Base Model: Z-Image-Turbo
Trigger Word: indian-girl-brown-eyes
This version marks a massive dataset expansion, moving to a curated collection of 97 high-fidelity images. By training for 3,120 steps across 10 epochs, the model has achieved a superior balance between strict likeness and stylistic flexibility.
Creative Capabilities
True Versatility: From traditional cinematic portraits to modern corporate or "street-style" photography, the 97-image foundation handles it all.
Dynamic Lighting: Whether it’s harsh studio lighting, neon-lit cityscapes, or soft "golden hour" sun, the model adapts naturally to the environment.
Consistent Identity: Even in complex scenes with multiple objects or text elements, the subject remains unmistakably consistent.
Prompting Guide (Z-Turbo Optimized)
This distilled model thrives on positive description. Since it's a Turbo variant, keep your prompts descriptive but concise.
Recommended Settings
LoRA Weight: 0.8 - 0.9 (Ideal for maintaining likeness while allowing style flexibility).
Sampling Steps: 8 - 14
Sampler: DPM++ SDE or Euler
CFG Scale: 1.0 - 2.5 (Keep it low for the cleanest results).
Clip Skip: 1
Sample Prompts
The Cinematic Traditional: indian-girl-brown-eyes, standing in a grand palace hallway, wearing an intricate lehenga with gold embroidery, soft candle lighting, cinematic shadows, 35mm lens, highly detailed.
Modern Lifestyle: indian-girl-brown-eyes, urban street style, wearing a chic blazer and glasses, walking through a modern tech park at sunset, blurred architectural background, realistic skin texture, 8k.
Soft Portrait: indian-girl-brown-eyes, extreme close up, wearing a simple cotton kurta, soft natural morning light from a window, focus on deep brown eyes and natural skin, bokeh, minimalist aesthetic.
