Download
1 variant available
This LoRA is compatible with the Z-Image-Turbo model.
Recommended weight: 0.75 - 0.8
Trigger word: wuwaams
Suggested prompt: wuwaams, a girl with long pink hair... (add your desired content)
wuwaams, dusk, a girl with long pink hair stands on the platform of an abandoned overpass, her hands gently resting on the rusty railings, the breeze blowing through her clothes and the ends of her hair. In the distance, city lights begin to flicker on as the sky transitions from orange to deep purple. She slightly tilts her head, gazing into the distance lost in thought. Heavy traffic flows beneath the overpass, emitting a low rumble. The air carries the distinct coolness of an autumn evening, with a few withered leaves drifting past her. The scene is captured with a cinematic wide-angle composition, backlit, creating a strong contrast between the figure's silhouette and the sunset afterglow. The overall color tone is warm yet tinged with melancholy, rich in detail, with exquisite texture, 8K ultra HD.
Perfect character detail restoration prompt (not recommended): A girl with incredibly eye-catching long pink hair. Her hair is styled into a high ponytail or high bun, with soft bangs hanging over her forehead and side strands framing her face on both sides. Her overall hair volume is fluffy and flows dynamically. Her eyes are large and bright, featuring a unique golden-yellow pupil. A translucent, high-tech white halo floats above her head at all times, with crystal-like decorations resembling wings or streamlined shapes on both sides. A shimmering pink four-pointed star sits at the very top center of the halo. On her chest, there is a highly distinctive heart-shaped luminous texture that blends seamlessly with her skin. She wields an incredibly imposing long spear, which appears ice-blue or translucent. The spearhead and shaft emit dazzling blue and pink starbursts and aurora-like effects during combat or when charging.
Using a full appearance prompt will make the generated images extremely close to the training images, which may prevent the model from following your prompt to generate the desired image effect (see image below).