import torch
from diffusers import DiffusionPipeline
# switch to "mps" for apple devices
pipe = DiffusionPipeline.from_pretrained("Tongyi-MAI/Z-Image-Turbo", dtype=torch.bfloat16, device_map="cuda")
pipe.load_lora_weights("jarod2212/Aetheria_Moonlight_Shadow")
prompt = "-"
image = pipe(prompt).images[0]Aetheria_Monnlight Shadow LoRA Z-image turbo
- Prompt
- -
- Prompt
- -
- Prompt
- -
- Prompt
- -
- Prompt
- -
- Prompt
- -
Model description
Aetheria_Moonlight Shadow is a concept‑driven LoRA designed to generate poetic moonlit scenes with flowing fabrics, glowing particles, and a strong cool/warm contrast. It is stable, consistent, and works perfectly with both simple and complex prompts.
What it can do integrated ambient particles (sparks, glowing dust, firefly‑like glow)
dress color changes with no collapse
fabric changes (silk, chiffon, cotton, heavy/light materials)
stable ethnicity variation
natural fabric motion
cinematic cool/warm moonlight contrast
consistent moonlit field environments
Dataset structure The dataset is divided into thematic folders, each teaching one specific visual element:
Cool portraits → moonlight lighting
Warm portraits → warm fabrics + warm particles
Half‑body motion → fabric movement
Half‑body particles → atmospheric glow
Close‑ups (warm/cool) → texture and detail consistency
Moonlit field environments → field + moon structure
Particles only → clean, integrated particle effects
This structure allows the model to generalize the concept instead of copying images, resulting in high stability and strong visual identity.
Trigger words
You should use aethx to trigger the image generation.
Download model
Download them in the Files & versions tab.
- Downloads last month
- 99
Model tree for jarod2212/Aetheria_Moonlight_Shadow
Base model
Tongyi-MAI/Z-Image-Turbo