NepotismXL: The Next Evolution in SDXL/Pony Modeling

NepotismXL is the latest iteration of the Nepotism™️ series, incorporating valuable user feedback to refine and enhance its capabilities. Building on the foundation of SupremeNepotism, NepotismXL features an improved VAE and superior performance in detail, prompt adherence, and artistic range.

INNATE CHARACTER LIST: CLICK HERE

Model Capabilities:

• Enhanced Detail and Quality: NepotismXL generates highly detailed images with improved visual quality, thanks to its refined VAE.

• Improved Prompt Adherence: The model demonstrates excellent adherence to user prompts, ensuring accurate and consistent outputs. (and yes it can do text without LoRA's just add text:"the text you want" followed by where you want it eg a sign)

• Wide Artistic Range: From hyper-realism and photography to anime and 3D models, NepotismXL handles various artistic styles effortlessly.

• LoRA Compatibility: While the model works exceptionally well on its own, it remains fully compatible with LoRAs for further customization and refinement.

Recommended Settings:

For optimal results, it is advised to use the following settings:

• Clip Skip: -2

• Steps: 30-75

• CFG (Classifier-Free Guidance): 4.0-7.0

• Sampler: dpmpp_2m_sde_gpu

• Scheduler: exponential

Using Simple Prompts:

NepotismXL excels with simple prompts. You don't need complex or detailed descriptions to achieve great results. Simple prompts like "fantasy landscape" or "anime character" can produce impressive images.

Example Usage:

• Simple Prompt: "Fantasy landscape"

• Detailed Prompt: "Hyper-realistic fantasy landscape with vibrant colors and intricate details, small birds fly overhead, and a tree is embedded with a cheerful face on the left"

• Art Style Prompt: "Anime, digital art, portrait of a character with expressive features and dynamic pose, red eye, yellow eye, heterochromia, holding a sign that reads "hello" text"

CLIP Vector Sculpting:

If you use CLIP vector sculpting, add the following settings for optimal results:

• Sculptor Intensity: 0.3-0.5 (Recommended: 0.35)

• Sculptor Method: Forward

• Token Normalization: mean

Feedback and Improvement:

Your feedback is crucial in refining Nepotism™️. Please share your experiences and suggestions to help us continue improving this model.

SupremeNepotism: Superior SDXL+Pony Model

Overview:

SupremeNepotism combines the best aspects of FurryNepotism and HentaiNepotism. This model sets a new standard in detail, prompt adherence, and artistic range, making it a versatile tool for creating high-quality images across various styles and themes.

Model Capabilities:

- Enhanced Detail: SupremeNepotism excels in generating highly detailed images, capturing intricate elements with precision and clarity.

- Improved Prompt Adherence: The model shows superior adherence to prompts, ensuring that the generated images closely match the user's specifications.

- Wide Artistic Range: From hyper-realism and photography to anime and 3D models, SupremeNepotism handles a broad spectrum of artistic styles with ease.

- LoRA Compatibility: While the model is fully compatible with LoRAs, all testing was conducted without any LoRAs, demonstrating its robust performance and versatility.

Recommended Settings:

For optimal results, it is advised to use the following settings:

- Clip Skip: -2

- Steps: 30-60

- CFG (Classifier-Free Guidance): 4.0-9.0

- Sampler: dpmpp_3m_sde_gpu

- Scheduler: exponential

These settings have been tested to ensure that SupremeNepotism works effectively even with shorter prompts and can generate text. Use these settings to maximize the model's performance and achieve the best possible outputs.

!!NOTE!! if you use CLIP vector sculpting, add the following

- sculptor intensity: 0.3-0.5 (I use 0.35)

- sculptor method: backward

- token normalization: mean

Furry Nepotism: A Comprehensive SDXL/Pony Model

Overview: Furry Nepotism is an advanced model merge, skillfully integrating multiple pony models into the XL framework. The primary objective behind this creation was to develop a versatile model proficient in utilizing pony tags, recognizing a wide array of pop culture characters and references, and excelling across various artistic styles, including hyper-realism, photography, anime, and 3D modeling.

Model Capabilities:

• Multifaceted Art Styles: Furry Nepotism adeptly handles diverse artistic styles, enabling users to generate hyper-realistic images, detailed photography, vibrant anime scenes, and intricate 3D models.

• Extensive Tag Recognition: Leveraging an extensive database of pony tags, the model ensures precise and contextually accurate image generation.

• Pop Culture Savvy: Equipped with knowledge of numerous pop culture characters and references, Furry Nepotism enhances creativity and relevance in generated content.

Compatibility:

• Pony and SDXL LoRAs: The model seamlessly integrates with both Pony and SDXL LoRAs, providing users with a wide range of customization options and enhancing the versatility of the generated outputs.

Usage Instructions:

1. (optional) Starting Point - Pony Tags: Begin by utilizing pony tags to guide the initial image generation process. This ensures the model accurately captures the desired context and theme.

   • Common Pony Tags:

     • Quality Modifiers: score_9, score_8_up, score_7_up, score_6_up

     • Sources and Ratings: source_pony, source_furry, source_cartoon, source_anime, rating_safe, rating_questionable, rating_explicit

     • Character and Setting Tags: 1girl, 1boy, 2girls, anthro, feral, monochrome, 3d, photo, hyperrealistic, realistic, rough sketch

2. Customization - LoRAs: Enhance and customize your creations by incorporating compatible pony or SDXL LoRAs. This allows for fine-tuning and personalizing the output to meet specific requirements.

3. Art Style Selection: Choose the preferred art style (hyper-realism, photography, anime, or 3D) to tailor the generated images to your creative vision.

4. NSFW Content: While the model is designed to produce high-quality images across various genres, explicit NSFW content can be generated if appropriately prompted. However, for hyper-explicit content, it is recommended to use Furry Nepotism as a refining tool to achieve the desired level of detail and accuracy.

Optimization: To maintain a balance between essential aspects and potential NSFW content, many hentai/porn poses have been removed. This optimization ensures the model focuses on delivering high-quality, versatile outputs suitable for a broad range of creative applications.