Diva-Lara

🌟 DIVA — Portrait LoRA for Flux 2.0 Klein

If you've been searching for a LoRA that captures raw femininity, presence, and power in every frame — this is it. DIVA is a character-focused portrait model trained to generate women with genuine depth, attitude, and visual impact.

✨ Core Specializations

• 💄 High-Fashion Character Identity — DIVA understands glamour at a foundational level. From bold editorial looks to intimate close-ups, it renders subjects with a consistent, striking aesthetic.

• 👁️ Expressive, Emotive Faces — Fine-tuned on 188 curated images to deliver eyes with real depth, natural skin texture, and emotional authenticity — no plasticky doll look.

• 🎬 Cinematic & Dramatic Lighting — Built-in understanding of moody, high-contrast lighting. Keywords like cinematic lighting, rim light, or dramatic shadows translate beautifully out of the box.

• 🔧 Clean Prompting — The model is biased toward quality. Short, direct prompts work. You don't need a wall of quality tags to get professional results.

⚙️ Training Configuration

ParameterValueBase ModelFlux 2.0 Klein 9BTrigger WorddivaDataset188 curated imagesSteps1,000Epochs10Rank32 (BF16)Softwareai-toolkit v0.8.0

🎛️ Recommended Parameters (Flux 2.0 Klein)

Optimal:

• Sampler: euler — Steps: 15+ — CFG: 1

Minimum:

• Sampler: euler — Steps: 5+ — CFG: 1

LoRA weight: 0.7 – 1.0

✍️ Example Prompts

Editorial / Fashion:

diva, glamorous woman, high fashion editorial, bold red lips, 
strong confident gaze, dramatic side lighting, luxury aesthetic, 
cinematic, film grain, sharp focus

Intimate Portrait:

diva, close up portrait, natural skin texture, expressive eyes, 
soft rim light, moody atmosphere, photorealistic, 8k

Dark & Cinematic:

diva, powerful woman, dark background, dramatic chiaroscuro lighting, 
high contrast, cinematic color grading, hyperdetailed skin

⚠️ Model is experimental — feedback welcome. Use trigger word diva at the start of your prompt for best activation.