Trigger Word

None required.
This LoRA is fully trigger-free and activates naturally through standard descriptive language (e.g., pig ears, short snout, tail growing through clothing, mid-transformation, etc.).

What This LoRA Does

This model enhances fantasy transformation themes, especially scenes involving a human gradually taking on pig-like features. It performs well across:

• Facial transformation stages (pig ears, forming snout while keeping human likeness)

• Clothing interaction (e.g., a small tail emerging through the back of a swimsuit or outfit)

• Limb transformation (early to mid-stage changes in hands, legs, and feet)

• Natural emotional reactions (confusion, surprise, curiosity)

• Indoor home/bedroom settings with soft lighting

• Full-body or portrait compositions

Base Model

Z-Image Turbo
(All sample images were generated on the base model alone.)

Recommended Weight

• 0.8–1.0 → Best detail and clearest transformation features

• 0.7 → Ideal when blending with character LoRAs while keeping this model fully active

• Still maintains stability even when combined with style LoRAs or lighting LoRAs

This LoRA plays very nicely with others , it doesn’t overpower or wash them out.

Generation Info for All Sample Images

• Model: Z-Image Turbo

• Sampler: Euler (or res_2s bong tengent)

• Steps: 9

• CFG: 1.0

• LoRA Strength: 0.8 - 1

Recommended Settings

• Base Model: Z-Image Turbo

• Sampler: Euler / res_2s bong tengent

• Steps: 9

• CFG: 1.0

• LoRA Strength:

  • 0.8–1.0 for strongest transformation clarity

  • when mixing with character or style LoRAs change make the character LoRA slightly weaker.

Known Limitations

• Feet / lower-leg transformation can be inconsistent.
  In some scenes the mid-transformation detail is strong, while in others the shaping may appear softer or less defined. This is simply due to dataset balance and will be improved in future updates.

• Refinement planned:
  This LoRA will be retrained once the full Z-Image model is released, allowing for more precise anatomical transitions.

• Upcoming project:
  I’m also planning a combined Transformation Pack that will introduce multiple creature- and fantasy-based transformation types trained together. This will provide stronger consistency across limbs, posture changes, and full-body sequences.

☕ Support

If you’re enjoying my LoRAs or want to request custom transformation concepts, consider supporting my work: → [Buy Me a Coffee]
Every donation helps fund new datasets, experiments, and model releases.

You can also help support future training runs through my RunPod referral link
we both receive credits toward GPU time → [RunPod]

Thank you for your support ❤️
More models and and updates are on the way!g