A LoRA designed to help style ******* videos with unique options for body types, clothing, and small details. Trained off of an old dataset of from a FLUX LoRa I left unfinished and later used on a Hunyuan Video lora. I didn't change any of the captions from the FLUX dataset.
This LoRA is INTENDED to give a slightly lower quality output more similar to cellphone photos in general. It also skews Z-Image away from it's default Asian tilt. Main Trigger Word: [amat_petite] (trained with the [] so they may be necessary) but honestly sometimes the output of the lora is better without the primary trigger word. See all training captions here:
PDF showing the same prompts without the lora, with the lora without a trigger word and lora with the trigger word:
What it's capable of:
******* photo styles such as cellphone video and selfie angles
Petite body types
Skinny, thin, short, tall, can get others out of it but it will skew towards petite.
Wide range of breasts
sizes but skews towards small breast sizes
Various poses
Kneeling, Standing with arms above ahead, reclined on bed/couch, running hands along body through hair, squeezing/holding breasts, hands on hips, general photography model posing, side profile
Tan lines
generally good at tan lines, lots of examples in the training set
Braces
The original dataset had a few training images with braces and honestly they work pretty well
Clothing in pulled down/up states
Not always perfect but works decently
Different lighting types
Best at natural light, should understand: "outdoor light, natural light, indoor light, overexposed
Sometimes causes weirdness with nipples but re-generating or changing the strength can help combat that.