designed for high-quality anime illustrations, emphasizing atmospheric quality.

While emphasizing the cuteness of characters,

it naturally adds soft lighting, background depth, and a dreamy, grainy atmosphere.

It elevates anime illustrations, which tend to be flat, to a 2.5D-like three-dimensionality and immersive depiction.

🎨 Compatible with a Wide Range of Illustration Styles

This LoRA is designed to blend naturally with a variety of illustration styles without leaning too strongly toward any particular style.

Anime-style (classic, light colors, soft)

Subtly illustrious coloring

Compatible with both fantasy and everyday styles

Variable from flat to 2.5D

Depending on checkpoints and prompts,

the atmosphere, density, and three-dimensionality can be flexibly adjusted,

allowing you to enjoy a wide range of expressions with just one LoRA.

🗣️ Strong support for natural language prompts

A key feature of LoRA is its ability to easily capture intent not only from tag lists but also from sentence-based natural language prompts.

It's easy to verbally describe mood, emotion, and atmosphere.

Abstract expressions like "gentle," "quiet," and "fantastic" are effective.

It's well-suited for scene descriptions and situation descriptions.

This makes it easy to use for those who want to construct images in sentences or write prompts intuitively.

⚙️ Recommended Settings (Guideline)

Recommended Weight: 0.6-0.9

CFG: 4-6

Base Model: Z-image base

Use: Single image/Atmospheric cut/Video frame

📝 Natural Language/Information Prompt Examples

①"A soft anime-style illustration of a girl in a quiet flower field,

gentle lighting, shallow depth of field,

tiny glowing particles floating in the air,

dreamy, calm, and warm atmosphere"

②"A fantasy anime illustration,

a small fairy sitting on a girl's hand,

soft glow, translucent wings,

natural lighting, flowers in the background,

peaceful and magical feeling"

Z-Anim Bloom＊Z-image base＊ - ｖ1