Model descriptionThis LoRA represents the character NINA, trained for FLUX models to generate a consistent identity across different scenes.What it does / what it’s forIt allows you to generate NINA in varied outfits, poses, and environments using natural language prompts, while keeping her visual identity stable.What it’s good at<ul><li>Consistent character identity</li><li>Responding well to detailed prompts (clothing, pose, camera)</li><li>Working across different compositions (close-up to full body)</li></ul>Limitations<ul><li>May reflect dataset biases (poses, lighting, style)</li><li>Less reliable with extreme angles or highly stylized prompts</li><li>Identity can weaken with conflicting or vague prompts</li></ul>Usage<ul><li>Use the trigger word NINA</li><li>Prompt in natural language (FLUX-style)</li><li>Explicitly describe non-identity elements (clothes, pose, setting)</li></ul>Not recommended<ul><li>Vague or contradictory prompts</li><li>Styles or scenarios not represented in training</li></ul>

Files with SHA256: 40fd1ede30abbde811109f350f5599710d8f72bc3376384807780ba37d308af0