Purpose

This model will generate a photo of a woman ??? ???????. It was trained using the same training set as the Chroma HD Lora.

Where Chroma tended towards overly ?????????????, Z-Image didn't, however through testing I noticed it really only had one body size for ???????. This training set did not fix that.

Prompting

This model was trained on descriptive prompts generated by JoyCaption Beta 1, with manual edits for accuracy and reducing wordiness. So descriptive prompts are best.

Trigger words

titfucking This is the main trigger.

closeup Focuses on the woman's face and breasts

Woman's face is not visible Combined with closeup zooms in on the woman's breasts.

There main difference in trigger words between Z-Image and the Chroma version is in variety of faces. While it includes small ???????, medium-sized ???????, large breasts, thin and skinny, in the prompts, it doesn't have as much affect on ?????? size. It seem to create more variety of faces especially when combined with thin or skinny

Strength

The LoRA is designed to be used at 1.0 strength. I haven't tested it with other strengths or other LoRAs due to not many yet.

I'll post examples later once I test more.

Notes

When I chose the appropriate model for output, I also included samples of ???? woman to make sure the ??????? came out correct. This had the accidental result of creating better ??????? than the base model. even if not prompting for the titfucking.