About this model
Introducing zgts-flux! No fancy keywords, since flux likes natural language.
Usage
Check the example images for example prompts. The model is still somewhat picky about details (eg the image where I said right hand, but it did left hand), but it still makes some pretty good images.
Info
All of my zgts models are:
Alpha-level quality (works, but can be tricky to control)
Baked from mostly non-nude images (99%)
Baked from non-watermarked images
Intended to be compatible with a wide range of checkpoints
Biased towards shrunken rather than mega giantess
Biased towards realism
Description
Initial release
Comments (1)
This does solve a lot of issues with shrunken people looking more like dolls than actual people. Really fun to play with this lora. One small complaint, sometimes tinies look off colorwise. Like they're pasted on the image, without adjusting their color to the rest of the image. Otherwise really good, thanks for this!




