Negative Embedding: BadPony v2 & BadPonyHD
Civitai Generator: Add BadPonyv2-neg or BadPonyHD-neg to the negative prompt field, depending on the version you are using.

Major Update: Version 2

Version 2 introduces full compatibility with Illustrious, alongside the release of BadPonyHD, which delivers enhanced resolution for even greater detail and fidelity.

New Release: BadPonyHD

• Training Resolution: 2000x2000

• Designed For: Creators seeking ultra-high-resolution image generation with detail and precision.

• Usage: Works seamlessly with SDXL, Pony, and Illustrious models.

Trigger Word for Illustrious-XL

To ensure proper functionality with Illustrious-XL, use the trigger word "ugly".

• For BadPony v2, your negative prompt should be:
  BadPonyv2, ugly

• For BadPonyHD, use the same format:
  BadPonyHD, ugly

Base Model Compatibility

• Supported Models: SDXL, Pony, and Illustrious

• Training Resolutions:

  • BadPony v2: 1324x1324

  • BadPonyHD: 2000x2000

Key Features

• Enhanced Detail: Both versions produce highly detailed and refined images, ideal for high-resolution outputs.

• VAE Integration: Includes a built-in VAE for improved quality and visual fidelity.

• Performance Enhancements: Addresses common issues such as body deformities and image blurriness, delivering consistent and precise results.

• Versatile Compatibility: Optimized for seamless performance across SDXL, Pony, and Illustrious models, supporting diverse artistic styles and workflows.

Whether you're using BadPony v2 for balanced high-quality images or BadPonyHD for ultra-high-resolution results, these embeddings are powerful tools for achieving superior clarity and accuracy in your generated content.

Update 2/26/2025
BadPonyUHD 4K ~soon we are still testing this embedding. trained at 4096x4096, color correction added and shadow depth field addedBadPonyHD 2000x2000

Known Issues spotted: 1/5/2025

We have identified a few issues with the current version of the embedding:

1. Trigger Word Conflicts:

   • This embedding may overwrite or conflict with trigger words on models that require specific special trigger words like "3D," "realistic," "anime," etc.

   • As a result, the embedding might override the intended behavior of these trigger words, leading to outputs that do not correspond correctly with the input prompt.

Planned Fix: This issue will be addressed in the next update, ensuring the embedding respects the intended functionality of trigger words while maintaining its effectiveness.

Support & Feedback

We value your feedback! If you encounter any issues or have suggestions for improvement, please leave a comment. Your input is crucial for addressing concerns and implementing improvements in future updates.