Ghost Mix Guide
Ghost Mix artfully incorporates the aesthetic of vintage animes, paying homage to its inspiration, Ghost in the Shell. By seamlessly integrating visually stunning magic and lighting effects and uniquely incorporating fractals into its image creation, Ghost Mix sets itself apart. Notably, even in non-fractal prompts, the model defaults to utilizing fractals during image creation, resulting in a cleaner and more stunning image. Below is a deeper look at Ghost Mix.
This Model Excels At
- Creating retro anime style images
- Creating women
- Creating magical effects
- Rendering fractal art in a unique style
What To Be Aware of About This Model
- Occasionally swaps male designs to female
- Maintains a stylistic approach and doesn't break into photorealistic
- Capable of semi-realism with proper prompting
- Encounters difficulties in generating 3D renders.
- Does not support style syntax.
- Generates multiple subjects and POV shots effectively when prompts follow the guidance in Prompting 101.
Prompting Style
Ghost Mix adheres to BOM and Furry Model's prompt style, necessitating "masterpiece" and "best image" in positives, and "worst quality" and "low quality" in negatives.
Recommended Settings
- Steps: 20-30
- CFG Scale: 5-7
- Sampler: DPM++ 2M Karras
Prompt Component Order
- Style/Genre
- Quality Terms
- Resolution Terms
- Subject Statement
- Major Details
- Minor Details
Example of Recommended Prompt Structure
masterpiece, best quality, high detail, 1woman, realistic, soft light, soft shadows, soft backlighting
Notable Keywords
fractal art: This phrase will distort the image by design. Precision weighting is a must.aestheticandbeautiful: Results in a more cohesive and visually appealing look for fractal art.haphazard: As a negative prompt, results in similar pixels being grouped together to provide a more cohesive composition.
Sample Prompts
Nurse Joy
Positive Prompt :
flat colors, (masterpiece, best quality:1.5), (high detail:1.5),(closeup portrait:1.4),(anime screencap:1.5),adult nurse joy,joyful, hospital background,(best line quality:1.5), official art,flat light,soft focus, (chiaroscuro:1.3), (absurdres:1.5)Negative Prompt :
jpeg artifact,(red eyes:1.3), bad anatomy,zoom layer, (traditional media:1.5),(lossless-lossy, poorly drawn:1.5), (worst quality:2), lowres, anatomical nonsense, (low quality:1.5), artist name, monochrome, grayscale
One Character Cosplaying Another
Positive Prompt :
flat colors, (masterpiece, best quality:1.5), (high detail:1.5),(closeup portrait:1.4),(anime screencap:1.5),(adult tifa lionhart:1.5), (mmorpg:1.5),cosplay,(borrowed clothes:1.5), (millia rage cosplay as tifa lionhart:1.5),(prehensile hair:1.2),(impossible white top:1.3), red gloves, joyful,(best line quality:1.5),official art,flat light,soft focus, (chiaroscuro:1.3), (absurdres:1.5)Negative Prompt :
animal ears,(hair ears:1.5),jpeg artifact,(red eyes:1.3), bad anatomy,zoom layer, (traditional media:1.5),(lossless lossy, poorly drawn:1.5), (worst quality:2), lowres, anatomical nonsense, (low quality:1.5), artist name, monochrome, grayscale
Mecha Battle Wing
Positive Prompt :
(masterpiece, best quality:1.5),(ultra detail:1.5),1male man, (battle mecha fractal art:1.2),portrait shot ,(universe:1.4), glowing, armor, glowing eyes, transparent mecha, large ethereal wings, soft shadows, soft light, backlighting, closeup photo, photorealistic, (chiaroscuro:1.3), best robot lineart, official art,(absurdres:1.5)Negative Prompt :
(haphazard:1.5),peg artifact,(red eyes:1.3), bad anatomy,zoom layer, (traditional media:1.5),(lossless lossy, poorly drawn:1.5), (worst quality:2), lowres, anatomical nonsense, (low quality:1.5), artist name, monochrome, grayscale
Prompt Style Comparison
Below you can see the difference in generation quality between applying only the general prompt style in Prompting 101 and applying the recommended style specific to this model.
| Result | Prompt |
|---|---|
 | closeup portrait shot, a woman standing in the forest, masterpiece, extremely detailed, soft light, soft shadows, soft backlighting, (best image, best quality:1.5) |
Final Thoughts
Ghost Mix excels at crafting mesmerizing anime visuals, imbuing them with otherworldly enchantments, evoking a sense of nostalgia through vintage anime styles, and seamlessly incorporating the convenience of danbooru tags.
While it shares a common limitation with the Hassaku Model, focusing primarily on female subjects, Ghost Mix distinguishes itself through its unique approach to fractal and image generation, earning its status as one of the premier models on sexy.ai. Surpassing Hassaku in terms of realistic rendering capabilities, Ghost Mix boasts one of the broadest spectrums for anime image creation.
The simplicity of using danbooru tags is akin to YiffieMix, although it may not consistently recognize all tags. Nevertheless, it provides a user-friendly means to bring your envisioned anime images to life.