Brunette-136 (Nerissa)
⚠️ FICTIONAL AI-GENERATED CHARACTER — NOT A REAL PERSON
Nerissa is a fictional AI-generated character created using ChatGPT. Her approved anchor portrait and subsequent training images were developed through iterative image generation and refinement in ChatGPT, with careful attention to facial consistency, anatomy, expressions, and photographic realism.
The final images were curated into a training dataset and used to train this LoRA for Stable Diffusion PONY.
This LoRA is intended to represent a fictional character, not to impersonate or reproduce the identity of any real person.
Character Description
Nerissa is an attractive adult brunette with warm hazel-brown eyes, long dark-brown hair, an elegant oval face, defined cheekbones, expressive eyebrows, and naturally full lips.
She has a feminine, fit, naturally curvy figure with realistic proportions, authentic skin texture, and subtle freckles. Her appearance combines refined beauty with a confident, composed, and approachable presence.
Nerissa works particularly well in natural lifestyle photography, elegant glamour portraits, fashion editorials, swimwear, lingerie, and artistic compositions.
Key Features
Long, dark-brown hair
Warm hazel-brown eyes
Elegant oval facial structure
Defined cheekbones and expressive eyebrows
Naturally full lips
Light freckles and realistic skin texture
Feminine, fit, naturally curvy physique
Mature adult appearance
Versatile loose, ponytail, and updo hairstyles
Strong results across frontal, three-quarter, and profile views
Recommended Settings
Base Model: Stable Diffusion PONY / PONY-compatible checkpoints
Training Checkpoint: Realism by Stable Yogi PONY
Trigger Word: brunette 136
Recommended LoRA Weight: 0.8
Suggested Weight Range: 0.65–1.0
CFG Scale: 5–7
Sampling Steps: 25–35
Sampler: DPM++ 2M Karras
Recommended Resolution: 832 × 1248 or 1024 × 1536
Preferred Aspect Ratio: 2:3
Results may vary depending on the checkpoint, sampler, prompt structure, and other LoRAs used.
Prompting Tips
For best results, place the trigger word near the beginning of the prompt and use descriptive language that reinforces Nerissa’s established appearance while allowing variation in clothing, hairstyles, poses, and environments.
Suggested descriptors:
brunette 136, adult woman, long dark brown hair, warm hazel-brown eyes, elegant oval face, defined cheekbones, expressive eyebrows, naturally full lips, subtle freckles, natural skin texture, realistic facial proportions, feminine fit curvy figure, photorealistic
Recommended Uses
Photorealistic glamour headshots
Casual lifestyle photography
Fashion and editorial portraits
Swimsuit photography
Lingerie fashion photography
Artistic and cinematic portraits
Indoor and outdoor photography
Varied expressions, hairstyles, and poses
For maximum facial consistency, begin with chest-up or waist-up compositions before experimenting with more complex full-body poses.
Creation & Training Process
Nerissa was developed through a multi-stage AI image-generation workflow using OpenAI’s ChatGPT.
The process included:
Creating and refining a primary facial anchor.
Establishing consistent facial features and natural proportions.
Generating controlled 2×2 training sets with varied facial angles, expressions, and hairstyles.
Expanding coverage to include waist-up, thigh-up, and full-body compositions.
Developing profile views, elevated and low viewpoints, seated poses, standing poses, and natural movement.
Reviewing the generated images for facial consistency, anatomical accuracy, and training suitability.
Curating 29 selected training images from the generated sets.
Preparing and captioning the selected images for LoRA training using Kohya SS.
Comparing saved training epochs and selecting epoch 9 for the published model.
The objective was to develop a recognizable and consistent fictional character capable of producing realistic photographic results across different settings and styles.
All character training imagery was generated using AI tools rather than a traditional photographic model shoot.
Disclaimer
Nerissa is an entirely fictional AI character. This LoRA was developed using AI-generated imagery created and refined through ChatGPT.
The character is not presented as a real individual, and generated images should not be interpreted as photographs documenting an actual person. Any resemblance to an actual person, living or deceased, is coincidental.
This model is intended for creative AI image generation. Users are responsible for complying with applicable laws, platform policies, and appropriate content guidelines.